Show More
@@ -1,1315 +1,1317 b'' | |||||
1 | # sqlitestore.py - Storage backend that uses SQLite |
|
1 | # sqlitestore.py - Storage backend that uses SQLite | |
2 | # |
|
2 | # | |
3 | # Copyright 2018 Gregory Szorc <gregory.szorc@gmail.com> |
|
3 | # Copyright 2018 Gregory Szorc <gregory.szorc@gmail.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | """store repository data in SQLite (EXPERIMENTAL) |
|
8 | """store repository data in SQLite (EXPERIMENTAL) | |
9 |
|
9 | |||
10 | The sqlitestore extension enables the storage of repository data in SQLite. |
|
10 | The sqlitestore extension enables the storage of repository data in SQLite. | |
11 |
|
11 | |||
12 | This extension is HIGHLY EXPERIMENTAL. There are NO BACKWARDS COMPATIBILITY |
|
12 | This extension is HIGHLY EXPERIMENTAL. There are NO BACKWARDS COMPATIBILITY | |
13 | GUARANTEES. This means that repositories created with this extension may |
|
13 | GUARANTEES. This means that repositories created with this extension may | |
14 | only be usable with the exact version of this extension/Mercurial that was |
|
14 | only be usable with the exact version of this extension/Mercurial that was | |
15 | used. The extension attempts to enforce this in order to prevent repository |
|
15 | used. The extension attempts to enforce this in order to prevent repository | |
16 | corruption. |
|
16 | corruption. | |
17 |
|
17 | |||
18 | In addition, several features are not yet supported or have known bugs: |
|
18 | In addition, several features are not yet supported or have known bugs: | |
19 |
|
19 | |||
20 | * Only some data is stored in SQLite. Changeset, manifest, and other repository |
|
20 | * Only some data is stored in SQLite. Changeset, manifest, and other repository | |
21 | data is not yet stored in SQLite. |
|
21 | data is not yet stored in SQLite. | |
22 | * Transactions are not robust. If the process is aborted at the right time |
|
22 | * Transactions are not robust. If the process is aborted at the right time | |
23 | during transaction close/rollback, the repository could be in an inconsistent |
|
23 | during transaction close/rollback, the repository could be in an inconsistent | |
24 | state. This problem will diminish once all repository data is tracked by |
|
24 | state. This problem will diminish once all repository data is tracked by | |
25 | SQLite. |
|
25 | SQLite. | |
26 | * Bundle repositories do not work (the ability to use e.g. |
|
26 | * Bundle repositories do not work (the ability to use e.g. | |
27 | `hg -R <bundle-file> log` to automatically overlay a bundle on top of the |
|
27 | `hg -R <bundle-file> log` to automatically overlay a bundle on top of the | |
28 | existing repository). |
|
28 | existing repository). | |
29 | * Various other features don't work. |
|
29 | * Various other features don't work. | |
30 |
|
30 | |||
31 | This extension should work for basic clone/pull, update, and commit workflows. |
|
31 | This extension should work for basic clone/pull, update, and commit workflows. | |
32 | Some history rewriting operations may fail due to lack of support for bundle |
|
32 | Some history rewriting operations may fail due to lack of support for bundle | |
33 | repositories. |
|
33 | repositories. | |
34 |
|
34 | |||
35 | To use, activate the extension and set the ``storage.new-repo-backend`` config |
|
35 | To use, activate the extension and set the ``storage.new-repo-backend`` config | |
36 | option to ``sqlite`` to enable new repositories to use SQLite for storage. |
|
36 | option to ``sqlite`` to enable new repositories to use SQLite for storage. | |
37 | """ |
|
37 | """ | |
38 |
|
38 | |||
39 | # To run the test suite with repos using SQLite by default, execute the |
|
39 | # To run the test suite with repos using SQLite by default, execute the | |
40 | # following: |
|
40 | # following: | |
41 | # |
|
41 | # | |
42 | # HGREPOFEATURES="sqlitestore" run-tests.py \ |
|
42 | # HGREPOFEATURES="sqlitestore" run-tests.py \ | |
43 | # --extra-config-opt extensions.sqlitestore= \ |
|
43 | # --extra-config-opt extensions.sqlitestore= \ | |
44 | # --extra-config-opt storage.new-repo-backend=sqlite |
|
44 | # --extra-config-opt storage.new-repo-backend=sqlite | |
45 |
|
45 | |||
46 | from __future__ import absolute_import |
|
46 | from __future__ import absolute_import | |
47 |
|
47 | |||
48 | import sqlite3 |
|
48 | import sqlite3 | |
49 | import struct |
|
49 | import struct | |
50 | import threading |
|
50 | import threading | |
51 | import zlib |
|
51 | import zlib | |
52 |
|
52 | |||
53 | from mercurial.i18n import _ |
|
53 | from mercurial.i18n import _ | |
54 | from mercurial.node import ( |
|
54 | from mercurial.node import ( | |
55 | nullid, |
|
55 | nullid, | |
56 | nullrev, |
|
56 | nullrev, | |
57 | short, |
|
57 | short, | |
58 | ) |
|
58 | ) | |
59 | from mercurial.thirdparty import attr |
|
59 | from mercurial.thirdparty import attr | |
60 | from mercurial import ( |
|
60 | from mercurial import ( | |
61 | ancestor, |
|
61 | ancestor, | |
62 | dagop, |
|
62 | dagop, | |
63 | encoding, |
|
63 | encoding, | |
64 | error, |
|
64 | error, | |
65 | extensions, |
|
65 | extensions, | |
66 | localrepo, |
|
66 | localrepo, | |
67 | mdiff, |
|
67 | mdiff, | |
68 | pycompat, |
|
68 | pycompat, | |
69 | registrar, |
|
69 | registrar, | |
70 | requirements, |
|
70 | requirements, | |
71 | util, |
|
71 | util, | |
72 | verify, |
|
72 | verify, | |
73 | ) |
|
73 | ) | |
74 | from mercurial.interfaces import ( |
|
74 | from mercurial.interfaces import ( | |
75 | repository, |
|
75 | repository, | |
76 | util as interfaceutil, |
|
76 | util as interfaceutil, | |
77 | ) |
|
77 | ) | |
78 | from mercurial.utils import ( |
|
78 | from mercurial.utils import ( | |
79 | hashutil, |
|
79 | hashutil, | |
80 | storageutil, |
|
80 | storageutil, | |
81 | ) |
|
81 | ) | |
82 |
|
82 | |||
83 | try: |
|
83 | try: | |
84 | from mercurial import zstd |
|
84 | from mercurial import zstd | |
85 |
|
85 | |||
86 | zstd.__version__ |
|
86 | zstd.__version__ | |
87 | except ImportError: |
|
87 | except ImportError: | |
88 | zstd = None |
|
88 | zstd = None | |
89 |
|
89 | |||
90 | configtable = {} |
|
90 | configtable = {} | |
91 | configitem = registrar.configitem(configtable) |
|
91 | configitem = registrar.configitem(configtable) | |
92 |
|
92 | |||
93 | # experimental config: storage.sqlite.compression |
|
93 | # experimental config: storage.sqlite.compression | |
94 | configitem( |
|
94 | configitem( | |
95 | b'storage', |
|
95 | b'storage', | |
96 | b'sqlite.compression', |
|
96 | b'sqlite.compression', | |
97 | default=b'zstd' if zstd else b'zlib', |
|
97 | default=b'zstd' if zstd else b'zlib', | |
98 | experimental=True, |
|
98 | experimental=True, | |
99 | ) |
|
99 | ) | |
100 |
|
100 | |||
101 | # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for |
|
101 | # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for | |
102 | # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should |
|
102 | # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should | |
103 | # be specifying the version(s) of Mercurial they are tested with, or |
|
103 | # be specifying the version(s) of Mercurial they are tested with, or | |
104 | # leave the attribute unspecified. |
|
104 | # leave the attribute unspecified. | |
105 | testedwith = b'ships-with-hg-core' |
|
105 | testedwith = b'ships-with-hg-core' | |
106 |
|
106 | |||
107 | REQUIREMENT = b'exp-sqlite-001' |
|
107 | REQUIREMENT = b'exp-sqlite-001' | |
108 | REQUIREMENT_ZSTD = b'exp-sqlite-comp-001=zstd' |
|
108 | REQUIREMENT_ZSTD = b'exp-sqlite-comp-001=zstd' | |
109 | REQUIREMENT_ZLIB = b'exp-sqlite-comp-001=zlib' |
|
109 | REQUIREMENT_ZLIB = b'exp-sqlite-comp-001=zlib' | |
110 | REQUIREMENT_NONE = b'exp-sqlite-comp-001=none' |
|
110 | REQUIREMENT_NONE = b'exp-sqlite-comp-001=none' | |
111 | REQUIREMENT_SHALLOW_FILES = b'exp-sqlite-shallow-files' |
|
111 | REQUIREMENT_SHALLOW_FILES = b'exp-sqlite-shallow-files' | |
112 |
|
112 | |||
113 | CURRENT_SCHEMA_VERSION = 1 |
|
113 | CURRENT_SCHEMA_VERSION = 1 | |
114 |
|
114 | |||
115 | COMPRESSION_NONE = 1 |
|
115 | COMPRESSION_NONE = 1 | |
116 | COMPRESSION_ZSTD = 2 |
|
116 | COMPRESSION_ZSTD = 2 | |
117 | COMPRESSION_ZLIB = 3 |
|
117 | COMPRESSION_ZLIB = 3 | |
118 |
|
118 | |||
119 | FLAG_CENSORED = 1 |
|
119 | FLAG_CENSORED = 1 | |
120 | FLAG_MISSING_P1 = 2 |
|
120 | FLAG_MISSING_P1 = 2 | |
121 | FLAG_MISSING_P2 = 4 |
|
121 | FLAG_MISSING_P2 = 4 | |
122 |
|
122 | |||
123 | CREATE_SCHEMA = [ |
|
123 | CREATE_SCHEMA = [ | |
124 | # Deltas are stored as content-indexed blobs. |
|
124 | # Deltas are stored as content-indexed blobs. | |
125 | # compression column holds COMPRESSION_* constant for how the |
|
125 | # compression column holds COMPRESSION_* constant for how the | |
126 | # delta is encoded. |
|
126 | # delta is encoded. | |
127 | 'CREATE TABLE delta (' |
|
127 | 'CREATE TABLE delta (' | |
128 | ' id INTEGER PRIMARY KEY, ' |
|
128 | ' id INTEGER PRIMARY KEY, ' | |
129 | ' compression INTEGER NOT NULL, ' |
|
129 | ' compression INTEGER NOT NULL, ' | |
130 | ' hash BLOB UNIQUE ON CONFLICT ABORT, ' |
|
130 | ' hash BLOB UNIQUE ON CONFLICT ABORT, ' | |
131 | ' delta BLOB NOT NULL ' |
|
131 | ' delta BLOB NOT NULL ' | |
132 | ')', |
|
132 | ')', | |
133 | # Tracked paths are denormalized to integers to avoid redundant |
|
133 | # Tracked paths are denormalized to integers to avoid redundant | |
134 | # storage of the path name. |
|
134 | # storage of the path name. | |
135 | 'CREATE TABLE filepath (' |
|
135 | 'CREATE TABLE filepath (' | |
136 | ' id INTEGER PRIMARY KEY, ' |
|
136 | ' id INTEGER PRIMARY KEY, ' | |
137 | ' path BLOB NOT NULL ' |
|
137 | ' path BLOB NOT NULL ' | |
138 | ')', |
|
138 | ')', | |
139 | 'CREATE UNIQUE INDEX filepath_path ON filepath (path)', |
|
139 | 'CREATE UNIQUE INDEX filepath_path ON filepath (path)', | |
140 | # We have a single table for all file revision data. |
|
140 | # We have a single table for all file revision data. | |
141 | # Each file revision is uniquely described by a (path, rev) and |
|
141 | # Each file revision is uniquely described by a (path, rev) and | |
142 | # (path, node). |
|
142 | # (path, node). | |
143 | # |
|
143 | # | |
144 | # Revision data is stored as a pointer to the delta producing this |
|
144 | # Revision data is stored as a pointer to the delta producing this | |
145 | # revision and the file revision whose delta should be applied before |
|
145 | # revision and the file revision whose delta should be applied before | |
146 | # that one. One can reconstruct the delta chain by recursively following |
|
146 | # that one. One can reconstruct the delta chain by recursively following | |
147 | # the delta base revision pointers until one encounters NULL. |
|
147 | # the delta base revision pointers until one encounters NULL. | |
148 | # |
|
148 | # | |
149 | # flags column holds bitwise integer flags controlling storage options. |
|
149 | # flags column holds bitwise integer flags controlling storage options. | |
150 | # These flags are defined by the FLAG_* constants. |
|
150 | # These flags are defined by the FLAG_* constants. | |
151 | 'CREATE TABLE fileindex (' |
|
151 | 'CREATE TABLE fileindex (' | |
152 | ' id INTEGER PRIMARY KEY, ' |
|
152 | ' id INTEGER PRIMARY KEY, ' | |
153 | ' pathid INTEGER REFERENCES filepath(id), ' |
|
153 | ' pathid INTEGER REFERENCES filepath(id), ' | |
154 | ' revnum INTEGER NOT NULL, ' |
|
154 | ' revnum INTEGER NOT NULL, ' | |
155 | ' p1rev INTEGER NOT NULL, ' |
|
155 | ' p1rev INTEGER NOT NULL, ' | |
156 | ' p2rev INTEGER NOT NULL, ' |
|
156 | ' p2rev INTEGER NOT NULL, ' | |
157 | ' linkrev INTEGER NOT NULL, ' |
|
157 | ' linkrev INTEGER NOT NULL, ' | |
158 | ' flags INTEGER NOT NULL, ' |
|
158 | ' flags INTEGER NOT NULL, ' | |
159 | ' deltaid INTEGER REFERENCES delta(id), ' |
|
159 | ' deltaid INTEGER REFERENCES delta(id), ' | |
160 | ' deltabaseid INTEGER REFERENCES fileindex(id), ' |
|
160 | ' deltabaseid INTEGER REFERENCES fileindex(id), ' | |
161 | ' node BLOB NOT NULL ' |
|
161 | ' node BLOB NOT NULL ' | |
162 | ')', |
|
162 | ')', | |
163 | 'CREATE UNIQUE INDEX fileindex_pathrevnum ' |
|
163 | 'CREATE UNIQUE INDEX fileindex_pathrevnum ' | |
164 | ' ON fileindex (pathid, revnum)', |
|
164 | ' ON fileindex (pathid, revnum)', | |
165 | 'CREATE UNIQUE INDEX fileindex_pathnode ON fileindex (pathid, node)', |
|
165 | 'CREATE UNIQUE INDEX fileindex_pathnode ON fileindex (pathid, node)', | |
166 | # Provide a view over all file data for convenience. |
|
166 | # Provide a view over all file data for convenience. | |
167 | 'CREATE VIEW filedata AS ' |
|
167 | 'CREATE VIEW filedata AS ' | |
168 | 'SELECT ' |
|
168 | 'SELECT ' | |
169 | ' fileindex.id AS id, ' |
|
169 | ' fileindex.id AS id, ' | |
170 | ' filepath.id AS pathid, ' |
|
170 | ' filepath.id AS pathid, ' | |
171 | ' filepath.path AS path, ' |
|
171 | ' filepath.path AS path, ' | |
172 | ' fileindex.revnum AS revnum, ' |
|
172 | ' fileindex.revnum AS revnum, ' | |
173 | ' fileindex.node AS node, ' |
|
173 | ' fileindex.node AS node, ' | |
174 | ' fileindex.p1rev AS p1rev, ' |
|
174 | ' fileindex.p1rev AS p1rev, ' | |
175 | ' fileindex.p2rev AS p2rev, ' |
|
175 | ' fileindex.p2rev AS p2rev, ' | |
176 | ' fileindex.linkrev AS linkrev, ' |
|
176 | ' fileindex.linkrev AS linkrev, ' | |
177 | ' fileindex.flags AS flags, ' |
|
177 | ' fileindex.flags AS flags, ' | |
178 | ' fileindex.deltaid AS deltaid, ' |
|
178 | ' fileindex.deltaid AS deltaid, ' | |
179 | ' fileindex.deltabaseid AS deltabaseid ' |
|
179 | ' fileindex.deltabaseid AS deltabaseid ' | |
180 | 'FROM filepath, fileindex ' |
|
180 | 'FROM filepath, fileindex ' | |
181 | 'WHERE fileindex.pathid=filepath.id', |
|
181 | 'WHERE fileindex.pathid=filepath.id', | |
182 | 'PRAGMA user_version=%d' % CURRENT_SCHEMA_VERSION, |
|
182 | 'PRAGMA user_version=%d' % CURRENT_SCHEMA_VERSION, | |
183 | ] |
|
183 | ] | |
184 |
|
184 | |||
185 |
|
185 | |||
186 | def resolvedeltachain(db, pathid, node, revisioncache, stoprids, zstddctx=None): |
|
186 | def resolvedeltachain(db, pathid, node, revisioncache, stoprids, zstddctx=None): | |
187 | """Resolve a delta chain for a file node.""" |
|
187 | """Resolve a delta chain for a file node.""" | |
188 |
|
188 | |||
189 | # TODO the "not in ({stops})" here is possibly slowing down the query |
|
189 | # TODO the "not in ({stops})" here is possibly slowing down the query | |
190 | # because it needs to perform the lookup on every recursive invocation. |
|
190 | # because it needs to perform the lookup on every recursive invocation. | |
191 | # This could possibly be faster if we created a temporary query with |
|
191 | # This could possibly be faster if we created a temporary query with | |
192 | # baseid "poisoned" to null and limited the recursive filter to |
|
192 | # baseid "poisoned" to null and limited the recursive filter to | |
193 | # "is not null". |
|
193 | # "is not null". | |
194 | res = db.execute( |
|
194 | res = db.execute( | |
195 | 'WITH RECURSIVE ' |
|
195 | 'WITH RECURSIVE ' | |
196 | ' deltachain(deltaid, baseid) AS (' |
|
196 | ' deltachain(deltaid, baseid) AS (' | |
197 | ' SELECT deltaid, deltabaseid FROM fileindex ' |
|
197 | ' SELECT deltaid, deltabaseid FROM fileindex ' | |
198 | ' WHERE pathid=? AND node=? ' |
|
198 | ' WHERE pathid=? AND node=? ' | |
199 | ' UNION ALL ' |
|
199 | ' UNION ALL ' | |
200 | ' SELECT fileindex.deltaid, deltabaseid ' |
|
200 | ' SELECT fileindex.deltaid, deltabaseid ' | |
201 | ' FROM fileindex, deltachain ' |
|
201 | ' FROM fileindex, deltachain ' | |
202 | ' WHERE ' |
|
202 | ' WHERE ' | |
203 | ' fileindex.id=deltachain.baseid ' |
|
203 | ' fileindex.id=deltachain.baseid ' | |
204 | ' AND deltachain.baseid IS NOT NULL ' |
|
204 | ' AND deltachain.baseid IS NOT NULL ' | |
205 | ' AND fileindex.id NOT IN ({stops}) ' |
|
205 | ' AND fileindex.id NOT IN ({stops}) ' | |
206 | ' ) ' |
|
206 | ' ) ' | |
207 | 'SELECT deltachain.baseid, compression, delta ' |
|
207 | 'SELECT deltachain.baseid, compression, delta ' | |
208 | 'FROM deltachain, delta ' |
|
208 | 'FROM deltachain, delta ' | |
209 | 'WHERE delta.id=deltachain.deltaid'.format( |
|
209 | 'WHERE delta.id=deltachain.deltaid'.format( | |
210 | stops=','.join(['?'] * len(stoprids)) |
|
210 | stops=','.join(['?'] * len(stoprids)) | |
211 | ), |
|
211 | ), | |
212 | tuple([pathid, node] + list(stoprids.keys())), |
|
212 | tuple([pathid, node] + list(stoprids.keys())), | |
213 | ) |
|
213 | ) | |
214 |
|
214 | |||
215 | deltas = [] |
|
215 | deltas = [] | |
216 | lastdeltabaseid = None |
|
216 | lastdeltabaseid = None | |
217 |
|
217 | |||
218 | for deltabaseid, compression, delta in res: |
|
218 | for deltabaseid, compression, delta in res: | |
219 | lastdeltabaseid = deltabaseid |
|
219 | lastdeltabaseid = deltabaseid | |
220 |
|
220 | |||
221 | if compression == COMPRESSION_ZSTD: |
|
221 | if compression == COMPRESSION_ZSTD: | |
222 | delta = zstddctx.decompress(delta) |
|
222 | delta = zstddctx.decompress(delta) | |
223 | elif compression == COMPRESSION_NONE: |
|
223 | elif compression == COMPRESSION_NONE: | |
224 | delta = delta |
|
224 | delta = delta | |
225 | elif compression == COMPRESSION_ZLIB: |
|
225 | elif compression == COMPRESSION_ZLIB: | |
226 | delta = zlib.decompress(delta) |
|
226 | delta = zlib.decompress(delta) | |
227 | else: |
|
227 | else: | |
228 | raise SQLiteStoreError( |
|
228 | raise SQLiteStoreError( | |
229 | b'unhandled compression type: %d' % compression |
|
229 | b'unhandled compression type: %d' % compression | |
230 | ) |
|
230 | ) | |
231 |
|
231 | |||
232 | deltas.append(delta) |
|
232 | deltas.append(delta) | |
233 |
|
233 | |||
234 | if lastdeltabaseid in stoprids: |
|
234 | if lastdeltabaseid in stoprids: | |
235 | basetext = revisioncache[stoprids[lastdeltabaseid]] |
|
235 | basetext = revisioncache[stoprids[lastdeltabaseid]] | |
236 | else: |
|
236 | else: | |
237 | basetext = deltas.pop() |
|
237 | basetext = deltas.pop() | |
238 |
|
238 | |||
239 | deltas.reverse() |
|
239 | deltas.reverse() | |
240 | fulltext = mdiff.patches(basetext, deltas) |
|
240 | fulltext = mdiff.patches(basetext, deltas) | |
241 |
|
241 | |||
242 | # SQLite returns buffer instances for blob columns on Python 2. This |
|
242 | # SQLite returns buffer instances for blob columns on Python 2. This | |
243 | # type can propagate through the delta application layer. Because |
|
243 | # type can propagate through the delta application layer. Because | |
244 | # downstream callers assume revisions are bytes, cast as needed. |
|
244 | # downstream callers assume revisions are bytes, cast as needed. | |
245 | if not isinstance(fulltext, bytes): |
|
245 | if not isinstance(fulltext, bytes): | |
246 | fulltext = bytes(delta) |
|
246 | fulltext = bytes(delta) | |
247 |
|
247 | |||
248 | return fulltext |
|
248 | return fulltext | |
249 |
|
249 | |||
250 |
|
250 | |||
251 | def insertdelta(db, compression, hash, delta): |
|
251 | def insertdelta(db, compression, hash, delta): | |
252 | try: |
|
252 | try: | |
253 | return db.execute( |
|
253 | return db.execute( | |
254 | 'INSERT INTO delta (compression, hash, delta) VALUES (?, ?, ?)', |
|
254 | 'INSERT INTO delta (compression, hash, delta) VALUES (?, ?, ?)', | |
255 | (compression, hash, delta), |
|
255 | (compression, hash, delta), | |
256 | ).lastrowid |
|
256 | ).lastrowid | |
257 | except sqlite3.IntegrityError: |
|
257 | except sqlite3.IntegrityError: | |
258 | return db.execute( |
|
258 | return db.execute( | |
259 | 'SELECT id FROM delta WHERE hash=?', (hash,) |
|
259 | 'SELECT id FROM delta WHERE hash=?', (hash,) | |
260 | ).fetchone()[0] |
|
260 | ).fetchone()[0] | |
261 |
|
261 | |||
262 |
|
262 | |||
263 | class SQLiteStoreError(error.StorageError): |
|
263 | class SQLiteStoreError(error.StorageError): | |
264 | pass |
|
264 | pass | |
265 |
|
265 | |||
266 |
|
266 | |||
267 | @attr.s |
|
267 | @attr.s | |
268 | class revisionentry(object): |
|
268 | class revisionentry(object): | |
269 | rid = attr.ib() |
|
269 | rid = attr.ib() | |
270 | rev = attr.ib() |
|
270 | rev = attr.ib() | |
271 | node = attr.ib() |
|
271 | node = attr.ib() | |
272 | p1rev = attr.ib() |
|
272 | p1rev = attr.ib() | |
273 | p2rev = attr.ib() |
|
273 | p2rev = attr.ib() | |
274 | p1node = attr.ib() |
|
274 | p1node = attr.ib() | |
275 | p2node = attr.ib() |
|
275 | p2node = attr.ib() | |
276 | linkrev = attr.ib() |
|
276 | linkrev = attr.ib() | |
277 | flags = attr.ib() |
|
277 | flags = attr.ib() | |
278 |
|
278 | |||
279 |
|
279 | |||
280 | @interfaceutil.implementer(repository.irevisiondelta) |
|
280 | @interfaceutil.implementer(repository.irevisiondelta) | |
281 | @attr.s(slots=True) |
|
281 | @attr.s(slots=True) | |
282 | class sqliterevisiondelta(object): |
|
282 | class sqliterevisiondelta(object): | |
283 | node = attr.ib() |
|
283 | node = attr.ib() | |
284 | p1node = attr.ib() |
|
284 | p1node = attr.ib() | |
285 | p2node = attr.ib() |
|
285 | p2node = attr.ib() | |
286 | basenode = attr.ib() |
|
286 | basenode = attr.ib() | |
287 | flags = attr.ib() |
|
287 | flags = attr.ib() | |
288 | baserevisionsize = attr.ib() |
|
288 | baserevisionsize = attr.ib() | |
289 | revision = attr.ib() |
|
289 | revision = attr.ib() | |
290 | delta = attr.ib() |
|
290 | delta = attr.ib() | |
291 | sidedata = attr.ib() |
|
291 | sidedata = attr.ib() | |
292 | linknode = attr.ib(default=None) |
|
292 | linknode = attr.ib(default=None) | |
293 |
|
293 | |||
294 |
|
294 | |||
295 | @interfaceutil.implementer(repository.iverifyproblem) |
|
295 | @interfaceutil.implementer(repository.iverifyproblem) | |
296 | @attr.s(frozen=True) |
|
296 | @attr.s(frozen=True) | |
297 | class sqliteproblem(object): |
|
297 | class sqliteproblem(object): | |
298 | warning = attr.ib(default=None) |
|
298 | warning = attr.ib(default=None) | |
299 | error = attr.ib(default=None) |
|
299 | error = attr.ib(default=None) | |
300 | node = attr.ib(default=None) |
|
300 | node = attr.ib(default=None) | |
301 |
|
301 | |||
302 |
|
302 | |||
303 | @interfaceutil.implementer(repository.ifilestorage) |
|
303 | @interfaceutil.implementer(repository.ifilestorage) | |
304 | class sqlitefilestore(object): |
|
304 | class sqlitefilestore(object): | |
305 | """Implements storage for an individual tracked path.""" |
|
305 | """Implements storage for an individual tracked path.""" | |
306 |
|
306 | |||
307 | def __init__(self, db, path, compression): |
|
307 | def __init__(self, db, path, compression): | |
308 | self._db = db |
|
308 | self._db = db | |
309 | self._path = path |
|
309 | self._path = path | |
310 |
|
310 | |||
311 | self._pathid = None |
|
311 | self._pathid = None | |
312 |
|
312 | |||
313 | # revnum -> node |
|
313 | # revnum -> node | |
314 | self._revtonode = {} |
|
314 | self._revtonode = {} | |
315 | # node -> revnum |
|
315 | # node -> revnum | |
316 | self._nodetorev = {} |
|
316 | self._nodetorev = {} | |
317 | # node -> data structure |
|
317 | # node -> data structure | |
318 | self._revisions = {} |
|
318 | self._revisions = {} | |
319 |
|
319 | |||
320 | self._revisioncache = util.lrucachedict(10) |
|
320 | self._revisioncache = util.lrucachedict(10) | |
321 |
|
321 | |||
322 | self._compengine = compression |
|
322 | self._compengine = compression | |
323 |
|
323 | |||
324 | if compression == b'zstd': |
|
324 | if compression == b'zstd': | |
325 | self._cctx = zstd.ZstdCompressor(level=3) |
|
325 | self._cctx = zstd.ZstdCompressor(level=3) | |
326 | self._dctx = zstd.ZstdDecompressor() |
|
326 | self._dctx = zstd.ZstdDecompressor() | |
327 | else: |
|
327 | else: | |
328 | self._cctx = None |
|
328 | self._cctx = None | |
329 | self._dctx = None |
|
329 | self._dctx = None | |
330 |
|
330 | |||
331 | self._refreshindex() |
|
331 | self._refreshindex() | |
332 |
|
332 | |||
333 | def _refreshindex(self): |
|
333 | def _refreshindex(self): | |
334 | self._revtonode = {} |
|
334 | self._revtonode = {} | |
335 | self._nodetorev = {} |
|
335 | self._nodetorev = {} | |
336 | self._revisions = {} |
|
336 | self._revisions = {} | |
337 |
|
337 | |||
338 | res = list( |
|
338 | res = list( | |
339 | self._db.execute( |
|
339 | self._db.execute( | |
340 | 'SELECT id FROM filepath WHERE path=?', (self._path,) |
|
340 | 'SELECT id FROM filepath WHERE path=?', (self._path,) | |
341 | ) |
|
341 | ) | |
342 | ) |
|
342 | ) | |
343 |
|
343 | |||
344 | if not res: |
|
344 | if not res: | |
345 | self._pathid = None |
|
345 | self._pathid = None | |
346 | return |
|
346 | return | |
347 |
|
347 | |||
348 | self._pathid = res[0][0] |
|
348 | self._pathid = res[0][0] | |
349 |
|
349 | |||
350 | res = self._db.execute( |
|
350 | res = self._db.execute( | |
351 | 'SELECT id, revnum, node, p1rev, p2rev, linkrev, flags ' |
|
351 | 'SELECT id, revnum, node, p1rev, p2rev, linkrev, flags ' | |
352 | 'FROM fileindex ' |
|
352 | 'FROM fileindex ' | |
353 | 'WHERE pathid=? ' |
|
353 | 'WHERE pathid=? ' | |
354 | 'ORDER BY revnum ASC', |
|
354 | 'ORDER BY revnum ASC', | |
355 | (self._pathid,), |
|
355 | (self._pathid,), | |
356 | ) |
|
356 | ) | |
357 |
|
357 | |||
358 | for i, row in enumerate(res): |
|
358 | for i, row in enumerate(res): | |
359 | rid, rev, node, p1rev, p2rev, linkrev, flags = row |
|
359 | rid, rev, node, p1rev, p2rev, linkrev, flags = row | |
360 |
|
360 | |||
361 | if i != rev: |
|
361 | if i != rev: | |
362 | raise SQLiteStoreError( |
|
362 | raise SQLiteStoreError( | |
363 | _(b'sqlite database has inconsistent revision numbers') |
|
363 | _(b'sqlite database has inconsistent revision numbers') | |
364 | ) |
|
364 | ) | |
365 |
|
365 | |||
366 | if p1rev == nullrev: |
|
366 | if p1rev == nullrev: | |
367 | p1node = nullid |
|
367 | p1node = nullid | |
368 | else: |
|
368 | else: | |
369 | p1node = self._revtonode[p1rev] |
|
369 | p1node = self._revtonode[p1rev] | |
370 |
|
370 | |||
371 | if p2rev == nullrev: |
|
371 | if p2rev == nullrev: | |
372 | p2node = nullid |
|
372 | p2node = nullid | |
373 | else: |
|
373 | else: | |
374 | p2node = self._revtonode[p2rev] |
|
374 | p2node = self._revtonode[p2rev] | |
375 |
|
375 | |||
376 | entry = revisionentry( |
|
376 | entry = revisionentry( | |
377 | rid=rid, |
|
377 | rid=rid, | |
378 | rev=rev, |
|
378 | rev=rev, | |
379 | node=node, |
|
379 | node=node, | |
380 | p1rev=p1rev, |
|
380 | p1rev=p1rev, | |
381 | p2rev=p2rev, |
|
381 | p2rev=p2rev, | |
382 | p1node=p1node, |
|
382 | p1node=p1node, | |
383 | p2node=p2node, |
|
383 | p2node=p2node, | |
384 | linkrev=linkrev, |
|
384 | linkrev=linkrev, | |
385 | flags=flags, |
|
385 | flags=flags, | |
386 | ) |
|
386 | ) | |
387 |
|
387 | |||
388 | self._revtonode[rev] = node |
|
388 | self._revtonode[rev] = node | |
389 | self._nodetorev[node] = rev |
|
389 | self._nodetorev[node] = rev | |
390 | self._revisions[node] = entry |
|
390 | self._revisions[node] = entry | |
391 |
|
391 | |||
392 | # Start of ifileindex interface. |
|
392 | # Start of ifileindex interface. | |
393 |
|
393 | |||
394 | def __len__(self): |
|
394 | def __len__(self): | |
395 | return len(self._revisions) |
|
395 | return len(self._revisions) | |
396 |
|
396 | |||
397 | def __iter__(self): |
|
397 | def __iter__(self): | |
398 | return iter(pycompat.xrange(len(self._revisions))) |
|
398 | return iter(pycompat.xrange(len(self._revisions))) | |
399 |
|
399 | |||
400 | def hasnode(self, node): |
|
400 | def hasnode(self, node): | |
401 | if node == nullid: |
|
401 | if node == nullid: | |
402 | return False |
|
402 | return False | |
403 |
|
403 | |||
404 | return node in self._nodetorev |
|
404 | return node in self._nodetorev | |
405 |
|
405 | |||
406 | def revs(self, start=0, stop=None): |
|
406 | def revs(self, start=0, stop=None): | |
407 | return storageutil.iterrevs( |
|
407 | return storageutil.iterrevs( | |
408 | len(self._revisions), start=start, stop=stop |
|
408 | len(self._revisions), start=start, stop=stop | |
409 | ) |
|
409 | ) | |
410 |
|
410 | |||
411 | def parents(self, node): |
|
411 | def parents(self, node): | |
412 | if node == nullid: |
|
412 | if node == nullid: | |
413 | return nullid, nullid |
|
413 | return nullid, nullid | |
414 |
|
414 | |||
415 | if node not in self._revisions: |
|
415 | if node not in self._revisions: | |
416 | raise error.LookupError(node, self._path, _(b'no node')) |
|
416 | raise error.LookupError(node, self._path, _(b'no node')) | |
417 |
|
417 | |||
418 | entry = self._revisions[node] |
|
418 | entry = self._revisions[node] | |
419 | return entry.p1node, entry.p2node |
|
419 | return entry.p1node, entry.p2node | |
420 |
|
420 | |||
421 | def parentrevs(self, rev): |
|
421 | def parentrevs(self, rev): | |
422 | if rev == nullrev: |
|
422 | if rev == nullrev: | |
423 | return nullrev, nullrev |
|
423 | return nullrev, nullrev | |
424 |
|
424 | |||
425 | if rev not in self._revtonode: |
|
425 | if rev not in self._revtonode: | |
426 | raise IndexError(rev) |
|
426 | raise IndexError(rev) | |
427 |
|
427 | |||
428 | entry = self._revisions[self._revtonode[rev]] |
|
428 | entry = self._revisions[self._revtonode[rev]] | |
429 | return entry.p1rev, entry.p2rev |
|
429 | return entry.p1rev, entry.p2rev | |
430 |
|
430 | |||
431 | def rev(self, node): |
|
431 | def rev(self, node): | |
432 | if node == nullid: |
|
432 | if node == nullid: | |
433 | return nullrev |
|
433 | return nullrev | |
434 |
|
434 | |||
435 | if node not in self._nodetorev: |
|
435 | if node not in self._nodetorev: | |
436 | raise error.LookupError(node, self._path, _(b'no node')) |
|
436 | raise error.LookupError(node, self._path, _(b'no node')) | |
437 |
|
437 | |||
438 | return self._nodetorev[node] |
|
438 | return self._nodetorev[node] | |
439 |
|
439 | |||
440 | def node(self, rev): |
|
440 | def node(self, rev): | |
441 | if rev == nullrev: |
|
441 | if rev == nullrev: | |
442 | return nullid |
|
442 | return nullid | |
443 |
|
443 | |||
444 | if rev not in self._revtonode: |
|
444 | if rev not in self._revtonode: | |
445 | raise IndexError(rev) |
|
445 | raise IndexError(rev) | |
446 |
|
446 | |||
447 | return self._revtonode[rev] |
|
447 | return self._revtonode[rev] | |
448 |
|
448 | |||
449 | def lookup(self, node): |
|
449 | def lookup(self, node): | |
450 | return storageutil.fileidlookup(self, node, self._path) |
|
450 | return storageutil.fileidlookup(self, node, self._path) | |
451 |
|
451 | |||
452 | def linkrev(self, rev): |
|
452 | def linkrev(self, rev): | |
453 | if rev == nullrev: |
|
453 | if rev == nullrev: | |
454 | return nullrev |
|
454 | return nullrev | |
455 |
|
455 | |||
456 | if rev not in self._revtonode: |
|
456 | if rev not in self._revtonode: | |
457 | raise IndexError(rev) |
|
457 | raise IndexError(rev) | |
458 |
|
458 | |||
459 | entry = self._revisions[self._revtonode[rev]] |
|
459 | entry = self._revisions[self._revtonode[rev]] | |
460 | return entry.linkrev |
|
460 | return entry.linkrev | |
461 |
|
461 | |||
462 | def iscensored(self, rev): |
|
462 | def iscensored(self, rev): | |
463 | if rev == nullrev: |
|
463 | if rev == nullrev: | |
464 | return False |
|
464 | return False | |
465 |
|
465 | |||
466 | if rev not in self._revtonode: |
|
466 | if rev not in self._revtonode: | |
467 | raise IndexError(rev) |
|
467 | raise IndexError(rev) | |
468 |
|
468 | |||
469 | return self._revisions[self._revtonode[rev]].flags & FLAG_CENSORED |
|
469 | return self._revisions[self._revtonode[rev]].flags & FLAG_CENSORED | |
470 |
|
470 | |||
471 | def commonancestorsheads(self, node1, node2): |
|
471 | def commonancestorsheads(self, node1, node2): | |
472 | rev1 = self.rev(node1) |
|
472 | rev1 = self.rev(node1) | |
473 | rev2 = self.rev(node2) |
|
473 | rev2 = self.rev(node2) | |
474 |
|
474 | |||
475 | ancestors = ancestor.commonancestorsheads(self.parentrevs, rev1, rev2) |
|
475 | ancestors = ancestor.commonancestorsheads(self.parentrevs, rev1, rev2) | |
476 | return pycompat.maplist(self.node, ancestors) |
|
476 | return pycompat.maplist(self.node, ancestors) | |
477 |
|
477 | |||
478 | def descendants(self, revs): |
|
478 | def descendants(self, revs): | |
479 | # TODO we could implement this using a recursive SQL query, which |
|
479 | # TODO we could implement this using a recursive SQL query, which | |
480 | # might be faster. |
|
480 | # might be faster. | |
481 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
481 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) | |
482 |
|
482 | |||
483 | def heads(self, start=None, stop=None): |
|
483 | def heads(self, start=None, stop=None): | |
484 | if start is None and stop is None: |
|
484 | if start is None and stop is None: | |
485 | if not len(self): |
|
485 | if not len(self): | |
486 | return [nullid] |
|
486 | return [nullid] | |
487 |
|
487 | |||
488 | startrev = self.rev(start) if start is not None else nullrev |
|
488 | startrev = self.rev(start) if start is not None else nullrev | |
489 | stoprevs = {self.rev(n) for n in stop or []} |
|
489 | stoprevs = {self.rev(n) for n in stop or []} | |
490 |
|
490 | |||
491 | revs = dagop.headrevssubset( |
|
491 | revs = dagop.headrevssubset( | |
492 | self.revs, self.parentrevs, startrev=startrev, stoprevs=stoprevs |
|
492 | self.revs, self.parentrevs, startrev=startrev, stoprevs=stoprevs | |
493 | ) |
|
493 | ) | |
494 |
|
494 | |||
495 | return [self.node(rev) for rev in revs] |
|
495 | return [self.node(rev) for rev in revs] | |
496 |
|
496 | |||
497 | def children(self, node): |
|
497 | def children(self, node): | |
498 | rev = self.rev(node) |
|
498 | rev = self.rev(node) | |
499 |
|
499 | |||
500 | res = self._db.execute( |
|
500 | res = self._db.execute( | |
501 | 'SELECT' |
|
501 | 'SELECT' | |
502 | ' node ' |
|
502 | ' node ' | |
503 | ' FROM filedata ' |
|
503 | ' FROM filedata ' | |
504 | ' WHERE path=? AND (p1rev=? OR p2rev=?) ' |
|
504 | ' WHERE path=? AND (p1rev=? OR p2rev=?) ' | |
505 | ' ORDER BY revnum ASC', |
|
505 | ' ORDER BY revnum ASC', | |
506 | (self._path, rev, rev), |
|
506 | (self._path, rev, rev), | |
507 | ) |
|
507 | ) | |
508 |
|
508 | |||
509 | return [row[0] for row in res] |
|
509 | return [row[0] for row in res] | |
510 |
|
510 | |||
511 | # End of ifileindex interface. |
|
511 | # End of ifileindex interface. | |
512 |
|
512 | |||
513 | # Start of ifiledata interface. |
|
513 | # Start of ifiledata interface. | |
514 |
|
514 | |||
515 | def size(self, rev): |
|
515 | def size(self, rev): | |
516 | if rev == nullrev: |
|
516 | if rev == nullrev: | |
517 | return 0 |
|
517 | return 0 | |
518 |
|
518 | |||
519 | if rev not in self._revtonode: |
|
519 | if rev not in self._revtonode: | |
520 | raise IndexError(rev) |
|
520 | raise IndexError(rev) | |
521 |
|
521 | |||
522 | node = self._revtonode[rev] |
|
522 | node = self._revtonode[rev] | |
523 |
|
523 | |||
524 | if self.renamed(node): |
|
524 | if self.renamed(node): | |
525 | return len(self.read(node)) |
|
525 | return len(self.read(node)) | |
526 |
|
526 | |||
527 | return len(self.revision(node)) |
|
527 | return len(self.revision(node)) | |
528 |
|
528 | |||
529 | def revision(self, node, raw=False, _verifyhash=True): |
|
529 | def revision(self, node, raw=False, _verifyhash=True): | |
530 | if node in (nullid, nullrev): |
|
530 | if node in (nullid, nullrev): | |
531 | return b'' |
|
531 | return b'' | |
532 |
|
532 | |||
533 | if isinstance(node, int): |
|
533 | if isinstance(node, int): | |
534 | node = self.node(node) |
|
534 | node = self.node(node) | |
535 |
|
535 | |||
536 | if node not in self._nodetorev: |
|
536 | if node not in self._nodetorev: | |
537 | raise error.LookupError(node, self._path, _(b'no node')) |
|
537 | raise error.LookupError(node, self._path, _(b'no node')) | |
538 |
|
538 | |||
539 | if node in self._revisioncache: |
|
539 | if node in self._revisioncache: | |
540 | return self._revisioncache[node] |
|
540 | return self._revisioncache[node] | |
541 |
|
541 | |||
542 | # Because we have a fulltext revision cache, we are able to |
|
542 | # Because we have a fulltext revision cache, we are able to | |
543 | # short-circuit delta chain traversal and decompression as soon as |
|
543 | # short-circuit delta chain traversal and decompression as soon as | |
544 | # we encounter a revision in the cache. |
|
544 | # we encounter a revision in the cache. | |
545 |
|
545 | |||
546 | stoprids = {self._revisions[n].rid: n for n in self._revisioncache} |
|
546 | stoprids = {self._revisions[n].rid: n for n in self._revisioncache} | |
547 |
|
547 | |||
548 | if not stoprids: |
|
548 | if not stoprids: | |
549 | stoprids[-1] = None |
|
549 | stoprids[-1] = None | |
550 |
|
550 | |||
551 | fulltext = resolvedeltachain( |
|
551 | fulltext = resolvedeltachain( | |
552 | self._db, |
|
552 | self._db, | |
553 | self._pathid, |
|
553 | self._pathid, | |
554 | node, |
|
554 | node, | |
555 | self._revisioncache, |
|
555 | self._revisioncache, | |
556 | stoprids, |
|
556 | stoprids, | |
557 | zstddctx=self._dctx, |
|
557 | zstddctx=self._dctx, | |
558 | ) |
|
558 | ) | |
559 |
|
559 | |||
560 | # Don't verify hashes if parent nodes were rewritten, as the hash |
|
560 | # Don't verify hashes if parent nodes were rewritten, as the hash | |
561 | # wouldn't verify. |
|
561 | # wouldn't verify. | |
562 | if self._revisions[node].flags & (FLAG_MISSING_P1 | FLAG_MISSING_P2): |
|
562 | if self._revisions[node].flags & (FLAG_MISSING_P1 | FLAG_MISSING_P2): | |
563 | _verifyhash = False |
|
563 | _verifyhash = False | |
564 |
|
564 | |||
565 | if _verifyhash: |
|
565 | if _verifyhash: | |
566 | self._checkhash(fulltext, node) |
|
566 | self._checkhash(fulltext, node) | |
567 | self._revisioncache[node] = fulltext |
|
567 | self._revisioncache[node] = fulltext | |
568 |
|
568 | |||
569 | return fulltext |
|
569 | return fulltext | |
570 |
|
570 | |||
571 | def rawdata(self, *args, **kwargs): |
|
571 | def rawdata(self, *args, **kwargs): | |
572 | return self.revision(*args, **kwargs) |
|
572 | return self.revision(*args, **kwargs) | |
573 |
|
573 | |||
574 | def read(self, node): |
|
574 | def read(self, node): | |
575 | return storageutil.filtermetadata(self.revision(node)) |
|
575 | return storageutil.filtermetadata(self.revision(node)) | |
576 |
|
576 | |||
577 | def renamed(self, node): |
|
577 | def renamed(self, node): | |
578 | return storageutil.filerevisioncopied(self, node) |
|
578 | return storageutil.filerevisioncopied(self, node) | |
579 |
|
579 | |||
580 | def cmp(self, node, fulltext): |
|
580 | def cmp(self, node, fulltext): | |
581 | return not storageutil.filedataequivalent(self, node, fulltext) |
|
581 | return not storageutil.filedataequivalent(self, node, fulltext) | |
582 |
|
582 | |||
583 | def emitrevisions( |
|
583 | def emitrevisions( | |
584 | self, |
|
584 | self, | |
585 | nodes, |
|
585 | nodes, | |
586 | nodesorder=None, |
|
586 | nodesorder=None, | |
587 | revisiondata=False, |
|
587 | revisiondata=False, | |
588 | assumehaveparentrevisions=False, |
|
588 | assumehaveparentrevisions=False, | |
589 | deltamode=repository.CG_DELTAMODE_STD, |
|
589 | deltamode=repository.CG_DELTAMODE_STD, | |
|
590 | sidedata_helpers=None, | |||
590 | ): |
|
591 | ): | |
591 | if nodesorder not in (b'nodes', b'storage', b'linear', None): |
|
592 | if nodesorder not in (b'nodes', b'storage', b'linear', None): | |
592 | raise error.ProgrammingError( |
|
593 | raise error.ProgrammingError( | |
593 | b'unhandled value for nodesorder: %s' % nodesorder |
|
594 | b'unhandled value for nodesorder: %s' % nodesorder | |
594 | ) |
|
595 | ) | |
595 |
|
596 | |||
596 | nodes = [n for n in nodes if n != nullid] |
|
597 | nodes = [n for n in nodes if n != nullid] | |
597 |
|
598 | |||
598 | if not nodes: |
|
599 | if not nodes: | |
599 | return |
|
600 | return | |
600 |
|
601 | |||
601 | # TODO perform in a single query. |
|
602 | # TODO perform in a single query. | |
602 | res = self._db.execute( |
|
603 | res = self._db.execute( | |
603 | 'SELECT revnum, deltaid FROM fileindex ' |
|
604 | 'SELECT revnum, deltaid FROM fileindex ' | |
604 | 'WHERE pathid=? ' |
|
605 | 'WHERE pathid=? ' | |
605 | ' AND node in (%s)' % (','.join(['?'] * len(nodes))), |
|
606 | ' AND node in (%s)' % (','.join(['?'] * len(nodes))), | |
606 | tuple([self._pathid] + nodes), |
|
607 | tuple([self._pathid] + nodes), | |
607 | ) |
|
608 | ) | |
608 |
|
609 | |||
609 | deltabases = {} |
|
610 | deltabases = {} | |
610 |
|
611 | |||
611 | for rev, deltaid in res: |
|
612 | for rev, deltaid in res: | |
612 | res = self._db.execute( |
|
613 | res = self._db.execute( | |
613 | 'SELECT revnum from fileindex WHERE pathid=? AND deltaid=?', |
|
614 | 'SELECT revnum from fileindex WHERE pathid=? AND deltaid=?', | |
614 | (self._pathid, deltaid), |
|
615 | (self._pathid, deltaid), | |
615 | ) |
|
616 | ) | |
616 | deltabases[rev] = res.fetchone()[0] |
|
617 | deltabases[rev] = res.fetchone()[0] | |
617 |
|
618 | |||
618 | # TODO define revdifffn so we can use delta from storage. |
|
619 | # TODO define revdifffn so we can use delta from storage. | |
619 | for delta in storageutil.emitrevisions( |
|
620 | for delta in storageutil.emitrevisions( | |
620 | self, |
|
621 | self, | |
621 | nodes, |
|
622 | nodes, | |
622 | nodesorder, |
|
623 | nodesorder, | |
623 | sqliterevisiondelta, |
|
624 | sqliterevisiondelta, | |
624 | deltaparentfn=deltabases.__getitem__, |
|
625 | deltaparentfn=deltabases.__getitem__, | |
625 | revisiondata=revisiondata, |
|
626 | revisiondata=revisiondata, | |
626 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
627 | assumehaveparentrevisions=assumehaveparentrevisions, | |
627 | deltamode=deltamode, |
|
628 | deltamode=deltamode, | |
|
629 | sidedata_helpers=sidedata_helpers, | |||
628 | ): |
|
630 | ): | |
629 |
|
631 | |||
630 | yield delta |
|
632 | yield delta | |
631 |
|
633 | |||
632 | # End of ifiledata interface. |
|
634 | # End of ifiledata interface. | |
633 |
|
635 | |||
634 | # Start of ifilemutation interface. |
|
636 | # Start of ifilemutation interface. | |
635 |
|
637 | |||
636 | def add(self, filedata, meta, transaction, linkrev, p1, p2): |
|
638 | def add(self, filedata, meta, transaction, linkrev, p1, p2): | |
637 | if meta or filedata.startswith(b'\x01\n'): |
|
639 | if meta or filedata.startswith(b'\x01\n'): | |
638 | filedata = storageutil.packmeta(meta, filedata) |
|
640 | filedata = storageutil.packmeta(meta, filedata) | |
639 |
|
641 | |||
640 | rev = self.addrevision(filedata, transaction, linkrev, p1, p2) |
|
642 | rev = self.addrevision(filedata, transaction, linkrev, p1, p2) | |
641 | return self.node(rev) |
|
643 | return self.node(rev) | |
642 |
|
644 | |||
643 | def addrevision( |
|
645 | def addrevision( | |
644 | self, |
|
646 | self, | |
645 | revisiondata, |
|
647 | revisiondata, | |
646 | transaction, |
|
648 | transaction, | |
647 | linkrev, |
|
649 | linkrev, | |
648 | p1, |
|
650 | p1, | |
649 | p2, |
|
651 | p2, | |
650 | node=None, |
|
652 | node=None, | |
651 | flags=0, |
|
653 | flags=0, | |
652 | cachedelta=None, |
|
654 | cachedelta=None, | |
653 | ): |
|
655 | ): | |
654 | if flags: |
|
656 | if flags: | |
655 | raise SQLiteStoreError(_(b'flags not supported on revisions')) |
|
657 | raise SQLiteStoreError(_(b'flags not supported on revisions')) | |
656 |
|
658 | |||
657 | validatehash = node is not None |
|
659 | validatehash = node is not None | |
658 | node = node or storageutil.hashrevisionsha1(revisiondata, p1, p2) |
|
660 | node = node or storageutil.hashrevisionsha1(revisiondata, p1, p2) | |
659 |
|
661 | |||
660 | if validatehash: |
|
662 | if validatehash: | |
661 | self._checkhash(revisiondata, node, p1, p2) |
|
663 | self._checkhash(revisiondata, node, p1, p2) | |
662 |
|
664 | |||
663 | rev = self._nodetorev.get(node) |
|
665 | rev = self._nodetorev.get(node) | |
664 | if rev is not None: |
|
666 | if rev is not None: | |
665 | return rev |
|
667 | return rev | |
666 |
|
668 | |||
667 | rev = self._addrawrevision( |
|
669 | rev = self._addrawrevision( | |
668 | node, revisiondata, transaction, linkrev, p1, p2 |
|
670 | node, revisiondata, transaction, linkrev, p1, p2 | |
669 | ) |
|
671 | ) | |
670 |
|
672 | |||
671 | self._revisioncache[node] = revisiondata |
|
673 | self._revisioncache[node] = revisiondata | |
672 | return rev |
|
674 | return rev | |
673 |
|
675 | |||
674 | def addgroup( |
|
676 | def addgroup( | |
675 | self, |
|
677 | self, | |
676 | deltas, |
|
678 | deltas, | |
677 | linkmapper, |
|
679 | linkmapper, | |
678 | transaction, |
|
680 | transaction, | |
679 | addrevisioncb=None, |
|
681 | addrevisioncb=None, | |
680 | duplicaterevisioncb=None, |
|
682 | duplicaterevisioncb=None, | |
681 | maybemissingparents=False, |
|
683 | maybemissingparents=False, | |
682 | ): |
|
684 | ): | |
683 | empty = True |
|
685 | empty = True | |
684 |
|
686 | |||
685 | for ( |
|
687 | for ( | |
686 | node, |
|
688 | node, | |
687 | p1, |
|
689 | p1, | |
688 | p2, |
|
690 | p2, | |
689 | linknode, |
|
691 | linknode, | |
690 | deltabase, |
|
692 | deltabase, | |
691 | delta, |
|
693 | delta, | |
692 | wireflags, |
|
694 | wireflags, | |
693 | sidedata, |
|
695 | sidedata, | |
694 | ) in deltas: |
|
696 | ) in deltas: | |
695 | storeflags = 0 |
|
697 | storeflags = 0 | |
696 |
|
698 | |||
697 | if wireflags & repository.REVISION_FLAG_CENSORED: |
|
699 | if wireflags & repository.REVISION_FLAG_CENSORED: | |
698 | storeflags |= FLAG_CENSORED |
|
700 | storeflags |= FLAG_CENSORED | |
699 |
|
701 | |||
700 | if wireflags & ~repository.REVISION_FLAG_CENSORED: |
|
702 | if wireflags & ~repository.REVISION_FLAG_CENSORED: | |
701 | raise SQLiteStoreError(b'unhandled revision flag') |
|
703 | raise SQLiteStoreError(b'unhandled revision flag') | |
702 |
|
704 | |||
703 | if maybemissingparents: |
|
705 | if maybemissingparents: | |
704 | if p1 != nullid and not self.hasnode(p1): |
|
706 | if p1 != nullid and not self.hasnode(p1): | |
705 | p1 = nullid |
|
707 | p1 = nullid | |
706 | storeflags |= FLAG_MISSING_P1 |
|
708 | storeflags |= FLAG_MISSING_P1 | |
707 |
|
709 | |||
708 | if p2 != nullid and not self.hasnode(p2): |
|
710 | if p2 != nullid and not self.hasnode(p2): | |
709 | p2 = nullid |
|
711 | p2 = nullid | |
710 | storeflags |= FLAG_MISSING_P2 |
|
712 | storeflags |= FLAG_MISSING_P2 | |
711 |
|
713 | |||
712 | baserev = self.rev(deltabase) |
|
714 | baserev = self.rev(deltabase) | |
713 |
|
715 | |||
714 | # If base is censored, delta must be full replacement in a single |
|
716 | # If base is censored, delta must be full replacement in a single | |
715 | # patch operation. |
|
717 | # patch operation. | |
716 | if baserev != nullrev and self.iscensored(baserev): |
|
718 | if baserev != nullrev and self.iscensored(baserev): | |
717 | hlen = struct.calcsize(b'>lll') |
|
719 | hlen = struct.calcsize(b'>lll') | |
718 | oldlen = len(self.rawdata(deltabase, _verifyhash=False)) |
|
720 | oldlen = len(self.rawdata(deltabase, _verifyhash=False)) | |
719 | newlen = len(delta) - hlen |
|
721 | newlen = len(delta) - hlen | |
720 |
|
722 | |||
721 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): |
|
723 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): | |
722 | raise error.CensoredBaseError(self._path, deltabase) |
|
724 | raise error.CensoredBaseError(self._path, deltabase) | |
723 |
|
725 | |||
724 | if not (storeflags & FLAG_CENSORED) and storageutil.deltaiscensored( |
|
726 | if not (storeflags & FLAG_CENSORED) and storageutil.deltaiscensored( | |
725 | delta, baserev, lambda x: len(self.rawdata(x)) |
|
727 | delta, baserev, lambda x: len(self.rawdata(x)) | |
726 | ): |
|
728 | ): | |
727 | storeflags |= FLAG_CENSORED |
|
729 | storeflags |= FLAG_CENSORED | |
728 |
|
730 | |||
729 | linkrev = linkmapper(linknode) |
|
731 | linkrev = linkmapper(linknode) | |
730 |
|
732 | |||
731 | if node in self._revisions: |
|
733 | if node in self._revisions: | |
732 | # Possibly reset parents to make them proper. |
|
734 | # Possibly reset parents to make them proper. | |
733 | entry = self._revisions[node] |
|
735 | entry = self._revisions[node] | |
734 |
|
736 | |||
735 | if entry.flags & FLAG_MISSING_P1 and p1 != nullid: |
|
737 | if entry.flags & FLAG_MISSING_P1 and p1 != nullid: | |
736 | entry.p1node = p1 |
|
738 | entry.p1node = p1 | |
737 | entry.p1rev = self._nodetorev[p1] |
|
739 | entry.p1rev = self._nodetorev[p1] | |
738 | entry.flags &= ~FLAG_MISSING_P1 |
|
740 | entry.flags &= ~FLAG_MISSING_P1 | |
739 |
|
741 | |||
740 | self._db.execute( |
|
742 | self._db.execute( | |
741 | 'UPDATE fileindex SET p1rev=?, flags=? WHERE id=?', |
|
743 | 'UPDATE fileindex SET p1rev=?, flags=? WHERE id=?', | |
742 | (self._nodetorev[p1], entry.flags, entry.rid), |
|
744 | (self._nodetorev[p1], entry.flags, entry.rid), | |
743 | ) |
|
745 | ) | |
744 |
|
746 | |||
745 | if entry.flags & FLAG_MISSING_P2 and p2 != nullid: |
|
747 | if entry.flags & FLAG_MISSING_P2 and p2 != nullid: | |
746 | entry.p2node = p2 |
|
748 | entry.p2node = p2 | |
747 | entry.p2rev = self._nodetorev[p2] |
|
749 | entry.p2rev = self._nodetorev[p2] | |
748 | entry.flags &= ~FLAG_MISSING_P2 |
|
750 | entry.flags &= ~FLAG_MISSING_P2 | |
749 |
|
751 | |||
750 | self._db.execute( |
|
752 | self._db.execute( | |
751 | 'UPDATE fileindex SET p2rev=?, flags=? WHERE id=?', |
|
753 | 'UPDATE fileindex SET p2rev=?, flags=? WHERE id=?', | |
752 | (self._nodetorev[p1], entry.flags, entry.rid), |
|
754 | (self._nodetorev[p1], entry.flags, entry.rid), | |
753 | ) |
|
755 | ) | |
754 |
|
756 | |||
755 | if duplicaterevisioncb: |
|
757 | if duplicaterevisioncb: | |
756 | duplicaterevisioncb(self, self.rev(node)) |
|
758 | duplicaterevisioncb(self, self.rev(node)) | |
757 | empty = False |
|
759 | empty = False | |
758 | continue |
|
760 | continue | |
759 |
|
761 | |||
760 | if deltabase == nullid: |
|
762 | if deltabase == nullid: | |
761 | text = mdiff.patch(b'', delta) |
|
763 | text = mdiff.patch(b'', delta) | |
762 | storedelta = None |
|
764 | storedelta = None | |
763 | else: |
|
765 | else: | |
764 | text = None |
|
766 | text = None | |
765 | storedelta = (deltabase, delta) |
|
767 | storedelta = (deltabase, delta) | |
766 |
|
768 | |||
767 | rev = self._addrawrevision( |
|
769 | rev = self._addrawrevision( | |
768 | node, |
|
770 | node, | |
769 | text, |
|
771 | text, | |
770 | transaction, |
|
772 | transaction, | |
771 | linkrev, |
|
773 | linkrev, | |
772 | p1, |
|
774 | p1, | |
773 | p2, |
|
775 | p2, | |
774 | storedelta=storedelta, |
|
776 | storedelta=storedelta, | |
775 | flags=storeflags, |
|
777 | flags=storeflags, | |
776 | ) |
|
778 | ) | |
777 |
|
779 | |||
778 | if addrevisioncb: |
|
780 | if addrevisioncb: | |
779 | addrevisioncb(self, rev) |
|
781 | addrevisioncb(self, rev) | |
780 | empty = False |
|
782 | empty = False | |
781 |
|
783 | |||
782 | return not empty |
|
784 | return not empty | |
783 |
|
785 | |||
784 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
786 | def censorrevision(self, tr, censornode, tombstone=b''): | |
785 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') |
|
787 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') | |
786 |
|
788 | |||
787 | # This restriction is cargo culted from revlogs and makes no sense for |
|
789 | # This restriction is cargo culted from revlogs and makes no sense for | |
788 | # SQLite, since columns can be resized at will. |
|
790 | # SQLite, since columns can be resized at will. | |
789 | if len(tombstone) > len(self.rawdata(censornode)): |
|
791 | if len(tombstone) > len(self.rawdata(censornode)): | |
790 | raise error.Abort( |
|
792 | raise error.Abort( | |
791 | _(b'censor tombstone must be no longer than censored data') |
|
793 | _(b'censor tombstone must be no longer than censored data') | |
792 | ) |
|
794 | ) | |
793 |
|
795 | |||
794 | # We need to replace the censored revision's data with the tombstone. |
|
796 | # We need to replace the censored revision's data with the tombstone. | |
795 | # But replacing that data will have implications for delta chains that |
|
797 | # But replacing that data will have implications for delta chains that | |
796 | # reference it. |
|
798 | # reference it. | |
797 | # |
|
799 | # | |
798 | # While "better," more complex strategies are possible, we do something |
|
800 | # While "better," more complex strategies are possible, we do something | |
799 | # simple: we find delta chain children of the censored revision and we |
|
801 | # simple: we find delta chain children of the censored revision and we | |
800 | # replace those incremental deltas with fulltexts of their corresponding |
|
802 | # replace those incremental deltas with fulltexts of their corresponding | |
801 | # revision. Then we delete the now-unreferenced delta and original |
|
803 | # revision. Then we delete the now-unreferenced delta and original | |
802 | # revision and insert a replacement. |
|
804 | # revision and insert a replacement. | |
803 |
|
805 | |||
804 | # Find the delta to be censored. |
|
806 | # Find the delta to be censored. | |
805 | censoreddeltaid = self._db.execute( |
|
807 | censoreddeltaid = self._db.execute( | |
806 | 'SELECT deltaid FROM fileindex WHERE id=?', |
|
808 | 'SELECT deltaid FROM fileindex WHERE id=?', | |
807 | (self._revisions[censornode].rid,), |
|
809 | (self._revisions[censornode].rid,), | |
808 | ).fetchone()[0] |
|
810 | ).fetchone()[0] | |
809 |
|
811 | |||
810 | # Find all its delta chain children. |
|
812 | # Find all its delta chain children. | |
811 | # TODO once we support storing deltas for !files, we'll need to look |
|
813 | # TODO once we support storing deltas for !files, we'll need to look | |
812 | # for those delta chains too. |
|
814 | # for those delta chains too. | |
813 | rows = list( |
|
815 | rows = list( | |
814 | self._db.execute( |
|
816 | self._db.execute( | |
815 | 'SELECT id, pathid, node FROM fileindex ' |
|
817 | 'SELECT id, pathid, node FROM fileindex ' | |
816 | 'WHERE deltabaseid=? OR deltaid=?', |
|
818 | 'WHERE deltabaseid=? OR deltaid=?', | |
817 | (censoreddeltaid, censoreddeltaid), |
|
819 | (censoreddeltaid, censoreddeltaid), | |
818 | ) |
|
820 | ) | |
819 | ) |
|
821 | ) | |
820 |
|
822 | |||
821 | for row in rows: |
|
823 | for row in rows: | |
822 | rid, pathid, node = row |
|
824 | rid, pathid, node = row | |
823 |
|
825 | |||
824 | fulltext = resolvedeltachain( |
|
826 | fulltext = resolvedeltachain( | |
825 | self._db, pathid, node, {}, {-1: None}, zstddctx=self._dctx |
|
827 | self._db, pathid, node, {}, {-1: None}, zstddctx=self._dctx | |
826 | ) |
|
828 | ) | |
827 |
|
829 | |||
828 | deltahash = hashutil.sha1(fulltext).digest() |
|
830 | deltahash = hashutil.sha1(fulltext).digest() | |
829 |
|
831 | |||
830 | if self._compengine == b'zstd': |
|
832 | if self._compengine == b'zstd': | |
831 | deltablob = self._cctx.compress(fulltext) |
|
833 | deltablob = self._cctx.compress(fulltext) | |
832 | compression = COMPRESSION_ZSTD |
|
834 | compression = COMPRESSION_ZSTD | |
833 | elif self._compengine == b'zlib': |
|
835 | elif self._compengine == b'zlib': | |
834 | deltablob = zlib.compress(fulltext) |
|
836 | deltablob = zlib.compress(fulltext) | |
835 | compression = COMPRESSION_ZLIB |
|
837 | compression = COMPRESSION_ZLIB | |
836 | elif self._compengine == b'none': |
|
838 | elif self._compengine == b'none': | |
837 | deltablob = fulltext |
|
839 | deltablob = fulltext | |
838 | compression = COMPRESSION_NONE |
|
840 | compression = COMPRESSION_NONE | |
839 | else: |
|
841 | else: | |
840 | raise error.ProgrammingError( |
|
842 | raise error.ProgrammingError( | |
841 | b'unhandled compression engine: %s' % self._compengine |
|
843 | b'unhandled compression engine: %s' % self._compengine | |
842 | ) |
|
844 | ) | |
843 |
|
845 | |||
844 | if len(deltablob) >= len(fulltext): |
|
846 | if len(deltablob) >= len(fulltext): | |
845 | deltablob = fulltext |
|
847 | deltablob = fulltext | |
846 | compression = COMPRESSION_NONE |
|
848 | compression = COMPRESSION_NONE | |
847 |
|
849 | |||
848 | deltaid = insertdelta(self._db, compression, deltahash, deltablob) |
|
850 | deltaid = insertdelta(self._db, compression, deltahash, deltablob) | |
849 |
|
851 | |||
850 | self._db.execute( |
|
852 | self._db.execute( | |
851 | 'UPDATE fileindex SET deltaid=?, deltabaseid=NULL ' |
|
853 | 'UPDATE fileindex SET deltaid=?, deltabaseid=NULL ' | |
852 | 'WHERE id=?', |
|
854 | 'WHERE id=?', | |
853 | (deltaid, rid), |
|
855 | (deltaid, rid), | |
854 | ) |
|
856 | ) | |
855 |
|
857 | |||
856 | # Now create the tombstone delta and replace the delta on the censored |
|
858 | # Now create the tombstone delta and replace the delta on the censored | |
857 | # node. |
|
859 | # node. | |
858 | deltahash = hashutil.sha1(tombstone).digest() |
|
860 | deltahash = hashutil.sha1(tombstone).digest() | |
859 | tombstonedeltaid = insertdelta( |
|
861 | tombstonedeltaid = insertdelta( | |
860 | self._db, COMPRESSION_NONE, deltahash, tombstone |
|
862 | self._db, COMPRESSION_NONE, deltahash, tombstone | |
861 | ) |
|
863 | ) | |
862 |
|
864 | |||
863 | flags = self._revisions[censornode].flags |
|
865 | flags = self._revisions[censornode].flags | |
864 | flags |= FLAG_CENSORED |
|
866 | flags |= FLAG_CENSORED | |
865 |
|
867 | |||
866 | self._db.execute( |
|
868 | self._db.execute( | |
867 | 'UPDATE fileindex SET flags=?, deltaid=?, deltabaseid=NULL ' |
|
869 | 'UPDATE fileindex SET flags=?, deltaid=?, deltabaseid=NULL ' | |
868 | 'WHERE pathid=? AND node=?', |
|
870 | 'WHERE pathid=? AND node=?', | |
869 | (flags, tombstonedeltaid, self._pathid, censornode), |
|
871 | (flags, tombstonedeltaid, self._pathid, censornode), | |
870 | ) |
|
872 | ) | |
871 |
|
873 | |||
872 | self._db.execute('DELETE FROM delta WHERE id=?', (censoreddeltaid,)) |
|
874 | self._db.execute('DELETE FROM delta WHERE id=?', (censoreddeltaid,)) | |
873 |
|
875 | |||
874 | self._refreshindex() |
|
876 | self._refreshindex() | |
875 | self._revisioncache.clear() |
|
877 | self._revisioncache.clear() | |
876 |
|
878 | |||
877 | def getstrippoint(self, minlink): |
|
879 | def getstrippoint(self, minlink): | |
878 | return storageutil.resolvestripinfo( |
|
880 | return storageutil.resolvestripinfo( | |
879 | minlink, |
|
881 | minlink, | |
880 | len(self) - 1, |
|
882 | len(self) - 1, | |
881 | [self.rev(n) for n in self.heads()], |
|
883 | [self.rev(n) for n in self.heads()], | |
882 | self.linkrev, |
|
884 | self.linkrev, | |
883 | self.parentrevs, |
|
885 | self.parentrevs, | |
884 | ) |
|
886 | ) | |
885 |
|
887 | |||
886 | def strip(self, minlink, transaction): |
|
888 | def strip(self, minlink, transaction): | |
887 | if not len(self): |
|
889 | if not len(self): | |
888 | return |
|
890 | return | |
889 |
|
891 | |||
890 | rev, _ignored = self.getstrippoint(minlink) |
|
892 | rev, _ignored = self.getstrippoint(minlink) | |
891 |
|
893 | |||
892 | if rev == len(self): |
|
894 | if rev == len(self): | |
893 | return |
|
895 | return | |
894 |
|
896 | |||
895 | for rev in self.revs(rev): |
|
897 | for rev in self.revs(rev): | |
896 | self._db.execute( |
|
898 | self._db.execute( | |
897 | 'DELETE FROM fileindex WHERE pathid=? AND node=?', |
|
899 | 'DELETE FROM fileindex WHERE pathid=? AND node=?', | |
898 | (self._pathid, self.node(rev)), |
|
900 | (self._pathid, self.node(rev)), | |
899 | ) |
|
901 | ) | |
900 |
|
902 | |||
901 | # TODO how should we garbage collect data in delta table? |
|
903 | # TODO how should we garbage collect data in delta table? | |
902 |
|
904 | |||
903 | self._refreshindex() |
|
905 | self._refreshindex() | |
904 |
|
906 | |||
905 | # End of ifilemutation interface. |
|
907 | # End of ifilemutation interface. | |
906 |
|
908 | |||
907 | # Start of ifilestorage interface. |
|
909 | # Start of ifilestorage interface. | |
908 |
|
910 | |||
909 | def files(self): |
|
911 | def files(self): | |
910 | return [] |
|
912 | return [] | |
911 |
|
913 | |||
912 | def sidedata(self, nodeorrev, _df=None): |
|
914 | def sidedata(self, nodeorrev, _df=None): | |
913 | # Not supported for now |
|
915 | # Not supported for now | |
914 | return {} |
|
916 | return {} | |
915 |
|
917 | |||
916 | def storageinfo( |
|
918 | def storageinfo( | |
917 | self, |
|
919 | self, | |
918 | exclusivefiles=False, |
|
920 | exclusivefiles=False, | |
919 | sharedfiles=False, |
|
921 | sharedfiles=False, | |
920 | revisionscount=False, |
|
922 | revisionscount=False, | |
921 | trackedsize=False, |
|
923 | trackedsize=False, | |
922 | storedsize=False, |
|
924 | storedsize=False, | |
923 | ): |
|
925 | ): | |
924 | d = {} |
|
926 | d = {} | |
925 |
|
927 | |||
926 | if exclusivefiles: |
|
928 | if exclusivefiles: | |
927 | d[b'exclusivefiles'] = [] |
|
929 | d[b'exclusivefiles'] = [] | |
928 |
|
930 | |||
929 | if sharedfiles: |
|
931 | if sharedfiles: | |
930 | # TODO list sqlite file(s) here. |
|
932 | # TODO list sqlite file(s) here. | |
931 | d[b'sharedfiles'] = [] |
|
933 | d[b'sharedfiles'] = [] | |
932 |
|
934 | |||
933 | if revisionscount: |
|
935 | if revisionscount: | |
934 | d[b'revisionscount'] = len(self) |
|
936 | d[b'revisionscount'] = len(self) | |
935 |
|
937 | |||
936 | if trackedsize: |
|
938 | if trackedsize: | |
937 | d[b'trackedsize'] = sum( |
|
939 | d[b'trackedsize'] = sum( | |
938 | len(self.revision(node)) for node in self._nodetorev |
|
940 | len(self.revision(node)) for node in self._nodetorev | |
939 | ) |
|
941 | ) | |
940 |
|
942 | |||
941 | if storedsize: |
|
943 | if storedsize: | |
942 | # TODO implement this? |
|
944 | # TODO implement this? | |
943 | d[b'storedsize'] = None |
|
945 | d[b'storedsize'] = None | |
944 |
|
946 | |||
945 | return d |
|
947 | return d | |
946 |
|
948 | |||
947 | def verifyintegrity(self, state): |
|
949 | def verifyintegrity(self, state): | |
948 | state[b'skipread'] = set() |
|
950 | state[b'skipread'] = set() | |
949 |
|
951 | |||
950 | for rev in self: |
|
952 | for rev in self: | |
951 | node = self.node(rev) |
|
953 | node = self.node(rev) | |
952 |
|
954 | |||
953 | try: |
|
955 | try: | |
954 | self.revision(node) |
|
956 | self.revision(node) | |
955 | except Exception as e: |
|
957 | except Exception as e: | |
956 | yield sqliteproblem( |
|
958 | yield sqliteproblem( | |
957 | error=_(b'unpacking %s: %s') % (short(node), e), node=node |
|
959 | error=_(b'unpacking %s: %s') % (short(node), e), node=node | |
958 | ) |
|
960 | ) | |
959 |
|
961 | |||
960 | state[b'skipread'].add(node) |
|
962 | state[b'skipread'].add(node) | |
961 |
|
963 | |||
962 | # End of ifilestorage interface. |
|
964 | # End of ifilestorage interface. | |
963 |
|
965 | |||
964 | def _checkhash(self, fulltext, node, p1=None, p2=None): |
|
966 | def _checkhash(self, fulltext, node, p1=None, p2=None): | |
965 | if p1 is None and p2 is None: |
|
967 | if p1 is None and p2 is None: | |
966 | p1, p2 = self.parents(node) |
|
968 | p1, p2 = self.parents(node) | |
967 |
|
969 | |||
968 | if node == storageutil.hashrevisionsha1(fulltext, p1, p2): |
|
970 | if node == storageutil.hashrevisionsha1(fulltext, p1, p2): | |
969 | return |
|
971 | return | |
970 |
|
972 | |||
971 | try: |
|
973 | try: | |
972 | del self._revisioncache[node] |
|
974 | del self._revisioncache[node] | |
973 | except KeyError: |
|
975 | except KeyError: | |
974 | pass |
|
976 | pass | |
975 |
|
977 | |||
976 | if storageutil.iscensoredtext(fulltext): |
|
978 | if storageutil.iscensoredtext(fulltext): | |
977 | raise error.CensoredNodeError(self._path, node, fulltext) |
|
979 | raise error.CensoredNodeError(self._path, node, fulltext) | |
978 |
|
980 | |||
979 | raise SQLiteStoreError(_(b'integrity check failed on %s') % self._path) |
|
981 | raise SQLiteStoreError(_(b'integrity check failed on %s') % self._path) | |
980 |
|
982 | |||
981 | def _addrawrevision( |
|
983 | def _addrawrevision( | |
982 | self, |
|
984 | self, | |
983 | node, |
|
985 | node, | |
984 | revisiondata, |
|
986 | revisiondata, | |
985 | transaction, |
|
987 | transaction, | |
986 | linkrev, |
|
988 | linkrev, | |
987 | p1, |
|
989 | p1, | |
988 | p2, |
|
990 | p2, | |
989 | storedelta=None, |
|
991 | storedelta=None, | |
990 | flags=0, |
|
992 | flags=0, | |
991 | ): |
|
993 | ): | |
992 | if self._pathid is None: |
|
994 | if self._pathid is None: | |
993 | res = self._db.execute( |
|
995 | res = self._db.execute( | |
994 | 'INSERT INTO filepath (path) VALUES (?)', (self._path,) |
|
996 | 'INSERT INTO filepath (path) VALUES (?)', (self._path,) | |
995 | ) |
|
997 | ) | |
996 | self._pathid = res.lastrowid |
|
998 | self._pathid = res.lastrowid | |
997 |
|
999 | |||
998 | # For simplicity, always store a delta against p1. |
|
1000 | # For simplicity, always store a delta against p1. | |
999 | # TODO we need a lot more logic here to make behavior reasonable. |
|
1001 | # TODO we need a lot more logic here to make behavior reasonable. | |
1000 |
|
1002 | |||
1001 | if storedelta: |
|
1003 | if storedelta: | |
1002 | deltabase, delta = storedelta |
|
1004 | deltabase, delta = storedelta | |
1003 |
|
1005 | |||
1004 | if isinstance(deltabase, int): |
|
1006 | if isinstance(deltabase, int): | |
1005 | deltabase = self.node(deltabase) |
|
1007 | deltabase = self.node(deltabase) | |
1006 |
|
1008 | |||
1007 | else: |
|
1009 | else: | |
1008 | assert revisiondata is not None |
|
1010 | assert revisiondata is not None | |
1009 | deltabase = p1 |
|
1011 | deltabase = p1 | |
1010 |
|
1012 | |||
1011 | if deltabase == nullid: |
|
1013 | if deltabase == nullid: | |
1012 | delta = revisiondata |
|
1014 | delta = revisiondata | |
1013 | else: |
|
1015 | else: | |
1014 | delta = mdiff.textdiff( |
|
1016 | delta = mdiff.textdiff( | |
1015 | self.revision(self.rev(deltabase)), revisiondata |
|
1017 | self.revision(self.rev(deltabase)), revisiondata | |
1016 | ) |
|
1018 | ) | |
1017 |
|
1019 | |||
1018 | # File index stores a pointer to its delta and the parent delta. |
|
1020 | # File index stores a pointer to its delta and the parent delta. | |
1019 | # The parent delta is stored via a pointer to the fileindex PK. |
|
1021 | # The parent delta is stored via a pointer to the fileindex PK. | |
1020 | if deltabase == nullid: |
|
1022 | if deltabase == nullid: | |
1021 | baseid = None |
|
1023 | baseid = None | |
1022 | else: |
|
1024 | else: | |
1023 | baseid = self._revisions[deltabase].rid |
|
1025 | baseid = self._revisions[deltabase].rid | |
1024 |
|
1026 | |||
1025 | # Deltas are stored with a hash of their content. This allows |
|
1027 | # Deltas are stored with a hash of their content. This allows | |
1026 | # us to de-duplicate. The table is configured to ignore conflicts |
|
1028 | # us to de-duplicate. The table is configured to ignore conflicts | |
1027 | # and it is faster to just insert and silently noop than to look |
|
1029 | # and it is faster to just insert and silently noop than to look | |
1028 | # first. |
|
1030 | # first. | |
1029 | deltahash = hashutil.sha1(delta).digest() |
|
1031 | deltahash = hashutil.sha1(delta).digest() | |
1030 |
|
1032 | |||
1031 | if self._compengine == b'zstd': |
|
1033 | if self._compengine == b'zstd': | |
1032 | deltablob = self._cctx.compress(delta) |
|
1034 | deltablob = self._cctx.compress(delta) | |
1033 | compression = COMPRESSION_ZSTD |
|
1035 | compression = COMPRESSION_ZSTD | |
1034 | elif self._compengine == b'zlib': |
|
1036 | elif self._compengine == b'zlib': | |
1035 | deltablob = zlib.compress(delta) |
|
1037 | deltablob = zlib.compress(delta) | |
1036 | compression = COMPRESSION_ZLIB |
|
1038 | compression = COMPRESSION_ZLIB | |
1037 | elif self._compengine == b'none': |
|
1039 | elif self._compengine == b'none': | |
1038 | deltablob = delta |
|
1040 | deltablob = delta | |
1039 | compression = COMPRESSION_NONE |
|
1041 | compression = COMPRESSION_NONE | |
1040 | else: |
|
1042 | else: | |
1041 | raise error.ProgrammingError( |
|
1043 | raise error.ProgrammingError( | |
1042 | b'unhandled compression engine: %s' % self._compengine |
|
1044 | b'unhandled compression engine: %s' % self._compengine | |
1043 | ) |
|
1045 | ) | |
1044 |
|
1046 | |||
1045 | # Don't store compressed data if it isn't practical. |
|
1047 | # Don't store compressed data if it isn't practical. | |
1046 | if len(deltablob) >= len(delta): |
|
1048 | if len(deltablob) >= len(delta): | |
1047 | deltablob = delta |
|
1049 | deltablob = delta | |
1048 | compression = COMPRESSION_NONE |
|
1050 | compression = COMPRESSION_NONE | |
1049 |
|
1051 | |||
1050 | deltaid = insertdelta(self._db, compression, deltahash, deltablob) |
|
1052 | deltaid = insertdelta(self._db, compression, deltahash, deltablob) | |
1051 |
|
1053 | |||
1052 | rev = len(self) |
|
1054 | rev = len(self) | |
1053 |
|
1055 | |||
1054 | if p1 == nullid: |
|
1056 | if p1 == nullid: | |
1055 | p1rev = nullrev |
|
1057 | p1rev = nullrev | |
1056 | else: |
|
1058 | else: | |
1057 | p1rev = self._nodetorev[p1] |
|
1059 | p1rev = self._nodetorev[p1] | |
1058 |
|
1060 | |||
1059 | if p2 == nullid: |
|
1061 | if p2 == nullid: | |
1060 | p2rev = nullrev |
|
1062 | p2rev = nullrev | |
1061 | else: |
|
1063 | else: | |
1062 | p2rev = self._nodetorev[p2] |
|
1064 | p2rev = self._nodetorev[p2] | |
1063 |
|
1065 | |||
1064 | rid = self._db.execute( |
|
1066 | rid = self._db.execute( | |
1065 | 'INSERT INTO fileindex (' |
|
1067 | 'INSERT INTO fileindex (' | |
1066 | ' pathid, revnum, node, p1rev, p2rev, linkrev, flags, ' |
|
1068 | ' pathid, revnum, node, p1rev, p2rev, linkrev, flags, ' | |
1067 | ' deltaid, deltabaseid) ' |
|
1069 | ' deltaid, deltabaseid) ' | |
1068 | ' VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)', |
|
1070 | ' VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)', | |
1069 | ( |
|
1071 | ( | |
1070 | self._pathid, |
|
1072 | self._pathid, | |
1071 | rev, |
|
1073 | rev, | |
1072 | node, |
|
1074 | node, | |
1073 | p1rev, |
|
1075 | p1rev, | |
1074 | p2rev, |
|
1076 | p2rev, | |
1075 | linkrev, |
|
1077 | linkrev, | |
1076 | flags, |
|
1078 | flags, | |
1077 | deltaid, |
|
1079 | deltaid, | |
1078 | baseid, |
|
1080 | baseid, | |
1079 | ), |
|
1081 | ), | |
1080 | ).lastrowid |
|
1082 | ).lastrowid | |
1081 |
|
1083 | |||
1082 | entry = revisionentry( |
|
1084 | entry = revisionentry( | |
1083 | rid=rid, |
|
1085 | rid=rid, | |
1084 | rev=rev, |
|
1086 | rev=rev, | |
1085 | node=node, |
|
1087 | node=node, | |
1086 | p1rev=p1rev, |
|
1088 | p1rev=p1rev, | |
1087 | p2rev=p2rev, |
|
1089 | p2rev=p2rev, | |
1088 | p1node=p1, |
|
1090 | p1node=p1, | |
1089 | p2node=p2, |
|
1091 | p2node=p2, | |
1090 | linkrev=linkrev, |
|
1092 | linkrev=linkrev, | |
1091 | flags=flags, |
|
1093 | flags=flags, | |
1092 | ) |
|
1094 | ) | |
1093 |
|
1095 | |||
1094 | self._nodetorev[node] = rev |
|
1096 | self._nodetorev[node] = rev | |
1095 | self._revtonode[rev] = node |
|
1097 | self._revtonode[rev] = node | |
1096 | self._revisions[node] = entry |
|
1098 | self._revisions[node] = entry | |
1097 |
|
1099 | |||
1098 | return rev |
|
1100 | return rev | |
1099 |
|
1101 | |||
1100 |
|
1102 | |||
1101 | class sqliterepository(localrepo.localrepository): |
|
1103 | class sqliterepository(localrepo.localrepository): | |
1102 | def cancopy(self): |
|
1104 | def cancopy(self): | |
1103 | return False |
|
1105 | return False | |
1104 |
|
1106 | |||
1105 | def transaction(self, *args, **kwargs): |
|
1107 | def transaction(self, *args, **kwargs): | |
1106 | current = self.currenttransaction() |
|
1108 | current = self.currenttransaction() | |
1107 |
|
1109 | |||
1108 | tr = super(sqliterepository, self).transaction(*args, **kwargs) |
|
1110 | tr = super(sqliterepository, self).transaction(*args, **kwargs) | |
1109 |
|
1111 | |||
1110 | if current: |
|
1112 | if current: | |
1111 | return tr |
|
1113 | return tr | |
1112 |
|
1114 | |||
1113 | self._dbconn.execute('BEGIN TRANSACTION') |
|
1115 | self._dbconn.execute('BEGIN TRANSACTION') | |
1114 |
|
1116 | |||
1115 | def committransaction(_): |
|
1117 | def committransaction(_): | |
1116 | self._dbconn.commit() |
|
1118 | self._dbconn.commit() | |
1117 |
|
1119 | |||
1118 | tr.addfinalize(b'sqlitestore', committransaction) |
|
1120 | tr.addfinalize(b'sqlitestore', committransaction) | |
1119 |
|
1121 | |||
1120 | return tr |
|
1122 | return tr | |
1121 |
|
1123 | |||
1122 | @property |
|
1124 | @property | |
1123 | def _dbconn(self): |
|
1125 | def _dbconn(self): | |
1124 | # SQLite connections can only be used on the thread that created |
|
1126 | # SQLite connections can only be used on the thread that created | |
1125 | # them. In most cases, this "just works." However, hgweb uses |
|
1127 | # them. In most cases, this "just works." However, hgweb uses | |
1126 | # multiple threads. |
|
1128 | # multiple threads. | |
1127 | tid = threading.current_thread().ident |
|
1129 | tid = threading.current_thread().ident | |
1128 |
|
1130 | |||
1129 | if self._db: |
|
1131 | if self._db: | |
1130 | if self._db[0] == tid: |
|
1132 | if self._db[0] == tid: | |
1131 | return self._db[1] |
|
1133 | return self._db[1] | |
1132 |
|
1134 | |||
1133 | db = makedb(self.svfs.join(b'db.sqlite')) |
|
1135 | db = makedb(self.svfs.join(b'db.sqlite')) | |
1134 | self._db = (tid, db) |
|
1136 | self._db = (tid, db) | |
1135 |
|
1137 | |||
1136 | return db |
|
1138 | return db | |
1137 |
|
1139 | |||
1138 |
|
1140 | |||
1139 | def makedb(path): |
|
1141 | def makedb(path): | |
1140 | """Construct a database handle for a database at path.""" |
|
1142 | """Construct a database handle for a database at path.""" | |
1141 |
|
1143 | |||
1142 | db = sqlite3.connect(encoding.strfromlocal(path)) |
|
1144 | db = sqlite3.connect(encoding.strfromlocal(path)) | |
1143 | db.text_factory = bytes |
|
1145 | db.text_factory = bytes | |
1144 |
|
1146 | |||
1145 | res = db.execute('PRAGMA user_version').fetchone()[0] |
|
1147 | res = db.execute('PRAGMA user_version').fetchone()[0] | |
1146 |
|
1148 | |||
1147 | # New database. |
|
1149 | # New database. | |
1148 | if res == 0: |
|
1150 | if res == 0: | |
1149 | for statement in CREATE_SCHEMA: |
|
1151 | for statement in CREATE_SCHEMA: | |
1150 | db.execute(statement) |
|
1152 | db.execute(statement) | |
1151 |
|
1153 | |||
1152 | db.commit() |
|
1154 | db.commit() | |
1153 |
|
1155 | |||
1154 | elif res == CURRENT_SCHEMA_VERSION: |
|
1156 | elif res == CURRENT_SCHEMA_VERSION: | |
1155 | pass |
|
1157 | pass | |
1156 |
|
1158 | |||
1157 | else: |
|
1159 | else: | |
1158 | raise error.Abort(_(b'sqlite database has unrecognized version')) |
|
1160 | raise error.Abort(_(b'sqlite database has unrecognized version')) | |
1159 |
|
1161 | |||
1160 | db.execute('PRAGMA journal_mode=WAL') |
|
1162 | db.execute('PRAGMA journal_mode=WAL') | |
1161 |
|
1163 | |||
1162 | return db |
|
1164 | return db | |
1163 |
|
1165 | |||
1164 |
|
1166 | |||
1165 | def featuresetup(ui, supported): |
|
1167 | def featuresetup(ui, supported): | |
1166 | supported.add(REQUIREMENT) |
|
1168 | supported.add(REQUIREMENT) | |
1167 |
|
1169 | |||
1168 | if zstd: |
|
1170 | if zstd: | |
1169 | supported.add(REQUIREMENT_ZSTD) |
|
1171 | supported.add(REQUIREMENT_ZSTD) | |
1170 |
|
1172 | |||
1171 | supported.add(REQUIREMENT_ZLIB) |
|
1173 | supported.add(REQUIREMENT_ZLIB) | |
1172 | supported.add(REQUIREMENT_NONE) |
|
1174 | supported.add(REQUIREMENT_NONE) | |
1173 | supported.add(REQUIREMENT_SHALLOW_FILES) |
|
1175 | supported.add(REQUIREMENT_SHALLOW_FILES) | |
1174 | supported.add(requirements.NARROW_REQUIREMENT) |
|
1176 | supported.add(requirements.NARROW_REQUIREMENT) | |
1175 |
|
1177 | |||
1176 |
|
1178 | |||
1177 | def newreporequirements(orig, ui, createopts): |
|
1179 | def newreporequirements(orig, ui, createopts): | |
1178 | if createopts[b'backend'] != b'sqlite': |
|
1180 | if createopts[b'backend'] != b'sqlite': | |
1179 | return orig(ui, createopts) |
|
1181 | return orig(ui, createopts) | |
1180 |
|
1182 | |||
1181 | # This restriction can be lifted once we have more confidence. |
|
1183 | # This restriction can be lifted once we have more confidence. | |
1182 | if b'sharedrepo' in createopts: |
|
1184 | if b'sharedrepo' in createopts: | |
1183 | raise error.Abort( |
|
1185 | raise error.Abort( | |
1184 | _(b'shared repositories not supported with SQLite store') |
|
1186 | _(b'shared repositories not supported with SQLite store') | |
1185 | ) |
|
1187 | ) | |
1186 |
|
1188 | |||
1187 | # This filtering is out of an abundance of caution: we want to ensure |
|
1189 | # This filtering is out of an abundance of caution: we want to ensure | |
1188 | # we honor creation options and we do that by annotating exactly the |
|
1190 | # we honor creation options and we do that by annotating exactly the | |
1189 | # creation options we recognize. |
|
1191 | # creation options we recognize. | |
1190 | known = { |
|
1192 | known = { | |
1191 | b'narrowfiles', |
|
1193 | b'narrowfiles', | |
1192 | b'backend', |
|
1194 | b'backend', | |
1193 | b'shallowfilestore', |
|
1195 | b'shallowfilestore', | |
1194 | } |
|
1196 | } | |
1195 |
|
1197 | |||
1196 | unsupported = set(createopts) - known |
|
1198 | unsupported = set(createopts) - known | |
1197 | if unsupported: |
|
1199 | if unsupported: | |
1198 | raise error.Abort( |
|
1200 | raise error.Abort( | |
1199 | _(b'SQLite store does not support repo creation option: %s') |
|
1201 | _(b'SQLite store does not support repo creation option: %s') | |
1200 | % b', '.join(sorted(unsupported)) |
|
1202 | % b', '.join(sorted(unsupported)) | |
1201 | ) |
|
1203 | ) | |
1202 |
|
1204 | |||
1203 | # Since we're a hybrid store that still relies on revlogs, we fall back |
|
1205 | # Since we're a hybrid store that still relies on revlogs, we fall back | |
1204 | # to using the revlogv1 backend's storage requirements then adding our |
|
1206 | # to using the revlogv1 backend's storage requirements then adding our | |
1205 | # own requirement. |
|
1207 | # own requirement. | |
1206 | createopts[b'backend'] = b'revlogv1' |
|
1208 | createopts[b'backend'] = b'revlogv1' | |
1207 | requirements = orig(ui, createopts) |
|
1209 | requirements = orig(ui, createopts) | |
1208 | requirements.add(REQUIREMENT) |
|
1210 | requirements.add(REQUIREMENT) | |
1209 |
|
1211 | |||
1210 | compression = ui.config(b'storage', b'sqlite.compression') |
|
1212 | compression = ui.config(b'storage', b'sqlite.compression') | |
1211 |
|
1213 | |||
1212 | if compression == b'zstd' and not zstd: |
|
1214 | if compression == b'zstd' and not zstd: | |
1213 | raise error.Abort( |
|
1215 | raise error.Abort( | |
1214 | _( |
|
1216 | _( | |
1215 | b'storage.sqlite.compression set to "zstd" but ' |
|
1217 | b'storage.sqlite.compression set to "zstd" but ' | |
1216 | b'zstandard compression not available to this ' |
|
1218 | b'zstandard compression not available to this ' | |
1217 | b'Mercurial install' |
|
1219 | b'Mercurial install' | |
1218 | ) |
|
1220 | ) | |
1219 | ) |
|
1221 | ) | |
1220 |
|
1222 | |||
1221 | if compression == b'zstd': |
|
1223 | if compression == b'zstd': | |
1222 | requirements.add(REQUIREMENT_ZSTD) |
|
1224 | requirements.add(REQUIREMENT_ZSTD) | |
1223 | elif compression == b'zlib': |
|
1225 | elif compression == b'zlib': | |
1224 | requirements.add(REQUIREMENT_ZLIB) |
|
1226 | requirements.add(REQUIREMENT_ZLIB) | |
1225 | elif compression == b'none': |
|
1227 | elif compression == b'none': | |
1226 | requirements.add(REQUIREMENT_NONE) |
|
1228 | requirements.add(REQUIREMENT_NONE) | |
1227 | else: |
|
1229 | else: | |
1228 | raise error.Abort( |
|
1230 | raise error.Abort( | |
1229 | _( |
|
1231 | _( | |
1230 | b'unknown compression engine defined in ' |
|
1232 | b'unknown compression engine defined in ' | |
1231 | b'storage.sqlite.compression: %s' |
|
1233 | b'storage.sqlite.compression: %s' | |
1232 | ) |
|
1234 | ) | |
1233 | % compression |
|
1235 | % compression | |
1234 | ) |
|
1236 | ) | |
1235 |
|
1237 | |||
1236 | if createopts.get(b'shallowfilestore'): |
|
1238 | if createopts.get(b'shallowfilestore'): | |
1237 | requirements.add(REQUIREMENT_SHALLOW_FILES) |
|
1239 | requirements.add(REQUIREMENT_SHALLOW_FILES) | |
1238 |
|
1240 | |||
1239 | return requirements |
|
1241 | return requirements | |
1240 |
|
1242 | |||
1241 |
|
1243 | |||
1242 | @interfaceutil.implementer(repository.ilocalrepositoryfilestorage) |
|
1244 | @interfaceutil.implementer(repository.ilocalrepositoryfilestorage) | |
1243 | class sqlitefilestorage(object): |
|
1245 | class sqlitefilestorage(object): | |
1244 | """Repository file storage backed by SQLite.""" |
|
1246 | """Repository file storage backed by SQLite.""" | |
1245 |
|
1247 | |||
1246 | def file(self, path): |
|
1248 | def file(self, path): | |
1247 | if path[0] == b'/': |
|
1249 | if path[0] == b'/': | |
1248 | path = path[1:] |
|
1250 | path = path[1:] | |
1249 |
|
1251 | |||
1250 | if REQUIREMENT_ZSTD in self.requirements: |
|
1252 | if REQUIREMENT_ZSTD in self.requirements: | |
1251 | compression = b'zstd' |
|
1253 | compression = b'zstd' | |
1252 | elif REQUIREMENT_ZLIB in self.requirements: |
|
1254 | elif REQUIREMENT_ZLIB in self.requirements: | |
1253 | compression = b'zlib' |
|
1255 | compression = b'zlib' | |
1254 | elif REQUIREMENT_NONE in self.requirements: |
|
1256 | elif REQUIREMENT_NONE in self.requirements: | |
1255 | compression = b'none' |
|
1257 | compression = b'none' | |
1256 | else: |
|
1258 | else: | |
1257 | raise error.Abort( |
|
1259 | raise error.Abort( | |
1258 | _( |
|
1260 | _( | |
1259 | b'unable to determine what compression engine ' |
|
1261 | b'unable to determine what compression engine ' | |
1260 | b'to use for SQLite storage' |
|
1262 | b'to use for SQLite storage' | |
1261 | ) |
|
1263 | ) | |
1262 | ) |
|
1264 | ) | |
1263 |
|
1265 | |||
1264 | return sqlitefilestore(self._dbconn, path, compression) |
|
1266 | return sqlitefilestore(self._dbconn, path, compression) | |
1265 |
|
1267 | |||
1266 |
|
1268 | |||
1267 | def makefilestorage(orig, requirements, features, **kwargs): |
|
1269 | def makefilestorage(orig, requirements, features, **kwargs): | |
1268 | """Produce a type conforming to ``ilocalrepositoryfilestorage``.""" |
|
1270 | """Produce a type conforming to ``ilocalrepositoryfilestorage``.""" | |
1269 | if REQUIREMENT in requirements: |
|
1271 | if REQUIREMENT in requirements: | |
1270 | if REQUIREMENT_SHALLOW_FILES in requirements: |
|
1272 | if REQUIREMENT_SHALLOW_FILES in requirements: | |
1271 | features.add(repository.REPO_FEATURE_SHALLOW_FILE_STORAGE) |
|
1273 | features.add(repository.REPO_FEATURE_SHALLOW_FILE_STORAGE) | |
1272 |
|
1274 | |||
1273 | return sqlitefilestorage |
|
1275 | return sqlitefilestorage | |
1274 | else: |
|
1276 | else: | |
1275 | return orig(requirements=requirements, features=features, **kwargs) |
|
1277 | return orig(requirements=requirements, features=features, **kwargs) | |
1276 |
|
1278 | |||
1277 |
|
1279 | |||
1278 | def makemain(orig, ui, requirements, **kwargs): |
|
1280 | def makemain(orig, ui, requirements, **kwargs): | |
1279 | if REQUIREMENT in requirements: |
|
1281 | if REQUIREMENT in requirements: | |
1280 | if REQUIREMENT_ZSTD in requirements and not zstd: |
|
1282 | if REQUIREMENT_ZSTD in requirements and not zstd: | |
1281 | raise error.Abort( |
|
1283 | raise error.Abort( | |
1282 | _( |
|
1284 | _( | |
1283 | b'repository uses zstandard compression, which ' |
|
1285 | b'repository uses zstandard compression, which ' | |
1284 | b'is not available to this Mercurial install' |
|
1286 | b'is not available to this Mercurial install' | |
1285 | ) |
|
1287 | ) | |
1286 | ) |
|
1288 | ) | |
1287 |
|
1289 | |||
1288 | return sqliterepository |
|
1290 | return sqliterepository | |
1289 |
|
1291 | |||
1290 | return orig(requirements=requirements, **kwargs) |
|
1292 | return orig(requirements=requirements, **kwargs) | |
1291 |
|
1293 | |||
1292 |
|
1294 | |||
1293 | def verifierinit(orig, self, *args, **kwargs): |
|
1295 | def verifierinit(orig, self, *args, **kwargs): | |
1294 | orig(self, *args, **kwargs) |
|
1296 | orig(self, *args, **kwargs) | |
1295 |
|
1297 | |||
1296 | # We don't care that files in the store don't align with what is |
|
1298 | # We don't care that files in the store don't align with what is | |
1297 | # advertised. So suppress these warnings. |
|
1299 | # advertised. So suppress these warnings. | |
1298 | self.warnorphanstorefiles = False |
|
1300 | self.warnorphanstorefiles = False | |
1299 |
|
1301 | |||
1300 |
|
1302 | |||
1301 | def extsetup(ui): |
|
1303 | def extsetup(ui): | |
1302 | localrepo.featuresetupfuncs.add(featuresetup) |
|
1304 | localrepo.featuresetupfuncs.add(featuresetup) | |
1303 | extensions.wrapfunction( |
|
1305 | extensions.wrapfunction( | |
1304 | localrepo, b'newreporequirements', newreporequirements |
|
1306 | localrepo, b'newreporequirements', newreporequirements | |
1305 | ) |
|
1307 | ) | |
1306 | extensions.wrapfunction(localrepo, b'makefilestorage', makefilestorage) |
|
1308 | extensions.wrapfunction(localrepo, b'makefilestorage', makefilestorage) | |
1307 | extensions.wrapfunction(localrepo, b'makemain', makemain) |
|
1309 | extensions.wrapfunction(localrepo, b'makemain', makemain) | |
1308 | extensions.wrapfunction(verify.verifier, b'__init__', verifierinit) |
|
1310 | extensions.wrapfunction(verify.verifier, b'__init__', verifierinit) | |
1309 |
|
1311 | |||
1310 |
|
1312 | |||
1311 | def reposetup(ui, repo): |
|
1313 | def reposetup(ui, repo): | |
1312 | if isinstance(repo, sqliterepository): |
|
1314 | if isinstance(repo, sqliterepository): | |
1313 | repo._db = None |
|
1315 | repo._db = None | |
1314 |
|
1316 | |||
1315 | # TODO check for bundlerepository? |
|
1317 | # TODO check for bundlerepository? |
@@ -1,1794 +1,1861 b'' | |||||
1 | # changegroup.py - Mercurial changegroup manipulation functions |
|
1 | # changegroup.py - Mercurial changegroup manipulation functions | |
2 | # |
|
2 | # | |
3 | # Copyright 2006 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2006 Matt Mackall <mpm@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import os |
|
10 | import os | |
11 | import struct |
|
11 | import struct | |
12 | import weakref |
|
12 | import weakref | |
13 |
|
13 | |||
14 | from .i18n import _ |
|
14 | from .i18n import _ | |
15 | from .node import ( |
|
15 | from .node import ( | |
16 | hex, |
|
16 | hex, | |
17 | nullid, |
|
17 | nullid, | |
18 | nullrev, |
|
18 | nullrev, | |
19 | short, |
|
19 | short, | |
20 | ) |
|
20 | ) | |
21 | from .pycompat import open |
|
21 | from .pycompat import open | |
22 |
|
22 | |||
23 | from . import ( |
|
23 | from . import ( | |
24 | error, |
|
24 | error, | |
25 | match as matchmod, |
|
25 | match as matchmod, | |
26 | mdiff, |
|
26 | mdiff, | |
27 | phases, |
|
27 | phases, | |
28 | pycompat, |
|
28 | pycompat, | |
29 | requirements, |
|
29 | requirements, | |
30 | scmutil, |
|
30 | scmutil, | |
31 | util, |
|
31 | util, | |
32 | ) |
|
32 | ) | |
33 |
|
33 | |||
34 | from .interfaces import repository |
|
34 | from .interfaces import repository | |
35 | from .revlogutils import sidedata as sidedatamod |
|
35 | from .revlogutils import sidedata as sidedatamod | |
36 |
|
36 | |||
37 | _CHANGEGROUPV1_DELTA_HEADER = struct.Struct(b"20s20s20s20s") |
|
37 | _CHANGEGROUPV1_DELTA_HEADER = struct.Struct(b"20s20s20s20s") | |
38 | _CHANGEGROUPV2_DELTA_HEADER = struct.Struct(b"20s20s20s20s20s") |
|
38 | _CHANGEGROUPV2_DELTA_HEADER = struct.Struct(b"20s20s20s20s20s") | |
39 | _CHANGEGROUPV3_DELTA_HEADER = struct.Struct(b">20s20s20s20s20sH") |
|
39 | _CHANGEGROUPV3_DELTA_HEADER = struct.Struct(b">20s20s20s20s20sH") | |
40 |
|
40 | |||
41 | LFS_REQUIREMENT = b'lfs' |
|
41 | LFS_REQUIREMENT = b'lfs' | |
42 |
|
42 | |||
43 | readexactly = util.readexactly |
|
43 | readexactly = util.readexactly | |
44 |
|
44 | |||
45 |
|
45 | |||
46 | def getchunk(stream): |
|
46 | def getchunk(stream): | |
47 | """return the next chunk from stream as a string""" |
|
47 | """return the next chunk from stream as a string""" | |
48 | d = readexactly(stream, 4) |
|
48 | d = readexactly(stream, 4) | |
49 | l = struct.unpack(b">l", d)[0] |
|
49 | l = struct.unpack(b">l", d)[0] | |
50 | if l <= 4: |
|
50 | if l <= 4: | |
51 | if l: |
|
51 | if l: | |
52 | raise error.Abort(_(b"invalid chunk length %d") % l) |
|
52 | raise error.Abort(_(b"invalid chunk length %d") % l) | |
53 | return b"" |
|
53 | return b"" | |
54 | return readexactly(stream, l - 4) |
|
54 | return readexactly(stream, l - 4) | |
55 |
|
55 | |||
56 |
|
56 | |||
57 | def chunkheader(length): |
|
57 | def chunkheader(length): | |
58 | """return a changegroup chunk header (string)""" |
|
58 | """return a changegroup chunk header (string)""" | |
59 | return struct.pack(b">l", length + 4) |
|
59 | return struct.pack(b">l", length + 4) | |
60 |
|
60 | |||
61 |
|
61 | |||
62 | def closechunk(): |
|
62 | def closechunk(): | |
63 | """return a changegroup chunk header (string) for a zero-length chunk""" |
|
63 | """return a changegroup chunk header (string) for a zero-length chunk""" | |
64 | return struct.pack(b">l", 0) |
|
64 | return struct.pack(b">l", 0) | |
65 |
|
65 | |||
66 |
|
66 | |||
67 | def _fileheader(path): |
|
67 | def _fileheader(path): | |
68 | """Obtain a changegroup chunk header for a named path.""" |
|
68 | """Obtain a changegroup chunk header for a named path.""" | |
69 | return chunkheader(len(path)) + path |
|
69 | return chunkheader(len(path)) + path | |
70 |
|
70 | |||
71 |
|
71 | |||
72 | def writechunks(ui, chunks, filename, vfs=None): |
|
72 | def writechunks(ui, chunks, filename, vfs=None): | |
73 | """Write chunks to a file and return its filename. |
|
73 | """Write chunks to a file and return its filename. | |
74 |
|
74 | |||
75 | The stream is assumed to be a bundle file. |
|
75 | The stream is assumed to be a bundle file. | |
76 | Existing files will not be overwritten. |
|
76 | Existing files will not be overwritten. | |
77 | If no filename is specified, a temporary file is created. |
|
77 | If no filename is specified, a temporary file is created. | |
78 | """ |
|
78 | """ | |
79 | fh = None |
|
79 | fh = None | |
80 | cleanup = None |
|
80 | cleanup = None | |
81 | try: |
|
81 | try: | |
82 | if filename: |
|
82 | if filename: | |
83 | if vfs: |
|
83 | if vfs: | |
84 | fh = vfs.open(filename, b"wb") |
|
84 | fh = vfs.open(filename, b"wb") | |
85 | else: |
|
85 | else: | |
86 | # Increase default buffer size because default is usually |
|
86 | # Increase default buffer size because default is usually | |
87 | # small (4k is common on Linux). |
|
87 | # small (4k is common on Linux). | |
88 | fh = open(filename, b"wb", 131072) |
|
88 | fh = open(filename, b"wb", 131072) | |
89 | else: |
|
89 | else: | |
90 | fd, filename = pycompat.mkstemp(prefix=b"hg-bundle-", suffix=b".hg") |
|
90 | fd, filename = pycompat.mkstemp(prefix=b"hg-bundle-", suffix=b".hg") | |
91 | fh = os.fdopen(fd, "wb") |
|
91 | fh = os.fdopen(fd, "wb") | |
92 | cleanup = filename |
|
92 | cleanup = filename | |
93 | for c in chunks: |
|
93 | for c in chunks: | |
94 | fh.write(c) |
|
94 | fh.write(c) | |
95 | cleanup = None |
|
95 | cleanup = None | |
96 | return filename |
|
96 | return filename | |
97 | finally: |
|
97 | finally: | |
98 | if fh is not None: |
|
98 | if fh is not None: | |
99 | fh.close() |
|
99 | fh.close() | |
100 | if cleanup is not None: |
|
100 | if cleanup is not None: | |
101 | if filename and vfs: |
|
101 | if filename and vfs: | |
102 | vfs.unlink(cleanup) |
|
102 | vfs.unlink(cleanup) | |
103 | else: |
|
103 | else: | |
104 | os.unlink(cleanup) |
|
104 | os.unlink(cleanup) | |
105 |
|
105 | |||
106 |
|
106 | |||
107 | class cg1unpacker(object): |
|
107 | class cg1unpacker(object): | |
108 | """Unpacker for cg1 changegroup streams. |
|
108 | """Unpacker for cg1 changegroup streams. | |
109 |
|
109 | |||
110 | A changegroup unpacker handles the framing of the revision data in |
|
110 | A changegroup unpacker handles the framing of the revision data in | |
111 | the wire format. Most consumers will want to use the apply() |
|
111 | the wire format. Most consumers will want to use the apply() | |
112 | method to add the changes from the changegroup to a repository. |
|
112 | method to add the changes from the changegroup to a repository. | |
113 |
|
113 | |||
114 | If you're forwarding a changegroup unmodified to another consumer, |
|
114 | If you're forwarding a changegroup unmodified to another consumer, | |
115 | use getchunks(), which returns an iterator of changegroup |
|
115 | use getchunks(), which returns an iterator of changegroup | |
116 | chunks. This is mostly useful for cases where you need to know the |
|
116 | chunks. This is mostly useful for cases where you need to know the | |
117 | data stream has ended by observing the end of the changegroup. |
|
117 | data stream has ended by observing the end of the changegroup. | |
118 |
|
118 | |||
119 | deltachunk() is useful only if you're applying delta data. Most |
|
119 | deltachunk() is useful only if you're applying delta data. Most | |
120 | consumers should prefer apply() instead. |
|
120 | consumers should prefer apply() instead. | |
121 |
|
121 | |||
122 | A few other public methods exist. Those are used only for |
|
122 | A few other public methods exist. Those are used only for | |
123 | bundlerepo and some debug commands - their use is discouraged. |
|
123 | bundlerepo and some debug commands - their use is discouraged. | |
124 | """ |
|
124 | """ | |
125 |
|
125 | |||
126 | deltaheader = _CHANGEGROUPV1_DELTA_HEADER |
|
126 | deltaheader = _CHANGEGROUPV1_DELTA_HEADER | |
127 | deltaheadersize = deltaheader.size |
|
127 | deltaheadersize = deltaheader.size | |
128 | version = b'01' |
|
128 | version = b'01' | |
129 | _grouplistcount = 1 # One list of files after the manifests |
|
129 | _grouplistcount = 1 # One list of files after the manifests | |
130 |
|
130 | |||
131 | def __init__(self, fh, alg, extras=None): |
|
131 | def __init__(self, fh, alg, extras=None): | |
132 | if alg is None: |
|
132 | if alg is None: | |
133 | alg = b'UN' |
|
133 | alg = b'UN' | |
134 | if alg not in util.compengines.supportedbundletypes: |
|
134 | if alg not in util.compengines.supportedbundletypes: | |
135 | raise error.Abort(_(b'unknown stream compression type: %s') % alg) |
|
135 | raise error.Abort(_(b'unknown stream compression type: %s') % alg) | |
136 | if alg == b'BZ': |
|
136 | if alg == b'BZ': | |
137 | alg = b'_truncatedBZ' |
|
137 | alg = b'_truncatedBZ' | |
138 |
|
138 | |||
139 | compengine = util.compengines.forbundletype(alg) |
|
139 | compengine = util.compengines.forbundletype(alg) | |
140 | self._stream = compengine.decompressorreader(fh) |
|
140 | self._stream = compengine.decompressorreader(fh) | |
141 | self._type = alg |
|
141 | self._type = alg | |
142 | self.extras = extras or {} |
|
142 | self.extras = extras or {} | |
143 | self.callback = None |
|
143 | self.callback = None | |
144 |
|
144 | |||
145 | # These methods (compressed, read, seek, tell) all appear to only |
|
145 | # These methods (compressed, read, seek, tell) all appear to only | |
146 | # be used by bundlerepo, but it's a little hard to tell. |
|
146 | # be used by bundlerepo, but it's a little hard to tell. | |
147 | def compressed(self): |
|
147 | def compressed(self): | |
148 | return self._type is not None and self._type != b'UN' |
|
148 | return self._type is not None and self._type != b'UN' | |
149 |
|
149 | |||
150 | def read(self, l): |
|
150 | def read(self, l): | |
151 | return self._stream.read(l) |
|
151 | return self._stream.read(l) | |
152 |
|
152 | |||
153 | def seek(self, pos): |
|
153 | def seek(self, pos): | |
154 | return self._stream.seek(pos) |
|
154 | return self._stream.seek(pos) | |
155 |
|
155 | |||
156 | def tell(self): |
|
156 | def tell(self): | |
157 | return self._stream.tell() |
|
157 | return self._stream.tell() | |
158 |
|
158 | |||
159 | def close(self): |
|
159 | def close(self): | |
160 | return self._stream.close() |
|
160 | return self._stream.close() | |
161 |
|
161 | |||
162 | def _chunklength(self): |
|
162 | def _chunklength(self): | |
163 | d = readexactly(self._stream, 4) |
|
163 | d = readexactly(self._stream, 4) | |
164 | l = struct.unpack(b">l", d)[0] |
|
164 | l = struct.unpack(b">l", d)[0] | |
165 | if l <= 4: |
|
165 | if l <= 4: | |
166 | if l: |
|
166 | if l: | |
167 | raise error.Abort(_(b"invalid chunk length %d") % l) |
|
167 | raise error.Abort(_(b"invalid chunk length %d") % l) | |
168 | return 0 |
|
168 | return 0 | |
169 | if self.callback: |
|
169 | if self.callback: | |
170 | self.callback() |
|
170 | self.callback() | |
171 | return l - 4 |
|
171 | return l - 4 | |
172 |
|
172 | |||
173 | def changelogheader(self): |
|
173 | def changelogheader(self): | |
174 | """v10 does not have a changelog header chunk""" |
|
174 | """v10 does not have a changelog header chunk""" | |
175 | return {} |
|
175 | return {} | |
176 |
|
176 | |||
177 | def manifestheader(self): |
|
177 | def manifestheader(self): | |
178 | """v10 does not have a manifest header chunk""" |
|
178 | """v10 does not have a manifest header chunk""" | |
179 | return {} |
|
179 | return {} | |
180 |
|
180 | |||
181 | def filelogheader(self): |
|
181 | def filelogheader(self): | |
182 | """return the header of the filelogs chunk, v10 only has the filename""" |
|
182 | """return the header of the filelogs chunk, v10 only has the filename""" | |
183 | l = self._chunklength() |
|
183 | l = self._chunklength() | |
184 | if not l: |
|
184 | if not l: | |
185 | return {} |
|
185 | return {} | |
186 | fname = readexactly(self._stream, l) |
|
186 | fname = readexactly(self._stream, l) | |
187 | return {b'filename': fname} |
|
187 | return {b'filename': fname} | |
188 |
|
188 | |||
189 | def _deltaheader(self, headertuple, prevnode): |
|
189 | def _deltaheader(self, headertuple, prevnode): | |
190 | node, p1, p2, cs = headertuple |
|
190 | node, p1, p2, cs = headertuple | |
191 | if prevnode is None: |
|
191 | if prevnode is None: | |
192 | deltabase = p1 |
|
192 | deltabase = p1 | |
193 | else: |
|
193 | else: | |
194 | deltabase = prevnode |
|
194 | deltabase = prevnode | |
195 | flags = 0 |
|
195 | flags = 0 | |
196 | return node, p1, p2, deltabase, cs, flags |
|
196 | return node, p1, p2, deltabase, cs, flags | |
197 |
|
197 | |||
198 | def deltachunk(self, prevnode): |
|
198 | def deltachunk(self, prevnode): | |
199 | l = self._chunklength() |
|
199 | l = self._chunklength() | |
200 | if not l: |
|
200 | if not l: | |
201 | return {} |
|
201 | return {} | |
202 | headerdata = readexactly(self._stream, self.deltaheadersize) |
|
202 | headerdata = readexactly(self._stream, self.deltaheadersize) | |
203 | header = self.deltaheader.unpack(headerdata) |
|
203 | header = self.deltaheader.unpack(headerdata) | |
204 | delta = readexactly(self._stream, l - self.deltaheadersize) |
|
204 | delta = readexactly(self._stream, l - self.deltaheadersize) | |
205 | node, p1, p2, deltabase, cs, flags = self._deltaheader(header, prevnode) |
|
205 | node, p1, p2, deltabase, cs, flags = self._deltaheader(header, prevnode) | |
206 | # cg4 forward-compat |
|
206 | # cg4 forward-compat | |
207 | sidedata = {} |
|
207 | sidedata = {} | |
208 | return (node, p1, p2, cs, deltabase, delta, flags, sidedata) |
|
208 | return (node, p1, p2, cs, deltabase, delta, flags, sidedata) | |
209 |
|
209 | |||
210 | def getchunks(self): |
|
210 | def getchunks(self): | |
211 | """returns all the chunks contains in the bundle |
|
211 | """returns all the chunks contains in the bundle | |
212 |
|
212 | |||
213 | Used when you need to forward the binary stream to a file or another |
|
213 | Used when you need to forward the binary stream to a file or another | |
214 | network API. To do so, it parse the changegroup data, otherwise it will |
|
214 | network API. To do so, it parse the changegroup data, otherwise it will | |
215 | block in case of sshrepo because it don't know the end of the stream. |
|
215 | block in case of sshrepo because it don't know the end of the stream. | |
216 | """ |
|
216 | """ | |
217 | # For changegroup 1 and 2, we expect 3 parts: changelog, manifestlog, |
|
217 | # For changegroup 1 and 2, we expect 3 parts: changelog, manifestlog, | |
218 | # and a list of filelogs. For changegroup 3, we expect 4 parts: |
|
218 | # and a list of filelogs. For changegroup 3, we expect 4 parts: | |
219 | # changelog, manifestlog, a list of tree manifestlogs, and a list of |
|
219 | # changelog, manifestlog, a list of tree manifestlogs, and a list of | |
220 | # filelogs. |
|
220 | # filelogs. | |
221 | # |
|
221 | # | |
222 | # Changelog and manifestlog parts are terminated with empty chunks. The |
|
222 | # Changelog and manifestlog parts are terminated with empty chunks. The | |
223 | # tree and file parts are a list of entry sections. Each entry section |
|
223 | # tree and file parts are a list of entry sections. Each entry section | |
224 | # is a series of chunks terminating in an empty chunk. The list of these |
|
224 | # is a series of chunks terminating in an empty chunk. The list of these | |
225 | # entry sections is terminated in yet another empty chunk, so we know |
|
225 | # entry sections is terminated in yet another empty chunk, so we know | |
226 | # we've reached the end of the tree/file list when we reach an empty |
|
226 | # we've reached the end of the tree/file list when we reach an empty | |
227 | # chunk that was proceeded by no non-empty chunks. |
|
227 | # chunk that was proceeded by no non-empty chunks. | |
228 |
|
228 | |||
229 | parts = 0 |
|
229 | parts = 0 | |
230 | while parts < 2 + self._grouplistcount: |
|
230 | while parts < 2 + self._grouplistcount: | |
231 | noentries = True |
|
231 | noentries = True | |
232 | while True: |
|
232 | while True: | |
233 | chunk = getchunk(self) |
|
233 | chunk = getchunk(self) | |
234 | if not chunk: |
|
234 | if not chunk: | |
235 | # The first two empty chunks represent the end of the |
|
235 | # The first two empty chunks represent the end of the | |
236 | # changelog and the manifestlog portions. The remaining |
|
236 | # changelog and the manifestlog portions. The remaining | |
237 | # empty chunks represent either A) the end of individual |
|
237 | # empty chunks represent either A) the end of individual | |
238 | # tree or file entries in the file list, or B) the end of |
|
238 | # tree or file entries in the file list, or B) the end of | |
239 | # the entire list. It's the end of the entire list if there |
|
239 | # the entire list. It's the end of the entire list if there | |
240 | # were no entries (i.e. noentries is True). |
|
240 | # were no entries (i.e. noentries is True). | |
241 | if parts < 2: |
|
241 | if parts < 2: | |
242 | parts += 1 |
|
242 | parts += 1 | |
243 | elif noentries: |
|
243 | elif noentries: | |
244 | parts += 1 |
|
244 | parts += 1 | |
245 | break |
|
245 | break | |
246 | noentries = False |
|
246 | noentries = False | |
247 | yield chunkheader(len(chunk)) |
|
247 | yield chunkheader(len(chunk)) | |
248 | pos = 0 |
|
248 | pos = 0 | |
249 | while pos < len(chunk): |
|
249 | while pos < len(chunk): | |
250 | next = pos + 2 ** 20 |
|
250 | next = pos + 2 ** 20 | |
251 | yield chunk[pos:next] |
|
251 | yield chunk[pos:next] | |
252 | pos = next |
|
252 | pos = next | |
253 | yield closechunk() |
|
253 | yield closechunk() | |
254 |
|
254 | |||
255 | def _unpackmanifests(self, repo, revmap, trp, prog): |
|
255 | def _unpackmanifests(self, repo, revmap, trp, prog): | |
256 | self.callback = prog.increment |
|
256 | self.callback = prog.increment | |
257 | # no need to check for empty manifest group here: |
|
257 | # no need to check for empty manifest group here: | |
258 | # if the result of the merge of 1 and 2 is the same in 3 and 4, |
|
258 | # if the result of the merge of 1 and 2 is the same in 3 and 4, | |
259 | # no new manifest will be created and the manifest group will |
|
259 | # no new manifest will be created and the manifest group will | |
260 | # be empty during the pull |
|
260 | # be empty during the pull | |
261 | self.manifestheader() |
|
261 | self.manifestheader() | |
262 | deltas = self.deltaiter() |
|
262 | deltas = self.deltaiter() | |
263 | repo.manifestlog.getstorage(b'').addgroup(deltas, revmap, trp) |
|
263 | repo.manifestlog.getstorage(b'').addgroup(deltas, revmap, trp) | |
264 | prog.complete() |
|
264 | prog.complete() | |
265 | self.callback = None |
|
265 | self.callback = None | |
266 |
|
266 | |||
267 | def apply( |
|
267 | def apply( | |
268 | self, |
|
268 | self, | |
269 | repo, |
|
269 | repo, | |
270 | tr, |
|
270 | tr, | |
271 | srctype, |
|
271 | srctype, | |
272 | url, |
|
272 | url, | |
273 | targetphase=phases.draft, |
|
273 | targetphase=phases.draft, | |
274 | expectedtotal=None, |
|
274 | expectedtotal=None, | |
|
275 | sidedata_categories=None, | |||
275 | ): |
|
276 | ): | |
276 | """Add the changegroup returned by source.read() to this repo. |
|
277 | """Add the changegroup returned by source.read() to this repo. | |
277 | srctype is a string like 'push', 'pull', or 'unbundle'. url is |
|
278 | srctype is a string like 'push', 'pull', or 'unbundle'. url is | |
278 | the URL of the repo where this changegroup is coming from. |
|
279 | the URL of the repo where this changegroup is coming from. | |
279 |
|
280 | |||
280 | Return an integer summarizing the change to this repo: |
|
281 | Return an integer summarizing the change to this repo: | |
281 | - nothing changed or no source: 0 |
|
282 | - nothing changed or no source: 0 | |
282 | - more heads than before: 1+added heads (2..n) |
|
283 | - more heads than before: 1+added heads (2..n) | |
283 | - fewer heads than before: -1-removed heads (-2..-n) |
|
284 | - fewer heads than before: -1-removed heads (-2..-n) | |
284 | - number of heads stays the same: 1 |
|
285 | - number of heads stays the same: 1 | |
|
286 | ||||
|
287 | `sidedata_categories` is an optional set of the remote's sidedata wanted | |||
|
288 | categories. | |||
285 | """ |
|
289 | """ | |
286 | repo = repo.unfiltered() |
|
290 | repo = repo.unfiltered() | |
287 |
|
291 | |||
|
292 | # Only useful if we're adding sidedata categories. If both peers have | |||
|
293 | # the same categories, then we simply don't do anything. | |||
|
294 | if self.version == b'04' and srctype == b'pull': | |||
|
295 | sidedata_helpers = get_sidedata_helpers( | |||
|
296 | repo, | |||
|
297 | sidedata_categories or set(), | |||
|
298 | pull=True, | |||
|
299 | ) | |||
|
300 | else: | |||
|
301 | sidedata_helpers = None | |||
|
302 | ||||
288 | def csmap(x): |
|
303 | def csmap(x): | |
289 | repo.ui.debug(b"add changeset %s\n" % short(x)) |
|
304 | repo.ui.debug(b"add changeset %s\n" % short(x)) | |
290 | return len(cl) |
|
305 | return len(cl) | |
291 |
|
306 | |||
292 | def revmap(x): |
|
307 | def revmap(x): | |
293 | return cl.rev(x) |
|
308 | return cl.rev(x) | |
294 |
|
309 | |||
295 | try: |
|
310 | try: | |
296 | # The transaction may already carry source information. In this |
|
311 | # The transaction may already carry source information. In this | |
297 | # case we use the top level data. We overwrite the argument |
|
312 | # case we use the top level data. We overwrite the argument | |
298 | # because we need to use the top level value (if they exist) |
|
313 | # because we need to use the top level value (if they exist) | |
299 | # in this function. |
|
314 | # in this function. | |
300 | srctype = tr.hookargs.setdefault(b'source', srctype) |
|
315 | srctype = tr.hookargs.setdefault(b'source', srctype) | |
301 | tr.hookargs.setdefault(b'url', url) |
|
316 | tr.hookargs.setdefault(b'url', url) | |
302 | repo.hook( |
|
317 | repo.hook( | |
303 | b'prechangegroup', throw=True, **pycompat.strkwargs(tr.hookargs) |
|
318 | b'prechangegroup', throw=True, **pycompat.strkwargs(tr.hookargs) | |
304 | ) |
|
319 | ) | |
305 |
|
320 | |||
306 | # write changelog data to temp files so concurrent readers |
|
321 | # write changelog data to temp files so concurrent readers | |
307 | # will not see an inconsistent view |
|
322 | # will not see an inconsistent view | |
308 | cl = repo.changelog |
|
323 | cl = repo.changelog | |
309 | cl.delayupdate(tr) |
|
324 | cl.delayupdate(tr) | |
310 | oldheads = set(cl.heads()) |
|
325 | oldheads = set(cl.heads()) | |
311 |
|
326 | |||
312 | trp = weakref.proxy(tr) |
|
327 | trp = weakref.proxy(tr) | |
313 | # pull off the changeset group |
|
328 | # pull off the changeset group | |
314 | repo.ui.status(_(b"adding changesets\n")) |
|
329 | repo.ui.status(_(b"adding changesets\n")) | |
315 | clstart = len(cl) |
|
330 | clstart = len(cl) | |
316 | progress = repo.ui.makeprogress( |
|
331 | progress = repo.ui.makeprogress( | |
317 | _(b'changesets'), unit=_(b'chunks'), total=expectedtotal |
|
332 | _(b'changesets'), unit=_(b'chunks'), total=expectedtotal | |
318 | ) |
|
333 | ) | |
319 | self.callback = progress.increment |
|
334 | self.callback = progress.increment | |
320 |
|
335 | |||
321 | efilesset = set() |
|
336 | efilesset = set() | |
322 | duprevs = [] |
|
337 | duprevs = [] | |
323 |
|
338 | |||
324 | def ondupchangelog(cl, rev): |
|
339 | def ondupchangelog(cl, rev): | |
325 | if rev < clstart: |
|
340 | if rev < clstart: | |
326 | duprevs.append(rev) |
|
341 | duprevs.append(rev) | |
327 |
|
342 | |||
328 | def onchangelog(cl, rev): |
|
343 | def onchangelog(cl, rev): | |
329 | ctx = cl.changelogrevision(rev) |
|
344 | ctx = cl.changelogrevision(rev) | |
330 | efilesset.update(ctx.files) |
|
345 | efilesset.update(ctx.files) | |
331 | repo.register_changeset(rev, ctx) |
|
346 | repo.register_changeset(rev, ctx) | |
332 |
|
347 | |||
333 | self.changelogheader() |
|
348 | self.changelogheader() | |
334 | deltas = self.deltaiter() |
|
349 | deltas = self.deltaiter() | |
335 | if not cl.addgroup( |
|
350 | if not cl.addgroup( | |
336 | deltas, |
|
351 | deltas, | |
337 | csmap, |
|
352 | csmap, | |
338 | trp, |
|
353 | trp, | |
339 | alwayscache=True, |
|
354 | alwayscache=True, | |
340 | addrevisioncb=onchangelog, |
|
355 | addrevisioncb=onchangelog, | |
341 | duplicaterevisioncb=ondupchangelog, |
|
356 | duplicaterevisioncb=ondupchangelog, | |
342 | ): |
|
357 | ): | |
343 | repo.ui.develwarn( |
|
358 | repo.ui.develwarn( | |
344 | b'applied empty changelog from changegroup', |
|
359 | b'applied empty changelog from changegroup', | |
345 | config=b'warn-empty-changegroup', |
|
360 | config=b'warn-empty-changegroup', | |
346 | ) |
|
361 | ) | |
347 | efiles = len(efilesset) |
|
362 | efiles = len(efilesset) | |
348 | clend = len(cl) |
|
363 | clend = len(cl) | |
349 | changesets = clend - clstart |
|
364 | changesets = clend - clstart | |
350 | progress.complete() |
|
365 | progress.complete() | |
351 | del deltas |
|
366 | del deltas | |
352 | # TODO Python 2.7 removal |
|
367 | # TODO Python 2.7 removal | |
353 | # del efilesset |
|
368 | # del efilesset | |
354 | efilesset = None |
|
369 | efilesset = None | |
355 | self.callback = None |
|
370 | self.callback = None | |
356 |
|
371 | |||
357 | # pull off the manifest group |
|
372 | # pull off the manifest group | |
358 | repo.ui.status(_(b"adding manifests\n")) |
|
373 | repo.ui.status(_(b"adding manifests\n")) | |
359 | # We know that we'll never have more manifests than we had |
|
374 | # We know that we'll never have more manifests than we had | |
360 | # changesets. |
|
375 | # changesets. | |
361 | progress = repo.ui.makeprogress( |
|
376 | progress = repo.ui.makeprogress( | |
362 | _(b'manifests'), unit=_(b'chunks'), total=changesets |
|
377 | _(b'manifests'), unit=_(b'chunks'), total=changesets | |
363 | ) |
|
378 | ) | |
364 | self._unpackmanifests(repo, revmap, trp, progress) |
|
379 | self._unpackmanifests(repo, revmap, trp, progress) | |
365 |
|
380 | |||
366 | needfiles = {} |
|
381 | needfiles = {} | |
367 | if repo.ui.configbool(b'server', b'validate'): |
|
382 | if repo.ui.configbool(b'server', b'validate'): | |
368 | cl = repo.changelog |
|
383 | cl = repo.changelog | |
369 | ml = repo.manifestlog |
|
384 | ml = repo.manifestlog | |
370 | # validate incoming csets have their manifests |
|
385 | # validate incoming csets have their manifests | |
371 | for cset in pycompat.xrange(clstart, clend): |
|
386 | for cset in pycompat.xrange(clstart, clend): | |
372 | mfnode = cl.changelogrevision(cset).manifest |
|
387 | mfnode = cl.changelogrevision(cset).manifest | |
373 | mfest = ml[mfnode].readdelta() |
|
388 | mfest = ml[mfnode].readdelta() | |
374 | # store file nodes we must see |
|
389 | # store file nodes we must see | |
375 | for f, n in pycompat.iteritems(mfest): |
|
390 | for f, n in pycompat.iteritems(mfest): | |
376 | needfiles.setdefault(f, set()).add(n) |
|
391 | needfiles.setdefault(f, set()).add(n) | |
377 |
|
392 | |||
378 | # process the files |
|
393 | # process the files | |
379 | repo.ui.status(_(b"adding file changes\n")) |
|
394 | repo.ui.status(_(b"adding file changes\n")) | |
380 | newrevs, newfiles = _addchangegroupfiles( |
|
395 | newrevs, newfiles = _addchangegroupfiles( | |
381 | repo, self, revmap, trp, efiles, needfiles |
|
396 | repo, self, revmap, trp, efiles, needfiles | |
382 | ) |
|
397 | ) | |
383 |
|
398 | |||
384 | # making sure the value exists |
|
399 | # making sure the value exists | |
385 | tr.changes.setdefault(b'changegroup-count-changesets', 0) |
|
400 | tr.changes.setdefault(b'changegroup-count-changesets', 0) | |
386 | tr.changes.setdefault(b'changegroup-count-revisions', 0) |
|
401 | tr.changes.setdefault(b'changegroup-count-revisions', 0) | |
387 | tr.changes.setdefault(b'changegroup-count-files', 0) |
|
402 | tr.changes.setdefault(b'changegroup-count-files', 0) | |
388 | tr.changes.setdefault(b'changegroup-count-heads', 0) |
|
403 | tr.changes.setdefault(b'changegroup-count-heads', 0) | |
389 |
|
404 | |||
390 | # some code use bundle operation for internal purpose. They usually |
|
405 | # some code use bundle operation for internal purpose. They usually | |
391 | # set `ui.quiet` to do this outside of user sight. Size the report |
|
406 | # set `ui.quiet` to do this outside of user sight. Size the report | |
392 | # of such operation now happens at the end of the transaction, that |
|
407 | # of such operation now happens at the end of the transaction, that | |
393 | # ui.quiet has not direct effect on the output. |
|
408 | # ui.quiet has not direct effect on the output. | |
394 | # |
|
409 | # | |
395 | # To preserve this intend use an inelegant hack, we fail to report |
|
410 | # To preserve this intend use an inelegant hack, we fail to report | |
396 | # the change if `quiet` is set. We should probably move to |
|
411 | # the change if `quiet` is set. We should probably move to | |
397 | # something better, but this is a good first step to allow the "end |
|
412 | # something better, but this is a good first step to allow the "end | |
398 | # of transaction report" to pass tests. |
|
413 | # of transaction report" to pass tests. | |
399 | if not repo.ui.quiet: |
|
414 | if not repo.ui.quiet: | |
400 | tr.changes[b'changegroup-count-changesets'] += changesets |
|
415 | tr.changes[b'changegroup-count-changesets'] += changesets | |
401 | tr.changes[b'changegroup-count-revisions'] += newrevs |
|
416 | tr.changes[b'changegroup-count-revisions'] += newrevs | |
402 | tr.changes[b'changegroup-count-files'] += newfiles |
|
417 | tr.changes[b'changegroup-count-files'] += newfiles | |
403 |
|
418 | |||
404 | deltaheads = 0 |
|
419 | deltaheads = 0 | |
405 | if oldheads: |
|
420 | if oldheads: | |
406 | heads = cl.heads() |
|
421 | heads = cl.heads() | |
407 | deltaheads += len(heads) - len(oldheads) |
|
422 | deltaheads += len(heads) - len(oldheads) | |
408 | for h in heads: |
|
423 | for h in heads: | |
409 | if h not in oldheads and repo[h].closesbranch(): |
|
424 | if h not in oldheads and repo[h].closesbranch(): | |
410 | deltaheads -= 1 |
|
425 | deltaheads -= 1 | |
411 |
|
426 | |||
412 | # see previous comment about checking ui.quiet |
|
427 | # see previous comment about checking ui.quiet | |
413 | if not repo.ui.quiet: |
|
428 | if not repo.ui.quiet: | |
414 | tr.changes[b'changegroup-count-heads'] += deltaheads |
|
429 | tr.changes[b'changegroup-count-heads'] += deltaheads | |
415 | repo.invalidatevolatilesets() |
|
430 | repo.invalidatevolatilesets() | |
416 |
|
431 | |||
417 | if changesets > 0: |
|
432 | if changesets > 0: | |
418 | if b'node' not in tr.hookargs: |
|
433 | if b'node' not in tr.hookargs: | |
419 | tr.hookargs[b'node'] = hex(cl.node(clstart)) |
|
434 | tr.hookargs[b'node'] = hex(cl.node(clstart)) | |
420 | tr.hookargs[b'node_last'] = hex(cl.node(clend - 1)) |
|
435 | tr.hookargs[b'node_last'] = hex(cl.node(clend - 1)) | |
421 | hookargs = dict(tr.hookargs) |
|
436 | hookargs = dict(tr.hookargs) | |
422 | else: |
|
437 | else: | |
423 | hookargs = dict(tr.hookargs) |
|
438 | hookargs = dict(tr.hookargs) | |
424 | hookargs[b'node'] = hex(cl.node(clstart)) |
|
439 | hookargs[b'node'] = hex(cl.node(clstart)) | |
425 | hookargs[b'node_last'] = hex(cl.node(clend - 1)) |
|
440 | hookargs[b'node_last'] = hex(cl.node(clend - 1)) | |
426 | repo.hook( |
|
441 | repo.hook( | |
427 | b'pretxnchangegroup', |
|
442 | b'pretxnchangegroup', | |
428 | throw=True, |
|
443 | throw=True, | |
429 | **pycompat.strkwargs(hookargs) |
|
444 | **pycompat.strkwargs(hookargs) | |
430 | ) |
|
445 | ) | |
431 |
|
446 | |||
432 | added = pycompat.xrange(clstart, clend) |
|
447 | added = pycompat.xrange(clstart, clend) | |
433 | phaseall = None |
|
448 | phaseall = None | |
434 | if srctype in (b'push', b'serve'): |
|
449 | if srctype in (b'push', b'serve'): | |
435 | # Old servers can not push the boundary themselves. |
|
450 | # Old servers can not push the boundary themselves. | |
436 | # New servers won't push the boundary if changeset already |
|
451 | # New servers won't push the boundary if changeset already | |
437 | # exists locally as secret |
|
452 | # exists locally as secret | |
438 | # |
|
453 | # | |
439 | # We should not use added here but the list of all change in |
|
454 | # We should not use added here but the list of all change in | |
440 | # the bundle |
|
455 | # the bundle | |
441 | if repo.publishing(): |
|
456 | if repo.publishing(): | |
442 | targetphase = phaseall = phases.public |
|
457 | targetphase = phaseall = phases.public | |
443 | else: |
|
458 | else: | |
444 | # closer target phase computation |
|
459 | # closer target phase computation | |
445 |
|
460 | |||
446 | # Those changesets have been pushed from the |
|
461 | # Those changesets have been pushed from the | |
447 | # outside, their phases are going to be pushed |
|
462 | # outside, their phases are going to be pushed | |
448 | # alongside. Therefor `targetphase` is |
|
463 | # alongside. Therefor `targetphase` is | |
449 | # ignored. |
|
464 | # ignored. | |
450 | targetphase = phaseall = phases.draft |
|
465 | targetphase = phaseall = phases.draft | |
451 | if added: |
|
466 | if added: | |
452 | phases.registernew(repo, tr, targetphase, added) |
|
467 | phases.registernew(repo, tr, targetphase, added) | |
453 | if phaseall is not None: |
|
468 | if phaseall is not None: | |
454 | if duprevs: |
|
469 | if duprevs: | |
455 | duprevs.extend(added) |
|
470 | duprevs.extend(added) | |
456 | else: |
|
471 | else: | |
457 | duprevs = added |
|
472 | duprevs = added | |
458 | phases.advanceboundary(repo, tr, phaseall, [], revs=duprevs) |
|
473 | phases.advanceboundary(repo, tr, phaseall, [], revs=duprevs) | |
459 | duprevs = [] |
|
474 | duprevs = [] | |
460 |
|
475 | |||
461 | if changesets > 0: |
|
476 | if changesets > 0: | |
462 |
|
477 | |||
463 | def runhooks(unused_success): |
|
478 | def runhooks(unused_success): | |
464 | # These hooks run when the lock releases, not when the |
|
479 | # These hooks run when the lock releases, not when the | |
465 | # transaction closes. So it's possible for the changelog |
|
480 | # transaction closes. So it's possible for the changelog | |
466 | # to have changed since we last saw it. |
|
481 | # to have changed since we last saw it. | |
467 | if clstart >= len(repo): |
|
482 | if clstart >= len(repo): | |
468 | return |
|
483 | return | |
469 |
|
484 | |||
470 | repo.hook(b"changegroup", **pycompat.strkwargs(hookargs)) |
|
485 | repo.hook(b"changegroup", **pycompat.strkwargs(hookargs)) | |
471 |
|
486 | |||
472 | for rev in added: |
|
487 | for rev in added: | |
473 | args = hookargs.copy() |
|
488 | args = hookargs.copy() | |
474 | args[b'node'] = hex(cl.node(rev)) |
|
489 | args[b'node'] = hex(cl.node(rev)) | |
475 | del args[b'node_last'] |
|
490 | del args[b'node_last'] | |
476 | repo.hook(b"incoming", **pycompat.strkwargs(args)) |
|
491 | repo.hook(b"incoming", **pycompat.strkwargs(args)) | |
477 |
|
492 | |||
478 | newheads = [h for h in repo.heads() if h not in oldheads] |
|
493 | newheads = [h for h in repo.heads() if h not in oldheads] | |
479 | repo.ui.log( |
|
494 | repo.ui.log( | |
480 | b"incoming", |
|
495 | b"incoming", | |
481 | b"%d incoming changes - new heads: %s\n", |
|
496 | b"%d incoming changes - new heads: %s\n", | |
482 | len(added), |
|
497 | len(added), | |
483 | b', '.join([hex(c[:6]) for c in newheads]), |
|
498 | b', '.join([hex(c[:6]) for c in newheads]), | |
484 | ) |
|
499 | ) | |
485 |
|
500 | |||
486 | tr.addpostclose( |
|
501 | tr.addpostclose( | |
487 | b'changegroup-runhooks-%020i' % clstart, |
|
502 | b'changegroup-runhooks-%020i' % clstart, | |
488 | lambda tr: repo._afterlock(runhooks), |
|
503 | lambda tr: repo._afterlock(runhooks), | |
489 | ) |
|
504 | ) | |
490 | finally: |
|
505 | finally: | |
491 | repo.ui.flush() |
|
506 | repo.ui.flush() | |
492 | # never return 0 here: |
|
507 | # never return 0 here: | |
493 | if deltaheads < 0: |
|
508 | if deltaheads < 0: | |
494 | ret = deltaheads - 1 |
|
509 | ret = deltaheads - 1 | |
495 | else: |
|
510 | else: | |
496 | ret = deltaheads + 1 |
|
511 | ret = deltaheads + 1 | |
497 | return ret |
|
512 | return ret | |
498 |
|
513 | |||
499 | def deltaiter(self): |
|
514 | def deltaiter(self): | |
500 | """ |
|
515 | """ | |
501 | returns an iterator of the deltas in this changegroup |
|
516 | returns an iterator of the deltas in this changegroup | |
502 |
|
517 | |||
503 | Useful for passing to the underlying storage system to be stored. |
|
518 | Useful for passing to the underlying storage system to be stored. | |
504 | """ |
|
519 | """ | |
505 | chain = None |
|
520 | chain = None | |
506 | for chunkdata in iter(lambda: self.deltachunk(chain), {}): |
|
521 | for chunkdata in iter(lambda: self.deltachunk(chain), {}): | |
507 | # Chunkdata: (node, p1, p2, cs, deltabase, delta, flags) |
|
522 | # Chunkdata: (node, p1, p2, cs, deltabase, delta, flags) | |
508 | yield chunkdata |
|
523 | yield chunkdata | |
509 | chain = chunkdata[0] |
|
524 | chain = chunkdata[0] | |
510 |
|
525 | |||
511 |
|
526 | |||
512 | class cg2unpacker(cg1unpacker): |
|
527 | class cg2unpacker(cg1unpacker): | |
513 | """Unpacker for cg2 streams. |
|
528 | """Unpacker for cg2 streams. | |
514 |
|
529 | |||
515 | cg2 streams add support for generaldelta, so the delta header |
|
530 | cg2 streams add support for generaldelta, so the delta header | |
516 | format is slightly different. All other features about the data |
|
531 | format is slightly different. All other features about the data | |
517 | remain the same. |
|
532 | remain the same. | |
518 | """ |
|
533 | """ | |
519 |
|
534 | |||
520 | deltaheader = _CHANGEGROUPV2_DELTA_HEADER |
|
535 | deltaheader = _CHANGEGROUPV2_DELTA_HEADER | |
521 | deltaheadersize = deltaheader.size |
|
536 | deltaheadersize = deltaheader.size | |
522 | version = b'02' |
|
537 | version = b'02' | |
523 |
|
538 | |||
524 | def _deltaheader(self, headertuple, prevnode): |
|
539 | def _deltaheader(self, headertuple, prevnode): | |
525 | node, p1, p2, deltabase, cs = headertuple |
|
540 | node, p1, p2, deltabase, cs = headertuple | |
526 | flags = 0 |
|
541 | flags = 0 | |
527 | return node, p1, p2, deltabase, cs, flags |
|
542 | return node, p1, p2, deltabase, cs, flags | |
528 |
|
543 | |||
529 |
|
544 | |||
530 | class cg3unpacker(cg2unpacker): |
|
545 | class cg3unpacker(cg2unpacker): | |
531 | """Unpacker for cg3 streams. |
|
546 | """Unpacker for cg3 streams. | |
532 |
|
547 | |||
533 | cg3 streams add support for exchanging treemanifests and revlog |
|
548 | cg3 streams add support for exchanging treemanifests and revlog | |
534 | flags. It adds the revlog flags to the delta header and an empty chunk |
|
549 | flags. It adds the revlog flags to the delta header and an empty chunk | |
535 | separating manifests and files. |
|
550 | separating manifests and files. | |
536 | """ |
|
551 | """ | |
537 |
|
552 | |||
538 | deltaheader = _CHANGEGROUPV3_DELTA_HEADER |
|
553 | deltaheader = _CHANGEGROUPV3_DELTA_HEADER | |
539 | deltaheadersize = deltaheader.size |
|
554 | deltaheadersize = deltaheader.size | |
540 | version = b'03' |
|
555 | version = b'03' | |
541 | _grouplistcount = 2 # One list of manifests and one list of files |
|
556 | _grouplistcount = 2 # One list of manifests and one list of files | |
542 |
|
557 | |||
543 | def _deltaheader(self, headertuple, prevnode): |
|
558 | def _deltaheader(self, headertuple, prevnode): | |
544 | node, p1, p2, deltabase, cs, flags = headertuple |
|
559 | node, p1, p2, deltabase, cs, flags = headertuple | |
545 | return node, p1, p2, deltabase, cs, flags |
|
560 | return node, p1, p2, deltabase, cs, flags | |
546 |
|
561 | |||
547 | def _unpackmanifests(self, repo, revmap, trp, prog): |
|
562 | def _unpackmanifests(self, repo, revmap, trp, prog): | |
548 | super(cg3unpacker, self)._unpackmanifests(repo, revmap, trp, prog) |
|
563 | super(cg3unpacker, self)._unpackmanifests(repo, revmap, trp, prog) | |
549 | for chunkdata in iter(self.filelogheader, {}): |
|
564 | for chunkdata in iter(self.filelogheader, {}): | |
550 | # If we get here, there are directory manifests in the changegroup |
|
565 | # If we get here, there are directory manifests in the changegroup | |
551 | d = chunkdata[b"filename"] |
|
566 | d = chunkdata[b"filename"] | |
552 | repo.ui.debug(b"adding %s revisions\n" % d) |
|
567 | repo.ui.debug(b"adding %s revisions\n" % d) | |
553 | deltas = self.deltaiter() |
|
568 | deltas = self.deltaiter() | |
554 | if not repo.manifestlog.getstorage(d).addgroup(deltas, revmap, trp): |
|
569 | if not repo.manifestlog.getstorage(d).addgroup(deltas, revmap, trp): | |
555 | raise error.Abort(_(b"received dir revlog group is empty")) |
|
570 | raise error.Abort(_(b"received dir revlog group is empty")) | |
556 |
|
571 | |||
557 |
|
572 | |||
558 | class cg4unpacker(cg3unpacker): |
|
573 | class cg4unpacker(cg3unpacker): | |
559 | """Unpacker for cg4 streams. |
|
574 | """Unpacker for cg4 streams. | |
560 |
|
575 | |||
561 | cg4 streams add support for exchanging sidedata. |
|
576 | cg4 streams add support for exchanging sidedata. | |
562 | """ |
|
577 | """ | |
563 |
|
578 | |||
564 | version = b'04' |
|
579 | version = b'04' | |
565 |
|
580 | |||
566 | def deltachunk(self, prevnode): |
|
581 | def deltachunk(self, prevnode): | |
567 | res = super(cg4unpacker, self).deltachunk(prevnode) |
|
582 | res = super(cg4unpacker, self).deltachunk(prevnode) | |
568 | if not res: |
|
583 | if not res: | |
569 | return res |
|
584 | return res | |
570 |
|
585 | |||
571 | (node, p1, p2, cs, deltabase, delta, flags, _sidedata) = res |
|
586 | (node, p1, p2, cs, deltabase, delta, flags, _sidedata) = res | |
572 |
|
587 | |||
573 | sidedata_raw = getchunk(self._stream) |
|
588 | sidedata_raw = getchunk(self._stream) | |
574 | sidedata = {} |
|
589 | sidedata = {} | |
575 | if len(sidedata_raw) > 0: |
|
590 | if len(sidedata_raw) > 0: | |
576 | sidedata = sidedatamod.deserialize_sidedata(sidedata_raw) |
|
591 | sidedata = sidedatamod.deserialize_sidedata(sidedata_raw) | |
577 |
|
592 | |||
578 | return node, p1, p2, cs, deltabase, delta, flags, sidedata |
|
593 | return node, p1, p2, cs, deltabase, delta, flags, sidedata | |
579 |
|
594 | |||
580 |
|
595 | |||
581 | class headerlessfixup(object): |
|
596 | class headerlessfixup(object): | |
582 | def __init__(self, fh, h): |
|
597 | def __init__(self, fh, h): | |
583 | self._h = h |
|
598 | self._h = h | |
584 | self._fh = fh |
|
599 | self._fh = fh | |
585 |
|
600 | |||
586 | def read(self, n): |
|
601 | def read(self, n): | |
587 | if self._h: |
|
602 | if self._h: | |
588 | d, self._h = self._h[:n], self._h[n:] |
|
603 | d, self._h = self._h[:n], self._h[n:] | |
589 | if len(d) < n: |
|
604 | if len(d) < n: | |
590 | d += readexactly(self._fh, n - len(d)) |
|
605 | d += readexactly(self._fh, n - len(d)) | |
591 | return d |
|
606 | return d | |
592 | return readexactly(self._fh, n) |
|
607 | return readexactly(self._fh, n) | |
593 |
|
608 | |||
594 |
|
609 | |||
595 | def _revisiondeltatochunks(delta, headerfn): |
|
610 | def _revisiondeltatochunks(delta, headerfn): | |
596 | """Serialize a revisiondelta to changegroup chunks.""" |
|
611 | """Serialize a revisiondelta to changegroup chunks.""" | |
597 |
|
612 | |||
598 | # The captured revision delta may be encoded as a delta against |
|
613 | # The captured revision delta may be encoded as a delta against | |
599 | # a base revision or as a full revision. The changegroup format |
|
614 | # a base revision or as a full revision. The changegroup format | |
600 | # requires that everything on the wire be deltas. So for full |
|
615 | # requires that everything on the wire be deltas. So for full | |
601 | # revisions, we need to invent a header that says to rewrite |
|
616 | # revisions, we need to invent a header that says to rewrite | |
602 | # data. |
|
617 | # data. | |
603 |
|
618 | |||
604 | if delta.delta is not None: |
|
619 | if delta.delta is not None: | |
605 | prefix, data = b'', delta.delta |
|
620 | prefix, data = b'', delta.delta | |
606 | elif delta.basenode == nullid: |
|
621 | elif delta.basenode == nullid: | |
607 | data = delta.revision |
|
622 | data = delta.revision | |
608 | prefix = mdiff.trivialdiffheader(len(data)) |
|
623 | prefix = mdiff.trivialdiffheader(len(data)) | |
609 | else: |
|
624 | else: | |
610 | data = delta.revision |
|
625 | data = delta.revision | |
611 | prefix = mdiff.replacediffheader(delta.baserevisionsize, len(data)) |
|
626 | prefix = mdiff.replacediffheader(delta.baserevisionsize, len(data)) | |
612 |
|
627 | |||
613 | meta = headerfn(delta) |
|
628 | meta = headerfn(delta) | |
614 |
|
629 | |||
615 | yield chunkheader(len(meta) + len(prefix) + len(data)) |
|
630 | yield chunkheader(len(meta) + len(prefix) + len(data)) | |
616 | yield meta |
|
631 | yield meta | |
617 | if prefix: |
|
632 | if prefix: | |
618 | yield prefix |
|
633 | yield prefix | |
619 | yield data |
|
634 | yield data | |
620 |
|
635 | |||
621 | sidedata = delta.sidedata |
|
636 | sidedata = delta.sidedata | |
622 | if sidedata is not None: |
|
637 | if sidedata is not None: | |
623 | # Need a separate chunk for sidedata to be able to differentiate |
|
638 | # Need a separate chunk for sidedata to be able to differentiate | |
624 | # "raw delta" length and sidedata length |
|
639 | # "raw delta" length and sidedata length | |
625 | yield chunkheader(len(sidedata)) |
|
640 | yield chunkheader(len(sidedata)) | |
626 | yield sidedata |
|
641 | yield sidedata | |
627 |
|
642 | |||
628 |
|
643 | |||
629 | def _sortnodesellipsis(store, nodes, cl, lookup): |
|
644 | def _sortnodesellipsis(store, nodes, cl, lookup): | |
630 | """Sort nodes for changegroup generation.""" |
|
645 | """Sort nodes for changegroup generation.""" | |
631 | # Ellipses serving mode. |
|
646 | # Ellipses serving mode. | |
632 | # |
|
647 | # | |
633 | # In a perfect world, we'd generate better ellipsis-ified graphs |
|
648 | # In a perfect world, we'd generate better ellipsis-ified graphs | |
634 | # for non-changelog revlogs. In practice, we haven't started doing |
|
649 | # for non-changelog revlogs. In practice, we haven't started doing | |
635 | # that yet, so the resulting DAGs for the manifestlog and filelogs |
|
650 | # that yet, so the resulting DAGs for the manifestlog and filelogs | |
636 | # are actually full of bogus parentage on all the ellipsis |
|
651 | # are actually full of bogus parentage on all the ellipsis | |
637 | # nodes. This has the side effect that, while the contents are |
|
652 | # nodes. This has the side effect that, while the contents are | |
638 | # correct, the individual DAGs might be completely out of whack in |
|
653 | # correct, the individual DAGs might be completely out of whack in | |
639 | # a case like 882681bc3166 and its ancestors (back about 10 |
|
654 | # a case like 882681bc3166 and its ancestors (back about 10 | |
640 | # revisions or so) in the main hg repo. |
|
655 | # revisions or so) in the main hg repo. | |
641 | # |
|
656 | # | |
642 | # The one invariant we *know* holds is that the new (potentially |
|
657 | # The one invariant we *know* holds is that the new (potentially | |
643 | # bogus) DAG shape will be valid if we order the nodes in the |
|
658 | # bogus) DAG shape will be valid if we order the nodes in the | |
644 | # order that they're introduced in dramatis personae by the |
|
659 | # order that they're introduced in dramatis personae by the | |
645 | # changelog, so what we do is we sort the non-changelog histories |
|
660 | # changelog, so what we do is we sort the non-changelog histories | |
646 | # by the order in which they are used by the changelog. |
|
661 | # by the order in which they are used by the changelog. | |
647 | key = lambda n: cl.rev(lookup(n)) |
|
662 | key = lambda n: cl.rev(lookup(n)) | |
648 | return sorted(nodes, key=key) |
|
663 | return sorted(nodes, key=key) | |
649 |
|
664 | |||
650 |
|
665 | |||
651 | def _resolvenarrowrevisioninfo( |
|
666 | def _resolvenarrowrevisioninfo( | |
652 | cl, |
|
667 | cl, | |
653 | store, |
|
668 | store, | |
654 | ischangelog, |
|
669 | ischangelog, | |
655 | rev, |
|
670 | rev, | |
656 | linkrev, |
|
671 | linkrev, | |
657 | linknode, |
|
672 | linknode, | |
658 | clrevtolocalrev, |
|
673 | clrevtolocalrev, | |
659 | fullclnodes, |
|
674 | fullclnodes, | |
660 | precomputedellipsis, |
|
675 | precomputedellipsis, | |
661 | ): |
|
676 | ): | |
662 | linkparents = precomputedellipsis[linkrev] |
|
677 | linkparents = precomputedellipsis[linkrev] | |
663 |
|
678 | |||
664 | def local(clrev): |
|
679 | def local(clrev): | |
665 | """Turn a changelog revnum into a local revnum. |
|
680 | """Turn a changelog revnum into a local revnum. | |
666 |
|
681 | |||
667 | The ellipsis dag is stored as revnums on the changelog, |
|
682 | The ellipsis dag is stored as revnums on the changelog, | |
668 | but when we're producing ellipsis entries for |
|
683 | but when we're producing ellipsis entries for | |
669 | non-changelog revlogs, we need to turn those numbers into |
|
684 | non-changelog revlogs, we need to turn those numbers into | |
670 | something local. This does that for us, and during the |
|
685 | something local. This does that for us, and during the | |
671 | changelog sending phase will also expand the stored |
|
686 | changelog sending phase will also expand the stored | |
672 | mappings as needed. |
|
687 | mappings as needed. | |
673 | """ |
|
688 | """ | |
674 | if clrev == nullrev: |
|
689 | if clrev == nullrev: | |
675 | return nullrev |
|
690 | return nullrev | |
676 |
|
691 | |||
677 | if ischangelog: |
|
692 | if ischangelog: | |
678 | return clrev |
|
693 | return clrev | |
679 |
|
694 | |||
680 | # Walk the ellipsis-ized changelog breadth-first looking for a |
|
695 | # Walk the ellipsis-ized changelog breadth-first looking for a | |
681 | # change that has been linked from the current revlog. |
|
696 | # change that has been linked from the current revlog. | |
682 | # |
|
697 | # | |
683 | # For a flat manifest revlog only a single step should be necessary |
|
698 | # For a flat manifest revlog only a single step should be necessary | |
684 | # as all relevant changelog entries are relevant to the flat |
|
699 | # as all relevant changelog entries are relevant to the flat | |
685 | # manifest. |
|
700 | # manifest. | |
686 | # |
|
701 | # | |
687 | # For a filelog or tree manifest dirlog however not every changelog |
|
702 | # For a filelog or tree manifest dirlog however not every changelog | |
688 | # entry will have been relevant, so we need to skip some changelog |
|
703 | # entry will have been relevant, so we need to skip some changelog | |
689 | # nodes even after ellipsis-izing. |
|
704 | # nodes even after ellipsis-izing. | |
690 | walk = [clrev] |
|
705 | walk = [clrev] | |
691 | while walk: |
|
706 | while walk: | |
692 | p = walk[0] |
|
707 | p = walk[0] | |
693 | walk = walk[1:] |
|
708 | walk = walk[1:] | |
694 | if p in clrevtolocalrev: |
|
709 | if p in clrevtolocalrev: | |
695 | return clrevtolocalrev[p] |
|
710 | return clrevtolocalrev[p] | |
696 | elif p in fullclnodes: |
|
711 | elif p in fullclnodes: | |
697 | walk.extend([pp for pp in cl.parentrevs(p) if pp != nullrev]) |
|
712 | walk.extend([pp for pp in cl.parentrevs(p) if pp != nullrev]) | |
698 | elif p in precomputedellipsis: |
|
713 | elif p in precomputedellipsis: | |
699 | walk.extend( |
|
714 | walk.extend( | |
700 | [pp for pp in precomputedellipsis[p] if pp != nullrev] |
|
715 | [pp for pp in precomputedellipsis[p] if pp != nullrev] | |
701 | ) |
|
716 | ) | |
702 | else: |
|
717 | else: | |
703 | # In this case, we've got an ellipsis with parents |
|
718 | # In this case, we've got an ellipsis with parents | |
704 | # outside the current bundle (likely an |
|
719 | # outside the current bundle (likely an | |
705 | # incremental pull). We "know" that we can use the |
|
720 | # incremental pull). We "know" that we can use the | |
706 | # value of this same revlog at whatever revision |
|
721 | # value of this same revlog at whatever revision | |
707 | # is pointed to by linknode. "Know" is in scare |
|
722 | # is pointed to by linknode. "Know" is in scare | |
708 | # quotes because I haven't done enough examination |
|
723 | # quotes because I haven't done enough examination | |
709 | # of edge cases to convince myself this is really |
|
724 | # of edge cases to convince myself this is really | |
710 | # a fact - it works for all the (admittedly |
|
725 | # a fact - it works for all the (admittedly | |
711 | # thorough) cases in our testsuite, but I would be |
|
726 | # thorough) cases in our testsuite, but I would be | |
712 | # somewhat unsurprised to find a case in the wild |
|
727 | # somewhat unsurprised to find a case in the wild | |
713 | # where this breaks down a bit. That said, I don't |
|
728 | # where this breaks down a bit. That said, I don't | |
714 | # know if it would hurt anything. |
|
729 | # know if it would hurt anything. | |
715 | for i in pycompat.xrange(rev, 0, -1): |
|
730 | for i in pycompat.xrange(rev, 0, -1): | |
716 | if store.linkrev(i) == clrev: |
|
731 | if store.linkrev(i) == clrev: | |
717 | return i |
|
732 | return i | |
718 | # We failed to resolve a parent for this node, so |
|
733 | # We failed to resolve a parent for this node, so | |
719 | # we crash the changegroup construction. |
|
734 | # we crash the changegroup construction. | |
720 | raise error.Abort( |
|
735 | raise error.Abort( | |
721 | b"unable to resolve parent while packing '%s' %r" |
|
736 | b"unable to resolve parent while packing '%s' %r" | |
722 | b' for changeset %r' % (store.indexfile, rev, clrev) |
|
737 | b' for changeset %r' % (store.indexfile, rev, clrev) | |
723 | ) |
|
738 | ) | |
724 |
|
739 | |||
725 | return nullrev |
|
740 | return nullrev | |
726 |
|
741 | |||
727 | if not linkparents or (store.parentrevs(rev) == (nullrev, nullrev)): |
|
742 | if not linkparents or (store.parentrevs(rev) == (nullrev, nullrev)): | |
728 | p1, p2 = nullrev, nullrev |
|
743 | p1, p2 = nullrev, nullrev | |
729 | elif len(linkparents) == 1: |
|
744 | elif len(linkparents) == 1: | |
730 | (p1,) = sorted(local(p) for p in linkparents) |
|
745 | (p1,) = sorted(local(p) for p in linkparents) | |
731 | p2 = nullrev |
|
746 | p2 = nullrev | |
732 | else: |
|
747 | else: | |
733 | p1, p2 = sorted(local(p) for p in linkparents) |
|
748 | p1, p2 = sorted(local(p) for p in linkparents) | |
734 |
|
749 | |||
735 | p1node, p2node = store.node(p1), store.node(p2) |
|
750 | p1node, p2node = store.node(p1), store.node(p2) | |
736 |
|
751 | |||
737 | return p1node, p2node, linknode |
|
752 | return p1node, p2node, linknode | |
738 |
|
753 | |||
739 |
|
754 | |||
740 | def deltagroup( |
|
755 | def deltagroup( | |
741 | repo, |
|
756 | repo, | |
742 | store, |
|
757 | store, | |
743 | nodes, |
|
758 | nodes, | |
744 | ischangelog, |
|
759 | ischangelog, | |
745 | lookup, |
|
760 | lookup, | |
746 | forcedeltaparentprev, |
|
761 | forcedeltaparentprev, | |
747 | topic=None, |
|
762 | topic=None, | |
748 | ellipses=False, |
|
763 | ellipses=False, | |
749 | clrevtolocalrev=None, |
|
764 | clrevtolocalrev=None, | |
750 | fullclnodes=None, |
|
765 | fullclnodes=None, | |
751 | precomputedellipsis=None, |
|
766 | precomputedellipsis=None, | |
|
767 | sidedata_helpers=None, | |||
752 | ): |
|
768 | ): | |
753 | """Calculate deltas for a set of revisions. |
|
769 | """Calculate deltas for a set of revisions. | |
754 |
|
770 | |||
755 | Is a generator of ``revisiondelta`` instances. |
|
771 | Is a generator of ``revisiondelta`` instances. | |
756 |
|
772 | |||
757 | If topic is not None, progress detail will be generated using this |
|
773 | If topic is not None, progress detail will be generated using this | |
758 | topic name (e.g. changesets, manifests, etc). |
|
774 | topic name (e.g. changesets, manifests, etc). | |
|
775 | ||||
|
776 | See `storageutil.emitrevisions` for the doc on `sidedata_helpers`. | |||
759 | """ |
|
777 | """ | |
760 | if not nodes: |
|
778 | if not nodes: | |
761 | return |
|
779 | return | |
762 |
|
780 | |||
763 | cl = repo.changelog |
|
781 | cl = repo.changelog | |
764 |
|
782 | |||
765 | if ischangelog: |
|
783 | if ischangelog: | |
766 | # `hg log` shows changesets in storage order. To preserve order |
|
784 | # `hg log` shows changesets in storage order. To preserve order | |
767 | # across clones, send out changesets in storage order. |
|
785 | # across clones, send out changesets in storage order. | |
768 | nodesorder = b'storage' |
|
786 | nodesorder = b'storage' | |
769 | elif ellipses: |
|
787 | elif ellipses: | |
770 | nodes = _sortnodesellipsis(store, nodes, cl, lookup) |
|
788 | nodes = _sortnodesellipsis(store, nodes, cl, lookup) | |
771 | nodesorder = b'nodes' |
|
789 | nodesorder = b'nodes' | |
772 | else: |
|
790 | else: | |
773 | nodesorder = None |
|
791 | nodesorder = None | |
774 |
|
792 | |||
775 | # Perform ellipses filtering and revision massaging. We do this before |
|
793 | # Perform ellipses filtering and revision massaging. We do this before | |
776 | # emitrevisions() because a) filtering out revisions creates less work |
|
794 | # emitrevisions() because a) filtering out revisions creates less work | |
777 | # for emitrevisions() b) dropping revisions would break emitrevisions()'s |
|
795 | # for emitrevisions() b) dropping revisions would break emitrevisions()'s | |
778 | # assumptions about delta choices and we would possibly send a delta |
|
796 | # assumptions about delta choices and we would possibly send a delta | |
779 | # referencing a missing base revision. |
|
797 | # referencing a missing base revision. | |
780 | # |
|
798 | # | |
781 | # Also, calling lookup() has side-effects with regards to populating |
|
799 | # Also, calling lookup() has side-effects with regards to populating | |
782 | # data structures. If we don't call lookup() for each node or if we call |
|
800 | # data structures. If we don't call lookup() for each node or if we call | |
783 | # lookup() after the first pass through each node, things can break - |
|
801 | # lookup() after the first pass through each node, things can break - | |
784 | # possibly intermittently depending on the python hash seed! For that |
|
802 | # possibly intermittently depending on the python hash seed! For that | |
785 | # reason, we store a mapping of all linknodes during the initial node |
|
803 | # reason, we store a mapping of all linknodes during the initial node | |
786 | # pass rather than use lookup() on the output side. |
|
804 | # pass rather than use lookup() on the output side. | |
787 | if ellipses: |
|
805 | if ellipses: | |
788 | filtered = [] |
|
806 | filtered = [] | |
789 | adjustedparents = {} |
|
807 | adjustedparents = {} | |
790 | linknodes = {} |
|
808 | linknodes = {} | |
791 |
|
809 | |||
792 | for node in nodes: |
|
810 | for node in nodes: | |
793 | rev = store.rev(node) |
|
811 | rev = store.rev(node) | |
794 | linknode = lookup(node) |
|
812 | linknode = lookup(node) | |
795 | linkrev = cl.rev(linknode) |
|
813 | linkrev = cl.rev(linknode) | |
796 | clrevtolocalrev[linkrev] = rev |
|
814 | clrevtolocalrev[linkrev] = rev | |
797 |
|
815 | |||
798 | # If linknode is in fullclnodes, it means the corresponding |
|
816 | # If linknode is in fullclnodes, it means the corresponding | |
799 | # changeset was a full changeset and is being sent unaltered. |
|
817 | # changeset was a full changeset and is being sent unaltered. | |
800 | if linknode in fullclnodes: |
|
818 | if linknode in fullclnodes: | |
801 | linknodes[node] = linknode |
|
819 | linknodes[node] = linknode | |
802 |
|
820 | |||
803 | # If the corresponding changeset wasn't in the set computed |
|
821 | # If the corresponding changeset wasn't in the set computed | |
804 | # as relevant to us, it should be dropped outright. |
|
822 | # as relevant to us, it should be dropped outright. | |
805 | elif linkrev not in precomputedellipsis: |
|
823 | elif linkrev not in precomputedellipsis: | |
806 | continue |
|
824 | continue | |
807 |
|
825 | |||
808 | else: |
|
826 | else: | |
809 | # We could probably do this later and avoid the dict |
|
827 | # We could probably do this later and avoid the dict | |
810 | # holding state. But it likely doesn't matter. |
|
828 | # holding state. But it likely doesn't matter. | |
811 | p1node, p2node, linknode = _resolvenarrowrevisioninfo( |
|
829 | p1node, p2node, linknode = _resolvenarrowrevisioninfo( | |
812 | cl, |
|
830 | cl, | |
813 | store, |
|
831 | store, | |
814 | ischangelog, |
|
832 | ischangelog, | |
815 | rev, |
|
833 | rev, | |
816 | linkrev, |
|
834 | linkrev, | |
817 | linknode, |
|
835 | linknode, | |
818 | clrevtolocalrev, |
|
836 | clrevtolocalrev, | |
819 | fullclnodes, |
|
837 | fullclnodes, | |
820 | precomputedellipsis, |
|
838 | precomputedellipsis, | |
821 | ) |
|
839 | ) | |
822 |
|
840 | |||
823 | adjustedparents[node] = (p1node, p2node) |
|
841 | adjustedparents[node] = (p1node, p2node) | |
824 | linknodes[node] = linknode |
|
842 | linknodes[node] = linknode | |
825 |
|
843 | |||
826 | filtered.append(node) |
|
844 | filtered.append(node) | |
827 |
|
845 | |||
828 | nodes = filtered |
|
846 | nodes = filtered | |
829 |
|
847 | |||
830 | # We expect the first pass to be fast, so we only engage the progress |
|
848 | # We expect the first pass to be fast, so we only engage the progress | |
831 | # meter for constructing the revision deltas. |
|
849 | # meter for constructing the revision deltas. | |
832 | progress = None |
|
850 | progress = None | |
833 | if topic is not None: |
|
851 | if topic is not None: | |
834 | progress = repo.ui.makeprogress( |
|
852 | progress = repo.ui.makeprogress( | |
835 | topic, unit=_(b'chunks'), total=len(nodes) |
|
853 | topic, unit=_(b'chunks'), total=len(nodes) | |
836 | ) |
|
854 | ) | |
837 |
|
855 | |||
838 | configtarget = repo.ui.config(b'devel', b'bundle.delta') |
|
856 | configtarget = repo.ui.config(b'devel', b'bundle.delta') | |
839 | if configtarget not in (b'', b'p1', b'full'): |
|
857 | if configtarget not in (b'', b'p1', b'full'): | |
840 | msg = _("""config "devel.bundle.delta" as unknown value: %s""") |
|
858 | msg = _("""config "devel.bundle.delta" as unknown value: %s""") | |
841 | repo.ui.warn(msg % configtarget) |
|
859 | repo.ui.warn(msg % configtarget) | |
842 |
|
860 | |||
843 | deltamode = repository.CG_DELTAMODE_STD |
|
861 | deltamode = repository.CG_DELTAMODE_STD | |
844 | if forcedeltaparentprev: |
|
862 | if forcedeltaparentprev: | |
845 | deltamode = repository.CG_DELTAMODE_PREV |
|
863 | deltamode = repository.CG_DELTAMODE_PREV | |
846 | elif configtarget == b'p1': |
|
864 | elif configtarget == b'p1': | |
847 | deltamode = repository.CG_DELTAMODE_P1 |
|
865 | deltamode = repository.CG_DELTAMODE_P1 | |
848 | elif configtarget == b'full': |
|
866 | elif configtarget == b'full': | |
849 | deltamode = repository.CG_DELTAMODE_FULL |
|
867 | deltamode = repository.CG_DELTAMODE_FULL | |
850 |
|
868 | |||
851 | revisions = store.emitrevisions( |
|
869 | revisions = store.emitrevisions( | |
852 | nodes, |
|
870 | nodes, | |
853 | nodesorder=nodesorder, |
|
871 | nodesorder=nodesorder, | |
854 | revisiondata=True, |
|
872 | revisiondata=True, | |
855 | assumehaveparentrevisions=not ellipses, |
|
873 | assumehaveparentrevisions=not ellipses, | |
856 | deltamode=deltamode, |
|
874 | deltamode=deltamode, | |
|
875 | sidedata_helpers=sidedata_helpers, | |||
857 | ) |
|
876 | ) | |
858 |
|
877 | |||
859 | for i, revision in enumerate(revisions): |
|
878 | for i, revision in enumerate(revisions): | |
860 | if progress: |
|
879 | if progress: | |
861 | progress.update(i + 1) |
|
880 | progress.update(i + 1) | |
862 |
|
881 | |||
863 | if ellipses: |
|
882 | if ellipses: | |
864 | linknode = linknodes[revision.node] |
|
883 | linknode = linknodes[revision.node] | |
865 |
|
884 | |||
866 | if revision.node in adjustedparents: |
|
885 | if revision.node in adjustedparents: | |
867 | p1node, p2node = adjustedparents[revision.node] |
|
886 | p1node, p2node = adjustedparents[revision.node] | |
868 | revision.p1node = p1node |
|
887 | revision.p1node = p1node | |
869 | revision.p2node = p2node |
|
888 | revision.p2node = p2node | |
870 | revision.flags |= repository.REVISION_FLAG_ELLIPSIS |
|
889 | revision.flags |= repository.REVISION_FLAG_ELLIPSIS | |
871 |
|
890 | |||
872 | else: |
|
891 | else: | |
873 | linknode = lookup(revision.node) |
|
892 | linknode = lookup(revision.node) | |
874 |
|
893 | |||
875 | revision.linknode = linknode |
|
894 | revision.linknode = linknode | |
876 | yield revision |
|
895 | yield revision | |
877 |
|
896 | |||
878 | if progress: |
|
897 | if progress: | |
879 | progress.complete() |
|
898 | progress.complete() | |
880 |
|
899 | |||
881 |
|
900 | |||
882 | class cgpacker(object): |
|
901 | class cgpacker(object): | |
883 | def __init__( |
|
902 | def __init__( | |
884 | self, |
|
903 | self, | |
885 | repo, |
|
904 | repo, | |
886 | oldmatcher, |
|
905 | oldmatcher, | |
887 | matcher, |
|
906 | matcher, | |
888 | version, |
|
907 | version, | |
889 | builddeltaheader, |
|
908 | builddeltaheader, | |
890 | manifestsend, |
|
909 | manifestsend, | |
891 | forcedeltaparentprev=False, |
|
910 | forcedeltaparentprev=False, | |
892 | bundlecaps=None, |
|
911 | bundlecaps=None, | |
893 | ellipses=False, |
|
912 | ellipses=False, | |
894 | shallow=False, |
|
913 | shallow=False, | |
895 | ellipsisroots=None, |
|
914 | ellipsisroots=None, | |
896 | fullnodes=None, |
|
915 | fullnodes=None, | |
897 | remote_sidedata=None, |
|
916 | remote_sidedata=None, | |
898 | ): |
|
917 | ): | |
899 | """Given a source repo, construct a bundler. |
|
918 | """Given a source repo, construct a bundler. | |
900 |
|
919 | |||
901 | oldmatcher is a matcher that matches on files the client already has. |
|
920 | oldmatcher is a matcher that matches on files the client already has. | |
902 | These will not be included in the changegroup. |
|
921 | These will not be included in the changegroup. | |
903 |
|
922 | |||
904 | matcher is a matcher that matches on files to include in the |
|
923 | matcher is a matcher that matches on files to include in the | |
905 | changegroup. Used to facilitate sparse changegroups. |
|
924 | changegroup. Used to facilitate sparse changegroups. | |
906 |
|
925 | |||
907 | forcedeltaparentprev indicates whether delta parents must be against |
|
926 | forcedeltaparentprev indicates whether delta parents must be against | |
908 | the previous revision in a delta group. This should only be used for |
|
927 | the previous revision in a delta group. This should only be used for | |
909 | compatibility with changegroup version 1. |
|
928 | compatibility with changegroup version 1. | |
910 |
|
929 | |||
911 | builddeltaheader is a callable that constructs the header for a group |
|
930 | builddeltaheader is a callable that constructs the header for a group | |
912 | delta. |
|
931 | delta. | |
913 |
|
932 | |||
914 | manifestsend is a chunk to send after manifests have been fully emitted. |
|
933 | manifestsend is a chunk to send after manifests have been fully emitted. | |
915 |
|
934 | |||
916 | ellipses indicates whether ellipsis serving mode is enabled. |
|
935 | ellipses indicates whether ellipsis serving mode is enabled. | |
917 |
|
936 | |||
918 | bundlecaps is optional and can be used to specify the set of |
|
937 | bundlecaps is optional and can be used to specify the set of | |
919 | capabilities which can be used to build the bundle. While bundlecaps is |
|
938 | capabilities which can be used to build the bundle. While bundlecaps is | |
920 | unused in core Mercurial, extensions rely on this feature to communicate |
|
939 | unused in core Mercurial, extensions rely on this feature to communicate | |
921 | capabilities to customize the changegroup packer. |
|
940 | capabilities to customize the changegroup packer. | |
922 |
|
941 | |||
923 | shallow indicates whether shallow data might be sent. The packer may |
|
942 | shallow indicates whether shallow data might be sent. The packer may | |
924 | need to pack file contents not introduced by the changes being packed. |
|
943 | need to pack file contents not introduced by the changes being packed. | |
925 |
|
944 | |||
926 | fullnodes is the set of changelog nodes which should not be ellipsis |
|
945 | fullnodes is the set of changelog nodes which should not be ellipsis | |
927 | nodes. We store this rather than the set of nodes that should be |
|
946 | nodes. We store this rather than the set of nodes that should be | |
928 | ellipsis because for very large histories we expect this to be |
|
947 | ellipsis because for very large histories we expect this to be | |
929 | significantly smaller. |
|
948 | significantly smaller. | |
930 |
|
949 | |||
931 | remote_sidedata is the set of sidedata categories wanted by the remote. |
|
950 | remote_sidedata is the set of sidedata categories wanted by the remote. | |
932 | """ |
|
951 | """ | |
933 | assert oldmatcher |
|
952 | assert oldmatcher | |
934 | assert matcher |
|
953 | assert matcher | |
935 | self._oldmatcher = oldmatcher |
|
954 | self._oldmatcher = oldmatcher | |
936 | self._matcher = matcher |
|
955 | self._matcher = matcher | |
937 |
|
956 | |||
938 | self.version = version |
|
957 | self.version = version | |
939 | self._forcedeltaparentprev = forcedeltaparentprev |
|
958 | self._forcedeltaparentprev = forcedeltaparentprev | |
940 | self._builddeltaheader = builddeltaheader |
|
959 | self._builddeltaheader = builddeltaheader | |
941 | self._manifestsend = manifestsend |
|
960 | self._manifestsend = manifestsend | |
942 | self._ellipses = ellipses |
|
961 | self._ellipses = ellipses | |
943 |
|
962 | |||
944 | # Set of capabilities we can use to build the bundle. |
|
963 | # Set of capabilities we can use to build the bundle. | |
945 | if bundlecaps is None: |
|
964 | if bundlecaps is None: | |
946 | bundlecaps = set() |
|
965 | bundlecaps = set() | |
947 | self._bundlecaps = bundlecaps |
|
966 | self._bundlecaps = bundlecaps | |
948 | if remote_sidedata is None: |
|
967 | if remote_sidedata is None: | |
949 | remote_sidedata = set() |
|
968 | remote_sidedata = set() | |
950 | self._remote_sidedata = remote_sidedata |
|
969 | self._remote_sidedata = remote_sidedata | |
951 | self._isshallow = shallow |
|
970 | self._isshallow = shallow | |
952 | self._fullclnodes = fullnodes |
|
971 | self._fullclnodes = fullnodes | |
953 |
|
972 | |||
954 | # Maps ellipsis revs to their roots at the changelog level. |
|
973 | # Maps ellipsis revs to their roots at the changelog level. | |
955 | self._precomputedellipsis = ellipsisroots |
|
974 | self._precomputedellipsis = ellipsisroots | |
956 |
|
975 | |||
957 | self._repo = repo |
|
976 | self._repo = repo | |
958 |
|
977 | |||
959 | if self._repo.ui.verbose and not self._repo.ui.debugflag: |
|
978 | if self._repo.ui.verbose and not self._repo.ui.debugflag: | |
960 | self._verbosenote = self._repo.ui.note |
|
979 | self._verbosenote = self._repo.ui.note | |
961 | else: |
|
980 | else: | |
962 | self._verbosenote = lambda s: None |
|
981 | self._verbosenote = lambda s: None | |
963 |
|
982 | |||
964 | def generate( |
|
983 | def generate( | |
965 | self, commonrevs, clnodes, fastpathlinkrev, source, changelog=True |
|
984 | self, commonrevs, clnodes, fastpathlinkrev, source, changelog=True | |
966 | ): |
|
985 | ): | |
967 | """Yield a sequence of changegroup byte chunks. |
|
986 | """Yield a sequence of changegroup byte chunks. | |
968 | If changelog is False, changelog data won't be added to changegroup |
|
987 | If changelog is False, changelog data won't be added to changegroup | |
969 | """ |
|
988 | """ | |
970 |
|
989 | |||
971 | repo = self._repo |
|
990 | repo = self._repo | |
972 | cl = repo.changelog |
|
991 | cl = repo.changelog | |
973 |
|
992 | |||
974 | self._verbosenote(_(b'uncompressed size of bundle content:\n')) |
|
993 | self._verbosenote(_(b'uncompressed size of bundle content:\n')) | |
975 | size = 0 |
|
994 | size = 0 | |
976 |
|
995 | |||
|
996 | sidedata_helpers = None | |||
|
997 | if self.version == b'04': | |||
|
998 | remote_sidedata = self._remote_sidedata | |||
|
999 | if source == b'strip': | |||
|
1000 | # We're our own remote when stripping, get the no-op helpers | |||
|
1001 | # TODO a better approach would be for the strip bundle to | |||
|
1002 | # correctly advertise its sidedata categories directly. | |||
|
1003 | remote_sidedata = repo._wanted_sidedata | |||
|
1004 | sidedata_helpers = get_sidedata_helpers(repo, remote_sidedata) | |||
|
1005 | ||||
977 | clstate, deltas = self._generatechangelog( |
|
1006 | clstate, deltas = self._generatechangelog( | |
978 | cl, clnodes, generate=changelog |
|
1007 | cl, | |
|
1008 | clnodes, | |||
|
1009 | generate=changelog, | |||
|
1010 | sidedata_helpers=sidedata_helpers, | |||
979 | ) |
|
1011 | ) | |
980 | for delta in deltas: |
|
1012 | for delta in deltas: | |
981 | for chunk in _revisiondeltatochunks(delta, self._builddeltaheader): |
|
1013 | for chunk in _revisiondeltatochunks(delta, self._builddeltaheader): | |
982 | size += len(chunk) |
|
1014 | size += len(chunk) | |
983 | yield chunk |
|
1015 | yield chunk | |
984 |
|
1016 | |||
985 | close = closechunk() |
|
1017 | close = closechunk() | |
986 | size += len(close) |
|
1018 | size += len(close) | |
987 | yield closechunk() |
|
1019 | yield closechunk() | |
988 |
|
1020 | |||
989 | self._verbosenote(_(b'%8.i (changelog)\n') % size) |
|
1021 | self._verbosenote(_(b'%8.i (changelog)\n') % size) | |
990 |
|
1022 | |||
991 | clrevorder = clstate[b'clrevorder'] |
|
1023 | clrevorder = clstate[b'clrevorder'] | |
992 | manifests = clstate[b'manifests'] |
|
1024 | manifests = clstate[b'manifests'] | |
993 | changedfiles = clstate[b'changedfiles'] |
|
1025 | changedfiles = clstate[b'changedfiles'] | |
994 |
|
1026 | |||
995 | # We need to make sure that the linkrev in the changegroup refers to |
|
1027 | # We need to make sure that the linkrev in the changegroup refers to | |
996 | # the first changeset that introduced the manifest or file revision. |
|
1028 | # the first changeset that introduced the manifest or file revision. | |
997 | # The fastpath is usually safer than the slowpath, because the filelogs |
|
1029 | # The fastpath is usually safer than the slowpath, because the filelogs | |
998 | # are walked in revlog order. |
|
1030 | # are walked in revlog order. | |
999 | # |
|
1031 | # | |
1000 | # When taking the slowpath when the manifest revlog uses generaldelta, |
|
1032 | # When taking the slowpath when the manifest revlog uses generaldelta, | |
1001 | # the manifest may be walked in the "wrong" order. Without 'clrevorder', |
|
1033 | # the manifest may be walked in the "wrong" order. Without 'clrevorder', | |
1002 | # we would get an incorrect linkrev (see fix in cc0ff93d0c0c). |
|
1034 | # we would get an incorrect linkrev (see fix in cc0ff93d0c0c). | |
1003 | # |
|
1035 | # | |
1004 | # When taking the fastpath, we are only vulnerable to reordering |
|
1036 | # When taking the fastpath, we are only vulnerable to reordering | |
1005 | # of the changelog itself. The changelog never uses generaldelta and is |
|
1037 | # of the changelog itself. The changelog never uses generaldelta and is | |
1006 | # never reordered. To handle this case, we simply take the slowpath, |
|
1038 | # never reordered. To handle this case, we simply take the slowpath, | |
1007 | # which already has the 'clrevorder' logic. This was also fixed in |
|
1039 | # which already has the 'clrevorder' logic. This was also fixed in | |
1008 | # cc0ff93d0c0c. |
|
1040 | # cc0ff93d0c0c. | |
1009 |
|
1041 | |||
1010 | # Treemanifests don't work correctly with fastpathlinkrev |
|
1042 | # Treemanifests don't work correctly with fastpathlinkrev | |
1011 | # either, because we don't discover which directory nodes to |
|
1043 | # either, because we don't discover which directory nodes to | |
1012 | # send along with files. This could probably be fixed. |
|
1044 | # send along with files. This could probably be fixed. | |
1013 | fastpathlinkrev = fastpathlinkrev and not scmutil.istreemanifest(repo) |
|
1045 | fastpathlinkrev = fastpathlinkrev and not scmutil.istreemanifest(repo) | |
1014 |
|
1046 | |||
1015 | fnodes = {} # needed file nodes |
|
1047 | fnodes = {} # needed file nodes | |
1016 |
|
1048 | |||
1017 | size = 0 |
|
1049 | size = 0 | |
1018 | it = self.generatemanifests( |
|
1050 | it = self.generatemanifests( | |
1019 | commonrevs, |
|
1051 | commonrevs, | |
1020 | clrevorder, |
|
1052 | clrevorder, | |
1021 | fastpathlinkrev, |
|
1053 | fastpathlinkrev, | |
1022 | manifests, |
|
1054 | manifests, | |
1023 | fnodes, |
|
1055 | fnodes, | |
1024 | source, |
|
1056 | source, | |
1025 | clstate[b'clrevtomanifestrev'], |
|
1057 | clstate[b'clrevtomanifestrev'], | |
|
1058 | sidedata_helpers=sidedata_helpers, | |||
1026 | ) |
|
1059 | ) | |
1027 |
|
1060 | |||
1028 | for tree, deltas in it: |
|
1061 | for tree, deltas in it: | |
1029 | if tree: |
|
1062 | if tree: | |
1030 | assert self.version in (b'03', b'04') |
|
1063 | assert self.version in (b'03', b'04') | |
1031 | chunk = _fileheader(tree) |
|
1064 | chunk = _fileheader(tree) | |
1032 | size += len(chunk) |
|
1065 | size += len(chunk) | |
1033 | yield chunk |
|
1066 | yield chunk | |
1034 |
|
1067 | |||
1035 | for delta in deltas: |
|
1068 | for delta in deltas: | |
1036 | chunks = _revisiondeltatochunks(delta, self._builddeltaheader) |
|
1069 | chunks = _revisiondeltatochunks(delta, self._builddeltaheader) | |
1037 | for chunk in chunks: |
|
1070 | for chunk in chunks: | |
1038 | size += len(chunk) |
|
1071 | size += len(chunk) | |
1039 | yield chunk |
|
1072 | yield chunk | |
1040 |
|
1073 | |||
1041 | close = closechunk() |
|
1074 | close = closechunk() | |
1042 | size += len(close) |
|
1075 | size += len(close) | |
1043 | yield close |
|
1076 | yield close | |
1044 |
|
1077 | |||
1045 | self._verbosenote(_(b'%8.i (manifests)\n') % size) |
|
1078 | self._verbosenote(_(b'%8.i (manifests)\n') % size) | |
1046 | yield self._manifestsend |
|
1079 | yield self._manifestsend | |
1047 |
|
1080 | |||
1048 | mfdicts = None |
|
1081 | mfdicts = None | |
1049 | if self._ellipses and self._isshallow: |
|
1082 | if self._ellipses and self._isshallow: | |
1050 | mfdicts = [ |
|
1083 | mfdicts = [ | |
1051 | (repo.manifestlog[n].read(), lr) |
|
1084 | (repo.manifestlog[n].read(), lr) | |
1052 | for (n, lr) in pycompat.iteritems(manifests) |
|
1085 | for (n, lr) in pycompat.iteritems(manifests) | |
1053 | ] |
|
1086 | ] | |
1054 |
|
1087 | |||
1055 | manifests.clear() |
|
1088 | manifests.clear() | |
1056 | clrevs = {cl.rev(x) for x in clnodes} |
|
1089 | clrevs = {cl.rev(x) for x in clnodes} | |
1057 |
|
1090 | |||
1058 | it = self.generatefiles( |
|
1091 | it = self.generatefiles( | |
1059 | changedfiles, |
|
1092 | changedfiles, | |
1060 | commonrevs, |
|
1093 | commonrevs, | |
1061 | source, |
|
1094 | source, | |
1062 | mfdicts, |
|
1095 | mfdicts, | |
1063 | fastpathlinkrev, |
|
1096 | fastpathlinkrev, | |
1064 | fnodes, |
|
1097 | fnodes, | |
1065 | clrevs, |
|
1098 | clrevs, | |
|
1099 | sidedata_helpers=sidedata_helpers, | |||
1066 | ) |
|
1100 | ) | |
1067 |
|
1101 | |||
1068 | for path, deltas in it: |
|
1102 | for path, deltas in it: | |
1069 | h = _fileheader(path) |
|
1103 | h = _fileheader(path) | |
1070 | size = len(h) |
|
1104 | size = len(h) | |
1071 | yield h |
|
1105 | yield h | |
1072 |
|
1106 | |||
1073 | for delta in deltas: |
|
1107 | for delta in deltas: | |
1074 | chunks = _revisiondeltatochunks(delta, self._builddeltaheader) |
|
1108 | chunks = _revisiondeltatochunks(delta, self._builddeltaheader) | |
1075 | for chunk in chunks: |
|
1109 | for chunk in chunks: | |
1076 | size += len(chunk) |
|
1110 | size += len(chunk) | |
1077 | yield chunk |
|
1111 | yield chunk | |
1078 |
|
1112 | |||
1079 | close = closechunk() |
|
1113 | close = closechunk() | |
1080 | size += len(close) |
|
1114 | size += len(close) | |
1081 | yield close |
|
1115 | yield close | |
1082 |
|
1116 | |||
1083 | self._verbosenote(_(b'%8.i %s\n') % (size, path)) |
|
1117 | self._verbosenote(_(b'%8.i %s\n') % (size, path)) | |
1084 |
|
1118 | |||
1085 | yield closechunk() |
|
1119 | yield closechunk() | |
1086 |
|
1120 | |||
1087 | if clnodes: |
|
1121 | if clnodes: | |
1088 | repo.hook(b'outgoing', node=hex(clnodes[0]), source=source) |
|
1122 | repo.hook(b'outgoing', node=hex(clnodes[0]), source=source) | |
1089 |
|
1123 | |||
1090 |
def _generatechangelog( |
|
1124 | def _generatechangelog( | |
|
1125 | self, cl, nodes, generate=True, sidedata_helpers=None | |||
|
1126 | ): | |||
1091 | """Generate data for changelog chunks. |
|
1127 | """Generate data for changelog chunks. | |
1092 |
|
1128 | |||
1093 | Returns a 2-tuple of a dict containing state and an iterable of |
|
1129 | Returns a 2-tuple of a dict containing state and an iterable of | |
1094 | byte chunks. The state will not be fully populated until the |
|
1130 | byte chunks. The state will not be fully populated until the | |
1095 | chunk stream has been fully consumed. |
|
1131 | chunk stream has been fully consumed. | |
1096 |
|
1132 | |||
1097 | if generate is False, the state will be fully populated and no chunk |
|
1133 | if generate is False, the state will be fully populated and no chunk | |
1098 | stream will be yielded |
|
1134 | stream will be yielded | |
|
1135 | ||||
|
1136 | See `storageutil.emitrevisions` for the doc on `sidedata_helpers`. | |||
1099 | """ |
|
1137 | """ | |
1100 | clrevorder = {} |
|
1138 | clrevorder = {} | |
1101 | manifests = {} |
|
1139 | manifests = {} | |
1102 | mfl = self._repo.manifestlog |
|
1140 | mfl = self._repo.manifestlog | |
1103 | changedfiles = set() |
|
1141 | changedfiles = set() | |
1104 | clrevtomanifestrev = {} |
|
1142 | clrevtomanifestrev = {} | |
1105 |
|
1143 | |||
1106 | state = { |
|
1144 | state = { | |
1107 | b'clrevorder': clrevorder, |
|
1145 | b'clrevorder': clrevorder, | |
1108 | b'manifests': manifests, |
|
1146 | b'manifests': manifests, | |
1109 | b'changedfiles': changedfiles, |
|
1147 | b'changedfiles': changedfiles, | |
1110 | b'clrevtomanifestrev': clrevtomanifestrev, |
|
1148 | b'clrevtomanifestrev': clrevtomanifestrev, | |
1111 | } |
|
1149 | } | |
1112 |
|
1150 | |||
1113 | if not (generate or self._ellipses): |
|
1151 | if not (generate or self._ellipses): | |
1114 | # sort the nodes in storage order |
|
1152 | # sort the nodes in storage order | |
1115 | nodes = sorted(nodes, key=cl.rev) |
|
1153 | nodes = sorted(nodes, key=cl.rev) | |
1116 | for node in nodes: |
|
1154 | for node in nodes: | |
1117 | c = cl.changelogrevision(node) |
|
1155 | c = cl.changelogrevision(node) | |
1118 | clrevorder[node] = len(clrevorder) |
|
1156 | clrevorder[node] = len(clrevorder) | |
1119 | # record the first changeset introducing this manifest version |
|
1157 | # record the first changeset introducing this manifest version | |
1120 | manifests.setdefault(c.manifest, node) |
|
1158 | manifests.setdefault(c.manifest, node) | |
1121 | # Record a complete list of potentially-changed files in |
|
1159 | # Record a complete list of potentially-changed files in | |
1122 | # this manifest. |
|
1160 | # this manifest. | |
1123 | changedfiles.update(c.files) |
|
1161 | changedfiles.update(c.files) | |
1124 |
|
1162 | |||
1125 | return state, () |
|
1163 | return state, () | |
1126 |
|
1164 | |||
1127 | # Callback for the changelog, used to collect changed files and |
|
1165 | # Callback for the changelog, used to collect changed files and | |
1128 | # manifest nodes. |
|
1166 | # manifest nodes. | |
1129 | # Returns the linkrev node (identity in the changelog case). |
|
1167 | # Returns the linkrev node (identity in the changelog case). | |
1130 | def lookupcl(x): |
|
1168 | def lookupcl(x): | |
1131 | c = cl.changelogrevision(x) |
|
1169 | c = cl.changelogrevision(x) | |
1132 | clrevorder[x] = len(clrevorder) |
|
1170 | clrevorder[x] = len(clrevorder) | |
1133 |
|
1171 | |||
1134 | if self._ellipses: |
|
1172 | if self._ellipses: | |
1135 | # Only update manifests if x is going to be sent. Otherwise we |
|
1173 | # Only update manifests if x is going to be sent. Otherwise we | |
1136 | # end up with bogus linkrevs specified for manifests and |
|
1174 | # end up with bogus linkrevs specified for manifests and | |
1137 | # we skip some manifest nodes that we should otherwise |
|
1175 | # we skip some manifest nodes that we should otherwise | |
1138 | # have sent. |
|
1176 | # have sent. | |
1139 | if ( |
|
1177 | if ( | |
1140 | x in self._fullclnodes |
|
1178 | x in self._fullclnodes | |
1141 | or cl.rev(x) in self._precomputedellipsis |
|
1179 | or cl.rev(x) in self._precomputedellipsis | |
1142 | ): |
|
1180 | ): | |
1143 |
|
1181 | |||
1144 | manifestnode = c.manifest |
|
1182 | manifestnode = c.manifest | |
1145 | # Record the first changeset introducing this manifest |
|
1183 | # Record the first changeset introducing this manifest | |
1146 | # version. |
|
1184 | # version. | |
1147 | manifests.setdefault(manifestnode, x) |
|
1185 | manifests.setdefault(manifestnode, x) | |
1148 | # Set this narrow-specific dict so we have the lowest |
|
1186 | # Set this narrow-specific dict so we have the lowest | |
1149 | # manifest revnum to look up for this cl revnum. (Part of |
|
1187 | # manifest revnum to look up for this cl revnum. (Part of | |
1150 | # mapping changelog ellipsis parents to manifest ellipsis |
|
1188 | # mapping changelog ellipsis parents to manifest ellipsis | |
1151 | # parents) |
|
1189 | # parents) | |
1152 | clrevtomanifestrev.setdefault( |
|
1190 | clrevtomanifestrev.setdefault( | |
1153 | cl.rev(x), mfl.rev(manifestnode) |
|
1191 | cl.rev(x), mfl.rev(manifestnode) | |
1154 | ) |
|
1192 | ) | |
1155 | # We can't trust the changed files list in the changeset if the |
|
1193 | # We can't trust the changed files list in the changeset if the | |
1156 | # client requested a shallow clone. |
|
1194 | # client requested a shallow clone. | |
1157 | if self._isshallow: |
|
1195 | if self._isshallow: | |
1158 | changedfiles.update(mfl[c.manifest].read().keys()) |
|
1196 | changedfiles.update(mfl[c.manifest].read().keys()) | |
1159 | else: |
|
1197 | else: | |
1160 | changedfiles.update(c.files) |
|
1198 | changedfiles.update(c.files) | |
1161 | else: |
|
1199 | else: | |
1162 | # record the first changeset introducing this manifest version |
|
1200 | # record the first changeset introducing this manifest version | |
1163 | manifests.setdefault(c.manifest, x) |
|
1201 | manifests.setdefault(c.manifest, x) | |
1164 | # Record a complete list of potentially-changed files in |
|
1202 | # Record a complete list of potentially-changed files in | |
1165 | # this manifest. |
|
1203 | # this manifest. | |
1166 | changedfiles.update(c.files) |
|
1204 | changedfiles.update(c.files) | |
1167 |
|
1205 | |||
1168 | return x |
|
1206 | return x | |
1169 |
|
1207 | |||
1170 | gen = deltagroup( |
|
1208 | gen = deltagroup( | |
1171 | self._repo, |
|
1209 | self._repo, | |
1172 | cl, |
|
1210 | cl, | |
1173 | nodes, |
|
1211 | nodes, | |
1174 | True, |
|
1212 | True, | |
1175 | lookupcl, |
|
1213 | lookupcl, | |
1176 | self._forcedeltaparentprev, |
|
1214 | self._forcedeltaparentprev, | |
1177 | ellipses=self._ellipses, |
|
1215 | ellipses=self._ellipses, | |
1178 | topic=_(b'changesets'), |
|
1216 | topic=_(b'changesets'), | |
1179 | clrevtolocalrev={}, |
|
1217 | clrevtolocalrev={}, | |
1180 | fullclnodes=self._fullclnodes, |
|
1218 | fullclnodes=self._fullclnodes, | |
1181 | precomputedellipsis=self._precomputedellipsis, |
|
1219 | precomputedellipsis=self._precomputedellipsis, | |
|
1220 | sidedata_helpers=sidedata_helpers, | |||
1182 | ) |
|
1221 | ) | |
1183 |
|
1222 | |||
1184 | return state, gen |
|
1223 | return state, gen | |
1185 |
|
1224 | |||
1186 | def generatemanifests( |
|
1225 | def generatemanifests( | |
1187 | self, |
|
1226 | self, | |
1188 | commonrevs, |
|
1227 | commonrevs, | |
1189 | clrevorder, |
|
1228 | clrevorder, | |
1190 | fastpathlinkrev, |
|
1229 | fastpathlinkrev, | |
1191 | manifests, |
|
1230 | manifests, | |
1192 | fnodes, |
|
1231 | fnodes, | |
1193 | source, |
|
1232 | source, | |
1194 | clrevtolocalrev, |
|
1233 | clrevtolocalrev, | |
|
1234 | sidedata_helpers=None, | |||
1195 | ): |
|
1235 | ): | |
1196 | """Returns an iterator of changegroup chunks containing manifests. |
|
1236 | """Returns an iterator of changegroup chunks containing manifests. | |
1197 |
|
1237 | |||
1198 | `source` is unused here, but is used by extensions like remotefilelog to |
|
1238 | `source` is unused here, but is used by extensions like remotefilelog to | |
1199 | change what is sent based in pulls vs pushes, etc. |
|
1239 | change what is sent based in pulls vs pushes, etc. | |
|
1240 | ||||
|
1241 | See `storageutil.emitrevisions` for the doc on `sidedata_helpers`. | |||
1200 | """ |
|
1242 | """ | |
1201 | repo = self._repo |
|
1243 | repo = self._repo | |
1202 | mfl = repo.manifestlog |
|
1244 | mfl = repo.manifestlog | |
1203 | tmfnodes = {b'': manifests} |
|
1245 | tmfnodes = {b'': manifests} | |
1204 |
|
1246 | |||
1205 | # Callback for the manifest, used to collect linkrevs for filelog |
|
1247 | # Callback for the manifest, used to collect linkrevs for filelog | |
1206 | # revisions. |
|
1248 | # revisions. | |
1207 | # Returns the linkrev node (collected in lookupcl). |
|
1249 | # Returns the linkrev node (collected in lookupcl). | |
1208 | def makelookupmflinknode(tree, nodes): |
|
1250 | def makelookupmflinknode(tree, nodes): | |
1209 | if fastpathlinkrev: |
|
1251 | if fastpathlinkrev: | |
1210 | assert not tree |
|
1252 | assert not tree | |
1211 | return ( |
|
1253 | return ( | |
1212 | manifests.__getitem__ |
|
1254 | manifests.__getitem__ | |
1213 | ) # pytype: disable=unsupported-operands |
|
1255 | ) # pytype: disable=unsupported-operands | |
1214 |
|
1256 | |||
1215 | def lookupmflinknode(x): |
|
1257 | def lookupmflinknode(x): | |
1216 | """Callback for looking up the linknode for manifests. |
|
1258 | """Callback for looking up the linknode for manifests. | |
1217 |
|
1259 | |||
1218 | Returns the linkrev node for the specified manifest. |
|
1260 | Returns the linkrev node for the specified manifest. | |
1219 |
|
1261 | |||
1220 | SIDE EFFECT: |
|
1262 | SIDE EFFECT: | |
1221 |
|
1263 | |||
1222 | 1) fclnodes gets populated with the list of relevant |
|
1264 | 1) fclnodes gets populated with the list of relevant | |
1223 | file nodes if we're not using fastpathlinkrev |
|
1265 | file nodes if we're not using fastpathlinkrev | |
1224 | 2) When treemanifests are in use, collects treemanifest nodes |
|
1266 | 2) When treemanifests are in use, collects treemanifest nodes | |
1225 | to send |
|
1267 | to send | |
1226 |
|
1268 | |||
1227 | Note that this means manifests must be completely sent to |
|
1269 | Note that this means manifests must be completely sent to | |
1228 | the client before you can trust the list of files and |
|
1270 | the client before you can trust the list of files and | |
1229 | treemanifests to send. |
|
1271 | treemanifests to send. | |
1230 | """ |
|
1272 | """ | |
1231 | clnode = nodes[x] |
|
1273 | clnode = nodes[x] | |
1232 | mdata = mfl.get(tree, x).readfast(shallow=True) |
|
1274 | mdata = mfl.get(tree, x).readfast(shallow=True) | |
1233 | for p, n, fl in mdata.iterentries(): |
|
1275 | for p, n, fl in mdata.iterentries(): | |
1234 | if fl == b't': # subdirectory manifest |
|
1276 | if fl == b't': # subdirectory manifest | |
1235 | subtree = tree + p + b'/' |
|
1277 | subtree = tree + p + b'/' | |
1236 | tmfclnodes = tmfnodes.setdefault(subtree, {}) |
|
1278 | tmfclnodes = tmfnodes.setdefault(subtree, {}) | |
1237 | tmfclnode = tmfclnodes.setdefault(n, clnode) |
|
1279 | tmfclnode = tmfclnodes.setdefault(n, clnode) | |
1238 | if clrevorder[clnode] < clrevorder[tmfclnode]: |
|
1280 | if clrevorder[clnode] < clrevorder[tmfclnode]: | |
1239 | tmfclnodes[n] = clnode |
|
1281 | tmfclnodes[n] = clnode | |
1240 | else: |
|
1282 | else: | |
1241 | f = tree + p |
|
1283 | f = tree + p | |
1242 | fclnodes = fnodes.setdefault(f, {}) |
|
1284 | fclnodes = fnodes.setdefault(f, {}) | |
1243 | fclnode = fclnodes.setdefault(n, clnode) |
|
1285 | fclnode = fclnodes.setdefault(n, clnode) | |
1244 | if clrevorder[clnode] < clrevorder[fclnode]: |
|
1286 | if clrevorder[clnode] < clrevorder[fclnode]: | |
1245 | fclnodes[n] = clnode |
|
1287 | fclnodes[n] = clnode | |
1246 | return clnode |
|
1288 | return clnode | |
1247 |
|
1289 | |||
1248 | return lookupmflinknode |
|
1290 | return lookupmflinknode | |
1249 |
|
1291 | |||
1250 | while tmfnodes: |
|
1292 | while tmfnodes: | |
1251 | tree, nodes = tmfnodes.popitem() |
|
1293 | tree, nodes = tmfnodes.popitem() | |
1252 |
|
1294 | |||
1253 | should_visit = self._matcher.visitdir(tree[:-1]) |
|
1295 | should_visit = self._matcher.visitdir(tree[:-1]) | |
1254 | if tree and not should_visit: |
|
1296 | if tree and not should_visit: | |
1255 | continue |
|
1297 | continue | |
1256 |
|
1298 | |||
1257 | store = mfl.getstorage(tree) |
|
1299 | store = mfl.getstorage(tree) | |
1258 |
|
1300 | |||
1259 | if not should_visit: |
|
1301 | if not should_visit: | |
1260 | # No nodes to send because this directory is out of |
|
1302 | # No nodes to send because this directory is out of | |
1261 | # the client's view of the repository (probably |
|
1303 | # the client's view of the repository (probably | |
1262 | # because of narrow clones). Do this even for the root |
|
1304 | # because of narrow clones). Do this even for the root | |
1263 | # directory (tree=='') |
|
1305 | # directory (tree=='') | |
1264 | prunednodes = [] |
|
1306 | prunednodes = [] | |
1265 | else: |
|
1307 | else: | |
1266 | # Avoid sending any manifest nodes we can prove the |
|
1308 | # Avoid sending any manifest nodes we can prove the | |
1267 | # client already has by checking linkrevs. See the |
|
1309 | # client already has by checking linkrevs. See the | |
1268 | # related comment in generatefiles(). |
|
1310 | # related comment in generatefiles(). | |
1269 | prunednodes = self._prunemanifests(store, nodes, commonrevs) |
|
1311 | prunednodes = self._prunemanifests(store, nodes, commonrevs) | |
1270 |
|
1312 | |||
1271 | if tree and not prunednodes: |
|
1313 | if tree and not prunednodes: | |
1272 | continue |
|
1314 | continue | |
1273 |
|
1315 | |||
1274 | lookupfn = makelookupmflinknode(tree, nodes) |
|
1316 | lookupfn = makelookupmflinknode(tree, nodes) | |
1275 |
|
1317 | |||
1276 | deltas = deltagroup( |
|
1318 | deltas = deltagroup( | |
1277 | self._repo, |
|
1319 | self._repo, | |
1278 | store, |
|
1320 | store, | |
1279 | prunednodes, |
|
1321 | prunednodes, | |
1280 | False, |
|
1322 | False, | |
1281 | lookupfn, |
|
1323 | lookupfn, | |
1282 | self._forcedeltaparentprev, |
|
1324 | self._forcedeltaparentprev, | |
1283 | ellipses=self._ellipses, |
|
1325 | ellipses=self._ellipses, | |
1284 | topic=_(b'manifests'), |
|
1326 | topic=_(b'manifests'), | |
1285 | clrevtolocalrev=clrevtolocalrev, |
|
1327 | clrevtolocalrev=clrevtolocalrev, | |
1286 | fullclnodes=self._fullclnodes, |
|
1328 | fullclnodes=self._fullclnodes, | |
1287 | precomputedellipsis=self._precomputedellipsis, |
|
1329 | precomputedellipsis=self._precomputedellipsis, | |
|
1330 | sidedata_helpers=sidedata_helpers, | |||
1288 | ) |
|
1331 | ) | |
1289 |
|
1332 | |||
1290 | if not self._oldmatcher.visitdir(store.tree[:-1]): |
|
1333 | if not self._oldmatcher.visitdir(store.tree[:-1]): | |
1291 | yield tree, deltas |
|
1334 | yield tree, deltas | |
1292 | else: |
|
1335 | else: | |
1293 | # 'deltas' is a generator and we need to consume it even if |
|
1336 | # 'deltas' is a generator and we need to consume it even if | |
1294 | # we are not going to send it because a side-effect is that |
|
1337 | # we are not going to send it because a side-effect is that | |
1295 | # it updates tmdnodes (via lookupfn) |
|
1338 | # it updates tmdnodes (via lookupfn) | |
1296 | for d in deltas: |
|
1339 | for d in deltas: | |
1297 | pass |
|
1340 | pass | |
1298 | if not tree: |
|
1341 | if not tree: | |
1299 | yield tree, [] |
|
1342 | yield tree, [] | |
1300 |
|
1343 | |||
1301 | def _prunemanifests(self, store, nodes, commonrevs): |
|
1344 | def _prunemanifests(self, store, nodes, commonrevs): | |
1302 | if not self._ellipses: |
|
1345 | if not self._ellipses: | |
1303 | # In non-ellipses case and large repositories, it is better to |
|
1346 | # In non-ellipses case and large repositories, it is better to | |
1304 | # prevent calling of store.rev and store.linkrev on a lot of |
|
1347 | # prevent calling of store.rev and store.linkrev on a lot of | |
1305 | # nodes as compared to sending some extra data |
|
1348 | # nodes as compared to sending some extra data | |
1306 | return nodes.copy() |
|
1349 | return nodes.copy() | |
1307 | # This is split out as a separate method to allow filtering |
|
1350 | # This is split out as a separate method to allow filtering | |
1308 | # commonrevs in extension code. |
|
1351 | # commonrevs in extension code. | |
1309 | # |
|
1352 | # | |
1310 | # TODO(augie): this shouldn't be required, instead we should |
|
1353 | # TODO(augie): this shouldn't be required, instead we should | |
1311 | # make filtering of revisions to send delegated to the store |
|
1354 | # make filtering of revisions to send delegated to the store | |
1312 | # layer. |
|
1355 | # layer. | |
1313 | frev, flr = store.rev, store.linkrev |
|
1356 | frev, flr = store.rev, store.linkrev | |
1314 | return [n for n in nodes if flr(frev(n)) not in commonrevs] |
|
1357 | return [n for n in nodes if flr(frev(n)) not in commonrevs] | |
1315 |
|
1358 | |||
1316 | # The 'source' parameter is useful for extensions |
|
1359 | # The 'source' parameter is useful for extensions | |
1317 | def generatefiles( |
|
1360 | def generatefiles( | |
1318 | self, |
|
1361 | self, | |
1319 | changedfiles, |
|
1362 | changedfiles, | |
1320 | commonrevs, |
|
1363 | commonrevs, | |
1321 | source, |
|
1364 | source, | |
1322 | mfdicts, |
|
1365 | mfdicts, | |
1323 | fastpathlinkrev, |
|
1366 | fastpathlinkrev, | |
1324 | fnodes, |
|
1367 | fnodes, | |
1325 | clrevs, |
|
1368 | clrevs, | |
|
1369 | sidedata_helpers=None, | |||
1326 | ): |
|
1370 | ): | |
1327 | changedfiles = [ |
|
1371 | changedfiles = [ | |
1328 | f |
|
1372 | f | |
1329 | for f in changedfiles |
|
1373 | for f in changedfiles | |
1330 | if self._matcher(f) and not self._oldmatcher(f) |
|
1374 | if self._matcher(f) and not self._oldmatcher(f) | |
1331 | ] |
|
1375 | ] | |
1332 |
|
1376 | |||
1333 | if not fastpathlinkrev: |
|
1377 | if not fastpathlinkrev: | |
1334 |
|
1378 | |||
1335 | def normallinknodes(unused, fname): |
|
1379 | def normallinknodes(unused, fname): | |
1336 | return fnodes.get(fname, {}) |
|
1380 | return fnodes.get(fname, {}) | |
1337 |
|
1381 | |||
1338 | else: |
|
1382 | else: | |
1339 | cln = self._repo.changelog.node |
|
1383 | cln = self._repo.changelog.node | |
1340 |
|
1384 | |||
1341 | def normallinknodes(store, fname): |
|
1385 | def normallinknodes(store, fname): | |
1342 | flinkrev = store.linkrev |
|
1386 | flinkrev = store.linkrev | |
1343 | fnode = store.node |
|
1387 | fnode = store.node | |
1344 | revs = ((r, flinkrev(r)) for r in store) |
|
1388 | revs = ((r, flinkrev(r)) for r in store) | |
1345 | return {fnode(r): cln(lr) for r, lr in revs if lr in clrevs} |
|
1389 | return {fnode(r): cln(lr) for r, lr in revs if lr in clrevs} | |
1346 |
|
1390 | |||
1347 | clrevtolocalrev = {} |
|
1391 | clrevtolocalrev = {} | |
1348 |
|
1392 | |||
1349 | if self._isshallow: |
|
1393 | if self._isshallow: | |
1350 | # In a shallow clone, the linknodes callback needs to also include |
|
1394 | # In a shallow clone, the linknodes callback needs to also include | |
1351 | # those file nodes that are in the manifests we sent but weren't |
|
1395 | # those file nodes that are in the manifests we sent but weren't | |
1352 | # introduced by those manifests. |
|
1396 | # introduced by those manifests. | |
1353 | commonctxs = [self._repo[c] for c in commonrevs] |
|
1397 | commonctxs = [self._repo[c] for c in commonrevs] | |
1354 | clrev = self._repo.changelog.rev |
|
1398 | clrev = self._repo.changelog.rev | |
1355 |
|
1399 | |||
1356 | def linknodes(flog, fname): |
|
1400 | def linknodes(flog, fname): | |
1357 | for c in commonctxs: |
|
1401 | for c in commonctxs: | |
1358 | try: |
|
1402 | try: | |
1359 | fnode = c.filenode(fname) |
|
1403 | fnode = c.filenode(fname) | |
1360 | clrevtolocalrev[c.rev()] = flog.rev(fnode) |
|
1404 | clrevtolocalrev[c.rev()] = flog.rev(fnode) | |
1361 | except error.ManifestLookupError: |
|
1405 | except error.ManifestLookupError: | |
1362 | pass |
|
1406 | pass | |
1363 | links = normallinknodes(flog, fname) |
|
1407 | links = normallinknodes(flog, fname) | |
1364 | if len(links) != len(mfdicts): |
|
1408 | if len(links) != len(mfdicts): | |
1365 | for mf, lr in mfdicts: |
|
1409 | for mf, lr in mfdicts: | |
1366 | fnode = mf.get(fname, None) |
|
1410 | fnode = mf.get(fname, None) | |
1367 | if fnode in links: |
|
1411 | if fnode in links: | |
1368 | links[fnode] = min(links[fnode], lr, key=clrev) |
|
1412 | links[fnode] = min(links[fnode], lr, key=clrev) | |
1369 | elif fnode: |
|
1413 | elif fnode: | |
1370 | links[fnode] = lr |
|
1414 | links[fnode] = lr | |
1371 | return links |
|
1415 | return links | |
1372 |
|
1416 | |||
1373 | else: |
|
1417 | else: | |
1374 | linknodes = normallinknodes |
|
1418 | linknodes = normallinknodes | |
1375 |
|
1419 | |||
1376 | repo = self._repo |
|
1420 | repo = self._repo | |
1377 | progress = repo.ui.makeprogress( |
|
1421 | progress = repo.ui.makeprogress( | |
1378 | _(b'files'), unit=_(b'files'), total=len(changedfiles) |
|
1422 | _(b'files'), unit=_(b'files'), total=len(changedfiles) | |
1379 | ) |
|
1423 | ) | |
1380 | for i, fname in enumerate(sorted(changedfiles)): |
|
1424 | for i, fname in enumerate(sorted(changedfiles)): | |
1381 | filerevlog = repo.file(fname) |
|
1425 | filerevlog = repo.file(fname) | |
1382 | if not filerevlog: |
|
1426 | if not filerevlog: | |
1383 | raise error.Abort( |
|
1427 | raise error.Abort( | |
1384 | _(b"empty or missing file data for %s") % fname |
|
1428 | _(b"empty or missing file data for %s") % fname | |
1385 | ) |
|
1429 | ) | |
1386 |
|
1430 | |||
1387 | clrevtolocalrev.clear() |
|
1431 | clrevtolocalrev.clear() | |
1388 |
|
1432 | |||
1389 | linkrevnodes = linknodes(filerevlog, fname) |
|
1433 | linkrevnodes = linknodes(filerevlog, fname) | |
1390 | # Lookup for filenodes, we collected the linkrev nodes above in the |
|
1434 | # Lookup for filenodes, we collected the linkrev nodes above in the | |
1391 | # fastpath case and with lookupmf in the slowpath case. |
|
1435 | # fastpath case and with lookupmf in the slowpath case. | |
1392 | def lookupfilelog(x): |
|
1436 | def lookupfilelog(x): | |
1393 | return linkrevnodes[x] |
|
1437 | return linkrevnodes[x] | |
1394 |
|
1438 | |||
1395 | frev, flr = filerevlog.rev, filerevlog.linkrev |
|
1439 | frev, flr = filerevlog.rev, filerevlog.linkrev | |
1396 | # Skip sending any filenode we know the client already |
|
1440 | # Skip sending any filenode we know the client already | |
1397 | # has. This avoids over-sending files relatively |
|
1441 | # has. This avoids over-sending files relatively | |
1398 | # inexpensively, so it's not a problem if we under-filter |
|
1442 | # inexpensively, so it's not a problem if we under-filter | |
1399 | # here. |
|
1443 | # here. | |
1400 | filenodes = [ |
|
1444 | filenodes = [ | |
1401 | n for n in linkrevnodes if flr(frev(n)) not in commonrevs |
|
1445 | n for n in linkrevnodes if flr(frev(n)) not in commonrevs | |
1402 | ] |
|
1446 | ] | |
1403 |
|
1447 | |||
1404 | if not filenodes: |
|
1448 | if not filenodes: | |
1405 | continue |
|
1449 | continue | |
1406 |
|
1450 | |||
1407 | progress.update(i + 1, item=fname) |
|
1451 | progress.update(i + 1, item=fname) | |
1408 |
|
1452 | |||
1409 | deltas = deltagroup( |
|
1453 | deltas = deltagroup( | |
1410 | self._repo, |
|
1454 | self._repo, | |
1411 | filerevlog, |
|
1455 | filerevlog, | |
1412 | filenodes, |
|
1456 | filenodes, | |
1413 | False, |
|
1457 | False, | |
1414 | lookupfilelog, |
|
1458 | lookupfilelog, | |
1415 | self._forcedeltaparentprev, |
|
1459 | self._forcedeltaparentprev, | |
1416 | ellipses=self._ellipses, |
|
1460 | ellipses=self._ellipses, | |
1417 | clrevtolocalrev=clrevtolocalrev, |
|
1461 | clrevtolocalrev=clrevtolocalrev, | |
1418 | fullclnodes=self._fullclnodes, |
|
1462 | fullclnodes=self._fullclnodes, | |
1419 | precomputedellipsis=self._precomputedellipsis, |
|
1463 | precomputedellipsis=self._precomputedellipsis, | |
|
1464 | sidedata_helpers=sidedata_helpers, | |||
1420 | ) |
|
1465 | ) | |
1421 |
|
1466 | |||
1422 | yield fname, deltas |
|
1467 | yield fname, deltas | |
1423 |
|
1468 | |||
1424 | progress.complete() |
|
1469 | progress.complete() | |
1425 |
|
1470 | |||
1426 |
|
1471 | |||
1427 | def _makecg1packer( |
|
1472 | def _makecg1packer( | |
1428 | repo, |
|
1473 | repo, | |
1429 | oldmatcher, |
|
1474 | oldmatcher, | |
1430 | matcher, |
|
1475 | matcher, | |
1431 | bundlecaps, |
|
1476 | bundlecaps, | |
1432 | ellipses=False, |
|
1477 | ellipses=False, | |
1433 | shallow=False, |
|
1478 | shallow=False, | |
1434 | ellipsisroots=None, |
|
1479 | ellipsisroots=None, | |
1435 | fullnodes=None, |
|
1480 | fullnodes=None, | |
1436 | remote_sidedata=None, |
|
1481 | remote_sidedata=None, | |
1437 | ): |
|
1482 | ): | |
1438 | builddeltaheader = lambda d: _CHANGEGROUPV1_DELTA_HEADER.pack( |
|
1483 | builddeltaheader = lambda d: _CHANGEGROUPV1_DELTA_HEADER.pack( | |
1439 | d.node, d.p1node, d.p2node, d.linknode |
|
1484 | d.node, d.p1node, d.p2node, d.linknode | |
1440 | ) |
|
1485 | ) | |
1441 |
|
1486 | |||
1442 | return cgpacker( |
|
1487 | return cgpacker( | |
1443 | repo, |
|
1488 | repo, | |
1444 | oldmatcher, |
|
1489 | oldmatcher, | |
1445 | matcher, |
|
1490 | matcher, | |
1446 | b'01', |
|
1491 | b'01', | |
1447 | builddeltaheader=builddeltaheader, |
|
1492 | builddeltaheader=builddeltaheader, | |
1448 | manifestsend=b'', |
|
1493 | manifestsend=b'', | |
1449 | forcedeltaparentprev=True, |
|
1494 | forcedeltaparentprev=True, | |
1450 | bundlecaps=bundlecaps, |
|
1495 | bundlecaps=bundlecaps, | |
1451 | ellipses=ellipses, |
|
1496 | ellipses=ellipses, | |
1452 | shallow=shallow, |
|
1497 | shallow=shallow, | |
1453 | ellipsisroots=ellipsisroots, |
|
1498 | ellipsisroots=ellipsisroots, | |
1454 | fullnodes=fullnodes, |
|
1499 | fullnodes=fullnodes, | |
1455 | ) |
|
1500 | ) | |
1456 |
|
1501 | |||
1457 |
|
1502 | |||
1458 | def _makecg2packer( |
|
1503 | def _makecg2packer( | |
1459 | repo, |
|
1504 | repo, | |
1460 | oldmatcher, |
|
1505 | oldmatcher, | |
1461 | matcher, |
|
1506 | matcher, | |
1462 | bundlecaps, |
|
1507 | bundlecaps, | |
1463 | ellipses=False, |
|
1508 | ellipses=False, | |
1464 | shallow=False, |
|
1509 | shallow=False, | |
1465 | ellipsisroots=None, |
|
1510 | ellipsisroots=None, | |
1466 | fullnodes=None, |
|
1511 | fullnodes=None, | |
1467 | remote_sidedata=None, |
|
1512 | remote_sidedata=None, | |
1468 | ): |
|
1513 | ): | |
1469 | builddeltaheader = lambda d: _CHANGEGROUPV2_DELTA_HEADER.pack( |
|
1514 | builddeltaheader = lambda d: _CHANGEGROUPV2_DELTA_HEADER.pack( | |
1470 | d.node, d.p1node, d.p2node, d.basenode, d.linknode |
|
1515 | d.node, d.p1node, d.p2node, d.basenode, d.linknode | |
1471 | ) |
|
1516 | ) | |
1472 |
|
1517 | |||
1473 | return cgpacker( |
|
1518 | return cgpacker( | |
1474 | repo, |
|
1519 | repo, | |
1475 | oldmatcher, |
|
1520 | oldmatcher, | |
1476 | matcher, |
|
1521 | matcher, | |
1477 | b'02', |
|
1522 | b'02', | |
1478 | builddeltaheader=builddeltaheader, |
|
1523 | builddeltaheader=builddeltaheader, | |
1479 | manifestsend=b'', |
|
1524 | manifestsend=b'', | |
1480 | bundlecaps=bundlecaps, |
|
1525 | bundlecaps=bundlecaps, | |
1481 | ellipses=ellipses, |
|
1526 | ellipses=ellipses, | |
1482 | shallow=shallow, |
|
1527 | shallow=shallow, | |
1483 | ellipsisroots=ellipsisroots, |
|
1528 | ellipsisroots=ellipsisroots, | |
1484 | fullnodes=fullnodes, |
|
1529 | fullnodes=fullnodes, | |
1485 | ) |
|
1530 | ) | |
1486 |
|
1531 | |||
1487 |
|
1532 | |||
1488 | def _makecg3packer( |
|
1533 | def _makecg3packer( | |
1489 | repo, |
|
1534 | repo, | |
1490 | oldmatcher, |
|
1535 | oldmatcher, | |
1491 | matcher, |
|
1536 | matcher, | |
1492 | bundlecaps, |
|
1537 | bundlecaps, | |
1493 | ellipses=False, |
|
1538 | ellipses=False, | |
1494 | shallow=False, |
|
1539 | shallow=False, | |
1495 | ellipsisroots=None, |
|
1540 | ellipsisroots=None, | |
1496 | fullnodes=None, |
|
1541 | fullnodes=None, | |
1497 | remote_sidedata=None, |
|
1542 | remote_sidedata=None, | |
1498 | ): |
|
1543 | ): | |
1499 | builddeltaheader = lambda d: _CHANGEGROUPV3_DELTA_HEADER.pack( |
|
1544 | builddeltaheader = lambda d: _CHANGEGROUPV3_DELTA_HEADER.pack( | |
1500 | d.node, d.p1node, d.p2node, d.basenode, d.linknode, d.flags |
|
1545 | d.node, d.p1node, d.p2node, d.basenode, d.linknode, d.flags | |
1501 | ) |
|
1546 | ) | |
1502 |
|
1547 | |||
1503 | return cgpacker( |
|
1548 | return cgpacker( | |
1504 | repo, |
|
1549 | repo, | |
1505 | oldmatcher, |
|
1550 | oldmatcher, | |
1506 | matcher, |
|
1551 | matcher, | |
1507 | b'03', |
|
1552 | b'03', | |
1508 | builddeltaheader=builddeltaheader, |
|
1553 | builddeltaheader=builddeltaheader, | |
1509 | manifestsend=closechunk(), |
|
1554 | manifestsend=closechunk(), | |
1510 | bundlecaps=bundlecaps, |
|
1555 | bundlecaps=bundlecaps, | |
1511 | ellipses=ellipses, |
|
1556 | ellipses=ellipses, | |
1512 | shallow=shallow, |
|
1557 | shallow=shallow, | |
1513 | ellipsisroots=ellipsisroots, |
|
1558 | ellipsisroots=ellipsisroots, | |
1514 | fullnodes=fullnodes, |
|
1559 | fullnodes=fullnodes, | |
1515 | ) |
|
1560 | ) | |
1516 |
|
1561 | |||
1517 |
|
1562 | |||
1518 | def _makecg4packer( |
|
1563 | def _makecg4packer( | |
1519 | repo, |
|
1564 | repo, | |
1520 | oldmatcher, |
|
1565 | oldmatcher, | |
1521 | matcher, |
|
1566 | matcher, | |
1522 | bundlecaps, |
|
1567 | bundlecaps, | |
1523 | ellipses=False, |
|
1568 | ellipses=False, | |
1524 | shallow=False, |
|
1569 | shallow=False, | |
1525 | ellipsisroots=None, |
|
1570 | ellipsisroots=None, | |
1526 | fullnodes=None, |
|
1571 | fullnodes=None, | |
1527 | remote_sidedata=None, |
|
1572 | remote_sidedata=None, | |
1528 | ): |
|
1573 | ): | |
1529 | # Same header func as cg3. Sidedata is in a separate chunk from the delta to |
|
1574 | # Same header func as cg3. Sidedata is in a separate chunk from the delta to | |
1530 | # differenciate "raw delta" and sidedata. |
|
1575 | # differenciate "raw delta" and sidedata. | |
1531 | builddeltaheader = lambda d: _CHANGEGROUPV3_DELTA_HEADER.pack( |
|
1576 | builddeltaheader = lambda d: _CHANGEGROUPV3_DELTA_HEADER.pack( | |
1532 | d.node, d.p1node, d.p2node, d.basenode, d.linknode, d.flags |
|
1577 | d.node, d.p1node, d.p2node, d.basenode, d.linknode, d.flags | |
1533 | ) |
|
1578 | ) | |
1534 |
|
1579 | |||
1535 | return cgpacker( |
|
1580 | return cgpacker( | |
1536 | repo, |
|
1581 | repo, | |
1537 | oldmatcher, |
|
1582 | oldmatcher, | |
1538 | matcher, |
|
1583 | matcher, | |
1539 | b'04', |
|
1584 | b'04', | |
1540 | builddeltaheader=builddeltaheader, |
|
1585 | builddeltaheader=builddeltaheader, | |
1541 | manifestsend=closechunk(), |
|
1586 | manifestsend=closechunk(), | |
1542 | bundlecaps=bundlecaps, |
|
1587 | bundlecaps=bundlecaps, | |
1543 | ellipses=ellipses, |
|
1588 | ellipses=ellipses, | |
1544 | shallow=shallow, |
|
1589 | shallow=shallow, | |
1545 | ellipsisroots=ellipsisroots, |
|
1590 | ellipsisroots=ellipsisroots, | |
1546 | fullnodes=fullnodes, |
|
1591 | fullnodes=fullnodes, | |
1547 | remote_sidedata=remote_sidedata, |
|
1592 | remote_sidedata=remote_sidedata, | |
1548 | ) |
|
1593 | ) | |
1549 |
|
1594 | |||
1550 |
|
1595 | |||
1551 | _packermap = { |
|
1596 | _packermap = { | |
1552 | b'01': (_makecg1packer, cg1unpacker), |
|
1597 | b'01': (_makecg1packer, cg1unpacker), | |
1553 | # cg2 adds support for exchanging generaldelta |
|
1598 | # cg2 adds support for exchanging generaldelta | |
1554 | b'02': (_makecg2packer, cg2unpacker), |
|
1599 | b'02': (_makecg2packer, cg2unpacker), | |
1555 | # cg3 adds support for exchanging revlog flags and treemanifests |
|
1600 | # cg3 adds support for exchanging revlog flags and treemanifests | |
1556 | b'03': (_makecg3packer, cg3unpacker), |
|
1601 | b'03': (_makecg3packer, cg3unpacker), | |
1557 | # ch4 adds support for exchanging sidedata |
|
1602 | # ch4 adds support for exchanging sidedata | |
1558 | b'04': (_makecg4packer, cg4unpacker), |
|
1603 | b'04': (_makecg4packer, cg4unpacker), | |
1559 | } |
|
1604 | } | |
1560 |
|
1605 | |||
1561 |
|
1606 | |||
1562 | def allsupportedversions(repo): |
|
1607 | def allsupportedversions(repo): | |
1563 | versions = set(_packermap.keys()) |
|
1608 | versions = set(_packermap.keys()) | |
1564 | needv03 = False |
|
1609 | needv03 = False | |
1565 | if ( |
|
1610 | if ( | |
1566 | repo.ui.configbool(b'experimental', b'changegroup3') |
|
1611 | repo.ui.configbool(b'experimental', b'changegroup3') | |
1567 | or repo.ui.configbool(b'experimental', b'treemanifest') |
|
1612 | or repo.ui.configbool(b'experimental', b'treemanifest') | |
1568 | or scmutil.istreemanifest(repo) |
|
1613 | or scmutil.istreemanifest(repo) | |
1569 | ): |
|
1614 | ): | |
1570 | # we keep version 03 because we need to to exchange treemanifest data |
|
1615 | # we keep version 03 because we need to to exchange treemanifest data | |
1571 | # |
|
1616 | # | |
1572 | # we also keep vresion 01 and 02, because it is possible for repo to |
|
1617 | # we also keep vresion 01 and 02, because it is possible for repo to | |
1573 | # contains both normal and tree manifest at the same time. so using |
|
1618 | # contains both normal and tree manifest at the same time. so using | |
1574 | # older version to pull data is viable |
|
1619 | # older version to pull data is viable | |
1575 | # |
|
1620 | # | |
1576 | # (or even to push subset of history) |
|
1621 | # (or even to push subset of history) | |
1577 | needv03 = True |
|
1622 | needv03 = True | |
1578 | has_revlogv2 = requirements.REVLOGV2_REQUIREMENT in repo.requirements |
|
1623 | has_revlogv2 = requirements.REVLOGV2_REQUIREMENT in repo.requirements | |
1579 | if not has_revlogv2: |
|
1624 | if not has_revlogv2: | |
1580 | versions.discard(b'04') |
|
1625 | versions.discard(b'04') | |
1581 | if not needv03: |
|
1626 | if not needv03: | |
1582 | versions.discard(b'03') |
|
1627 | versions.discard(b'03') | |
1583 | return versions |
|
1628 | return versions | |
1584 |
|
1629 | |||
1585 |
|
1630 | |||
1586 | # Changegroup versions that can be applied to the repo |
|
1631 | # Changegroup versions that can be applied to the repo | |
1587 | def supportedincomingversions(repo): |
|
1632 | def supportedincomingversions(repo): | |
1588 | return allsupportedversions(repo) |
|
1633 | return allsupportedversions(repo) | |
1589 |
|
1634 | |||
1590 |
|
1635 | |||
1591 | # Changegroup versions that can be created from the repo |
|
1636 | # Changegroup versions that can be created from the repo | |
1592 | def supportedoutgoingversions(repo): |
|
1637 | def supportedoutgoingversions(repo): | |
1593 | versions = allsupportedversions(repo) |
|
1638 | versions = allsupportedversions(repo) | |
1594 | if scmutil.istreemanifest(repo): |
|
1639 | if scmutil.istreemanifest(repo): | |
1595 | # Versions 01 and 02 support only flat manifests and it's just too |
|
1640 | # Versions 01 and 02 support only flat manifests and it's just too | |
1596 | # expensive to convert between the flat manifest and tree manifest on |
|
1641 | # expensive to convert between the flat manifest and tree manifest on | |
1597 | # the fly. Since tree manifests are hashed differently, all of history |
|
1642 | # the fly. Since tree manifests are hashed differently, all of history | |
1598 | # would have to be converted. Instead, we simply don't even pretend to |
|
1643 | # would have to be converted. Instead, we simply don't even pretend to | |
1599 | # support versions 01 and 02. |
|
1644 | # support versions 01 and 02. | |
1600 | versions.discard(b'01') |
|
1645 | versions.discard(b'01') | |
1601 | versions.discard(b'02') |
|
1646 | versions.discard(b'02') | |
1602 | if requirements.NARROW_REQUIREMENT in repo.requirements: |
|
1647 | if requirements.NARROW_REQUIREMENT in repo.requirements: | |
1603 | # Versions 01 and 02 don't support revlog flags, and we need to |
|
1648 | # Versions 01 and 02 don't support revlog flags, and we need to | |
1604 | # support that for stripping and unbundling to work. |
|
1649 | # support that for stripping and unbundling to work. | |
1605 | versions.discard(b'01') |
|
1650 | versions.discard(b'01') | |
1606 | versions.discard(b'02') |
|
1651 | versions.discard(b'02') | |
1607 | if LFS_REQUIREMENT in repo.requirements: |
|
1652 | if LFS_REQUIREMENT in repo.requirements: | |
1608 | # Versions 01 and 02 don't support revlog flags, and we need to |
|
1653 | # Versions 01 and 02 don't support revlog flags, and we need to | |
1609 | # mark LFS entries with REVIDX_EXTSTORED. |
|
1654 | # mark LFS entries with REVIDX_EXTSTORED. | |
1610 | versions.discard(b'01') |
|
1655 | versions.discard(b'01') | |
1611 | versions.discard(b'02') |
|
1656 | versions.discard(b'02') | |
1612 |
|
1657 | |||
1613 | return versions |
|
1658 | return versions | |
1614 |
|
1659 | |||
1615 |
|
1660 | |||
1616 | def localversion(repo): |
|
1661 | def localversion(repo): | |
1617 | # Finds the best version to use for bundles that are meant to be used |
|
1662 | # Finds the best version to use for bundles that are meant to be used | |
1618 | # locally, such as those from strip and shelve, and temporary bundles. |
|
1663 | # locally, such as those from strip and shelve, and temporary bundles. | |
1619 | return max(supportedoutgoingversions(repo)) |
|
1664 | return max(supportedoutgoingversions(repo)) | |
1620 |
|
1665 | |||
1621 |
|
1666 | |||
1622 | def safeversion(repo): |
|
1667 | def safeversion(repo): | |
1623 | # Finds the smallest version that it's safe to assume clients of the repo |
|
1668 | # Finds the smallest version that it's safe to assume clients of the repo | |
1624 | # will support. For example, all hg versions that support generaldelta also |
|
1669 | # will support. For example, all hg versions that support generaldelta also | |
1625 | # support changegroup 02. |
|
1670 | # support changegroup 02. | |
1626 | versions = supportedoutgoingversions(repo) |
|
1671 | versions = supportedoutgoingversions(repo) | |
1627 | if requirements.GENERALDELTA_REQUIREMENT in repo.requirements: |
|
1672 | if requirements.GENERALDELTA_REQUIREMENT in repo.requirements: | |
1628 | versions.discard(b'01') |
|
1673 | versions.discard(b'01') | |
1629 | assert versions |
|
1674 | assert versions | |
1630 | return min(versions) |
|
1675 | return min(versions) | |
1631 |
|
1676 | |||
1632 |
|
1677 | |||
1633 | def getbundler( |
|
1678 | def getbundler( | |
1634 | version, |
|
1679 | version, | |
1635 | repo, |
|
1680 | repo, | |
1636 | bundlecaps=None, |
|
1681 | bundlecaps=None, | |
1637 | oldmatcher=None, |
|
1682 | oldmatcher=None, | |
1638 | matcher=None, |
|
1683 | matcher=None, | |
1639 | ellipses=False, |
|
1684 | ellipses=False, | |
1640 | shallow=False, |
|
1685 | shallow=False, | |
1641 | ellipsisroots=None, |
|
1686 | ellipsisroots=None, | |
1642 | fullnodes=None, |
|
1687 | fullnodes=None, | |
1643 | remote_sidedata=None, |
|
1688 | remote_sidedata=None, | |
1644 | ): |
|
1689 | ): | |
1645 | assert version in supportedoutgoingversions(repo) |
|
1690 | assert version in supportedoutgoingversions(repo) | |
1646 |
|
1691 | |||
1647 | if matcher is None: |
|
1692 | if matcher is None: | |
1648 | matcher = matchmod.always() |
|
1693 | matcher = matchmod.always() | |
1649 | if oldmatcher is None: |
|
1694 | if oldmatcher is None: | |
1650 | oldmatcher = matchmod.never() |
|
1695 | oldmatcher = matchmod.never() | |
1651 |
|
1696 | |||
1652 | if version == b'01' and not matcher.always(): |
|
1697 | if version == b'01' and not matcher.always(): | |
1653 | raise error.ProgrammingError( |
|
1698 | raise error.ProgrammingError( | |
1654 | b'version 01 changegroups do not support sparse file matchers' |
|
1699 | b'version 01 changegroups do not support sparse file matchers' | |
1655 | ) |
|
1700 | ) | |
1656 |
|
1701 | |||
1657 | if ellipses and version in (b'01', b'02'): |
|
1702 | if ellipses and version in (b'01', b'02'): | |
1658 | raise error.Abort( |
|
1703 | raise error.Abort( | |
1659 | _( |
|
1704 | _( | |
1660 | b'ellipsis nodes require at least cg3 on client and server, ' |
|
1705 | b'ellipsis nodes require at least cg3 on client and server, ' | |
1661 | b'but negotiated version %s' |
|
1706 | b'but negotiated version %s' | |
1662 | ) |
|
1707 | ) | |
1663 | % version |
|
1708 | % version | |
1664 | ) |
|
1709 | ) | |
1665 |
|
1710 | |||
1666 | # Requested files could include files not in the local store. So |
|
1711 | # Requested files could include files not in the local store. So | |
1667 | # filter those out. |
|
1712 | # filter those out. | |
1668 | matcher = repo.narrowmatch(matcher) |
|
1713 | matcher = repo.narrowmatch(matcher) | |
1669 |
|
1714 | |||
1670 | fn = _packermap[version][0] |
|
1715 | fn = _packermap[version][0] | |
1671 | return fn( |
|
1716 | return fn( | |
1672 | repo, |
|
1717 | repo, | |
1673 | oldmatcher, |
|
1718 | oldmatcher, | |
1674 | matcher, |
|
1719 | matcher, | |
1675 | bundlecaps, |
|
1720 | bundlecaps, | |
1676 | ellipses=ellipses, |
|
1721 | ellipses=ellipses, | |
1677 | shallow=shallow, |
|
1722 | shallow=shallow, | |
1678 | ellipsisroots=ellipsisroots, |
|
1723 | ellipsisroots=ellipsisroots, | |
1679 | fullnodes=fullnodes, |
|
1724 | fullnodes=fullnodes, | |
1680 | remote_sidedata=remote_sidedata, |
|
1725 | remote_sidedata=remote_sidedata, | |
1681 | ) |
|
1726 | ) | |
1682 |
|
1727 | |||
1683 |
|
1728 | |||
1684 | def getunbundler(version, fh, alg, extras=None): |
|
1729 | def getunbundler(version, fh, alg, extras=None): | |
1685 | return _packermap[version][1](fh, alg, extras=extras) |
|
1730 | return _packermap[version][1](fh, alg, extras=extras) | |
1686 |
|
1731 | |||
1687 |
|
1732 | |||
1688 | def _changegroupinfo(repo, nodes, source): |
|
1733 | def _changegroupinfo(repo, nodes, source): | |
1689 | if repo.ui.verbose or source == b'bundle': |
|
1734 | if repo.ui.verbose or source == b'bundle': | |
1690 | repo.ui.status(_(b"%d changesets found\n") % len(nodes)) |
|
1735 | repo.ui.status(_(b"%d changesets found\n") % len(nodes)) | |
1691 | if repo.ui.debugflag: |
|
1736 | if repo.ui.debugflag: | |
1692 | repo.ui.debug(b"list of changesets:\n") |
|
1737 | repo.ui.debug(b"list of changesets:\n") | |
1693 | for node in nodes: |
|
1738 | for node in nodes: | |
1694 | repo.ui.debug(b"%s\n" % hex(node)) |
|
1739 | repo.ui.debug(b"%s\n" % hex(node)) | |
1695 |
|
1740 | |||
1696 |
|
1741 | |||
1697 | def makechangegroup( |
|
1742 | def makechangegroup( | |
1698 | repo, outgoing, version, source, fastpath=False, bundlecaps=None |
|
1743 | repo, outgoing, version, source, fastpath=False, bundlecaps=None | |
1699 | ): |
|
1744 | ): | |
1700 | cgstream = makestream( |
|
1745 | cgstream = makestream( | |
1701 | repo, |
|
1746 | repo, | |
1702 | outgoing, |
|
1747 | outgoing, | |
1703 | version, |
|
1748 | version, | |
1704 | source, |
|
1749 | source, | |
1705 | fastpath=fastpath, |
|
1750 | fastpath=fastpath, | |
1706 | bundlecaps=bundlecaps, |
|
1751 | bundlecaps=bundlecaps, | |
1707 | ) |
|
1752 | ) | |
1708 | return getunbundler( |
|
1753 | return getunbundler( | |
1709 | version, |
|
1754 | version, | |
1710 | util.chunkbuffer(cgstream), |
|
1755 | util.chunkbuffer(cgstream), | |
1711 | None, |
|
1756 | None, | |
1712 | {b'clcount': len(outgoing.missing)}, |
|
1757 | {b'clcount': len(outgoing.missing)}, | |
1713 | ) |
|
1758 | ) | |
1714 |
|
1759 | |||
1715 |
|
1760 | |||
1716 | def makestream( |
|
1761 | def makestream( | |
1717 | repo, |
|
1762 | repo, | |
1718 | outgoing, |
|
1763 | outgoing, | |
1719 | version, |
|
1764 | version, | |
1720 | source, |
|
1765 | source, | |
1721 | fastpath=False, |
|
1766 | fastpath=False, | |
1722 | bundlecaps=None, |
|
1767 | bundlecaps=None, | |
1723 | matcher=None, |
|
1768 | matcher=None, | |
1724 | remote_sidedata=None, |
|
1769 | remote_sidedata=None, | |
1725 | ): |
|
1770 | ): | |
1726 | bundler = getbundler( |
|
1771 | bundler = getbundler( | |
1727 | version, |
|
1772 | version, | |
1728 | repo, |
|
1773 | repo, | |
1729 | bundlecaps=bundlecaps, |
|
1774 | bundlecaps=bundlecaps, | |
1730 | matcher=matcher, |
|
1775 | matcher=matcher, | |
1731 | remote_sidedata=remote_sidedata, |
|
1776 | remote_sidedata=remote_sidedata, | |
1732 | ) |
|
1777 | ) | |
1733 |
|
1778 | |||
1734 | repo = repo.unfiltered() |
|
1779 | repo = repo.unfiltered() | |
1735 | commonrevs = outgoing.common |
|
1780 | commonrevs = outgoing.common | |
1736 | csets = outgoing.missing |
|
1781 | csets = outgoing.missing | |
1737 | heads = outgoing.ancestorsof |
|
1782 | heads = outgoing.ancestorsof | |
1738 | # We go through the fast path if we get told to, or if all (unfiltered |
|
1783 | # We go through the fast path if we get told to, or if all (unfiltered | |
1739 | # heads have been requested (since we then know there all linkrevs will |
|
1784 | # heads have been requested (since we then know there all linkrevs will | |
1740 | # be pulled by the client). |
|
1785 | # be pulled by the client). | |
1741 | heads.sort() |
|
1786 | heads.sort() | |
1742 | fastpathlinkrev = fastpath or ( |
|
1787 | fastpathlinkrev = fastpath or ( | |
1743 | repo.filtername is None and heads == sorted(repo.heads()) |
|
1788 | repo.filtername is None and heads == sorted(repo.heads()) | |
1744 | ) |
|
1789 | ) | |
1745 |
|
1790 | |||
1746 | repo.hook(b'preoutgoing', throw=True, source=source) |
|
1791 | repo.hook(b'preoutgoing', throw=True, source=source) | |
1747 | _changegroupinfo(repo, csets, source) |
|
1792 | _changegroupinfo(repo, csets, source) | |
1748 | return bundler.generate(commonrevs, csets, fastpathlinkrev, source) |
|
1793 | return bundler.generate(commonrevs, csets, fastpathlinkrev, source) | |
1749 |
|
1794 | |||
1750 |
|
1795 | |||
1751 | def _addchangegroupfiles(repo, source, revmap, trp, expectedfiles, needfiles): |
|
1796 | def _addchangegroupfiles(repo, source, revmap, trp, expectedfiles, needfiles): | |
1752 | revisions = 0 |
|
1797 | revisions = 0 | |
1753 | files = 0 |
|
1798 | files = 0 | |
1754 | progress = repo.ui.makeprogress( |
|
1799 | progress = repo.ui.makeprogress( | |
1755 | _(b'files'), unit=_(b'files'), total=expectedfiles |
|
1800 | _(b'files'), unit=_(b'files'), total=expectedfiles | |
1756 | ) |
|
1801 | ) | |
1757 | for chunkdata in iter(source.filelogheader, {}): |
|
1802 | for chunkdata in iter(source.filelogheader, {}): | |
1758 | files += 1 |
|
1803 | files += 1 | |
1759 | f = chunkdata[b"filename"] |
|
1804 | f = chunkdata[b"filename"] | |
1760 | repo.ui.debug(b"adding %s revisions\n" % f) |
|
1805 | repo.ui.debug(b"adding %s revisions\n" % f) | |
1761 | progress.increment() |
|
1806 | progress.increment() | |
1762 | fl = repo.file(f) |
|
1807 | fl = repo.file(f) | |
1763 | o = len(fl) |
|
1808 | o = len(fl) | |
1764 | try: |
|
1809 | try: | |
1765 | deltas = source.deltaiter() |
|
1810 | deltas = source.deltaiter() | |
1766 | if not fl.addgroup(deltas, revmap, trp): |
|
1811 | if not fl.addgroup(deltas, revmap, trp): | |
1767 | raise error.Abort(_(b"received file revlog group is empty")) |
|
1812 | raise error.Abort(_(b"received file revlog group is empty")) | |
1768 | except error.CensoredBaseError as e: |
|
1813 | except error.CensoredBaseError as e: | |
1769 | raise error.Abort(_(b"received delta base is censored: %s") % e) |
|
1814 | raise error.Abort(_(b"received delta base is censored: %s") % e) | |
1770 | revisions += len(fl) - o |
|
1815 | revisions += len(fl) - o | |
1771 | if f in needfiles: |
|
1816 | if f in needfiles: | |
1772 | needs = needfiles[f] |
|
1817 | needs = needfiles[f] | |
1773 | for new in pycompat.xrange(o, len(fl)): |
|
1818 | for new in pycompat.xrange(o, len(fl)): | |
1774 | n = fl.node(new) |
|
1819 | n = fl.node(new) | |
1775 | if n in needs: |
|
1820 | if n in needs: | |
1776 | needs.remove(n) |
|
1821 | needs.remove(n) | |
1777 | else: |
|
1822 | else: | |
1778 | raise error.Abort(_(b"received spurious file revlog entry")) |
|
1823 | raise error.Abort(_(b"received spurious file revlog entry")) | |
1779 | if not needs: |
|
1824 | if not needs: | |
1780 | del needfiles[f] |
|
1825 | del needfiles[f] | |
1781 | progress.complete() |
|
1826 | progress.complete() | |
1782 |
|
1827 | |||
1783 | for f, needs in pycompat.iteritems(needfiles): |
|
1828 | for f, needs in pycompat.iteritems(needfiles): | |
1784 | fl = repo.file(f) |
|
1829 | fl = repo.file(f) | |
1785 | for n in needs: |
|
1830 | for n in needs: | |
1786 | try: |
|
1831 | try: | |
1787 | fl.rev(n) |
|
1832 | fl.rev(n) | |
1788 | except error.LookupError: |
|
1833 | except error.LookupError: | |
1789 | raise error.Abort( |
|
1834 | raise error.Abort( | |
1790 | _(b'missing file data for %s:%s - run hg verify') |
|
1835 | _(b'missing file data for %s:%s - run hg verify') | |
1791 | % (f, hex(n)) |
|
1836 | % (f, hex(n)) | |
1792 | ) |
|
1837 | ) | |
1793 |
|
1838 | |||
1794 | return revisions, files |
|
1839 | return revisions, files | |
|
1840 | ||||
|
1841 | ||||
|
1842 | def get_sidedata_helpers(repo, remote_sd_categories, pull=False): | |||
|
1843 | # Computers for computing sidedata on-the-fly | |||
|
1844 | sd_computers = collections.defaultdict(list) | |||
|
1845 | # Computers for categories to remove from sidedata | |||
|
1846 | sd_removers = collections.defaultdict(list) | |||
|
1847 | ||||
|
1848 | to_generate = remote_sd_categories - repo._wanted_sidedata | |||
|
1849 | to_remove = repo._wanted_sidedata - remote_sd_categories | |||
|
1850 | if pull: | |||
|
1851 | to_generate, to_remove = to_remove, to_generate | |||
|
1852 | ||||
|
1853 | for revlog_kind, computers in repo._sidedata_computers.items(): | |||
|
1854 | for category, computer in computers.items(): | |||
|
1855 | if category in to_generate: | |||
|
1856 | sd_computers[revlog_kind].append(computer) | |||
|
1857 | if category in to_remove: | |||
|
1858 | sd_removers[revlog_kind].append(computer) | |||
|
1859 | ||||
|
1860 | sidedata_helpers = (repo, sd_computers, sd_removers) | |||
|
1861 | return sidedata_helpers |
@@ -1,292 +1,294 b'' | |||||
1 | # filelog.py - file history class for mercurial |
|
1 | # filelog.py - file history class for mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | from .i18n import _ |
|
10 | from .i18n import _ | |
11 | from .node import ( |
|
11 | from .node import ( | |
12 | nullid, |
|
12 | nullid, | |
13 | nullrev, |
|
13 | nullrev, | |
14 | ) |
|
14 | ) | |
15 | from . import ( |
|
15 | from . import ( | |
16 | error, |
|
16 | error, | |
17 | revlog, |
|
17 | revlog, | |
18 | ) |
|
18 | ) | |
19 | from .interfaces import ( |
|
19 | from .interfaces import ( | |
20 | repository, |
|
20 | repository, | |
21 | util as interfaceutil, |
|
21 | util as interfaceutil, | |
22 | ) |
|
22 | ) | |
23 | from .utils import storageutil |
|
23 | from .utils import storageutil | |
24 |
|
24 | |||
25 |
|
25 | |||
26 | @interfaceutil.implementer(repository.ifilestorage) |
|
26 | @interfaceutil.implementer(repository.ifilestorage) | |
27 | class filelog(object): |
|
27 | class filelog(object): | |
28 | def __init__(self, opener, path): |
|
28 | def __init__(self, opener, path): | |
29 | self._revlog = revlog.revlog( |
|
29 | self._revlog = revlog.revlog( | |
30 | opener, b'/'.join((b'data', path + b'.i')), censorable=True |
|
30 | opener, b'/'.join((b'data', path + b'.i')), censorable=True | |
31 | ) |
|
31 | ) | |
32 | # Full name of the user visible file, relative to the repository root. |
|
32 | # Full name of the user visible file, relative to the repository root. | |
33 | # Used by LFS. |
|
33 | # Used by LFS. | |
34 | self._revlog.filename = path |
|
34 | self._revlog.filename = path | |
35 | self._revlog.revlog_kind = b'filelog' |
|
35 | self._revlog.revlog_kind = b'filelog' | |
36 |
|
36 | |||
37 | def __len__(self): |
|
37 | def __len__(self): | |
38 | return len(self._revlog) |
|
38 | return len(self._revlog) | |
39 |
|
39 | |||
40 | def __iter__(self): |
|
40 | def __iter__(self): | |
41 | return self._revlog.__iter__() |
|
41 | return self._revlog.__iter__() | |
42 |
|
42 | |||
43 | def hasnode(self, node): |
|
43 | def hasnode(self, node): | |
44 | if node in (nullid, nullrev): |
|
44 | if node in (nullid, nullrev): | |
45 | return False |
|
45 | return False | |
46 |
|
46 | |||
47 | try: |
|
47 | try: | |
48 | self._revlog.rev(node) |
|
48 | self._revlog.rev(node) | |
49 | return True |
|
49 | return True | |
50 | except (TypeError, ValueError, IndexError, error.LookupError): |
|
50 | except (TypeError, ValueError, IndexError, error.LookupError): | |
51 | return False |
|
51 | return False | |
52 |
|
52 | |||
53 | def revs(self, start=0, stop=None): |
|
53 | def revs(self, start=0, stop=None): | |
54 | return self._revlog.revs(start=start, stop=stop) |
|
54 | return self._revlog.revs(start=start, stop=stop) | |
55 |
|
55 | |||
56 | def parents(self, node): |
|
56 | def parents(self, node): | |
57 | return self._revlog.parents(node) |
|
57 | return self._revlog.parents(node) | |
58 |
|
58 | |||
59 | def parentrevs(self, rev): |
|
59 | def parentrevs(self, rev): | |
60 | return self._revlog.parentrevs(rev) |
|
60 | return self._revlog.parentrevs(rev) | |
61 |
|
61 | |||
62 | def rev(self, node): |
|
62 | def rev(self, node): | |
63 | return self._revlog.rev(node) |
|
63 | return self._revlog.rev(node) | |
64 |
|
64 | |||
65 | def node(self, rev): |
|
65 | def node(self, rev): | |
66 | return self._revlog.node(rev) |
|
66 | return self._revlog.node(rev) | |
67 |
|
67 | |||
68 | def lookup(self, node): |
|
68 | def lookup(self, node): | |
69 | return storageutil.fileidlookup( |
|
69 | return storageutil.fileidlookup( | |
70 | self._revlog, node, self._revlog.indexfile |
|
70 | self._revlog, node, self._revlog.indexfile | |
71 | ) |
|
71 | ) | |
72 |
|
72 | |||
73 | def linkrev(self, rev): |
|
73 | def linkrev(self, rev): | |
74 | return self._revlog.linkrev(rev) |
|
74 | return self._revlog.linkrev(rev) | |
75 |
|
75 | |||
76 | def commonancestorsheads(self, node1, node2): |
|
76 | def commonancestorsheads(self, node1, node2): | |
77 | return self._revlog.commonancestorsheads(node1, node2) |
|
77 | return self._revlog.commonancestorsheads(node1, node2) | |
78 |
|
78 | |||
79 | # Used by dagop.blockdescendants(). |
|
79 | # Used by dagop.blockdescendants(). | |
80 | def descendants(self, revs): |
|
80 | def descendants(self, revs): | |
81 | return self._revlog.descendants(revs) |
|
81 | return self._revlog.descendants(revs) | |
82 |
|
82 | |||
83 | def heads(self, start=None, stop=None): |
|
83 | def heads(self, start=None, stop=None): | |
84 | return self._revlog.heads(start, stop) |
|
84 | return self._revlog.heads(start, stop) | |
85 |
|
85 | |||
86 | # Used by hgweb, children extension. |
|
86 | # Used by hgweb, children extension. | |
87 | def children(self, node): |
|
87 | def children(self, node): | |
88 | return self._revlog.children(node) |
|
88 | return self._revlog.children(node) | |
89 |
|
89 | |||
90 | def iscensored(self, rev): |
|
90 | def iscensored(self, rev): | |
91 | return self._revlog.iscensored(rev) |
|
91 | return self._revlog.iscensored(rev) | |
92 |
|
92 | |||
93 | def revision(self, node, _df=None, raw=False): |
|
93 | def revision(self, node, _df=None, raw=False): | |
94 | return self._revlog.revision(node, _df=_df, raw=raw) |
|
94 | return self._revlog.revision(node, _df=_df, raw=raw) | |
95 |
|
95 | |||
96 | def rawdata(self, node, _df=None): |
|
96 | def rawdata(self, node, _df=None): | |
97 | return self._revlog.rawdata(node, _df=_df) |
|
97 | return self._revlog.rawdata(node, _df=_df) | |
98 |
|
98 | |||
99 | def emitrevisions( |
|
99 | def emitrevisions( | |
100 | self, |
|
100 | self, | |
101 | nodes, |
|
101 | nodes, | |
102 | nodesorder=None, |
|
102 | nodesorder=None, | |
103 | revisiondata=False, |
|
103 | revisiondata=False, | |
104 | assumehaveparentrevisions=False, |
|
104 | assumehaveparentrevisions=False, | |
105 | deltamode=repository.CG_DELTAMODE_STD, |
|
105 | deltamode=repository.CG_DELTAMODE_STD, | |
|
106 | sidedata_helpers=None, | |||
106 | ): |
|
107 | ): | |
107 | return self._revlog.emitrevisions( |
|
108 | return self._revlog.emitrevisions( | |
108 | nodes, |
|
109 | nodes, | |
109 | nodesorder=nodesorder, |
|
110 | nodesorder=nodesorder, | |
110 | revisiondata=revisiondata, |
|
111 | revisiondata=revisiondata, | |
111 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
112 | assumehaveparentrevisions=assumehaveparentrevisions, | |
112 | deltamode=deltamode, |
|
113 | deltamode=deltamode, | |
|
114 | sidedata_helpers=sidedata_helpers, | |||
113 | ) |
|
115 | ) | |
114 |
|
116 | |||
115 | def addrevision( |
|
117 | def addrevision( | |
116 | self, |
|
118 | self, | |
117 | revisiondata, |
|
119 | revisiondata, | |
118 | transaction, |
|
120 | transaction, | |
119 | linkrev, |
|
121 | linkrev, | |
120 | p1, |
|
122 | p1, | |
121 | p2, |
|
123 | p2, | |
122 | node=None, |
|
124 | node=None, | |
123 | flags=revlog.REVIDX_DEFAULT_FLAGS, |
|
125 | flags=revlog.REVIDX_DEFAULT_FLAGS, | |
124 | cachedelta=None, |
|
126 | cachedelta=None, | |
125 | ): |
|
127 | ): | |
126 | return self._revlog.addrevision( |
|
128 | return self._revlog.addrevision( | |
127 | revisiondata, |
|
129 | revisiondata, | |
128 | transaction, |
|
130 | transaction, | |
129 | linkrev, |
|
131 | linkrev, | |
130 | p1, |
|
132 | p1, | |
131 | p2, |
|
133 | p2, | |
132 | node=node, |
|
134 | node=node, | |
133 | flags=flags, |
|
135 | flags=flags, | |
134 | cachedelta=cachedelta, |
|
136 | cachedelta=cachedelta, | |
135 | ) |
|
137 | ) | |
136 |
|
138 | |||
137 | def addgroup( |
|
139 | def addgroup( | |
138 | self, |
|
140 | self, | |
139 | deltas, |
|
141 | deltas, | |
140 | linkmapper, |
|
142 | linkmapper, | |
141 | transaction, |
|
143 | transaction, | |
142 | addrevisioncb=None, |
|
144 | addrevisioncb=None, | |
143 | duplicaterevisioncb=None, |
|
145 | duplicaterevisioncb=None, | |
144 | maybemissingparents=False, |
|
146 | maybemissingparents=False, | |
145 | ): |
|
147 | ): | |
146 | if maybemissingparents: |
|
148 | if maybemissingparents: | |
147 | raise error.Abort( |
|
149 | raise error.Abort( | |
148 | _( |
|
150 | _( | |
149 | b'revlog storage does not support missing ' |
|
151 | b'revlog storage does not support missing ' | |
150 | b'parents write mode' |
|
152 | b'parents write mode' | |
151 | ) |
|
153 | ) | |
152 | ) |
|
154 | ) | |
153 |
|
155 | |||
154 | return self._revlog.addgroup( |
|
156 | return self._revlog.addgroup( | |
155 | deltas, |
|
157 | deltas, | |
156 | linkmapper, |
|
158 | linkmapper, | |
157 | transaction, |
|
159 | transaction, | |
158 | addrevisioncb=addrevisioncb, |
|
160 | addrevisioncb=addrevisioncb, | |
159 | duplicaterevisioncb=duplicaterevisioncb, |
|
161 | duplicaterevisioncb=duplicaterevisioncb, | |
160 | ) |
|
162 | ) | |
161 |
|
163 | |||
162 | def getstrippoint(self, minlink): |
|
164 | def getstrippoint(self, minlink): | |
163 | return self._revlog.getstrippoint(minlink) |
|
165 | return self._revlog.getstrippoint(minlink) | |
164 |
|
166 | |||
165 | def strip(self, minlink, transaction): |
|
167 | def strip(self, minlink, transaction): | |
166 | return self._revlog.strip(minlink, transaction) |
|
168 | return self._revlog.strip(minlink, transaction) | |
167 |
|
169 | |||
168 | def censorrevision(self, tr, node, tombstone=b''): |
|
170 | def censorrevision(self, tr, node, tombstone=b''): | |
169 | return self._revlog.censorrevision(tr, node, tombstone=tombstone) |
|
171 | return self._revlog.censorrevision(tr, node, tombstone=tombstone) | |
170 |
|
172 | |||
171 | def files(self): |
|
173 | def files(self): | |
172 | return self._revlog.files() |
|
174 | return self._revlog.files() | |
173 |
|
175 | |||
174 | def read(self, node): |
|
176 | def read(self, node): | |
175 | return storageutil.filtermetadata(self.revision(node)) |
|
177 | return storageutil.filtermetadata(self.revision(node)) | |
176 |
|
178 | |||
177 | def add(self, text, meta, transaction, link, p1=None, p2=None): |
|
179 | def add(self, text, meta, transaction, link, p1=None, p2=None): | |
178 | if meta or text.startswith(b'\1\n'): |
|
180 | if meta or text.startswith(b'\1\n'): | |
179 | text = storageutil.packmeta(meta, text) |
|
181 | text = storageutil.packmeta(meta, text) | |
180 | rev = self.addrevision(text, transaction, link, p1, p2) |
|
182 | rev = self.addrevision(text, transaction, link, p1, p2) | |
181 | return self.node(rev) |
|
183 | return self.node(rev) | |
182 |
|
184 | |||
183 | def renamed(self, node): |
|
185 | def renamed(self, node): | |
184 | return storageutil.filerevisioncopied(self, node) |
|
186 | return storageutil.filerevisioncopied(self, node) | |
185 |
|
187 | |||
186 | def size(self, rev): |
|
188 | def size(self, rev): | |
187 | """return the size of a given revision""" |
|
189 | """return the size of a given revision""" | |
188 |
|
190 | |||
189 | # for revisions with renames, we have to go the slow way |
|
191 | # for revisions with renames, we have to go the slow way | |
190 | node = self.node(rev) |
|
192 | node = self.node(rev) | |
191 | if self.renamed(node): |
|
193 | if self.renamed(node): | |
192 | return len(self.read(node)) |
|
194 | return len(self.read(node)) | |
193 | if self.iscensored(rev): |
|
195 | if self.iscensored(rev): | |
194 | return 0 |
|
196 | return 0 | |
195 |
|
197 | |||
196 | # XXX if self.read(node).startswith("\1\n"), this returns (size+4) |
|
198 | # XXX if self.read(node).startswith("\1\n"), this returns (size+4) | |
197 | return self._revlog.size(rev) |
|
199 | return self._revlog.size(rev) | |
198 |
|
200 | |||
199 | def cmp(self, node, text): |
|
201 | def cmp(self, node, text): | |
200 | """compare text with a given file revision |
|
202 | """compare text with a given file revision | |
201 |
|
203 | |||
202 | returns True if text is different than what is stored. |
|
204 | returns True if text is different than what is stored. | |
203 | """ |
|
205 | """ | |
204 | return not storageutil.filedataequivalent(self, node, text) |
|
206 | return not storageutil.filedataequivalent(self, node, text) | |
205 |
|
207 | |||
206 | def verifyintegrity(self, state): |
|
208 | def verifyintegrity(self, state): | |
207 | return self._revlog.verifyintegrity(state) |
|
209 | return self._revlog.verifyintegrity(state) | |
208 |
|
210 | |||
209 | def storageinfo( |
|
211 | def storageinfo( | |
210 | self, |
|
212 | self, | |
211 | exclusivefiles=False, |
|
213 | exclusivefiles=False, | |
212 | sharedfiles=False, |
|
214 | sharedfiles=False, | |
213 | revisionscount=False, |
|
215 | revisionscount=False, | |
214 | trackedsize=False, |
|
216 | trackedsize=False, | |
215 | storedsize=False, |
|
217 | storedsize=False, | |
216 | ): |
|
218 | ): | |
217 | return self._revlog.storageinfo( |
|
219 | return self._revlog.storageinfo( | |
218 | exclusivefiles=exclusivefiles, |
|
220 | exclusivefiles=exclusivefiles, | |
219 | sharedfiles=sharedfiles, |
|
221 | sharedfiles=sharedfiles, | |
220 | revisionscount=revisionscount, |
|
222 | revisionscount=revisionscount, | |
221 | trackedsize=trackedsize, |
|
223 | trackedsize=trackedsize, | |
222 | storedsize=storedsize, |
|
224 | storedsize=storedsize, | |
223 | ) |
|
225 | ) | |
224 |
|
226 | |||
225 | # TODO these aren't part of the interface and aren't internal methods. |
|
227 | # TODO these aren't part of the interface and aren't internal methods. | |
226 | # Callers should be fixed to not use them. |
|
228 | # Callers should be fixed to not use them. | |
227 |
|
229 | |||
228 | # Used by bundlefilelog, unionfilelog. |
|
230 | # Used by bundlefilelog, unionfilelog. | |
229 | @property |
|
231 | @property | |
230 | def indexfile(self): |
|
232 | def indexfile(self): | |
231 | return self._revlog.indexfile |
|
233 | return self._revlog.indexfile | |
232 |
|
234 | |||
233 | @indexfile.setter |
|
235 | @indexfile.setter | |
234 | def indexfile(self, value): |
|
236 | def indexfile(self, value): | |
235 | self._revlog.indexfile = value |
|
237 | self._revlog.indexfile = value | |
236 |
|
238 | |||
237 | # Used by repo upgrade. |
|
239 | # Used by repo upgrade. | |
238 | def clone(self, tr, destrevlog, **kwargs): |
|
240 | def clone(self, tr, destrevlog, **kwargs): | |
239 | if not isinstance(destrevlog, filelog): |
|
241 | if not isinstance(destrevlog, filelog): | |
240 | raise error.ProgrammingError(b'expected filelog to clone()') |
|
242 | raise error.ProgrammingError(b'expected filelog to clone()') | |
241 |
|
243 | |||
242 | return self._revlog.clone(tr, destrevlog._revlog, **kwargs) |
|
244 | return self._revlog.clone(tr, destrevlog._revlog, **kwargs) | |
243 |
|
245 | |||
244 |
|
246 | |||
245 | class narrowfilelog(filelog): |
|
247 | class narrowfilelog(filelog): | |
246 | """Filelog variation to be used with narrow stores.""" |
|
248 | """Filelog variation to be used with narrow stores.""" | |
247 |
|
249 | |||
248 | def __init__(self, opener, path, narrowmatch): |
|
250 | def __init__(self, opener, path, narrowmatch): | |
249 | super(narrowfilelog, self).__init__(opener, path) |
|
251 | super(narrowfilelog, self).__init__(opener, path) | |
250 | self._narrowmatch = narrowmatch |
|
252 | self._narrowmatch = narrowmatch | |
251 |
|
253 | |||
252 | def renamed(self, node): |
|
254 | def renamed(self, node): | |
253 | res = super(narrowfilelog, self).renamed(node) |
|
255 | res = super(narrowfilelog, self).renamed(node) | |
254 |
|
256 | |||
255 | # Renames that come from outside the narrowspec are problematic |
|
257 | # Renames that come from outside the narrowspec are problematic | |
256 | # because we may lack the base text for the rename. This can result |
|
258 | # because we may lack the base text for the rename. This can result | |
257 | # in code attempting to walk the ancestry or compute a diff |
|
259 | # in code attempting to walk the ancestry or compute a diff | |
258 | # encountering a missing revision. We address this by silently |
|
260 | # encountering a missing revision. We address this by silently | |
259 | # removing rename metadata if the source file is outside the |
|
261 | # removing rename metadata if the source file is outside the | |
260 | # narrow spec. |
|
262 | # narrow spec. | |
261 | # |
|
263 | # | |
262 | # A better solution would be to see if the base revision is available, |
|
264 | # A better solution would be to see if the base revision is available, | |
263 | # rather than assuming it isn't. |
|
265 | # rather than assuming it isn't. | |
264 | # |
|
266 | # | |
265 | # An even better solution would be to teach all consumers of rename |
|
267 | # An even better solution would be to teach all consumers of rename | |
266 | # metadata that the base revision may not be available. |
|
268 | # metadata that the base revision may not be available. | |
267 | # |
|
269 | # | |
268 | # TODO consider better ways of doing this. |
|
270 | # TODO consider better ways of doing this. | |
269 | if res and not self._narrowmatch(res[0]): |
|
271 | if res and not self._narrowmatch(res[0]): | |
270 | return None |
|
272 | return None | |
271 |
|
273 | |||
272 | return res |
|
274 | return res | |
273 |
|
275 | |||
274 | def size(self, rev): |
|
276 | def size(self, rev): | |
275 | # Because we have a custom renamed() that may lie, we need to call |
|
277 | # Because we have a custom renamed() that may lie, we need to call | |
276 | # the base renamed() to report accurate results. |
|
278 | # the base renamed() to report accurate results. | |
277 | node = self.node(rev) |
|
279 | node = self.node(rev) | |
278 | if super(narrowfilelog, self).renamed(node): |
|
280 | if super(narrowfilelog, self).renamed(node): | |
279 | return len(self.read(node)) |
|
281 | return len(self.read(node)) | |
280 | else: |
|
282 | else: | |
281 | return super(narrowfilelog, self).size(rev) |
|
283 | return super(narrowfilelog, self).size(rev) | |
282 |
|
284 | |||
283 | def cmp(self, node, text): |
|
285 | def cmp(self, node, text): | |
284 | # We don't call `super` because narrow parents can be buggy in case of a |
|
286 | # We don't call `super` because narrow parents can be buggy in case of a | |
285 | # ambiguous dirstate. Always take the slow path until there is a better |
|
287 | # ambiguous dirstate. Always take the slow path until there is a better | |
286 | # fix, see issue6150. |
|
288 | # fix, see issue6150. | |
287 |
|
289 | |||
288 | # Censored files compare against the empty file. |
|
290 | # Censored files compare against the empty file. | |
289 | if self.iscensored(self.rev(node)): |
|
291 | if self.iscensored(self.rev(node)): | |
290 | return text != b'' |
|
292 | return text != b'' | |
291 |
|
293 | |||
292 | return self.read(node) != text |
|
294 | return self.read(node) != text |
@@ -1,2355 +1,2357 b'' | |||||
1 | # manifest.py - manifest revision class for mercurial |
|
1 | # manifest.py - manifest revision class for mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import heapq |
|
10 | import heapq | |
11 | import itertools |
|
11 | import itertools | |
12 | import struct |
|
12 | import struct | |
13 | import weakref |
|
13 | import weakref | |
14 |
|
14 | |||
15 | from .i18n import _ |
|
15 | from .i18n import _ | |
16 | from .node import ( |
|
16 | from .node import ( | |
17 | bin, |
|
17 | bin, | |
18 | hex, |
|
18 | hex, | |
19 | nullid, |
|
19 | nullid, | |
20 | nullrev, |
|
20 | nullrev, | |
21 | ) |
|
21 | ) | |
22 | from .pycompat import getattr |
|
22 | from .pycompat import getattr | |
23 | from . import ( |
|
23 | from . import ( | |
24 | encoding, |
|
24 | encoding, | |
25 | error, |
|
25 | error, | |
26 | match as matchmod, |
|
26 | match as matchmod, | |
27 | mdiff, |
|
27 | mdiff, | |
28 | pathutil, |
|
28 | pathutil, | |
29 | policy, |
|
29 | policy, | |
30 | pycompat, |
|
30 | pycompat, | |
31 | revlog, |
|
31 | revlog, | |
32 | util, |
|
32 | util, | |
33 | ) |
|
33 | ) | |
34 | from .interfaces import ( |
|
34 | from .interfaces import ( | |
35 | repository, |
|
35 | repository, | |
36 | util as interfaceutil, |
|
36 | util as interfaceutil, | |
37 | ) |
|
37 | ) | |
38 |
|
38 | |||
39 | parsers = policy.importmod('parsers') |
|
39 | parsers = policy.importmod('parsers') | |
40 | propertycache = util.propertycache |
|
40 | propertycache = util.propertycache | |
41 |
|
41 | |||
42 | # Allow tests to more easily test the alternate path in manifestdict.fastdelta() |
|
42 | # Allow tests to more easily test the alternate path in manifestdict.fastdelta() | |
43 | FASTDELTA_TEXTDIFF_THRESHOLD = 1000 |
|
43 | FASTDELTA_TEXTDIFF_THRESHOLD = 1000 | |
44 |
|
44 | |||
45 |
|
45 | |||
46 | def _parse(data): |
|
46 | def _parse(data): | |
47 | # This method does a little bit of excessive-looking |
|
47 | # This method does a little bit of excessive-looking | |
48 | # precondition checking. This is so that the behavior of this |
|
48 | # precondition checking. This is so that the behavior of this | |
49 | # class exactly matches its C counterpart to try and help |
|
49 | # class exactly matches its C counterpart to try and help | |
50 | # prevent surprise breakage for anyone that develops against |
|
50 | # prevent surprise breakage for anyone that develops against | |
51 | # the pure version. |
|
51 | # the pure version. | |
52 | if data and data[-1:] != b'\n': |
|
52 | if data and data[-1:] != b'\n': | |
53 | raise ValueError(b'Manifest did not end in a newline.') |
|
53 | raise ValueError(b'Manifest did not end in a newline.') | |
54 | prev = None |
|
54 | prev = None | |
55 | for l in data.splitlines(): |
|
55 | for l in data.splitlines(): | |
56 | if prev is not None and prev > l: |
|
56 | if prev is not None and prev > l: | |
57 | raise ValueError(b'Manifest lines not in sorted order.') |
|
57 | raise ValueError(b'Manifest lines not in sorted order.') | |
58 | prev = l |
|
58 | prev = l | |
59 | f, n = l.split(b'\0') |
|
59 | f, n = l.split(b'\0') | |
60 | nl = len(n) |
|
60 | nl = len(n) | |
61 | flags = n[-1:] |
|
61 | flags = n[-1:] | |
62 | if flags in _manifestflags: |
|
62 | if flags in _manifestflags: | |
63 | n = n[:-1] |
|
63 | n = n[:-1] | |
64 | nl -= 1 |
|
64 | nl -= 1 | |
65 | else: |
|
65 | else: | |
66 | flags = b'' |
|
66 | flags = b'' | |
67 | if nl not in (40, 64): |
|
67 | if nl not in (40, 64): | |
68 | raise ValueError(b'Invalid manifest line') |
|
68 | raise ValueError(b'Invalid manifest line') | |
69 |
|
69 | |||
70 | yield f, bin(n), flags |
|
70 | yield f, bin(n), flags | |
71 |
|
71 | |||
72 |
|
72 | |||
73 | def _text(it): |
|
73 | def _text(it): | |
74 | files = [] |
|
74 | files = [] | |
75 | lines = [] |
|
75 | lines = [] | |
76 | for f, n, fl in it: |
|
76 | for f, n, fl in it: | |
77 | files.append(f) |
|
77 | files.append(f) | |
78 | # if this is changed to support newlines in filenames, |
|
78 | # if this is changed to support newlines in filenames, | |
79 | # be sure to check the templates/ dir again (especially *-raw.tmpl) |
|
79 | # be sure to check the templates/ dir again (especially *-raw.tmpl) | |
80 | lines.append(b"%s\0%s%s\n" % (f, hex(n), fl)) |
|
80 | lines.append(b"%s\0%s%s\n" % (f, hex(n), fl)) | |
81 |
|
81 | |||
82 | _checkforbidden(files) |
|
82 | _checkforbidden(files) | |
83 | return b''.join(lines) |
|
83 | return b''.join(lines) | |
84 |
|
84 | |||
85 |
|
85 | |||
86 | class lazymanifestiter(object): |
|
86 | class lazymanifestiter(object): | |
87 | def __init__(self, lm): |
|
87 | def __init__(self, lm): | |
88 | self.pos = 0 |
|
88 | self.pos = 0 | |
89 | self.lm = lm |
|
89 | self.lm = lm | |
90 |
|
90 | |||
91 | def __iter__(self): |
|
91 | def __iter__(self): | |
92 | return self |
|
92 | return self | |
93 |
|
93 | |||
94 | def next(self): |
|
94 | def next(self): | |
95 | try: |
|
95 | try: | |
96 | data, pos = self.lm._get(self.pos) |
|
96 | data, pos = self.lm._get(self.pos) | |
97 | except IndexError: |
|
97 | except IndexError: | |
98 | raise StopIteration |
|
98 | raise StopIteration | |
99 | if pos == -1: |
|
99 | if pos == -1: | |
100 | self.pos += 1 |
|
100 | self.pos += 1 | |
101 | return data[0] |
|
101 | return data[0] | |
102 | self.pos += 1 |
|
102 | self.pos += 1 | |
103 | zeropos = data.find(b'\x00', pos) |
|
103 | zeropos = data.find(b'\x00', pos) | |
104 | return data[pos:zeropos] |
|
104 | return data[pos:zeropos] | |
105 |
|
105 | |||
106 | __next__ = next |
|
106 | __next__ = next | |
107 |
|
107 | |||
108 |
|
108 | |||
109 | class lazymanifestiterentries(object): |
|
109 | class lazymanifestiterentries(object): | |
110 | def __init__(self, lm): |
|
110 | def __init__(self, lm): | |
111 | self.lm = lm |
|
111 | self.lm = lm | |
112 | self.pos = 0 |
|
112 | self.pos = 0 | |
113 |
|
113 | |||
114 | def __iter__(self): |
|
114 | def __iter__(self): | |
115 | return self |
|
115 | return self | |
116 |
|
116 | |||
117 | def next(self): |
|
117 | def next(self): | |
118 | try: |
|
118 | try: | |
119 | data, pos = self.lm._get(self.pos) |
|
119 | data, pos = self.lm._get(self.pos) | |
120 | except IndexError: |
|
120 | except IndexError: | |
121 | raise StopIteration |
|
121 | raise StopIteration | |
122 | if pos == -1: |
|
122 | if pos == -1: | |
123 | self.pos += 1 |
|
123 | self.pos += 1 | |
124 | return data |
|
124 | return data | |
125 | zeropos = data.find(b'\x00', pos) |
|
125 | zeropos = data.find(b'\x00', pos) | |
126 | nlpos = data.find(b'\n', pos) |
|
126 | nlpos = data.find(b'\n', pos) | |
127 | if zeropos == -1 or nlpos == -1 or nlpos < zeropos: |
|
127 | if zeropos == -1 or nlpos == -1 or nlpos < zeropos: | |
128 | raise error.StorageError(b'Invalid manifest line') |
|
128 | raise error.StorageError(b'Invalid manifest line') | |
129 | flags = data[nlpos - 1 : nlpos] |
|
129 | flags = data[nlpos - 1 : nlpos] | |
130 | if flags in _manifestflags: |
|
130 | if flags in _manifestflags: | |
131 | hlen = nlpos - zeropos - 2 |
|
131 | hlen = nlpos - zeropos - 2 | |
132 | else: |
|
132 | else: | |
133 | hlen = nlpos - zeropos - 1 |
|
133 | hlen = nlpos - zeropos - 1 | |
134 | flags = b'' |
|
134 | flags = b'' | |
135 | if hlen not in (40, 64): |
|
135 | if hlen not in (40, 64): | |
136 | raise error.StorageError(b'Invalid manifest line') |
|
136 | raise error.StorageError(b'Invalid manifest line') | |
137 | hashval = unhexlify( |
|
137 | hashval = unhexlify( | |
138 | data, self.lm.extrainfo[self.pos], zeropos + 1, hlen |
|
138 | data, self.lm.extrainfo[self.pos], zeropos + 1, hlen | |
139 | ) |
|
139 | ) | |
140 | self.pos += 1 |
|
140 | self.pos += 1 | |
141 | return (data[pos:zeropos], hashval, flags) |
|
141 | return (data[pos:zeropos], hashval, flags) | |
142 |
|
142 | |||
143 | __next__ = next |
|
143 | __next__ = next | |
144 |
|
144 | |||
145 |
|
145 | |||
146 | def unhexlify(data, extra, pos, length): |
|
146 | def unhexlify(data, extra, pos, length): | |
147 | s = bin(data[pos : pos + length]) |
|
147 | s = bin(data[pos : pos + length]) | |
148 | if extra: |
|
148 | if extra: | |
149 | s += chr(extra & 0xFF) |
|
149 | s += chr(extra & 0xFF) | |
150 | return s |
|
150 | return s | |
151 |
|
151 | |||
152 |
|
152 | |||
153 | def _cmp(a, b): |
|
153 | def _cmp(a, b): | |
154 | return (a > b) - (a < b) |
|
154 | return (a > b) - (a < b) | |
155 |
|
155 | |||
156 |
|
156 | |||
157 | _manifestflags = {b'', b'l', b't', b'x'} |
|
157 | _manifestflags = {b'', b'l', b't', b'x'} | |
158 |
|
158 | |||
159 |
|
159 | |||
160 | class _lazymanifest(object): |
|
160 | class _lazymanifest(object): | |
161 | """A pure python manifest backed by a byte string. It is supplimented with |
|
161 | """A pure python manifest backed by a byte string. It is supplimented with | |
162 | internal lists as it is modified, until it is compacted back to a pure byte |
|
162 | internal lists as it is modified, until it is compacted back to a pure byte | |
163 | string. |
|
163 | string. | |
164 |
|
164 | |||
165 | ``data`` is the initial manifest data. |
|
165 | ``data`` is the initial manifest data. | |
166 |
|
166 | |||
167 | ``positions`` is a list of offsets, one per manifest entry. Positive |
|
167 | ``positions`` is a list of offsets, one per manifest entry. Positive | |
168 | values are offsets into ``data``, negative values are offsets into the |
|
168 | values are offsets into ``data``, negative values are offsets into the | |
169 | ``extradata`` list. When an entry is removed, its entry is dropped from |
|
169 | ``extradata`` list. When an entry is removed, its entry is dropped from | |
170 | ``positions``. The values are encoded such that when walking the list and |
|
170 | ``positions``. The values are encoded such that when walking the list and | |
171 | indexing into ``data`` or ``extradata`` as appropriate, the entries are |
|
171 | indexing into ``data`` or ``extradata`` as appropriate, the entries are | |
172 | sorted by filename. |
|
172 | sorted by filename. | |
173 |
|
173 | |||
174 | ``extradata`` is a list of (key, hash, flags) for entries that were added or |
|
174 | ``extradata`` is a list of (key, hash, flags) for entries that were added or | |
175 | modified since the manifest was created or compacted. |
|
175 | modified since the manifest was created or compacted. | |
176 | """ |
|
176 | """ | |
177 |
|
177 | |||
178 | def __init__( |
|
178 | def __init__( | |
179 | self, |
|
179 | self, | |
180 | data, |
|
180 | data, | |
181 | positions=None, |
|
181 | positions=None, | |
182 | extrainfo=None, |
|
182 | extrainfo=None, | |
183 | extradata=None, |
|
183 | extradata=None, | |
184 | hasremovals=False, |
|
184 | hasremovals=False, | |
185 | ): |
|
185 | ): | |
186 | if positions is None: |
|
186 | if positions is None: | |
187 | self.positions = self.findlines(data) |
|
187 | self.positions = self.findlines(data) | |
188 | self.extrainfo = [0] * len(self.positions) |
|
188 | self.extrainfo = [0] * len(self.positions) | |
189 | self.data = data |
|
189 | self.data = data | |
190 | self.extradata = [] |
|
190 | self.extradata = [] | |
191 | self.hasremovals = False |
|
191 | self.hasremovals = False | |
192 | else: |
|
192 | else: | |
193 | self.positions = positions[:] |
|
193 | self.positions = positions[:] | |
194 | self.extrainfo = extrainfo[:] |
|
194 | self.extrainfo = extrainfo[:] | |
195 | self.extradata = extradata[:] |
|
195 | self.extradata = extradata[:] | |
196 | self.data = data |
|
196 | self.data = data | |
197 | self.hasremovals = hasremovals |
|
197 | self.hasremovals = hasremovals | |
198 |
|
198 | |||
199 | def findlines(self, data): |
|
199 | def findlines(self, data): | |
200 | if not data: |
|
200 | if not data: | |
201 | return [] |
|
201 | return [] | |
202 | pos = data.find(b"\n") |
|
202 | pos = data.find(b"\n") | |
203 | if pos == -1 or data[-1:] != b'\n': |
|
203 | if pos == -1 or data[-1:] != b'\n': | |
204 | raise ValueError(b"Manifest did not end in a newline.") |
|
204 | raise ValueError(b"Manifest did not end in a newline.") | |
205 | positions = [0] |
|
205 | positions = [0] | |
206 | prev = data[: data.find(b'\x00')] |
|
206 | prev = data[: data.find(b'\x00')] | |
207 | while pos < len(data) - 1 and pos != -1: |
|
207 | while pos < len(data) - 1 and pos != -1: | |
208 | positions.append(pos + 1) |
|
208 | positions.append(pos + 1) | |
209 | nexts = data[pos + 1 : data.find(b'\x00', pos + 1)] |
|
209 | nexts = data[pos + 1 : data.find(b'\x00', pos + 1)] | |
210 | if nexts < prev: |
|
210 | if nexts < prev: | |
211 | raise ValueError(b"Manifest lines not in sorted order.") |
|
211 | raise ValueError(b"Manifest lines not in sorted order.") | |
212 | prev = nexts |
|
212 | prev = nexts | |
213 | pos = data.find(b"\n", pos + 1) |
|
213 | pos = data.find(b"\n", pos + 1) | |
214 | return positions |
|
214 | return positions | |
215 |
|
215 | |||
216 | def _get(self, index): |
|
216 | def _get(self, index): | |
217 | # get the position encoded in pos: |
|
217 | # get the position encoded in pos: | |
218 | # positive number is an index in 'data' |
|
218 | # positive number is an index in 'data' | |
219 | # negative number is in extrapieces |
|
219 | # negative number is in extrapieces | |
220 | pos = self.positions[index] |
|
220 | pos = self.positions[index] | |
221 | if pos >= 0: |
|
221 | if pos >= 0: | |
222 | return self.data, pos |
|
222 | return self.data, pos | |
223 | return self.extradata[-pos - 1], -1 |
|
223 | return self.extradata[-pos - 1], -1 | |
224 |
|
224 | |||
225 | def _getkey(self, pos): |
|
225 | def _getkey(self, pos): | |
226 | if pos >= 0: |
|
226 | if pos >= 0: | |
227 | return self.data[pos : self.data.find(b'\x00', pos + 1)] |
|
227 | return self.data[pos : self.data.find(b'\x00', pos + 1)] | |
228 | return self.extradata[-pos - 1][0] |
|
228 | return self.extradata[-pos - 1][0] | |
229 |
|
229 | |||
230 | def bsearch(self, key): |
|
230 | def bsearch(self, key): | |
231 | first = 0 |
|
231 | first = 0 | |
232 | last = len(self.positions) - 1 |
|
232 | last = len(self.positions) - 1 | |
233 |
|
233 | |||
234 | while first <= last: |
|
234 | while first <= last: | |
235 | midpoint = (first + last) // 2 |
|
235 | midpoint = (first + last) // 2 | |
236 | nextpos = self.positions[midpoint] |
|
236 | nextpos = self.positions[midpoint] | |
237 | candidate = self._getkey(nextpos) |
|
237 | candidate = self._getkey(nextpos) | |
238 | r = _cmp(key, candidate) |
|
238 | r = _cmp(key, candidate) | |
239 | if r == 0: |
|
239 | if r == 0: | |
240 | return midpoint |
|
240 | return midpoint | |
241 | else: |
|
241 | else: | |
242 | if r < 0: |
|
242 | if r < 0: | |
243 | last = midpoint - 1 |
|
243 | last = midpoint - 1 | |
244 | else: |
|
244 | else: | |
245 | first = midpoint + 1 |
|
245 | first = midpoint + 1 | |
246 | return -1 |
|
246 | return -1 | |
247 |
|
247 | |||
248 | def bsearch2(self, key): |
|
248 | def bsearch2(self, key): | |
249 | # same as the above, but will always return the position |
|
249 | # same as the above, but will always return the position | |
250 | # done for performance reasons |
|
250 | # done for performance reasons | |
251 | first = 0 |
|
251 | first = 0 | |
252 | last = len(self.positions) - 1 |
|
252 | last = len(self.positions) - 1 | |
253 |
|
253 | |||
254 | while first <= last: |
|
254 | while first <= last: | |
255 | midpoint = (first + last) // 2 |
|
255 | midpoint = (first + last) // 2 | |
256 | nextpos = self.positions[midpoint] |
|
256 | nextpos = self.positions[midpoint] | |
257 | candidate = self._getkey(nextpos) |
|
257 | candidate = self._getkey(nextpos) | |
258 | r = _cmp(key, candidate) |
|
258 | r = _cmp(key, candidate) | |
259 | if r == 0: |
|
259 | if r == 0: | |
260 | return (midpoint, True) |
|
260 | return (midpoint, True) | |
261 | else: |
|
261 | else: | |
262 | if r < 0: |
|
262 | if r < 0: | |
263 | last = midpoint - 1 |
|
263 | last = midpoint - 1 | |
264 | else: |
|
264 | else: | |
265 | first = midpoint + 1 |
|
265 | first = midpoint + 1 | |
266 | return (first, False) |
|
266 | return (first, False) | |
267 |
|
267 | |||
268 | def __contains__(self, key): |
|
268 | def __contains__(self, key): | |
269 | return self.bsearch(key) != -1 |
|
269 | return self.bsearch(key) != -1 | |
270 |
|
270 | |||
271 | def __getitem__(self, key): |
|
271 | def __getitem__(self, key): | |
272 | if not isinstance(key, bytes): |
|
272 | if not isinstance(key, bytes): | |
273 | raise TypeError(b"getitem: manifest keys must be a bytes.") |
|
273 | raise TypeError(b"getitem: manifest keys must be a bytes.") | |
274 | needle = self.bsearch(key) |
|
274 | needle = self.bsearch(key) | |
275 | if needle == -1: |
|
275 | if needle == -1: | |
276 | raise KeyError |
|
276 | raise KeyError | |
277 | data, pos = self._get(needle) |
|
277 | data, pos = self._get(needle) | |
278 | if pos == -1: |
|
278 | if pos == -1: | |
279 | return (data[1], data[2]) |
|
279 | return (data[1], data[2]) | |
280 | zeropos = data.find(b'\x00', pos) |
|
280 | zeropos = data.find(b'\x00', pos) | |
281 | nlpos = data.find(b'\n', zeropos) |
|
281 | nlpos = data.find(b'\n', zeropos) | |
282 | assert 0 <= needle <= len(self.positions) |
|
282 | assert 0 <= needle <= len(self.positions) | |
283 | assert len(self.extrainfo) == len(self.positions) |
|
283 | assert len(self.extrainfo) == len(self.positions) | |
284 | if zeropos == -1 or nlpos == -1 or nlpos < zeropos: |
|
284 | if zeropos == -1 or nlpos == -1 or nlpos < zeropos: | |
285 | raise error.StorageError(b'Invalid manifest line') |
|
285 | raise error.StorageError(b'Invalid manifest line') | |
286 | hlen = nlpos - zeropos - 1 |
|
286 | hlen = nlpos - zeropos - 1 | |
287 | flags = data[nlpos - 1 : nlpos] |
|
287 | flags = data[nlpos - 1 : nlpos] | |
288 | if flags in _manifestflags: |
|
288 | if flags in _manifestflags: | |
289 | hlen -= 1 |
|
289 | hlen -= 1 | |
290 | else: |
|
290 | else: | |
291 | flags = b'' |
|
291 | flags = b'' | |
292 | if hlen not in (40, 64): |
|
292 | if hlen not in (40, 64): | |
293 | raise error.StorageError(b'Invalid manifest line') |
|
293 | raise error.StorageError(b'Invalid manifest line') | |
294 | hashval = unhexlify(data, self.extrainfo[needle], zeropos + 1, hlen) |
|
294 | hashval = unhexlify(data, self.extrainfo[needle], zeropos + 1, hlen) | |
295 | return (hashval, flags) |
|
295 | return (hashval, flags) | |
296 |
|
296 | |||
297 | def __delitem__(self, key): |
|
297 | def __delitem__(self, key): | |
298 | needle, found = self.bsearch2(key) |
|
298 | needle, found = self.bsearch2(key) | |
299 | if not found: |
|
299 | if not found: | |
300 | raise KeyError |
|
300 | raise KeyError | |
301 | cur = self.positions[needle] |
|
301 | cur = self.positions[needle] | |
302 | self.positions = self.positions[:needle] + self.positions[needle + 1 :] |
|
302 | self.positions = self.positions[:needle] + self.positions[needle + 1 :] | |
303 | self.extrainfo = self.extrainfo[:needle] + self.extrainfo[needle + 1 :] |
|
303 | self.extrainfo = self.extrainfo[:needle] + self.extrainfo[needle + 1 :] | |
304 | if cur >= 0: |
|
304 | if cur >= 0: | |
305 | # This does NOT unsort the list as far as the search functions are |
|
305 | # This does NOT unsort the list as far as the search functions are | |
306 | # concerned, as they only examine lines mapped by self.positions. |
|
306 | # concerned, as they only examine lines mapped by self.positions. | |
307 | self.data = self.data[:cur] + b'\x00' + self.data[cur + 1 :] |
|
307 | self.data = self.data[:cur] + b'\x00' + self.data[cur + 1 :] | |
308 | self.hasremovals = True |
|
308 | self.hasremovals = True | |
309 |
|
309 | |||
310 | def __setitem__(self, key, value): |
|
310 | def __setitem__(self, key, value): | |
311 | if not isinstance(key, bytes): |
|
311 | if not isinstance(key, bytes): | |
312 | raise TypeError(b"setitem: manifest keys must be a byte string.") |
|
312 | raise TypeError(b"setitem: manifest keys must be a byte string.") | |
313 | if not isinstance(value, tuple) or len(value) != 2: |
|
313 | if not isinstance(value, tuple) or len(value) != 2: | |
314 | raise TypeError( |
|
314 | raise TypeError( | |
315 | b"Manifest values must be a tuple of (node, flags)." |
|
315 | b"Manifest values must be a tuple of (node, flags)." | |
316 | ) |
|
316 | ) | |
317 | hashval = value[0] |
|
317 | hashval = value[0] | |
318 | if not isinstance(hashval, bytes) or len(hashval) not in (20, 32): |
|
318 | if not isinstance(hashval, bytes) or len(hashval) not in (20, 32): | |
319 | raise TypeError(b"node must be a 20-byte or 32-byte byte string") |
|
319 | raise TypeError(b"node must be a 20-byte or 32-byte byte string") | |
320 | flags = value[1] |
|
320 | flags = value[1] | |
321 | if not isinstance(flags, bytes) or len(flags) > 1: |
|
321 | if not isinstance(flags, bytes) or len(flags) > 1: | |
322 | raise TypeError(b"flags must a 0 or 1 byte string, got %r", flags) |
|
322 | raise TypeError(b"flags must a 0 or 1 byte string, got %r", flags) | |
323 | needle, found = self.bsearch2(key) |
|
323 | needle, found = self.bsearch2(key) | |
324 | if found: |
|
324 | if found: | |
325 | # put the item |
|
325 | # put the item | |
326 | pos = self.positions[needle] |
|
326 | pos = self.positions[needle] | |
327 | if pos < 0: |
|
327 | if pos < 0: | |
328 | self.extradata[-pos - 1] = (key, hashval, value[1]) |
|
328 | self.extradata[-pos - 1] = (key, hashval, value[1]) | |
329 | else: |
|
329 | else: | |
330 | # just don't bother |
|
330 | # just don't bother | |
331 | self.extradata.append((key, hashval, value[1])) |
|
331 | self.extradata.append((key, hashval, value[1])) | |
332 | self.positions[needle] = -len(self.extradata) |
|
332 | self.positions[needle] = -len(self.extradata) | |
333 | else: |
|
333 | else: | |
334 | # not found, put it in with extra positions |
|
334 | # not found, put it in with extra positions | |
335 | self.extradata.append((key, hashval, value[1])) |
|
335 | self.extradata.append((key, hashval, value[1])) | |
336 | self.positions = ( |
|
336 | self.positions = ( | |
337 | self.positions[:needle] |
|
337 | self.positions[:needle] | |
338 | + [-len(self.extradata)] |
|
338 | + [-len(self.extradata)] | |
339 | + self.positions[needle:] |
|
339 | + self.positions[needle:] | |
340 | ) |
|
340 | ) | |
341 | self.extrainfo = ( |
|
341 | self.extrainfo = ( | |
342 | self.extrainfo[:needle] + [0] + self.extrainfo[needle:] |
|
342 | self.extrainfo[:needle] + [0] + self.extrainfo[needle:] | |
343 | ) |
|
343 | ) | |
344 |
|
344 | |||
345 | def copy(self): |
|
345 | def copy(self): | |
346 | # XXX call _compact like in C? |
|
346 | # XXX call _compact like in C? | |
347 | return _lazymanifest( |
|
347 | return _lazymanifest( | |
348 | self.data, |
|
348 | self.data, | |
349 | self.positions, |
|
349 | self.positions, | |
350 | self.extrainfo, |
|
350 | self.extrainfo, | |
351 | self.extradata, |
|
351 | self.extradata, | |
352 | self.hasremovals, |
|
352 | self.hasremovals, | |
353 | ) |
|
353 | ) | |
354 |
|
354 | |||
355 | def _compact(self): |
|
355 | def _compact(self): | |
356 | # hopefully not called TOO often |
|
356 | # hopefully not called TOO often | |
357 | if len(self.extradata) == 0 and not self.hasremovals: |
|
357 | if len(self.extradata) == 0 and not self.hasremovals: | |
358 | return |
|
358 | return | |
359 | l = [] |
|
359 | l = [] | |
360 | i = 0 |
|
360 | i = 0 | |
361 | offset = 0 |
|
361 | offset = 0 | |
362 | self.extrainfo = [0] * len(self.positions) |
|
362 | self.extrainfo = [0] * len(self.positions) | |
363 | while i < len(self.positions): |
|
363 | while i < len(self.positions): | |
364 | if self.positions[i] >= 0: |
|
364 | if self.positions[i] >= 0: | |
365 | cur = self.positions[i] |
|
365 | cur = self.positions[i] | |
366 | last_cut = cur |
|
366 | last_cut = cur | |
367 |
|
367 | |||
368 | # Collect all contiguous entries in the buffer at the current |
|
368 | # Collect all contiguous entries in the buffer at the current | |
369 | # offset, breaking out only for added/modified items held in |
|
369 | # offset, breaking out only for added/modified items held in | |
370 | # extradata, or a deleted line prior to the next position. |
|
370 | # extradata, or a deleted line prior to the next position. | |
371 | while True: |
|
371 | while True: | |
372 | self.positions[i] = offset |
|
372 | self.positions[i] = offset | |
373 | i += 1 |
|
373 | i += 1 | |
374 | if i == len(self.positions) or self.positions[i] < 0: |
|
374 | if i == len(self.positions) or self.positions[i] < 0: | |
375 | break |
|
375 | break | |
376 |
|
376 | |||
377 | # A removed file has no positions[] entry, but does have an |
|
377 | # A removed file has no positions[] entry, but does have an | |
378 | # overwritten first byte. Break out and find the end of the |
|
378 | # overwritten first byte. Break out and find the end of the | |
379 | # current good entry/entries if there is a removed file |
|
379 | # current good entry/entries if there is a removed file | |
380 | # before the next position. |
|
380 | # before the next position. | |
381 | if ( |
|
381 | if ( | |
382 | self.hasremovals |
|
382 | self.hasremovals | |
383 | and self.data.find(b'\n\x00', cur, self.positions[i]) |
|
383 | and self.data.find(b'\n\x00', cur, self.positions[i]) | |
384 | != -1 |
|
384 | != -1 | |
385 | ): |
|
385 | ): | |
386 | break |
|
386 | break | |
387 |
|
387 | |||
388 | offset += self.positions[i] - cur |
|
388 | offset += self.positions[i] - cur | |
389 | cur = self.positions[i] |
|
389 | cur = self.positions[i] | |
390 | end_cut = self.data.find(b'\n', cur) |
|
390 | end_cut = self.data.find(b'\n', cur) | |
391 | if end_cut != -1: |
|
391 | if end_cut != -1: | |
392 | end_cut += 1 |
|
392 | end_cut += 1 | |
393 | offset += end_cut - cur |
|
393 | offset += end_cut - cur | |
394 | l.append(self.data[last_cut:end_cut]) |
|
394 | l.append(self.data[last_cut:end_cut]) | |
395 | else: |
|
395 | else: | |
396 | while i < len(self.positions) and self.positions[i] < 0: |
|
396 | while i < len(self.positions) and self.positions[i] < 0: | |
397 | cur = self.positions[i] |
|
397 | cur = self.positions[i] | |
398 | t = self.extradata[-cur - 1] |
|
398 | t = self.extradata[-cur - 1] | |
399 | l.append(self._pack(t)) |
|
399 | l.append(self._pack(t)) | |
400 | self.positions[i] = offset |
|
400 | self.positions[i] = offset | |
401 | # Hashes are either 20 bytes (old sha1s) or 32 |
|
401 | # Hashes are either 20 bytes (old sha1s) or 32 | |
402 | # bytes (new non-sha1). |
|
402 | # bytes (new non-sha1). | |
403 | hlen = 20 |
|
403 | hlen = 20 | |
404 | if len(t[1]) > 25: |
|
404 | if len(t[1]) > 25: | |
405 | hlen = 32 |
|
405 | hlen = 32 | |
406 | if len(t[1]) > hlen: |
|
406 | if len(t[1]) > hlen: | |
407 | self.extrainfo[i] = ord(t[1][hlen + 1]) |
|
407 | self.extrainfo[i] = ord(t[1][hlen + 1]) | |
408 | offset += len(l[-1]) |
|
408 | offset += len(l[-1]) | |
409 | i += 1 |
|
409 | i += 1 | |
410 | self.data = b''.join(l) |
|
410 | self.data = b''.join(l) | |
411 | self.hasremovals = False |
|
411 | self.hasremovals = False | |
412 | self.extradata = [] |
|
412 | self.extradata = [] | |
413 |
|
413 | |||
414 | def _pack(self, d): |
|
414 | def _pack(self, d): | |
415 | n = d[1] |
|
415 | n = d[1] | |
416 | assert len(n) in (20, 32) |
|
416 | assert len(n) in (20, 32) | |
417 | return d[0] + b'\x00' + hex(n) + d[2] + b'\n' |
|
417 | return d[0] + b'\x00' + hex(n) + d[2] + b'\n' | |
418 |
|
418 | |||
419 | def text(self): |
|
419 | def text(self): | |
420 | self._compact() |
|
420 | self._compact() | |
421 | return self.data |
|
421 | return self.data | |
422 |
|
422 | |||
423 | def diff(self, m2, clean=False): |
|
423 | def diff(self, m2, clean=False): | |
424 | '''Finds changes between the current manifest and m2.''' |
|
424 | '''Finds changes between the current manifest and m2.''' | |
425 | # XXX think whether efficiency matters here |
|
425 | # XXX think whether efficiency matters here | |
426 | diff = {} |
|
426 | diff = {} | |
427 |
|
427 | |||
428 | for fn, e1, flags in self.iterentries(): |
|
428 | for fn, e1, flags in self.iterentries(): | |
429 | if fn not in m2: |
|
429 | if fn not in m2: | |
430 | diff[fn] = (e1, flags), (None, b'') |
|
430 | diff[fn] = (e1, flags), (None, b'') | |
431 | else: |
|
431 | else: | |
432 | e2 = m2[fn] |
|
432 | e2 = m2[fn] | |
433 | if (e1, flags) != e2: |
|
433 | if (e1, flags) != e2: | |
434 | diff[fn] = (e1, flags), e2 |
|
434 | diff[fn] = (e1, flags), e2 | |
435 | elif clean: |
|
435 | elif clean: | |
436 | diff[fn] = None |
|
436 | diff[fn] = None | |
437 |
|
437 | |||
438 | for fn, e2, flags in m2.iterentries(): |
|
438 | for fn, e2, flags in m2.iterentries(): | |
439 | if fn not in self: |
|
439 | if fn not in self: | |
440 | diff[fn] = (None, b''), (e2, flags) |
|
440 | diff[fn] = (None, b''), (e2, flags) | |
441 |
|
441 | |||
442 | return diff |
|
442 | return diff | |
443 |
|
443 | |||
444 | def iterentries(self): |
|
444 | def iterentries(self): | |
445 | return lazymanifestiterentries(self) |
|
445 | return lazymanifestiterentries(self) | |
446 |
|
446 | |||
447 | def iterkeys(self): |
|
447 | def iterkeys(self): | |
448 | return lazymanifestiter(self) |
|
448 | return lazymanifestiter(self) | |
449 |
|
449 | |||
450 | def __iter__(self): |
|
450 | def __iter__(self): | |
451 | return lazymanifestiter(self) |
|
451 | return lazymanifestiter(self) | |
452 |
|
452 | |||
453 | def __len__(self): |
|
453 | def __len__(self): | |
454 | return len(self.positions) |
|
454 | return len(self.positions) | |
455 |
|
455 | |||
456 | def filtercopy(self, filterfn): |
|
456 | def filtercopy(self, filterfn): | |
457 | # XXX should be optimized |
|
457 | # XXX should be optimized | |
458 | c = _lazymanifest(b'') |
|
458 | c = _lazymanifest(b'') | |
459 | for f, n, fl in self.iterentries(): |
|
459 | for f, n, fl in self.iterentries(): | |
460 | if filterfn(f): |
|
460 | if filterfn(f): | |
461 | c[f] = n, fl |
|
461 | c[f] = n, fl | |
462 | return c |
|
462 | return c | |
463 |
|
463 | |||
464 |
|
464 | |||
465 | try: |
|
465 | try: | |
466 | _lazymanifest = parsers.lazymanifest |
|
466 | _lazymanifest = parsers.lazymanifest | |
467 | except AttributeError: |
|
467 | except AttributeError: | |
468 | pass |
|
468 | pass | |
469 |
|
469 | |||
470 |
|
470 | |||
471 | @interfaceutil.implementer(repository.imanifestdict) |
|
471 | @interfaceutil.implementer(repository.imanifestdict) | |
472 | class manifestdict(object): |
|
472 | class manifestdict(object): | |
473 | def __init__(self, data=b''): |
|
473 | def __init__(self, data=b''): | |
474 | self._lm = _lazymanifest(data) |
|
474 | self._lm = _lazymanifest(data) | |
475 |
|
475 | |||
476 | def __getitem__(self, key): |
|
476 | def __getitem__(self, key): | |
477 | return self._lm[key][0] |
|
477 | return self._lm[key][0] | |
478 |
|
478 | |||
479 | def find(self, key): |
|
479 | def find(self, key): | |
480 | return self._lm[key] |
|
480 | return self._lm[key] | |
481 |
|
481 | |||
482 | def __len__(self): |
|
482 | def __len__(self): | |
483 | return len(self._lm) |
|
483 | return len(self._lm) | |
484 |
|
484 | |||
485 | def __nonzero__(self): |
|
485 | def __nonzero__(self): | |
486 | # nonzero is covered by the __len__ function, but implementing it here |
|
486 | # nonzero is covered by the __len__ function, but implementing it here | |
487 | # makes it easier for extensions to override. |
|
487 | # makes it easier for extensions to override. | |
488 | return len(self._lm) != 0 |
|
488 | return len(self._lm) != 0 | |
489 |
|
489 | |||
490 | __bool__ = __nonzero__ |
|
490 | __bool__ = __nonzero__ | |
491 |
|
491 | |||
492 | def __setitem__(self, key, node): |
|
492 | def __setitem__(self, key, node): | |
493 | self._lm[key] = node, self.flags(key) |
|
493 | self._lm[key] = node, self.flags(key) | |
494 |
|
494 | |||
495 | def __contains__(self, key): |
|
495 | def __contains__(self, key): | |
496 | if key is None: |
|
496 | if key is None: | |
497 | return False |
|
497 | return False | |
498 | return key in self._lm |
|
498 | return key in self._lm | |
499 |
|
499 | |||
500 | def __delitem__(self, key): |
|
500 | def __delitem__(self, key): | |
501 | del self._lm[key] |
|
501 | del self._lm[key] | |
502 |
|
502 | |||
503 | def __iter__(self): |
|
503 | def __iter__(self): | |
504 | return self._lm.__iter__() |
|
504 | return self._lm.__iter__() | |
505 |
|
505 | |||
506 | def iterkeys(self): |
|
506 | def iterkeys(self): | |
507 | return self._lm.iterkeys() |
|
507 | return self._lm.iterkeys() | |
508 |
|
508 | |||
509 | def keys(self): |
|
509 | def keys(self): | |
510 | return list(self.iterkeys()) |
|
510 | return list(self.iterkeys()) | |
511 |
|
511 | |||
512 | def filesnotin(self, m2, match=None): |
|
512 | def filesnotin(self, m2, match=None): | |
513 | '''Set of files in this manifest that are not in the other''' |
|
513 | '''Set of files in this manifest that are not in the other''' | |
514 | if match is not None: |
|
514 | if match is not None: | |
515 | match = matchmod.badmatch(match, lambda path, msg: None) |
|
515 | match = matchmod.badmatch(match, lambda path, msg: None) | |
516 | sm2 = set(m2.walk(match)) |
|
516 | sm2 = set(m2.walk(match)) | |
517 | return {f for f in self.walk(match) if f not in sm2} |
|
517 | return {f for f in self.walk(match) if f not in sm2} | |
518 | return {f for f in self if f not in m2} |
|
518 | return {f for f in self if f not in m2} | |
519 |
|
519 | |||
520 | @propertycache |
|
520 | @propertycache | |
521 | def _dirs(self): |
|
521 | def _dirs(self): | |
522 | return pathutil.dirs(self) |
|
522 | return pathutil.dirs(self) | |
523 |
|
523 | |||
524 | def dirs(self): |
|
524 | def dirs(self): | |
525 | return self._dirs |
|
525 | return self._dirs | |
526 |
|
526 | |||
527 | def hasdir(self, dir): |
|
527 | def hasdir(self, dir): | |
528 | return dir in self._dirs |
|
528 | return dir in self._dirs | |
529 |
|
529 | |||
530 | def _filesfastpath(self, match): |
|
530 | def _filesfastpath(self, match): | |
531 | """Checks whether we can correctly and quickly iterate over matcher |
|
531 | """Checks whether we can correctly and quickly iterate over matcher | |
532 | files instead of over manifest files.""" |
|
532 | files instead of over manifest files.""" | |
533 | files = match.files() |
|
533 | files = match.files() | |
534 | return len(files) < 100 and ( |
|
534 | return len(files) < 100 and ( | |
535 | match.isexact() |
|
535 | match.isexact() | |
536 | or (match.prefix() and all(fn in self for fn in files)) |
|
536 | or (match.prefix() and all(fn in self for fn in files)) | |
537 | ) |
|
537 | ) | |
538 |
|
538 | |||
539 | def walk(self, match): |
|
539 | def walk(self, match): | |
540 | """Generates matching file names. |
|
540 | """Generates matching file names. | |
541 |
|
541 | |||
542 | Equivalent to manifest.matches(match).iterkeys(), but without creating |
|
542 | Equivalent to manifest.matches(match).iterkeys(), but without creating | |
543 | an entirely new manifest. |
|
543 | an entirely new manifest. | |
544 |
|
544 | |||
545 | It also reports nonexistent files by marking them bad with match.bad(). |
|
545 | It also reports nonexistent files by marking them bad with match.bad(). | |
546 | """ |
|
546 | """ | |
547 | if match.always(): |
|
547 | if match.always(): | |
548 | for f in iter(self): |
|
548 | for f in iter(self): | |
549 | yield f |
|
549 | yield f | |
550 | return |
|
550 | return | |
551 |
|
551 | |||
552 | fset = set(match.files()) |
|
552 | fset = set(match.files()) | |
553 |
|
553 | |||
554 | # avoid the entire walk if we're only looking for specific files |
|
554 | # avoid the entire walk if we're only looking for specific files | |
555 | if self._filesfastpath(match): |
|
555 | if self._filesfastpath(match): | |
556 | for fn in sorted(fset): |
|
556 | for fn in sorted(fset): | |
557 | if fn in self: |
|
557 | if fn in self: | |
558 | yield fn |
|
558 | yield fn | |
559 | return |
|
559 | return | |
560 |
|
560 | |||
561 | for fn in self: |
|
561 | for fn in self: | |
562 | if fn in fset: |
|
562 | if fn in fset: | |
563 | # specified pattern is the exact name |
|
563 | # specified pattern is the exact name | |
564 | fset.remove(fn) |
|
564 | fset.remove(fn) | |
565 | if match(fn): |
|
565 | if match(fn): | |
566 | yield fn |
|
566 | yield fn | |
567 |
|
567 | |||
568 | # for dirstate.walk, files=[''] means "walk the whole tree". |
|
568 | # for dirstate.walk, files=[''] means "walk the whole tree". | |
569 | # follow that here, too |
|
569 | # follow that here, too | |
570 | fset.discard(b'') |
|
570 | fset.discard(b'') | |
571 |
|
571 | |||
572 | for fn in sorted(fset): |
|
572 | for fn in sorted(fset): | |
573 | if not self.hasdir(fn): |
|
573 | if not self.hasdir(fn): | |
574 | match.bad(fn, None) |
|
574 | match.bad(fn, None) | |
575 |
|
575 | |||
576 | def _matches(self, match): |
|
576 | def _matches(self, match): | |
577 | '''generate a new manifest filtered by the match argument''' |
|
577 | '''generate a new manifest filtered by the match argument''' | |
578 | if match.always(): |
|
578 | if match.always(): | |
579 | return self.copy() |
|
579 | return self.copy() | |
580 |
|
580 | |||
581 | if self._filesfastpath(match): |
|
581 | if self._filesfastpath(match): | |
582 | m = manifestdict() |
|
582 | m = manifestdict() | |
583 | lm = self._lm |
|
583 | lm = self._lm | |
584 | for fn in match.files(): |
|
584 | for fn in match.files(): | |
585 | if fn in lm: |
|
585 | if fn in lm: | |
586 | m._lm[fn] = lm[fn] |
|
586 | m._lm[fn] = lm[fn] | |
587 | return m |
|
587 | return m | |
588 |
|
588 | |||
589 | m = manifestdict() |
|
589 | m = manifestdict() | |
590 | m._lm = self._lm.filtercopy(match) |
|
590 | m._lm = self._lm.filtercopy(match) | |
591 | return m |
|
591 | return m | |
592 |
|
592 | |||
593 | def diff(self, m2, match=None, clean=False): |
|
593 | def diff(self, m2, match=None, clean=False): | |
594 | """Finds changes between the current manifest and m2. |
|
594 | """Finds changes between the current manifest and m2. | |
595 |
|
595 | |||
596 | Args: |
|
596 | Args: | |
597 | m2: the manifest to which this manifest should be compared. |
|
597 | m2: the manifest to which this manifest should be compared. | |
598 | clean: if true, include files unchanged between these manifests |
|
598 | clean: if true, include files unchanged between these manifests | |
599 | with a None value in the returned dictionary. |
|
599 | with a None value in the returned dictionary. | |
600 |
|
600 | |||
601 | The result is returned as a dict with filename as key and |
|
601 | The result is returned as a dict with filename as key and | |
602 | values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the |
|
602 | values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the | |
603 | nodeid in the current/other manifest and fl1/fl2 is the flag |
|
603 | nodeid in the current/other manifest and fl1/fl2 is the flag | |
604 | in the current/other manifest. Where the file does not exist, |
|
604 | in the current/other manifest. Where the file does not exist, | |
605 | the nodeid will be None and the flags will be the empty |
|
605 | the nodeid will be None and the flags will be the empty | |
606 | string. |
|
606 | string. | |
607 | """ |
|
607 | """ | |
608 | if match: |
|
608 | if match: | |
609 | m1 = self._matches(match) |
|
609 | m1 = self._matches(match) | |
610 | m2 = m2._matches(match) |
|
610 | m2 = m2._matches(match) | |
611 | return m1.diff(m2, clean=clean) |
|
611 | return m1.diff(m2, clean=clean) | |
612 | return self._lm.diff(m2._lm, clean) |
|
612 | return self._lm.diff(m2._lm, clean) | |
613 |
|
613 | |||
614 | def setflag(self, key, flag): |
|
614 | def setflag(self, key, flag): | |
615 | if flag not in _manifestflags: |
|
615 | if flag not in _manifestflags: | |
616 | raise TypeError(b"Invalid manifest flag set.") |
|
616 | raise TypeError(b"Invalid manifest flag set.") | |
617 | self._lm[key] = self[key], flag |
|
617 | self._lm[key] = self[key], flag | |
618 |
|
618 | |||
619 | def get(self, key, default=None): |
|
619 | def get(self, key, default=None): | |
620 | try: |
|
620 | try: | |
621 | return self._lm[key][0] |
|
621 | return self._lm[key][0] | |
622 | except KeyError: |
|
622 | except KeyError: | |
623 | return default |
|
623 | return default | |
624 |
|
624 | |||
625 | def flags(self, key): |
|
625 | def flags(self, key): | |
626 | try: |
|
626 | try: | |
627 | return self._lm[key][1] |
|
627 | return self._lm[key][1] | |
628 | except KeyError: |
|
628 | except KeyError: | |
629 | return b'' |
|
629 | return b'' | |
630 |
|
630 | |||
631 | def copy(self): |
|
631 | def copy(self): | |
632 | c = manifestdict() |
|
632 | c = manifestdict() | |
633 | c._lm = self._lm.copy() |
|
633 | c._lm = self._lm.copy() | |
634 | return c |
|
634 | return c | |
635 |
|
635 | |||
636 | def items(self): |
|
636 | def items(self): | |
637 | return (x[:2] for x in self._lm.iterentries()) |
|
637 | return (x[:2] for x in self._lm.iterentries()) | |
638 |
|
638 | |||
639 | def iteritems(self): |
|
639 | def iteritems(self): | |
640 | return (x[:2] for x in self._lm.iterentries()) |
|
640 | return (x[:2] for x in self._lm.iterentries()) | |
641 |
|
641 | |||
642 | def iterentries(self): |
|
642 | def iterentries(self): | |
643 | return self._lm.iterentries() |
|
643 | return self._lm.iterentries() | |
644 |
|
644 | |||
645 | def text(self): |
|
645 | def text(self): | |
646 | # most likely uses native version |
|
646 | # most likely uses native version | |
647 | return self._lm.text() |
|
647 | return self._lm.text() | |
648 |
|
648 | |||
649 | def fastdelta(self, base, changes): |
|
649 | def fastdelta(self, base, changes): | |
650 | """Given a base manifest text as a bytearray and a list of changes |
|
650 | """Given a base manifest text as a bytearray and a list of changes | |
651 | relative to that text, compute a delta that can be used by revlog. |
|
651 | relative to that text, compute a delta that can be used by revlog. | |
652 | """ |
|
652 | """ | |
653 | delta = [] |
|
653 | delta = [] | |
654 | dstart = None |
|
654 | dstart = None | |
655 | dend = None |
|
655 | dend = None | |
656 | dline = [b""] |
|
656 | dline = [b""] | |
657 | start = 0 |
|
657 | start = 0 | |
658 | # zero copy representation of base as a buffer |
|
658 | # zero copy representation of base as a buffer | |
659 | addbuf = util.buffer(base) |
|
659 | addbuf = util.buffer(base) | |
660 |
|
660 | |||
661 | changes = list(changes) |
|
661 | changes = list(changes) | |
662 | if len(changes) < FASTDELTA_TEXTDIFF_THRESHOLD: |
|
662 | if len(changes) < FASTDELTA_TEXTDIFF_THRESHOLD: | |
663 | # start with a readonly loop that finds the offset of |
|
663 | # start with a readonly loop that finds the offset of | |
664 | # each line and creates the deltas |
|
664 | # each line and creates the deltas | |
665 | for f, todelete in changes: |
|
665 | for f, todelete in changes: | |
666 | # bs will either be the index of the item or the insert point |
|
666 | # bs will either be the index of the item or the insert point | |
667 | start, end = _msearch(addbuf, f, start) |
|
667 | start, end = _msearch(addbuf, f, start) | |
668 | if not todelete: |
|
668 | if not todelete: | |
669 | h, fl = self._lm[f] |
|
669 | h, fl = self._lm[f] | |
670 | l = b"%s\0%s%s\n" % (f, hex(h), fl) |
|
670 | l = b"%s\0%s%s\n" % (f, hex(h), fl) | |
671 | else: |
|
671 | else: | |
672 | if start == end: |
|
672 | if start == end: | |
673 | # item we want to delete was not found, error out |
|
673 | # item we want to delete was not found, error out | |
674 | raise AssertionError( |
|
674 | raise AssertionError( | |
675 | _(b"failed to remove %s from manifest") % f |
|
675 | _(b"failed to remove %s from manifest") % f | |
676 | ) |
|
676 | ) | |
677 | l = b"" |
|
677 | l = b"" | |
678 | if dstart is not None and dstart <= start and dend >= start: |
|
678 | if dstart is not None and dstart <= start and dend >= start: | |
679 | if dend < end: |
|
679 | if dend < end: | |
680 | dend = end |
|
680 | dend = end | |
681 | if l: |
|
681 | if l: | |
682 | dline.append(l) |
|
682 | dline.append(l) | |
683 | else: |
|
683 | else: | |
684 | if dstart is not None: |
|
684 | if dstart is not None: | |
685 | delta.append([dstart, dend, b"".join(dline)]) |
|
685 | delta.append([dstart, dend, b"".join(dline)]) | |
686 | dstart = start |
|
686 | dstart = start | |
687 | dend = end |
|
687 | dend = end | |
688 | dline = [l] |
|
688 | dline = [l] | |
689 |
|
689 | |||
690 | if dstart is not None: |
|
690 | if dstart is not None: | |
691 | delta.append([dstart, dend, b"".join(dline)]) |
|
691 | delta.append([dstart, dend, b"".join(dline)]) | |
692 | # apply the delta to the base, and get a delta for addrevision |
|
692 | # apply the delta to the base, and get a delta for addrevision | |
693 | deltatext, arraytext = _addlistdelta(base, delta) |
|
693 | deltatext, arraytext = _addlistdelta(base, delta) | |
694 | else: |
|
694 | else: | |
695 | # For large changes, it's much cheaper to just build the text and |
|
695 | # For large changes, it's much cheaper to just build the text and | |
696 | # diff it. |
|
696 | # diff it. | |
697 | arraytext = bytearray(self.text()) |
|
697 | arraytext = bytearray(self.text()) | |
698 | deltatext = mdiff.textdiff( |
|
698 | deltatext = mdiff.textdiff( | |
699 | util.buffer(base), util.buffer(arraytext) |
|
699 | util.buffer(base), util.buffer(arraytext) | |
700 | ) |
|
700 | ) | |
701 |
|
701 | |||
702 | return arraytext, deltatext |
|
702 | return arraytext, deltatext | |
703 |
|
703 | |||
704 |
|
704 | |||
705 | def _msearch(m, s, lo=0, hi=None): |
|
705 | def _msearch(m, s, lo=0, hi=None): | |
706 | """return a tuple (start, end) that says where to find s within m. |
|
706 | """return a tuple (start, end) that says where to find s within m. | |
707 |
|
707 | |||
708 | If the string is found m[start:end] are the line containing |
|
708 | If the string is found m[start:end] are the line containing | |
709 | that string. If start == end the string was not found and |
|
709 | that string. If start == end the string was not found and | |
710 | they indicate the proper sorted insertion point. |
|
710 | they indicate the proper sorted insertion point. | |
711 |
|
711 | |||
712 | m should be a buffer, a memoryview or a byte string. |
|
712 | m should be a buffer, a memoryview or a byte string. | |
713 | s is a byte string""" |
|
713 | s is a byte string""" | |
714 |
|
714 | |||
715 | def advance(i, c): |
|
715 | def advance(i, c): | |
716 | while i < lenm and m[i : i + 1] != c: |
|
716 | while i < lenm and m[i : i + 1] != c: | |
717 | i += 1 |
|
717 | i += 1 | |
718 | return i |
|
718 | return i | |
719 |
|
719 | |||
720 | if not s: |
|
720 | if not s: | |
721 | return (lo, lo) |
|
721 | return (lo, lo) | |
722 | lenm = len(m) |
|
722 | lenm = len(m) | |
723 | if not hi: |
|
723 | if not hi: | |
724 | hi = lenm |
|
724 | hi = lenm | |
725 | while lo < hi: |
|
725 | while lo < hi: | |
726 | mid = (lo + hi) // 2 |
|
726 | mid = (lo + hi) // 2 | |
727 | start = mid |
|
727 | start = mid | |
728 | while start > 0 and m[start - 1 : start] != b'\n': |
|
728 | while start > 0 and m[start - 1 : start] != b'\n': | |
729 | start -= 1 |
|
729 | start -= 1 | |
730 | end = advance(start, b'\0') |
|
730 | end = advance(start, b'\0') | |
731 | if bytes(m[start:end]) < s: |
|
731 | if bytes(m[start:end]) < s: | |
732 | # we know that after the null there are 40 bytes of sha1 |
|
732 | # we know that after the null there are 40 bytes of sha1 | |
733 | # this translates to the bisect lo = mid + 1 |
|
733 | # this translates to the bisect lo = mid + 1 | |
734 | lo = advance(end + 40, b'\n') + 1 |
|
734 | lo = advance(end + 40, b'\n') + 1 | |
735 | else: |
|
735 | else: | |
736 | # this translates to the bisect hi = mid |
|
736 | # this translates to the bisect hi = mid | |
737 | hi = start |
|
737 | hi = start | |
738 | end = advance(lo, b'\0') |
|
738 | end = advance(lo, b'\0') | |
739 | found = m[lo:end] |
|
739 | found = m[lo:end] | |
740 | if s == found: |
|
740 | if s == found: | |
741 | # we know that after the null there are 40 bytes of sha1 |
|
741 | # we know that after the null there are 40 bytes of sha1 | |
742 | end = advance(end + 40, b'\n') |
|
742 | end = advance(end + 40, b'\n') | |
743 | return (lo, end + 1) |
|
743 | return (lo, end + 1) | |
744 | else: |
|
744 | else: | |
745 | return (lo, lo) |
|
745 | return (lo, lo) | |
746 |
|
746 | |||
747 |
|
747 | |||
748 | def _checkforbidden(l): |
|
748 | def _checkforbidden(l): | |
749 | """Check filenames for illegal characters.""" |
|
749 | """Check filenames for illegal characters.""" | |
750 | for f in l: |
|
750 | for f in l: | |
751 | if b'\n' in f or b'\r' in f: |
|
751 | if b'\n' in f or b'\r' in f: | |
752 | raise error.StorageError( |
|
752 | raise error.StorageError( | |
753 | _(b"'\\n' and '\\r' disallowed in filenames: %r") |
|
753 | _(b"'\\n' and '\\r' disallowed in filenames: %r") | |
754 | % pycompat.bytestr(f) |
|
754 | % pycompat.bytestr(f) | |
755 | ) |
|
755 | ) | |
756 |
|
756 | |||
757 |
|
757 | |||
758 | # apply the changes collected during the bisect loop to our addlist |
|
758 | # apply the changes collected during the bisect loop to our addlist | |
759 | # return a delta suitable for addrevision |
|
759 | # return a delta suitable for addrevision | |
760 | def _addlistdelta(addlist, x): |
|
760 | def _addlistdelta(addlist, x): | |
761 | # for large addlist arrays, building a new array is cheaper |
|
761 | # for large addlist arrays, building a new array is cheaper | |
762 | # than repeatedly modifying the existing one |
|
762 | # than repeatedly modifying the existing one | |
763 | currentposition = 0 |
|
763 | currentposition = 0 | |
764 | newaddlist = bytearray() |
|
764 | newaddlist = bytearray() | |
765 |
|
765 | |||
766 | for start, end, content in x: |
|
766 | for start, end, content in x: | |
767 | newaddlist += addlist[currentposition:start] |
|
767 | newaddlist += addlist[currentposition:start] | |
768 | if content: |
|
768 | if content: | |
769 | newaddlist += bytearray(content) |
|
769 | newaddlist += bytearray(content) | |
770 |
|
770 | |||
771 | currentposition = end |
|
771 | currentposition = end | |
772 |
|
772 | |||
773 | newaddlist += addlist[currentposition:] |
|
773 | newaddlist += addlist[currentposition:] | |
774 |
|
774 | |||
775 | deltatext = b"".join( |
|
775 | deltatext = b"".join( | |
776 | struct.pack(b">lll", start, end, len(content)) + content |
|
776 | struct.pack(b">lll", start, end, len(content)) + content | |
777 | for start, end, content in x |
|
777 | for start, end, content in x | |
778 | ) |
|
778 | ) | |
779 | return deltatext, newaddlist |
|
779 | return deltatext, newaddlist | |
780 |
|
780 | |||
781 |
|
781 | |||
782 | def _splittopdir(f): |
|
782 | def _splittopdir(f): | |
783 | if b'/' in f: |
|
783 | if b'/' in f: | |
784 | dir, subpath = f.split(b'/', 1) |
|
784 | dir, subpath = f.split(b'/', 1) | |
785 | return dir + b'/', subpath |
|
785 | return dir + b'/', subpath | |
786 | else: |
|
786 | else: | |
787 | return b'', f |
|
787 | return b'', f | |
788 |
|
788 | |||
789 |
|
789 | |||
790 | _noop = lambda s: None |
|
790 | _noop = lambda s: None | |
791 |
|
791 | |||
792 |
|
792 | |||
793 | @interfaceutil.implementer(repository.imanifestdict) |
|
793 | @interfaceutil.implementer(repository.imanifestdict) | |
794 | class treemanifest(object): |
|
794 | class treemanifest(object): | |
795 | def __init__(self, dir=b'', text=b''): |
|
795 | def __init__(self, dir=b'', text=b''): | |
796 | self._dir = dir |
|
796 | self._dir = dir | |
797 | self._node = nullid |
|
797 | self._node = nullid | |
798 | self._loadfunc = _noop |
|
798 | self._loadfunc = _noop | |
799 | self._copyfunc = _noop |
|
799 | self._copyfunc = _noop | |
800 | self._dirty = False |
|
800 | self._dirty = False | |
801 | self._dirs = {} |
|
801 | self._dirs = {} | |
802 | self._lazydirs = {} |
|
802 | self._lazydirs = {} | |
803 | # Using _lazymanifest here is a little slower than plain old dicts |
|
803 | # Using _lazymanifest here is a little slower than plain old dicts | |
804 | self._files = {} |
|
804 | self._files = {} | |
805 | self._flags = {} |
|
805 | self._flags = {} | |
806 | if text: |
|
806 | if text: | |
807 |
|
807 | |||
808 | def readsubtree(subdir, subm): |
|
808 | def readsubtree(subdir, subm): | |
809 | raise AssertionError( |
|
809 | raise AssertionError( | |
810 | b'treemanifest constructor only accepts flat manifests' |
|
810 | b'treemanifest constructor only accepts flat manifests' | |
811 | ) |
|
811 | ) | |
812 |
|
812 | |||
813 | self.parse(text, readsubtree) |
|
813 | self.parse(text, readsubtree) | |
814 | self._dirty = True # Mark flat manifest dirty after parsing |
|
814 | self._dirty = True # Mark flat manifest dirty after parsing | |
815 |
|
815 | |||
816 | def _subpath(self, path): |
|
816 | def _subpath(self, path): | |
817 | return self._dir + path |
|
817 | return self._dir + path | |
818 |
|
818 | |||
819 | def _loadalllazy(self): |
|
819 | def _loadalllazy(self): | |
820 | selfdirs = self._dirs |
|
820 | selfdirs = self._dirs | |
821 | subpath = self._subpath |
|
821 | subpath = self._subpath | |
822 | for d, (node, readsubtree, docopy) in pycompat.iteritems( |
|
822 | for d, (node, readsubtree, docopy) in pycompat.iteritems( | |
823 | self._lazydirs |
|
823 | self._lazydirs | |
824 | ): |
|
824 | ): | |
825 | if docopy: |
|
825 | if docopy: | |
826 | selfdirs[d] = readsubtree(subpath(d), node).copy() |
|
826 | selfdirs[d] = readsubtree(subpath(d), node).copy() | |
827 | else: |
|
827 | else: | |
828 | selfdirs[d] = readsubtree(subpath(d), node) |
|
828 | selfdirs[d] = readsubtree(subpath(d), node) | |
829 | self._lazydirs = {} |
|
829 | self._lazydirs = {} | |
830 |
|
830 | |||
831 | def _loadlazy(self, d): |
|
831 | def _loadlazy(self, d): | |
832 | v = self._lazydirs.get(d) |
|
832 | v = self._lazydirs.get(d) | |
833 | if v: |
|
833 | if v: | |
834 | node, readsubtree, docopy = v |
|
834 | node, readsubtree, docopy = v | |
835 | if docopy: |
|
835 | if docopy: | |
836 | self._dirs[d] = readsubtree(self._subpath(d), node).copy() |
|
836 | self._dirs[d] = readsubtree(self._subpath(d), node).copy() | |
837 | else: |
|
837 | else: | |
838 | self._dirs[d] = readsubtree(self._subpath(d), node) |
|
838 | self._dirs[d] = readsubtree(self._subpath(d), node) | |
839 | del self._lazydirs[d] |
|
839 | del self._lazydirs[d] | |
840 |
|
840 | |||
841 | def _loadchildrensetlazy(self, visit): |
|
841 | def _loadchildrensetlazy(self, visit): | |
842 | if not visit: |
|
842 | if not visit: | |
843 | return None |
|
843 | return None | |
844 | if visit == b'all' or visit == b'this': |
|
844 | if visit == b'all' or visit == b'this': | |
845 | self._loadalllazy() |
|
845 | self._loadalllazy() | |
846 | return None |
|
846 | return None | |
847 |
|
847 | |||
848 | loadlazy = self._loadlazy |
|
848 | loadlazy = self._loadlazy | |
849 | for k in visit: |
|
849 | for k in visit: | |
850 | loadlazy(k + b'/') |
|
850 | loadlazy(k + b'/') | |
851 | return visit |
|
851 | return visit | |
852 |
|
852 | |||
853 | def _loaddifflazy(self, t1, t2): |
|
853 | def _loaddifflazy(self, t1, t2): | |
854 | """load items in t1 and t2 if they're needed for diffing. |
|
854 | """load items in t1 and t2 if they're needed for diffing. | |
855 |
|
855 | |||
856 | The criteria currently is: |
|
856 | The criteria currently is: | |
857 | - if it's not present in _lazydirs in either t1 or t2, load it in the |
|
857 | - if it's not present in _lazydirs in either t1 or t2, load it in the | |
858 | other (it may already be loaded or it may not exist, doesn't matter) |
|
858 | other (it may already be loaded or it may not exist, doesn't matter) | |
859 | - if it's present in _lazydirs in both, compare the nodeid; if it |
|
859 | - if it's present in _lazydirs in both, compare the nodeid; if it | |
860 | differs, load it in both |
|
860 | differs, load it in both | |
861 | """ |
|
861 | """ | |
862 | toloadlazy = [] |
|
862 | toloadlazy = [] | |
863 | for d, v1 in pycompat.iteritems(t1._lazydirs): |
|
863 | for d, v1 in pycompat.iteritems(t1._lazydirs): | |
864 | v2 = t2._lazydirs.get(d) |
|
864 | v2 = t2._lazydirs.get(d) | |
865 | if not v2 or v2[0] != v1[0]: |
|
865 | if not v2 or v2[0] != v1[0]: | |
866 | toloadlazy.append(d) |
|
866 | toloadlazy.append(d) | |
867 | for d, v1 in pycompat.iteritems(t2._lazydirs): |
|
867 | for d, v1 in pycompat.iteritems(t2._lazydirs): | |
868 | if d not in t1._lazydirs: |
|
868 | if d not in t1._lazydirs: | |
869 | toloadlazy.append(d) |
|
869 | toloadlazy.append(d) | |
870 |
|
870 | |||
871 | for d in toloadlazy: |
|
871 | for d in toloadlazy: | |
872 | t1._loadlazy(d) |
|
872 | t1._loadlazy(d) | |
873 | t2._loadlazy(d) |
|
873 | t2._loadlazy(d) | |
874 |
|
874 | |||
875 | def __len__(self): |
|
875 | def __len__(self): | |
876 | self._load() |
|
876 | self._load() | |
877 | size = len(self._files) |
|
877 | size = len(self._files) | |
878 | self._loadalllazy() |
|
878 | self._loadalllazy() | |
879 | for m in self._dirs.values(): |
|
879 | for m in self._dirs.values(): | |
880 | size += m.__len__() |
|
880 | size += m.__len__() | |
881 | return size |
|
881 | return size | |
882 |
|
882 | |||
883 | def __nonzero__(self): |
|
883 | def __nonzero__(self): | |
884 | # Faster than "__len() != 0" since it avoids loading sub-manifests |
|
884 | # Faster than "__len() != 0" since it avoids loading sub-manifests | |
885 | return not self._isempty() |
|
885 | return not self._isempty() | |
886 |
|
886 | |||
887 | __bool__ = __nonzero__ |
|
887 | __bool__ = __nonzero__ | |
888 |
|
888 | |||
889 | def _isempty(self): |
|
889 | def _isempty(self): | |
890 | self._load() # for consistency; already loaded by all callers |
|
890 | self._load() # for consistency; already loaded by all callers | |
891 | # See if we can skip loading everything. |
|
891 | # See if we can skip loading everything. | |
892 | if self._files or ( |
|
892 | if self._files or ( | |
893 | self._dirs and any(not m._isempty() for m in self._dirs.values()) |
|
893 | self._dirs and any(not m._isempty() for m in self._dirs.values()) | |
894 | ): |
|
894 | ): | |
895 | return False |
|
895 | return False | |
896 | self._loadalllazy() |
|
896 | self._loadalllazy() | |
897 | return not self._dirs or all(m._isempty() for m in self._dirs.values()) |
|
897 | return not self._dirs or all(m._isempty() for m in self._dirs.values()) | |
898 |
|
898 | |||
899 | @encoding.strmethod |
|
899 | @encoding.strmethod | |
900 | def __repr__(self): |
|
900 | def __repr__(self): | |
901 | return ( |
|
901 | return ( | |
902 | b'<treemanifest dir=%s, node=%s, loaded=%r, dirty=%r at 0x%x>' |
|
902 | b'<treemanifest dir=%s, node=%s, loaded=%r, dirty=%r at 0x%x>' | |
903 | % ( |
|
903 | % ( | |
904 | self._dir, |
|
904 | self._dir, | |
905 | hex(self._node), |
|
905 | hex(self._node), | |
906 | bool(self._loadfunc is _noop), |
|
906 | bool(self._loadfunc is _noop), | |
907 | self._dirty, |
|
907 | self._dirty, | |
908 | id(self), |
|
908 | id(self), | |
909 | ) |
|
909 | ) | |
910 | ) |
|
910 | ) | |
911 |
|
911 | |||
912 | def dir(self): |
|
912 | def dir(self): | |
913 | """The directory that this tree manifest represents, including a |
|
913 | """The directory that this tree manifest represents, including a | |
914 | trailing '/'. Empty string for the repo root directory.""" |
|
914 | trailing '/'. Empty string for the repo root directory.""" | |
915 | return self._dir |
|
915 | return self._dir | |
916 |
|
916 | |||
917 | def node(self): |
|
917 | def node(self): | |
918 | """This node of this instance. nullid for unsaved instances. Should |
|
918 | """This node of this instance. nullid for unsaved instances. Should | |
919 | be updated when the instance is read or written from a revlog. |
|
919 | be updated when the instance is read or written from a revlog. | |
920 | """ |
|
920 | """ | |
921 | assert not self._dirty |
|
921 | assert not self._dirty | |
922 | return self._node |
|
922 | return self._node | |
923 |
|
923 | |||
924 | def setnode(self, node): |
|
924 | def setnode(self, node): | |
925 | self._node = node |
|
925 | self._node = node | |
926 | self._dirty = False |
|
926 | self._dirty = False | |
927 |
|
927 | |||
928 | def iterentries(self): |
|
928 | def iterentries(self): | |
929 | self._load() |
|
929 | self._load() | |
930 | self._loadalllazy() |
|
930 | self._loadalllazy() | |
931 | for p, n in sorted( |
|
931 | for p, n in sorted( | |
932 | itertools.chain(self._dirs.items(), self._files.items()) |
|
932 | itertools.chain(self._dirs.items(), self._files.items()) | |
933 | ): |
|
933 | ): | |
934 | if p in self._files: |
|
934 | if p in self._files: | |
935 | yield self._subpath(p), n, self._flags.get(p, b'') |
|
935 | yield self._subpath(p), n, self._flags.get(p, b'') | |
936 | else: |
|
936 | else: | |
937 | for x in n.iterentries(): |
|
937 | for x in n.iterentries(): | |
938 | yield x |
|
938 | yield x | |
939 |
|
939 | |||
940 | def items(self): |
|
940 | def items(self): | |
941 | self._load() |
|
941 | self._load() | |
942 | self._loadalllazy() |
|
942 | self._loadalllazy() | |
943 | for p, n in sorted( |
|
943 | for p, n in sorted( | |
944 | itertools.chain(self._dirs.items(), self._files.items()) |
|
944 | itertools.chain(self._dirs.items(), self._files.items()) | |
945 | ): |
|
945 | ): | |
946 | if p in self._files: |
|
946 | if p in self._files: | |
947 | yield self._subpath(p), n |
|
947 | yield self._subpath(p), n | |
948 | else: |
|
948 | else: | |
949 | for f, sn in pycompat.iteritems(n): |
|
949 | for f, sn in pycompat.iteritems(n): | |
950 | yield f, sn |
|
950 | yield f, sn | |
951 |
|
951 | |||
952 | iteritems = items |
|
952 | iteritems = items | |
953 |
|
953 | |||
954 | def iterkeys(self): |
|
954 | def iterkeys(self): | |
955 | self._load() |
|
955 | self._load() | |
956 | self._loadalllazy() |
|
956 | self._loadalllazy() | |
957 | for p in sorted(itertools.chain(self._dirs, self._files)): |
|
957 | for p in sorted(itertools.chain(self._dirs, self._files)): | |
958 | if p in self._files: |
|
958 | if p in self._files: | |
959 | yield self._subpath(p) |
|
959 | yield self._subpath(p) | |
960 | else: |
|
960 | else: | |
961 | for f in self._dirs[p]: |
|
961 | for f in self._dirs[p]: | |
962 | yield f |
|
962 | yield f | |
963 |
|
963 | |||
964 | def keys(self): |
|
964 | def keys(self): | |
965 | return list(self.iterkeys()) |
|
965 | return list(self.iterkeys()) | |
966 |
|
966 | |||
967 | def __iter__(self): |
|
967 | def __iter__(self): | |
968 | return self.iterkeys() |
|
968 | return self.iterkeys() | |
969 |
|
969 | |||
970 | def __contains__(self, f): |
|
970 | def __contains__(self, f): | |
971 | if f is None: |
|
971 | if f is None: | |
972 | return False |
|
972 | return False | |
973 | self._load() |
|
973 | self._load() | |
974 | dir, subpath = _splittopdir(f) |
|
974 | dir, subpath = _splittopdir(f) | |
975 | if dir: |
|
975 | if dir: | |
976 | self._loadlazy(dir) |
|
976 | self._loadlazy(dir) | |
977 |
|
977 | |||
978 | if dir not in self._dirs: |
|
978 | if dir not in self._dirs: | |
979 | return False |
|
979 | return False | |
980 |
|
980 | |||
981 | return self._dirs[dir].__contains__(subpath) |
|
981 | return self._dirs[dir].__contains__(subpath) | |
982 | else: |
|
982 | else: | |
983 | return f in self._files |
|
983 | return f in self._files | |
984 |
|
984 | |||
985 | def get(self, f, default=None): |
|
985 | def get(self, f, default=None): | |
986 | self._load() |
|
986 | self._load() | |
987 | dir, subpath = _splittopdir(f) |
|
987 | dir, subpath = _splittopdir(f) | |
988 | if dir: |
|
988 | if dir: | |
989 | self._loadlazy(dir) |
|
989 | self._loadlazy(dir) | |
990 |
|
990 | |||
991 | if dir not in self._dirs: |
|
991 | if dir not in self._dirs: | |
992 | return default |
|
992 | return default | |
993 | return self._dirs[dir].get(subpath, default) |
|
993 | return self._dirs[dir].get(subpath, default) | |
994 | else: |
|
994 | else: | |
995 | return self._files.get(f, default) |
|
995 | return self._files.get(f, default) | |
996 |
|
996 | |||
997 | def __getitem__(self, f): |
|
997 | def __getitem__(self, f): | |
998 | self._load() |
|
998 | self._load() | |
999 | dir, subpath = _splittopdir(f) |
|
999 | dir, subpath = _splittopdir(f) | |
1000 | if dir: |
|
1000 | if dir: | |
1001 | self._loadlazy(dir) |
|
1001 | self._loadlazy(dir) | |
1002 |
|
1002 | |||
1003 | return self._dirs[dir].__getitem__(subpath) |
|
1003 | return self._dirs[dir].__getitem__(subpath) | |
1004 | else: |
|
1004 | else: | |
1005 | return self._files[f] |
|
1005 | return self._files[f] | |
1006 |
|
1006 | |||
1007 | def flags(self, f): |
|
1007 | def flags(self, f): | |
1008 | self._load() |
|
1008 | self._load() | |
1009 | dir, subpath = _splittopdir(f) |
|
1009 | dir, subpath = _splittopdir(f) | |
1010 | if dir: |
|
1010 | if dir: | |
1011 | self._loadlazy(dir) |
|
1011 | self._loadlazy(dir) | |
1012 |
|
1012 | |||
1013 | if dir not in self._dirs: |
|
1013 | if dir not in self._dirs: | |
1014 | return b'' |
|
1014 | return b'' | |
1015 | return self._dirs[dir].flags(subpath) |
|
1015 | return self._dirs[dir].flags(subpath) | |
1016 | else: |
|
1016 | else: | |
1017 | if f in self._lazydirs or f in self._dirs: |
|
1017 | if f in self._lazydirs or f in self._dirs: | |
1018 | return b'' |
|
1018 | return b'' | |
1019 | return self._flags.get(f, b'') |
|
1019 | return self._flags.get(f, b'') | |
1020 |
|
1020 | |||
1021 | def find(self, f): |
|
1021 | def find(self, f): | |
1022 | self._load() |
|
1022 | self._load() | |
1023 | dir, subpath = _splittopdir(f) |
|
1023 | dir, subpath = _splittopdir(f) | |
1024 | if dir: |
|
1024 | if dir: | |
1025 | self._loadlazy(dir) |
|
1025 | self._loadlazy(dir) | |
1026 |
|
1026 | |||
1027 | return self._dirs[dir].find(subpath) |
|
1027 | return self._dirs[dir].find(subpath) | |
1028 | else: |
|
1028 | else: | |
1029 | return self._files[f], self._flags.get(f, b'') |
|
1029 | return self._files[f], self._flags.get(f, b'') | |
1030 |
|
1030 | |||
1031 | def __delitem__(self, f): |
|
1031 | def __delitem__(self, f): | |
1032 | self._load() |
|
1032 | self._load() | |
1033 | dir, subpath = _splittopdir(f) |
|
1033 | dir, subpath = _splittopdir(f) | |
1034 | if dir: |
|
1034 | if dir: | |
1035 | self._loadlazy(dir) |
|
1035 | self._loadlazy(dir) | |
1036 |
|
1036 | |||
1037 | self._dirs[dir].__delitem__(subpath) |
|
1037 | self._dirs[dir].__delitem__(subpath) | |
1038 | # If the directory is now empty, remove it |
|
1038 | # If the directory is now empty, remove it | |
1039 | if self._dirs[dir]._isempty(): |
|
1039 | if self._dirs[dir]._isempty(): | |
1040 | del self._dirs[dir] |
|
1040 | del self._dirs[dir] | |
1041 | else: |
|
1041 | else: | |
1042 | del self._files[f] |
|
1042 | del self._files[f] | |
1043 | if f in self._flags: |
|
1043 | if f in self._flags: | |
1044 | del self._flags[f] |
|
1044 | del self._flags[f] | |
1045 | self._dirty = True |
|
1045 | self._dirty = True | |
1046 |
|
1046 | |||
1047 | def __setitem__(self, f, n): |
|
1047 | def __setitem__(self, f, n): | |
1048 | assert n is not None |
|
1048 | assert n is not None | |
1049 | self._load() |
|
1049 | self._load() | |
1050 | dir, subpath = _splittopdir(f) |
|
1050 | dir, subpath = _splittopdir(f) | |
1051 | if dir: |
|
1051 | if dir: | |
1052 | self._loadlazy(dir) |
|
1052 | self._loadlazy(dir) | |
1053 | if dir not in self._dirs: |
|
1053 | if dir not in self._dirs: | |
1054 | self._dirs[dir] = treemanifest(self._subpath(dir)) |
|
1054 | self._dirs[dir] = treemanifest(self._subpath(dir)) | |
1055 | self._dirs[dir].__setitem__(subpath, n) |
|
1055 | self._dirs[dir].__setitem__(subpath, n) | |
1056 | else: |
|
1056 | else: | |
1057 | # manifest nodes are either 20 bytes or 32 bytes, |
|
1057 | # manifest nodes are either 20 bytes or 32 bytes, | |
1058 | # depending on the hash in use. Assert this as historically |
|
1058 | # depending on the hash in use. Assert this as historically | |
1059 | # sometimes extra bytes were added. |
|
1059 | # sometimes extra bytes were added. | |
1060 | assert len(n) in (20, 32) |
|
1060 | assert len(n) in (20, 32) | |
1061 | self._files[f] = n |
|
1061 | self._files[f] = n | |
1062 | self._dirty = True |
|
1062 | self._dirty = True | |
1063 |
|
1063 | |||
1064 | def _load(self): |
|
1064 | def _load(self): | |
1065 | if self._loadfunc is not _noop: |
|
1065 | if self._loadfunc is not _noop: | |
1066 | lf, self._loadfunc = self._loadfunc, _noop |
|
1066 | lf, self._loadfunc = self._loadfunc, _noop | |
1067 | lf(self) |
|
1067 | lf(self) | |
1068 | elif self._copyfunc is not _noop: |
|
1068 | elif self._copyfunc is not _noop: | |
1069 | cf, self._copyfunc = self._copyfunc, _noop |
|
1069 | cf, self._copyfunc = self._copyfunc, _noop | |
1070 | cf(self) |
|
1070 | cf(self) | |
1071 |
|
1071 | |||
1072 | def setflag(self, f, flags): |
|
1072 | def setflag(self, f, flags): | |
1073 | """Set the flags (symlink, executable) for path f.""" |
|
1073 | """Set the flags (symlink, executable) for path f.""" | |
1074 | if flags not in _manifestflags: |
|
1074 | if flags not in _manifestflags: | |
1075 | raise TypeError(b"Invalid manifest flag set.") |
|
1075 | raise TypeError(b"Invalid manifest flag set.") | |
1076 | self._load() |
|
1076 | self._load() | |
1077 | dir, subpath = _splittopdir(f) |
|
1077 | dir, subpath = _splittopdir(f) | |
1078 | if dir: |
|
1078 | if dir: | |
1079 | self._loadlazy(dir) |
|
1079 | self._loadlazy(dir) | |
1080 | if dir not in self._dirs: |
|
1080 | if dir not in self._dirs: | |
1081 | self._dirs[dir] = treemanifest(self._subpath(dir)) |
|
1081 | self._dirs[dir] = treemanifest(self._subpath(dir)) | |
1082 | self._dirs[dir].setflag(subpath, flags) |
|
1082 | self._dirs[dir].setflag(subpath, flags) | |
1083 | else: |
|
1083 | else: | |
1084 | self._flags[f] = flags |
|
1084 | self._flags[f] = flags | |
1085 | self._dirty = True |
|
1085 | self._dirty = True | |
1086 |
|
1086 | |||
1087 | def copy(self): |
|
1087 | def copy(self): | |
1088 | copy = treemanifest(self._dir) |
|
1088 | copy = treemanifest(self._dir) | |
1089 | copy._node = self._node |
|
1089 | copy._node = self._node | |
1090 | copy._dirty = self._dirty |
|
1090 | copy._dirty = self._dirty | |
1091 | if self._copyfunc is _noop: |
|
1091 | if self._copyfunc is _noop: | |
1092 |
|
1092 | |||
1093 | def _copyfunc(s): |
|
1093 | def _copyfunc(s): | |
1094 | self._load() |
|
1094 | self._load() | |
1095 | s._lazydirs = { |
|
1095 | s._lazydirs = { | |
1096 | d: (n, r, True) |
|
1096 | d: (n, r, True) | |
1097 | for d, (n, r, c) in pycompat.iteritems(self._lazydirs) |
|
1097 | for d, (n, r, c) in pycompat.iteritems(self._lazydirs) | |
1098 | } |
|
1098 | } | |
1099 | sdirs = s._dirs |
|
1099 | sdirs = s._dirs | |
1100 | for d, v in pycompat.iteritems(self._dirs): |
|
1100 | for d, v in pycompat.iteritems(self._dirs): | |
1101 | sdirs[d] = v.copy() |
|
1101 | sdirs[d] = v.copy() | |
1102 | s._files = dict.copy(self._files) |
|
1102 | s._files = dict.copy(self._files) | |
1103 | s._flags = dict.copy(self._flags) |
|
1103 | s._flags = dict.copy(self._flags) | |
1104 |
|
1104 | |||
1105 | if self._loadfunc is _noop: |
|
1105 | if self._loadfunc is _noop: | |
1106 | _copyfunc(copy) |
|
1106 | _copyfunc(copy) | |
1107 | else: |
|
1107 | else: | |
1108 | copy._copyfunc = _copyfunc |
|
1108 | copy._copyfunc = _copyfunc | |
1109 | else: |
|
1109 | else: | |
1110 | copy._copyfunc = self._copyfunc |
|
1110 | copy._copyfunc = self._copyfunc | |
1111 | return copy |
|
1111 | return copy | |
1112 |
|
1112 | |||
1113 | def filesnotin(self, m2, match=None): |
|
1113 | def filesnotin(self, m2, match=None): | |
1114 | '''Set of files in this manifest that are not in the other''' |
|
1114 | '''Set of files in this manifest that are not in the other''' | |
1115 | if match and not match.always(): |
|
1115 | if match and not match.always(): | |
1116 | m1 = self._matches(match) |
|
1116 | m1 = self._matches(match) | |
1117 | m2 = m2._matches(match) |
|
1117 | m2 = m2._matches(match) | |
1118 | return m1.filesnotin(m2) |
|
1118 | return m1.filesnotin(m2) | |
1119 |
|
1119 | |||
1120 | files = set() |
|
1120 | files = set() | |
1121 |
|
1121 | |||
1122 | def _filesnotin(t1, t2): |
|
1122 | def _filesnotin(t1, t2): | |
1123 | if t1._node == t2._node and not t1._dirty and not t2._dirty: |
|
1123 | if t1._node == t2._node and not t1._dirty and not t2._dirty: | |
1124 | return |
|
1124 | return | |
1125 | t1._load() |
|
1125 | t1._load() | |
1126 | t2._load() |
|
1126 | t2._load() | |
1127 | self._loaddifflazy(t1, t2) |
|
1127 | self._loaddifflazy(t1, t2) | |
1128 | for d, m1 in pycompat.iteritems(t1._dirs): |
|
1128 | for d, m1 in pycompat.iteritems(t1._dirs): | |
1129 | if d in t2._dirs: |
|
1129 | if d in t2._dirs: | |
1130 | m2 = t2._dirs[d] |
|
1130 | m2 = t2._dirs[d] | |
1131 | _filesnotin(m1, m2) |
|
1131 | _filesnotin(m1, m2) | |
1132 | else: |
|
1132 | else: | |
1133 | files.update(m1.iterkeys()) |
|
1133 | files.update(m1.iterkeys()) | |
1134 |
|
1134 | |||
1135 | for fn in t1._files: |
|
1135 | for fn in t1._files: | |
1136 | if fn not in t2._files: |
|
1136 | if fn not in t2._files: | |
1137 | files.add(t1._subpath(fn)) |
|
1137 | files.add(t1._subpath(fn)) | |
1138 |
|
1138 | |||
1139 | _filesnotin(self, m2) |
|
1139 | _filesnotin(self, m2) | |
1140 | return files |
|
1140 | return files | |
1141 |
|
1141 | |||
1142 | @propertycache |
|
1142 | @propertycache | |
1143 | def _alldirs(self): |
|
1143 | def _alldirs(self): | |
1144 | return pathutil.dirs(self) |
|
1144 | return pathutil.dirs(self) | |
1145 |
|
1145 | |||
1146 | def dirs(self): |
|
1146 | def dirs(self): | |
1147 | return self._alldirs |
|
1147 | return self._alldirs | |
1148 |
|
1148 | |||
1149 | def hasdir(self, dir): |
|
1149 | def hasdir(self, dir): | |
1150 | self._load() |
|
1150 | self._load() | |
1151 | topdir, subdir = _splittopdir(dir) |
|
1151 | topdir, subdir = _splittopdir(dir) | |
1152 | if topdir: |
|
1152 | if topdir: | |
1153 | self._loadlazy(topdir) |
|
1153 | self._loadlazy(topdir) | |
1154 | if topdir in self._dirs: |
|
1154 | if topdir in self._dirs: | |
1155 | return self._dirs[topdir].hasdir(subdir) |
|
1155 | return self._dirs[topdir].hasdir(subdir) | |
1156 | return False |
|
1156 | return False | |
1157 | dirslash = dir + b'/' |
|
1157 | dirslash = dir + b'/' | |
1158 | return dirslash in self._dirs or dirslash in self._lazydirs |
|
1158 | return dirslash in self._dirs or dirslash in self._lazydirs | |
1159 |
|
1159 | |||
1160 | def walk(self, match): |
|
1160 | def walk(self, match): | |
1161 | """Generates matching file names. |
|
1161 | """Generates matching file names. | |
1162 |
|
1162 | |||
1163 | It also reports nonexistent files by marking them bad with match.bad(). |
|
1163 | It also reports nonexistent files by marking them bad with match.bad(). | |
1164 | """ |
|
1164 | """ | |
1165 | if match.always(): |
|
1165 | if match.always(): | |
1166 | for f in iter(self): |
|
1166 | for f in iter(self): | |
1167 | yield f |
|
1167 | yield f | |
1168 | return |
|
1168 | return | |
1169 |
|
1169 | |||
1170 | fset = set(match.files()) |
|
1170 | fset = set(match.files()) | |
1171 |
|
1171 | |||
1172 | for fn in self._walk(match): |
|
1172 | for fn in self._walk(match): | |
1173 | if fn in fset: |
|
1173 | if fn in fset: | |
1174 | # specified pattern is the exact name |
|
1174 | # specified pattern is the exact name | |
1175 | fset.remove(fn) |
|
1175 | fset.remove(fn) | |
1176 | yield fn |
|
1176 | yield fn | |
1177 |
|
1177 | |||
1178 | # for dirstate.walk, files=[''] means "walk the whole tree". |
|
1178 | # for dirstate.walk, files=[''] means "walk the whole tree". | |
1179 | # follow that here, too |
|
1179 | # follow that here, too | |
1180 | fset.discard(b'') |
|
1180 | fset.discard(b'') | |
1181 |
|
1181 | |||
1182 | for fn in sorted(fset): |
|
1182 | for fn in sorted(fset): | |
1183 | if not self.hasdir(fn): |
|
1183 | if not self.hasdir(fn): | |
1184 | match.bad(fn, None) |
|
1184 | match.bad(fn, None) | |
1185 |
|
1185 | |||
1186 | def _walk(self, match): |
|
1186 | def _walk(self, match): | |
1187 | '''Recursively generates matching file names for walk().''' |
|
1187 | '''Recursively generates matching file names for walk().''' | |
1188 | visit = match.visitchildrenset(self._dir[:-1]) |
|
1188 | visit = match.visitchildrenset(self._dir[:-1]) | |
1189 | if not visit: |
|
1189 | if not visit: | |
1190 | return |
|
1190 | return | |
1191 |
|
1191 | |||
1192 | # yield this dir's files and walk its submanifests |
|
1192 | # yield this dir's files and walk its submanifests | |
1193 | self._load() |
|
1193 | self._load() | |
1194 | visit = self._loadchildrensetlazy(visit) |
|
1194 | visit = self._loadchildrensetlazy(visit) | |
1195 | for p in sorted(list(self._dirs) + list(self._files)): |
|
1195 | for p in sorted(list(self._dirs) + list(self._files)): | |
1196 | if p in self._files: |
|
1196 | if p in self._files: | |
1197 | fullp = self._subpath(p) |
|
1197 | fullp = self._subpath(p) | |
1198 | if match(fullp): |
|
1198 | if match(fullp): | |
1199 | yield fullp |
|
1199 | yield fullp | |
1200 | else: |
|
1200 | else: | |
1201 | if not visit or p[:-1] in visit: |
|
1201 | if not visit or p[:-1] in visit: | |
1202 | for f in self._dirs[p]._walk(match): |
|
1202 | for f in self._dirs[p]._walk(match): | |
1203 | yield f |
|
1203 | yield f | |
1204 |
|
1204 | |||
1205 | def _matches(self, match): |
|
1205 | def _matches(self, match): | |
1206 | """recursively generate a new manifest filtered by the match argument.""" |
|
1206 | """recursively generate a new manifest filtered by the match argument.""" | |
1207 | if match.always(): |
|
1207 | if match.always(): | |
1208 | return self.copy() |
|
1208 | return self.copy() | |
1209 | return self._matches_inner(match) |
|
1209 | return self._matches_inner(match) | |
1210 |
|
1210 | |||
1211 | def _matches_inner(self, match): |
|
1211 | def _matches_inner(self, match): | |
1212 | if match.always(): |
|
1212 | if match.always(): | |
1213 | return self.copy() |
|
1213 | return self.copy() | |
1214 |
|
1214 | |||
1215 | visit = match.visitchildrenset(self._dir[:-1]) |
|
1215 | visit = match.visitchildrenset(self._dir[:-1]) | |
1216 | if visit == b'all': |
|
1216 | if visit == b'all': | |
1217 | return self.copy() |
|
1217 | return self.copy() | |
1218 | ret = treemanifest(self._dir) |
|
1218 | ret = treemanifest(self._dir) | |
1219 | if not visit: |
|
1219 | if not visit: | |
1220 | return ret |
|
1220 | return ret | |
1221 |
|
1221 | |||
1222 | self._load() |
|
1222 | self._load() | |
1223 | for fn in self._files: |
|
1223 | for fn in self._files: | |
1224 | # While visitchildrenset *usually* lists only subdirs, this is |
|
1224 | # While visitchildrenset *usually* lists only subdirs, this is | |
1225 | # actually up to the matcher and may have some files in the set(). |
|
1225 | # actually up to the matcher and may have some files in the set(). | |
1226 | # If visit == 'this', we should obviously look at the files in this |
|
1226 | # If visit == 'this', we should obviously look at the files in this | |
1227 | # directory; if visit is a set, and fn is in it, we should inspect |
|
1227 | # directory; if visit is a set, and fn is in it, we should inspect | |
1228 | # fn (but no need to inspect things not in the set). |
|
1228 | # fn (but no need to inspect things not in the set). | |
1229 | if visit != b'this' and fn not in visit: |
|
1229 | if visit != b'this' and fn not in visit: | |
1230 | continue |
|
1230 | continue | |
1231 | fullp = self._subpath(fn) |
|
1231 | fullp = self._subpath(fn) | |
1232 | # visitchildrenset isn't perfect, we still need to call the regular |
|
1232 | # visitchildrenset isn't perfect, we still need to call the regular | |
1233 | # matcher code to further filter results. |
|
1233 | # matcher code to further filter results. | |
1234 | if not match(fullp): |
|
1234 | if not match(fullp): | |
1235 | continue |
|
1235 | continue | |
1236 | ret._files[fn] = self._files[fn] |
|
1236 | ret._files[fn] = self._files[fn] | |
1237 | if fn in self._flags: |
|
1237 | if fn in self._flags: | |
1238 | ret._flags[fn] = self._flags[fn] |
|
1238 | ret._flags[fn] = self._flags[fn] | |
1239 |
|
1239 | |||
1240 | visit = self._loadchildrensetlazy(visit) |
|
1240 | visit = self._loadchildrensetlazy(visit) | |
1241 | for dir, subm in pycompat.iteritems(self._dirs): |
|
1241 | for dir, subm in pycompat.iteritems(self._dirs): | |
1242 | if visit and dir[:-1] not in visit: |
|
1242 | if visit and dir[:-1] not in visit: | |
1243 | continue |
|
1243 | continue | |
1244 | m = subm._matches_inner(match) |
|
1244 | m = subm._matches_inner(match) | |
1245 | if not m._isempty(): |
|
1245 | if not m._isempty(): | |
1246 | ret._dirs[dir] = m |
|
1246 | ret._dirs[dir] = m | |
1247 |
|
1247 | |||
1248 | if not ret._isempty(): |
|
1248 | if not ret._isempty(): | |
1249 | ret._dirty = True |
|
1249 | ret._dirty = True | |
1250 | return ret |
|
1250 | return ret | |
1251 |
|
1251 | |||
1252 | def fastdelta(self, base, changes): |
|
1252 | def fastdelta(self, base, changes): | |
1253 | raise FastdeltaUnavailable() |
|
1253 | raise FastdeltaUnavailable() | |
1254 |
|
1254 | |||
1255 | def diff(self, m2, match=None, clean=False): |
|
1255 | def diff(self, m2, match=None, clean=False): | |
1256 | """Finds changes between the current manifest and m2. |
|
1256 | """Finds changes between the current manifest and m2. | |
1257 |
|
1257 | |||
1258 | Args: |
|
1258 | Args: | |
1259 | m2: the manifest to which this manifest should be compared. |
|
1259 | m2: the manifest to which this manifest should be compared. | |
1260 | clean: if true, include files unchanged between these manifests |
|
1260 | clean: if true, include files unchanged between these manifests | |
1261 | with a None value in the returned dictionary. |
|
1261 | with a None value in the returned dictionary. | |
1262 |
|
1262 | |||
1263 | The result is returned as a dict with filename as key and |
|
1263 | The result is returned as a dict with filename as key and | |
1264 | values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the |
|
1264 | values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the | |
1265 | nodeid in the current/other manifest and fl1/fl2 is the flag |
|
1265 | nodeid in the current/other manifest and fl1/fl2 is the flag | |
1266 | in the current/other manifest. Where the file does not exist, |
|
1266 | in the current/other manifest. Where the file does not exist, | |
1267 | the nodeid will be None and the flags will be the empty |
|
1267 | the nodeid will be None and the flags will be the empty | |
1268 | string. |
|
1268 | string. | |
1269 | """ |
|
1269 | """ | |
1270 | if match and not match.always(): |
|
1270 | if match and not match.always(): | |
1271 | m1 = self._matches(match) |
|
1271 | m1 = self._matches(match) | |
1272 | m2 = m2._matches(match) |
|
1272 | m2 = m2._matches(match) | |
1273 | return m1.diff(m2, clean=clean) |
|
1273 | return m1.diff(m2, clean=clean) | |
1274 | result = {} |
|
1274 | result = {} | |
1275 | emptytree = treemanifest() |
|
1275 | emptytree = treemanifest() | |
1276 |
|
1276 | |||
1277 | def _iterativediff(t1, t2, stack): |
|
1277 | def _iterativediff(t1, t2, stack): | |
1278 | """compares two tree manifests and append new tree-manifests which |
|
1278 | """compares two tree manifests and append new tree-manifests which | |
1279 | needs to be compared to stack""" |
|
1279 | needs to be compared to stack""" | |
1280 | if t1._node == t2._node and not t1._dirty and not t2._dirty: |
|
1280 | if t1._node == t2._node and not t1._dirty and not t2._dirty: | |
1281 | return |
|
1281 | return | |
1282 | t1._load() |
|
1282 | t1._load() | |
1283 | t2._load() |
|
1283 | t2._load() | |
1284 | self._loaddifflazy(t1, t2) |
|
1284 | self._loaddifflazy(t1, t2) | |
1285 |
|
1285 | |||
1286 | for d, m1 in pycompat.iteritems(t1._dirs): |
|
1286 | for d, m1 in pycompat.iteritems(t1._dirs): | |
1287 | m2 = t2._dirs.get(d, emptytree) |
|
1287 | m2 = t2._dirs.get(d, emptytree) | |
1288 | stack.append((m1, m2)) |
|
1288 | stack.append((m1, m2)) | |
1289 |
|
1289 | |||
1290 | for d, m2 in pycompat.iteritems(t2._dirs): |
|
1290 | for d, m2 in pycompat.iteritems(t2._dirs): | |
1291 | if d not in t1._dirs: |
|
1291 | if d not in t1._dirs: | |
1292 | stack.append((emptytree, m2)) |
|
1292 | stack.append((emptytree, m2)) | |
1293 |
|
1293 | |||
1294 | for fn, n1 in pycompat.iteritems(t1._files): |
|
1294 | for fn, n1 in pycompat.iteritems(t1._files): | |
1295 | fl1 = t1._flags.get(fn, b'') |
|
1295 | fl1 = t1._flags.get(fn, b'') | |
1296 | n2 = t2._files.get(fn, None) |
|
1296 | n2 = t2._files.get(fn, None) | |
1297 | fl2 = t2._flags.get(fn, b'') |
|
1297 | fl2 = t2._flags.get(fn, b'') | |
1298 | if n1 != n2 or fl1 != fl2: |
|
1298 | if n1 != n2 or fl1 != fl2: | |
1299 | result[t1._subpath(fn)] = ((n1, fl1), (n2, fl2)) |
|
1299 | result[t1._subpath(fn)] = ((n1, fl1), (n2, fl2)) | |
1300 | elif clean: |
|
1300 | elif clean: | |
1301 | result[t1._subpath(fn)] = None |
|
1301 | result[t1._subpath(fn)] = None | |
1302 |
|
1302 | |||
1303 | for fn, n2 in pycompat.iteritems(t2._files): |
|
1303 | for fn, n2 in pycompat.iteritems(t2._files): | |
1304 | if fn not in t1._files: |
|
1304 | if fn not in t1._files: | |
1305 | fl2 = t2._flags.get(fn, b'') |
|
1305 | fl2 = t2._flags.get(fn, b'') | |
1306 | result[t2._subpath(fn)] = ((None, b''), (n2, fl2)) |
|
1306 | result[t2._subpath(fn)] = ((None, b''), (n2, fl2)) | |
1307 |
|
1307 | |||
1308 | stackls = [] |
|
1308 | stackls = [] | |
1309 | _iterativediff(self, m2, stackls) |
|
1309 | _iterativediff(self, m2, stackls) | |
1310 | while stackls: |
|
1310 | while stackls: | |
1311 | t1, t2 = stackls.pop() |
|
1311 | t1, t2 = stackls.pop() | |
1312 | # stackls is populated in the function call |
|
1312 | # stackls is populated in the function call | |
1313 | _iterativediff(t1, t2, stackls) |
|
1313 | _iterativediff(t1, t2, stackls) | |
1314 | return result |
|
1314 | return result | |
1315 |
|
1315 | |||
1316 | def unmodifiedsince(self, m2): |
|
1316 | def unmodifiedsince(self, m2): | |
1317 | return not self._dirty and not m2._dirty and self._node == m2._node |
|
1317 | return not self._dirty and not m2._dirty and self._node == m2._node | |
1318 |
|
1318 | |||
1319 | def parse(self, text, readsubtree): |
|
1319 | def parse(self, text, readsubtree): | |
1320 | selflazy = self._lazydirs |
|
1320 | selflazy = self._lazydirs | |
1321 | for f, n, fl in _parse(text): |
|
1321 | for f, n, fl in _parse(text): | |
1322 | if fl == b't': |
|
1322 | if fl == b't': | |
1323 | f = f + b'/' |
|
1323 | f = f + b'/' | |
1324 | # False below means "doesn't need to be copied" and can use the |
|
1324 | # False below means "doesn't need to be copied" and can use the | |
1325 | # cached value from readsubtree directly. |
|
1325 | # cached value from readsubtree directly. | |
1326 | selflazy[f] = (n, readsubtree, False) |
|
1326 | selflazy[f] = (n, readsubtree, False) | |
1327 | elif b'/' in f: |
|
1327 | elif b'/' in f: | |
1328 | # This is a flat manifest, so use __setitem__ and setflag rather |
|
1328 | # This is a flat manifest, so use __setitem__ and setflag rather | |
1329 | # than assigning directly to _files and _flags, so we can |
|
1329 | # than assigning directly to _files and _flags, so we can | |
1330 | # assign a path in a subdirectory, and to mark dirty (compared |
|
1330 | # assign a path in a subdirectory, and to mark dirty (compared | |
1331 | # to nullid). |
|
1331 | # to nullid). | |
1332 | self[f] = n |
|
1332 | self[f] = n | |
1333 | if fl: |
|
1333 | if fl: | |
1334 | self.setflag(f, fl) |
|
1334 | self.setflag(f, fl) | |
1335 | else: |
|
1335 | else: | |
1336 | # Assigning to _files and _flags avoids marking as dirty, |
|
1336 | # Assigning to _files and _flags avoids marking as dirty, | |
1337 | # and should be a little faster. |
|
1337 | # and should be a little faster. | |
1338 | self._files[f] = n |
|
1338 | self._files[f] = n | |
1339 | if fl: |
|
1339 | if fl: | |
1340 | self._flags[f] = fl |
|
1340 | self._flags[f] = fl | |
1341 |
|
1341 | |||
1342 | def text(self): |
|
1342 | def text(self): | |
1343 | """Get the full data of this manifest as a bytestring.""" |
|
1343 | """Get the full data of this manifest as a bytestring.""" | |
1344 | self._load() |
|
1344 | self._load() | |
1345 | return _text(self.iterentries()) |
|
1345 | return _text(self.iterentries()) | |
1346 |
|
1346 | |||
1347 | def dirtext(self): |
|
1347 | def dirtext(self): | |
1348 | """Get the full data of this directory as a bytestring. Make sure that |
|
1348 | """Get the full data of this directory as a bytestring. Make sure that | |
1349 | any submanifests have been written first, so their nodeids are correct. |
|
1349 | any submanifests have been written first, so their nodeids are correct. | |
1350 | """ |
|
1350 | """ | |
1351 | self._load() |
|
1351 | self._load() | |
1352 | flags = self.flags |
|
1352 | flags = self.flags | |
1353 | lazydirs = [ |
|
1353 | lazydirs = [ | |
1354 | (d[:-1], v[0], b't') for d, v in pycompat.iteritems(self._lazydirs) |
|
1354 | (d[:-1], v[0], b't') for d, v in pycompat.iteritems(self._lazydirs) | |
1355 | ] |
|
1355 | ] | |
1356 | dirs = [(d[:-1], self._dirs[d]._node, b't') for d in self._dirs] |
|
1356 | dirs = [(d[:-1], self._dirs[d]._node, b't') for d in self._dirs] | |
1357 | files = [(f, self._files[f], flags(f)) for f in self._files] |
|
1357 | files = [(f, self._files[f], flags(f)) for f in self._files] | |
1358 | return _text(sorted(dirs + files + lazydirs)) |
|
1358 | return _text(sorted(dirs + files + lazydirs)) | |
1359 |
|
1359 | |||
1360 | def read(self, gettext, readsubtree): |
|
1360 | def read(self, gettext, readsubtree): | |
1361 | def _load_for_read(s): |
|
1361 | def _load_for_read(s): | |
1362 | s.parse(gettext(), readsubtree) |
|
1362 | s.parse(gettext(), readsubtree) | |
1363 | s._dirty = False |
|
1363 | s._dirty = False | |
1364 |
|
1364 | |||
1365 | self._loadfunc = _load_for_read |
|
1365 | self._loadfunc = _load_for_read | |
1366 |
|
1366 | |||
1367 | def writesubtrees(self, m1, m2, writesubtree, match): |
|
1367 | def writesubtrees(self, m1, m2, writesubtree, match): | |
1368 | self._load() # for consistency; should never have any effect here |
|
1368 | self._load() # for consistency; should never have any effect here | |
1369 | m1._load() |
|
1369 | m1._load() | |
1370 | m2._load() |
|
1370 | m2._load() | |
1371 | emptytree = treemanifest() |
|
1371 | emptytree = treemanifest() | |
1372 |
|
1372 | |||
1373 | def getnode(m, d): |
|
1373 | def getnode(m, d): | |
1374 | ld = m._lazydirs.get(d) |
|
1374 | ld = m._lazydirs.get(d) | |
1375 | if ld: |
|
1375 | if ld: | |
1376 | return ld[0] |
|
1376 | return ld[0] | |
1377 | return m._dirs.get(d, emptytree)._node |
|
1377 | return m._dirs.get(d, emptytree)._node | |
1378 |
|
1378 | |||
1379 | # let's skip investigating things that `match` says we do not need. |
|
1379 | # let's skip investigating things that `match` says we do not need. | |
1380 | visit = match.visitchildrenset(self._dir[:-1]) |
|
1380 | visit = match.visitchildrenset(self._dir[:-1]) | |
1381 | visit = self._loadchildrensetlazy(visit) |
|
1381 | visit = self._loadchildrensetlazy(visit) | |
1382 | if visit == b'this' or visit == b'all': |
|
1382 | if visit == b'this' or visit == b'all': | |
1383 | visit = None |
|
1383 | visit = None | |
1384 | for d, subm in pycompat.iteritems(self._dirs): |
|
1384 | for d, subm in pycompat.iteritems(self._dirs): | |
1385 | if visit and d[:-1] not in visit: |
|
1385 | if visit and d[:-1] not in visit: | |
1386 | continue |
|
1386 | continue | |
1387 | subp1 = getnode(m1, d) |
|
1387 | subp1 = getnode(m1, d) | |
1388 | subp2 = getnode(m2, d) |
|
1388 | subp2 = getnode(m2, d) | |
1389 | if subp1 == nullid: |
|
1389 | if subp1 == nullid: | |
1390 | subp1, subp2 = subp2, subp1 |
|
1390 | subp1, subp2 = subp2, subp1 | |
1391 | writesubtree(subm, subp1, subp2, match) |
|
1391 | writesubtree(subm, subp1, subp2, match) | |
1392 |
|
1392 | |||
1393 | def walksubtrees(self, matcher=None): |
|
1393 | def walksubtrees(self, matcher=None): | |
1394 | """Returns an iterator of the subtrees of this manifest, including this |
|
1394 | """Returns an iterator of the subtrees of this manifest, including this | |
1395 | manifest itself. |
|
1395 | manifest itself. | |
1396 |
|
1396 | |||
1397 | If `matcher` is provided, it only returns subtrees that match. |
|
1397 | If `matcher` is provided, it only returns subtrees that match. | |
1398 | """ |
|
1398 | """ | |
1399 | if matcher and not matcher.visitdir(self._dir[:-1]): |
|
1399 | if matcher and not matcher.visitdir(self._dir[:-1]): | |
1400 | return |
|
1400 | return | |
1401 | if not matcher or matcher(self._dir[:-1]): |
|
1401 | if not matcher or matcher(self._dir[:-1]): | |
1402 | yield self |
|
1402 | yield self | |
1403 |
|
1403 | |||
1404 | self._load() |
|
1404 | self._load() | |
1405 | # OPT: use visitchildrenset to avoid loading everything. |
|
1405 | # OPT: use visitchildrenset to avoid loading everything. | |
1406 | self._loadalllazy() |
|
1406 | self._loadalllazy() | |
1407 | for d, subm in pycompat.iteritems(self._dirs): |
|
1407 | for d, subm in pycompat.iteritems(self._dirs): | |
1408 | for subtree in subm.walksubtrees(matcher=matcher): |
|
1408 | for subtree in subm.walksubtrees(matcher=matcher): | |
1409 | yield subtree |
|
1409 | yield subtree | |
1410 |
|
1410 | |||
1411 |
|
1411 | |||
1412 | class manifestfulltextcache(util.lrucachedict): |
|
1412 | class manifestfulltextcache(util.lrucachedict): | |
1413 | """File-backed LRU cache for the manifest cache |
|
1413 | """File-backed LRU cache for the manifest cache | |
1414 |
|
1414 | |||
1415 | File consists of entries, up to EOF: |
|
1415 | File consists of entries, up to EOF: | |
1416 |
|
1416 | |||
1417 | - 20 bytes node, 4 bytes length, <length> manifest data |
|
1417 | - 20 bytes node, 4 bytes length, <length> manifest data | |
1418 |
|
1418 | |||
1419 | These are written in reverse cache order (oldest to newest). |
|
1419 | These are written in reverse cache order (oldest to newest). | |
1420 |
|
1420 | |||
1421 | """ |
|
1421 | """ | |
1422 |
|
1422 | |||
1423 | _file = b'manifestfulltextcache' |
|
1423 | _file = b'manifestfulltextcache' | |
1424 |
|
1424 | |||
1425 | def __init__(self, max): |
|
1425 | def __init__(self, max): | |
1426 | super(manifestfulltextcache, self).__init__(max) |
|
1426 | super(manifestfulltextcache, self).__init__(max) | |
1427 | self._dirty = False |
|
1427 | self._dirty = False | |
1428 | self._read = False |
|
1428 | self._read = False | |
1429 | self._opener = None |
|
1429 | self._opener = None | |
1430 |
|
1430 | |||
1431 | def read(self): |
|
1431 | def read(self): | |
1432 | if self._read or self._opener is None: |
|
1432 | if self._read or self._opener is None: | |
1433 | return |
|
1433 | return | |
1434 |
|
1434 | |||
1435 | try: |
|
1435 | try: | |
1436 | with self._opener(self._file) as fp: |
|
1436 | with self._opener(self._file) as fp: | |
1437 | set = super(manifestfulltextcache, self).__setitem__ |
|
1437 | set = super(manifestfulltextcache, self).__setitem__ | |
1438 | # ignore trailing data, this is a cache, corruption is skipped |
|
1438 | # ignore trailing data, this is a cache, corruption is skipped | |
1439 | while True: |
|
1439 | while True: | |
1440 | # TODO do we need to do work here for sha1 portability? |
|
1440 | # TODO do we need to do work here for sha1 portability? | |
1441 | node = fp.read(20) |
|
1441 | node = fp.read(20) | |
1442 | if len(node) < 20: |
|
1442 | if len(node) < 20: | |
1443 | break |
|
1443 | break | |
1444 | try: |
|
1444 | try: | |
1445 | size = struct.unpack(b'>L', fp.read(4))[0] |
|
1445 | size = struct.unpack(b'>L', fp.read(4))[0] | |
1446 | except struct.error: |
|
1446 | except struct.error: | |
1447 | break |
|
1447 | break | |
1448 | value = bytearray(fp.read(size)) |
|
1448 | value = bytearray(fp.read(size)) | |
1449 | if len(value) != size: |
|
1449 | if len(value) != size: | |
1450 | break |
|
1450 | break | |
1451 | set(node, value) |
|
1451 | set(node, value) | |
1452 | except IOError: |
|
1452 | except IOError: | |
1453 | # the file is allowed to be missing |
|
1453 | # the file is allowed to be missing | |
1454 | pass |
|
1454 | pass | |
1455 |
|
1455 | |||
1456 | self._read = True |
|
1456 | self._read = True | |
1457 | self._dirty = False |
|
1457 | self._dirty = False | |
1458 |
|
1458 | |||
1459 | def write(self): |
|
1459 | def write(self): | |
1460 | if not self._dirty or self._opener is None: |
|
1460 | if not self._dirty or self._opener is None: | |
1461 | return |
|
1461 | return | |
1462 | # rotate backwards to the first used node |
|
1462 | # rotate backwards to the first used node | |
1463 | try: |
|
1463 | try: | |
1464 | with self._opener( |
|
1464 | with self._opener( | |
1465 | self._file, b'w', atomictemp=True, checkambig=True |
|
1465 | self._file, b'w', atomictemp=True, checkambig=True | |
1466 | ) as fp: |
|
1466 | ) as fp: | |
1467 | node = self._head.prev |
|
1467 | node = self._head.prev | |
1468 | while True: |
|
1468 | while True: | |
1469 | if node.key in self._cache: |
|
1469 | if node.key in self._cache: | |
1470 | fp.write(node.key) |
|
1470 | fp.write(node.key) | |
1471 | fp.write(struct.pack(b'>L', len(node.value))) |
|
1471 | fp.write(struct.pack(b'>L', len(node.value))) | |
1472 | fp.write(node.value) |
|
1472 | fp.write(node.value) | |
1473 | if node is self._head: |
|
1473 | if node is self._head: | |
1474 | break |
|
1474 | break | |
1475 | node = node.prev |
|
1475 | node = node.prev | |
1476 | except IOError: |
|
1476 | except IOError: | |
1477 | # We could not write the cache (eg: permission error) |
|
1477 | # We could not write the cache (eg: permission error) | |
1478 | # the content can be missing. |
|
1478 | # the content can be missing. | |
1479 | # |
|
1479 | # | |
1480 | # We could try harder and see if we could recreate a wcache |
|
1480 | # We could try harder and see if we could recreate a wcache | |
1481 | # directory were we coudl write too. |
|
1481 | # directory were we coudl write too. | |
1482 | # |
|
1482 | # | |
1483 | # XXX the error pass silently, having some way to issue an error |
|
1483 | # XXX the error pass silently, having some way to issue an error | |
1484 | # log `ui.log` would be nice. |
|
1484 | # log `ui.log` would be nice. | |
1485 | pass |
|
1485 | pass | |
1486 |
|
1486 | |||
1487 | def __len__(self): |
|
1487 | def __len__(self): | |
1488 | if not self._read: |
|
1488 | if not self._read: | |
1489 | self.read() |
|
1489 | self.read() | |
1490 | return super(manifestfulltextcache, self).__len__() |
|
1490 | return super(manifestfulltextcache, self).__len__() | |
1491 |
|
1491 | |||
1492 | def __contains__(self, k): |
|
1492 | def __contains__(self, k): | |
1493 | if not self._read: |
|
1493 | if not self._read: | |
1494 | self.read() |
|
1494 | self.read() | |
1495 | return super(manifestfulltextcache, self).__contains__(k) |
|
1495 | return super(manifestfulltextcache, self).__contains__(k) | |
1496 |
|
1496 | |||
1497 | def __iter__(self): |
|
1497 | def __iter__(self): | |
1498 | if not self._read: |
|
1498 | if not self._read: | |
1499 | self.read() |
|
1499 | self.read() | |
1500 | return super(manifestfulltextcache, self).__iter__() |
|
1500 | return super(manifestfulltextcache, self).__iter__() | |
1501 |
|
1501 | |||
1502 | def __getitem__(self, k): |
|
1502 | def __getitem__(self, k): | |
1503 | if not self._read: |
|
1503 | if not self._read: | |
1504 | self.read() |
|
1504 | self.read() | |
1505 | # the cache lru order can change on read |
|
1505 | # the cache lru order can change on read | |
1506 | setdirty = self._cache.get(k) is not self._head |
|
1506 | setdirty = self._cache.get(k) is not self._head | |
1507 | value = super(manifestfulltextcache, self).__getitem__(k) |
|
1507 | value = super(manifestfulltextcache, self).__getitem__(k) | |
1508 | if setdirty: |
|
1508 | if setdirty: | |
1509 | self._dirty = True |
|
1509 | self._dirty = True | |
1510 | return value |
|
1510 | return value | |
1511 |
|
1511 | |||
1512 | def __setitem__(self, k, v): |
|
1512 | def __setitem__(self, k, v): | |
1513 | if not self._read: |
|
1513 | if not self._read: | |
1514 | self.read() |
|
1514 | self.read() | |
1515 | super(manifestfulltextcache, self).__setitem__(k, v) |
|
1515 | super(manifestfulltextcache, self).__setitem__(k, v) | |
1516 | self._dirty = True |
|
1516 | self._dirty = True | |
1517 |
|
1517 | |||
1518 | def __delitem__(self, k): |
|
1518 | def __delitem__(self, k): | |
1519 | if not self._read: |
|
1519 | if not self._read: | |
1520 | self.read() |
|
1520 | self.read() | |
1521 | super(manifestfulltextcache, self).__delitem__(k) |
|
1521 | super(manifestfulltextcache, self).__delitem__(k) | |
1522 | self._dirty = True |
|
1522 | self._dirty = True | |
1523 |
|
1523 | |||
1524 | def get(self, k, default=None): |
|
1524 | def get(self, k, default=None): | |
1525 | if not self._read: |
|
1525 | if not self._read: | |
1526 | self.read() |
|
1526 | self.read() | |
1527 | return super(manifestfulltextcache, self).get(k, default=default) |
|
1527 | return super(manifestfulltextcache, self).get(k, default=default) | |
1528 |
|
1528 | |||
1529 | def clear(self, clear_persisted_data=False): |
|
1529 | def clear(self, clear_persisted_data=False): | |
1530 | super(manifestfulltextcache, self).clear() |
|
1530 | super(manifestfulltextcache, self).clear() | |
1531 | if clear_persisted_data: |
|
1531 | if clear_persisted_data: | |
1532 | self._dirty = True |
|
1532 | self._dirty = True | |
1533 | self.write() |
|
1533 | self.write() | |
1534 | self._read = False |
|
1534 | self._read = False | |
1535 |
|
1535 | |||
1536 |
|
1536 | |||
1537 | # and upper bound of what we expect from compression |
|
1537 | # and upper bound of what we expect from compression | |
1538 | # (real live value seems to be "3") |
|
1538 | # (real live value seems to be "3") | |
1539 | MAXCOMPRESSION = 3 |
|
1539 | MAXCOMPRESSION = 3 | |
1540 |
|
1540 | |||
1541 |
|
1541 | |||
1542 | class FastdeltaUnavailable(Exception): |
|
1542 | class FastdeltaUnavailable(Exception): | |
1543 | """Exception raised when fastdelta isn't usable on a manifest.""" |
|
1543 | """Exception raised when fastdelta isn't usable on a manifest.""" | |
1544 |
|
1544 | |||
1545 |
|
1545 | |||
1546 | @interfaceutil.implementer(repository.imanifeststorage) |
|
1546 | @interfaceutil.implementer(repository.imanifeststorage) | |
1547 | class manifestrevlog(object): |
|
1547 | class manifestrevlog(object): | |
1548 | """A revlog that stores manifest texts. This is responsible for caching the |
|
1548 | """A revlog that stores manifest texts. This is responsible for caching the | |
1549 | full-text manifest contents. |
|
1549 | full-text manifest contents. | |
1550 | """ |
|
1550 | """ | |
1551 |
|
1551 | |||
1552 | def __init__( |
|
1552 | def __init__( | |
1553 | self, |
|
1553 | self, | |
1554 | opener, |
|
1554 | opener, | |
1555 | tree=b'', |
|
1555 | tree=b'', | |
1556 | dirlogcache=None, |
|
1556 | dirlogcache=None, | |
1557 | indexfile=None, |
|
1557 | indexfile=None, | |
1558 | treemanifest=False, |
|
1558 | treemanifest=False, | |
1559 | ): |
|
1559 | ): | |
1560 | """Constructs a new manifest revlog |
|
1560 | """Constructs a new manifest revlog | |
1561 |
|
1561 | |||
1562 | `indexfile` - used by extensions to have two manifests at once, like |
|
1562 | `indexfile` - used by extensions to have two manifests at once, like | |
1563 | when transitioning between flatmanifeset and treemanifests. |
|
1563 | when transitioning between flatmanifeset and treemanifests. | |
1564 |
|
1564 | |||
1565 | `treemanifest` - used to indicate this is a tree manifest revlog. Opener |
|
1565 | `treemanifest` - used to indicate this is a tree manifest revlog. Opener | |
1566 | options can also be used to make this a tree manifest revlog. The opener |
|
1566 | options can also be used to make this a tree manifest revlog. The opener | |
1567 | option takes precedence, so if it is set to True, we ignore whatever |
|
1567 | option takes precedence, so if it is set to True, we ignore whatever | |
1568 | value is passed in to the constructor. |
|
1568 | value is passed in to the constructor. | |
1569 | """ |
|
1569 | """ | |
1570 | # During normal operations, we expect to deal with not more than four |
|
1570 | # During normal operations, we expect to deal with not more than four | |
1571 | # revs at a time (such as during commit --amend). When rebasing large |
|
1571 | # revs at a time (such as during commit --amend). When rebasing large | |
1572 | # stacks of commits, the number can go up, hence the config knob below. |
|
1572 | # stacks of commits, the number can go up, hence the config knob below. | |
1573 | cachesize = 4 |
|
1573 | cachesize = 4 | |
1574 | optiontreemanifest = False |
|
1574 | optiontreemanifest = False | |
1575 | opts = getattr(opener, 'options', None) |
|
1575 | opts = getattr(opener, 'options', None) | |
1576 | if opts is not None: |
|
1576 | if opts is not None: | |
1577 | cachesize = opts.get(b'manifestcachesize', cachesize) |
|
1577 | cachesize = opts.get(b'manifestcachesize', cachesize) | |
1578 | optiontreemanifest = opts.get(b'treemanifest', False) |
|
1578 | optiontreemanifest = opts.get(b'treemanifest', False) | |
1579 |
|
1579 | |||
1580 | self._treeondisk = optiontreemanifest or treemanifest |
|
1580 | self._treeondisk = optiontreemanifest or treemanifest | |
1581 |
|
1581 | |||
1582 | self._fulltextcache = manifestfulltextcache(cachesize) |
|
1582 | self._fulltextcache = manifestfulltextcache(cachesize) | |
1583 |
|
1583 | |||
1584 | if tree: |
|
1584 | if tree: | |
1585 | assert self._treeondisk, b'opts is %r' % opts |
|
1585 | assert self._treeondisk, b'opts is %r' % opts | |
1586 |
|
1586 | |||
1587 | if indexfile is None: |
|
1587 | if indexfile is None: | |
1588 | indexfile = b'00manifest.i' |
|
1588 | indexfile = b'00manifest.i' | |
1589 | if tree: |
|
1589 | if tree: | |
1590 | indexfile = b"meta/" + tree + indexfile |
|
1590 | indexfile = b"meta/" + tree + indexfile | |
1591 |
|
1591 | |||
1592 | self.tree = tree |
|
1592 | self.tree = tree | |
1593 |
|
1593 | |||
1594 | # The dirlogcache is kept on the root manifest log |
|
1594 | # The dirlogcache is kept on the root manifest log | |
1595 | if tree: |
|
1595 | if tree: | |
1596 | self._dirlogcache = dirlogcache |
|
1596 | self._dirlogcache = dirlogcache | |
1597 | else: |
|
1597 | else: | |
1598 | self._dirlogcache = {b'': self} |
|
1598 | self._dirlogcache = {b'': self} | |
1599 |
|
1599 | |||
1600 | self._revlog = revlog.revlog( |
|
1600 | self._revlog = revlog.revlog( | |
1601 | opener, |
|
1601 | opener, | |
1602 | indexfile, |
|
1602 | indexfile, | |
1603 | # only root indexfile is cached |
|
1603 | # only root indexfile is cached | |
1604 | checkambig=not bool(tree), |
|
1604 | checkambig=not bool(tree), | |
1605 | mmaplargeindex=True, |
|
1605 | mmaplargeindex=True, | |
1606 | upperboundcomp=MAXCOMPRESSION, |
|
1606 | upperboundcomp=MAXCOMPRESSION, | |
1607 | persistentnodemap=opener.options.get(b'persistent-nodemap', False), |
|
1607 | persistentnodemap=opener.options.get(b'persistent-nodemap', False), | |
1608 | ) |
|
1608 | ) | |
1609 |
|
1609 | |||
1610 | self.index = self._revlog.index |
|
1610 | self.index = self._revlog.index | |
1611 | self.version = self._revlog.version |
|
1611 | self.version = self._revlog.version | |
1612 | self._generaldelta = self._revlog._generaldelta |
|
1612 | self._generaldelta = self._revlog._generaldelta | |
1613 | self._revlog.revlog_kind = b'manifest' |
|
1613 | self._revlog.revlog_kind = b'manifest' | |
1614 |
|
1614 | |||
1615 | def _setupmanifestcachehooks(self, repo): |
|
1615 | def _setupmanifestcachehooks(self, repo): | |
1616 | """Persist the manifestfulltextcache on lock release""" |
|
1616 | """Persist the manifestfulltextcache on lock release""" | |
1617 | if not util.safehasattr(repo, b'_wlockref'): |
|
1617 | if not util.safehasattr(repo, b'_wlockref'): | |
1618 | return |
|
1618 | return | |
1619 |
|
1619 | |||
1620 | self._fulltextcache._opener = repo.wcachevfs |
|
1620 | self._fulltextcache._opener = repo.wcachevfs | |
1621 | if repo._currentlock(repo._wlockref) is None: |
|
1621 | if repo._currentlock(repo._wlockref) is None: | |
1622 | return |
|
1622 | return | |
1623 |
|
1623 | |||
1624 | reporef = weakref.ref(repo) |
|
1624 | reporef = weakref.ref(repo) | |
1625 | manifestrevlogref = weakref.ref(self) |
|
1625 | manifestrevlogref = weakref.ref(self) | |
1626 |
|
1626 | |||
1627 | def persistmanifestcache(success): |
|
1627 | def persistmanifestcache(success): | |
1628 | # Repo is in an unknown state, do not persist. |
|
1628 | # Repo is in an unknown state, do not persist. | |
1629 | if not success: |
|
1629 | if not success: | |
1630 | return |
|
1630 | return | |
1631 |
|
1631 | |||
1632 | repo = reporef() |
|
1632 | repo = reporef() | |
1633 | self = manifestrevlogref() |
|
1633 | self = manifestrevlogref() | |
1634 | if repo is None or self is None: |
|
1634 | if repo is None or self is None: | |
1635 | return |
|
1635 | return | |
1636 | if repo.manifestlog.getstorage(b'') is not self: |
|
1636 | if repo.manifestlog.getstorage(b'') is not self: | |
1637 | # there's a different manifest in play now, abort |
|
1637 | # there's a different manifest in play now, abort | |
1638 | return |
|
1638 | return | |
1639 | self._fulltextcache.write() |
|
1639 | self._fulltextcache.write() | |
1640 |
|
1640 | |||
1641 | repo._afterlock(persistmanifestcache) |
|
1641 | repo._afterlock(persistmanifestcache) | |
1642 |
|
1642 | |||
1643 | @property |
|
1643 | @property | |
1644 | def fulltextcache(self): |
|
1644 | def fulltextcache(self): | |
1645 | return self._fulltextcache |
|
1645 | return self._fulltextcache | |
1646 |
|
1646 | |||
1647 | def clearcaches(self, clear_persisted_data=False): |
|
1647 | def clearcaches(self, clear_persisted_data=False): | |
1648 | self._revlog.clearcaches() |
|
1648 | self._revlog.clearcaches() | |
1649 | self._fulltextcache.clear(clear_persisted_data=clear_persisted_data) |
|
1649 | self._fulltextcache.clear(clear_persisted_data=clear_persisted_data) | |
1650 | self._dirlogcache = {self.tree: self} |
|
1650 | self._dirlogcache = {self.tree: self} | |
1651 |
|
1651 | |||
1652 | def dirlog(self, d): |
|
1652 | def dirlog(self, d): | |
1653 | if d: |
|
1653 | if d: | |
1654 | assert self._treeondisk |
|
1654 | assert self._treeondisk | |
1655 | if d not in self._dirlogcache: |
|
1655 | if d not in self._dirlogcache: | |
1656 | mfrevlog = manifestrevlog( |
|
1656 | mfrevlog = manifestrevlog( | |
1657 | self.opener, d, self._dirlogcache, treemanifest=self._treeondisk |
|
1657 | self.opener, d, self._dirlogcache, treemanifest=self._treeondisk | |
1658 | ) |
|
1658 | ) | |
1659 | self._dirlogcache[d] = mfrevlog |
|
1659 | self._dirlogcache[d] = mfrevlog | |
1660 | return self._dirlogcache[d] |
|
1660 | return self._dirlogcache[d] | |
1661 |
|
1661 | |||
1662 | def add( |
|
1662 | def add( | |
1663 | self, |
|
1663 | self, | |
1664 | m, |
|
1664 | m, | |
1665 | transaction, |
|
1665 | transaction, | |
1666 | link, |
|
1666 | link, | |
1667 | p1, |
|
1667 | p1, | |
1668 | p2, |
|
1668 | p2, | |
1669 | added, |
|
1669 | added, | |
1670 | removed, |
|
1670 | removed, | |
1671 | readtree=None, |
|
1671 | readtree=None, | |
1672 | match=None, |
|
1672 | match=None, | |
1673 | ): |
|
1673 | ): | |
1674 | """add some manifest entry in to the manifest log |
|
1674 | """add some manifest entry in to the manifest log | |
1675 |
|
1675 | |||
1676 | input: |
|
1676 | input: | |
1677 |
|
1677 | |||
1678 | m: the manifest dict we want to store |
|
1678 | m: the manifest dict we want to store | |
1679 | transaction: the open transaction |
|
1679 | transaction: the open transaction | |
1680 | p1: manifest-node of p1 |
|
1680 | p1: manifest-node of p1 | |
1681 | p2: manifest-node of p2 |
|
1681 | p2: manifest-node of p2 | |
1682 | added: file added/changed compared to parent |
|
1682 | added: file added/changed compared to parent | |
1683 | removed: file removed compared to parent |
|
1683 | removed: file removed compared to parent | |
1684 |
|
1684 | |||
1685 | tree manifest input: |
|
1685 | tree manifest input: | |
1686 |
|
1686 | |||
1687 | readtree: a function to read a subtree |
|
1687 | readtree: a function to read a subtree | |
1688 | match: a filematcher for the subpart of the tree manifest |
|
1688 | match: a filematcher for the subpart of the tree manifest | |
1689 | """ |
|
1689 | """ | |
1690 | try: |
|
1690 | try: | |
1691 | if p1 not in self.fulltextcache: |
|
1691 | if p1 not in self.fulltextcache: | |
1692 | raise FastdeltaUnavailable() |
|
1692 | raise FastdeltaUnavailable() | |
1693 | # If our first parent is in the manifest cache, we can |
|
1693 | # If our first parent is in the manifest cache, we can | |
1694 | # compute a delta here using properties we know about the |
|
1694 | # compute a delta here using properties we know about the | |
1695 | # manifest up-front, which may save time later for the |
|
1695 | # manifest up-front, which may save time later for the | |
1696 | # revlog layer. |
|
1696 | # revlog layer. | |
1697 |
|
1697 | |||
1698 | _checkforbidden(added) |
|
1698 | _checkforbidden(added) | |
1699 | # combine the changed lists into one sorted iterator |
|
1699 | # combine the changed lists into one sorted iterator | |
1700 | work = heapq.merge( |
|
1700 | work = heapq.merge( | |
1701 | [(x, False) for x in sorted(added)], |
|
1701 | [(x, False) for x in sorted(added)], | |
1702 | [(x, True) for x in sorted(removed)], |
|
1702 | [(x, True) for x in sorted(removed)], | |
1703 | ) |
|
1703 | ) | |
1704 |
|
1704 | |||
1705 | arraytext, deltatext = m.fastdelta(self.fulltextcache[p1], work) |
|
1705 | arraytext, deltatext = m.fastdelta(self.fulltextcache[p1], work) | |
1706 | cachedelta = self._revlog.rev(p1), deltatext |
|
1706 | cachedelta = self._revlog.rev(p1), deltatext | |
1707 | text = util.buffer(arraytext) |
|
1707 | text = util.buffer(arraytext) | |
1708 | rev = self._revlog.addrevision( |
|
1708 | rev = self._revlog.addrevision( | |
1709 | text, transaction, link, p1, p2, cachedelta |
|
1709 | text, transaction, link, p1, p2, cachedelta | |
1710 | ) |
|
1710 | ) | |
1711 | n = self._revlog.node(rev) |
|
1711 | n = self._revlog.node(rev) | |
1712 | except FastdeltaUnavailable: |
|
1712 | except FastdeltaUnavailable: | |
1713 | # The first parent manifest isn't already loaded or the |
|
1713 | # The first parent manifest isn't already loaded or the | |
1714 | # manifest implementation doesn't support fastdelta, so |
|
1714 | # manifest implementation doesn't support fastdelta, so | |
1715 | # we'll just encode a fulltext of the manifest and pass |
|
1715 | # we'll just encode a fulltext of the manifest and pass | |
1716 | # that through to the revlog layer, and let it handle the |
|
1716 | # that through to the revlog layer, and let it handle the | |
1717 | # delta process. |
|
1717 | # delta process. | |
1718 | if self._treeondisk: |
|
1718 | if self._treeondisk: | |
1719 | assert readtree, b"readtree must be set for treemanifest writes" |
|
1719 | assert readtree, b"readtree must be set for treemanifest writes" | |
1720 | assert match, b"match must be specified for treemanifest writes" |
|
1720 | assert match, b"match must be specified for treemanifest writes" | |
1721 | m1 = readtree(self.tree, p1) |
|
1721 | m1 = readtree(self.tree, p1) | |
1722 | m2 = readtree(self.tree, p2) |
|
1722 | m2 = readtree(self.tree, p2) | |
1723 | n = self._addtree( |
|
1723 | n = self._addtree( | |
1724 | m, transaction, link, m1, m2, readtree, match=match |
|
1724 | m, transaction, link, m1, m2, readtree, match=match | |
1725 | ) |
|
1725 | ) | |
1726 | arraytext = None |
|
1726 | arraytext = None | |
1727 | else: |
|
1727 | else: | |
1728 | text = m.text() |
|
1728 | text = m.text() | |
1729 | rev = self._revlog.addrevision(text, transaction, link, p1, p2) |
|
1729 | rev = self._revlog.addrevision(text, transaction, link, p1, p2) | |
1730 | n = self._revlog.node(rev) |
|
1730 | n = self._revlog.node(rev) | |
1731 | arraytext = bytearray(text) |
|
1731 | arraytext = bytearray(text) | |
1732 |
|
1732 | |||
1733 | if arraytext is not None: |
|
1733 | if arraytext is not None: | |
1734 | self.fulltextcache[n] = arraytext |
|
1734 | self.fulltextcache[n] = arraytext | |
1735 |
|
1735 | |||
1736 | return n |
|
1736 | return n | |
1737 |
|
1737 | |||
1738 | def _addtree(self, m, transaction, link, m1, m2, readtree, match): |
|
1738 | def _addtree(self, m, transaction, link, m1, m2, readtree, match): | |
1739 | # If the manifest is unchanged compared to one parent, |
|
1739 | # If the manifest is unchanged compared to one parent, | |
1740 | # don't write a new revision |
|
1740 | # don't write a new revision | |
1741 | if self.tree != b'' and ( |
|
1741 | if self.tree != b'' and ( | |
1742 | m.unmodifiedsince(m1) or m.unmodifiedsince(m2) |
|
1742 | m.unmodifiedsince(m1) or m.unmodifiedsince(m2) | |
1743 | ): |
|
1743 | ): | |
1744 | return m.node() |
|
1744 | return m.node() | |
1745 |
|
1745 | |||
1746 | def writesubtree(subm, subp1, subp2, match): |
|
1746 | def writesubtree(subm, subp1, subp2, match): | |
1747 | sublog = self.dirlog(subm.dir()) |
|
1747 | sublog = self.dirlog(subm.dir()) | |
1748 | sublog.add( |
|
1748 | sublog.add( | |
1749 | subm, |
|
1749 | subm, | |
1750 | transaction, |
|
1750 | transaction, | |
1751 | link, |
|
1751 | link, | |
1752 | subp1, |
|
1752 | subp1, | |
1753 | subp2, |
|
1753 | subp2, | |
1754 | None, |
|
1754 | None, | |
1755 | None, |
|
1755 | None, | |
1756 | readtree=readtree, |
|
1756 | readtree=readtree, | |
1757 | match=match, |
|
1757 | match=match, | |
1758 | ) |
|
1758 | ) | |
1759 |
|
1759 | |||
1760 | m.writesubtrees(m1, m2, writesubtree, match) |
|
1760 | m.writesubtrees(m1, m2, writesubtree, match) | |
1761 | text = m.dirtext() |
|
1761 | text = m.dirtext() | |
1762 | n = None |
|
1762 | n = None | |
1763 | if self.tree != b'': |
|
1763 | if self.tree != b'': | |
1764 | # Double-check whether contents are unchanged to one parent |
|
1764 | # Double-check whether contents are unchanged to one parent | |
1765 | if text == m1.dirtext(): |
|
1765 | if text == m1.dirtext(): | |
1766 | n = m1.node() |
|
1766 | n = m1.node() | |
1767 | elif text == m2.dirtext(): |
|
1767 | elif text == m2.dirtext(): | |
1768 | n = m2.node() |
|
1768 | n = m2.node() | |
1769 |
|
1769 | |||
1770 | if not n: |
|
1770 | if not n: | |
1771 | rev = self._revlog.addrevision( |
|
1771 | rev = self._revlog.addrevision( | |
1772 | text, transaction, link, m1.node(), m2.node() |
|
1772 | text, transaction, link, m1.node(), m2.node() | |
1773 | ) |
|
1773 | ) | |
1774 | n = self._revlog.node(rev) |
|
1774 | n = self._revlog.node(rev) | |
1775 |
|
1775 | |||
1776 | # Save nodeid so parent manifest can calculate its nodeid |
|
1776 | # Save nodeid so parent manifest can calculate its nodeid | |
1777 | m.setnode(n) |
|
1777 | m.setnode(n) | |
1778 | return n |
|
1778 | return n | |
1779 |
|
1779 | |||
1780 | def __len__(self): |
|
1780 | def __len__(self): | |
1781 | return len(self._revlog) |
|
1781 | return len(self._revlog) | |
1782 |
|
1782 | |||
1783 | def __iter__(self): |
|
1783 | def __iter__(self): | |
1784 | return self._revlog.__iter__() |
|
1784 | return self._revlog.__iter__() | |
1785 |
|
1785 | |||
1786 | def rev(self, node): |
|
1786 | def rev(self, node): | |
1787 | return self._revlog.rev(node) |
|
1787 | return self._revlog.rev(node) | |
1788 |
|
1788 | |||
1789 | def node(self, rev): |
|
1789 | def node(self, rev): | |
1790 | return self._revlog.node(rev) |
|
1790 | return self._revlog.node(rev) | |
1791 |
|
1791 | |||
1792 | def lookup(self, value): |
|
1792 | def lookup(self, value): | |
1793 | return self._revlog.lookup(value) |
|
1793 | return self._revlog.lookup(value) | |
1794 |
|
1794 | |||
1795 | def parentrevs(self, rev): |
|
1795 | def parentrevs(self, rev): | |
1796 | return self._revlog.parentrevs(rev) |
|
1796 | return self._revlog.parentrevs(rev) | |
1797 |
|
1797 | |||
1798 | def parents(self, node): |
|
1798 | def parents(self, node): | |
1799 | return self._revlog.parents(node) |
|
1799 | return self._revlog.parents(node) | |
1800 |
|
1800 | |||
1801 | def linkrev(self, rev): |
|
1801 | def linkrev(self, rev): | |
1802 | return self._revlog.linkrev(rev) |
|
1802 | return self._revlog.linkrev(rev) | |
1803 |
|
1803 | |||
1804 | def checksize(self): |
|
1804 | def checksize(self): | |
1805 | return self._revlog.checksize() |
|
1805 | return self._revlog.checksize() | |
1806 |
|
1806 | |||
1807 | def revision(self, node, _df=None, raw=False): |
|
1807 | def revision(self, node, _df=None, raw=False): | |
1808 | return self._revlog.revision(node, _df=_df, raw=raw) |
|
1808 | return self._revlog.revision(node, _df=_df, raw=raw) | |
1809 |
|
1809 | |||
1810 | def rawdata(self, node, _df=None): |
|
1810 | def rawdata(self, node, _df=None): | |
1811 | return self._revlog.rawdata(node, _df=_df) |
|
1811 | return self._revlog.rawdata(node, _df=_df) | |
1812 |
|
1812 | |||
1813 | def revdiff(self, rev1, rev2): |
|
1813 | def revdiff(self, rev1, rev2): | |
1814 | return self._revlog.revdiff(rev1, rev2) |
|
1814 | return self._revlog.revdiff(rev1, rev2) | |
1815 |
|
1815 | |||
1816 | def cmp(self, node, text): |
|
1816 | def cmp(self, node, text): | |
1817 | return self._revlog.cmp(node, text) |
|
1817 | return self._revlog.cmp(node, text) | |
1818 |
|
1818 | |||
1819 | def deltaparent(self, rev): |
|
1819 | def deltaparent(self, rev): | |
1820 | return self._revlog.deltaparent(rev) |
|
1820 | return self._revlog.deltaparent(rev) | |
1821 |
|
1821 | |||
1822 | def emitrevisions( |
|
1822 | def emitrevisions( | |
1823 | self, |
|
1823 | self, | |
1824 | nodes, |
|
1824 | nodes, | |
1825 | nodesorder=None, |
|
1825 | nodesorder=None, | |
1826 | revisiondata=False, |
|
1826 | revisiondata=False, | |
1827 | assumehaveparentrevisions=False, |
|
1827 | assumehaveparentrevisions=False, | |
1828 | deltamode=repository.CG_DELTAMODE_STD, |
|
1828 | deltamode=repository.CG_DELTAMODE_STD, | |
|
1829 | sidedata_helpers=None, | |||
1829 | ): |
|
1830 | ): | |
1830 | return self._revlog.emitrevisions( |
|
1831 | return self._revlog.emitrevisions( | |
1831 | nodes, |
|
1832 | nodes, | |
1832 | nodesorder=nodesorder, |
|
1833 | nodesorder=nodesorder, | |
1833 | revisiondata=revisiondata, |
|
1834 | revisiondata=revisiondata, | |
1834 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
1835 | assumehaveparentrevisions=assumehaveparentrevisions, | |
1835 | deltamode=deltamode, |
|
1836 | deltamode=deltamode, | |
|
1837 | sidedata_helpers=sidedata_helpers, | |||
1836 | ) |
|
1838 | ) | |
1837 |
|
1839 | |||
1838 | def addgroup( |
|
1840 | def addgroup( | |
1839 | self, |
|
1841 | self, | |
1840 | deltas, |
|
1842 | deltas, | |
1841 | linkmapper, |
|
1843 | linkmapper, | |
1842 | transaction, |
|
1844 | transaction, | |
1843 | alwayscache=False, |
|
1845 | alwayscache=False, | |
1844 | addrevisioncb=None, |
|
1846 | addrevisioncb=None, | |
1845 | duplicaterevisioncb=None, |
|
1847 | duplicaterevisioncb=None, | |
1846 | ): |
|
1848 | ): | |
1847 | return self._revlog.addgroup( |
|
1849 | return self._revlog.addgroup( | |
1848 | deltas, |
|
1850 | deltas, | |
1849 | linkmapper, |
|
1851 | linkmapper, | |
1850 | transaction, |
|
1852 | transaction, | |
1851 | alwayscache=alwayscache, |
|
1853 | alwayscache=alwayscache, | |
1852 | addrevisioncb=addrevisioncb, |
|
1854 | addrevisioncb=addrevisioncb, | |
1853 | duplicaterevisioncb=duplicaterevisioncb, |
|
1855 | duplicaterevisioncb=duplicaterevisioncb, | |
1854 | ) |
|
1856 | ) | |
1855 |
|
1857 | |||
1856 | def rawsize(self, rev): |
|
1858 | def rawsize(self, rev): | |
1857 | return self._revlog.rawsize(rev) |
|
1859 | return self._revlog.rawsize(rev) | |
1858 |
|
1860 | |||
1859 | def getstrippoint(self, minlink): |
|
1861 | def getstrippoint(self, minlink): | |
1860 | return self._revlog.getstrippoint(minlink) |
|
1862 | return self._revlog.getstrippoint(minlink) | |
1861 |
|
1863 | |||
1862 | def strip(self, minlink, transaction): |
|
1864 | def strip(self, minlink, transaction): | |
1863 | return self._revlog.strip(minlink, transaction) |
|
1865 | return self._revlog.strip(minlink, transaction) | |
1864 |
|
1866 | |||
1865 | def files(self): |
|
1867 | def files(self): | |
1866 | return self._revlog.files() |
|
1868 | return self._revlog.files() | |
1867 |
|
1869 | |||
1868 | def clone(self, tr, destrevlog, **kwargs): |
|
1870 | def clone(self, tr, destrevlog, **kwargs): | |
1869 | if not isinstance(destrevlog, manifestrevlog): |
|
1871 | if not isinstance(destrevlog, manifestrevlog): | |
1870 | raise error.ProgrammingError(b'expected manifestrevlog to clone()') |
|
1872 | raise error.ProgrammingError(b'expected manifestrevlog to clone()') | |
1871 |
|
1873 | |||
1872 | return self._revlog.clone(tr, destrevlog._revlog, **kwargs) |
|
1874 | return self._revlog.clone(tr, destrevlog._revlog, **kwargs) | |
1873 |
|
1875 | |||
1874 | def storageinfo( |
|
1876 | def storageinfo( | |
1875 | self, |
|
1877 | self, | |
1876 | exclusivefiles=False, |
|
1878 | exclusivefiles=False, | |
1877 | sharedfiles=False, |
|
1879 | sharedfiles=False, | |
1878 | revisionscount=False, |
|
1880 | revisionscount=False, | |
1879 | trackedsize=False, |
|
1881 | trackedsize=False, | |
1880 | storedsize=False, |
|
1882 | storedsize=False, | |
1881 | ): |
|
1883 | ): | |
1882 | return self._revlog.storageinfo( |
|
1884 | return self._revlog.storageinfo( | |
1883 | exclusivefiles=exclusivefiles, |
|
1885 | exclusivefiles=exclusivefiles, | |
1884 | sharedfiles=sharedfiles, |
|
1886 | sharedfiles=sharedfiles, | |
1885 | revisionscount=revisionscount, |
|
1887 | revisionscount=revisionscount, | |
1886 | trackedsize=trackedsize, |
|
1888 | trackedsize=trackedsize, | |
1887 | storedsize=storedsize, |
|
1889 | storedsize=storedsize, | |
1888 | ) |
|
1890 | ) | |
1889 |
|
1891 | |||
1890 | @property |
|
1892 | @property | |
1891 | def indexfile(self): |
|
1893 | def indexfile(self): | |
1892 | return self._revlog.indexfile |
|
1894 | return self._revlog.indexfile | |
1893 |
|
1895 | |||
1894 | @indexfile.setter |
|
1896 | @indexfile.setter | |
1895 | def indexfile(self, value): |
|
1897 | def indexfile(self, value): | |
1896 | self._revlog.indexfile = value |
|
1898 | self._revlog.indexfile = value | |
1897 |
|
1899 | |||
1898 | @property |
|
1900 | @property | |
1899 | def opener(self): |
|
1901 | def opener(self): | |
1900 | return self._revlog.opener |
|
1902 | return self._revlog.opener | |
1901 |
|
1903 | |||
1902 | @opener.setter |
|
1904 | @opener.setter | |
1903 | def opener(self, value): |
|
1905 | def opener(self, value): | |
1904 | self._revlog.opener = value |
|
1906 | self._revlog.opener = value | |
1905 |
|
1907 | |||
1906 |
|
1908 | |||
1907 | @interfaceutil.implementer(repository.imanifestlog) |
|
1909 | @interfaceutil.implementer(repository.imanifestlog) | |
1908 | class manifestlog(object): |
|
1910 | class manifestlog(object): | |
1909 | """A collection class representing the collection of manifest snapshots |
|
1911 | """A collection class representing the collection of manifest snapshots | |
1910 | referenced by commits in the repository. |
|
1912 | referenced by commits in the repository. | |
1911 |
|
1913 | |||
1912 | In this situation, 'manifest' refers to the abstract concept of a snapshot |
|
1914 | In this situation, 'manifest' refers to the abstract concept of a snapshot | |
1913 | of the list of files in the given commit. Consumers of the output of this |
|
1915 | of the list of files in the given commit. Consumers of the output of this | |
1914 | class do not care about the implementation details of the actual manifests |
|
1916 | class do not care about the implementation details of the actual manifests | |
1915 | they receive (i.e. tree or flat or lazily loaded, etc).""" |
|
1917 | they receive (i.e. tree or flat or lazily loaded, etc).""" | |
1916 |
|
1918 | |||
1917 | def __init__(self, opener, repo, rootstore, narrowmatch): |
|
1919 | def __init__(self, opener, repo, rootstore, narrowmatch): | |
1918 | usetreemanifest = False |
|
1920 | usetreemanifest = False | |
1919 | cachesize = 4 |
|
1921 | cachesize = 4 | |
1920 |
|
1922 | |||
1921 | opts = getattr(opener, 'options', None) |
|
1923 | opts = getattr(opener, 'options', None) | |
1922 | if opts is not None: |
|
1924 | if opts is not None: | |
1923 | usetreemanifest = opts.get(b'treemanifest', usetreemanifest) |
|
1925 | usetreemanifest = opts.get(b'treemanifest', usetreemanifest) | |
1924 | cachesize = opts.get(b'manifestcachesize', cachesize) |
|
1926 | cachesize = opts.get(b'manifestcachesize', cachesize) | |
1925 |
|
1927 | |||
1926 | self._treemanifests = usetreemanifest |
|
1928 | self._treemanifests = usetreemanifest | |
1927 |
|
1929 | |||
1928 | self._rootstore = rootstore |
|
1930 | self._rootstore = rootstore | |
1929 | self._rootstore._setupmanifestcachehooks(repo) |
|
1931 | self._rootstore._setupmanifestcachehooks(repo) | |
1930 | self._narrowmatch = narrowmatch |
|
1932 | self._narrowmatch = narrowmatch | |
1931 |
|
1933 | |||
1932 | # A cache of the manifestctx or treemanifestctx for each directory |
|
1934 | # A cache of the manifestctx or treemanifestctx for each directory | |
1933 | self._dirmancache = {} |
|
1935 | self._dirmancache = {} | |
1934 | self._dirmancache[b''] = util.lrucachedict(cachesize) |
|
1936 | self._dirmancache[b''] = util.lrucachedict(cachesize) | |
1935 |
|
1937 | |||
1936 | self._cachesize = cachesize |
|
1938 | self._cachesize = cachesize | |
1937 |
|
1939 | |||
1938 | def __getitem__(self, node): |
|
1940 | def __getitem__(self, node): | |
1939 | """Retrieves the manifest instance for the given node. Throws a |
|
1941 | """Retrieves the manifest instance for the given node. Throws a | |
1940 | LookupError if not found. |
|
1942 | LookupError if not found. | |
1941 | """ |
|
1943 | """ | |
1942 | return self.get(b'', node) |
|
1944 | return self.get(b'', node) | |
1943 |
|
1945 | |||
1944 | def get(self, tree, node, verify=True): |
|
1946 | def get(self, tree, node, verify=True): | |
1945 | """Retrieves the manifest instance for the given node. Throws a |
|
1947 | """Retrieves the manifest instance for the given node. Throws a | |
1946 | LookupError if not found. |
|
1948 | LookupError if not found. | |
1947 |
|
1949 | |||
1948 | `verify` - if True an exception will be thrown if the node is not in |
|
1950 | `verify` - if True an exception will be thrown if the node is not in | |
1949 | the revlog |
|
1951 | the revlog | |
1950 | """ |
|
1952 | """ | |
1951 | if node in self._dirmancache.get(tree, ()): |
|
1953 | if node in self._dirmancache.get(tree, ()): | |
1952 | return self._dirmancache[tree][node] |
|
1954 | return self._dirmancache[tree][node] | |
1953 |
|
1955 | |||
1954 | if not self._narrowmatch.always(): |
|
1956 | if not self._narrowmatch.always(): | |
1955 | if not self._narrowmatch.visitdir(tree[:-1]): |
|
1957 | if not self._narrowmatch.visitdir(tree[:-1]): | |
1956 | return excludeddirmanifestctx(tree, node) |
|
1958 | return excludeddirmanifestctx(tree, node) | |
1957 | if tree: |
|
1959 | if tree: | |
1958 | if self._rootstore._treeondisk: |
|
1960 | if self._rootstore._treeondisk: | |
1959 | if verify: |
|
1961 | if verify: | |
1960 | # Side-effect is LookupError is raised if node doesn't |
|
1962 | # Side-effect is LookupError is raised if node doesn't | |
1961 | # exist. |
|
1963 | # exist. | |
1962 | self.getstorage(tree).rev(node) |
|
1964 | self.getstorage(tree).rev(node) | |
1963 |
|
1965 | |||
1964 | m = treemanifestctx(self, tree, node) |
|
1966 | m = treemanifestctx(self, tree, node) | |
1965 | else: |
|
1967 | else: | |
1966 | raise error.Abort( |
|
1968 | raise error.Abort( | |
1967 | _( |
|
1969 | _( | |
1968 | b"cannot ask for manifest directory '%s' in a flat " |
|
1970 | b"cannot ask for manifest directory '%s' in a flat " | |
1969 | b"manifest" |
|
1971 | b"manifest" | |
1970 | ) |
|
1972 | ) | |
1971 | % tree |
|
1973 | % tree | |
1972 | ) |
|
1974 | ) | |
1973 | else: |
|
1975 | else: | |
1974 | if verify: |
|
1976 | if verify: | |
1975 | # Side-effect is LookupError is raised if node doesn't exist. |
|
1977 | # Side-effect is LookupError is raised if node doesn't exist. | |
1976 | self._rootstore.rev(node) |
|
1978 | self._rootstore.rev(node) | |
1977 |
|
1979 | |||
1978 | if self._treemanifests: |
|
1980 | if self._treemanifests: | |
1979 | m = treemanifestctx(self, b'', node) |
|
1981 | m = treemanifestctx(self, b'', node) | |
1980 | else: |
|
1982 | else: | |
1981 | m = manifestctx(self, node) |
|
1983 | m = manifestctx(self, node) | |
1982 |
|
1984 | |||
1983 | if node != nullid: |
|
1985 | if node != nullid: | |
1984 | mancache = self._dirmancache.get(tree) |
|
1986 | mancache = self._dirmancache.get(tree) | |
1985 | if not mancache: |
|
1987 | if not mancache: | |
1986 | mancache = util.lrucachedict(self._cachesize) |
|
1988 | mancache = util.lrucachedict(self._cachesize) | |
1987 | self._dirmancache[tree] = mancache |
|
1989 | self._dirmancache[tree] = mancache | |
1988 | mancache[node] = m |
|
1990 | mancache[node] = m | |
1989 | return m |
|
1991 | return m | |
1990 |
|
1992 | |||
1991 | def getstorage(self, tree): |
|
1993 | def getstorage(self, tree): | |
1992 | return self._rootstore.dirlog(tree) |
|
1994 | return self._rootstore.dirlog(tree) | |
1993 |
|
1995 | |||
1994 | def clearcaches(self, clear_persisted_data=False): |
|
1996 | def clearcaches(self, clear_persisted_data=False): | |
1995 | self._dirmancache.clear() |
|
1997 | self._dirmancache.clear() | |
1996 | self._rootstore.clearcaches(clear_persisted_data=clear_persisted_data) |
|
1998 | self._rootstore.clearcaches(clear_persisted_data=clear_persisted_data) | |
1997 |
|
1999 | |||
1998 | def rev(self, node): |
|
2000 | def rev(self, node): | |
1999 | return self._rootstore.rev(node) |
|
2001 | return self._rootstore.rev(node) | |
2000 |
|
2002 | |||
2001 | def update_caches(self, transaction): |
|
2003 | def update_caches(self, transaction): | |
2002 | return self._rootstore._revlog.update_caches(transaction=transaction) |
|
2004 | return self._rootstore._revlog.update_caches(transaction=transaction) | |
2003 |
|
2005 | |||
2004 |
|
2006 | |||
2005 | @interfaceutil.implementer(repository.imanifestrevisionwritable) |
|
2007 | @interfaceutil.implementer(repository.imanifestrevisionwritable) | |
2006 | class memmanifestctx(object): |
|
2008 | class memmanifestctx(object): | |
2007 | def __init__(self, manifestlog): |
|
2009 | def __init__(self, manifestlog): | |
2008 | self._manifestlog = manifestlog |
|
2010 | self._manifestlog = manifestlog | |
2009 | self._manifestdict = manifestdict() |
|
2011 | self._manifestdict = manifestdict() | |
2010 |
|
2012 | |||
2011 | def _storage(self): |
|
2013 | def _storage(self): | |
2012 | return self._manifestlog.getstorage(b'') |
|
2014 | return self._manifestlog.getstorage(b'') | |
2013 |
|
2015 | |||
2014 | def copy(self): |
|
2016 | def copy(self): | |
2015 | memmf = memmanifestctx(self._manifestlog) |
|
2017 | memmf = memmanifestctx(self._manifestlog) | |
2016 | memmf._manifestdict = self.read().copy() |
|
2018 | memmf._manifestdict = self.read().copy() | |
2017 | return memmf |
|
2019 | return memmf | |
2018 |
|
2020 | |||
2019 | def read(self): |
|
2021 | def read(self): | |
2020 | return self._manifestdict |
|
2022 | return self._manifestdict | |
2021 |
|
2023 | |||
2022 | def write(self, transaction, link, p1, p2, added, removed, match=None): |
|
2024 | def write(self, transaction, link, p1, p2, added, removed, match=None): | |
2023 | return self._storage().add( |
|
2025 | return self._storage().add( | |
2024 | self._manifestdict, |
|
2026 | self._manifestdict, | |
2025 | transaction, |
|
2027 | transaction, | |
2026 | link, |
|
2028 | link, | |
2027 | p1, |
|
2029 | p1, | |
2028 | p2, |
|
2030 | p2, | |
2029 | added, |
|
2031 | added, | |
2030 | removed, |
|
2032 | removed, | |
2031 | match=match, |
|
2033 | match=match, | |
2032 | ) |
|
2034 | ) | |
2033 |
|
2035 | |||
2034 |
|
2036 | |||
2035 | @interfaceutil.implementer(repository.imanifestrevisionstored) |
|
2037 | @interfaceutil.implementer(repository.imanifestrevisionstored) | |
2036 | class manifestctx(object): |
|
2038 | class manifestctx(object): | |
2037 | """A class representing a single revision of a manifest, including its |
|
2039 | """A class representing a single revision of a manifest, including its | |
2038 | contents, its parent revs, and its linkrev. |
|
2040 | contents, its parent revs, and its linkrev. | |
2039 | """ |
|
2041 | """ | |
2040 |
|
2042 | |||
2041 | def __init__(self, manifestlog, node): |
|
2043 | def __init__(self, manifestlog, node): | |
2042 | self._manifestlog = manifestlog |
|
2044 | self._manifestlog = manifestlog | |
2043 | self._data = None |
|
2045 | self._data = None | |
2044 |
|
2046 | |||
2045 | self._node = node |
|
2047 | self._node = node | |
2046 |
|
2048 | |||
2047 | # TODO: We eventually want p1, p2, and linkrev exposed on this class, |
|
2049 | # TODO: We eventually want p1, p2, and linkrev exposed on this class, | |
2048 | # but let's add it later when something needs it and we can load it |
|
2050 | # but let's add it later when something needs it and we can load it | |
2049 | # lazily. |
|
2051 | # lazily. | |
2050 | # self.p1, self.p2 = store.parents(node) |
|
2052 | # self.p1, self.p2 = store.parents(node) | |
2051 | # rev = store.rev(node) |
|
2053 | # rev = store.rev(node) | |
2052 | # self.linkrev = store.linkrev(rev) |
|
2054 | # self.linkrev = store.linkrev(rev) | |
2053 |
|
2055 | |||
2054 | def _storage(self): |
|
2056 | def _storage(self): | |
2055 | return self._manifestlog.getstorage(b'') |
|
2057 | return self._manifestlog.getstorage(b'') | |
2056 |
|
2058 | |||
2057 | def node(self): |
|
2059 | def node(self): | |
2058 | return self._node |
|
2060 | return self._node | |
2059 |
|
2061 | |||
2060 | def copy(self): |
|
2062 | def copy(self): | |
2061 | memmf = memmanifestctx(self._manifestlog) |
|
2063 | memmf = memmanifestctx(self._manifestlog) | |
2062 | memmf._manifestdict = self.read().copy() |
|
2064 | memmf._manifestdict = self.read().copy() | |
2063 | return memmf |
|
2065 | return memmf | |
2064 |
|
2066 | |||
2065 | @propertycache |
|
2067 | @propertycache | |
2066 | def parents(self): |
|
2068 | def parents(self): | |
2067 | return self._storage().parents(self._node) |
|
2069 | return self._storage().parents(self._node) | |
2068 |
|
2070 | |||
2069 | def read(self): |
|
2071 | def read(self): | |
2070 | if self._data is None: |
|
2072 | if self._data is None: | |
2071 | if self._node == nullid: |
|
2073 | if self._node == nullid: | |
2072 | self._data = manifestdict() |
|
2074 | self._data = manifestdict() | |
2073 | else: |
|
2075 | else: | |
2074 | store = self._storage() |
|
2076 | store = self._storage() | |
2075 | if self._node in store.fulltextcache: |
|
2077 | if self._node in store.fulltextcache: | |
2076 | text = pycompat.bytestr(store.fulltextcache[self._node]) |
|
2078 | text = pycompat.bytestr(store.fulltextcache[self._node]) | |
2077 | else: |
|
2079 | else: | |
2078 | text = store.revision(self._node) |
|
2080 | text = store.revision(self._node) | |
2079 | arraytext = bytearray(text) |
|
2081 | arraytext = bytearray(text) | |
2080 | store.fulltextcache[self._node] = arraytext |
|
2082 | store.fulltextcache[self._node] = arraytext | |
2081 | self._data = manifestdict(text) |
|
2083 | self._data = manifestdict(text) | |
2082 | return self._data |
|
2084 | return self._data | |
2083 |
|
2085 | |||
2084 | def readfast(self, shallow=False): |
|
2086 | def readfast(self, shallow=False): | |
2085 | """Calls either readdelta or read, based on which would be less work. |
|
2087 | """Calls either readdelta or read, based on which would be less work. | |
2086 | readdelta is called if the delta is against the p1, and therefore can be |
|
2088 | readdelta is called if the delta is against the p1, and therefore can be | |
2087 | read quickly. |
|
2089 | read quickly. | |
2088 |
|
2090 | |||
2089 | If `shallow` is True, nothing changes since this is a flat manifest. |
|
2091 | If `shallow` is True, nothing changes since this is a flat manifest. | |
2090 | """ |
|
2092 | """ | |
2091 | store = self._storage() |
|
2093 | store = self._storage() | |
2092 | r = store.rev(self._node) |
|
2094 | r = store.rev(self._node) | |
2093 | deltaparent = store.deltaparent(r) |
|
2095 | deltaparent = store.deltaparent(r) | |
2094 | if deltaparent != nullrev and deltaparent in store.parentrevs(r): |
|
2096 | if deltaparent != nullrev and deltaparent in store.parentrevs(r): | |
2095 | return self.readdelta() |
|
2097 | return self.readdelta() | |
2096 | return self.read() |
|
2098 | return self.read() | |
2097 |
|
2099 | |||
2098 | def readdelta(self, shallow=False): |
|
2100 | def readdelta(self, shallow=False): | |
2099 | """Returns a manifest containing just the entries that are present |
|
2101 | """Returns a manifest containing just the entries that are present | |
2100 | in this manifest, but not in its p1 manifest. This is efficient to read |
|
2102 | in this manifest, but not in its p1 manifest. This is efficient to read | |
2101 | if the revlog delta is already p1. |
|
2103 | if the revlog delta is already p1. | |
2102 |
|
2104 | |||
2103 | Changing the value of `shallow` has no effect on flat manifests. |
|
2105 | Changing the value of `shallow` has no effect on flat manifests. | |
2104 | """ |
|
2106 | """ | |
2105 | store = self._storage() |
|
2107 | store = self._storage() | |
2106 | r = store.rev(self._node) |
|
2108 | r = store.rev(self._node) | |
2107 | d = mdiff.patchtext(store.revdiff(store.deltaparent(r), r)) |
|
2109 | d = mdiff.patchtext(store.revdiff(store.deltaparent(r), r)) | |
2108 | return manifestdict(d) |
|
2110 | return manifestdict(d) | |
2109 |
|
2111 | |||
2110 | def find(self, key): |
|
2112 | def find(self, key): | |
2111 | return self.read().find(key) |
|
2113 | return self.read().find(key) | |
2112 |
|
2114 | |||
2113 |
|
2115 | |||
2114 | @interfaceutil.implementer(repository.imanifestrevisionwritable) |
|
2116 | @interfaceutil.implementer(repository.imanifestrevisionwritable) | |
2115 | class memtreemanifestctx(object): |
|
2117 | class memtreemanifestctx(object): | |
2116 | def __init__(self, manifestlog, dir=b''): |
|
2118 | def __init__(self, manifestlog, dir=b''): | |
2117 | self._manifestlog = manifestlog |
|
2119 | self._manifestlog = manifestlog | |
2118 | self._dir = dir |
|
2120 | self._dir = dir | |
2119 | self._treemanifest = treemanifest() |
|
2121 | self._treemanifest = treemanifest() | |
2120 |
|
2122 | |||
2121 | def _storage(self): |
|
2123 | def _storage(self): | |
2122 | return self._manifestlog.getstorage(b'') |
|
2124 | return self._manifestlog.getstorage(b'') | |
2123 |
|
2125 | |||
2124 | def copy(self): |
|
2126 | def copy(self): | |
2125 | memmf = memtreemanifestctx(self._manifestlog, dir=self._dir) |
|
2127 | memmf = memtreemanifestctx(self._manifestlog, dir=self._dir) | |
2126 | memmf._treemanifest = self._treemanifest.copy() |
|
2128 | memmf._treemanifest = self._treemanifest.copy() | |
2127 | return memmf |
|
2129 | return memmf | |
2128 |
|
2130 | |||
2129 | def read(self): |
|
2131 | def read(self): | |
2130 | return self._treemanifest |
|
2132 | return self._treemanifest | |
2131 |
|
2133 | |||
2132 | def write(self, transaction, link, p1, p2, added, removed, match=None): |
|
2134 | def write(self, transaction, link, p1, p2, added, removed, match=None): | |
2133 | def readtree(dir, node): |
|
2135 | def readtree(dir, node): | |
2134 | return self._manifestlog.get(dir, node).read() |
|
2136 | return self._manifestlog.get(dir, node).read() | |
2135 |
|
2137 | |||
2136 | return self._storage().add( |
|
2138 | return self._storage().add( | |
2137 | self._treemanifest, |
|
2139 | self._treemanifest, | |
2138 | transaction, |
|
2140 | transaction, | |
2139 | link, |
|
2141 | link, | |
2140 | p1, |
|
2142 | p1, | |
2141 | p2, |
|
2143 | p2, | |
2142 | added, |
|
2144 | added, | |
2143 | removed, |
|
2145 | removed, | |
2144 | readtree=readtree, |
|
2146 | readtree=readtree, | |
2145 | match=match, |
|
2147 | match=match, | |
2146 | ) |
|
2148 | ) | |
2147 |
|
2149 | |||
2148 |
|
2150 | |||
2149 | @interfaceutil.implementer(repository.imanifestrevisionstored) |
|
2151 | @interfaceutil.implementer(repository.imanifestrevisionstored) | |
2150 | class treemanifestctx(object): |
|
2152 | class treemanifestctx(object): | |
2151 | def __init__(self, manifestlog, dir, node): |
|
2153 | def __init__(self, manifestlog, dir, node): | |
2152 | self._manifestlog = manifestlog |
|
2154 | self._manifestlog = manifestlog | |
2153 | self._dir = dir |
|
2155 | self._dir = dir | |
2154 | self._data = None |
|
2156 | self._data = None | |
2155 |
|
2157 | |||
2156 | self._node = node |
|
2158 | self._node = node | |
2157 |
|
2159 | |||
2158 | # TODO: Load p1/p2/linkrev lazily. They need to be lazily loaded so that |
|
2160 | # TODO: Load p1/p2/linkrev lazily. They need to be lazily loaded so that | |
2159 | # we can instantiate treemanifestctx objects for directories we don't |
|
2161 | # we can instantiate treemanifestctx objects for directories we don't | |
2160 | # have on disk. |
|
2162 | # have on disk. | |
2161 | # self.p1, self.p2 = store.parents(node) |
|
2163 | # self.p1, self.p2 = store.parents(node) | |
2162 | # rev = store.rev(node) |
|
2164 | # rev = store.rev(node) | |
2163 | # self.linkrev = store.linkrev(rev) |
|
2165 | # self.linkrev = store.linkrev(rev) | |
2164 |
|
2166 | |||
2165 | def _storage(self): |
|
2167 | def _storage(self): | |
2166 | narrowmatch = self._manifestlog._narrowmatch |
|
2168 | narrowmatch = self._manifestlog._narrowmatch | |
2167 | if not narrowmatch.always(): |
|
2169 | if not narrowmatch.always(): | |
2168 | if not narrowmatch.visitdir(self._dir[:-1]): |
|
2170 | if not narrowmatch.visitdir(self._dir[:-1]): | |
2169 | return excludedmanifestrevlog(self._dir) |
|
2171 | return excludedmanifestrevlog(self._dir) | |
2170 | return self._manifestlog.getstorage(self._dir) |
|
2172 | return self._manifestlog.getstorage(self._dir) | |
2171 |
|
2173 | |||
2172 | def read(self): |
|
2174 | def read(self): | |
2173 | if self._data is None: |
|
2175 | if self._data is None: | |
2174 | store = self._storage() |
|
2176 | store = self._storage() | |
2175 | if self._node == nullid: |
|
2177 | if self._node == nullid: | |
2176 | self._data = treemanifest() |
|
2178 | self._data = treemanifest() | |
2177 | # TODO accessing non-public API |
|
2179 | # TODO accessing non-public API | |
2178 | elif store._treeondisk: |
|
2180 | elif store._treeondisk: | |
2179 | m = treemanifest(dir=self._dir) |
|
2181 | m = treemanifest(dir=self._dir) | |
2180 |
|
2182 | |||
2181 | def gettext(): |
|
2183 | def gettext(): | |
2182 | return store.revision(self._node) |
|
2184 | return store.revision(self._node) | |
2183 |
|
2185 | |||
2184 | def readsubtree(dir, subm): |
|
2186 | def readsubtree(dir, subm): | |
2185 | # Set verify to False since we need to be able to create |
|
2187 | # Set verify to False since we need to be able to create | |
2186 | # subtrees for trees that don't exist on disk. |
|
2188 | # subtrees for trees that don't exist on disk. | |
2187 | return self._manifestlog.get(dir, subm, verify=False).read() |
|
2189 | return self._manifestlog.get(dir, subm, verify=False).read() | |
2188 |
|
2190 | |||
2189 | m.read(gettext, readsubtree) |
|
2191 | m.read(gettext, readsubtree) | |
2190 | m.setnode(self._node) |
|
2192 | m.setnode(self._node) | |
2191 | self._data = m |
|
2193 | self._data = m | |
2192 | else: |
|
2194 | else: | |
2193 | if self._node in store.fulltextcache: |
|
2195 | if self._node in store.fulltextcache: | |
2194 | text = pycompat.bytestr(store.fulltextcache[self._node]) |
|
2196 | text = pycompat.bytestr(store.fulltextcache[self._node]) | |
2195 | else: |
|
2197 | else: | |
2196 | text = store.revision(self._node) |
|
2198 | text = store.revision(self._node) | |
2197 | arraytext = bytearray(text) |
|
2199 | arraytext = bytearray(text) | |
2198 | store.fulltextcache[self._node] = arraytext |
|
2200 | store.fulltextcache[self._node] = arraytext | |
2199 | self._data = treemanifest(dir=self._dir, text=text) |
|
2201 | self._data = treemanifest(dir=self._dir, text=text) | |
2200 |
|
2202 | |||
2201 | return self._data |
|
2203 | return self._data | |
2202 |
|
2204 | |||
2203 | def node(self): |
|
2205 | def node(self): | |
2204 | return self._node |
|
2206 | return self._node | |
2205 |
|
2207 | |||
2206 | def copy(self): |
|
2208 | def copy(self): | |
2207 | memmf = memtreemanifestctx(self._manifestlog, dir=self._dir) |
|
2209 | memmf = memtreemanifestctx(self._manifestlog, dir=self._dir) | |
2208 | memmf._treemanifest = self.read().copy() |
|
2210 | memmf._treemanifest = self.read().copy() | |
2209 | return memmf |
|
2211 | return memmf | |
2210 |
|
2212 | |||
2211 | @propertycache |
|
2213 | @propertycache | |
2212 | def parents(self): |
|
2214 | def parents(self): | |
2213 | return self._storage().parents(self._node) |
|
2215 | return self._storage().parents(self._node) | |
2214 |
|
2216 | |||
2215 | def readdelta(self, shallow=False): |
|
2217 | def readdelta(self, shallow=False): | |
2216 | """Returns a manifest containing just the entries that are present |
|
2218 | """Returns a manifest containing just the entries that are present | |
2217 | in this manifest, but not in its p1 manifest. This is efficient to read |
|
2219 | in this manifest, but not in its p1 manifest. This is efficient to read | |
2218 | if the revlog delta is already p1. |
|
2220 | if the revlog delta is already p1. | |
2219 |
|
2221 | |||
2220 | If `shallow` is True, this will read the delta for this directory, |
|
2222 | If `shallow` is True, this will read the delta for this directory, | |
2221 | without recursively reading subdirectory manifests. Instead, any |
|
2223 | without recursively reading subdirectory manifests. Instead, any | |
2222 | subdirectory entry will be reported as it appears in the manifest, i.e. |
|
2224 | subdirectory entry will be reported as it appears in the manifest, i.e. | |
2223 | the subdirectory will be reported among files and distinguished only by |
|
2225 | the subdirectory will be reported among files and distinguished only by | |
2224 | its 't' flag. |
|
2226 | its 't' flag. | |
2225 | """ |
|
2227 | """ | |
2226 | store = self._storage() |
|
2228 | store = self._storage() | |
2227 | if shallow: |
|
2229 | if shallow: | |
2228 | r = store.rev(self._node) |
|
2230 | r = store.rev(self._node) | |
2229 | d = mdiff.patchtext(store.revdiff(store.deltaparent(r), r)) |
|
2231 | d = mdiff.patchtext(store.revdiff(store.deltaparent(r), r)) | |
2230 | return manifestdict(d) |
|
2232 | return manifestdict(d) | |
2231 | else: |
|
2233 | else: | |
2232 | # Need to perform a slow delta |
|
2234 | # Need to perform a slow delta | |
2233 | r0 = store.deltaparent(store.rev(self._node)) |
|
2235 | r0 = store.deltaparent(store.rev(self._node)) | |
2234 | m0 = self._manifestlog.get(self._dir, store.node(r0)).read() |
|
2236 | m0 = self._manifestlog.get(self._dir, store.node(r0)).read() | |
2235 | m1 = self.read() |
|
2237 | m1 = self.read() | |
2236 | md = treemanifest(dir=self._dir) |
|
2238 | md = treemanifest(dir=self._dir) | |
2237 | for f, ((n0, fl0), (n1, fl1)) in pycompat.iteritems(m0.diff(m1)): |
|
2239 | for f, ((n0, fl0), (n1, fl1)) in pycompat.iteritems(m0.diff(m1)): | |
2238 | if n1: |
|
2240 | if n1: | |
2239 | md[f] = n1 |
|
2241 | md[f] = n1 | |
2240 | if fl1: |
|
2242 | if fl1: | |
2241 | md.setflag(f, fl1) |
|
2243 | md.setflag(f, fl1) | |
2242 | return md |
|
2244 | return md | |
2243 |
|
2245 | |||
2244 | def readfast(self, shallow=False): |
|
2246 | def readfast(self, shallow=False): | |
2245 | """Calls either readdelta or read, based on which would be less work. |
|
2247 | """Calls either readdelta or read, based on which would be less work. | |
2246 | readdelta is called if the delta is against the p1, and therefore can be |
|
2248 | readdelta is called if the delta is against the p1, and therefore can be | |
2247 | read quickly. |
|
2249 | read quickly. | |
2248 |
|
2250 | |||
2249 | If `shallow` is True, it only returns the entries from this manifest, |
|
2251 | If `shallow` is True, it only returns the entries from this manifest, | |
2250 | and not any submanifests. |
|
2252 | and not any submanifests. | |
2251 | """ |
|
2253 | """ | |
2252 | store = self._storage() |
|
2254 | store = self._storage() | |
2253 | r = store.rev(self._node) |
|
2255 | r = store.rev(self._node) | |
2254 | deltaparent = store.deltaparent(r) |
|
2256 | deltaparent = store.deltaparent(r) | |
2255 | if deltaparent != nullrev and deltaparent in store.parentrevs(r): |
|
2257 | if deltaparent != nullrev and deltaparent in store.parentrevs(r): | |
2256 | return self.readdelta(shallow=shallow) |
|
2258 | return self.readdelta(shallow=shallow) | |
2257 |
|
2259 | |||
2258 | if shallow: |
|
2260 | if shallow: | |
2259 | return manifestdict(store.revision(self._node)) |
|
2261 | return manifestdict(store.revision(self._node)) | |
2260 | else: |
|
2262 | else: | |
2261 | return self.read() |
|
2263 | return self.read() | |
2262 |
|
2264 | |||
2263 | def find(self, key): |
|
2265 | def find(self, key): | |
2264 | return self.read().find(key) |
|
2266 | return self.read().find(key) | |
2265 |
|
2267 | |||
2266 |
|
2268 | |||
2267 | class excludeddir(treemanifest): |
|
2269 | class excludeddir(treemanifest): | |
2268 | """Stand-in for a directory that is excluded from the repository. |
|
2270 | """Stand-in for a directory that is excluded from the repository. | |
2269 |
|
2271 | |||
2270 | With narrowing active on a repository that uses treemanifests, |
|
2272 | With narrowing active on a repository that uses treemanifests, | |
2271 | some of the directory revlogs will be excluded from the resulting |
|
2273 | some of the directory revlogs will be excluded from the resulting | |
2272 | clone. This is a huge storage win for clients, but means we need |
|
2274 | clone. This is a huge storage win for clients, but means we need | |
2273 | some sort of pseudo-manifest to surface to internals so we can |
|
2275 | some sort of pseudo-manifest to surface to internals so we can | |
2274 | detect a merge conflict outside the narrowspec. That's what this |
|
2276 | detect a merge conflict outside the narrowspec. That's what this | |
2275 | class is: it stands in for a directory whose node is known, but |
|
2277 | class is: it stands in for a directory whose node is known, but | |
2276 | whose contents are unknown. |
|
2278 | whose contents are unknown. | |
2277 | """ |
|
2279 | """ | |
2278 |
|
2280 | |||
2279 | def __init__(self, dir, node): |
|
2281 | def __init__(self, dir, node): | |
2280 | super(excludeddir, self).__init__(dir) |
|
2282 | super(excludeddir, self).__init__(dir) | |
2281 | self._node = node |
|
2283 | self._node = node | |
2282 | # Add an empty file, which will be included by iterators and such, |
|
2284 | # Add an empty file, which will be included by iterators and such, | |
2283 | # appearing as the directory itself (i.e. something like "dir/") |
|
2285 | # appearing as the directory itself (i.e. something like "dir/") | |
2284 | self._files[b''] = node |
|
2286 | self._files[b''] = node | |
2285 | self._flags[b''] = b't' |
|
2287 | self._flags[b''] = b't' | |
2286 |
|
2288 | |||
2287 | # Manifests outside the narrowspec should never be modified, so avoid |
|
2289 | # Manifests outside the narrowspec should never be modified, so avoid | |
2288 | # copying. This makes a noticeable difference when there are very many |
|
2290 | # copying. This makes a noticeable difference when there are very many | |
2289 | # directories outside the narrowspec. Also, it makes sense for the copy to |
|
2291 | # directories outside the narrowspec. Also, it makes sense for the copy to | |
2290 | # be of the same type as the original, which would not happen with the |
|
2292 | # be of the same type as the original, which would not happen with the | |
2291 | # super type's copy(). |
|
2293 | # super type's copy(). | |
2292 | def copy(self): |
|
2294 | def copy(self): | |
2293 | return self |
|
2295 | return self | |
2294 |
|
2296 | |||
2295 |
|
2297 | |||
2296 | class excludeddirmanifestctx(treemanifestctx): |
|
2298 | class excludeddirmanifestctx(treemanifestctx): | |
2297 | """context wrapper for excludeddir - see that docstring for rationale""" |
|
2299 | """context wrapper for excludeddir - see that docstring for rationale""" | |
2298 |
|
2300 | |||
2299 | def __init__(self, dir, node): |
|
2301 | def __init__(self, dir, node): | |
2300 | self._dir = dir |
|
2302 | self._dir = dir | |
2301 | self._node = node |
|
2303 | self._node = node | |
2302 |
|
2304 | |||
2303 | def read(self): |
|
2305 | def read(self): | |
2304 | return excludeddir(self._dir, self._node) |
|
2306 | return excludeddir(self._dir, self._node) | |
2305 |
|
2307 | |||
2306 | def readfast(self, shallow=False): |
|
2308 | def readfast(self, shallow=False): | |
2307 | # special version of readfast since we don't have underlying storage |
|
2309 | # special version of readfast since we don't have underlying storage | |
2308 | return self.read() |
|
2310 | return self.read() | |
2309 |
|
2311 | |||
2310 | def write(self, *args): |
|
2312 | def write(self, *args): | |
2311 | raise error.ProgrammingError( |
|
2313 | raise error.ProgrammingError( | |
2312 | b'attempt to write manifest from excluded dir %s' % self._dir |
|
2314 | b'attempt to write manifest from excluded dir %s' % self._dir | |
2313 | ) |
|
2315 | ) | |
2314 |
|
2316 | |||
2315 |
|
2317 | |||
2316 | class excludedmanifestrevlog(manifestrevlog): |
|
2318 | class excludedmanifestrevlog(manifestrevlog): | |
2317 | """Stand-in for excluded treemanifest revlogs. |
|
2319 | """Stand-in for excluded treemanifest revlogs. | |
2318 |
|
2320 | |||
2319 | When narrowing is active on a treemanifest repository, we'll have |
|
2321 | When narrowing is active on a treemanifest repository, we'll have | |
2320 | references to directories we can't see due to the revlog being |
|
2322 | references to directories we can't see due to the revlog being | |
2321 | skipped. This class exists to conform to the manifestrevlog |
|
2323 | skipped. This class exists to conform to the manifestrevlog | |
2322 | interface for those directories and proactively prevent writes to |
|
2324 | interface for those directories and proactively prevent writes to | |
2323 | outside the narrowspec. |
|
2325 | outside the narrowspec. | |
2324 | """ |
|
2326 | """ | |
2325 |
|
2327 | |||
2326 | def __init__(self, dir): |
|
2328 | def __init__(self, dir): | |
2327 | self._dir = dir |
|
2329 | self._dir = dir | |
2328 |
|
2330 | |||
2329 | def __len__(self): |
|
2331 | def __len__(self): | |
2330 | raise error.ProgrammingError( |
|
2332 | raise error.ProgrammingError( | |
2331 | b'attempt to get length of excluded dir %s' % self._dir |
|
2333 | b'attempt to get length of excluded dir %s' % self._dir | |
2332 | ) |
|
2334 | ) | |
2333 |
|
2335 | |||
2334 | def rev(self, node): |
|
2336 | def rev(self, node): | |
2335 | raise error.ProgrammingError( |
|
2337 | raise error.ProgrammingError( | |
2336 | b'attempt to get rev from excluded dir %s' % self._dir |
|
2338 | b'attempt to get rev from excluded dir %s' % self._dir | |
2337 | ) |
|
2339 | ) | |
2338 |
|
2340 | |||
2339 | def linkrev(self, node): |
|
2341 | def linkrev(self, node): | |
2340 | raise error.ProgrammingError( |
|
2342 | raise error.ProgrammingError( | |
2341 | b'attempt to get linkrev from excluded dir %s' % self._dir |
|
2343 | b'attempt to get linkrev from excluded dir %s' % self._dir | |
2342 | ) |
|
2344 | ) | |
2343 |
|
2345 | |||
2344 | def node(self, rev): |
|
2346 | def node(self, rev): | |
2345 | raise error.ProgrammingError( |
|
2347 | raise error.ProgrammingError( | |
2346 | b'attempt to get node from excluded dir %s' % self._dir |
|
2348 | b'attempt to get node from excluded dir %s' % self._dir | |
2347 | ) |
|
2349 | ) | |
2348 |
|
2350 | |||
2349 | def add(self, *args, **kwargs): |
|
2351 | def add(self, *args, **kwargs): | |
2350 | # We should never write entries in dirlogs outside the narrow clone. |
|
2352 | # We should never write entries in dirlogs outside the narrow clone. | |
2351 | # However, the method still gets called from writesubtree() in |
|
2353 | # However, the method still gets called from writesubtree() in | |
2352 | # _addtree(), so we need to handle it. We should possibly make that |
|
2354 | # _addtree(), so we need to handle it. We should possibly make that | |
2353 | # avoid calling add() with a clean manifest (_dirty is always False |
|
2355 | # avoid calling add() with a clean manifest (_dirty is always False | |
2354 | # in excludeddir instances). |
|
2356 | # in excludeddir instances). | |
2355 | pass |
|
2357 | pass |
@@ -1,3201 +1,3203 b'' | |||||
1 | # revlog.py - storage back-end for mercurial |
|
1 | # revlog.py - storage back-end for mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | """Storage back-end for Mercurial. |
|
8 | """Storage back-end for Mercurial. | |
9 |
|
9 | |||
10 | This provides efficient delta storage with O(1) retrieve and append |
|
10 | This provides efficient delta storage with O(1) retrieve and append | |
11 | and O(changes) merge between branches. |
|
11 | and O(changes) merge between branches. | |
12 | """ |
|
12 | """ | |
13 |
|
13 | |||
14 | from __future__ import absolute_import |
|
14 | from __future__ import absolute_import | |
15 |
|
15 | |||
16 | import collections |
|
16 | import collections | |
17 | import contextlib |
|
17 | import contextlib | |
18 | import errno |
|
18 | import errno | |
19 | import io |
|
19 | import io | |
20 | import os |
|
20 | import os | |
21 | import struct |
|
21 | import struct | |
22 | import zlib |
|
22 | import zlib | |
23 |
|
23 | |||
24 | # import stuff from node for others to import from revlog |
|
24 | # import stuff from node for others to import from revlog | |
25 | from .node import ( |
|
25 | from .node import ( | |
26 | bin, |
|
26 | bin, | |
27 | hex, |
|
27 | hex, | |
28 | nullhex, |
|
28 | nullhex, | |
29 | nullid, |
|
29 | nullid, | |
30 | nullrev, |
|
30 | nullrev, | |
31 | short, |
|
31 | short, | |
32 | wdirfilenodeids, |
|
32 | wdirfilenodeids, | |
33 | wdirhex, |
|
33 | wdirhex, | |
34 | wdirid, |
|
34 | wdirid, | |
35 | wdirrev, |
|
35 | wdirrev, | |
36 | ) |
|
36 | ) | |
37 | from .i18n import _ |
|
37 | from .i18n import _ | |
38 | from .pycompat import getattr |
|
38 | from .pycompat import getattr | |
39 | from .revlogutils.constants import ( |
|
39 | from .revlogutils.constants import ( | |
40 | FLAG_GENERALDELTA, |
|
40 | FLAG_GENERALDELTA, | |
41 | FLAG_INLINE_DATA, |
|
41 | FLAG_INLINE_DATA, | |
42 | REVLOGV0, |
|
42 | REVLOGV0, | |
43 | REVLOGV1, |
|
43 | REVLOGV1, | |
44 | REVLOGV1_FLAGS, |
|
44 | REVLOGV1_FLAGS, | |
45 | REVLOGV2, |
|
45 | REVLOGV2, | |
46 | REVLOGV2_FLAGS, |
|
46 | REVLOGV2_FLAGS, | |
47 | REVLOG_DEFAULT_FLAGS, |
|
47 | REVLOG_DEFAULT_FLAGS, | |
48 | REVLOG_DEFAULT_FORMAT, |
|
48 | REVLOG_DEFAULT_FORMAT, | |
49 | REVLOG_DEFAULT_VERSION, |
|
49 | REVLOG_DEFAULT_VERSION, | |
50 | ) |
|
50 | ) | |
51 | from .revlogutils.flagutil import ( |
|
51 | from .revlogutils.flagutil import ( | |
52 | REVIDX_DEFAULT_FLAGS, |
|
52 | REVIDX_DEFAULT_FLAGS, | |
53 | REVIDX_ELLIPSIS, |
|
53 | REVIDX_ELLIPSIS, | |
54 | REVIDX_EXTSTORED, |
|
54 | REVIDX_EXTSTORED, | |
55 | REVIDX_FLAGS_ORDER, |
|
55 | REVIDX_FLAGS_ORDER, | |
56 | REVIDX_HASCOPIESINFO, |
|
56 | REVIDX_HASCOPIESINFO, | |
57 | REVIDX_ISCENSORED, |
|
57 | REVIDX_ISCENSORED, | |
58 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
58 | REVIDX_RAWTEXT_CHANGING_FLAGS, | |
59 | REVIDX_SIDEDATA, |
|
59 | REVIDX_SIDEDATA, | |
60 | ) |
|
60 | ) | |
61 | from .thirdparty import attr |
|
61 | from .thirdparty import attr | |
62 | from . import ( |
|
62 | from . import ( | |
63 | ancestor, |
|
63 | ancestor, | |
64 | dagop, |
|
64 | dagop, | |
65 | error, |
|
65 | error, | |
66 | mdiff, |
|
66 | mdiff, | |
67 | policy, |
|
67 | policy, | |
68 | pycompat, |
|
68 | pycompat, | |
69 | templatefilters, |
|
69 | templatefilters, | |
70 | util, |
|
70 | util, | |
71 | ) |
|
71 | ) | |
72 | from .interfaces import ( |
|
72 | from .interfaces import ( | |
73 | repository, |
|
73 | repository, | |
74 | util as interfaceutil, |
|
74 | util as interfaceutil, | |
75 | ) |
|
75 | ) | |
76 | from .revlogutils import ( |
|
76 | from .revlogutils import ( | |
77 | deltas as deltautil, |
|
77 | deltas as deltautil, | |
78 | flagutil, |
|
78 | flagutil, | |
79 | nodemap as nodemaputil, |
|
79 | nodemap as nodemaputil, | |
80 | sidedata as sidedatautil, |
|
80 | sidedata as sidedatautil, | |
81 | ) |
|
81 | ) | |
82 | from .utils import ( |
|
82 | from .utils import ( | |
83 | storageutil, |
|
83 | storageutil, | |
84 | stringutil, |
|
84 | stringutil, | |
85 | ) |
|
85 | ) | |
86 | from .pure import parsers as pureparsers |
|
86 | from .pure import parsers as pureparsers | |
87 |
|
87 | |||
88 | # blanked usage of all the name to prevent pyflakes constraints |
|
88 | # blanked usage of all the name to prevent pyflakes constraints | |
89 | # We need these name available in the module for extensions. |
|
89 | # We need these name available in the module for extensions. | |
90 | REVLOGV0 |
|
90 | REVLOGV0 | |
91 | REVLOGV1 |
|
91 | REVLOGV1 | |
92 | REVLOGV2 |
|
92 | REVLOGV2 | |
93 | FLAG_INLINE_DATA |
|
93 | FLAG_INLINE_DATA | |
94 | FLAG_GENERALDELTA |
|
94 | FLAG_GENERALDELTA | |
95 | REVLOG_DEFAULT_FLAGS |
|
95 | REVLOG_DEFAULT_FLAGS | |
96 | REVLOG_DEFAULT_FORMAT |
|
96 | REVLOG_DEFAULT_FORMAT | |
97 | REVLOG_DEFAULT_VERSION |
|
97 | REVLOG_DEFAULT_VERSION | |
98 | REVLOGV1_FLAGS |
|
98 | REVLOGV1_FLAGS | |
99 | REVLOGV2_FLAGS |
|
99 | REVLOGV2_FLAGS | |
100 | REVIDX_ISCENSORED |
|
100 | REVIDX_ISCENSORED | |
101 | REVIDX_ELLIPSIS |
|
101 | REVIDX_ELLIPSIS | |
102 | REVIDX_SIDEDATA |
|
102 | REVIDX_SIDEDATA | |
103 | REVIDX_HASCOPIESINFO |
|
103 | REVIDX_HASCOPIESINFO | |
104 | REVIDX_EXTSTORED |
|
104 | REVIDX_EXTSTORED | |
105 | REVIDX_DEFAULT_FLAGS |
|
105 | REVIDX_DEFAULT_FLAGS | |
106 | REVIDX_FLAGS_ORDER |
|
106 | REVIDX_FLAGS_ORDER | |
107 | REVIDX_RAWTEXT_CHANGING_FLAGS |
|
107 | REVIDX_RAWTEXT_CHANGING_FLAGS | |
108 |
|
108 | |||
109 | parsers = policy.importmod('parsers') |
|
109 | parsers = policy.importmod('parsers') | |
110 | rustancestor = policy.importrust('ancestor') |
|
110 | rustancestor = policy.importrust('ancestor') | |
111 | rustdagop = policy.importrust('dagop') |
|
111 | rustdagop = policy.importrust('dagop') | |
112 | rustrevlog = policy.importrust('revlog') |
|
112 | rustrevlog = policy.importrust('revlog') | |
113 |
|
113 | |||
114 | # Aliased for performance. |
|
114 | # Aliased for performance. | |
115 | _zlibdecompress = zlib.decompress |
|
115 | _zlibdecompress = zlib.decompress | |
116 |
|
116 | |||
117 | # max size of revlog with inline data |
|
117 | # max size of revlog with inline data | |
118 | _maxinline = 131072 |
|
118 | _maxinline = 131072 | |
119 | _chunksize = 1048576 |
|
119 | _chunksize = 1048576 | |
120 |
|
120 | |||
121 | # Flag processors for REVIDX_ELLIPSIS. |
|
121 | # Flag processors for REVIDX_ELLIPSIS. | |
122 | def ellipsisreadprocessor(rl, text): |
|
122 | def ellipsisreadprocessor(rl, text): | |
123 | return text, False |
|
123 | return text, False | |
124 |
|
124 | |||
125 |
|
125 | |||
126 | def ellipsiswriteprocessor(rl, text): |
|
126 | def ellipsiswriteprocessor(rl, text): | |
127 | return text, False |
|
127 | return text, False | |
128 |
|
128 | |||
129 |
|
129 | |||
130 | def ellipsisrawprocessor(rl, text): |
|
130 | def ellipsisrawprocessor(rl, text): | |
131 | return False |
|
131 | return False | |
132 |
|
132 | |||
133 |
|
133 | |||
134 | ellipsisprocessor = ( |
|
134 | ellipsisprocessor = ( | |
135 | ellipsisreadprocessor, |
|
135 | ellipsisreadprocessor, | |
136 | ellipsiswriteprocessor, |
|
136 | ellipsiswriteprocessor, | |
137 | ellipsisrawprocessor, |
|
137 | ellipsisrawprocessor, | |
138 | ) |
|
138 | ) | |
139 |
|
139 | |||
140 |
|
140 | |||
141 | def getoffset(q): |
|
141 | def getoffset(q): | |
142 | return int(q >> 16) |
|
142 | return int(q >> 16) | |
143 |
|
143 | |||
144 |
|
144 | |||
145 | def gettype(q): |
|
145 | def gettype(q): | |
146 | return int(q & 0xFFFF) |
|
146 | return int(q & 0xFFFF) | |
147 |
|
147 | |||
148 |
|
148 | |||
149 | def offset_type(offset, type): |
|
149 | def offset_type(offset, type): | |
150 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: |
|
150 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: | |
151 | raise ValueError(b'unknown revlog index flags') |
|
151 | raise ValueError(b'unknown revlog index flags') | |
152 | return int(int(offset) << 16 | type) |
|
152 | return int(int(offset) << 16 | type) | |
153 |
|
153 | |||
154 |
|
154 | |||
155 | def _verify_revision(rl, skipflags, state, node): |
|
155 | def _verify_revision(rl, skipflags, state, node): | |
156 | """Verify the integrity of the given revlog ``node`` while providing a hook |
|
156 | """Verify the integrity of the given revlog ``node`` while providing a hook | |
157 | point for extensions to influence the operation.""" |
|
157 | point for extensions to influence the operation.""" | |
158 | if skipflags: |
|
158 | if skipflags: | |
159 | state[b'skipread'].add(node) |
|
159 | state[b'skipread'].add(node) | |
160 | else: |
|
160 | else: | |
161 | # Side-effect: read content and verify hash. |
|
161 | # Side-effect: read content and verify hash. | |
162 | rl.revision(node) |
|
162 | rl.revision(node) | |
163 |
|
163 | |||
164 |
|
164 | |||
165 | # True if a fast implementation for persistent-nodemap is available |
|
165 | # True if a fast implementation for persistent-nodemap is available | |
166 | # |
|
166 | # | |
167 | # We also consider we have a "fast" implementation in "pure" python because |
|
167 | # We also consider we have a "fast" implementation in "pure" python because | |
168 | # people using pure don't really have performance consideration (and a |
|
168 | # people using pure don't really have performance consideration (and a | |
169 | # wheelbarrow of other slowness source) |
|
169 | # wheelbarrow of other slowness source) | |
170 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or util.safehasattr( |
|
170 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or util.safehasattr( | |
171 | parsers, 'BaseIndexObject' |
|
171 | parsers, 'BaseIndexObject' | |
172 | ) |
|
172 | ) | |
173 |
|
173 | |||
174 |
|
174 | |||
175 | @attr.s(slots=True, frozen=True) |
|
175 | @attr.s(slots=True, frozen=True) | |
176 | class _revisioninfo(object): |
|
176 | class _revisioninfo(object): | |
177 | """Information about a revision that allows building its fulltext |
|
177 | """Information about a revision that allows building its fulltext | |
178 | node: expected hash of the revision |
|
178 | node: expected hash of the revision | |
179 | p1, p2: parent revs of the revision |
|
179 | p1, p2: parent revs of the revision | |
180 | btext: built text cache consisting of a one-element list |
|
180 | btext: built text cache consisting of a one-element list | |
181 | cachedelta: (baserev, uncompressed_delta) or None |
|
181 | cachedelta: (baserev, uncompressed_delta) or None | |
182 | flags: flags associated to the revision storage |
|
182 | flags: flags associated to the revision storage | |
183 |
|
183 | |||
184 | One of btext[0] or cachedelta must be set. |
|
184 | One of btext[0] or cachedelta must be set. | |
185 | """ |
|
185 | """ | |
186 |
|
186 | |||
187 | node = attr.ib() |
|
187 | node = attr.ib() | |
188 | p1 = attr.ib() |
|
188 | p1 = attr.ib() | |
189 | p2 = attr.ib() |
|
189 | p2 = attr.ib() | |
190 | btext = attr.ib() |
|
190 | btext = attr.ib() | |
191 | textlen = attr.ib() |
|
191 | textlen = attr.ib() | |
192 | cachedelta = attr.ib() |
|
192 | cachedelta = attr.ib() | |
193 | flags = attr.ib() |
|
193 | flags = attr.ib() | |
194 |
|
194 | |||
195 |
|
195 | |||
196 | @interfaceutil.implementer(repository.irevisiondelta) |
|
196 | @interfaceutil.implementer(repository.irevisiondelta) | |
197 | @attr.s(slots=True) |
|
197 | @attr.s(slots=True) | |
198 | class revlogrevisiondelta(object): |
|
198 | class revlogrevisiondelta(object): | |
199 | node = attr.ib() |
|
199 | node = attr.ib() | |
200 | p1node = attr.ib() |
|
200 | p1node = attr.ib() | |
201 | p2node = attr.ib() |
|
201 | p2node = attr.ib() | |
202 | basenode = attr.ib() |
|
202 | basenode = attr.ib() | |
203 | flags = attr.ib() |
|
203 | flags = attr.ib() | |
204 | baserevisionsize = attr.ib() |
|
204 | baserevisionsize = attr.ib() | |
205 | revision = attr.ib() |
|
205 | revision = attr.ib() | |
206 | delta = attr.ib() |
|
206 | delta = attr.ib() | |
207 | sidedata = attr.ib() |
|
207 | sidedata = attr.ib() | |
208 | linknode = attr.ib(default=None) |
|
208 | linknode = attr.ib(default=None) | |
209 |
|
209 | |||
210 |
|
210 | |||
211 | @interfaceutil.implementer(repository.iverifyproblem) |
|
211 | @interfaceutil.implementer(repository.iverifyproblem) | |
212 | @attr.s(frozen=True) |
|
212 | @attr.s(frozen=True) | |
213 | class revlogproblem(object): |
|
213 | class revlogproblem(object): | |
214 | warning = attr.ib(default=None) |
|
214 | warning = attr.ib(default=None) | |
215 | error = attr.ib(default=None) |
|
215 | error = attr.ib(default=None) | |
216 | node = attr.ib(default=None) |
|
216 | node = attr.ib(default=None) | |
217 |
|
217 | |||
218 |
|
218 | |||
219 | # index v0: |
|
219 | # index v0: | |
220 | # 4 bytes: offset |
|
220 | # 4 bytes: offset | |
221 | # 4 bytes: compressed length |
|
221 | # 4 bytes: compressed length | |
222 | # 4 bytes: base rev |
|
222 | # 4 bytes: base rev | |
223 | # 4 bytes: link rev |
|
223 | # 4 bytes: link rev | |
224 | # 20 bytes: parent 1 nodeid |
|
224 | # 20 bytes: parent 1 nodeid | |
225 | # 20 bytes: parent 2 nodeid |
|
225 | # 20 bytes: parent 2 nodeid | |
226 | # 20 bytes: nodeid |
|
226 | # 20 bytes: nodeid | |
227 | indexformatv0 = struct.Struct(b">4l20s20s20s") |
|
227 | indexformatv0 = struct.Struct(b">4l20s20s20s") | |
228 | indexformatv0_pack = indexformatv0.pack |
|
228 | indexformatv0_pack = indexformatv0.pack | |
229 | indexformatv0_unpack = indexformatv0.unpack |
|
229 | indexformatv0_unpack = indexformatv0.unpack | |
230 |
|
230 | |||
231 |
|
231 | |||
232 | class revlogoldindex(list): |
|
232 | class revlogoldindex(list): | |
233 | @property |
|
233 | @property | |
234 | def nodemap(self): |
|
234 | def nodemap(self): | |
235 | msg = b"index.nodemap is deprecated, use index.[has_node|rev|get_rev]" |
|
235 | msg = b"index.nodemap is deprecated, use index.[has_node|rev|get_rev]" | |
236 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
236 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) | |
237 | return self._nodemap |
|
237 | return self._nodemap | |
238 |
|
238 | |||
239 | @util.propertycache |
|
239 | @util.propertycache | |
240 | def _nodemap(self): |
|
240 | def _nodemap(self): | |
241 | nodemap = nodemaputil.NodeMap({nullid: nullrev}) |
|
241 | nodemap = nodemaputil.NodeMap({nullid: nullrev}) | |
242 | for r in range(0, len(self)): |
|
242 | for r in range(0, len(self)): | |
243 | n = self[r][7] |
|
243 | n = self[r][7] | |
244 | nodemap[n] = r |
|
244 | nodemap[n] = r | |
245 | return nodemap |
|
245 | return nodemap | |
246 |
|
246 | |||
247 | def has_node(self, node): |
|
247 | def has_node(self, node): | |
248 | """return True if the node exist in the index""" |
|
248 | """return True if the node exist in the index""" | |
249 | return node in self._nodemap |
|
249 | return node in self._nodemap | |
250 |
|
250 | |||
251 | def rev(self, node): |
|
251 | def rev(self, node): | |
252 | """return a revision for a node |
|
252 | """return a revision for a node | |
253 |
|
253 | |||
254 | If the node is unknown, raise a RevlogError""" |
|
254 | If the node is unknown, raise a RevlogError""" | |
255 | return self._nodemap[node] |
|
255 | return self._nodemap[node] | |
256 |
|
256 | |||
257 | def get_rev(self, node): |
|
257 | def get_rev(self, node): | |
258 | """return a revision for a node |
|
258 | """return a revision for a node | |
259 |
|
259 | |||
260 | If the node is unknown, return None""" |
|
260 | If the node is unknown, return None""" | |
261 | return self._nodemap.get(node) |
|
261 | return self._nodemap.get(node) | |
262 |
|
262 | |||
263 | def append(self, tup): |
|
263 | def append(self, tup): | |
264 | self._nodemap[tup[7]] = len(self) |
|
264 | self._nodemap[tup[7]] = len(self) | |
265 | super(revlogoldindex, self).append(tup) |
|
265 | super(revlogoldindex, self).append(tup) | |
266 |
|
266 | |||
267 | def __delitem__(self, i): |
|
267 | def __delitem__(self, i): | |
268 | if not isinstance(i, slice) or not i.stop == -1 or i.step is not None: |
|
268 | if not isinstance(i, slice) or not i.stop == -1 or i.step is not None: | |
269 | raise ValueError(b"deleting slices only supports a:-1 with step 1") |
|
269 | raise ValueError(b"deleting slices only supports a:-1 with step 1") | |
270 | for r in pycompat.xrange(i.start, len(self)): |
|
270 | for r in pycompat.xrange(i.start, len(self)): | |
271 | del self._nodemap[self[r][7]] |
|
271 | del self._nodemap[self[r][7]] | |
272 | super(revlogoldindex, self).__delitem__(i) |
|
272 | super(revlogoldindex, self).__delitem__(i) | |
273 |
|
273 | |||
274 | def clearcaches(self): |
|
274 | def clearcaches(self): | |
275 | self.__dict__.pop('_nodemap', None) |
|
275 | self.__dict__.pop('_nodemap', None) | |
276 |
|
276 | |||
277 | def __getitem__(self, i): |
|
277 | def __getitem__(self, i): | |
278 | if i == -1: |
|
278 | if i == -1: | |
279 | return (0, 0, 0, -1, -1, -1, -1, nullid) |
|
279 | return (0, 0, 0, -1, -1, -1, -1, nullid) | |
280 | return list.__getitem__(self, i) |
|
280 | return list.__getitem__(self, i) | |
281 |
|
281 | |||
282 |
|
282 | |||
283 | class revlogoldio(object): |
|
283 | class revlogoldio(object): | |
284 | def __init__(self): |
|
284 | def __init__(self): | |
285 | self.size = indexformatv0.size |
|
285 | self.size = indexformatv0.size | |
286 |
|
286 | |||
287 | def parseindex(self, data, inline): |
|
287 | def parseindex(self, data, inline): | |
288 | s = self.size |
|
288 | s = self.size | |
289 | index = [] |
|
289 | index = [] | |
290 | nodemap = nodemaputil.NodeMap({nullid: nullrev}) |
|
290 | nodemap = nodemaputil.NodeMap({nullid: nullrev}) | |
291 | n = off = 0 |
|
291 | n = off = 0 | |
292 | l = len(data) |
|
292 | l = len(data) | |
293 | while off + s <= l: |
|
293 | while off + s <= l: | |
294 | cur = data[off : off + s] |
|
294 | cur = data[off : off + s] | |
295 | off += s |
|
295 | off += s | |
296 | e = indexformatv0_unpack(cur) |
|
296 | e = indexformatv0_unpack(cur) | |
297 | # transform to revlogv1 format |
|
297 | # transform to revlogv1 format | |
298 | e2 = ( |
|
298 | e2 = ( | |
299 | offset_type(e[0], 0), |
|
299 | offset_type(e[0], 0), | |
300 | e[1], |
|
300 | e[1], | |
301 | -1, |
|
301 | -1, | |
302 | e[2], |
|
302 | e[2], | |
303 | e[3], |
|
303 | e[3], | |
304 | nodemap.get(e[4], nullrev), |
|
304 | nodemap.get(e[4], nullrev), | |
305 | nodemap.get(e[5], nullrev), |
|
305 | nodemap.get(e[5], nullrev), | |
306 | e[6], |
|
306 | e[6], | |
307 | ) |
|
307 | ) | |
308 | index.append(e2) |
|
308 | index.append(e2) | |
309 | nodemap[e[6]] = n |
|
309 | nodemap[e[6]] = n | |
310 | n += 1 |
|
310 | n += 1 | |
311 |
|
311 | |||
312 | index = revlogoldindex(index) |
|
312 | index = revlogoldindex(index) | |
313 | return index, None |
|
313 | return index, None | |
314 |
|
314 | |||
315 | def packentry(self, entry, node, version, rev): |
|
315 | def packentry(self, entry, node, version, rev): | |
316 | if gettype(entry[0]): |
|
316 | if gettype(entry[0]): | |
317 | raise error.RevlogError( |
|
317 | raise error.RevlogError( | |
318 | _(b'index entry flags need revlog version 1') |
|
318 | _(b'index entry flags need revlog version 1') | |
319 | ) |
|
319 | ) | |
320 | e2 = ( |
|
320 | e2 = ( | |
321 | getoffset(entry[0]), |
|
321 | getoffset(entry[0]), | |
322 | entry[1], |
|
322 | entry[1], | |
323 | entry[3], |
|
323 | entry[3], | |
324 | entry[4], |
|
324 | entry[4], | |
325 | node(entry[5]), |
|
325 | node(entry[5]), | |
326 | node(entry[6]), |
|
326 | node(entry[6]), | |
327 | entry[7], |
|
327 | entry[7], | |
328 | ) |
|
328 | ) | |
329 | return indexformatv0_pack(*e2) |
|
329 | return indexformatv0_pack(*e2) | |
330 |
|
330 | |||
331 |
|
331 | |||
332 | # index ng: |
|
332 | # index ng: | |
333 | # 6 bytes: offset |
|
333 | # 6 bytes: offset | |
334 | # 2 bytes: flags |
|
334 | # 2 bytes: flags | |
335 | # 4 bytes: compressed length |
|
335 | # 4 bytes: compressed length | |
336 | # 4 bytes: uncompressed length |
|
336 | # 4 bytes: uncompressed length | |
337 | # 4 bytes: base rev |
|
337 | # 4 bytes: base rev | |
338 | # 4 bytes: link rev |
|
338 | # 4 bytes: link rev | |
339 | # 4 bytes: parent 1 rev |
|
339 | # 4 bytes: parent 1 rev | |
340 | # 4 bytes: parent 2 rev |
|
340 | # 4 bytes: parent 2 rev | |
341 | # 32 bytes: nodeid |
|
341 | # 32 bytes: nodeid | |
342 | indexformatng = struct.Struct(b">Qiiiiii20s12x") |
|
342 | indexformatng = struct.Struct(b">Qiiiiii20s12x") | |
343 | indexformatng_pack = indexformatng.pack |
|
343 | indexformatng_pack = indexformatng.pack | |
344 | versionformat = struct.Struct(b">I") |
|
344 | versionformat = struct.Struct(b">I") | |
345 | versionformat_pack = versionformat.pack |
|
345 | versionformat_pack = versionformat.pack | |
346 | versionformat_unpack = versionformat.unpack |
|
346 | versionformat_unpack = versionformat.unpack | |
347 |
|
347 | |||
348 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte |
|
348 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte | |
349 | # signed integer) |
|
349 | # signed integer) | |
350 | _maxentrysize = 0x7FFFFFFF |
|
350 | _maxentrysize = 0x7FFFFFFF | |
351 |
|
351 | |||
352 |
|
352 | |||
353 | class revlogio(object): |
|
353 | class revlogio(object): | |
354 | def __init__(self): |
|
354 | def __init__(self): | |
355 | self.size = indexformatng.size |
|
355 | self.size = indexformatng.size | |
356 |
|
356 | |||
357 | def parseindex(self, data, inline): |
|
357 | def parseindex(self, data, inline): | |
358 | # call the C implementation to parse the index data |
|
358 | # call the C implementation to parse the index data | |
359 | index, cache = parsers.parse_index2(data, inline) |
|
359 | index, cache = parsers.parse_index2(data, inline) | |
360 | return index, cache |
|
360 | return index, cache | |
361 |
|
361 | |||
362 | def packentry(self, entry, node, version, rev): |
|
362 | def packentry(self, entry, node, version, rev): | |
363 | p = indexformatng_pack(*entry) |
|
363 | p = indexformatng_pack(*entry) | |
364 | if rev == 0: |
|
364 | if rev == 0: | |
365 | p = versionformat_pack(version) + p[4:] |
|
365 | p = versionformat_pack(version) + p[4:] | |
366 | return p |
|
366 | return p | |
367 |
|
367 | |||
368 |
|
368 | |||
369 | indexformatv2 = struct.Struct(pureparsers.Index2Mixin.index_format) |
|
369 | indexformatv2 = struct.Struct(pureparsers.Index2Mixin.index_format) | |
370 | indexformatv2_pack = indexformatv2.pack |
|
370 | indexformatv2_pack = indexformatv2.pack | |
371 |
|
371 | |||
372 |
|
372 | |||
373 | class revlogv2io(object): |
|
373 | class revlogv2io(object): | |
374 | def __init__(self): |
|
374 | def __init__(self): | |
375 | self.size = indexformatv2.size |
|
375 | self.size = indexformatv2.size | |
376 |
|
376 | |||
377 | def parseindex(self, data, inline): |
|
377 | def parseindex(self, data, inline): | |
378 | index, cache = parsers.parse_index2(data, inline, revlogv2=True) |
|
378 | index, cache = parsers.parse_index2(data, inline, revlogv2=True) | |
379 | return index, cache |
|
379 | return index, cache | |
380 |
|
380 | |||
381 | def packentry(self, entry, node, version, rev): |
|
381 | def packentry(self, entry, node, version, rev): | |
382 | p = indexformatv2_pack(*entry) |
|
382 | p = indexformatv2_pack(*entry) | |
383 | if rev == 0: |
|
383 | if rev == 0: | |
384 | p = versionformat_pack(version) + p[4:] |
|
384 | p = versionformat_pack(version) + p[4:] | |
385 | return p |
|
385 | return p | |
386 |
|
386 | |||
387 |
|
387 | |||
388 | NodemapRevlogIO = None |
|
388 | NodemapRevlogIO = None | |
389 |
|
389 | |||
390 | if util.safehasattr(parsers, 'parse_index_devel_nodemap'): |
|
390 | if util.safehasattr(parsers, 'parse_index_devel_nodemap'): | |
391 |
|
391 | |||
392 | class NodemapRevlogIO(revlogio): |
|
392 | class NodemapRevlogIO(revlogio): | |
393 | """A debug oriented IO class that return a PersistentNodeMapIndexObject |
|
393 | """A debug oriented IO class that return a PersistentNodeMapIndexObject | |
394 |
|
394 | |||
395 | The PersistentNodeMapIndexObject object is meant to test the persistent nodemap feature. |
|
395 | The PersistentNodeMapIndexObject object is meant to test the persistent nodemap feature. | |
396 | """ |
|
396 | """ | |
397 |
|
397 | |||
398 | def parseindex(self, data, inline): |
|
398 | def parseindex(self, data, inline): | |
399 | index, cache = parsers.parse_index_devel_nodemap(data, inline) |
|
399 | index, cache = parsers.parse_index_devel_nodemap(data, inline) | |
400 | return index, cache |
|
400 | return index, cache | |
401 |
|
401 | |||
402 |
|
402 | |||
403 | class rustrevlogio(revlogio): |
|
403 | class rustrevlogio(revlogio): | |
404 | def parseindex(self, data, inline): |
|
404 | def parseindex(self, data, inline): | |
405 | index, cache = super(rustrevlogio, self).parseindex(data, inline) |
|
405 | index, cache = super(rustrevlogio, self).parseindex(data, inline) | |
406 | return rustrevlog.MixedIndex(index), cache |
|
406 | return rustrevlog.MixedIndex(index), cache | |
407 |
|
407 | |||
408 |
|
408 | |||
409 | class revlog(object): |
|
409 | class revlog(object): | |
410 | """ |
|
410 | """ | |
411 | the underlying revision storage object |
|
411 | the underlying revision storage object | |
412 |
|
412 | |||
413 | A revlog consists of two parts, an index and the revision data. |
|
413 | A revlog consists of two parts, an index and the revision data. | |
414 |
|
414 | |||
415 | The index is a file with a fixed record size containing |
|
415 | The index is a file with a fixed record size containing | |
416 | information on each revision, including its nodeid (hash), the |
|
416 | information on each revision, including its nodeid (hash), the | |
417 | nodeids of its parents, the position and offset of its data within |
|
417 | nodeids of its parents, the position and offset of its data within | |
418 | the data file, and the revision it's based on. Finally, each entry |
|
418 | the data file, and the revision it's based on. Finally, each entry | |
419 | contains a linkrev entry that can serve as a pointer to external |
|
419 | contains a linkrev entry that can serve as a pointer to external | |
420 | data. |
|
420 | data. | |
421 |
|
421 | |||
422 | The revision data itself is a linear collection of data chunks. |
|
422 | The revision data itself is a linear collection of data chunks. | |
423 | Each chunk represents a revision and is usually represented as a |
|
423 | Each chunk represents a revision and is usually represented as a | |
424 | delta against the previous chunk. To bound lookup time, runs of |
|
424 | delta against the previous chunk. To bound lookup time, runs of | |
425 | deltas are limited to about 2 times the length of the original |
|
425 | deltas are limited to about 2 times the length of the original | |
426 | version data. This makes retrieval of a version proportional to |
|
426 | version data. This makes retrieval of a version proportional to | |
427 | its size, or O(1) relative to the number of revisions. |
|
427 | its size, or O(1) relative to the number of revisions. | |
428 |
|
428 | |||
429 | Both pieces of the revlog are written to in an append-only |
|
429 | Both pieces of the revlog are written to in an append-only | |
430 | fashion, which means we never need to rewrite a file to insert or |
|
430 | fashion, which means we never need to rewrite a file to insert or | |
431 | remove data, and can use some simple techniques to avoid the need |
|
431 | remove data, and can use some simple techniques to avoid the need | |
432 | for locking while reading. |
|
432 | for locking while reading. | |
433 |
|
433 | |||
434 | If checkambig, indexfile is opened with checkambig=True at |
|
434 | If checkambig, indexfile is opened with checkambig=True at | |
435 | writing, to avoid file stat ambiguity. |
|
435 | writing, to avoid file stat ambiguity. | |
436 |
|
436 | |||
437 | If mmaplargeindex is True, and an mmapindexthreshold is set, the |
|
437 | If mmaplargeindex is True, and an mmapindexthreshold is set, the | |
438 | index will be mmapped rather than read if it is larger than the |
|
438 | index will be mmapped rather than read if it is larger than the | |
439 | configured threshold. |
|
439 | configured threshold. | |
440 |
|
440 | |||
441 | If censorable is True, the revlog can have censored revisions. |
|
441 | If censorable is True, the revlog can have censored revisions. | |
442 |
|
442 | |||
443 | If `upperboundcomp` is not None, this is the expected maximal gain from |
|
443 | If `upperboundcomp` is not None, this is the expected maximal gain from | |
444 | compression for the data content. |
|
444 | compression for the data content. | |
445 |
|
445 | |||
446 | `concurrencychecker` is an optional function that receives 3 arguments: a |
|
446 | `concurrencychecker` is an optional function that receives 3 arguments: a | |
447 | file handle, a filename, and an expected position. It should check whether |
|
447 | file handle, a filename, and an expected position. It should check whether | |
448 | the current position in the file handle is valid, and log/warn/fail (by |
|
448 | the current position in the file handle is valid, and log/warn/fail (by | |
449 | raising). |
|
449 | raising). | |
450 | """ |
|
450 | """ | |
451 |
|
451 | |||
452 | _flagserrorclass = error.RevlogError |
|
452 | _flagserrorclass = error.RevlogError | |
453 |
|
453 | |||
454 | def __init__( |
|
454 | def __init__( | |
455 | self, |
|
455 | self, | |
456 | opener, |
|
456 | opener, | |
457 | indexfile, |
|
457 | indexfile, | |
458 | datafile=None, |
|
458 | datafile=None, | |
459 | checkambig=False, |
|
459 | checkambig=False, | |
460 | mmaplargeindex=False, |
|
460 | mmaplargeindex=False, | |
461 | censorable=False, |
|
461 | censorable=False, | |
462 | upperboundcomp=None, |
|
462 | upperboundcomp=None, | |
463 | persistentnodemap=False, |
|
463 | persistentnodemap=False, | |
464 | concurrencychecker=None, |
|
464 | concurrencychecker=None, | |
465 | ): |
|
465 | ): | |
466 | """ |
|
466 | """ | |
467 | create a revlog object |
|
467 | create a revlog object | |
468 |
|
468 | |||
469 | opener is a function that abstracts the file opening operation |
|
469 | opener is a function that abstracts the file opening operation | |
470 | and can be used to implement COW semantics or the like. |
|
470 | and can be used to implement COW semantics or the like. | |
471 |
|
471 | |||
472 | """ |
|
472 | """ | |
473 | self.upperboundcomp = upperboundcomp |
|
473 | self.upperboundcomp = upperboundcomp | |
474 | self.indexfile = indexfile |
|
474 | self.indexfile = indexfile | |
475 | self.datafile = datafile or (indexfile[:-2] + b".d") |
|
475 | self.datafile = datafile or (indexfile[:-2] + b".d") | |
476 | self.nodemap_file = None |
|
476 | self.nodemap_file = None | |
477 | if persistentnodemap: |
|
477 | if persistentnodemap: | |
478 | self.nodemap_file = nodemaputil.get_nodemap_file( |
|
478 | self.nodemap_file = nodemaputil.get_nodemap_file( | |
479 | opener, self.indexfile |
|
479 | opener, self.indexfile | |
480 | ) |
|
480 | ) | |
481 |
|
481 | |||
482 | self.opener = opener |
|
482 | self.opener = opener | |
483 | # When True, indexfile is opened with checkambig=True at writing, to |
|
483 | # When True, indexfile is opened with checkambig=True at writing, to | |
484 | # avoid file stat ambiguity. |
|
484 | # avoid file stat ambiguity. | |
485 | self._checkambig = checkambig |
|
485 | self._checkambig = checkambig | |
486 | self._mmaplargeindex = mmaplargeindex |
|
486 | self._mmaplargeindex = mmaplargeindex | |
487 | self._censorable = censorable |
|
487 | self._censorable = censorable | |
488 | # 3-tuple of (node, rev, text) for a raw revision. |
|
488 | # 3-tuple of (node, rev, text) for a raw revision. | |
489 | self._revisioncache = None |
|
489 | self._revisioncache = None | |
490 | # Maps rev to chain base rev. |
|
490 | # Maps rev to chain base rev. | |
491 | self._chainbasecache = util.lrucachedict(100) |
|
491 | self._chainbasecache = util.lrucachedict(100) | |
492 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. |
|
492 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. | |
493 | self._chunkcache = (0, b'') |
|
493 | self._chunkcache = (0, b'') | |
494 | # How much data to read and cache into the raw revlog data cache. |
|
494 | # How much data to read and cache into the raw revlog data cache. | |
495 | self._chunkcachesize = 65536 |
|
495 | self._chunkcachesize = 65536 | |
496 | self._maxchainlen = None |
|
496 | self._maxchainlen = None | |
497 | self._deltabothparents = True |
|
497 | self._deltabothparents = True | |
498 | self.index = None |
|
498 | self.index = None | |
499 | self._nodemap_docket = None |
|
499 | self._nodemap_docket = None | |
500 | # Mapping of partial identifiers to full nodes. |
|
500 | # Mapping of partial identifiers to full nodes. | |
501 | self._pcache = {} |
|
501 | self._pcache = {} | |
502 | # Mapping of revision integer to full node. |
|
502 | # Mapping of revision integer to full node. | |
503 | self._compengine = b'zlib' |
|
503 | self._compengine = b'zlib' | |
504 | self._compengineopts = {} |
|
504 | self._compengineopts = {} | |
505 | self._maxdeltachainspan = -1 |
|
505 | self._maxdeltachainspan = -1 | |
506 | self._withsparseread = False |
|
506 | self._withsparseread = False | |
507 | self._sparserevlog = False |
|
507 | self._sparserevlog = False | |
508 | self._srdensitythreshold = 0.50 |
|
508 | self._srdensitythreshold = 0.50 | |
509 | self._srmingapsize = 262144 |
|
509 | self._srmingapsize = 262144 | |
510 |
|
510 | |||
511 | # Make copy of flag processors so each revlog instance can support |
|
511 | # Make copy of flag processors so each revlog instance can support | |
512 | # custom flags. |
|
512 | # custom flags. | |
513 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
513 | self._flagprocessors = dict(flagutil.flagprocessors) | |
514 |
|
514 | |||
515 | # 2-tuple of file handles being used for active writing. |
|
515 | # 2-tuple of file handles being used for active writing. | |
516 | self._writinghandles = None |
|
516 | self._writinghandles = None | |
517 |
|
517 | |||
518 | self._loadindex() |
|
518 | self._loadindex() | |
519 |
|
519 | |||
520 | self._concurrencychecker = concurrencychecker |
|
520 | self._concurrencychecker = concurrencychecker | |
521 |
|
521 | |||
522 | def _loadindex(self): |
|
522 | def _loadindex(self): | |
523 | mmapindexthreshold = None |
|
523 | mmapindexthreshold = None | |
524 | opts = self.opener.options |
|
524 | opts = self.opener.options | |
525 |
|
525 | |||
526 | if b'revlogv2' in opts: |
|
526 | if b'revlogv2' in opts: | |
527 | newversionflags = REVLOGV2 | FLAG_INLINE_DATA |
|
527 | newversionflags = REVLOGV2 | FLAG_INLINE_DATA | |
528 | elif b'revlogv1' in opts: |
|
528 | elif b'revlogv1' in opts: | |
529 | newversionflags = REVLOGV1 | FLAG_INLINE_DATA |
|
529 | newversionflags = REVLOGV1 | FLAG_INLINE_DATA | |
530 | if b'generaldelta' in opts: |
|
530 | if b'generaldelta' in opts: | |
531 | newversionflags |= FLAG_GENERALDELTA |
|
531 | newversionflags |= FLAG_GENERALDELTA | |
532 | elif b'revlogv0' in self.opener.options: |
|
532 | elif b'revlogv0' in self.opener.options: | |
533 | newversionflags = REVLOGV0 |
|
533 | newversionflags = REVLOGV0 | |
534 | else: |
|
534 | else: | |
535 | newversionflags = REVLOG_DEFAULT_VERSION |
|
535 | newversionflags = REVLOG_DEFAULT_VERSION | |
536 |
|
536 | |||
537 | if b'chunkcachesize' in opts: |
|
537 | if b'chunkcachesize' in opts: | |
538 | self._chunkcachesize = opts[b'chunkcachesize'] |
|
538 | self._chunkcachesize = opts[b'chunkcachesize'] | |
539 | if b'maxchainlen' in opts: |
|
539 | if b'maxchainlen' in opts: | |
540 | self._maxchainlen = opts[b'maxchainlen'] |
|
540 | self._maxchainlen = opts[b'maxchainlen'] | |
541 | if b'deltabothparents' in opts: |
|
541 | if b'deltabothparents' in opts: | |
542 | self._deltabothparents = opts[b'deltabothparents'] |
|
542 | self._deltabothparents = opts[b'deltabothparents'] | |
543 | self._lazydelta = bool(opts.get(b'lazydelta', True)) |
|
543 | self._lazydelta = bool(opts.get(b'lazydelta', True)) | |
544 | self._lazydeltabase = False |
|
544 | self._lazydeltabase = False | |
545 | if self._lazydelta: |
|
545 | if self._lazydelta: | |
546 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) |
|
546 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) | |
547 | if b'compengine' in opts: |
|
547 | if b'compengine' in opts: | |
548 | self._compengine = opts[b'compengine'] |
|
548 | self._compengine = opts[b'compengine'] | |
549 | if b'zlib.level' in opts: |
|
549 | if b'zlib.level' in opts: | |
550 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] |
|
550 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] | |
551 | if b'zstd.level' in opts: |
|
551 | if b'zstd.level' in opts: | |
552 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] |
|
552 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] | |
553 | if b'maxdeltachainspan' in opts: |
|
553 | if b'maxdeltachainspan' in opts: | |
554 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] |
|
554 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] | |
555 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: |
|
555 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: | |
556 | mmapindexthreshold = opts[b'mmapindexthreshold'] |
|
556 | mmapindexthreshold = opts[b'mmapindexthreshold'] | |
557 | self.hassidedata = bool(opts.get(b'side-data', False)) |
|
557 | self.hassidedata = bool(opts.get(b'side-data', False)) | |
558 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) |
|
558 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) | |
559 | withsparseread = bool(opts.get(b'with-sparse-read', False)) |
|
559 | withsparseread = bool(opts.get(b'with-sparse-read', False)) | |
560 | # sparse-revlog forces sparse-read |
|
560 | # sparse-revlog forces sparse-read | |
561 | self._withsparseread = self._sparserevlog or withsparseread |
|
561 | self._withsparseread = self._sparserevlog or withsparseread | |
562 | if b'sparse-read-density-threshold' in opts: |
|
562 | if b'sparse-read-density-threshold' in opts: | |
563 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] |
|
563 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] | |
564 | if b'sparse-read-min-gap-size' in opts: |
|
564 | if b'sparse-read-min-gap-size' in opts: | |
565 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] |
|
565 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] | |
566 | if opts.get(b'enableellipsis'): |
|
566 | if opts.get(b'enableellipsis'): | |
567 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor |
|
567 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor | |
568 |
|
568 | |||
569 | # revlog v0 doesn't have flag processors |
|
569 | # revlog v0 doesn't have flag processors | |
570 | for flag, processor in pycompat.iteritems( |
|
570 | for flag, processor in pycompat.iteritems( | |
571 | opts.get(b'flagprocessors', {}) |
|
571 | opts.get(b'flagprocessors', {}) | |
572 | ): |
|
572 | ): | |
573 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) |
|
573 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) | |
574 |
|
574 | |||
575 | if self._chunkcachesize <= 0: |
|
575 | if self._chunkcachesize <= 0: | |
576 | raise error.RevlogError( |
|
576 | raise error.RevlogError( | |
577 | _(b'revlog chunk cache size %r is not greater than 0') |
|
577 | _(b'revlog chunk cache size %r is not greater than 0') | |
578 | % self._chunkcachesize |
|
578 | % self._chunkcachesize | |
579 | ) |
|
579 | ) | |
580 | elif self._chunkcachesize & (self._chunkcachesize - 1): |
|
580 | elif self._chunkcachesize & (self._chunkcachesize - 1): | |
581 | raise error.RevlogError( |
|
581 | raise error.RevlogError( | |
582 | _(b'revlog chunk cache size %r is not a power of 2') |
|
582 | _(b'revlog chunk cache size %r is not a power of 2') | |
583 | % self._chunkcachesize |
|
583 | % self._chunkcachesize | |
584 | ) |
|
584 | ) | |
585 |
|
585 | |||
586 | indexdata = b'' |
|
586 | indexdata = b'' | |
587 | self._initempty = True |
|
587 | self._initempty = True | |
588 | try: |
|
588 | try: | |
589 | with self._indexfp() as f: |
|
589 | with self._indexfp() as f: | |
590 | if ( |
|
590 | if ( | |
591 | mmapindexthreshold is not None |
|
591 | mmapindexthreshold is not None | |
592 | and self.opener.fstat(f).st_size >= mmapindexthreshold |
|
592 | and self.opener.fstat(f).st_size >= mmapindexthreshold | |
593 | ): |
|
593 | ): | |
594 | # TODO: should .close() to release resources without |
|
594 | # TODO: should .close() to release resources without | |
595 | # relying on Python GC |
|
595 | # relying on Python GC | |
596 | indexdata = util.buffer(util.mmapread(f)) |
|
596 | indexdata = util.buffer(util.mmapread(f)) | |
597 | else: |
|
597 | else: | |
598 | indexdata = f.read() |
|
598 | indexdata = f.read() | |
599 | if len(indexdata) > 0: |
|
599 | if len(indexdata) > 0: | |
600 | versionflags = versionformat_unpack(indexdata[:4])[0] |
|
600 | versionflags = versionformat_unpack(indexdata[:4])[0] | |
601 | self._initempty = False |
|
601 | self._initempty = False | |
602 | else: |
|
602 | else: | |
603 | versionflags = newversionflags |
|
603 | versionflags = newversionflags | |
604 | except IOError as inst: |
|
604 | except IOError as inst: | |
605 | if inst.errno != errno.ENOENT: |
|
605 | if inst.errno != errno.ENOENT: | |
606 | raise |
|
606 | raise | |
607 |
|
607 | |||
608 | versionflags = newversionflags |
|
608 | versionflags = newversionflags | |
609 |
|
609 | |||
610 | self.version = versionflags |
|
610 | self.version = versionflags | |
611 |
|
611 | |||
612 | flags = versionflags & ~0xFFFF |
|
612 | flags = versionflags & ~0xFFFF | |
613 | fmt = versionflags & 0xFFFF |
|
613 | fmt = versionflags & 0xFFFF | |
614 |
|
614 | |||
615 | if fmt == REVLOGV0: |
|
615 | if fmt == REVLOGV0: | |
616 | if flags: |
|
616 | if flags: | |
617 | raise error.RevlogError( |
|
617 | raise error.RevlogError( | |
618 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
618 | _(b'unknown flags (%#04x) in version %d revlog %s') | |
619 | % (flags >> 16, fmt, self.indexfile) |
|
619 | % (flags >> 16, fmt, self.indexfile) | |
620 | ) |
|
620 | ) | |
621 |
|
621 | |||
622 | self._inline = False |
|
622 | self._inline = False | |
623 | self._generaldelta = False |
|
623 | self._generaldelta = False | |
624 |
|
624 | |||
625 | elif fmt == REVLOGV1: |
|
625 | elif fmt == REVLOGV1: | |
626 | if flags & ~REVLOGV1_FLAGS: |
|
626 | if flags & ~REVLOGV1_FLAGS: | |
627 | raise error.RevlogError( |
|
627 | raise error.RevlogError( | |
628 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
628 | _(b'unknown flags (%#04x) in version %d revlog %s') | |
629 | % (flags >> 16, fmt, self.indexfile) |
|
629 | % (flags >> 16, fmt, self.indexfile) | |
630 | ) |
|
630 | ) | |
631 |
|
631 | |||
632 | self._inline = versionflags & FLAG_INLINE_DATA |
|
632 | self._inline = versionflags & FLAG_INLINE_DATA | |
633 | self._generaldelta = versionflags & FLAG_GENERALDELTA |
|
633 | self._generaldelta = versionflags & FLAG_GENERALDELTA | |
634 |
|
634 | |||
635 | elif fmt == REVLOGV2: |
|
635 | elif fmt == REVLOGV2: | |
636 | if flags & ~REVLOGV2_FLAGS: |
|
636 | if flags & ~REVLOGV2_FLAGS: | |
637 | raise error.RevlogError( |
|
637 | raise error.RevlogError( | |
638 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
638 | _(b'unknown flags (%#04x) in version %d revlog %s') | |
639 | % (flags >> 16, fmt, self.indexfile) |
|
639 | % (flags >> 16, fmt, self.indexfile) | |
640 | ) |
|
640 | ) | |
641 |
|
641 | |||
642 | self._inline = versionflags & FLAG_INLINE_DATA |
|
642 | self._inline = versionflags & FLAG_INLINE_DATA | |
643 | # generaldelta implied by version 2 revlogs. |
|
643 | # generaldelta implied by version 2 revlogs. | |
644 | self._generaldelta = True |
|
644 | self._generaldelta = True | |
645 |
|
645 | |||
646 | else: |
|
646 | else: | |
647 | raise error.RevlogError( |
|
647 | raise error.RevlogError( | |
648 | _(b'unknown version (%d) in revlog %s') % (fmt, self.indexfile) |
|
648 | _(b'unknown version (%d) in revlog %s') % (fmt, self.indexfile) | |
649 | ) |
|
649 | ) | |
650 | # sparse-revlog can't be on without general-delta (issue6056) |
|
650 | # sparse-revlog can't be on without general-delta (issue6056) | |
651 | if not self._generaldelta: |
|
651 | if not self._generaldelta: | |
652 | self._sparserevlog = False |
|
652 | self._sparserevlog = False | |
653 |
|
653 | |||
654 | self._storedeltachains = True |
|
654 | self._storedeltachains = True | |
655 |
|
655 | |||
656 | devel_nodemap = ( |
|
656 | devel_nodemap = ( | |
657 | self.nodemap_file |
|
657 | self.nodemap_file | |
658 | and opts.get(b'devel-force-nodemap', False) |
|
658 | and opts.get(b'devel-force-nodemap', False) | |
659 | and NodemapRevlogIO is not None |
|
659 | and NodemapRevlogIO is not None | |
660 | ) |
|
660 | ) | |
661 |
|
661 | |||
662 | use_rust_index = False |
|
662 | use_rust_index = False | |
663 | if rustrevlog is not None: |
|
663 | if rustrevlog is not None: | |
664 | if self.nodemap_file is not None: |
|
664 | if self.nodemap_file is not None: | |
665 | use_rust_index = True |
|
665 | use_rust_index = True | |
666 | else: |
|
666 | else: | |
667 | use_rust_index = self.opener.options.get(b'rust.index') |
|
667 | use_rust_index = self.opener.options.get(b'rust.index') | |
668 |
|
668 | |||
669 | self._io = revlogio() |
|
669 | self._io = revlogio() | |
670 | if self.version == REVLOGV0: |
|
670 | if self.version == REVLOGV0: | |
671 | self._io = revlogoldio() |
|
671 | self._io = revlogoldio() | |
672 | elif fmt == REVLOGV2: |
|
672 | elif fmt == REVLOGV2: | |
673 | self._io = revlogv2io() |
|
673 | self._io = revlogv2io() | |
674 | elif devel_nodemap: |
|
674 | elif devel_nodemap: | |
675 | self._io = NodemapRevlogIO() |
|
675 | self._io = NodemapRevlogIO() | |
676 | elif use_rust_index: |
|
676 | elif use_rust_index: | |
677 | self._io = rustrevlogio() |
|
677 | self._io = rustrevlogio() | |
678 | try: |
|
678 | try: | |
679 | d = self._io.parseindex(indexdata, self._inline) |
|
679 | d = self._io.parseindex(indexdata, self._inline) | |
680 | index, _chunkcache = d |
|
680 | index, _chunkcache = d | |
681 | use_nodemap = ( |
|
681 | use_nodemap = ( | |
682 | not self._inline |
|
682 | not self._inline | |
683 | and self.nodemap_file is not None |
|
683 | and self.nodemap_file is not None | |
684 | and util.safehasattr(index, 'update_nodemap_data') |
|
684 | and util.safehasattr(index, 'update_nodemap_data') | |
685 | ) |
|
685 | ) | |
686 | if use_nodemap: |
|
686 | if use_nodemap: | |
687 | nodemap_data = nodemaputil.persisted_data(self) |
|
687 | nodemap_data = nodemaputil.persisted_data(self) | |
688 | if nodemap_data is not None: |
|
688 | if nodemap_data is not None: | |
689 | docket = nodemap_data[0] |
|
689 | docket = nodemap_data[0] | |
690 | if ( |
|
690 | if ( | |
691 | len(d[0]) > docket.tip_rev |
|
691 | len(d[0]) > docket.tip_rev | |
692 | and d[0][docket.tip_rev][7] == docket.tip_node |
|
692 | and d[0][docket.tip_rev][7] == docket.tip_node | |
693 | ): |
|
693 | ): | |
694 | # no changelog tampering |
|
694 | # no changelog tampering | |
695 | self._nodemap_docket = docket |
|
695 | self._nodemap_docket = docket | |
696 | index.update_nodemap_data(*nodemap_data) |
|
696 | index.update_nodemap_data(*nodemap_data) | |
697 | except (ValueError, IndexError): |
|
697 | except (ValueError, IndexError): | |
698 | raise error.RevlogError( |
|
698 | raise error.RevlogError( | |
699 | _(b"index %s is corrupted") % self.indexfile |
|
699 | _(b"index %s is corrupted") % self.indexfile | |
700 | ) |
|
700 | ) | |
701 | self.index, self._chunkcache = d |
|
701 | self.index, self._chunkcache = d | |
702 | if not self._chunkcache: |
|
702 | if not self._chunkcache: | |
703 | self._chunkclear() |
|
703 | self._chunkclear() | |
704 | # revnum -> (chain-length, sum-delta-length) |
|
704 | # revnum -> (chain-length, sum-delta-length) | |
705 | self._chaininfocache = util.lrucachedict(500) |
|
705 | self._chaininfocache = util.lrucachedict(500) | |
706 | # revlog header -> revlog compressor |
|
706 | # revlog header -> revlog compressor | |
707 | self._decompressors = {} |
|
707 | self._decompressors = {} | |
708 |
|
708 | |||
709 | @util.propertycache |
|
709 | @util.propertycache | |
710 | def _compressor(self): |
|
710 | def _compressor(self): | |
711 | engine = util.compengines[self._compengine] |
|
711 | engine = util.compengines[self._compengine] | |
712 | return engine.revlogcompressor(self._compengineopts) |
|
712 | return engine.revlogcompressor(self._compengineopts) | |
713 |
|
713 | |||
714 | def _indexfp(self, mode=b'r'): |
|
714 | def _indexfp(self, mode=b'r'): | |
715 | """file object for the revlog's index file""" |
|
715 | """file object for the revlog's index file""" | |
716 | args = {'mode': mode} |
|
716 | args = {'mode': mode} | |
717 | if mode != b'r': |
|
717 | if mode != b'r': | |
718 | args['checkambig'] = self._checkambig |
|
718 | args['checkambig'] = self._checkambig | |
719 | if mode == b'w': |
|
719 | if mode == b'w': | |
720 | args['atomictemp'] = True |
|
720 | args['atomictemp'] = True | |
721 | return self.opener(self.indexfile, **args) |
|
721 | return self.opener(self.indexfile, **args) | |
722 |
|
722 | |||
723 | def _datafp(self, mode=b'r'): |
|
723 | def _datafp(self, mode=b'r'): | |
724 | """file object for the revlog's data file""" |
|
724 | """file object for the revlog's data file""" | |
725 | return self.opener(self.datafile, mode=mode) |
|
725 | return self.opener(self.datafile, mode=mode) | |
726 |
|
726 | |||
727 | @contextlib.contextmanager |
|
727 | @contextlib.contextmanager | |
728 | def _datareadfp(self, existingfp=None): |
|
728 | def _datareadfp(self, existingfp=None): | |
729 | """file object suitable to read data""" |
|
729 | """file object suitable to read data""" | |
730 | # Use explicit file handle, if given. |
|
730 | # Use explicit file handle, if given. | |
731 | if existingfp is not None: |
|
731 | if existingfp is not None: | |
732 | yield existingfp |
|
732 | yield existingfp | |
733 |
|
733 | |||
734 | # Use a file handle being actively used for writes, if available. |
|
734 | # Use a file handle being actively used for writes, if available. | |
735 | # There is some danger to doing this because reads will seek the |
|
735 | # There is some danger to doing this because reads will seek the | |
736 | # file. However, _writeentry() performs a SEEK_END before all writes, |
|
736 | # file. However, _writeentry() performs a SEEK_END before all writes, | |
737 | # so we should be safe. |
|
737 | # so we should be safe. | |
738 | elif self._writinghandles: |
|
738 | elif self._writinghandles: | |
739 | if self._inline: |
|
739 | if self._inline: | |
740 | yield self._writinghandles[0] |
|
740 | yield self._writinghandles[0] | |
741 | else: |
|
741 | else: | |
742 | yield self._writinghandles[1] |
|
742 | yield self._writinghandles[1] | |
743 |
|
743 | |||
744 | # Otherwise open a new file handle. |
|
744 | # Otherwise open a new file handle. | |
745 | else: |
|
745 | else: | |
746 | if self._inline: |
|
746 | if self._inline: | |
747 | func = self._indexfp |
|
747 | func = self._indexfp | |
748 | else: |
|
748 | else: | |
749 | func = self._datafp |
|
749 | func = self._datafp | |
750 | with func() as fp: |
|
750 | with func() as fp: | |
751 | yield fp |
|
751 | yield fp | |
752 |
|
752 | |||
753 | def tiprev(self): |
|
753 | def tiprev(self): | |
754 | return len(self.index) - 1 |
|
754 | return len(self.index) - 1 | |
755 |
|
755 | |||
756 | def tip(self): |
|
756 | def tip(self): | |
757 | return self.node(self.tiprev()) |
|
757 | return self.node(self.tiprev()) | |
758 |
|
758 | |||
759 | def __contains__(self, rev): |
|
759 | def __contains__(self, rev): | |
760 | return 0 <= rev < len(self) |
|
760 | return 0 <= rev < len(self) | |
761 |
|
761 | |||
762 | def __len__(self): |
|
762 | def __len__(self): | |
763 | return len(self.index) |
|
763 | return len(self.index) | |
764 |
|
764 | |||
765 | def __iter__(self): |
|
765 | def __iter__(self): | |
766 | return iter(pycompat.xrange(len(self))) |
|
766 | return iter(pycompat.xrange(len(self))) | |
767 |
|
767 | |||
768 | def revs(self, start=0, stop=None): |
|
768 | def revs(self, start=0, stop=None): | |
769 | """iterate over all rev in this revlog (from start to stop)""" |
|
769 | """iterate over all rev in this revlog (from start to stop)""" | |
770 | return storageutil.iterrevs(len(self), start=start, stop=stop) |
|
770 | return storageutil.iterrevs(len(self), start=start, stop=stop) | |
771 |
|
771 | |||
772 | @property |
|
772 | @property | |
773 | def nodemap(self): |
|
773 | def nodemap(self): | |
774 | msg = ( |
|
774 | msg = ( | |
775 | b"revlog.nodemap is deprecated, " |
|
775 | b"revlog.nodemap is deprecated, " | |
776 | b"use revlog.index.[has_node|rev|get_rev]" |
|
776 | b"use revlog.index.[has_node|rev|get_rev]" | |
777 | ) |
|
777 | ) | |
778 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
778 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) | |
779 | return self.index.nodemap |
|
779 | return self.index.nodemap | |
780 |
|
780 | |||
781 | @property |
|
781 | @property | |
782 | def _nodecache(self): |
|
782 | def _nodecache(self): | |
783 | msg = b"revlog._nodecache is deprecated, use revlog.index.nodemap" |
|
783 | msg = b"revlog._nodecache is deprecated, use revlog.index.nodemap" | |
784 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
784 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) | |
785 | return self.index.nodemap |
|
785 | return self.index.nodemap | |
786 |
|
786 | |||
787 | def hasnode(self, node): |
|
787 | def hasnode(self, node): | |
788 | try: |
|
788 | try: | |
789 | self.rev(node) |
|
789 | self.rev(node) | |
790 | return True |
|
790 | return True | |
791 | except KeyError: |
|
791 | except KeyError: | |
792 | return False |
|
792 | return False | |
793 |
|
793 | |||
794 | def candelta(self, baserev, rev): |
|
794 | def candelta(self, baserev, rev): | |
795 | """whether two revisions (baserev, rev) can be delta-ed or not""" |
|
795 | """whether two revisions (baserev, rev) can be delta-ed or not""" | |
796 | # Disable delta if either rev requires a content-changing flag |
|
796 | # Disable delta if either rev requires a content-changing flag | |
797 | # processor (ex. LFS). This is because such flag processor can alter |
|
797 | # processor (ex. LFS). This is because such flag processor can alter | |
798 | # the rawtext content that the delta will be based on, and two clients |
|
798 | # the rawtext content that the delta will be based on, and two clients | |
799 | # could have a same revlog node with different flags (i.e. different |
|
799 | # could have a same revlog node with different flags (i.e. different | |
800 | # rawtext contents) and the delta could be incompatible. |
|
800 | # rawtext contents) and the delta could be incompatible. | |
801 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
801 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( | |
802 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS |
|
802 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS | |
803 | ): |
|
803 | ): | |
804 | return False |
|
804 | return False | |
805 | return True |
|
805 | return True | |
806 |
|
806 | |||
807 | def update_caches(self, transaction): |
|
807 | def update_caches(self, transaction): | |
808 | if self.nodemap_file is not None: |
|
808 | if self.nodemap_file is not None: | |
809 | if transaction is None: |
|
809 | if transaction is None: | |
810 | nodemaputil.update_persistent_nodemap(self) |
|
810 | nodemaputil.update_persistent_nodemap(self) | |
811 | else: |
|
811 | else: | |
812 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
812 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
813 |
|
813 | |||
814 | def clearcaches(self): |
|
814 | def clearcaches(self): | |
815 | self._revisioncache = None |
|
815 | self._revisioncache = None | |
816 | self._chainbasecache.clear() |
|
816 | self._chainbasecache.clear() | |
817 | self._chunkcache = (0, b'') |
|
817 | self._chunkcache = (0, b'') | |
818 | self._pcache = {} |
|
818 | self._pcache = {} | |
819 | self._nodemap_docket = None |
|
819 | self._nodemap_docket = None | |
820 | self.index.clearcaches() |
|
820 | self.index.clearcaches() | |
821 | # The python code is the one responsible for validating the docket, we |
|
821 | # The python code is the one responsible for validating the docket, we | |
822 | # end up having to refresh it here. |
|
822 | # end up having to refresh it here. | |
823 | use_nodemap = ( |
|
823 | use_nodemap = ( | |
824 | not self._inline |
|
824 | not self._inline | |
825 | and self.nodemap_file is not None |
|
825 | and self.nodemap_file is not None | |
826 | and util.safehasattr(self.index, 'update_nodemap_data') |
|
826 | and util.safehasattr(self.index, 'update_nodemap_data') | |
827 | ) |
|
827 | ) | |
828 | if use_nodemap: |
|
828 | if use_nodemap: | |
829 | nodemap_data = nodemaputil.persisted_data(self) |
|
829 | nodemap_data = nodemaputil.persisted_data(self) | |
830 | if nodemap_data is not None: |
|
830 | if nodemap_data is not None: | |
831 | self._nodemap_docket = nodemap_data[0] |
|
831 | self._nodemap_docket = nodemap_data[0] | |
832 | self.index.update_nodemap_data(*nodemap_data) |
|
832 | self.index.update_nodemap_data(*nodemap_data) | |
833 |
|
833 | |||
834 | def rev(self, node): |
|
834 | def rev(self, node): | |
835 | try: |
|
835 | try: | |
836 | return self.index.rev(node) |
|
836 | return self.index.rev(node) | |
837 | except TypeError: |
|
837 | except TypeError: | |
838 | raise |
|
838 | raise | |
839 | except error.RevlogError: |
|
839 | except error.RevlogError: | |
840 | # parsers.c radix tree lookup failed |
|
840 | # parsers.c radix tree lookup failed | |
841 | if node == wdirid or node in wdirfilenodeids: |
|
841 | if node == wdirid or node in wdirfilenodeids: | |
842 | raise error.WdirUnsupported |
|
842 | raise error.WdirUnsupported | |
843 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
843 | raise error.LookupError(node, self.indexfile, _(b'no node')) | |
844 |
|
844 | |||
845 | # Accessors for index entries. |
|
845 | # Accessors for index entries. | |
846 |
|
846 | |||
847 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes |
|
847 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes | |
848 | # are flags. |
|
848 | # are flags. | |
849 | def start(self, rev): |
|
849 | def start(self, rev): | |
850 | return int(self.index[rev][0] >> 16) |
|
850 | return int(self.index[rev][0] >> 16) | |
851 |
|
851 | |||
852 | def flags(self, rev): |
|
852 | def flags(self, rev): | |
853 | return self.index[rev][0] & 0xFFFF |
|
853 | return self.index[rev][0] & 0xFFFF | |
854 |
|
854 | |||
855 | def length(self, rev): |
|
855 | def length(self, rev): | |
856 | return self.index[rev][1] |
|
856 | return self.index[rev][1] | |
857 |
|
857 | |||
858 | def sidedata_length(self, rev): |
|
858 | def sidedata_length(self, rev): | |
859 | if self.version & 0xFFFF != REVLOGV2: |
|
859 | if self.version & 0xFFFF != REVLOGV2: | |
860 | return 0 |
|
860 | return 0 | |
861 | return self.index[rev][9] |
|
861 | return self.index[rev][9] | |
862 |
|
862 | |||
863 | def rawsize(self, rev): |
|
863 | def rawsize(self, rev): | |
864 | """return the length of the uncompressed text for a given revision""" |
|
864 | """return the length of the uncompressed text for a given revision""" | |
865 | l = self.index[rev][2] |
|
865 | l = self.index[rev][2] | |
866 | if l >= 0: |
|
866 | if l >= 0: | |
867 | return l |
|
867 | return l | |
868 |
|
868 | |||
869 | t = self.rawdata(rev) |
|
869 | t = self.rawdata(rev) | |
870 | return len(t) |
|
870 | return len(t) | |
871 |
|
871 | |||
872 | def size(self, rev): |
|
872 | def size(self, rev): | |
873 | """length of non-raw text (processed by a "read" flag processor)""" |
|
873 | """length of non-raw text (processed by a "read" flag processor)""" | |
874 | # fast path: if no "read" flag processor could change the content, |
|
874 | # fast path: if no "read" flag processor could change the content, | |
875 | # size is rawsize. note: ELLIPSIS is known to not change the content. |
|
875 | # size is rawsize. note: ELLIPSIS is known to not change the content. | |
876 | flags = self.flags(rev) |
|
876 | flags = self.flags(rev) | |
877 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: |
|
877 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: | |
878 | return self.rawsize(rev) |
|
878 | return self.rawsize(rev) | |
879 |
|
879 | |||
880 | return len(self.revision(rev, raw=False)) |
|
880 | return len(self.revision(rev, raw=False)) | |
881 |
|
881 | |||
882 | def chainbase(self, rev): |
|
882 | def chainbase(self, rev): | |
883 | base = self._chainbasecache.get(rev) |
|
883 | base = self._chainbasecache.get(rev) | |
884 | if base is not None: |
|
884 | if base is not None: | |
885 | return base |
|
885 | return base | |
886 |
|
886 | |||
887 | index = self.index |
|
887 | index = self.index | |
888 | iterrev = rev |
|
888 | iterrev = rev | |
889 | base = index[iterrev][3] |
|
889 | base = index[iterrev][3] | |
890 | while base != iterrev: |
|
890 | while base != iterrev: | |
891 | iterrev = base |
|
891 | iterrev = base | |
892 | base = index[iterrev][3] |
|
892 | base = index[iterrev][3] | |
893 |
|
893 | |||
894 | self._chainbasecache[rev] = base |
|
894 | self._chainbasecache[rev] = base | |
895 | return base |
|
895 | return base | |
896 |
|
896 | |||
897 | def linkrev(self, rev): |
|
897 | def linkrev(self, rev): | |
898 | return self.index[rev][4] |
|
898 | return self.index[rev][4] | |
899 |
|
899 | |||
900 | def parentrevs(self, rev): |
|
900 | def parentrevs(self, rev): | |
901 | try: |
|
901 | try: | |
902 | entry = self.index[rev] |
|
902 | entry = self.index[rev] | |
903 | except IndexError: |
|
903 | except IndexError: | |
904 | if rev == wdirrev: |
|
904 | if rev == wdirrev: | |
905 | raise error.WdirUnsupported |
|
905 | raise error.WdirUnsupported | |
906 | raise |
|
906 | raise | |
907 |
|
907 | |||
908 | return entry[5], entry[6] |
|
908 | return entry[5], entry[6] | |
909 |
|
909 | |||
910 | # fast parentrevs(rev) where rev isn't filtered |
|
910 | # fast parentrevs(rev) where rev isn't filtered | |
911 | _uncheckedparentrevs = parentrevs |
|
911 | _uncheckedparentrevs = parentrevs | |
912 |
|
912 | |||
913 | def node(self, rev): |
|
913 | def node(self, rev): | |
914 | try: |
|
914 | try: | |
915 | return self.index[rev][7] |
|
915 | return self.index[rev][7] | |
916 | except IndexError: |
|
916 | except IndexError: | |
917 | if rev == wdirrev: |
|
917 | if rev == wdirrev: | |
918 | raise error.WdirUnsupported |
|
918 | raise error.WdirUnsupported | |
919 | raise |
|
919 | raise | |
920 |
|
920 | |||
921 | # Derived from index values. |
|
921 | # Derived from index values. | |
922 |
|
922 | |||
923 | def end(self, rev): |
|
923 | def end(self, rev): | |
924 | return self.start(rev) + self.length(rev) |
|
924 | return self.start(rev) + self.length(rev) | |
925 |
|
925 | |||
926 | def parents(self, node): |
|
926 | def parents(self, node): | |
927 | i = self.index |
|
927 | i = self.index | |
928 | d = i[self.rev(node)] |
|
928 | d = i[self.rev(node)] | |
929 | return i[d[5]][7], i[d[6]][7] # map revisions to nodes inline |
|
929 | return i[d[5]][7], i[d[6]][7] # map revisions to nodes inline | |
930 |
|
930 | |||
931 | def chainlen(self, rev): |
|
931 | def chainlen(self, rev): | |
932 | return self._chaininfo(rev)[0] |
|
932 | return self._chaininfo(rev)[0] | |
933 |
|
933 | |||
934 | def _chaininfo(self, rev): |
|
934 | def _chaininfo(self, rev): | |
935 | chaininfocache = self._chaininfocache |
|
935 | chaininfocache = self._chaininfocache | |
936 | if rev in chaininfocache: |
|
936 | if rev in chaininfocache: | |
937 | return chaininfocache[rev] |
|
937 | return chaininfocache[rev] | |
938 | index = self.index |
|
938 | index = self.index | |
939 | generaldelta = self._generaldelta |
|
939 | generaldelta = self._generaldelta | |
940 | iterrev = rev |
|
940 | iterrev = rev | |
941 | e = index[iterrev] |
|
941 | e = index[iterrev] | |
942 | clen = 0 |
|
942 | clen = 0 | |
943 | compresseddeltalen = 0 |
|
943 | compresseddeltalen = 0 | |
944 | while iterrev != e[3]: |
|
944 | while iterrev != e[3]: | |
945 | clen += 1 |
|
945 | clen += 1 | |
946 | compresseddeltalen += e[1] |
|
946 | compresseddeltalen += e[1] | |
947 | if generaldelta: |
|
947 | if generaldelta: | |
948 | iterrev = e[3] |
|
948 | iterrev = e[3] | |
949 | else: |
|
949 | else: | |
950 | iterrev -= 1 |
|
950 | iterrev -= 1 | |
951 | if iterrev in chaininfocache: |
|
951 | if iterrev in chaininfocache: | |
952 | t = chaininfocache[iterrev] |
|
952 | t = chaininfocache[iterrev] | |
953 | clen += t[0] |
|
953 | clen += t[0] | |
954 | compresseddeltalen += t[1] |
|
954 | compresseddeltalen += t[1] | |
955 | break |
|
955 | break | |
956 | e = index[iterrev] |
|
956 | e = index[iterrev] | |
957 | else: |
|
957 | else: | |
958 | # Add text length of base since decompressing that also takes |
|
958 | # Add text length of base since decompressing that also takes | |
959 | # work. For cache hits the length is already included. |
|
959 | # work. For cache hits the length is already included. | |
960 | compresseddeltalen += e[1] |
|
960 | compresseddeltalen += e[1] | |
961 | r = (clen, compresseddeltalen) |
|
961 | r = (clen, compresseddeltalen) | |
962 | chaininfocache[rev] = r |
|
962 | chaininfocache[rev] = r | |
963 | return r |
|
963 | return r | |
964 |
|
964 | |||
965 | def _deltachain(self, rev, stoprev=None): |
|
965 | def _deltachain(self, rev, stoprev=None): | |
966 | """Obtain the delta chain for a revision. |
|
966 | """Obtain the delta chain for a revision. | |
967 |
|
967 | |||
968 | ``stoprev`` specifies a revision to stop at. If not specified, we |
|
968 | ``stoprev`` specifies a revision to stop at. If not specified, we | |
969 | stop at the base of the chain. |
|
969 | stop at the base of the chain. | |
970 |
|
970 | |||
971 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of |
|
971 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of | |
972 | revs in ascending order and ``stopped`` is a bool indicating whether |
|
972 | revs in ascending order and ``stopped`` is a bool indicating whether | |
973 | ``stoprev`` was hit. |
|
973 | ``stoprev`` was hit. | |
974 | """ |
|
974 | """ | |
975 | # Try C implementation. |
|
975 | # Try C implementation. | |
976 | try: |
|
976 | try: | |
977 | return self.index.deltachain(rev, stoprev, self._generaldelta) |
|
977 | return self.index.deltachain(rev, stoprev, self._generaldelta) | |
978 | except AttributeError: |
|
978 | except AttributeError: | |
979 | pass |
|
979 | pass | |
980 |
|
980 | |||
981 | chain = [] |
|
981 | chain = [] | |
982 |
|
982 | |||
983 | # Alias to prevent attribute lookup in tight loop. |
|
983 | # Alias to prevent attribute lookup in tight loop. | |
984 | index = self.index |
|
984 | index = self.index | |
985 | generaldelta = self._generaldelta |
|
985 | generaldelta = self._generaldelta | |
986 |
|
986 | |||
987 | iterrev = rev |
|
987 | iterrev = rev | |
988 | e = index[iterrev] |
|
988 | e = index[iterrev] | |
989 | while iterrev != e[3] and iterrev != stoprev: |
|
989 | while iterrev != e[3] and iterrev != stoprev: | |
990 | chain.append(iterrev) |
|
990 | chain.append(iterrev) | |
991 | if generaldelta: |
|
991 | if generaldelta: | |
992 | iterrev = e[3] |
|
992 | iterrev = e[3] | |
993 | else: |
|
993 | else: | |
994 | iterrev -= 1 |
|
994 | iterrev -= 1 | |
995 | e = index[iterrev] |
|
995 | e = index[iterrev] | |
996 |
|
996 | |||
997 | if iterrev == stoprev: |
|
997 | if iterrev == stoprev: | |
998 | stopped = True |
|
998 | stopped = True | |
999 | else: |
|
999 | else: | |
1000 | chain.append(iterrev) |
|
1000 | chain.append(iterrev) | |
1001 | stopped = False |
|
1001 | stopped = False | |
1002 |
|
1002 | |||
1003 | chain.reverse() |
|
1003 | chain.reverse() | |
1004 | return chain, stopped |
|
1004 | return chain, stopped | |
1005 |
|
1005 | |||
1006 | def ancestors(self, revs, stoprev=0, inclusive=False): |
|
1006 | def ancestors(self, revs, stoprev=0, inclusive=False): | |
1007 | """Generate the ancestors of 'revs' in reverse revision order. |
|
1007 | """Generate the ancestors of 'revs' in reverse revision order. | |
1008 | Does not generate revs lower than stoprev. |
|
1008 | Does not generate revs lower than stoprev. | |
1009 |
|
1009 | |||
1010 | See the documentation for ancestor.lazyancestors for more details.""" |
|
1010 | See the documentation for ancestor.lazyancestors for more details.""" | |
1011 |
|
1011 | |||
1012 | # first, make sure start revisions aren't filtered |
|
1012 | # first, make sure start revisions aren't filtered | |
1013 | revs = list(revs) |
|
1013 | revs = list(revs) | |
1014 | checkrev = self.node |
|
1014 | checkrev = self.node | |
1015 | for r in revs: |
|
1015 | for r in revs: | |
1016 | checkrev(r) |
|
1016 | checkrev(r) | |
1017 | # and we're sure ancestors aren't filtered as well |
|
1017 | # and we're sure ancestors aren't filtered as well | |
1018 |
|
1018 | |||
1019 | if rustancestor is not None: |
|
1019 | if rustancestor is not None: | |
1020 | lazyancestors = rustancestor.LazyAncestors |
|
1020 | lazyancestors = rustancestor.LazyAncestors | |
1021 | arg = self.index |
|
1021 | arg = self.index | |
1022 | else: |
|
1022 | else: | |
1023 | lazyancestors = ancestor.lazyancestors |
|
1023 | lazyancestors = ancestor.lazyancestors | |
1024 | arg = self._uncheckedparentrevs |
|
1024 | arg = self._uncheckedparentrevs | |
1025 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) |
|
1025 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) | |
1026 |
|
1026 | |||
1027 | def descendants(self, revs): |
|
1027 | def descendants(self, revs): | |
1028 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
1028 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) | |
1029 |
|
1029 | |||
1030 | def findcommonmissing(self, common=None, heads=None): |
|
1030 | def findcommonmissing(self, common=None, heads=None): | |
1031 | """Return a tuple of the ancestors of common and the ancestors of heads |
|
1031 | """Return a tuple of the ancestors of common and the ancestors of heads | |
1032 | that are not ancestors of common. In revset terminology, we return the |
|
1032 | that are not ancestors of common. In revset terminology, we return the | |
1033 | tuple: |
|
1033 | tuple: | |
1034 |
|
1034 | |||
1035 | ::common, (::heads) - (::common) |
|
1035 | ::common, (::heads) - (::common) | |
1036 |
|
1036 | |||
1037 | The list is sorted by revision number, meaning it is |
|
1037 | The list is sorted by revision number, meaning it is | |
1038 | topologically sorted. |
|
1038 | topologically sorted. | |
1039 |
|
1039 | |||
1040 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1040 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1041 | not supplied, uses all of the revlog's heads. If common is not |
|
1041 | not supplied, uses all of the revlog's heads. If common is not | |
1042 | supplied, uses nullid.""" |
|
1042 | supplied, uses nullid.""" | |
1043 | if common is None: |
|
1043 | if common is None: | |
1044 | common = [nullid] |
|
1044 | common = [nullid] | |
1045 | if heads is None: |
|
1045 | if heads is None: | |
1046 | heads = self.heads() |
|
1046 | heads = self.heads() | |
1047 |
|
1047 | |||
1048 | common = [self.rev(n) for n in common] |
|
1048 | common = [self.rev(n) for n in common] | |
1049 | heads = [self.rev(n) for n in heads] |
|
1049 | heads = [self.rev(n) for n in heads] | |
1050 |
|
1050 | |||
1051 | # we want the ancestors, but inclusive |
|
1051 | # we want the ancestors, but inclusive | |
1052 | class lazyset(object): |
|
1052 | class lazyset(object): | |
1053 | def __init__(self, lazyvalues): |
|
1053 | def __init__(self, lazyvalues): | |
1054 | self.addedvalues = set() |
|
1054 | self.addedvalues = set() | |
1055 | self.lazyvalues = lazyvalues |
|
1055 | self.lazyvalues = lazyvalues | |
1056 |
|
1056 | |||
1057 | def __contains__(self, value): |
|
1057 | def __contains__(self, value): | |
1058 | return value in self.addedvalues or value in self.lazyvalues |
|
1058 | return value in self.addedvalues or value in self.lazyvalues | |
1059 |
|
1059 | |||
1060 | def __iter__(self): |
|
1060 | def __iter__(self): | |
1061 | added = self.addedvalues |
|
1061 | added = self.addedvalues | |
1062 | for r in added: |
|
1062 | for r in added: | |
1063 | yield r |
|
1063 | yield r | |
1064 | for r in self.lazyvalues: |
|
1064 | for r in self.lazyvalues: | |
1065 | if not r in added: |
|
1065 | if not r in added: | |
1066 | yield r |
|
1066 | yield r | |
1067 |
|
1067 | |||
1068 | def add(self, value): |
|
1068 | def add(self, value): | |
1069 | self.addedvalues.add(value) |
|
1069 | self.addedvalues.add(value) | |
1070 |
|
1070 | |||
1071 | def update(self, values): |
|
1071 | def update(self, values): | |
1072 | self.addedvalues.update(values) |
|
1072 | self.addedvalues.update(values) | |
1073 |
|
1073 | |||
1074 | has = lazyset(self.ancestors(common)) |
|
1074 | has = lazyset(self.ancestors(common)) | |
1075 | has.add(nullrev) |
|
1075 | has.add(nullrev) | |
1076 | has.update(common) |
|
1076 | has.update(common) | |
1077 |
|
1077 | |||
1078 | # take all ancestors from heads that aren't in has |
|
1078 | # take all ancestors from heads that aren't in has | |
1079 | missing = set() |
|
1079 | missing = set() | |
1080 | visit = collections.deque(r for r in heads if r not in has) |
|
1080 | visit = collections.deque(r for r in heads if r not in has) | |
1081 | while visit: |
|
1081 | while visit: | |
1082 | r = visit.popleft() |
|
1082 | r = visit.popleft() | |
1083 | if r in missing: |
|
1083 | if r in missing: | |
1084 | continue |
|
1084 | continue | |
1085 | else: |
|
1085 | else: | |
1086 | missing.add(r) |
|
1086 | missing.add(r) | |
1087 | for p in self.parentrevs(r): |
|
1087 | for p in self.parentrevs(r): | |
1088 | if p not in has: |
|
1088 | if p not in has: | |
1089 | visit.append(p) |
|
1089 | visit.append(p) | |
1090 | missing = list(missing) |
|
1090 | missing = list(missing) | |
1091 | missing.sort() |
|
1091 | missing.sort() | |
1092 | return has, [self.node(miss) for miss in missing] |
|
1092 | return has, [self.node(miss) for miss in missing] | |
1093 |
|
1093 | |||
1094 | def incrementalmissingrevs(self, common=None): |
|
1094 | def incrementalmissingrevs(self, common=None): | |
1095 | """Return an object that can be used to incrementally compute the |
|
1095 | """Return an object that can be used to incrementally compute the | |
1096 | revision numbers of the ancestors of arbitrary sets that are not |
|
1096 | revision numbers of the ancestors of arbitrary sets that are not | |
1097 | ancestors of common. This is an ancestor.incrementalmissingancestors |
|
1097 | ancestors of common. This is an ancestor.incrementalmissingancestors | |
1098 | object. |
|
1098 | object. | |
1099 |
|
1099 | |||
1100 | 'common' is a list of revision numbers. If common is not supplied, uses |
|
1100 | 'common' is a list of revision numbers. If common is not supplied, uses | |
1101 | nullrev. |
|
1101 | nullrev. | |
1102 | """ |
|
1102 | """ | |
1103 | if common is None: |
|
1103 | if common is None: | |
1104 | common = [nullrev] |
|
1104 | common = [nullrev] | |
1105 |
|
1105 | |||
1106 | if rustancestor is not None: |
|
1106 | if rustancestor is not None: | |
1107 | return rustancestor.MissingAncestors(self.index, common) |
|
1107 | return rustancestor.MissingAncestors(self.index, common) | |
1108 | return ancestor.incrementalmissingancestors(self.parentrevs, common) |
|
1108 | return ancestor.incrementalmissingancestors(self.parentrevs, common) | |
1109 |
|
1109 | |||
1110 | def findmissingrevs(self, common=None, heads=None): |
|
1110 | def findmissingrevs(self, common=None, heads=None): | |
1111 | """Return the revision numbers of the ancestors of heads that |
|
1111 | """Return the revision numbers of the ancestors of heads that | |
1112 | are not ancestors of common. |
|
1112 | are not ancestors of common. | |
1113 |
|
1113 | |||
1114 | More specifically, return a list of revision numbers corresponding to |
|
1114 | More specifically, return a list of revision numbers corresponding to | |
1115 | nodes N such that every N satisfies the following constraints: |
|
1115 | nodes N such that every N satisfies the following constraints: | |
1116 |
|
1116 | |||
1117 | 1. N is an ancestor of some node in 'heads' |
|
1117 | 1. N is an ancestor of some node in 'heads' | |
1118 | 2. N is not an ancestor of any node in 'common' |
|
1118 | 2. N is not an ancestor of any node in 'common' | |
1119 |
|
1119 | |||
1120 | The list is sorted by revision number, meaning it is |
|
1120 | The list is sorted by revision number, meaning it is | |
1121 | topologically sorted. |
|
1121 | topologically sorted. | |
1122 |
|
1122 | |||
1123 | 'heads' and 'common' are both lists of revision numbers. If heads is |
|
1123 | 'heads' and 'common' are both lists of revision numbers. If heads is | |
1124 | not supplied, uses all of the revlog's heads. If common is not |
|
1124 | not supplied, uses all of the revlog's heads. If common is not | |
1125 | supplied, uses nullid.""" |
|
1125 | supplied, uses nullid.""" | |
1126 | if common is None: |
|
1126 | if common is None: | |
1127 | common = [nullrev] |
|
1127 | common = [nullrev] | |
1128 | if heads is None: |
|
1128 | if heads is None: | |
1129 | heads = self.headrevs() |
|
1129 | heads = self.headrevs() | |
1130 |
|
1130 | |||
1131 | inc = self.incrementalmissingrevs(common=common) |
|
1131 | inc = self.incrementalmissingrevs(common=common) | |
1132 | return inc.missingancestors(heads) |
|
1132 | return inc.missingancestors(heads) | |
1133 |
|
1133 | |||
1134 | def findmissing(self, common=None, heads=None): |
|
1134 | def findmissing(self, common=None, heads=None): | |
1135 | """Return the ancestors of heads that are not ancestors of common. |
|
1135 | """Return the ancestors of heads that are not ancestors of common. | |
1136 |
|
1136 | |||
1137 | More specifically, return a list of nodes N such that every N |
|
1137 | More specifically, return a list of nodes N such that every N | |
1138 | satisfies the following constraints: |
|
1138 | satisfies the following constraints: | |
1139 |
|
1139 | |||
1140 | 1. N is an ancestor of some node in 'heads' |
|
1140 | 1. N is an ancestor of some node in 'heads' | |
1141 | 2. N is not an ancestor of any node in 'common' |
|
1141 | 2. N is not an ancestor of any node in 'common' | |
1142 |
|
1142 | |||
1143 | The list is sorted by revision number, meaning it is |
|
1143 | The list is sorted by revision number, meaning it is | |
1144 | topologically sorted. |
|
1144 | topologically sorted. | |
1145 |
|
1145 | |||
1146 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1146 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1147 | not supplied, uses all of the revlog's heads. If common is not |
|
1147 | not supplied, uses all of the revlog's heads. If common is not | |
1148 | supplied, uses nullid.""" |
|
1148 | supplied, uses nullid.""" | |
1149 | if common is None: |
|
1149 | if common is None: | |
1150 | common = [nullid] |
|
1150 | common = [nullid] | |
1151 | if heads is None: |
|
1151 | if heads is None: | |
1152 | heads = self.heads() |
|
1152 | heads = self.heads() | |
1153 |
|
1153 | |||
1154 | common = [self.rev(n) for n in common] |
|
1154 | common = [self.rev(n) for n in common] | |
1155 | heads = [self.rev(n) for n in heads] |
|
1155 | heads = [self.rev(n) for n in heads] | |
1156 |
|
1156 | |||
1157 | inc = self.incrementalmissingrevs(common=common) |
|
1157 | inc = self.incrementalmissingrevs(common=common) | |
1158 | return [self.node(r) for r in inc.missingancestors(heads)] |
|
1158 | return [self.node(r) for r in inc.missingancestors(heads)] | |
1159 |
|
1159 | |||
1160 | def nodesbetween(self, roots=None, heads=None): |
|
1160 | def nodesbetween(self, roots=None, heads=None): | |
1161 | """Return a topological path from 'roots' to 'heads'. |
|
1161 | """Return a topological path from 'roots' to 'heads'. | |
1162 |
|
1162 | |||
1163 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a |
|
1163 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a | |
1164 | topologically sorted list of all nodes N that satisfy both of |
|
1164 | topologically sorted list of all nodes N that satisfy both of | |
1165 | these constraints: |
|
1165 | these constraints: | |
1166 |
|
1166 | |||
1167 | 1. N is a descendant of some node in 'roots' |
|
1167 | 1. N is a descendant of some node in 'roots' | |
1168 | 2. N is an ancestor of some node in 'heads' |
|
1168 | 2. N is an ancestor of some node in 'heads' | |
1169 |
|
1169 | |||
1170 | Every node is considered to be both a descendant and an ancestor |
|
1170 | Every node is considered to be both a descendant and an ancestor | |
1171 | of itself, so every reachable node in 'roots' and 'heads' will be |
|
1171 | of itself, so every reachable node in 'roots' and 'heads' will be | |
1172 | included in 'nodes'. |
|
1172 | included in 'nodes'. | |
1173 |
|
1173 | |||
1174 | 'outroots' is the list of reachable nodes in 'roots', i.e., the |
|
1174 | 'outroots' is the list of reachable nodes in 'roots', i.e., the | |
1175 | subset of 'roots' that is returned in 'nodes'. Likewise, |
|
1175 | subset of 'roots' that is returned in 'nodes'. Likewise, | |
1176 | 'outheads' is the subset of 'heads' that is also in 'nodes'. |
|
1176 | 'outheads' is the subset of 'heads' that is also in 'nodes'. | |
1177 |
|
1177 | |||
1178 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is |
|
1178 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is | |
1179 | unspecified, uses nullid as the only root. If 'heads' is |
|
1179 | unspecified, uses nullid as the only root. If 'heads' is | |
1180 | unspecified, uses list of all of the revlog's heads.""" |
|
1180 | unspecified, uses list of all of the revlog's heads.""" | |
1181 | nonodes = ([], [], []) |
|
1181 | nonodes = ([], [], []) | |
1182 | if roots is not None: |
|
1182 | if roots is not None: | |
1183 | roots = list(roots) |
|
1183 | roots = list(roots) | |
1184 | if not roots: |
|
1184 | if not roots: | |
1185 | return nonodes |
|
1185 | return nonodes | |
1186 | lowestrev = min([self.rev(n) for n in roots]) |
|
1186 | lowestrev = min([self.rev(n) for n in roots]) | |
1187 | else: |
|
1187 | else: | |
1188 | roots = [nullid] # Everybody's a descendant of nullid |
|
1188 | roots = [nullid] # Everybody's a descendant of nullid | |
1189 | lowestrev = nullrev |
|
1189 | lowestrev = nullrev | |
1190 | if (lowestrev == nullrev) and (heads is None): |
|
1190 | if (lowestrev == nullrev) and (heads is None): | |
1191 | # We want _all_ the nodes! |
|
1191 | # We want _all_ the nodes! | |
1192 | return ([self.node(r) for r in self], [nullid], list(self.heads())) |
|
1192 | return ([self.node(r) for r in self], [nullid], list(self.heads())) | |
1193 | if heads is None: |
|
1193 | if heads is None: | |
1194 | # All nodes are ancestors, so the latest ancestor is the last |
|
1194 | # All nodes are ancestors, so the latest ancestor is the last | |
1195 | # node. |
|
1195 | # node. | |
1196 | highestrev = len(self) - 1 |
|
1196 | highestrev = len(self) - 1 | |
1197 | # Set ancestors to None to signal that every node is an ancestor. |
|
1197 | # Set ancestors to None to signal that every node is an ancestor. | |
1198 | ancestors = None |
|
1198 | ancestors = None | |
1199 | # Set heads to an empty dictionary for later discovery of heads |
|
1199 | # Set heads to an empty dictionary for later discovery of heads | |
1200 | heads = {} |
|
1200 | heads = {} | |
1201 | else: |
|
1201 | else: | |
1202 | heads = list(heads) |
|
1202 | heads = list(heads) | |
1203 | if not heads: |
|
1203 | if not heads: | |
1204 | return nonodes |
|
1204 | return nonodes | |
1205 | ancestors = set() |
|
1205 | ancestors = set() | |
1206 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
1206 | # Turn heads into a dictionary so we can remove 'fake' heads. | |
1207 | # Also, later we will be using it to filter out the heads we can't |
|
1207 | # Also, later we will be using it to filter out the heads we can't | |
1208 | # find from roots. |
|
1208 | # find from roots. | |
1209 | heads = dict.fromkeys(heads, False) |
|
1209 | heads = dict.fromkeys(heads, False) | |
1210 | # Start at the top and keep marking parents until we're done. |
|
1210 | # Start at the top and keep marking parents until we're done. | |
1211 | nodestotag = set(heads) |
|
1211 | nodestotag = set(heads) | |
1212 | # Remember where the top was so we can use it as a limit later. |
|
1212 | # Remember where the top was so we can use it as a limit later. | |
1213 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
1213 | highestrev = max([self.rev(n) for n in nodestotag]) | |
1214 | while nodestotag: |
|
1214 | while nodestotag: | |
1215 | # grab a node to tag |
|
1215 | # grab a node to tag | |
1216 | n = nodestotag.pop() |
|
1216 | n = nodestotag.pop() | |
1217 | # Never tag nullid |
|
1217 | # Never tag nullid | |
1218 | if n == nullid: |
|
1218 | if n == nullid: | |
1219 | continue |
|
1219 | continue | |
1220 | # A node's revision number represents its place in a |
|
1220 | # A node's revision number represents its place in a | |
1221 | # topologically sorted list of nodes. |
|
1221 | # topologically sorted list of nodes. | |
1222 | r = self.rev(n) |
|
1222 | r = self.rev(n) | |
1223 | if r >= lowestrev: |
|
1223 | if r >= lowestrev: | |
1224 | if n not in ancestors: |
|
1224 | if n not in ancestors: | |
1225 | # If we are possibly a descendant of one of the roots |
|
1225 | # If we are possibly a descendant of one of the roots | |
1226 | # and we haven't already been marked as an ancestor |
|
1226 | # and we haven't already been marked as an ancestor | |
1227 | ancestors.add(n) # Mark as ancestor |
|
1227 | ancestors.add(n) # Mark as ancestor | |
1228 | # Add non-nullid parents to list of nodes to tag. |
|
1228 | # Add non-nullid parents to list of nodes to tag. | |
1229 | nodestotag.update( |
|
1229 | nodestotag.update( | |
1230 | [p for p in self.parents(n) if p != nullid] |
|
1230 | [p for p in self.parents(n) if p != nullid] | |
1231 | ) |
|
1231 | ) | |
1232 | elif n in heads: # We've seen it before, is it a fake head? |
|
1232 | elif n in heads: # We've seen it before, is it a fake head? | |
1233 | # So it is, real heads should not be the ancestors of |
|
1233 | # So it is, real heads should not be the ancestors of | |
1234 | # any other heads. |
|
1234 | # any other heads. | |
1235 | heads.pop(n) |
|
1235 | heads.pop(n) | |
1236 | if not ancestors: |
|
1236 | if not ancestors: | |
1237 | return nonodes |
|
1237 | return nonodes | |
1238 | # Now that we have our set of ancestors, we want to remove any |
|
1238 | # Now that we have our set of ancestors, we want to remove any | |
1239 | # roots that are not ancestors. |
|
1239 | # roots that are not ancestors. | |
1240 |
|
1240 | |||
1241 | # If one of the roots was nullid, everything is included anyway. |
|
1241 | # If one of the roots was nullid, everything is included anyway. | |
1242 | if lowestrev > nullrev: |
|
1242 | if lowestrev > nullrev: | |
1243 | # But, since we weren't, let's recompute the lowest rev to not |
|
1243 | # But, since we weren't, let's recompute the lowest rev to not | |
1244 | # include roots that aren't ancestors. |
|
1244 | # include roots that aren't ancestors. | |
1245 |
|
1245 | |||
1246 | # Filter out roots that aren't ancestors of heads |
|
1246 | # Filter out roots that aren't ancestors of heads | |
1247 | roots = [root for root in roots if root in ancestors] |
|
1247 | roots = [root for root in roots if root in ancestors] | |
1248 | # Recompute the lowest revision |
|
1248 | # Recompute the lowest revision | |
1249 | if roots: |
|
1249 | if roots: | |
1250 | lowestrev = min([self.rev(root) for root in roots]) |
|
1250 | lowestrev = min([self.rev(root) for root in roots]) | |
1251 | else: |
|
1251 | else: | |
1252 | # No more roots? Return empty list |
|
1252 | # No more roots? Return empty list | |
1253 | return nonodes |
|
1253 | return nonodes | |
1254 | else: |
|
1254 | else: | |
1255 | # We are descending from nullid, and don't need to care about |
|
1255 | # We are descending from nullid, and don't need to care about | |
1256 | # any other roots. |
|
1256 | # any other roots. | |
1257 | lowestrev = nullrev |
|
1257 | lowestrev = nullrev | |
1258 | roots = [nullid] |
|
1258 | roots = [nullid] | |
1259 | # Transform our roots list into a set. |
|
1259 | # Transform our roots list into a set. | |
1260 | descendants = set(roots) |
|
1260 | descendants = set(roots) | |
1261 | # Also, keep the original roots so we can filter out roots that aren't |
|
1261 | # Also, keep the original roots so we can filter out roots that aren't | |
1262 | # 'real' roots (i.e. are descended from other roots). |
|
1262 | # 'real' roots (i.e. are descended from other roots). | |
1263 | roots = descendants.copy() |
|
1263 | roots = descendants.copy() | |
1264 | # Our topologically sorted list of output nodes. |
|
1264 | # Our topologically sorted list of output nodes. | |
1265 | orderedout = [] |
|
1265 | orderedout = [] | |
1266 | # Don't start at nullid since we don't want nullid in our output list, |
|
1266 | # Don't start at nullid since we don't want nullid in our output list, | |
1267 | # and if nullid shows up in descendants, empty parents will look like |
|
1267 | # and if nullid shows up in descendants, empty parents will look like | |
1268 | # they're descendants. |
|
1268 | # they're descendants. | |
1269 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): |
|
1269 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): | |
1270 | n = self.node(r) |
|
1270 | n = self.node(r) | |
1271 | isdescendant = False |
|
1271 | isdescendant = False | |
1272 | if lowestrev == nullrev: # Everybody is a descendant of nullid |
|
1272 | if lowestrev == nullrev: # Everybody is a descendant of nullid | |
1273 | isdescendant = True |
|
1273 | isdescendant = True | |
1274 | elif n in descendants: |
|
1274 | elif n in descendants: | |
1275 | # n is already a descendant |
|
1275 | # n is already a descendant | |
1276 | isdescendant = True |
|
1276 | isdescendant = True | |
1277 | # This check only needs to be done here because all the roots |
|
1277 | # This check only needs to be done here because all the roots | |
1278 | # will start being marked is descendants before the loop. |
|
1278 | # will start being marked is descendants before the loop. | |
1279 | if n in roots: |
|
1279 | if n in roots: | |
1280 | # If n was a root, check if it's a 'real' root. |
|
1280 | # If n was a root, check if it's a 'real' root. | |
1281 | p = tuple(self.parents(n)) |
|
1281 | p = tuple(self.parents(n)) | |
1282 | # If any of its parents are descendants, it's not a root. |
|
1282 | # If any of its parents are descendants, it's not a root. | |
1283 | if (p[0] in descendants) or (p[1] in descendants): |
|
1283 | if (p[0] in descendants) or (p[1] in descendants): | |
1284 | roots.remove(n) |
|
1284 | roots.remove(n) | |
1285 | else: |
|
1285 | else: | |
1286 | p = tuple(self.parents(n)) |
|
1286 | p = tuple(self.parents(n)) | |
1287 | # A node is a descendant if either of its parents are |
|
1287 | # A node is a descendant if either of its parents are | |
1288 | # descendants. (We seeded the dependents list with the roots |
|
1288 | # descendants. (We seeded the dependents list with the roots | |
1289 | # up there, remember?) |
|
1289 | # up there, remember?) | |
1290 | if (p[0] in descendants) or (p[1] in descendants): |
|
1290 | if (p[0] in descendants) or (p[1] in descendants): | |
1291 | descendants.add(n) |
|
1291 | descendants.add(n) | |
1292 | isdescendant = True |
|
1292 | isdescendant = True | |
1293 | if isdescendant and ((ancestors is None) or (n in ancestors)): |
|
1293 | if isdescendant and ((ancestors is None) or (n in ancestors)): | |
1294 | # Only include nodes that are both descendants and ancestors. |
|
1294 | # Only include nodes that are both descendants and ancestors. | |
1295 | orderedout.append(n) |
|
1295 | orderedout.append(n) | |
1296 | if (ancestors is not None) and (n in heads): |
|
1296 | if (ancestors is not None) and (n in heads): | |
1297 | # We're trying to figure out which heads are reachable |
|
1297 | # We're trying to figure out which heads are reachable | |
1298 | # from roots. |
|
1298 | # from roots. | |
1299 | # Mark this head as having been reached |
|
1299 | # Mark this head as having been reached | |
1300 | heads[n] = True |
|
1300 | heads[n] = True | |
1301 | elif ancestors is None: |
|
1301 | elif ancestors is None: | |
1302 | # Otherwise, we're trying to discover the heads. |
|
1302 | # Otherwise, we're trying to discover the heads. | |
1303 | # Assume this is a head because if it isn't, the next step |
|
1303 | # Assume this is a head because if it isn't, the next step | |
1304 | # will eventually remove it. |
|
1304 | # will eventually remove it. | |
1305 | heads[n] = True |
|
1305 | heads[n] = True | |
1306 | # But, obviously its parents aren't. |
|
1306 | # But, obviously its parents aren't. | |
1307 | for p in self.parents(n): |
|
1307 | for p in self.parents(n): | |
1308 | heads.pop(p, None) |
|
1308 | heads.pop(p, None) | |
1309 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] |
|
1309 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] | |
1310 | roots = list(roots) |
|
1310 | roots = list(roots) | |
1311 | assert orderedout |
|
1311 | assert orderedout | |
1312 | assert roots |
|
1312 | assert roots | |
1313 | assert heads |
|
1313 | assert heads | |
1314 | return (orderedout, roots, heads) |
|
1314 | return (orderedout, roots, heads) | |
1315 |
|
1315 | |||
1316 | def headrevs(self, revs=None): |
|
1316 | def headrevs(self, revs=None): | |
1317 | if revs is None: |
|
1317 | if revs is None: | |
1318 | try: |
|
1318 | try: | |
1319 | return self.index.headrevs() |
|
1319 | return self.index.headrevs() | |
1320 | except AttributeError: |
|
1320 | except AttributeError: | |
1321 | return self._headrevs() |
|
1321 | return self._headrevs() | |
1322 | if rustdagop is not None: |
|
1322 | if rustdagop is not None: | |
1323 | return rustdagop.headrevs(self.index, revs) |
|
1323 | return rustdagop.headrevs(self.index, revs) | |
1324 | return dagop.headrevs(revs, self._uncheckedparentrevs) |
|
1324 | return dagop.headrevs(revs, self._uncheckedparentrevs) | |
1325 |
|
1325 | |||
1326 | def computephases(self, roots): |
|
1326 | def computephases(self, roots): | |
1327 | return self.index.computephasesmapsets(roots) |
|
1327 | return self.index.computephasesmapsets(roots) | |
1328 |
|
1328 | |||
1329 | def _headrevs(self): |
|
1329 | def _headrevs(self): | |
1330 | count = len(self) |
|
1330 | count = len(self) | |
1331 | if not count: |
|
1331 | if not count: | |
1332 | return [nullrev] |
|
1332 | return [nullrev] | |
1333 | # we won't iter over filtered rev so nobody is a head at start |
|
1333 | # we won't iter over filtered rev so nobody is a head at start | |
1334 | ishead = [0] * (count + 1) |
|
1334 | ishead = [0] * (count + 1) | |
1335 | index = self.index |
|
1335 | index = self.index | |
1336 | for r in self: |
|
1336 | for r in self: | |
1337 | ishead[r] = 1 # I may be an head |
|
1337 | ishead[r] = 1 # I may be an head | |
1338 | e = index[r] |
|
1338 | e = index[r] | |
1339 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not |
|
1339 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not | |
1340 | return [r for r, val in enumerate(ishead) if val] |
|
1340 | return [r for r, val in enumerate(ishead) if val] | |
1341 |
|
1341 | |||
1342 | def heads(self, start=None, stop=None): |
|
1342 | def heads(self, start=None, stop=None): | |
1343 | """return the list of all nodes that have no children |
|
1343 | """return the list of all nodes that have no children | |
1344 |
|
1344 | |||
1345 | if start is specified, only heads that are descendants of |
|
1345 | if start is specified, only heads that are descendants of | |
1346 | start will be returned |
|
1346 | start will be returned | |
1347 | if stop is specified, it will consider all the revs from stop |
|
1347 | if stop is specified, it will consider all the revs from stop | |
1348 | as if they had no children |
|
1348 | as if they had no children | |
1349 | """ |
|
1349 | """ | |
1350 | if start is None and stop is None: |
|
1350 | if start is None and stop is None: | |
1351 | if not len(self): |
|
1351 | if not len(self): | |
1352 | return [nullid] |
|
1352 | return [nullid] | |
1353 | return [self.node(r) for r in self.headrevs()] |
|
1353 | return [self.node(r) for r in self.headrevs()] | |
1354 |
|
1354 | |||
1355 | if start is None: |
|
1355 | if start is None: | |
1356 | start = nullrev |
|
1356 | start = nullrev | |
1357 | else: |
|
1357 | else: | |
1358 | start = self.rev(start) |
|
1358 | start = self.rev(start) | |
1359 |
|
1359 | |||
1360 | stoprevs = {self.rev(n) for n in stop or []} |
|
1360 | stoprevs = {self.rev(n) for n in stop or []} | |
1361 |
|
1361 | |||
1362 | revs = dagop.headrevssubset( |
|
1362 | revs = dagop.headrevssubset( | |
1363 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs |
|
1363 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs | |
1364 | ) |
|
1364 | ) | |
1365 |
|
1365 | |||
1366 | return [self.node(rev) for rev in revs] |
|
1366 | return [self.node(rev) for rev in revs] | |
1367 |
|
1367 | |||
1368 | def children(self, node): |
|
1368 | def children(self, node): | |
1369 | """find the children of a given node""" |
|
1369 | """find the children of a given node""" | |
1370 | c = [] |
|
1370 | c = [] | |
1371 | p = self.rev(node) |
|
1371 | p = self.rev(node) | |
1372 | for r in self.revs(start=p + 1): |
|
1372 | for r in self.revs(start=p + 1): | |
1373 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
1373 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
1374 | if prevs: |
|
1374 | if prevs: | |
1375 | for pr in prevs: |
|
1375 | for pr in prevs: | |
1376 | if pr == p: |
|
1376 | if pr == p: | |
1377 | c.append(self.node(r)) |
|
1377 | c.append(self.node(r)) | |
1378 | elif p == nullrev: |
|
1378 | elif p == nullrev: | |
1379 | c.append(self.node(r)) |
|
1379 | c.append(self.node(r)) | |
1380 | return c |
|
1380 | return c | |
1381 |
|
1381 | |||
1382 | def commonancestorsheads(self, a, b): |
|
1382 | def commonancestorsheads(self, a, b): | |
1383 | """calculate all the heads of the common ancestors of nodes a and b""" |
|
1383 | """calculate all the heads of the common ancestors of nodes a and b""" | |
1384 | a, b = self.rev(a), self.rev(b) |
|
1384 | a, b = self.rev(a), self.rev(b) | |
1385 | ancs = self._commonancestorsheads(a, b) |
|
1385 | ancs = self._commonancestorsheads(a, b) | |
1386 | return pycompat.maplist(self.node, ancs) |
|
1386 | return pycompat.maplist(self.node, ancs) | |
1387 |
|
1387 | |||
1388 | def _commonancestorsheads(self, *revs): |
|
1388 | def _commonancestorsheads(self, *revs): | |
1389 | """calculate all the heads of the common ancestors of revs""" |
|
1389 | """calculate all the heads of the common ancestors of revs""" | |
1390 | try: |
|
1390 | try: | |
1391 | ancs = self.index.commonancestorsheads(*revs) |
|
1391 | ancs = self.index.commonancestorsheads(*revs) | |
1392 | except (AttributeError, OverflowError): # C implementation failed |
|
1392 | except (AttributeError, OverflowError): # C implementation failed | |
1393 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) |
|
1393 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) | |
1394 | return ancs |
|
1394 | return ancs | |
1395 |
|
1395 | |||
1396 | def isancestor(self, a, b): |
|
1396 | def isancestor(self, a, b): | |
1397 | """return True if node a is an ancestor of node b |
|
1397 | """return True if node a is an ancestor of node b | |
1398 |
|
1398 | |||
1399 | A revision is considered an ancestor of itself.""" |
|
1399 | A revision is considered an ancestor of itself.""" | |
1400 | a, b = self.rev(a), self.rev(b) |
|
1400 | a, b = self.rev(a), self.rev(b) | |
1401 | return self.isancestorrev(a, b) |
|
1401 | return self.isancestorrev(a, b) | |
1402 |
|
1402 | |||
1403 | def isancestorrev(self, a, b): |
|
1403 | def isancestorrev(self, a, b): | |
1404 | """return True if revision a is an ancestor of revision b |
|
1404 | """return True if revision a is an ancestor of revision b | |
1405 |
|
1405 | |||
1406 | A revision is considered an ancestor of itself. |
|
1406 | A revision is considered an ancestor of itself. | |
1407 |
|
1407 | |||
1408 | The implementation of this is trivial but the use of |
|
1408 | The implementation of this is trivial but the use of | |
1409 | reachableroots is not.""" |
|
1409 | reachableroots is not.""" | |
1410 | if a == nullrev: |
|
1410 | if a == nullrev: | |
1411 | return True |
|
1411 | return True | |
1412 | elif a == b: |
|
1412 | elif a == b: | |
1413 | return True |
|
1413 | return True | |
1414 | elif a > b: |
|
1414 | elif a > b: | |
1415 | return False |
|
1415 | return False | |
1416 | return bool(self.reachableroots(a, [b], [a], includepath=False)) |
|
1416 | return bool(self.reachableroots(a, [b], [a], includepath=False)) | |
1417 |
|
1417 | |||
1418 | def reachableroots(self, minroot, heads, roots, includepath=False): |
|
1418 | def reachableroots(self, minroot, heads, roots, includepath=False): | |
1419 | """return (heads(::(<roots> and <roots>::<heads>))) |
|
1419 | """return (heads(::(<roots> and <roots>::<heads>))) | |
1420 |
|
1420 | |||
1421 | If includepath is True, return (<roots>::<heads>).""" |
|
1421 | If includepath is True, return (<roots>::<heads>).""" | |
1422 | try: |
|
1422 | try: | |
1423 | return self.index.reachableroots2( |
|
1423 | return self.index.reachableroots2( | |
1424 | minroot, heads, roots, includepath |
|
1424 | minroot, heads, roots, includepath | |
1425 | ) |
|
1425 | ) | |
1426 | except AttributeError: |
|
1426 | except AttributeError: | |
1427 | return dagop._reachablerootspure( |
|
1427 | return dagop._reachablerootspure( | |
1428 | self.parentrevs, minroot, roots, heads, includepath |
|
1428 | self.parentrevs, minroot, roots, heads, includepath | |
1429 | ) |
|
1429 | ) | |
1430 |
|
1430 | |||
1431 | def ancestor(self, a, b): |
|
1431 | def ancestor(self, a, b): | |
1432 | """calculate the "best" common ancestor of nodes a and b""" |
|
1432 | """calculate the "best" common ancestor of nodes a and b""" | |
1433 |
|
1433 | |||
1434 | a, b = self.rev(a), self.rev(b) |
|
1434 | a, b = self.rev(a), self.rev(b) | |
1435 | try: |
|
1435 | try: | |
1436 | ancs = self.index.ancestors(a, b) |
|
1436 | ancs = self.index.ancestors(a, b) | |
1437 | except (AttributeError, OverflowError): |
|
1437 | except (AttributeError, OverflowError): | |
1438 | ancs = ancestor.ancestors(self.parentrevs, a, b) |
|
1438 | ancs = ancestor.ancestors(self.parentrevs, a, b) | |
1439 | if ancs: |
|
1439 | if ancs: | |
1440 | # choose a consistent winner when there's a tie |
|
1440 | # choose a consistent winner when there's a tie | |
1441 | return min(map(self.node, ancs)) |
|
1441 | return min(map(self.node, ancs)) | |
1442 | return nullid |
|
1442 | return nullid | |
1443 |
|
1443 | |||
1444 | def _match(self, id): |
|
1444 | def _match(self, id): | |
1445 | if isinstance(id, int): |
|
1445 | if isinstance(id, int): | |
1446 | # rev |
|
1446 | # rev | |
1447 | return self.node(id) |
|
1447 | return self.node(id) | |
1448 | if len(id) == 20: |
|
1448 | if len(id) == 20: | |
1449 | # possibly a binary node |
|
1449 | # possibly a binary node | |
1450 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
1450 | # odds of a binary node being all hex in ASCII are 1 in 10**25 | |
1451 | try: |
|
1451 | try: | |
1452 | node = id |
|
1452 | node = id | |
1453 | self.rev(node) # quick search the index |
|
1453 | self.rev(node) # quick search the index | |
1454 | return node |
|
1454 | return node | |
1455 | except error.LookupError: |
|
1455 | except error.LookupError: | |
1456 | pass # may be partial hex id |
|
1456 | pass # may be partial hex id | |
1457 | try: |
|
1457 | try: | |
1458 | # str(rev) |
|
1458 | # str(rev) | |
1459 | rev = int(id) |
|
1459 | rev = int(id) | |
1460 | if b"%d" % rev != id: |
|
1460 | if b"%d" % rev != id: | |
1461 | raise ValueError |
|
1461 | raise ValueError | |
1462 | if rev < 0: |
|
1462 | if rev < 0: | |
1463 | rev = len(self) + rev |
|
1463 | rev = len(self) + rev | |
1464 | if rev < 0 or rev >= len(self): |
|
1464 | if rev < 0 or rev >= len(self): | |
1465 | raise ValueError |
|
1465 | raise ValueError | |
1466 | return self.node(rev) |
|
1466 | return self.node(rev) | |
1467 | except (ValueError, OverflowError): |
|
1467 | except (ValueError, OverflowError): | |
1468 | pass |
|
1468 | pass | |
1469 | if len(id) == 40: |
|
1469 | if len(id) == 40: | |
1470 | try: |
|
1470 | try: | |
1471 | # a full hex nodeid? |
|
1471 | # a full hex nodeid? | |
1472 | node = bin(id) |
|
1472 | node = bin(id) | |
1473 | self.rev(node) |
|
1473 | self.rev(node) | |
1474 | return node |
|
1474 | return node | |
1475 | except (TypeError, error.LookupError): |
|
1475 | except (TypeError, error.LookupError): | |
1476 | pass |
|
1476 | pass | |
1477 |
|
1477 | |||
1478 | def _partialmatch(self, id): |
|
1478 | def _partialmatch(self, id): | |
1479 | # we don't care wdirfilenodeids as they should be always full hash |
|
1479 | # we don't care wdirfilenodeids as they should be always full hash | |
1480 | maybewdir = wdirhex.startswith(id) |
|
1480 | maybewdir = wdirhex.startswith(id) | |
1481 | try: |
|
1481 | try: | |
1482 | partial = self.index.partialmatch(id) |
|
1482 | partial = self.index.partialmatch(id) | |
1483 | if partial and self.hasnode(partial): |
|
1483 | if partial and self.hasnode(partial): | |
1484 | if maybewdir: |
|
1484 | if maybewdir: | |
1485 | # single 'ff...' match in radix tree, ambiguous with wdir |
|
1485 | # single 'ff...' match in radix tree, ambiguous with wdir | |
1486 | raise error.RevlogError |
|
1486 | raise error.RevlogError | |
1487 | return partial |
|
1487 | return partial | |
1488 | if maybewdir: |
|
1488 | if maybewdir: | |
1489 | # no 'ff...' match in radix tree, wdir identified |
|
1489 | # no 'ff...' match in radix tree, wdir identified | |
1490 | raise error.WdirUnsupported |
|
1490 | raise error.WdirUnsupported | |
1491 | return None |
|
1491 | return None | |
1492 | except error.RevlogError: |
|
1492 | except error.RevlogError: | |
1493 | # parsers.c radix tree lookup gave multiple matches |
|
1493 | # parsers.c radix tree lookup gave multiple matches | |
1494 | # fast path: for unfiltered changelog, radix tree is accurate |
|
1494 | # fast path: for unfiltered changelog, radix tree is accurate | |
1495 | if not getattr(self, 'filteredrevs', None): |
|
1495 | if not getattr(self, 'filteredrevs', None): | |
1496 | raise error.AmbiguousPrefixLookupError( |
|
1496 | raise error.AmbiguousPrefixLookupError( | |
1497 | id, self.indexfile, _(b'ambiguous identifier') |
|
1497 | id, self.indexfile, _(b'ambiguous identifier') | |
1498 | ) |
|
1498 | ) | |
1499 | # fall through to slow path that filters hidden revisions |
|
1499 | # fall through to slow path that filters hidden revisions | |
1500 | except (AttributeError, ValueError): |
|
1500 | except (AttributeError, ValueError): | |
1501 | # we are pure python, or key was too short to search radix tree |
|
1501 | # we are pure python, or key was too short to search radix tree | |
1502 | pass |
|
1502 | pass | |
1503 |
|
1503 | |||
1504 | if id in self._pcache: |
|
1504 | if id in self._pcache: | |
1505 | return self._pcache[id] |
|
1505 | return self._pcache[id] | |
1506 |
|
1506 | |||
1507 | if len(id) <= 40: |
|
1507 | if len(id) <= 40: | |
1508 | try: |
|
1508 | try: | |
1509 | # hex(node)[:...] |
|
1509 | # hex(node)[:...] | |
1510 | l = len(id) // 2 # grab an even number of digits |
|
1510 | l = len(id) // 2 # grab an even number of digits | |
1511 | prefix = bin(id[: l * 2]) |
|
1511 | prefix = bin(id[: l * 2]) | |
1512 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] |
|
1512 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] | |
1513 | nl = [ |
|
1513 | nl = [ | |
1514 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) |
|
1514 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) | |
1515 | ] |
|
1515 | ] | |
1516 | if nullhex.startswith(id): |
|
1516 | if nullhex.startswith(id): | |
1517 | nl.append(nullid) |
|
1517 | nl.append(nullid) | |
1518 | if len(nl) > 0: |
|
1518 | if len(nl) > 0: | |
1519 | if len(nl) == 1 and not maybewdir: |
|
1519 | if len(nl) == 1 and not maybewdir: | |
1520 | self._pcache[id] = nl[0] |
|
1520 | self._pcache[id] = nl[0] | |
1521 | return nl[0] |
|
1521 | return nl[0] | |
1522 | raise error.AmbiguousPrefixLookupError( |
|
1522 | raise error.AmbiguousPrefixLookupError( | |
1523 | id, self.indexfile, _(b'ambiguous identifier') |
|
1523 | id, self.indexfile, _(b'ambiguous identifier') | |
1524 | ) |
|
1524 | ) | |
1525 | if maybewdir: |
|
1525 | if maybewdir: | |
1526 | raise error.WdirUnsupported |
|
1526 | raise error.WdirUnsupported | |
1527 | return None |
|
1527 | return None | |
1528 | except TypeError: |
|
1528 | except TypeError: | |
1529 | pass |
|
1529 | pass | |
1530 |
|
1530 | |||
1531 | def lookup(self, id): |
|
1531 | def lookup(self, id): | |
1532 | """locate a node based on: |
|
1532 | """locate a node based on: | |
1533 | - revision number or str(revision number) |
|
1533 | - revision number or str(revision number) | |
1534 | - nodeid or subset of hex nodeid |
|
1534 | - nodeid or subset of hex nodeid | |
1535 | """ |
|
1535 | """ | |
1536 | n = self._match(id) |
|
1536 | n = self._match(id) | |
1537 | if n is not None: |
|
1537 | if n is not None: | |
1538 | return n |
|
1538 | return n | |
1539 | n = self._partialmatch(id) |
|
1539 | n = self._partialmatch(id) | |
1540 | if n: |
|
1540 | if n: | |
1541 | return n |
|
1541 | return n | |
1542 |
|
1542 | |||
1543 | raise error.LookupError(id, self.indexfile, _(b'no match found')) |
|
1543 | raise error.LookupError(id, self.indexfile, _(b'no match found')) | |
1544 |
|
1544 | |||
1545 | def shortest(self, node, minlength=1): |
|
1545 | def shortest(self, node, minlength=1): | |
1546 | """Find the shortest unambiguous prefix that matches node.""" |
|
1546 | """Find the shortest unambiguous prefix that matches node.""" | |
1547 |
|
1547 | |||
1548 | def isvalid(prefix): |
|
1548 | def isvalid(prefix): | |
1549 | try: |
|
1549 | try: | |
1550 | matchednode = self._partialmatch(prefix) |
|
1550 | matchednode = self._partialmatch(prefix) | |
1551 | except error.AmbiguousPrefixLookupError: |
|
1551 | except error.AmbiguousPrefixLookupError: | |
1552 | return False |
|
1552 | return False | |
1553 | except error.WdirUnsupported: |
|
1553 | except error.WdirUnsupported: | |
1554 | # single 'ff...' match |
|
1554 | # single 'ff...' match | |
1555 | return True |
|
1555 | return True | |
1556 | if matchednode is None: |
|
1556 | if matchednode is None: | |
1557 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
1557 | raise error.LookupError(node, self.indexfile, _(b'no node')) | |
1558 | return True |
|
1558 | return True | |
1559 |
|
1559 | |||
1560 | def maybewdir(prefix): |
|
1560 | def maybewdir(prefix): | |
1561 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) |
|
1561 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) | |
1562 |
|
1562 | |||
1563 | hexnode = hex(node) |
|
1563 | hexnode = hex(node) | |
1564 |
|
1564 | |||
1565 | def disambiguate(hexnode, minlength): |
|
1565 | def disambiguate(hexnode, minlength): | |
1566 | """Disambiguate against wdirid.""" |
|
1566 | """Disambiguate against wdirid.""" | |
1567 | for length in range(minlength, len(hexnode) + 1): |
|
1567 | for length in range(minlength, len(hexnode) + 1): | |
1568 | prefix = hexnode[:length] |
|
1568 | prefix = hexnode[:length] | |
1569 | if not maybewdir(prefix): |
|
1569 | if not maybewdir(prefix): | |
1570 | return prefix |
|
1570 | return prefix | |
1571 |
|
1571 | |||
1572 | if not getattr(self, 'filteredrevs', None): |
|
1572 | if not getattr(self, 'filteredrevs', None): | |
1573 | try: |
|
1573 | try: | |
1574 | length = max(self.index.shortest(node), minlength) |
|
1574 | length = max(self.index.shortest(node), minlength) | |
1575 | return disambiguate(hexnode, length) |
|
1575 | return disambiguate(hexnode, length) | |
1576 | except error.RevlogError: |
|
1576 | except error.RevlogError: | |
1577 | if node != wdirid: |
|
1577 | if node != wdirid: | |
1578 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
1578 | raise error.LookupError(node, self.indexfile, _(b'no node')) | |
1579 | except AttributeError: |
|
1579 | except AttributeError: | |
1580 | # Fall through to pure code |
|
1580 | # Fall through to pure code | |
1581 | pass |
|
1581 | pass | |
1582 |
|
1582 | |||
1583 | if node == wdirid: |
|
1583 | if node == wdirid: | |
1584 | for length in range(minlength, len(hexnode) + 1): |
|
1584 | for length in range(minlength, len(hexnode) + 1): | |
1585 | prefix = hexnode[:length] |
|
1585 | prefix = hexnode[:length] | |
1586 | if isvalid(prefix): |
|
1586 | if isvalid(prefix): | |
1587 | return prefix |
|
1587 | return prefix | |
1588 |
|
1588 | |||
1589 | for length in range(minlength, len(hexnode) + 1): |
|
1589 | for length in range(minlength, len(hexnode) + 1): | |
1590 | prefix = hexnode[:length] |
|
1590 | prefix = hexnode[:length] | |
1591 | if isvalid(prefix): |
|
1591 | if isvalid(prefix): | |
1592 | return disambiguate(hexnode, length) |
|
1592 | return disambiguate(hexnode, length) | |
1593 |
|
1593 | |||
1594 | def cmp(self, node, text): |
|
1594 | def cmp(self, node, text): | |
1595 | """compare text with a given file revision |
|
1595 | """compare text with a given file revision | |
1596 |
|
1596 | |||
1597 | returns True if text is different than what is stored. |
|
1597 | returns True if text is different than what is stored. | |
1598 | """ |
|
1598 | """ | |
1599 | p1, p2 = self.parents(node) |
|
1599 | p1, p2 = self.parents(node) | |
1600 | return storageutil.hashrevisionsha1(text, p1, p2) != node |
|
1600 | return storageutil.hashrevisionsha1(text, p1, p2) != node | |
1601 |
|
1601 | |||
1602 | def _cachesegment(self, offset, data): |
|
1602 | def _cachesegment(self, offset, data): | |
1603 | """Add a segment to the revlog cache. |
|
1603 | """Add a segment to the revlog cache. | |
1604 |
|
1604 | |||
1605 | Accepts an absolute offset and the data that is at that location. |
|
1605 | Accepts an absolute offset and the data that is at that location. | |
1606 | """ |
|
1606 | """ | |
1607 | o, d = self._chunkcache |
|
1607 | o, d = self._chunkcache | |
1608 | # try to add to existing cache |
|
1608 | # try to add to existing cache | |
1609 | if o + len(d) == offset and len(d) + len(data) < _chunksize: |
|
1609 | if o + len(d) == offset and len(d) + len(data) < _chunksize: | |
1610 | self._chunkcache = o, d + data |
|
1610 | self._chunkcache = o, d + data | |
1611 | else: |
|
1611 | else: | |
1612 | self._chunkcache = offset, data |
|
1612 | self._chunkcache = offset, data | |
1613 |
|
1613 | |||
1614 | def _readsegment(self, offset, length, df=None): |
|
1614 | def _readsegment(self, offset, length, df=None): | |
1615 | """Load a segment of raw data from the revlog. |
|
1615 | """Load a segment of raw data from the revlog. | |
1616 |
|
1616 | |||
1617 | Accepts an absolute offset, length to read, and an optional existing |
|
1617 | Accepts an absolute offset, length to read, and an optional existing | |
1618 | file handle to read from. |
|
1618 | file handle to read from. | |
1619 |
|
1619 | |||
1620 | If an existing file handle is passed, it will be seeked and the |
|
1620 | If an existing file handle is passed, it will be seeked and the | |
1621 | original seek position will NOT be restored. |
|
1621 | original seek position will NOT be restored. | |
1622 |
|
1622 | |||
1623 | Returns a str or buffer of raw byte data. |
|
1623 | Returns a str or buffer of raw byte data. | |
1624 |
|
1624 | |||
1625 | Raises if the requested number of bytes could not be read. |
|
1625 | Raises if the requested number of bytes could not be read. | |
1626 | """ |
|
1626 | """ | |
1627 | # Cache data both forward and backward around the requested |
|
1627 | # Cache data both forward and backward around the requested | |
1628 | # data, in a fixed size window. This helps speed up operations |
|
1628 | # data, in a fixed size window. This helps speed up operations | |
1629 | # involving reading the revlog backwards. |
|
1629 | # involving reading the revlog backwards. | |
1630 | cachesize = self._chunkcachesize |
|
1630 | cachesize = self._chunkcachesize | |
1631 | realoffset = offset & ~(cachesize - 1) |
|
1631 | realoffset = offset & ~(cachesize - 1) | |
1632 | reallength = ( |
|
1632 | reallength = ( | |
1633 | (offset + length + cachesize) & ~(cachesize - 1) |
|
1633 | (offset + length + cachesize) & ~(cachesize - 1) | |
1634 | ) - realoffset |
|
1634 | ) - realoffset | |
1635 | with self._datareadfp(df) as df: |
|
1635 | with self._datareadfp(df) as df: | |
1636 | df.seek(realoffset) |
|
1636 | df.seek(realoffset) | |
1637 | d = df.read(reallength) |
|
1637 | d = df.read(reallength) | |
1638 |
|
1638 | |||
1639 | self._cachesegment(realoffset, d) |
|
1639 | self._cachesegment(realoffset, d) | |
1640 | if offset != realoffset or reallength != length: |
|
1640 | if offset != realoffset or reallength != length: | |
1641 | startoffset = offset - realoffset |
|
1641 | startoffset = offset - realoffset | |
1642 | if len(d) - startoffset < length: |
|
1642 | if len(d) - startoffset < length: | |
1643 | raise error.RevlogError( |
|
1643 | raise error.RevlogError( | |
1644 | _( |
|
1644 | _( | |
1645 | b'partial read of revlog %s; expected %d bytes from ' |
|
1645 | b'partial read of revlog %s; expected %d bytes from ' | |
1646 | b'offset %d, got %d' |
|
1646 | b'offset %d, got %d' | |
1647 | ) |
|
1647 | ) | |
1648 | % ( |
|
1648 | % ( | |
1649 | self.indexfile if self._inline else self.datafile, |
|
1649 | self.indexfile if self._inline else self.datafile, | |
1650 | length, |
|
1650 | length, | |
1651 | realoffset, |
|
1651 | realoffset, | |
1652 | len(d) - startoffset, |
|
1652 | len(d) - startoffset, | |
1653 | ) |
|
1653 | ) | |
1654 | ) |
|
1654 | ) | |
1655 |
|
1655 | |||
1656 | return util.buffer(d, startoffset, length) |
|
1656 | return util.buffer(d, startoffset, length) | |
1657 |
|
1657 | |||
1658 | if len(d) < length: |
|
1658 | if len(d) < length: | |
1659 | raise error.RevlogError( |
|
1659 | raise error.RevlogError( | |
1660 | _( |
|
1660 | _( | |
1661 | b'partial read of revlog %s; expected %d bytes from offset ' |
|
1661 | b'partial read of revlog %s; expected %d bytes from offset ' | |
1662 | b'%d, got %d' |
|
1662 | b'%d, got %d' | |
1663 | ) |
|
1663 | ) | |
1664 | % ( |
|
1664 | % ( | |
1665 | self.indexfile if self._inline else self.datafile, |
|
1665 | self.indexfile if self._inline else self.datafile, | |
1666 | length, |
|
1666 | length, | |
1667 | offset, |
|
1667 | offset, | |
1668 | len(d), |
|
1668 | len(d), | |
1669 | ) |
|
1669 | ) | |
1670 | ) |
|
1670 | ) | |
1671 |
|
1671 | |||
1672 | return d |
|
1672 | return d | |
1673 |
|
1673 | |||
1674 | def _getsegment(self, offset, length, df=None): |
|
1674 | def _getsegment(self, offset, length, df=None): | |
1675 | """Obtain a segment of raw data from the revlog. |
|
1675 | """Obtain a segment of raw data from the revlog. | |
1676 |
|
1676 | |||
1677 | Accepts an absolute offset, length of bytes to obtain, and an |
|
1677 | Accepts an absolute offset, length of bytes to obtain, and an | |
1678 | optional file handle to the already-opened revlog. If the file |
|
1678 | optional file handle to the already-opened revlog. If the file | |
1679 | handle is used, it's original seek position will not be preserved. |
|
1679 | handle is used, it's original seek position will not be preserved. | |
1680 |
|
1680 | |||
1681 | Requests for data may be returned from a cache. |
|
1681 | Requests for data may be returned from a cache. | |
1682 |
|
1682 | |||
1683 | Returns a str or a buffer instance of raw byte data. |
|
1683 | Returns a str or a buffer instance of raw byte data. | |
1684 | """ |
|
1684 | """ | |
1685 | o, d = self._chunkcache |
|
1685 | o, d = self._chunkcache | |
1686 | l = len(d) |
|
1686 | l = len(d) | |
1687 |
|
1687 | |||
1688 | # is it in the cache? |
|
1688 | # is it in the cache? | |
1689 | cachestart = offset - o |
|
1689 | cachestart = offset - o | |
1690 | cacheend = cachestart + length |
|
1690 | cacheend = cachestart + length | |
1691 | if cachestart >= 0 and cacheend <= l: |
|
1691 | if cachestart >= 0 and cacheend <= l: | |
1692 | if cachestart == 0 and cacheend == l: |
|
1692 | if cachestart == 0 and cacheend == l: | |
1693 | return d # avoid a copy |
|
1693 | return d # avoid a copy | |
1694 | return util.buffer(d, cachestart, cacheend - cachestart) |
|
1694 | return util.buffer(d, cachestart, cacheend - cachestart) | |
1695 |
|
1695 | |||
1696 | return self._readsegment(offset, length, df=df) |
|
1696 | return self._readsegment(offset, length, df=df) | |
1697 |
|
1697 | |||
1698 | def _getsegmentforrevs(self, startrev, endrev, df=None): |
|
1698 | def _getsegmentforrevs(self, startrev, endrev, df=None): | |
1699 | """Obtain a segment of raw data corresponding to a range of revisions. |
|
1699 | """Obtain a segment of raw data corresponding to a range of revisions. | |
1700 |
|
1700 | |||
1701 | Accepts the start and end revisions and an optional already-open |
|
1701 | Accepts the start and end revisions and an optional already-open | |
1702 | file handle to be used for reading. If the file handle is read, its |
|
1702 | file handle to be used for reading. If the file handle is read, its | |
1703 | seek position will not be preserved. |
|
1703 | seek position will not be preserved. | |
1704 |
|
1704 | |||
1705 | Requests for data may be satisfied by a cache. |
|
1705 | Requests for data may be satisfied by a cache. | |
1706 |
|
1706 | |||
1707 | Returns a 2-tuple of (offset, data) for the requested range of |
|
1707 | Returns a 2-tuple of (offset, data) for the requested range of | |
1708 | revisions. Offset is the integer offset from the beginning of the |
|
1708 | revisions. Offset is the integer offset from the beginning of the | |
1709 | revlog and data is a str or buffer of the raw byte data. |
|
1709 | revlog and data is a str or buffer of the raw byte data. | |
1710 |
|
1710 | |||
1711 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` |
|
1711 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` | |
1712 | to determine where each revision's data begins and ends. |
|
1712 | to determine where each revision's data begins and ends. | |
1713 | """ |
|
1713 | """ | |
1714 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons |
|
1714 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons | |
1715 | # (functions are expensive). |
|
1715 | # (functions are expensive). | |
1716 | index = self.index |
|
1716 | index = self.index | |
1717 | istart = index[startrev] |
|
1717 | istart = index[startrev] | |
1718 | start = int(istart[0] >> 16) |
|
1718 | start = int(istart[0] >> 16) | |
1719 | if startrev == endrev: |
|
1719 | if startrev == endrev: | |
1720 | end = start + istart[1] |
|
1720 | end = start + istart[1] | |
1721 | else: |
|
1721 | else: | |
1722 | iend = index[endrev] |
|
1722 | iend = index[endrev] | |
1723 | end = int(iend[0] >> 16) + iend[1] |
|
1723 | end = int(iend[0] >> 16) + iend[1] | |
1724 |
|
1724 | |||
1725 | if self._inline: |
|
1725 | if self._inline: | |
1726 | start += (startrev + 1) * self._io.size |
|
1726 | start += (startrev + 1) * self._io.size | |
1727 | end += (endrev + 1) * self._io.size |
|
1727 | end += (endrev + 1) * self._io.size | |
1728 | length = end - start |
|
1728 | length = end - start | |
1729 |
|
1729 | |||
1730 | return start, self._getsegment(start, length, df=df) |
|
1730 | return start, self._getsegment(start, length, df=df) | |
1731 |
|
1731 | |||
1732 | def _chunk(self, rev, df=None): |
|
1732 | def _chunk(self, rev, df=None): | |
1733 | """Obtain a single decompressed chunk for a revision. |
|
1733 | """Obtain a single decompressed chunk for a revision. | |
1734 |
|
1734 | |||
1735 | Accepts an integer revision and an optional already-open file handle |
|
1735 | Accepts an integer revision and an optional already-open file handle | |
1736 | to be used for reading. If used, the seek position of the file will not |
|
1736 | to be used for reading. If used, the seek position of the file will not | |
1737 | be preserved. |
|
1737 | be preserved. | |
1738 |
|
1738 | |||
1739 | Returns a str holding uncompressed data for the requested revision. |
|
1739 | Returns a str holding uncompressed data for the requested revision. | |
1740 | """ |
|
1740 | """ | |
1741 | return self.decompress(self._getsegmentforrevs(rev, rev, df=df)[1]) |
|
1741 | return self.decompress(self._getsegmentforrevs(rev, rev, df=df)[1]) | |
1742 |
|
1742 | |||
1743 | def _chunks(self, revs, df=None, targetsize=None): |
|
1743 | def _chunks(self, revs, df=None, targetsize=None): | |
1744 | """Obtain decompressed chunks for the specified revisions. |
|
1744 | """Obtain decompressed chunks for the specified revisions. | |
1745 |
|
1745 | |||
1746 | Accepts an iterable of numeric revisions that are assumed to be in |
|
1746 | Accepts an iterable of numeric revisions that are assumed to be in | |
1747 | ascending order. Also accepts an optional already-open file handle |
|
1747 | ascending order. Also accepts an optional already-open file handle | |
1748 | to be used for reading. If used, the seek position of the file will |
|
1748 | to be used for reading. If used, the seek position of the file will | |
1749 | not be preserved. |
|
1749 | not be preserved. | |
1750 |
|
1750 | |||
1751 | This function is similar to calling ``self._chunk()`` multiple times, |
|
1751 | This function is similar to calling ``self._chunk()`` multiple times, | |
1752 | but is faster. |
|
1752 | but is faster. | |
1753 |
|
1753 | |||
1754 | Returns a list with decompressed data for each requested revision. |
|
1754 | Returns a list with decompressed data for each requested revision. | |
1755 | """ |
|
1755 | """ | |
1756 | if not revs: |
|
1756 | if not revs: | |
1757 | return [] |
|
1757 | return [] | |
1758 | start = self.start |
|
1758 | start = self.start | |
1759 | length = self.length |
|
1759 | length = self.length | |
1760 | inline = self._inline |
|
1760 | inline = self._inline | |
1761 | iosize = self._io.size |
|
1761 | iosize = self._io.size | |
1762 | buffer = util.buffer |
|
1762 | buffer = util.buffer | |
1763 |
|
1763 | |||
1764 | l = [] |
|
1764 | l = [] | |
1765 | ladd = l.append |
|
1765 | ladd = l.append | |
1766 |
|
1766 | |||
1767 | if not self._withsparseread: |
|
1767 | if not self._withsparseread: | |
1768 | slicedchunks = (revs,) |
|
1768 | slicedchunks = (revs,) | |
1769 | else: |
|
1769 | else: | |
1770 | slicedchunks = deltautil.slicechunk( |
|
1770 | slicedchunks = deltautil.slicechunk( | |
1771 | self, revs, targetsize=targetsize |
|
1771 | self, revs, targetsize=targetsize | |
1772 | ) |
|
1772 | ) | |
1773 |
|
1773 | |||
1774 | for revschunk in slicedchunks: |
|
1774 | for revschunk in slicedchunks: | |
1775 | firstrev = revschunk[0] |
|
1775 | firstrev = revschunk[0] | |
1776 | # Skip trailing revisions with empty diff |
|
1776 | # Skip trailing revisions with empty diff | |
1777 | for lastrev in revschunk[::-1]: |
|
1777 | for lastrev in revschunk[::-1]: | |
1778 | if length(lastrev) != 0: |
|
1778 | if length(lastrev) != 0: | |
1779 | break |
|
1779 | break | |
1780 |
|
1780 | |||
1781 | try: |
|
1781 | try: | |
1782 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) |
|
1782 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) | |
1783 | except OverflowError: |
|
1783 | except OverflowError: | |
1784 | # issue4215 - we can't cache a run of chunks greater than |
|
1784 | # issue4215 - we can't cache a run of chunks greater than | |
1785 | # 2G on Windows |
|
1785 | # 2G on Windows | |
1786 | return [self._chunk(rev, df=df) for rev in revschunk] |
|
1786 | return [self._chunk(rev, df=df) for rev in revschunk] | |
1787 |
|
1787 | |||
1788 | decomp = self.decompress |
|
1788 | decomp = self.decompress | |
1789 | for rev in revschunk: |
|
1789 | for rev in revschunk: | |
1790 | chunkstart = start(rev) |
|
1790 | chunkstart = start(rev) | |
1791 | if inline: |
|
1791 | if inline: | |
1792 | chunkstart += (rev + 1) * iosize |
|
1792 | chunkstart += (rev + 1) * iosize | |
1793 | chunklength = length(rev) |
|
1793 | chunklength = length(rev) | |
1794 | ladd(decomp(buffer(data, chunkstart - offset, chunklength))) |
|
1794 | ladd(decomp(buffer(data, chunkstart - offset, chunklength))) | |
1795 |
|
1795 | |||
1796 | return l |
|
1796 | return l | |
1797 |
|
1797 | |||
1798 | def _chunkclear(self): |
|
1798 | def _chunkclear(self): | |
1799 | """Clear the raw chunk cache.""" |
|
1799 | """Clear the raw chunk cache.""" | |
1800 | self._chunkcache = (0, b'') |
|
1800 | self._chunkcache = (0, b'') | |
1801 |
|
1801 | |||
1802 | def deltaparent(self, rev): |
|
1802 | def deltaparent(self, rev): | |
1803 | """return deltaparent of the given revision""" |
|
1803 | """return deltaparent of the given revision""" | |
1804 | base = self.index[rev][3] |
|
1804 | base = self.index[rev][3] | |
1805 | if base == rev: |
|
1805 | if base == rev: | |
1806 | return nullrev |
|
1806 | return nullrev | |
1807 | elif self._generaldelta: |
|
1807 | elif self._generaldelta: | |
1808 | return base |
|
1808 | return base | |
1809 | else: |
|
1809 | else: | |
1810 | return rev - 1 |
|
1810 | return rev - 1 | |
1811 |
|
1811 | |||
1812 | def issnapshot(self, rev): |
|
1812 | def issnapshot(self, rev): | |
1813 | """tells whether rev is a snapshot""" |
|
1813 | """tells whether rev is a snapshot""" | |
1814 | if not self._sparserevlog: |
|
1814 | if not self._sparserevlog: | |
1815 | return self.deltaparent(rev) == nullrev |
|
1815 | return self.deltaparent(rev) == nullrev | |
1816 | elif util.safehasattr(self.index, b'issnapshot'): |
|
1816 | elif util.safehasattr(self.index, b'issnapshot'): | |
1817 | # directly assign the method to cache the testing and access |
|
1817 | # directly assign the method to cache the testing and access | |
1818 | self.issnapshot = self.index.issnapshot |
|
1818 | self.issnapshot = self.index.issnapshot | |
1819 | return self.issnapshot(rev) |
|
1819 | return self.issnapshot(rev) | |
1820 | if rev == nullrev: |
|
1820 | if rev == nullrev: | |
1821 | return True |
|
1821 | return True | |
1822 | entry = self.index[rev] |
|
1822 | entry = self.index[rev] | |
1823 | base = entry[3] |
|
1823 | base = entry[3] | |
1824 | if base == rev: |
|
1824 | if base == rev: | |
1825 | return True |
|
1825 | return True | |
1826 | if base == nullrev: |
|
1826 | if base == nullrev: | |
1827 | return True |
|
1827 | return True | |
1828 | p1 = entry[5] |
|
1828 | p1 = entry[5] | |
1829 | p2 = entry[6] |
|
1829 | p2 = entry[6] | |
1830 | if base == p1 or base == p2: |
|
1830 | if base == p1 or base == p2: | |
1831 | return False |
|
1831 | return False | |
1832 | return self.issnapshot(base) |
|
1832 | return self.issnapshot(base) | |
1833 |
|
1833 | |||
1834 | def snapshotdepth(self, rev): |
|
1834 | def snapshotdepth(self, rev): | |
1835 | """number of snapshot in the chain before this one""" |
|
1835 | """number of snapshot in the chain before this one""" | |
1836 | if not self.issnapshot(rev): |
|
1836 | if not self.issnapshot(rev): | |
1837 | raise error.ProgrammingError(b'revision %d not a snapshot') |
|
1837 | raise error.ProgrammingError(b'revision %d not a snapshot') | |
1838 | return len(self._deltachain(rev)[0]) - 1 |
|
1838 | return len(self._deltachain(rev)[0]) - 1 | |
1839 |
|
1839 | |||
1840 | def revdiff(self, rev1, rev2): |
|
1840 | def revdiff(self, rev1, rev2): | |
1841 | """return or calculate a delta between two revisions |
|
1841 | """return or calculate a delta between two revisions | |
1842 |
|
1842 | |||
1843 | The delta calculated is in binary form and is intended to be written to |
|
1843 | The delta calculated is in binary form and is intended to be written to | |
1844 | revlog data directly. So this function needs raw revision data. |
|
1844 | revlog data directly. So this function needs raw revision data. | |
1845 | """ |
|
1845 | """ | |
1846 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: |
|
1846 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: | |
1847 | return bytes(self._chunk(rev2)) |
|
1847 | return bytes(self._chunk(rev2)) | |
1848 |
|
1848 | |||
1849 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) |
|
1849 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) | |
1850 |
|
1850 | |||
1851 | def _processflags(self, text, flags, operation, raw=False): |
|
1851 | def _processflags(self, text, flags, operation, raw=False): | |
1852 | """deprecated entry point to access flag processors""" |
|
1852 | """deprecated entry point to access flag processors""" | |
1853 | msg = b'_processflag(...) use the specialized variant' |
|
1853 | msg = b'_processflag(...) use the specialized variant' | |
1854 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1854 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) | |
1855 | if raw: |
|
1855 | if raw: | |
1856 | return text, flagutil.processflagsraw(self, text, flags) |
|
1856 | return text, flagutil.processflagsraw(self, text, flags) | |
1857 | elif operation == b'read': |
|
1857 | elif operation == b'read': | |
1858 | return flagutil.processflagsread(self, text, flags) |
|
1858 | return flagutil.processflagsread(self, text, flags) | |
1859 | else: # write operation |
|
1859 | else: # write operation | |
1860 | return flagutil.processflagswrite(self, text, flags) |
|
1860 | return flagutil.processflagswrite(self, text, flags) | |
1861 |
|
1861 | |||
1862 | def revision(self, nodeorrev, _df=None, raw=False): |
|
1862 | def revision(self, nodeorrev, _df=None, raw=False): | |
1863 | """return an uncompressed revision of a given node or revision |
|
1863 | """return an uncompressed revision of a given node or revision | |
1864 | number. |
|
1864 | number. | |
1865 |
|
1865 | |||
1866 | _df - an existing file handle to read from. (internal-only) |
|
1866 | _df - an existing file handle to read from. (internal-only) | |
1867 | raw - an optional argument specifying if the revision data is to be |
|
1867 | raw - an optional argument specifying if the revision data is to be | |
1868 | treated as raw data when applying flag transforms. 'raw' should be set |
|
1868 | treated as raw data when applying flag transforms. 'raw' should be set | |
1869 | to True when generating changegroups or in debug commands. |
|
1869 | to True when generating changegroups or in debug commands. | |
1870 | """ |
|
1870 | """ | |
1871 | if raw: |
|
1871 | if raw: | |
1872 | msg = ( |
|
1872 | msg = ( | |
1873 | b'revlog.revision(..., raw=True) is deprecated, ' |
|
1873 | b'revlog.revision(..., raw=True) is deprecated, ' | |
1874 | b'use revlog.rawdata(...)' |
|
1874 | b'use revlog.rawdata(...)' | |
1875 | ) |
|
1875 | ) | |
1876 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1876 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) | |
1877 | return self._revisiondata(nodeorrev, _df, raw=raw)[0] |
|
1877 | return self._revisiondata(nodeorrev, _df, raw=raw)[0] | |
1878 |
|
1878 | |||
1879 | def sidedata(self, nodeorrev, _df=None): |
|
1879 | def sidedata(self, nodeorrev, _df=None): | |
1880 | """a map of extra data related to the changeset but not part of the hash |
|
1880 | """a map of extra data related to the changeset but not part of the hash | |
1881 |
|
1881 | |||
1882 | This function currently return a dictionary. However, more advanced |
|
1882 | This function currently return a dictionary. However, more advanced | |
1883 | mapping object will likely be used in the future for a more |
|
1883 | mapping object will likely be used in the future for a more | |
1884 | efficient/lazy code. |
|
1884 | efficient/lazy code. | |
1885 | """ |
|
1885 | """ | |
1886 | return self._revisiondata(nodeorrev, _df)[1] |
|
1886 | return self._revisiondata(nodeorrev, _df)[1] | |
1887 |
|
1887 | |||
1888 | def _revisiondata(self, nodeorrev, _df=None, raw=False): |
|
1888 | def _revisiondata(self, nodeorrev, _df=None, raw=False): | |
1889 | # deal with <nodeorrev> argument type |
|
1889 | # deal with <nodeorrev> argument type | |
1890 | if isinstance(nodeorrev, int): |
|
1890 | if isinstance(nodeorrev, int): | |
1891 | rev = nodeorrev |
|
1891 | rev = nodeorrev | |
1892 | node = self.node(rev) |
|
1892 | node = self.node(rev) | |
1893 | else: |
|
1893 | else: | |
1894 | node = nodeorrev |
|
1894 | node = nodeorrev | |
1895 | rev = None |
|
1895 | rev = None | |
1896 |
|
1896 | |||
1897 | # fast path the special `nullid` rev |
|
1897 | # fast path the special `nullid` rev | |
1898 | if node == nullid: |
|
1898 | if node == nullid: | |
1899 | return b"", {} |
|
1899 | return b"", {} | |
1900 |
|
1900 | |||
1901 | # ``rawtext`` is the text as stored inside the revlog. Might be the |
|
1901 | # ``rawtext`` is the text as stored inside the revlog. Might be the | |
1902 | # revision or might need to be processed to retrieve the revision. |
|
1902 | # revision or might need to be processed to retrieve the revision. | |
1903 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) |
|
1903 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) | |
1904 |
|
1904 | |||
1905 | if self.version & 0xFFFF == REVLOGV2: |
|
1905 | if self.version & 0xFFFF == REVLOGV2: | |
1906 | if rev is None: |
|
1906 | if rev is None: | |
1907 | rev = self.rev(node) |
|
1907 | rev = self.rev(node) | |
1908 | sidedata = self._sidedata(rev) |
|
1908 | sidedata = self._sidedata(rev) | |
1909 | else: |
|
1909 | else: | |
1910 | sidedata = {} |
|
1910 | sidedata = {} | |
1911 |
|
1911 | |||
1912 | if raw and validated: |
|
1912 | if raw and validated: | |
1913 | # if we don't want to process the raw text and that raw |
|
1913 | # if we don't want to process the raw text and that raw | |
1914 | # text is cached, we can exit early. |
|
1914 | # text is cached, we can exit early. | |
1915 | return rawtext, sidedata |
|
1915 | return rawtext, sidedata | |
1916 | if rev is None: |
|
1916 | if rev is None: | |
1917 | rev = self.rev(node) |
|
1917 | rev = self.rev(node) | |
1918 | # the revlog's flag for this revision |
|
1918 | # the revlog's flag for this revision | |
1919 | # (usually alter its state or content) |
|
1919 | # (usually alter its state or content) | |
1920 | flags = self.flags(rev) |
|
1920 | flags = self.flags(rev) | |
1921 |
|
1921 | |||
1922 | if validated and flags == REVIDX_DEFAULT_FLAGS: |
|
1922 | if validated and flags == REVIDX_DEFAULT_FLAGS: | |
1923 | # no extra flags set, no flag processor runs, text = rawtext |
|
1923 | # no extra flags set, no flag processor runs, text = rawtext | |
1924 | return rawtext, sidedata |
|
1924 | return rawtext, sidedata | |
1925 |
|
1925 | |||
1926 | if raw: |
|
1926 | if raw: | |
1927 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
1927 | validatehash = flagutil.processflagsraw(self, rawtext, flags) | |
1928 | text = rawtext |
|
1928 | text = rawtext | |
1929 | else: |
|
1929 | else: | |
1930 | r = flagutil.processflagsread(self, rawtext, flags) |
|
1930 | r = flagutil.processflagsread(self, rawtext, flags) | |
1931 | text, validatehash = r |
|
1931 | text, validatehash = r | |
1932 | if validatehash: |
|
1932 | if validatehash: | |
1933 | self.checkhash(text, node, rev=rev) |
|
1933 | self.checkhash(text, node, rev=rev) | |
1934 | if not validated: |
|
1934 | if not validated: | |
1935 | self._revisioncache = (node, rev, rawtext) |
|
1935 | self._revisioncache = (node, rev, rawtext) | |
1936 |
|
1936 | |||
1937 | return text, sidedata |
|
1937 | return text, sidedata | |
1938 |
|
1938 | |||
1939 | def _rawtext(self, node, rev, _df=None): |
|
1939 | def _rawtext(self, node, rev, _df=None): | |
1940 | """return the possibly unvalidated rawtext for a revision |
|
1940 | """return the possibly unvalidated rawtext for a revision | |
1941 |
|
1941 | |||
1942 | returns (rev, rawtext, validated) |
|
1942 | returns (rev, rawtext, validated) | |
1943 | """ |
|
1943 | """ | |
1944 |
|
1944 | |||
1945 | # revision in the cache (could be useful to apply delta) |
|
1945 | # revision in the cache (could be useful to apply delta) | |
1946 | cachedrev = None |
|
1946 | cachedrev = None | |
1947 | # An intermediate text to apply deltas to |
|
1947 | # An intermediate text to apply deltas to | |
1948 | basetext = None |
|
1948 | basetext = None | |
1949 |
|
1949 | |||
1950 | # Check if we have the entry in cache |
|
1950 | # Check if we have the entry in cache | |
1951 | # The cache entry looks like (node, rev, rawtext) |
|
1951 | # The cache entry looks like (node, rev, rawtext) | |
1952 | if self._revisioncache: |
|
1952 | if self._revisioncache: | |
1953 | if self._revisioncache[0] == node: |
|
1953 | if self._revisioncache[0] == node: | |
1954 | return (rev, self._revisioncache[2], True) |
|
1954 | return (rev, self._revisioncache[2], True) | |
1955 | cachedrev = self._revisioncache[1] |
|
1955 | cachedrev = self._revisioncache[1] | |
1956 |
|
1956 | |||
1957 | if rev is None: |
|
1957 | if rev is None: | |
1958 | rev = self.rev(node) |
|
1958 | rev = self.rev(node) | |
1959 |
|
1959 | |||
1960 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) |
|
1960 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) | |
1961 | if stopped: |
|
1961 | if stopped: | |
1962 | basetext = self._revisioncache[2] |
|
1962 | basetext = self._revisioncache[2] | |
1963 |
|
1963 | |||
1964 | # drop cache to save memory, the caller is expected to |
|
1964 | # drop cache to save memory, the caller is expected to | |
1965 | # update self._revisioncache after validating the text |
|
1965 | # update self._revisioncache after validating the text | |
1966 | self._revisioncache = None |
|
1966 | self._revisioncache = None | |
1967 |
|
1967 | |||
1968 | targetsize = None |
|
1968 | targetsize = None | |
1969 | rawsize = self.index[rev][2] |
|
1969 | rawsize = self.index[rev][2] | |
1970 | if 0 <= rawsize: |
|
1970 | if 0 <= rawsize: | |
1971 | targetsize = 4 * rawsize |
|
1971 | targetsize = 4 * rawsize | |
1972 |
|
1972 | |||
1973 | bins = self._chunks(chain, df=_df, targetsize=targetsize) |
|
1973 | bins = self._chunks(chain, df=_df, targetsize=targetsize) | |
1974 | if basetext is None: |
|
1974 | if basetext is None: | |
1975 | basetext = bytes(bins[0]) |
|
1975 | basetext = bytes(bins[0]) | |
1976 | bins = bins[1:] |
|
1976 | bins = bins[1:] | |
1977 |
|
1977 | |||
1978 | rawtext = mdiff.patches(basetext, bins) |
|
1978 | rawtext = mdiff.patches(basetext, bins) | |
1979 | del basetext # let us have a chance to free memory early |
|
1979 | del basetext # let us have a chance to free memory early | |
1980 | return (rev, rawtext, False) |
|
1980 | return (rev, rawtext, False) | |
1981 |
|
1981 | |||
1982 | def _sidedata(self, rev): |
|
1982 | def _sidedata(self, rev): | |
1983 | """Return the sidedata for a given revision number.""" |
|
1983 | """Return the sidedata for a given revision number.""" | |
1984 | index_entry = self.index[rev] |
|
1984 | index_entry = self.index[rev] | |
1985 | sidedata_offset = index_entry[8] |
|
1985 | sidedata_offset = index_entry[8] | |
1986 | sidedata_size = index_entry[9] |
|
1986 | sidedata_size = index_entry[9] | |
1987 |
|
1987 | |||
1988 | if self._inline: |
|
1988 | if self._inline: | |
1989 | sidedata_offset += self._io.size * (1 + rev) |
|
1989 | sidedata_offset += self._io.size * (1 + rev) | |
1990 | if sidedata_size == 0: |
|
1990 | if sidedata_size == 0: | |
1991 | return {} |
|
1991 | return {} | |
1992 |
|
1992 | |||
1993 | segment = self._getsegment(sidedata_offset, sidedata_size) |
|
1993 | segment = self._getsegment(sidedata_offset, sidedata_size) | |
1994 | sidedata = sidedatautil.deserialize_sidedata(segment) |
|
1994 | sidedata = sidedatautil.deserialize_sidedata(segment) | |
1995 | return sidedata |
|
1995 | return sidedata | |
1996 |
|
1996 | |||
1997 | def rawdata(self, nodeorrev, _df=None): |
|
1997 | def rawdata(self, nodeorrev, _df=None): | |
1998 | """return an uncompressed raw data of a given node or revision number. |
|
1998 | """return an uncompressed raw data of a given node or revision number. | |
1999 |
|
1999 | |||
2000 | _df - an existing file handle to read from. (internal-only) |
|
2000 | _df - an existing file handle to read from. (internal-only) | |
2001 | """ |
|
2001 | """ | |
2002 | return self._revisiondata(nodeorrev, _df, raw=True)[0] |
|
2002 | return self._revisiondata(nodeorrev, _df, raw=True)[0] | |
2003 |
|
2003 | |||
2004 | def hash(self, text, p1, p2): |
|
2004 | def hash(self, text, p1, p2): | |
2005 | """Compute a node hash. |
|
2005 | """Compute a node hash. | |
2006 |
|
2006 | |||
2007 | Available as a function so that subclasses can replace the hash |
|
2007 | Available as a function so that subclasses can replace the hash | |
2008 | as needed. |
|
2008 | as needed. | |
2009 | """ |
|
2009 | """ | |
2010 | return storageutil.hashrevisionsha1(text, p1, p2) |
|
2010 | return storageutil.hashrevisionsha1(text, p1, p2) | |
2011 |
|
2011 | |||
2012 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
2012 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
2013 | """Check node hash integrity. |
|
2013 | """Check node hash integrity. | |
2014 |
|
2014 | |||
2015 | Available as a function so that subclasses can extend hash mismatch |
|
2015 | Available as a function so that subclasses can extend hash mismatch | |
2016 | behaviors as needed. |
|
2016 | behaviors as needed. | |
2017 | """ |
|
2017 | """ | |
2018 | try: |
|
2018 | try: | |
2019 | if p1 is None and p2 is None: |
|
2019 | if p1 is None and p2 is None: | |
2020 | p1, p2 = self.parents(node) |
|
2020 | p1, p2 = self.parents(node) | |
2021 | if node != self.hash(text, p1, p2): |
|
2021 | if node != self.hash(text, p1, p2): | |
2022 | # Clear the revision cache on hash failure. The revision cache |
|
2022 | # Clear the revision cache on hash failure. The revision cache | |
2023 | # only stores the raw revision and clearing the cache does have |
|
2023 | # only stores the raw revision and clearing the cache does have | |
2024 | # the side-effect that we won't have a cache hit when the raw |
|
2024 | # the side-effect that we won't have a cache hit when the raw | |
2025 | # revision data is accessed. But this case should be rare and |
|
2025 | # revision data is accessed. But this case should be rare and | |
2026 | # it is extra work to teach the cache about the hash |
|
2026 | # it is extra work to teach the cache about the hash | |
2027 | # verification state. |
|
2027 | # verification state. | |
2028 | if self._revisioncache and self._revisioncache[0] == node: |
|
2028 | if self._revisioncache and self._revisioncache[0] == node: | |
2029 | self._revisioncache = None |
|
2029 | self._revisioncache = None | |
2030 |
|
2030 | |||
2031 | revornode = rev |
|
2031 | revornode = rev | |
2032 | if revornode is None: |
|
2032 | if revornode is None: | |
2033 | revornode = templatefilters.short(hex(node)) |
|
2033 | revornode = templatefilters.short(hex(node)) | |
2034 | raise error.RevlogError( |
|
2034 | raise error.RevlogError( | |
2035 | _(b"integrity check failed on %s:%s") |
|
2035 | _(b"integrity check failed on %s:%s") | |
2036 | % (self.indexfile, pycompat.bytestr(revornode)) |
|
2036 | % (self.indexfile, pycompat.bytestr(revornode)) | |
2037 | ) |
|
2037 | ) | |
2038 | except error.RevlogError: |
|
2038 | except error.RevlogError: | |
2039 | if self._censorable and storageutil.iscensoredtext(text): |
|
2039 | if self._censorable and storageutil.iscensoredtext(text): | |
2040 | raise error.CensoredNodeError(self.indexfile, node, text) |
|
2040 | raise error.CensoredNodeError(self.indexfile, node, text) | |
2041 | raise |
|
2041 | raise | |
2042 |
|
2042 | |||
2043 | def _enforceinlinesize(self, tr, fp=None): |
|
2043 | def _enforceinlinesize(self, tr, fp=None): | |
2044 | """Check if the revlog is too big for inline and convert if so. |
|
2044 | """Check if the revlog is too big for inline and convert if so. | |
2045 |
|
2045 | |||
2046 | This should be called after revisions are added to the revlog. If the |
|
2046 | This should be called after revisions are added to the revlog. If the | |
2047 | revlog has grown too large to be an inline revlog, it will convert it |
|
2047 | revlog has grown too large to be an inline revlog, it will convert it | |
2048 | to use multiple index and data files. |
|
2048 | to use multiple index and data files. | |
2049 | """ |
|
2049 | """ | |
2050 | tiprev = len(self) - 1 |
|
2050 | tiprev = len(self) - 1 | |
2051 | if ( |
|
2051 | if ( | |
2052 | not self._inline |
|
2052 | not self._inline | |
2053 | or (self.start(tiprev) + self.length(tiprev)) < _maxinline |
|
2053 | or (self.start(tiprev) + self.length(tiprev)) < _maxinline | |
2054 | ): |
|
2054 | ): | |
2055 | return |
|
2055 | return | |
2056 |
|
2056 | |||
2057 | troffset = tr.findoffset(self.indexfile) |
|
2057 | troffset = tr.findoffset(self.indexfile) | |
2058 | if troffset is None: |
|
2058 | if troffset is None: | |
2059 | raise error.RevlogError( |
|
2059 | raise error.RevlogError( | |
2060 | _(b"%s not found in the transaction") % self.indexfile |
|
2060 | _(b"%s not found in the transaction") % self.indexfile | |
2061 | ) |
|
2061 | ) | |
2062 | trindex = 0 |
|
2062 | trindex = 0 | |
2063 | tr.add(self.datafile, 0) |
|
2063 | tr.add(self.datafile, 0) | |
2064 |
|
2064 | |||
2065 | if fp: |
|
2065 | if fp: | |
2066 | fp.flush() |
|
2066 | fp.flush() | |
2067 | fp.close() |
|
2067 | fp.close() | |
2068 | # We can't use the cached file handle after close(). So prevent |
|
2068 | # We can't use the cached file handle after close(). So prevent | |
2069 | # its usage. |
|
2069 | # its usage. | |
2070 | self._writinghandles = None |
|
2070 | self._writinghandles = None | |
2071 |
|
2071 | |||
2072 | with self._indexfp(b'r') as ifh, self._datafp(b'w') as dfh: |
|
2072 | with self._indexfp(b'r') as ifh, self._datafp(b'w') as dfh: | |
2073 | for r in self: |
|
2073 | for r in self: | |
2074 | dfh.write(self._getsegmentforrevs(r, r, df=ifh)[1]) |
|
2074 | dfh.write(self._getsegmentforrevs(r, r, df=ifh)[1]) | |
2075 | if troffset <= self.start(r): |
|
2075 | if troffset <= self.start(r): | |
2076 | trindex = r |
|
2076 | trindex = r | |
2077 |
|
2077 | |||
2078 | with self._indexfp(b'w') as fp: |
|
2078 | with self._indexfp(b'w') as fp: | |
2079 | self.version &= ~FLAG_INLINE_DATA |
|
2079 | self.version &= ~FLAG_INLINE_DATA | |
2080 | self._inline = False |
|
2080 | self._inline = False | |
2081 | io = self._io |
|
2081 | io = self._io | |
2082 | for i in self: |
|
2082 | for i in self: | |
2083 | e = io.packentry(self.index[i], self.node, self.version, i) |
|
2083 | e = io.packentry(self.index[i], self.node, self.version, i) | |
2084 | fp.write(e) |
|
2084 | fp.write(e) | |
2085 |
|
2085 | |||
2086 | # the temp file replace the real index when we exit the context |
|
2086 | # the temp file replace the real index when we exit the context | |
2087 | # manager |
|
2087 | # manager | |
2088 |
|
2088 | |||
2089 | tr.replace(self.indexfile, trindex * self._io.size) |
|
2089 | tr.replace(self.indexfile, trindex * self._io.size) | |
2090 | nodemaputil.setup_persistent_nodemap(tr, self) |
|
2090 | nodemaputil.setup_persistent_nodemap(tr, self) | |
2091 | self._chunkclear() |
|
2091 | self._chunkclear() | |
2092 |
|
2092 | |||
2093 | def _nodeduplicatecallback(self, transaction, node): |
|
2093 | def _nodeduplicatecallback(self, transaction, node): | |
2094 | """called when trying to add a node already stored.""" |
|
2094 | """called when trying to add a node already stored.""" | |
2095 |
|
2095 | |||
2096 | def addrevision( |
|
2096 | def addrevision( | |
2097 | self, |
|
2097 | self, | |
2098 | text, |
|
2098 | text, | |
2099 | transaction, |
|
2099 | transaction, | |
2100 | link, |
|
2100 | link, | |
2101 | p1, |
|
2101 | p1, | |
2102 | p2, |
|
2102 | p2, | |
2103 | cachedelta=None, |
|
2103 | cachedelta=None, | |
2104 | node=None, |
|
2104 | node=None, | |
2105 | flags=REVIDX_DEFAULT_FLAGS, |
|
2105 | flags=REVIDX_DEFAULT_FLAGS, | |
2106 | deltacomputer=None, |
|
2106 | deltacomputer=None, | |
2107 | sidedata=None, |
|
2107 | sidedata=None, | |
2108 | ): |
|
2108 | ): | |
2109 | """add a revision to the log |
|
2109 | """add a revision to the log | |
2110 |
|
2110 | |||
2111 | text - the revision data to add |
|
2111 | text - the revision data to add | |
2112 | transaction - the transaction object used for rollback |
|
2112 | transaction - the transaction object used for rollback | |
2113 | link - the linkrev data to add |
|
2113 | link - the linkrev data to add | |
2114 | p1, p2 - the parent nodeids of the revision |
|
2114 | p1, p2 - the parent nodeids of the revision | |
2115 | cachedelta - an optional precomputed delta |
|
2115 | cachedelta - an optional precomputed delta | |
2116 | node - nodeid of revision; typically node is not specified, and it is |
|
2116 | node - nodeid of revision; typically node is not specified, and it is | |
2117 | computed by default as hash(text, p1, p2), however subclasses might |
|
2117 | computed by default as hash(text, p1, p2), however subclasses might | |
2118 | use different hashing method (and override checkhash() in such case) |
|
2118 | use different hashing method (and override checkhash() in such case) | |
2119 | flags - the known flags to set on the revision |
|
2119 | flags - the known flags to set on the revision | |
2120 | deltacomputer - an optional deltacomputer instance shared between |
|
2120 | deltacomputer - an optional deltacomputer instance shared between | |
2121 | multiple calls |
|
2121 | multiple calls | |
2122 | """ |
|
2122 | """ | |
2123 | if link == nullrev: |
|
2123 | if link == nullrev: | |
2124 | raise error.RevlogError( |
|
2124 | raise error.RevlogError( | |
2125 | _(b"attempted to add linkrev -1 to %s") % self.indexfile |
|
2125 | _(b"attempted to add linkrev -1 to %s") % self.indexfile | |
2126 | ) |
|
2126 | ) | |
2127 |
|
2127 | |||
2128 | if sidedata is None: |
|
2128 | if sidedata is None: | |
2129 | sidedata = {} |
|
2129 | sidedata = {} | |
2130 | elif not self.hassidedata: |
|
2130 | elif not self.hassidedata: | |
2131 | raise error.ProgrammingError( |
|
2131 | raise error.ProgrammingError( | |
2132 | _(b"trying to add sidedata to a revlog who don't support them") |
|
2132 | _(b"trying to add sidedata to a revlog who don't support them") | |
2133 | ) |
|
2133 | ) | |
2134 |
|
2134 | |||
2135 | if flags: |
|
2135 | if flags: | |
2136 | node = node or self.hash(text, p1, p2) |
|
2136 | node = node or self.hash(text, p1, p2) | |
2137 |
|
2137 | |||
2138 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) |
|
2138 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) | |
2139 |
|
2139 | |||
2140 | # If the flag processor modifies the revision data, ignore any provided |
|
2140 | # If the flag processor modifies the revision data, ignore any provided | |
2141 | # cachedelta. |
|
2141 | # cachedelta. | |
2142 | if rawtext != text: |
|
2142 | if rawtext != text: | |
2143 | cachedelta = None |
|
2143 | cachedelta = None | |
2144 |
|
2144 | |||
2145 | if len(rawtext) > _maxentrysize: |
|
2145 | if len(rawtext) > _maxentrysize: | |
2146 | raise error.RevlogError( |
|
2146 | raise error.RevlogError( | |
2147 | _( |
|
2147 | _( | |
2148 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" |
|
2148 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" | |
2149 | ) |
|
2149 | ) | |
2150 | % (self.indexfile, len(rawtext)) |
|
2150 | % (self.indexfile, len(rawtext)) | |
2151 | ) |
|
2151 | ) | |
2152 |
|
2152 | |||
2153 | node = node or self.hash(rawtext, p1, p2) |
|
2153 | node = node or self.hash(rawtext, p1, p2) | |
2154 | rev = self.index.get_rev(node) |
|
2154 | rev = self.index.get_rev(node) | |
2155 | if rev is not None: |
|
2155 | if rev is not None: | |
2156 | return rev |
|
2156 | return rev | |
2157 |
|
2157 | |||
2158 | if validatehash: |
|
2158 | if validatehash: | |
2159 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
2159 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
2160 |
|
2160 | |||
2161 | return self.addrawrevision( |
|
2161 | return self.addrawrevision( | |
2162 | rawtext, |
|
2162 | rawtext, | |
2163 | transaction, |
|
2163 | transaction, | |
2164 | link, |
|
2164 | link, | |
2165 | p1, |
|
2165 | p1, | |
2166 | p2, |
|
2166 | p2, | |
2167 | node, |
|
2167 | node, | |
2168 | flags, |
|
2168 | flags, | |
2169 | cachedelta=cachedelta, |
|
2169 | cachedelta=cachedelta, | |
2170 | deltacomputer=deltacomputer, |
|
2170 | deltacomputer=deltacomputer, | |
2171 | sidedata=sidedata, |
|
2171 | sidedata=sidedata, | |
2172 | ) |
|
2172 | ) | |
2173 |
|
2173 | |||
2174 | def addrawrevision( |
|
2174 | def addrawrevision( | |
2175 | self, |
|
2175 | self, | |
2176 | rawtext, |
|
2176 | rawtext, | |
2177 | transaction, |
|
2177 | transaction, | |
2178 | link, |
|
2178 | link, | |
2179 | p1, |
|
2179 | p1, | |
2180 | p2, |
|
2180 | p2, | |
2181 | node, |
|
2181 | node, | |
2182 | flags, |
|
2182 | flags, | |
2183 | cachedelta=None, |
|
2183 | cachedelta=None, | |
2184 | deltacomputer=None, |
|
2184 | deltacomputer=None, | |
2185 | sidedata=None, |
|
2185 | sidedata=None, | |
2186 | ): |
|
2186 | ): | |
2187 | """add a raw revision with known flags, node and parents |
|
2187 | """add a raw revision with known flags, node and parents | |
2188 | useful when reusing a revision not stored in this revlog (ex: received |
|
2188 | useful when reusing a revision not stored in this revlog (ex: received | |
2189 | over wire, or read from an external bundle). |
|
2189 | over wire, or read from an external bundle). | |
2190 | """ |
|
2190 | """ | |
2191 | dfh = None |
|
2191 | dfh = None | |
2192 | if not self._inline: |
|
2192 | if not self._inline: | |
2193 | dfh = self._datafp(b"a+") |
|
2193 | dfh = self._datafp(b"a+") | |
2194 | ifh = self._indexfp(b"a+") |
|
2194 | ifh = self._indexfp(b"a+") | |
2195 | try: |
|
2195 | try: | |
2196 | return self._addrevision( |
|
2196 | return self._addrevision( | |
2197 | node, |
|
2197 | node, | |
2198 | rawtext, |
|
2198 | rawtext, | |
2199 | transaction, |
|
2199 | transaction, | |
2200 | link, |
|
2200 | link, | |
2201 | p1, |
|
2201 | p1, | |
2202 | p2, |
|
2202 | p2, | |
2203 | flags, |
|
2203 | flags, | |
2204 | cachedelta, |
|
2204 | cachedelta, | |
2205 | ifh, |
|
2205 | ifh, | |
2206 | dfh, |
|
2206 | dfh, | |
2207 | deltacomputer=deltacomputer, |
|
2207 | deltacomputer=deltacomputer, | |
2208 | sidedata=sidedata, |
|
2208 | sidedata=sidedata, | |
2209 | ) |
|
2209 | ) | |
2210 | finally: |
|
2210 | finally: | |
2211 | if dfh: |
|
2211 | if dfh: | |
2212 | dfh.close() |
|
2212 | dfh.close() | |
2213 | ifh.close() |
|
2213 | ifh.close() | |
2214 |
|
2214 | |||
2215 | def compress(self, data): |
|
2215 | def compress(self, data): | |
2216 | """Generate a possibly-compressed representation of data.""" |
|
2216 | """Generate a possibly-compressed representation of data.""" | |
2217 | if not data: |
|
2217 | if not data: | |
2218 | return b'', data |
|
2218 | return b'', data | |
2219 |
|
2219 | |||
2220 | compressed = self._compressor.compress(data) |
|
2220 | compressed = self._compressor.compress(data) | |
2221 |
|
2221 | |||
2222 | if compressed: |
|
2222 | if compressed: | |
2223 | # The revlog compressor added the header in the returned data. |
|
2223 | # The revlog compressor added the header in the returned data. | |
2224 | return b'', compressed |
|
2224 | return b'', compressed | |
2225 |
|
2225 | |||
2226 | if data[0:1] == b'\0': |
|
2226 | if data[0:1] == b'\0': | |
2227 | return b'', data |
|
2227 | return b'', data | |
2228 | return b'u', data |
|
2228 | return b'u', data | |
2229 |
|
2229 | |||
2230 | def decompress(self, data): |
|
2230 | def decompress(self, data): | |
2231 | """Decompress a revlog chunk. |
|
2231 | """Decompress a revlog chunk. | |
2232 |
|
2232 | |||
2233 | The chunk is expected to begin with a header identifying the |
|
2233 | The chunk is expected to begin with a header identifying the | |
2234 | format type so it can be routed to an appropriate decompressor. |
|
2234 | format type so it can be routed to an appropriate decompressor. | |
2235 | """ |
|
2235 | """ | |
2236 | if not data: |
|
2236 | if not data: | |
2237 | return data |
|
2237 | return data | |
2238 |
|
2238 | |||
2239 | # Revlogs are read much more frequently than they are written and many |
|
2239 | # Revlogs are read much more frequently than they are written and many | |
2240 | # chunks only take microseconds to decompress, so performance is |
|
2240 | # chunks only take microseconds to decompress, so performance is | |
2241 | # important here. |
|
2241 | # important here. | |
2242 | # |
|
2242 | # | |
2243 | # We can make a few assumptions about revlogs: |
|
2243 | # We can make a few assumptions about revlogs: | |
2244 | # |
|
2244 | # | |
2245 | # 1) the majority of chunks will be compressed (as opposed to inline |
|
2245 | # 1) the majority of chunks will be compressed (as opposed to inline | |
2246 | # raw data). |
|
2246 | # raw data). | |
2247 | # 2) decompressing *any* data will likely by at least 10x slower than |
|
2247 | # 2) decompressing *any* data will likely by at least 10x slower than | |
2248 | # returning raw inline data. |
|
2248 | # returning raw inline data. | |
2249 | # 3) we want to prioritize common and officially supported compression |
|
2249 | # 3) we want to prioritize common and officially supported compression | |
2250 | # engines |
|
2250 | # engines | |
2251 | # |
|
2251 | # | |
2252 | # It follows that we want to optimize for "decompress compressed data |
|
2252 | # It follows that we want to optimize for "decompress compressed data | |
2253 | # when encoded with common and officially supported compression engines" |
|
2253 | # when encoded with common and officially supported compression engines" | |
2254 | # case over "raw data" and "data encoded by less common or non-official |
|
2254 | # case over "raw data" and "data encoded by less common or non-official | |
2255 | # compression engines." That is why we have the inline lookup first |
|
2255 | # compression engines." That is why we have the inline lookup first | |
2256 | # followed by the compengines lookup. |
|
2256 | # followed by the compengines lookup. | |
2257 | # |
|
2257 | # | |
2258 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib |
|
2258 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib | |
2259 | # compressed chunks. And this matters for changelog and manifest reads. |
|
2259 | # compressed chunks. And this matters for changelog and manifest reads. | |
2260 | t = data[0:1] |
|
2260 | t = data[0:1] | |
2261 |
|
2261 | |||
2262 | if t == b'x': |
|
2262 | if t == b'x': | |
2263 | try: |
|
2263 | try: | |
2264 | return _zlibdecompress(data) |
|
2264 | return _zlibdecompress(data) | |
2265 | except zlib.error as e: |
|
2265 | except zlib.error as e: | |
2266 | raise error.RevlogError( |
|
2266 | raise error.RevlogError( | |
2267 | _(b'revlog decompress error: %s') |
|
2267 | _(b'revlog decompress error: %s') | |
2268 | % stringutil.forcebytestr(e) |
|
2268 | % stringutil.forcebytestr(e) | |
2269 | ) |
|
2269 | ) | |
2270 | # '\0' is more common than 'u' so it goes first. |
|
2270 | # '\0' is more common than 'u' so it goes first. | |
2271 | elif t == b'\0': |
|
2271 | elif t == b'\0': | |
2272 | return data |
|
2272 | return data | |
2273 | elif t == b'u': |
|
2273 | elif t == b'u': | |
2274 | return util.buffer(data, 1) |
|
2274 | return util.buffer(data, 1) | |
2275 |
|
2275 | |||
2276 | try: |
|
2276 | try: | |
2277 | compressor = self._decompressors[t] |
|
2277 | compressor = self._decompressors[t] | |
2278 | except KeyError: |
|
2278 | except KeyError: | |
2279 | try: |
|
2279 | try: | |
2280 | engine = util.compengines.forrevlogheader(t) |
|
2280 | engine = util.compengines.forrevlogheader(t) | |
2281 | compressor = engine.revlogcompressor(self._compengineopts) |
|
2281 | compressor = engine.revlogcompressor(self._compengineopts) | |
2282 | self._decompressors[t] = compressor |
|
2282 | self._decompressors[t] = compressor | |
2283 | except KeyError: |
|
2283 | except KeyError: | |
2284 | raise error.RevlogError(_(b'unknown compression type %r') % t) |
|
2284 | raise error.RevlogError(_(b'unknown compression type %r') % t) | |
2285 |
|
2285 | |||
2286 | return compressor.decompress(data) |
|
2286 | return compressor.decompress(data) | |
2287 |
|
2287 | |||
2288 | def _addrevision( |
|
2288 | def _addrevision( | |
2289 | self, |
|
2289 | self, | |
2290 | node, |
|
2290 | node, | |
2291 | rawtext, |
|
2291 | rawtext, | |
2292 | transaction, |
|
2292 | transaction, | |
2293 | link, |
|
2293 | link, | |
2294 | p1, |
|
2294 | p1, | |
2295 | p2, |
|
2295 | p2, | |
2296 | flags, |
|
2296 | flags, | |
2297 | cachedelta, |
|
2297 | cachedelta, | |
2298 | ifh, |
|
2298 | ifh, | |
2299 | dfh, |
|
2299 | dfh, | |
2300 | alwayscache=False, |
|
2300 | alwayscache=False, | |
2301 | deltacomputer=None, |
|
2301 | deltacomputer=None, | |
2302 | sidedata=None, |
|
2302 | sidedata=None, | |
2303 | ): |
|
2303 | ): | |
2304 | """internal function to add revisions to the log |
|
2304 | """internal function to add revisions to the log | |
2305 |
|
2305 | |||
2306 | see addrevision for argument descriptions. |
|
2306 | see addrevision for argument descriptions. | |
2307 |
|
2307 | |||
2308 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. |
|
2308 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. | |
2309 |
|
2309 | |||
2310 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will |
|
2310 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will | |
2311 | be used. |
|
2311 | be used. | |
2312 |
|
2312 | |||
2313 | invariants: |
|
2313 | invariants: | |
2314 | - rawtext is optional (can be None); if not set, cachedelta must be set. |
|
2314 | - rawtext is optional (can be None); if not set, cachedelta must be set. | |
2315 | if both are set, they must correspond to each other. |
|
2315 | if both are set, they must correspond to each other. | |
2316 | """ |
|
2316 | """ | |
2317 | if node == nullid: |
|
2317 | if node == nullid: | |
2318 | raise error.RevlogError( |
|
2318 | raise error.RevlogError( | |
2319 | _(b"%s: attempt to add null revision") % self.indexfile |
|
2319 | _(b"%s: attempt to add null revision") % self.indexfile | |
2320 | ) |
|
2320 | ) | |
2321 | if node == wdirid or node in wdirfilenodeids: |
|
2321 | if node == wdirid or node in wdirfilenodeids: | |
2322 | raise error.RevlogError( |
|
2322 | raise error.RevlogError( | |
2323 | _(b"%s: attempt to add wdir revision") % self.indexfile |
|
2323 | _(b"%s: attempt to add wdir revision") % self.indexfile | |
2324 | ) |
|
2324 | ) | |
2325 |
|
2325 | |||
2326 | if self._inline: |
|
2326 | if self._inline: | |
2327 | fh = ifh |
|
2327 | fh = ifh | |
2328 | else: |
|
2328 | else: | |
2329 | fh = dfh |
|
2329 | fh = dfh | |
2330 |
|
2330 | |||
2331 | btext = [rawtext] |
|
2331 | btext = [rawtext] | |
2332 |
|
2332 | |||
2333 | curr = len(self) |
|
2333 | curr = len(self) | |
2334 | prev = curr - 1 |
|
2334 | prev = curr - 1 | |
2335 |
|
2335 | |||
2336 | offset = self._get_data_offset(prev) |
|
2336 | offset = self._get_data_offset(prev) | |
2337 |
|
2337 | |||
2338 | if self._concurrencychecker: |
|
2338 | if self._concurrencychecker: | |
2339 | if self._inline: |
|
2339 | if self._inline: | |
2340 | # offset is "as if" it were in the .d file, so we need to add on |
|
2340 | # offset is "as if" it were in the .d file, so we need to add on | |
2341 | # the size of the entry metadata. |
|
2341 | # the size of the entry metadata. | |
2342 | self._concurrencychecker( |
|
2342 | self._concurrencychecker( | |
2343 | ifh, self.indexfile, offset + curr * self._io.size |
|
2343 | ifh, self.indexfile, offset + curr * self._io.size | |
2344 | ) |
|
2344 | ) | |
2345 | else: |
|
2345 | else: | |
2346 | # Entries in the .i are a consistent size. |
|
2346 | # Entries in the .i are a consistent size. | |
2347 | self._concurrencychecker( |
|
2347 | self._concurrencychecker( | |
2348 | ifh, self.indexfile, curr * self._io.size |
|
2348 | ifh, self.indexfile, curr * self._io.size | |
2349 | ) |
|
2349 | ) | |
2350 | self._concurrencychecker(dfh, self.datafile, offset) |
|
2350 | self._concurrencychecker(dfh, self.datafile, offset) | |
2351 |
|
2351 | |||
2352 | p1r, p2r = self.rev(p1), self.rev(p2) |
|
2352 | p1r, p2r = self.rev(p1), self.rev(p2) | |
2353 |
|
2353 | |||
2354 | # full versions are inserted when the needed deltas |
|
2354 | # full versions are inserted when the needed deltas | |
2355 | # become comparable to the uncompressed text |
|
2355 | # become comparable to the uncompressed text | |
2356 | if rawtext is None: |
|
2356 | if rawtext is None: | |
2357 | # need rawtext size, before changed by flag processors, which is |
|
2357 | # need rawtext size, before changed by flag processors, which is | |
2358 | # the non-raw size. use revlog explicitly to avoid filelog's extra |
|
2358 | # the non-raw size. use revlog explicitly to avoid filelog's extra | |
2359 | # logic that might remove metadata size. |
|
2359 | # logic that might remove metadata size. | |
2360 | textlen = mdiff.patchedsize( |
|
2360 | textlen = mdiff.patchedsize( | |
2361 | revlog.size(self, cachedelta[0]), cachedelta[1] |
|
2361 | revlog.size(self, cachedelta[0]), cachedelta[1] | |
2362 | ) |
|
2362 | ) | |
2363 | else: |
|
2363 | else: | |
2364 | textlen = len(rawtext) |
|
2364 | textlen = len(rawtext) | |
2365 |
|
2365 | |||
2366 | if deltacomputer is None: |
|
2366 | if deltacomputer is None: | |
2367 | deltacomputer = deltautil.deltacomputer(self) |
|
2367 | deltacomputer = deltautil.deltacomputer(self) | |
2368 |
|
2368 | |||
2369 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) |
|
2369 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) | |
2370 |
|
2370 | |||
2371 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) |
|
2371 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) | |
2372 |
|
2372 | |||
2373 | if sidedata: |
|
2373 | if sidedata: | |
2374 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) |
|
2374 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) | |
2375 | sidedata_offset = offset + deltainfo.deltalen |
|
2375 | sidedata_offset = offset + deltainfo.deltalen | |
2376 | else: |
|
2376 | else: | |
2377 | serialized_sidedata = b"" |
|
2377 | serialized_sidedata = b"" | |
2378 | # Don't store the offset if the sidedata is empty, that way |
|
2378 | # Don't store the offset if the sidedata is empty, that way | |
2379 | # we can easily detect empty sidedata and they will be no different |
|
2379 | # we can easily detect empty sidedata and they will be no different | |
2380 | # than ones we manually add. |
|
2380 | # than ones we manually add. | |
2381 | sidedata_offset = 0 |
|
2381 | sidedata_offset = 0 | |
2382 |
|
2382 | |||
2383 | e = ( |
|
2383 | e = ( | |
2384 | offset_type(offset, flags), |
|
2384 | offset_type(offset, flags), | |
2385 | deltainfo.deltalen, |
|
2385 | deltainfo.deltalen, | |
2386 | textlen, |
|
2386 | textlen, | |
2387 | deltainfo.base, |
|
2387 | deltainfo.base, | |
2388 | link, |
|
2388 | link, | |
2389 | p1r, |
|
2389 | p1r, | |
2390 | p2r, |
|
2390 | p2r, | |
2391 | node, |
|
2391 | node, | |
2392 | sidedata_offset, |
|
2392 | sidedata_offset, | |
2393 | len(serialized_sidedata), |
|
2393 | len(serialized_sidedata), | |
2394 | ) |
|
2394 | ) | |
2395 |
|
2395 | |||
2396 | if self.version & 0xFFFF != REVLOGV2: |
|
2396 | if self.version & 0xFFFF != REVLOGV2: | |
2397 | e = e[:8] |
|
2397 | e = e[:8] | |
2398 |
|
2398 | |||
2399 | self.index.append(e) |
|
2399 | self.index.append(e) | |
2400 | entry = self._io.packentry(e, self.node, self.version, curr) |
|
2400 | entry = self._io.packentry(e, self.node, self.version, curr) | |
2401 | self._writeentry( |
|
2401 | self._writeentry( | |
2402 | transaction, |
|
2402 | transaction, | |
2403 | ifh, |
|
2403 | ifh, | |
2404 | dfh, |
|
2404 | dfh, | |
2405 | entry, |
|
2405 | entry, | |
2406 | deltainfo.data, |
|
2406 | deltainfo.data, | |
2407 | link, |
|
2407 | link, | |
2408 | offset, |
|
2408 | offset, | |
2409 | serialized_sidedata, |
|
2409 | serialized_sidedata, | |
2410 | ) |
|
2410 | ) | |
2411 |
|
2411 | |||
2412 | rawtext = btext[0] |
|
2412 | rawtext = btext[0] | |
2413 |
|
2413 | |||
2414 | if alwayscache and rawtext is None: |
|
2414 | if alwayscache and rawtext is None: | |
2415 | rawtext = deltacomputer.buildtext(revinfo, fh) |
|
2415 | rawtext = deltacomputer.buildtext(revinfo, fh) | |
2416 |
|
2416 | |||
2417 | if type(rawtext) == bytes: # only accept immutable objects |
|
2417 | if type(rawtext) == bytes: # only accept immutable objects | |
2418 | self._revisioncache = (node, curr, rawtext) |
|
2418 | self._revisioncache = (node, curr, rawtext) | |
2419 | self._chainbasecache[curr] = deltainfo.chainbase |
|
2419 | self._chainbasecache[curr] = deltainfo.chainbase | |
2420 | return curr |
|
2420 | return curr | |
2421 |
|
2421 | |||
2422 | def _get_data_offset(self, prev): |
|
2422 | def _get_data_offset(self, prev): | |
2423 | """Returns the current offset in the (in-transaction) data file. |
|
2423 | """Returns the current offset in the (in-transaction) data file. | |
2424 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket |
|
2424 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket | |
2425 | file to store that information: since sidedata can be rewritten to the |
|
2425 | file to store that information: since sidedata can be rewritten to the | |
2426 | end of the data file within a transaction, you can have cases where, for |
|
2426 | end of the data file within a transaction, you can have cases where, for | |
2427 | example, rev `n` does not have sidedata while rev `n - 1` does, leading |
|
2427 | example, rev `n` does not have sidedata while rev `n - 1` does, leading | |
2428 | to `n - 1`'s sidedata being written after `n`'s data. |
|
2428 | to `n - 1`'s sidedata being written after `n`'s data. | |
2429 |
|
2429 | |||
2430 | TODO cache this in a docket file before getting out of experimental.""" |
|
2430 | TODO cache this in a docket file before getting out of experimental.""" | |
2431 | if self.version & 0xFFFF != REVLOGV2: |
|
2431 | if self.version & 0xFFFF != REVLOGV2: | |
2432 | return self.end(prev) |
|
2432 | return self.end(prev) | |
2433 |
|
2433 | |||
2434 | offset = 0 |
|
2434 | offset = 0 | |
2435 | for rev, entry in enumerate(self.index): |
|
2435 | for rev, entry in enumerate(self.index): | |
2436 | sidedata_end = entry[8] + entry[9] |
|
2436 | sidedata_end = entry[8] + entry[9] | |
2437 | # Sidedata for a previous rev has potentially been written after |
|
2437 | # Sidedata for a previous rev has potentially been written after | |
2438 | # this rev's end, so take the max. |
|
2438 | # this rev's end, so take the max. | |
2439 | offset = max(self.end(rev), offset, sidedata_end) |
|
2439 | offset = max(self.end(rev), offset, sidedata_end) | |
2440 | return offset |
|
2440 | return offset | |
2441 |
|
2441 | |||
2442 | def _writeentry( |
|
2442 | def _writeentry( | |
2443 | self, transaction, ifh, dfh, entry, data, link, offset, sidedata |
|
2443 | self, transaction, ifh, dfh, entry, data, link, offset, sidedata | |
2444 | ): |
|
2444 | ): | |
2445 | # Files opened in a+ mode have inconsistent behavior on various |
|
2445 | # Files opened in a+ mode have inconsistent behavior on various | |
2446 | # platforms. Windows requires that a file positioning call be made |
|
2446 | # platforms. Windows requires that a file positioning call be made | |
2447 | # when the file handle transitions between reads and writes. See |
|
2447 | # when the file handle transitions between reads and writes. See | |
2448 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
2448 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other | |
2449 | # platforms, Python or the platform itself can be buggy. Some versions |
|
2449 | # platforms, Python or the platform itself can be buggy. Some versions | |
2450 | # of Solaris have been observed to not append at the end of the file |
|
2450 | # of Solaris have been observed to not append at the end of the file | |
2451 | # if the file was seeked to before the end. See issue4943 for more. |
|
2451 | # if the file was seeked to before the end. See issue4943 for more. | |
2452 | # |
|
2452 | # | |
2453 | # We work around this issue by inserting a seek() before writing. |
|
2453 | # We work around this issue by inserting a seek() before writing. | |
2454 | # Note: This is likely not necessary on Python 3. However, because |
|
2454 | # Note: This is likely not necessary on Python 3. However, because | |
2455 | # the file handle is reused for reads and may be seeked there, we need |
|
2455 | # the file handle is reused for reads and may be seeked there, we need | |
2456 | # to be careful before changing this. |
|
2456 | # to be careful before changing this. | |
2457 | ifh.seek(0, os.SEEK_END) |
|
2457 | ifh.seek(0, os.SEEK_END) | |
2458 | if dfh: |
|
2458 | if dfh: | |
2459 | dfh.seek(0, os.SEEK_END) |
|
2459 | dfh.seek(0, os.SEEK_END) | |
2460 |
|
2460 | |||
2461 | curr = len(self) - 1 |
|
2461 | curr = len(self) - 1 | |
2462 | if not self._inline: |
|
2462 | if not self._inline: | |
2463 | transaction.add(self.datafile, offset) |
|
2463 | transaction.add(self.datafile, offset) | |
2464 | transaction.add(self.indexfile, curr * len(entry)) |
|
2464 | transaction.add(self.indexfile, curr * len(entry)) | |
2465 | if data[0]: |
|
2465 | if data[0]: | |
2466 | dfh.write(data[0]) |
|
2466 | dfh.write(data[0]) | |
2467 | dfh.write(data[1]) |
|
2467 | dfh.write(data[1]) | |
2468 | if sidedata: |
|
2468 | if sidedata: | |
2469 | dfh.write(sidedata) |
|
2469 | dfh.write(sidedata) | |
2470 | ifh.write(entry) |
|
2470 | ifh.write(entry) | |
2471 | else: |
|
2471 | else: | |
2472 | offset += curr * self._io.size |
|
2472 | offset += curr * self._io.size | |
2473 | transaction.add(self.indexfile, offset) |
|
2473 | transaction.add(self.indexfile, offset) | |
2474 | ifh.write(entry) |
|
2474 | ifh.write(entry) | |
2475 | ifh.write(data[0]) |
|
2475 | ifh.write(data[0]) | |
2476 | ifh.write(data[1]) |
|
2476 | ifh.write(data[1]) | |
2477 | if sidedata: |
|
2477 | if sidedata: | |
2478 | ifh.write(sidedata) |
|
2478 | ifh.write(sidedata) | |
2479 | self._enforceinlinesize(transaction, ifh) |
|
2479 | self._enforceinlinesize(transaction, ifh) | |
2480 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
2480 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
2481 |
|
2481 | |||
2482 | def addgroup( |
|
2482 | def addgroup( | |
2483 | self, |
|
2483 | self, | |
2484 | deltas, |
|
2484 | deltas, | |
2485 | linkmapper, |
|
2485 | linkmapper, | |
2486 | transaction, |
|
2486 | transaction, | |
2487 | alwayscache=False, |
|
2487 | alwayscache=False, | |
2488 | addrevisioncb=None, |
|
2488 | addrevisioncb=None, | |
2489 | duplicaterevisioncb=None, |
|
2489 | duplicaterevisioncb=None, | |
2490 | ): |
|
2490 | ): | |
2491 | """ |
|
2491 | """ | |
2492 | add a delta group |
|
2492 | add a delta group | |
2493 |
|
2493 | |||
2494 | given a set of deltas, add them to the revision log. the |
|
2494 | given a set of deltas, add them to the revision log. the | |
2495 | first delta is against its parent, which should be in our |
|
2495 | first delta is against its parent, which should be in our | |
2496 | log, the rest are against the previous delta. |
|
2496 | log, the rest are against the previous delta. | |
2497 |
|
2497 | |||
2498 | If ``addrevisioncb`` is defined, it will be called with arguments of |
|
2498 | If ``addrevisioncb`` is defined, it will be called with arguments of | |
2499 | this revlog and the node that was added. |
|
2499 | this revlog and the node that was added. | |
2500 | """ |
|
2500 | """ | |
2501 |
|
2501 | |||
2502 | if self._writinghandles: |
|
2502 | if self._writinghandles: | |
2503 | raise error.ProgrammingError(b'cannot nest addgroup() calls') |
|
2503 | raise error.ProgrammingError(b'cannot nest addgroup() calls') | |
2504 |
|
2504 | |||
2505 | r = len(self) |
|
2505 | r = len(self) | |
2506 | end = 0 |
|
2506 | end = 0 | |
2507 | if r: |
|
2507 | if r: | |
2508 | end = self.end(r - 1) |
|
2508 | end = self.end(r - 1) | |
2509 | ifh = self._indexfp(b"a+") |
|
2509 | ifh = self._indexfp(b"a+") | |
2510 | isize = r * self._io.size |
|
2510 | isize = r * self._io.size | |
2511 | if self._inline: |
|
2511 | if self._inline: | |
2512 | transaction.add(self.indexfile, end + isize) |
|
2512 | transaction.add(self.indexfile, end + isize) | |
2513 | dfh = None |
|
2513 | dfh = None | |
2514 | else: |
|
2514 | else: | |
2515 | transaction.add(self.indexfile, isize) |
|
2515 | transaction.add(self.indexfile, isize) | |
2516 | transaction.add(self.datafile, end) |
|
2516 | transaction.add(self.datafile, end) | |
2517 | dfh = self._datafp(b"a+") |
|
2517 | dfh = self._datafp(b"a+") | |
2518 |
|
2518 | |||
2519 | def flush(): |
|
2519 | def flush(): | |
2520 | if dfh: |
|
2520 | if dfh: | |
2521 | dfh.flush() |
|
2521 | dfh.flush() | |
2522 | ifh.flush() |
|
2522 | ifh.flush() | |
2523 |
|
2523 | |||
2524 | self._writinghandles = (ifh, dfh) |
|
2524 | self._writinghandles = (ifh, dfh) | |
2525 | empty = True |
|
2525 | empty = True | |
2526 |
|
2526 | |||
2527 | try: |
|
2527 | try: | |
2528 | deltacomputer = deltautil.deltacomputer(self) |
|
2528 | deltacomputer = deltautil.deltacomputer(self) | |
2529 | # loop through our set of deltas |
|
2529 | # loop through our set of deltas | |
2530 | for data in deltas: |
|
2530 | for data in deltas: | |
2531 | node, p1, p2, linknode, deltabase, delta, flags, sidedata = data |
|
2531 | node, p1, p2, linknode, deltabase, delta, flags, sidedata = data | |
2532 | link = linkmapper(linknode) |
|
2532 | link = linkmapper(linknode) | |
2533 | flags = flags or REVIDX_DEFAULT_FLAGS |
|
2533 | flags = flags or REVIDX_DEFAULT_FLAGS | |
2534 |
|
2534 | |||
2535 | rev = self.index.get_rev(node) |
|
2535 | rev = self.index.get_rev(node) | |
2536 | if rev is not None: |
|
2536 | if rev is not None: | |
2537 | # this can happen if two branches make the same change |
|
2537 | # this can happen if two branches make the same change | |
2538 | self._nodeduplicatecallback(transaction, rev) |
|
2538 | self._nodeduplicatecallback(transaction, rev) | |
2539 | if duplicaterevisioncb: |
|
2539 | if duplicaterevisioncb: | |
2540 | duplicaterevisioncb(self, rev) |
|
2540 | duplicaterevisioncb(self, rev) | |
2541 | empty = False |
|
2541 | empty = False | |
2542 | continue |
|
2542 | continue | |
2543 |
|
2543 | |||
2544 | for p in (p1, p2): |
|
2544 | for p in (p1, p2): | |
2545 | if not self.index.has_node(p): |
|
2545 | if not self.index.has_node(p): | |
2546 | raise error.LookupError( |
|
2546 | raise error.LookupError( | |
2547 | p, self.indexfile, _(b'unknown parent') |
|
2547 | p, self.indexfile, _(b'unknown parent') | |
2548 | ) |
|
2548 | ) | |
2549 |
|
2549 | |||
2550 | if not self.index.has_node(deltabase): |
|
2550 | if not self.index.has_node(deltabase): | |
2551 | raise error.LookupError( |
|
2551 | raise error.LookupError( | |
2552 | deltabase, self.indexfile, _(b'unknown delta base') |
|
2552 | deltabase, self.indexfile, _(b'unknown delta base') | |
2553 | ) |
|
2553 | ) | |
2554 |
|
2554 | |||
2555 | baserev = self.rev(deltabase) |
|
2555 | baserev = self.rev(deltabase) | |
2556 |
|
2556 | |||
2557 | if baserev != nullrev and self.iscensored(baserev): |
|
2557 | if baserev != nullrev and self.iscensored(baserev): | |
2558 | # if base is censored, delta must be full replacement in a |
|
2558 | # if base is censored, delta must be full replacement in a | |
2559 | # single patch operation |
|
2559 | # single patch operation | |
2560 | hlen = struct.calcsize(b">lll") |
|
2560 | hlen = struct.calcsize(b">lll") | |
2561 | oldlen = self.rawsize(baserev) |
|
2561 | oldlen = self.rawsize(baserev) | |
2562 | newlen = len(delta) - hlen |
|
2562 | newlen = len(delta) - hlen | |
2563 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): |
|
2563 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): | |
2564 | raise error.CensoredBaseError( |
|
2564 | raise error.CensoredBaseError( | |
2565 | self.indexfile, self.node(baserev) |
|
2565 | self.indexfile, self.node(baserev) | |
2566 | ) |
|
2566 | ) | |
2567 |
|
2567 | |||
2568 | if not flags and self._peek_iscensored(baserev, delta, flush): |
|
2568 | if not flags and self._peek_iscensored(baserev, delta, flush): | |
2569 | flags |= REVIDX_ISCENSORED |
|
2569 | flags |= REVIDX_ISCENSORED | |
2570 |
|
2570 | |||
2571 | # We assume consumers of addrevisioncb will want to retrieve |
|
2571 | # We assume consumers of addrevisioncb will want to retrieve | |
2572 | # the added revision, which will require a call to |
|
2572 | # the added revision, which will require a call to | |
2573 | # revision(). revision() will fast path if there is a cache |
|
2573 | # revision(). revision() will fast path if there is a cache | |
2574 | # hit. So, we tell _addrevision() to always cache in this case. |
|
2574 | # hit. So, we tell _addrevision() to always cache in this case. | |
2575 | # We're only using addgroup() in the context of changegroup |
|
2575 | # We're only using addgroup() in the context of changegroup | |
2576 | # generation so the revision data can always be handled as raw |
|
2576 | # generation so the revision data can always be handled as raw | |
2577 | # by the flagprocessor. |
|
2577 | # by the flagprocessor. | |
2578 | rev = self._addrevision( |
|
2578 | rev = self._addrevision( | |
2579 | node, |
|
2579 | node, | |
2580 | None, |
|
2580 | None, | |
2581 | transaction, |
|
2581 | transaction, | |
2582 | link, |
|
2582 | link, | |
2583 | p1, |
|
2583 | p1, | |
2584 | p2, |
|
2584 | p2, | |
2585 | flags, |
|
2585 | flags, | |
2586 | (baserev, delta), |
|
2586 | (baserev, delta), | |
2587 | ifh, |
|
2587 | ifh, | |
2588 | dfh, |
|
2588 | dfh, | |
2589 | alwayscache=alwayscache, |
|
2589 | alwayscache=alwayscache, | |
2590 | deltacomputer=deltacomputer, |
|
2590 | deltacomputer=deltacomputer, | |
2591 | sidedata=sidedata, |
|
2591 | sidedata=sidedata, | |
2592 | ) |
|
2592 | ) | |
2593 |
|
2593 | |||
2594 | if addrevisioncb: |
|
2594 | if addrevisioncb: | |
2595 | addrevisioncb(self, rev) |
|
2595 | addrevisioncb(self, rev) | |
2596 | empty = False |
|
2596 | empty = False | |
2597 |
|
2597 | |||
2598 | if not dfh and not self._inline: |
|
2598 | if not dfh and not self._inline: | |
2599 | # addrevision switched from inline to conventional |
|
2599 | # addrevision switched from inline to conventional | |
2600 | # reopen the index |
|
2600 | # reopen the index | |
2601 | ifh.close() |
|
2601 | ifh.close() | |
2602 | dfh = self._datafp(b"a+") |
|
2602 | dfh = self._datafp(b"a+") | |
2603 | ifh = self._indexfp(b"a+") |
|
2603 | ifh = self._indexfp(b"a+") | |
2604 | self._writinghandles = (ifh, dfh) |
|
2604 | self._writinghandles = (ifh, dfh) | |
2605 | finally: |
|
2605 | finally: | |
2606 | self._writinghandles = None |
|
2606 | self._writinghandles = None | |
2607 |
|
2607 | |||
2608 | if dfh: |
|
2608 | if dfh: | |
2609 | dfh.close() |
|
2609 | dfh.close() | |
2610 | ifh.close() |
|
2610 | ifh.close() | |
2611 | return not empty |
|
2611 | return not empty | |
2612 |
|
2612 | |||
2613 | def iscensored(self, rev): |
|
2613 | def iscensored(self, rev): | |
2614 | """Check if a file revision is censored.""" |
|
2614 | """Check if a file revision is censored.""" | |
2615 | if not self._censorable: |
|
2615 | if not self._censorable: | |
2616 | return False |
|
2616 | return False | |
2617 |
|
2617 | |||
2618 | return self.flags(rev) & REVIDX_ISCENSORED |
|
2618 | return self.flags(rev) & REVIDX_ISCENSORED | |
2619 |
|
2619 | |||
2620 | def _peek_iscensored(self, baserev, delta, flush): |
|
2620 | def _peek_iscensored(self, baserev, delta, flush): | |
2621 | """Quickly check if a delta produces a censored revision.""" |
|
2621 | """Quickly check if a delta produces a censored revision.""" | |
2622 | if not self._censorable: |
|
2622 | if not self._censorable: | |
2623 | return False |
|
2623 | return False | |
2624 |
|
2624 | |||
2625 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) |
|
2625 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) | |
2626 |
|
2626 | |||
2627 | def getstrippoint(self, minlink): |
|
2627 | def getstrippoint(self, minlink): | |
2628 | """find the minimum rev that must be stripped to strip the linkrev |
|
2628 | """find the minimum rev that must be stripped to strip the linkrev | |
2629 |
|
2629 | |||
2630 | Returns a tuple containing the minimum rev and a set of all revs that |
|
2630 | Returns a tuple containing the minimum rev and a set of all revs that | |
2631 | have linkrevs that will be broken by this strip. |
|
2631 | have linkrevs that will be broken by this strip. | |
2632 | """ |
|
2632 | """ | |
2633 | return storageutil.resolvestripinfo( |
|
2633 | return storageutil.resolvestripinfo( | |
2634 | minlink, |
|
2634 | minlink, | |
2635 | len(self) - 1, |
|
2635 | len(self) - 1, | |
2636 | self.headrevs(), |
|
2636 | self.headrevs(), | |
2637 | self.linkrev, |
|
2637 | self.linkrev, | |
2638 | self.parentrevs, |
|
2638 | self.parentrevs, | |
2639 | ) |
|
2639 | ) | |
2640 |
|
2640 | |||
2641 | def strip(self, minlink, transaction): |
|
2641 | def strip(self, minlink, transaction): | |
2642 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
2642 | """truncate the revlog on the first revision with a linkrev >= minlink | |
2643 |
|
2643 | |||
2644 | This function is called when we're stripping revision minlink and |
|
2644 | This function is called when we're stripping revision minlink and | |
2645 | its descendants from the repository. |
|
2645 | its descendants from the repository. | |
2646 |
|
2646 | |||
2647 | We have to remove all revisions with linkrev >= minlink, because |
|
2647 | We have to remove all revisions with linkrev >= minlink, because | |
2648 | the equivalent changelog revisions will be renumbered after the |
|
2648 | the equivalent changelog revisions will be renumbered after the | |
2649 | strip. |
|
2649 | strip. | |
2650 |
|
2650 | |||
2651 | So we truncate the revlog on the first of these revisions, and |
|
2651 | So we truncate the revlog on the first of these revisions, and | |
2652 | trust that the caller has saved the revisions that shouldn't be |
|
2652 | trust that the caller has saved the revisions that shouldn't be | |
2653 | removed and that it'll re-add them after this truncation. |
|
2653 | removed and that it'll re-add them after this truncation. | |
2654 | """ |
|
2654 | """ | |
2655 | if len(self) == 0: |
|
2655 | if len(self) == 0: | |
2656 | return |
|
2656 | return | |
2657 |
|
2657 | |||
2658 | rev, _ = self.getstrippoint(minlink) |
|
2658 | rev, _ = self.getstrippoint(minlink) | |
2659 | if rev == len(self): |
|
2659 | if rev == len(self): | |
2660 | return |
|
2660 | return | |
2661 |
|
2661 | |||
2662 | # first truncate the files on disk |
|
2662 | # first truncate the files on disk | |
2663 | end = self.start(rev) |
|
2663 | end = self.start(rev) | |
2664 | if not self._inline: |
|
2664 | if not self._inline: | |
2665 | transaction.add(self.datafile, end) |
|
2665 | transaction.add(self.datafile, end) | |
2666 | end = rev * self._io.size |
|
2666 | end = rev * self._io.size | |
2667 | else: |
|
2667 | else: | |
2668 | end += rev * self._io.size |
|
2668 | end += rev * self._io.size | |
2669 |
|
2669 | |||
2670 | transaction.add(self.indexfile, end) |
|
2670 | transaction.add(self.indexfile, end) | |
2671 |
|
2671 | |||
2672 | # then reset internal state in memory to forget those revisions |
|
2672 | # then reset internal state in memory to forget those revisions | |
2673 | self._revisioncache = None |
|
2673 | self._revisioncache = None | |
2674 | self._chaininfocache = util.lrucachedict(500) |
|
2674 | self._chaininfocache = util.lrucachedict(500) | |
2675 | self._chunkclear() |
|
2675 | self._chunkclear() | |
2676 |
|
2676 | |||
2677 | del self.index[rev:-1] |
|
2677 | del self.index[rev:-1] | |
2678 |
|
2678 | |||
2679 | def checksize(self): |
|
2679 | def checksize(self): | |
2680 | """Check size of index and data files |
|
2680 | """Check size of index and data files | |
2681 |
|
2681 | |||
2682 | return a (dd, di) tuple. |
|
2682 | return a (dd, di) tuple. | |
2683 | - dd: extra bytes for the "data" file |
|
2683 | - dd: extra bytes for the "data" file | |
2684 | - di: extra bytes for the "index" file |
|
2684 | - di: extra bytes for the "index" file | |
2685 |
|
2685 | |||
2686 | A healthy revlog will return (0, 0). |
|
2686 | A healthy revlog will return (0, 0). | |
2687 | """ |
|
2687 | """ | |
2688 | expected = 0 |
|
2688 | expected = 0 | |
2689 | if len(self): |
|
2689 | if len(self): | |
2690 | expected = max(0, self.end(len(self) - 1)) |
|
2690 | expected = max(0, self.end(len(self) - 1)) | |
2691 |
|
2691 | |||
2692 | try: |
|
2692 | try: | |
2693 | with self._datafp() as f: |
|
2693 | with self._datafp() as f: | |
2694 | f.seek(0, io.SEEK_END) |
|
2694 | f.seek(0, io.SEEK_END) | |
2695 | actual = f.tell() |
|
2695 | actual = f.tell() | |
2696 | dd = actual - expected |
|
2696 | dd = actual - expected | |
2697 | except IOError as inst: |
|
2697 | except IOError as inst: | |
2698 | if inst.errno != errno.ENOENT: |
|
2698 | if inst.errno != errno.ENOENT: | |
2699 | raise |
|
2699 | raise | |
2700 | dd = 0 |
|
2700 | dd = 0 | |
2701 |
|
2701 | |||
2702 | try: |
|
2702 | try: | |
2703 | f = self.opener(self.indexfile) |
|
2703 | f = self.opener(self.indexfile) | |
2704 | f.seek(0, io.SEEK_END) |
|
2704 | f.seek(0, io.SEEK_END) | |
2705 | actual = f.tell() |
|
2705 | actual = f.tell() | |
2706 | f.close() |
|
2706 | f.close() | |
2707 | s = self._io.size |
|
2707 | s = self._io.size | |
2708 | i = max(0, actual // s) |
|
2708 | i = max(0, actual // s) | |
2709 | di = actual - (i * s) |
|
2709 | di = actual - (i * s) | |
2710 | if self._inline: |
|
2710 | if self._inline: | |
2711 | databytes = 0 |
|
2711 | databytes = 0 | |
2712 | for r in self: |
|
2712 | for r in self: | |
2713 | databytes += max(0, self.length(r)) |
|
2713 | databytes += max(0, self.length(r)) | |
2714 | dd = 0 |
|
2714 | dd = 0 | |
2715 | di = actual - len(self) * s - databytes |
|
2715 | di = actual - len(self) * s - databytes | |
2716 | except IOError as inst: |
|
2716 | except IOError as inst: | |
2717 | if inst.errno != errno.ENOENT: |
|
2717 | if inst.errno != errno.ENOENT: | |
2718 | raise |
|
2718 | raise | |
2719 | di = 0 |
|
2719 | di = 0 | |
2720 |
|
2720 | |||
2721 | return (dd, di) |
|
2721 | return (dd, di) | |
2722 |
|
2722 | |||
2723 | def files(self): |
|
2723 | def files(self): | |
2724 | res = [self.indexfile] |
|
2724 | res = [self.indexfile] | |
2725 | if not self._inline: |
|
2725 | if not self._inline: | |
2726 | res.append(self.datafile) |
|
2726 | res.append(self.datafile) | |
2727 | return res |
|
2727 | return res | |
2728 |
|
2728 | |||
2729 | def emitrevisions( |
|
2729 | def emitrevisions( | |
2730 | self, |
|
2730 | self, | |
2731 | nodes, |
|
2731 | nodes, | |
2732 | nodesorder=None, |
|
2732 | nodesorder=None, | |
2733 | revisiondata=False, |
|
2733 | revisiondata=False, | |
2734 | assumehaveparentrevisions=False, |
|
2734 | assumehaveparentrevisions=False, | |
2735 | deltamode=repository.CG_DELTAMODE_STD, |
|
2735 | deltamode=repository.CG_DELTAMODE_STD, | |
|
2736 | sidedata_helpers=None, | |||
2736 | ): |
|
2737 | ): | |
2737 | if nodesorder not in (b'nodes', b'storage', b'linear', None): |
|
2738 | if nodesorder not in (b'nodes', b'storage', b'linear', None): | |
2738 | raise error.ProgrammingError( |
|
2739 | raise error.ProgrammingError( | |
2739 | b'unhandled value for nodesorder: %s' % nodesorder |
|
2740 | b'unhandled value for nodesorder: %s' % nodesorder | |
2740 | ) |
|
2741 | ) | |
2741 |
|
2742 | |||
2742 | if nodesorder is None and not self._generaldelta: |
|
2743 | if nodesorder is None and not self._generaldelta: | |
2743 | nodesorder = b'storage' |
|
2744 | nodesorder = b'storage' | |
2744 |
|
2745 | |||
2745 | if ( |
|
2746 | if ( | |
2746 | not self._storedeltachains |
|
2747 | not self._storedeltachains | |
2747 | and deltamode != repository.CG_DELTAMODE_PREV |
|
2748 | and deltamode != repository.CG_DELTAMODE_PREV | |
2748 | ): |
|
2749 | ): | |
2749 | deltamode = repository.CG_DELTAMODE_FULL |
|
2750 | deltamode = repository.CG_DELTAMODE_FULL | |
2750 |
|
2751 | |||
2751 | return storageutil.emitrevisions( |
|
2752 | return storageutil.emitrevisions( | |
2752 | self, |
|
2753 | self, | |
2753 | nodes, |
|
2754 | nodes, | |
2754 | nodesorder, |
|
2755 | nodesorder, | |
2755 | revlogrevisiondelta, |
|
2756 | revlogrevisiondelta, | |
2756 | deltaparentfn=self.deltaparent, |
|
2757 | deltaparentfn=self.deltaparent, | |
2757 | candeltafn=self.candelta, |
|
2758 | candeltafn=self.candelta, | |
2758 | rawsizefn=self.rawsize, |
|
2759 | rawsizefn=self.rawsize, | |
2759 | revdifffn=self.revdiff, |
|
2760 | revdifffn=self.revdiff, | |
2760 | flagsfn=self.flags, |
|
2761 | flagsfn=self.flags, | |
2761 | deltamode=deltamode, |
|
2762 | deltamode=deltamode, | |
2762 | revisiondata=revisiondata, |
|
2763 | revisiondata=revisiondata, | |
2763 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
2764 | assumehaveparentrevisions=assumehaveparentrevisions, | |
|
2765 | sidedata_helpers=sidedata_helpers, | |||
2764 | ) |
|
2766 | ) | |
2765 |
|
2767 | |||
2766 | DELTAREUSEALWAYS = b'always' |
|
2768 | DELTAREUSEALWAYS = b'always' | |
2767 | DELTAREUSESAMEREVS = b'samerevs' |
|
2769 | DELTAREUSESAMEREVS = b'samerevs' | |
2768 | DELTAREUSENEVER = b'never' |
|
2770 | DELTAREUSENEVER = b'never' | |
2769 |
|
2771 | |||
2770 | DELTAREUSEFULLADD = b'fulladd' |
|
2772 | DELTAREUSEFULLADD = b'fulladd' | |
2771 |
|
2773 | |||
2772 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} |
|
2774 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} | |
2773 |
|
2775 | |||
2774 | def clone( |
|
2776 | def clone( | |
2775 | self, |
|
2777 | self, | |
2776 | tr, |
|
2778 | tr, | |
2777 | destrevlog, |
|
2779 | destrevlog, | |
2778 | addrevisioncb=None, |
|
2780 | addrevisioncb=None, | |
2779 | deltareuse=DELTAREUSESAMEREVS, |
|
2781 | deltareuse=DELTAREUSESAMEREVS, | |
2780 | forcedeltabothparents=None, |
|
2782 | forcedeltabothparents=None, | |
2781 | sidedatacompanion=None, |
|
2783 | sidedatacompanion=None, | |
2782 | ): |
|
2784 | ): | |
2783 | """Copy this revlog to another, possibly with format changes. |
|
2785 | """Copy this revlog to another, possibly with format changes. | |
2784 |
|
2786 | |||
2785 | The destination revlog will contain the same revisions and nodes. |
|
2787 | The destination revlog will contain the same revisions and nodes. | |
2786 | However, it may not be bit-for-bit identical due to e.g. delta encoding |
|
2788 | However, it may not be bit-for-bit identical due to e.g. delta encoding | |
2787 | differences. |
|
2789 | differences. | |
2788 |
|
2790 | |||
2789 | The ``deltareuse`` argument control how deltas from the existing revlog |
|
2791 | The ``deltareuse`` argument control how deltas from the existing revlog | |
2790 | are preserved in the destination revlog. The argument can have the |
|
2792 | are preserved in the destination revlog. The argument can have the | |
2791 | following values: |
|
2793 | following values: | |
2792 |
|
2794 | |||
2793 | DELTAREUSEALWAYS |
|
2795 | DELTAREUSEALWAYS | |
2794 | Deltas will always be reused (if possible), even if the destination |
|
2796 | Deltas will always be reused (if possible), even if the destination | |
2795 | revlog would not select the same revisions for the delta. This is the |
|
2797 | revlog would not select the same revisions for the delta. This is the | |
2796 | fastest mode of operation. |
|
2798 | fastest mode of operation. | |
2797 | DELTAREUSESAMEREVS |
|
2799 | DELTAREUSESAMEREVS | |
2798 | Deltas will be reused if the destination revlog would pick the same |
|
2800 | Deltas will be reused if the destination revlog would pick the same | |
2799 | revisions for the delta. This mode strikes a balance between speed |
|
2801 | revisions for the delta. This mode strikes a balance between speed | |
2800 | and optimization. |
|
2802 | and optimization. | |
2801 | DELTAREUSENEVER |
|
2803 | DELTAREUSENEVER | |
2802 | Deltas will never be reused. This is the slowest mode of execution. |
|
2804 | Deltas will never be reused. This is the slowest mode of execution. | |
2803 | This mode can be used to recompute deltas (e.g. if the diff/delta |
|
2805 | This mode can be used to recompute deltas (e.g. if the diff/delta | |
2804 | algorithm changes). |
|
2806 | algorithm changes). | |
2805 | DELTAREUSEFULLADD |
|
2807 | DELTAREUSEFULLADD | |
2806 | Revision will be re-added as if their were new content. This is |
|
2808 | Revision will be re-added as if their were new content. This is | |
2807 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. |
|
2809 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. | |
2808 | eg: large file detection and handling. |
|
2810 | eg: large file detection and handling. | |
2809 |
|
2811 | |||
2810 | Delta computation can be slow, so the choice of delta reuse policy can |
|
2812 | Delta computation can be slow, so the choice of delta reuse policy can | |
2811 | significantly affect run time. |
|
2813 | significantly affect run time. | |
2812 |
|
2814 | |||
2813 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between |
|
2815 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between | |
2814 | two extremes. Deltas will be reused if they are appropriate. But if the |
|
2816 | two extremes. Deltas will be reused if they are appropriate. But if the | |
2815 | delta could choose a better revision, it will do so. This means if you |
|
2817 | delta could choose a better revision, it will do so. This means if you | |
2816 | are converting a non-generaldelta revlog to a generaldelta revlog, |
|
2818 | are converting a non-generaldelta revlog to a generaldelta revlog, | |
2817 | deltas will be recomputed if the delta's parent isn't a parent of the |
|
2819 | deltas will be recomputed if the delta's parent isn't a parent of the | |
2818 | revision. |
|
2820 | revision. | |
2819 |
|
2821 | |||
2820 | In addition to the delta policy, the ``forcedeltabothparents`` |
|
2822 | In addition to the delta policy, the ``forcedeltabothparents`` | |
2821 | argument controls whether to force compute deltas against both parents |
|
2823 | argument controls whether to force compute deltas against both parents | |
2822 | for merges. By default, the current default is used. |
|
2824 | for merges. By default, the current default is used. | |
2823 |
|
2825 | |||
2824 | If not None, the `sidedatacompanion` is callable that accept two |
|
2826 | If not None, the `sidedatacompanion` is callable that accept two | |
2825 | arguments: |
|
2827 | arguments: | |
2826 |
|
2828 | |||
2827 | (srcrevlog, rev) |
|
2829 | (srcrevlog, rev) | |
2828 |
|
2830 | |||
2829 | and return a quintet that control changes to sidedata content from the |
|
2831 | and return a quintet that control changes to sidedata content from the | |
2830 | old revision to the new clone result: |
|
2832 | old revision to the new clone result: | |
2831 |
|
2833 | |||
2832 | (dropall, filterout, update, new_flags, dropped_flags) |
|
2834 | (dropall, filterout, update, new_flags, dropped_flags) | |
2833 |
|
2835 | |||
2834 | * if `dropall` is True, all sidedata should be dropped |
|
2836 | * if `dropall` is True, all sidedata should be dropped | |
2835 | * `filterout` is a set of sidedata keys that should be dropped |
|
2837 | * `filterout` is a set of sidedata keys that should be dropped | |
2836 | * `update` is a mapping of additionnal/new key -> value |
|
2838 | * `update` is a mapping of additionnal/new key -> value | |
2837 | * new_flags is a bitfields of new flags that the revision should get |
|
2839 | * new_flags is a bitfields of new flags that the revision should get | |
2838 | * dropped_flags is a bitfields of new flags that the revision shoudl not longer have |
|
2840 | * dropped_flags is a bitfields of new flags that the revision shoudl not longer have | |
2839 | """ |
|
2841 | """ | |
2840 | if deltareuse not in self.DELTAREUSEALL: |
|
2842 | if deltareuse not in self.DELTAREUSEALL: | |
2841 | raise ValueError( |
|
2843 | raise ValueError( | |
2842 | _(b'value for deltareuse invalid: %s') % deltareuse |
|
2844 | _(b'value for deltareuse invalid: %s') % deltareuse | |
2843 | ) |
|
2845 | ) | |
2844 |
|
2846 | |||
2845 | if len(destrevlog): |
|
2847 | if len(destrevlog): | |
2846 | raise ValueError(_(b'destination revlog is not empty')) |
|
2848 | raise ValueError(_(b'destination revlog is not empty')) | |
2847 |
|
2849 | |||
2848 | if getattr(self, 'filteredrevs', None): |
|
2850 | if getattr(self, 'filteredrevs', None): | |
2849 | raise ValueError(_(b'source revlog has filtered revisions')) |
|
2851 | raise ValueError(_(b'source revlog has filtered revisions')) | |
2850 | if getattr(destrevlog, 'filteredrevs', None): |
|
2852 | if getattr(destrevlog, 'filteredrevs', None): | |
2851 | raise ValueError(_(b'destination revlog has filtered revisions')) |
|
2853 | raise ValueError(_(b'destination revlog has filtered revisions')) | |
2852 |
|
2854 | |||
2853 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, |
|
2855 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, | |
2854 | # if possible. |
|
2856 | # if possible. | |
2855 | oldlazydelta = destrevlog._lazydelta |
|
2857 | oldlazydelta = destrevlog._lazydelta | |
2856 | oldlazydeltabase = destrevlog._lazydeltabase |
|
2858 | oldlazydeltabase = destrevlog._lazydeltabase | |
2857 | oldamd = destrevlog._deltabothparents |
|
2859 | oldamd = destrevlog._deltabothparents | |
2858 |
|
2860 | |||
2859 | try: |
|
2861 | try: | |
2860 | if deltareuse == self.DELTAREUSEALWAYS: |
|
2862 | if deltareuse == self.DELTAREUSEALWAYS: | |
2861 | destrevlog._lazydeltabase = True |
|
2863 | destrevlog._lazydeltabase = True | |
2862 | destrevlog._lazydelta = True |
|
2864 | destrevlog._lazydelta = True | |
2863 | elif deltareuse == self.DELTAREUSESAMEREVS: |
|
2865 | elif deltareuse == self.DELTAREUSESAMEREVS: | |
2864 | destrevlog._lazydeltabase = False |
|
2866 | destrevlog._lazydeltabase = False | |
2865 | destrevlog._lazydelta = True |
|
2867 | destrevlog._lazydelta = True | |
2866 | elif deltareuse == self.DELTAREUSENEVER: |
|
2868 | elif deltareuse == self.DELTAREUSENEVER: | |
2867 | destrevlog._lazydeltabase = False |
|
2869 | destrevlog._lazydeltabase = False | |
2868 | destrevlog._lazydelta = False |
|
2870 | destrevlog._lazydelta = False | |
2869 |
|
2871 | |||
2870 | destrevlog._deltabothparents = forcedeltabothparents or oldamd |
|
2872 | destrevlog._deltabothparents = forcedeltabothparents or oldamd | |
2871 |
|
2873 | |||
2872 | self._clone( |
|
2874 | self._clone( | |
2873 | tr, |
|
2875 | tr, | |
2874 | destrevlog, |
|
2876 | destrevlog, | |
2875 | addrevisioncb, |
|
2877 | addrevisioncb, | |
2876 | deltareuse, |
|
2878 | deltareuse, | |
2877 | forcedeltabothparents, |
|
2879 | forcedeltabothparents, | |
2878 | sidedatacompanion, |
|
2880 | sidedatacompanion, | |
2879 | ) |
|
2881 | ) | |
2880 |
|
2882 | |||
2881 | finally: |
|
2883 | finally: | |
2882 | destrevlog._lazydelta = oldlazydelta |
|
2884 | destrevlog._lazydelta = oldlazydelta | |
2883 | destrevlog._lazydeltabase = oldlazydeltabase |
|
2885 | destrevlog._lazydeltabase = oldlazydeltabase | |
2884 | destrevlog._deltabothparents = oldamd |
|
2886 | destrevlog._deltabothparents = oldamd | |
2885 |
|
2887 | |||
2886 | def _clone( |
|
2888 | def _clone( | |
2887 | self, |
|
2889 | self, | |
2888 | tr, |
|
2890 | tr, | |
2889 | destrevlog, |
|
2891 | destrevlog, | |
2890 | addrevisioncb, |
|
2892 | addrevisioncb, | |
2891 | deltareuse, |
|
2893 | deltareuse, | |
2892 | forcedeltabothparents, |
|
2894 | forcedeltabothparents, | |
2893 | sidedatacompanion, |
|
2895 | sidedatacompanion, | |
2894 | ): |
|
2896 | ): | |
2895 | """perform the core duty of `revlog.clone` after parameter processing""" |
|
2897 | """perform the core duty of `revlog.clone` after parameter processing""" | |
2896 | deltacomputer = deltautil.deltacomputer(destrevlog) |
|
2898 | deltacomputer = deltautil.deltacomputer(destrevlog) | |
2897 | index = self.index |
|
2899 | index = self.index | |
2898 | for rev in self: |
|
2900 | for rev in self: | |
2899 | entry = index[rev] |
|
2901 | entry = index[rev] | |
2900 |
|
2902 | |||
2901 | # Some classes override linkrev to take filtered revs into |
|
2903 | # Some classes override linkrev to take filtered revs into | |
2902 | # account. Use raw entry from index. |
|
2904 | # account. Use raw entry from index. | |
2903 | flags = entry[0] & 0xFFFF |
|
2905 | flags = entry[0] & 0xFFFF | |
2904 | linkrev = entry[4] |
|
2906 | linkrev = entry[4] | |
2905 | p1 = index[entry[5]][7] |
|
2907 | p1 = index[entry[5]][7] | |
2906 | p2 = index[entry[6]][7] |
|
2908 | p2 = index[entry[6]][7] | |
2907 | node = entry[7] |
|
2909 | node = entry[7] | |
2908 |
|
2910 | |||
2909 | sidedataactions = (False, [], {}, 0, 0) |
|
2911 | sidedataactions = (False, [], {}, 0, 0) | |
2910 | if sidedatacompanion is not None: |
|
2912 | if sidedatacompanion is not None: | |
2911 | sidedataactions = sidedatacompanion(self, rev) |
|
2913 | sidedataactions = sidedatacompanion(self, rev) | |
2912 |
|
2914 | |||
2913 | # (Possibly) reuse the delta from the revlog if allowed and |
|
2915 | # (Possibly) reuse the delta from the revlog if allowed and | |
2914 | # the revlog chunk is a delta. |
|
2916 | # the revlog chunk is a delta. | |
2915 | cachedelta = None |
|
2917 | cachedelta = None | |
2916 | rawtext = None |
|
2918 | rawtext = None | |
2917 | if any(sidedataactions) or deltareuse == self.DELTAREUSEFULLADD: |
|
2919 | if any(sidedataactions) or deltareuse == self.DELTAREUSEFULLADD: | |
2918 | dropall = sidedataactions[0] |
|
2920 | dropall = sidedataactions[0] | |
2919 | filterout = sidedataactions[1] |
|
2921 | filterout = sidedataactions[1] | |
2920 | update = sidedataactions[2] |
|
2922 | update = sidedataactions[2] | |
2921 | new_flags = sidedataactions[3] |
|
2923 | new_flags = sidedataactions[3] | |
2922 | dropped_flags = sidedataactions[4] |
|
2924 | dropped_flags = sidedataactions[4] | |
2923 | text, sidedata = self._revisiondata(rev) |
|
2925 | text, sidedata = self._revisiondata(rev) | |
2924 | if dropall: |
|
2926 | if dropall: | |
2925 | sidedata = {} |
|
2927 | sidedata = {} | |
2926 | for key in filterout: |
|
2928 | for key in filterout: | |
2927 | sidedata.pop(key, None) |
|
2929 | sidedata.pop(key, None) | |
2928 | sidedata.update(update) |
|
2930 | sidedata.update(update) | |
2929 | if not sidedata: |
|
2931 | if not sidedata: | |
2930 | sidedata = None |
|
2932 | sidedata = None | |
2931 |
|
2933 | |||
2932 | flags |= new_flags |
|
2934 | flags |= new_flags | |
2933 | flags &= ~dropped_flags |
|
2935 | flags &= ~dropped_flags | |
2934 |
|
2936 | |||
2935 | destrevlog.addrevision( |
|
2937 | destrevlog.addrevision( | |
2936 | text, |
|
2938 | text, | |
2937 | tr, |
|
2939 | tr, | |
2938 | linkrev, |
|
2940 | linkrev, | |
2939 | p1, |
|
2941 | p1, | |
2940 | p2, |
|
2942 | p2, | |
2941 | cachedelta=cachedelta, |
|
2943 | cachedelta=cachedelta, | |
2942 | node=node, |
|
2944 | node=node, | |
2943 | flags=flags, |
|
2945 | flags=flags, | |
2944 | deltacomputer=deltacomputer, |
|
2946 | deltacomputer=deltacomputer, | |
2945 | sidedata=sidedata, |
|
2947 | sidedata=sidedata, | |
2946 | ) |
|
2948 | ) | |
2947 | else: |
|
2949 | else: | |
2948 | if destrevlog._lazydelta: |
|
2950 | if destrevlog._lazydelta: | |
2949 | dp = self.deltaparent(rev) |
|
2951 | dp = self.deltaparent(rev) | |
2950 | if dp != nullrev: |
|
2952 | if dp != nullrev: | |
2951 | cachedelta = (dp, bytes(self._chunk(rev))) |
|
2953 | cachedelta = (dp, bytes(self._chunk(rev))) | |
2952 |
|
2954 | |||
2953 | if not cachedelta: |
|
2955 | if not cachedelta: | |
2954 | rawtext = self.rawdata(rev) |
|
2956 | rawtext = self.rawdata(rev) | |
2955 |
|
2957 | |||
2956 | ifh = destrevlog.opener( |
|
2958 | ifh = destrevlog.opener( | |
2957 | destrevlog.indexfile, b'a+', checkambig=False |
|
2959 | destrevlog.indexfile, b'a+', checkambig=False | |
2958 | ) |
|
2960 | ) | |
2959 | dfh = None |
|
2961 | dfh = None | |
2960 | if not destrevlog._inline: |
|
2962 | if not destrevlog._inline: | |
2961 | dfh = destrevlog.opener(destrevlog.datafile, b'a+') |
|
2963 | dfh = destrevlog.opener(destrevlog.datafile, b'a+') | |
2962 | try: |
|
2964 | try: | |
2963 | destrevlog._addrevision( |
|
2965 | destrevlog._addrevision( | |
2964 | node, |
|
2966 | node, | |
2965 | rawtext, |
|
2967 | rawtext, | |
2966 | tr, |
|
2968 | tr, | |
2967 | linkrev, |
|
2969 | linkrev, | |
2968 | p1, |
|
2970 | p1, | |
2969 | p2, |
|
2971 | p2, | |
2970 | flags, |
|
2972 | flags, | |
2971 | cachedelta, |
|
2973 | cachedelta, | |
2972 | ifh, |
|
2974 | ifh, | |
2973 | dfh, |
|
2975 | dfh, | |
2974 | deltacomputer=deltacomputer, |
|
2976 | deltacomputer=deltacomputer, | |
2975 | ) |
|
2977 | ) | |
2976 | finally: |
|
2978 | finally: | |
2977 | if dfh: |
|
2979 | if dfh: | |
2978 | dfh.close() |
|
2980 | dfh.close() | |
2979 | ifh.close() |
|
2981 | ifh.close() | |
2980 |
|
2982 | |||
2981 | if addrevisioncb: |
|
2983 | if addrevisioncb: | |
2982 | addrevisioncb(self, rev, node) |
|
2984 | addrevisioncb(self, rev, node) | |
2983 |
|
2985 | |||
2984 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
2986 | def censorrevision(self, tr, censornode, tombstone=b''): | |
2985 | if (self.version & 0xFFFF) == REVLOGV0: |
|
2987 | if (self.version & 0xFFFF) == REVLOGV0: | |
2986 | raise error.RevlogError( |
|
2988 | raise error.RevlogError( | |
2987 | _(b'cannot censor with version %d revlogs') % self.version |
|
2989 | _(b'cannot censor with version %d revlogs') % self.version | |
2988 | ) |
|
2990 | ) | |
2989 |
|
2991 | |||
2990 | censorrev = self.rev(censornode) |
|
2992 | censorrev = self.rev(censornode) | |
2991 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') |
|
2993 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') | |
2992 |
|
2994 | |||
2993 | if len(tombstone) > self.rawsize(censorrev): |
|
2995 | if len(tombstone) > self.rawsize(censorrev): | |
2994 | raise error.Abort( |
|
2996 | raise error.Abort( | |
2995 | _(b'censor tombstone must be no longer than censored data') |
|
2997 | _(b'censor tombstone must be no longer than censored data') | |
2996 | ) |
|
2998 | ) | |
2997 |
|
2999 | |||
2998 | # Rewriting the revlog in place is hard. Our strategy for censoring is |
|
3000 | # Rewriting the revlog in place is hard. Our strategy for censoring is | |
2999 | # to create a new revlog, copy all revisions to it, then replace the |
|
3001 | # to create a new revlog, copy all revisions to it, then replace the | |
3000 | # revlogs on transaction close. |
|
3002 | # revlogs on transaction close. | |
3001 |
|
3003 | |||
3002 | newindexfile = self.indexfile + b'.tmpcensored' |
|
3004 | newindexfile = self.indexfile + b'.tmpcensored' | |
3003 | newdatafile = self.datafile + b'.tmpcensored' |
|
3005 | newdatafile = self.datafile + b'.tmpcensored' | |
3004 |
|
3006 | |||
3005 | # This is a bit dangerous. We could easily have a mismatch of state. |
|
3007 | # This is a bit dangerous. We could easily have a mismatch of state. | |
3006 | newrl = revlog(self.opener, newindexfile, newdatafile, censorable=True) |
|
3008 | newrl = revlog(self.opener, newindexfile, newdatafile, censorable=True) | |
3007 | newrl.version = self.version |
|
3009 | newrl.version = self.version | |
3008 | newrl._generaldelta = self._generaldelta |
|
3010 | newrl._generaldelta = self._generaldelta | |
3009 | newrl._io = self._io |
|
3011 | newrl._io = self._io | |
3010 |
|
3012 | |||
3011 | for rev in self.revs(): |
|
3013 | for rev in self.revs(): | |
3012 | node = self.node(rev) |
|
3014 | node = self.node(rev) | |
3013 | p1, p2 = self.parents(node) |
|
3015 | p1, p2 = self.parents(node) | |
3014 |
|
3016 | |||
3015 | if rev == censorrev: |
|
3017 | if rev == censorrev: | |
3016 | newrl.addrawrevision( |
|
3018 | newrl.addrawrevision( | |
3017 | tombstone, |
|
3019 | tombstone, | |
3018 | tr, |
|
3020 | tr, | |
3019 | self.linkrev(censorrev), |
|
3021 | self.linkrev(censorrev), | |
3020 | p1, |
|
3022 | p1, | |
3021 | p2, |
|
3023 | p2, | |
3022 | censornode, |
|
3024 | censornode, | |
3023 | REVIDX_ISCENSORED, |
|
3025 | REVIDX_ISCENSORED, | |
3024 | ) |
|
3026 | ) | |
3025 |
|
3027 | |||
3026 | if newrl.deltaparent(rev) != nullrev: |
|
3028 | if newrl.deltaparent(rev) != nullrev: | |
3027 | raise error.Abort( |
|
3029 | raise error.Abort( | |
3028 | _( |
|
3030 | _( | |
3029 | b'censored revision stored as delta; ' |
|
3031 | b'censored revision stored as delta; ' | |
3030 | b'cannot censor' |
|
3032 | b'cannot censor' | |
3031 | ), |
|
3033 | ), | |
3032 | hint=_( |
|
3034 | hint=_( | |
3033 | b'censoring of revlogs is not ' |
|
3035 | b'censoring of revlogs is not ' | |
3034 | b'fully implemented; please report ' |
|
3036 | b'fully implemented; please report ' | |
3035 | b'this bug' |
|
3037 | b'this bug' | |
3036 | ), |
|
3038 | ), | |
3037 | ) |
|
3039 | ) | |
3038 | continue |
|
3040 | continue | |
3039 |
|
3041 | |||
3040 | if self.iscensored(rev): |
|
3042 | if self.iscensored(rev): | |
3041 | if self.deltaparent(rev) != nullrev: |
|
3043 | if self.deltaparent(rev) != nullrev: | |
3042 | raise error.Abort( |
|
3044 | raise error.Abort( | |
3043 | _( |
|
3045 | _( | |
3044 | b'cannot censor due to censored ' |
|
3046 | b'cannot censor due to censored ' | |
3045 | b'revision having delta stored' |
|
3047 | b'revision having delta stored' | |
3046 | ) |
|
3048 | ) | |
3047 | ) |
|
3049 | ) | |
3048 | rawtext = self._chunk(rev) |
|
3050 | rawtext = self._chunk(rev) | |
3049 | else: |
|
3051 | else: | |
3050 | rawtext = self.rawdata(rev) |
|
3052 | rawtext = self.rawdata(rev) | |
3051 |
|
3053 | |||
3052 | newrl.addrawrevision( |
|
3054 | newrl.addrawrevision( | |
3053 | rawtext, tr, self.linkrev(rev), p1, p2, node, self.flags(rev) |
|
3055 | rawtext, tr, self.linkrev(rev), p1, p2, node, self.flags(rev) | |
3054 | ) |
|
3056 | ) | |
3055 |
|
3057 | |||
3056 | tr.addbackup(self.indexfile, location=b'store') |
|
3058 | tr.addbackup(self.indexfile, location=b'store') | |
3057 | if not self._inline: |
|
3059 | if not self._inline: | |
3058 | tr.addbackup(self.datafile, location=b'store') |
|
3060 | tr.addbackup(self.datafile, location=b'store') | |
3059 |
|
3061 | |||
3060 | self.opener.rename(newrl.indexfile, self.indexfile) |
|
3062 | self.opener.rename(newrl.indexfile, self.indexfile) | |
3061 | if not self._inline: |
|
3063 | if not self._inline: | |
3062 | self.opener.rename(newrl.datafile, self.datafile) |
|
3064 | self.opener.rename(newrl.datafile, self.datafile) | |
3063 |
|
3065 | |||
3064 | self.clearcaches() |
|
3066 | self.clearcaches() | |
3065 | self._loadindex() |
|
3067 | self._loadindex() | |
3066 |
|
3068 | |||
3067 | def verifyintegrity(self, state): |
|
3069 | def verifyintegrity(self, state): | |
3068 | """Verifies the integrity of the revlog. |
|
3070 | """Verifies the integrity of the revlog. | |
3069 |
|
3071 | |||
3070 | Yields ``revlogproblem`` instances describing problems that are |
|
3072 | Yields ``revlogproblem`` instances describing problems that are | |
3071 | found. |
|
3073 | found. | |
3072 | """ |
|
3074 | """ | |
3073 | dd, di = self.checksize() |
|
3075 | dd, di = self.checksize() | |
3074 | if dd: |
|
3076 | if dd: | |
3075 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) |
|
3077 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) | |
3076 | if di: |
|
3078 | if di: | |
3077 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) |
|
3079 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) | |
3078 |
|
3080 | |||
3079 | version = self.version & 0xFFFF |
|
3081 | version = self.version & 0xFFFF | |
3080 |
|
3082 | |||
3081 | # The verifier tells us what version revlog we should be. |
|
3083 | # The verifier tells us what version revlog we should be. | |
3082 | if version != state[b'expectedversion']: |
|
3084 | if version != state[b'expectedversion']: | |
3083 | yield revlogproblem( |
|
3085 | yield revlogproblem( | |
3084 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") |
|
3086 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") | |
3085 | % (self.indexfile, version, state[b'expectedversion']) |
|
3087 | % (self.indexfile, version, state[b'expectedversion']) | |
3086 | ) |
|
3088 | ) | |
3087 |
|
3089 | |||
3088 | state[b'skipread'] = set() |
|
3090 | state[b'skipread'] = set() | |
3089 | state[b'safe_renamed'] = set() |
|
3091 | state[b'safe_renamed'] = set() | |
3090 |
|
3092 | |||
3091 | for rev in self: |
|
3093 | for rev in self: | |
3092 | node = self.node(rev) |
|
3094 | node = self.node(rev) | |
3093 |
|
3095 | |||
3094 | # Verify contents. 4 cases to care about: |
|
3096 | # Verify contents. 4 cases to care about: | |
3095 | # |
|
3097 | # | |
3096 | # common: the most common case |
|
3098 | # common: the most common case | |
3097 | # rename: with a rename |
|
3099 | # rename: with a rename | |
3098 | # meta: file content starts with b'\1\n', the metadata |
|
3100 | # meta: file content starts with b'\1\n', the metadata | |
3099 | # header defined in filelog.py, but without a rename |
|
3101 | # header defined in filelog.py, but without a rename | |
3100 | # ext: content stored externally |
|
3102 | # ext: content stored externally | |
3101 | # |
|
3103 | # | |
3102 | # More formally, their differences are shown below: |
|
3104 | # More formally, their differences are shown below: | |
3103 | # |
|
3105 | # | |
3104 | # | common | rename | meta | ext |
|
3106 | # | common | rename | meta | ext | |
3105 | # ------------------------------------------------------- |
|
3107 | # ------------------------------------------------------- | |
3106 | # flags() | 0 | 0 | 0 | not 0 |
|
3108 | # flags() | 0 | 0 | 0 | not 0 | |
3107 | # renamed() | False | True | False | ? |
|
3109 | # renamed() | False | True | False | ? | |
3108 | # rawtext[0:2]=='\1\n'| False | True | True | ? |
|
3110 | # rawtext[0:2]=='\1\n'| False | True | True | ? | |
3109 | # |
|
3111 | # | |
3110 | # "rawtext" means the raw text stored in revlog data, which |
|
3112 | # "rawtext" means the raw text stored in revlog data, which | |
3111 | # could be retrieved by "rawdata(rev)". "text" |
|
3113 | # could be retrieved by "rawdata(rev)". "text" | |
3112 | # mentioned below is "revision(rev)". |
|
3114 | # mentioned below is "revision(rev)". | |
3113 | # |
|
3115 | # | |
3114 | # There are 3 different lengths stored physically: |
|
3116 | # There are 3 different lengths stored physically: | |
3115 | # 1. L1: rawsize, stored in revlog index |
|
3117 | # 1. L1: rawsize, stored in revlog index | |
3116 | # 2. L2: len(rawtext), stored in revlog data |
|
3118 | # 2. L2: len(rawtext), stored in revlog data | |
3117 | # 3. L3: len(text), stored in revlog data if flags==0, or |
|
3119 | # 3. L3: len(text), stored in revlog data if flags==0, or | |
3118 | # possibly somewhere else if flags!=0 |
|
3120 | # possibly somewhere else if flags!=0 | |
3119 | # |
|
3121 | # | |
3120 | # L1 should be equal to L2. L3 could be different from them. |
|
3122 | # L1 should be equal to L2. L3 could be different from them. | |
3121 | # "text" may or may not affect commit hash depending on flag |
|
3123 | # "text" may or may not affect commit hash depending on flag | |
3122 | # processors (see flagutil.addflagprocessor). |
|
3124 | # processors (see flagutil.addflagprocessor). | |
3123 | # |
|
3125 | # | |
3124 | # | common | rename | meta | ext |
|
3126 | # | common | rename | meta | ext | |
3125 | # ------------------------------------------------- |
|
3127 | # ------------------------------------------------- | |
3126 | # rawsize() | L1 | L1 | L1 | L1 |
|
3128 | # rawsize() | L1 | L1 | L1 | L1 | |
3127 | # size() | L1 | L2-LM | L1(*) | L1 (?) |
|
3129 | # size() | L1 | L2-LM | L1(*) | L1 (?) | |
3128 | # len(rawtext) | L2 | L2 | L2 | L2 |
|
3130 | # len(rawtext) | L2 | L2 | L2 | L2 | |
3129 | # len(text) | L2 | L2 | L2 | L3 |
|
3131 | # len(text) | L2 | L2 | L2 | L3 | |
3130 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) |
|
3132 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) | |
3131 | # |
|
3133 | # | |
3132 | # LM: length of metadata, depending on rawtext |
|
3134 | # LM: length of metadata, depending on rawtext | |
3133 | # (*): not ideal, see comment in filelog.size |
|
3135 | # (*): not ideal, see comment in filelog.size | |
3134 | # (?): could be "- len(meta)" if the resolved content has |
|
3136 | # (?): could be "- len(meta)" if the resolved content has | |
3135 | # rename metadata |
|
3137 | # rename metadata | |
3136 | # |
|
3138 | # | |
3137 | # Checks needed to be done: |
|
3139 | # Checks needed to be done: | |
3138 | # 1. length check: L1 == L2, in all cases. |
|
3140 | # 1. length check: L1 == L2, in all cases. | |
3139 | # 2. hash check: depending on flag processor, we may need to |
|
3141 | # 2. hash check: depending on flag processor, we may need to | |
3140 | # use either "text" (external), or "rawtext" (in revlog). |
|
3142 | # use either "text" (external), or "rawtext" (in revlog). | |
3141 |
|
3143 | |||
3142 | try: |
|
3144 | try: | |
3143 | skipflags = state.get(b'skipflags', 0) |
|
3145 | skipflags = state.get(b'skipflags', 0) | |
3144 | if skipflags: |
|
3146 | if skipflags: | |
3145 | skipflags &= self.flags(rev) |
|
3147 | skipflags &= self.flags(rev) | |
3146 |
|
3148 | |||
3147 | _verify_revision(self, skipflags, state, node) |
|
3149 | _verify_revision(self, skipflags, state, node) | |
3148 |
|
3150 | |||
3149 | l1 = self.rawsize(rev) |
|
3151 | l1 = self.rawsize(rev) | |
3150 | l2 = len(self.rawdata(node)) |
|
3152 | l2 = len(self.rawdata(node)) | |
3151 |
|
3153 | |||
3152 | if l1 != l2: |
|
3154 | if l1 != l2: | |
3153 | yield revlogproblem( |
|
3155 | yield revlogproblem( | |
3154 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), |
|
3156 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), | |
3155 | node=node, |
|
3157 | node=node, | |
3156 | ) |
|
3158 | ) | |
3157 |
|
3159 | |||
3158 | except error.CensoredNodeError: |
|
3160 | except error.CensoredNodeError: | |
3159 | if state[b'erroroncensored']: |
|
3161 | if state[b'erroroncensored']: | |
3160 | yield revlogproblem( |
|
3162 | yield revlogproblem( | |
3161 | error=_(b'censored file data'), node=node |
|
3163 | error=_(b'censored file data'), node=node | |
3162 | ) |
|
3164 | ) | |
3163 | state[b'skipread'].add(node) |
|
3165 | state[b'skipread'].add(node) | |
3164 | except Exception as e: |
|
3166 | except Exception as e: | |
3165 | yield revlogproblem( |
|
3167 | yield revlogproblem( | |
3166 | error=_(b'unpacking %s: %s') |
|
3168 | error=_(b'unpacking %s: %s') | |
3167 | % (short(node), stringutil.forcebytestr(e)), |
|
3169 | % (short(node), stringutil.forcebytestr(e)), | |
3168 | node=node, |
|
3170 | node=node, | |
3169 | ) |
|
3171 | ) | |
3170 | state[b'skipread'].add(node) |
|
3172 | state[b'skipread'].add(node) | |
3171 |
|
3173 | |||
3172 | def storageinfo( |
|
3174 | def storageinfo( | |
3173 | self, |
|
3175 | self, | |
3174 | exclusivefiles=False, |
|
3176 | exclusivefiles=False, | |
3175 | sharedfiles=False, |
|
3177 | sharedfiles=False, | |
3176 | revisionscount=False, |
|
3178 | revisionscount=False, | |
3177 | trackedsize=False, |
|
3179 | trackedsize=False, | |
3178 | storedsize=False, |
|
3180 | storedsize=False, | |
3179 | ): |
|
3181 | ): | |
3180 | d = {} |
|
3182 | d = {} | |
3181 |
|
3183 | |||
3182 | if exclusivefiles: |
|
3184 | if exclusivefiles: | |
3183 | d[b'exclusivefiles'] = [(self.opener, self.indexfile)] |
|
3185 | d[b'exclusivefiles'] = [(self.opener, self.indexfile)] | |
3184 | if not self._inline: |
|
3186 | if not self._inline: | |
3185 | d[b'exclusivefiles'].append((self.opener, self.datafile)) |
|
3187 | d[b'exclusivefiles'].append((self.opener, self.datafile)) | |
3186 |
|
3188 | |||
3187 | if sharedfiles: |
|
3189 | if sharedfiles: | |
3188 | d[b'sharedfiles'] = [] |
|
3190 | d[b'sharedfiles'] = [] | |
3189 |
|
3191 | |||
3190 | if revisionscount: |
|
3192 | if revisionscount: | |
3191 | d[b'revisionscount'] = len(self) |
|
3193 | d[b'revisionscount'] = len(self) | |
3192 |
|
3194 | |||
3193 | if trackedsize: |
|
3195 | if trackedsize: | |
3194 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) |
|
3196 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) | |
3195 |
|
3197 | |||
3196 | if storedsize: |
|
3198 | if storedsize: | |
3197 | d[b'storedsize'] = sum( |
|
3199 | d[b'storedsize'] = sum( | |
3198 | self.opener.stat(path).st_size for path in self.files() |
|
3200 | self.opener.stat(path).st_size for path in self.files() | |
3199 | ) |
|
3201 | ) | |
3200 |
|
3202 | |||
3201 | return d |
|
3203 | return d |
@@ -1,513 +1,560 b'' | |||||
1 | # storageutil.py - Storage functionality agnostic of backend implementation. |
|
1 | # storageutil.py - Storage functionality agnostic of backend implementation. | |
2 | # |
|
2 | # | |
3 | # Copyright 2018 Gregory Szorc <gregory.szorc@gmail.com> |
|
3 | # Copyright 2018 Gregory Szorc <gregory.szorc@gmail.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import re |
|
10 | import re | |
11 | import struct |
|
11 | import struct | |
12 |
|
12 | |||
13 | from ..i18n import _ |
|
13 | from ..i18n import _ | |
14 | from ..node import ( |
|
14 | from ..node import ( | |
15 | bin, |
|
15 | bin, | |
16 | nullid, |
|
16 | nullid, | |
17 | nullrev, |
|
17 | nullrev, | |
18 | ) |
|
18 | ) | |
19 | from .. import ( |
|
19 | from .. import ( | |
20 | dagop, |
|
20 | dagop, | |
21 | error, |
|
21 | error, | |
22 | mdiff, |
|
22 | mdiff, | |
23 | pycompat, |
|
23 | pycompat, | |
24 | ) |
|
24 | ) | |
25 | from ..interfaces import repository |
|
25 | from ..interfaces import repository | |
|
26 | from ..revlogutils import sidedata as sidedatamod | |||
26 | from ..utils import hashutil |
|
27 | from ..utils import hashutil | |
27 |
|
28 | |||
28 | _nullhash = hashutil.sha1(nullid) |
|
29 | _nullhash = hashutil.sha1(nullid) | |
29 |
|
30 | |||
30 |
|
31 | |||
31 | def hashrevisionsha1(text, p1, p2): |
|
32 | def hashrevisionsha1(text, p1, p2): | |
32 | """Compute the SHA-1 for revision data and its parents. |
|
33 | """Compute the SHA-1 for revision data and its parents. | |
33 |
|
34 | |||
34 | This hash combines both the current file contents and its history |
|
35 | This hash combines both the current file contents and its history | |
35 | in a manner that makes it easy to distinguish nodes with the same |
|
36 | in a manner that makes it easy to distinguish nodes with the same | |
36 | content in the revision graph. |
|
37 | content in the revision graph. | |
37 | """ |
|
38 | """ | |
38 | # As of now, if one of the parent node is null, p2 is null |
|
39 | # As of now, if one of the parent node is null, p2 is null | |
39 | if p2 == nullid: |
|
40 | if p2 == nullid: | |
40 | # deep copy of a hash is faster than creating one |
|
41 | # deep copy of a hash is faster than creating one | |
41 | s = _nullhash.copy() |
|
42 | s = _nullhash.copy() | |
42 | s.update(p1) |
|
43 | s.update(p1) | |
43 | else: |
|
44 | else: | |
44 | # none of the parent nodes are nullid |
|
45 | # none of the parent nodes are nullid | |
45 | if p1 < p2: |
|
46 | if p1 < p2: | |
46 | a = p1 |
|
47 | a = p1 | |
47 | b = p2 |
|
48 | b = p2 | |
48 | else: |
|
49 | else: | |
49 | a = p2 |
|
50 | a = p2 | |
50 | b = p1 |
|
51 | b = p1 | |
51 | s = hashutil.sha1(a) |
|
52 | s = hashutil.sha1(a) | |
52 | s.update(b) |
|
53 | s.update(b) | |
53 | s.update(text) |
|
54 | s.update(text) | |
54 | return s.digest() |
|
55 | return s.digest() | |
55 |
|
56 | |||
56 |
|
57 | |||
57 | METADATA_RE = re.compile(b'\x01\n') |
|
58 | METADATA_RE = re.compile(b'\x01\n') | |
58 |
|
59 | |||
59 |
|
60 | |||
60 | def parsemeta(text): |
|
61 | def parsemeta(text): | |
61 | """Parse metadata header from revision data. |
|
62 | """Parse metadata header from revision data. | |
62 |
|
63 | |||
63 | Returns a 2-tuple of (metadata, offset), where both can be None if there |
|
64 | Returns a 2-tuple of (metadata, offset), where both can be None if there | |
64 | is no metadata. |
|
65 | is no metadata. | |
65 | """ |
|
66 | """ | |
66 | # text can be buffer, so we can't use .startswith or .index |
|
67 | # text can be buffer, so we can't use .startswith or .index | |
67 | if text[:2] != b'\x01\n': |
|
68 | if text[:2] != b'\x01\n': | |
68 | return None, None |
|
69 | return None, None | |
69 | s = METADATA_RE.search(text, 2).start() |
|
70 | s = METADATA_RE.search(text, 2).start() | |
70 | mtext = text[2:s] |
|
71 | mtext = text[2:s] | |
71 | meta = {} |
|
72 | meta = {} | |
72 | for l in mtext.splitlines(): |
|
73 | for l in mtext.splitlines(): | |
73 | k, v = l.split(b': ', 1) |
|
74 | k, v = l.split(b': ', 1) | |
74 | meta[k] = v |
|
75 | meta[k] = v | |
75 | return meta, s + 2 |
|
76 | return meta, s + 2 | |
76 |
|
77 | |||
77 |
|
78 | |||
78 | def packmeta(meta, text): |
|
79 | def packmeta(meta, text): | |
79 | """Add metadata to fulltext to produce revision text.""" |
|
80 | """Add metadata to fulltext to produce revision text.""" | |
80 | keys = sorted(meta) |
|
81 | keys = sorted(meta) | |
81 | metatext = b''.join(b'%s: %s\n' % (k, meta[k]) for k in keys) |
|
82 | metatext = b''.join(b'%s: %s\n' % (k, meta[k]) for k in keys) | |
82 | return b'\x01\n%s\x01\n%s' % (metatext, text) |
|
83 | return b'\x01\n%s\x01\n%s' % (metatext, text) | |
83 |
|
84 | |||
84 |
|
85 | |||
85 | def iscensoredtext(text): |
|
86 | def iscensoredtext(text): | |
86 | meta = parsemeta(text)[0] |
|
87 | meta = parsemeta(text)[0] | |
87 | return meta and b'censored' in meta |
|
88 | return meta and b'censored' in meta | |
88 |
|
89 | |||
89 |
|
90 | |||
90 | def filtermetadata(text): |
|
91 | def filtermetadata(text): | |
91 | """Extract just the revision data from source text. |
|
92 | """Extract just the revision data from source text. | |
92 |
|
93 | |||
93 | Returns ``text`` unless it has a metadata header, in which case we return |
|
94 | Returns ``text`` unless it has a metadata header, in which case we return | |
94 | a new buffer without hte metadata. |
|
95 | a new buffer without hte metadata. | |
95 | """ |
|
96 | """ | |
96 | if not text.startswith(b'\x01\n'): |
|
97 | if not text.startswith(b'\x01\n'): | |
97 | return text |
|
98 | return text | |
98 |
|
99 | |||
99 | offset = text.index(b'\x01\n', 2) |
|
100 | offset = text.index(b'\x01\n', 2) | |
100 | return text[offset + 2 :] |
|
101 | return text[offset + 2 :] | |
101 |
|
102 | |||
102 |
|
103 | |||
103 | def filerevisioncopied(store, node): |
|
104 | def filerevisioncopied(store, node): | |
104 | """Resolve file revision copy metadata. |
|
105 | """Resolve file revision copy metadata. | |
105 |
|
106 | |||
106 | Returns ``False`` if the file has no copy metadata. Otherwise a |
|
107 | Returns ``False`` if the file has no copy metadata. Otherwise a | |
107 | 2-tuple of the source filename and node. |
|
108 | 2-tuple of the source filename and node. | |
108 | """ |
|
109 | """ | |
109 | if store.parents(node)[0] != nullid: |
|
110 | if store.parents(node)[0] != nullid: | |
110 | return False |
|
111 | return False | |
111 |
|
112 | |||
112 | meta = parsemeta(store.revision(node))[0] |
|
113 | meta = parsemeta(store.revision(node))[0] | |
113 |
|
114 | |||
114 | # copy and copyrev occur in pairs. In rare cases due to old bugs, |
|
115 | # copy and copyrev occur in pairs. In rare cases due to old bugs, | |
115 | # one can occur without the other. So ensure both are present to flag |
|
116 | # one can occur without the other. So ensure both are present to flag | |
116 | # as a copy. |
|
117 | # as a copy. | |
117 | if meta and b'copy' in meta and b'copyrev' in meta: |
|
118 | if meta and b'copy' in meta and b'copyrev' in meta: | |
118 | return meta[b'copy'], bin(meta[b'copyrev']) |
|
119 | return meta[b'copy'], bin(meta[b'copyrev']) | |
119 |
|
120 | |||
120 | return False |
|
121 | return False | |
121 |
|
122 | |||
122 |
|
123 | |||
123 | def filedataequivalent(store, node, filedata): |
|
124 | def filedataequivalent(store, node, filedata): | |
124 | """Determines whether file data is equivalent to a stored node. |
|
125 | """Determines whether file data is equivalent to a stored node. | |
125 |
|
126 | |||
126 | Returns True if the passed file data would hash to the same value |
|
127 | Returns True if the passed file data would hash to the same value | |
127 | as a stored revision and False otherwise. |
|
128 | as a stored revision and False otherwise. | |
128 |
|
129 | |||
129 | When a stored revision is censored, filedata must be empty to have |
|
130 | When a stored revision is censored, filedata must be empty to have | |
130 | equivalence. |
|
131 | equivalence. | |
131 |
|
132 | |||
132 | When a stored revision has copy metadata, it is ignored as part |
|
133 | When a stored revision has copy metadata, it is ignored as part | |
133 | of the compare. |
|
134 | of the compare. | |
134 | """ |
|
135 | """ | |
135 |
|
136 | |||
136 | if filedata.startswith(b'\x01\n'): |
|
137 | if filedata.startswith(b'\x01\n'): | |
137 | revisiontext = b'\x01\n\x01\n' + filedata |
|
138 | revisiontext = b'\x01\n\x01\n' + filedata | |
138 | else: |
|
139 | else: | |
139 | revisiontext = filedata |
|
140 | revisiontext = filedata | |
140 |
|
141 | |||
141 | p1, p2 = store.parents(node) |
|
142 | p1, p2 = store.parents(node) | |
142 |
|
143 | |||
143 | computednode = hashrevisionsha1(revisiontext, p1, p2) |
|
144 | computednode = hashrevisionsha1(revisiontext, p1, p2) | |
144 |
|
145 | |||
145 | if computednode == node: |
|
146 | if computednode == node: | |
146 | return True |
|
147 | return True | |
147 |
|
148 | |||
148 | # Censored files compare against the empty file. |
|
149 | # Censored files compare against the empty file. | |
149 | if store.iscensored(store.rev(node)): |
|
150 | if store.iscensored(store.rev(node)): | |
150 | return filedata == b'' |
|
151 | return filedata == b'' | |
151 |
|
152 | |||
152 | # Renaming a file produces a different hash, even if the data |
|
153 | # Renaming a file produces a different hash, even if the data | |
153 | # remains unchanged. Check if that's the case. |
|
154 | # remains unchanged. Check if that's the case. | |
154 | if store.renamed(node): |
|
155 | if store.renamed(node): | |
155 | return store.read(node) == filedata |
|
156 | return store.read(node) == filedata | |
156 |
|
157 | |||
157 | return False |
|
158 | return False | |
158 |
|
159 | |||
159 |
|
160 | |||
160 | def iterrevs(storelen, start=0, stop=None): |
|
161 | def iterrevs(storelen, start=0, stop=None): | |
161 | """Iterate over revision numbers in a store.""" |
|
162 | """Iterate over revision numbers in a store.""" | |
162 | step = 1 |
|
163 | step = 1 | |
163 |
|
164 | |||
164 | if stop is not None: |
|
165 | if stop is not None: | |
165 | if start > stop: |
|
166 | if start > stop: | |
166 | step = -1 |
|
167 | step = -1 | |
167 | stop += step |
|
168 | stop += step | |
168 | if stop > storelen: |
|
169 | if stop > storelen: | |
169 | stop = storelen |
|
170 | stop = storelen | |
170 | else: |
|
171 | else: | |
171 | stop = storelen |
|
172 | stop = storelen | |
172 |
|
173 | |||
173 | return pycompat.xrange(start, stop, step) |
|
174 | return pycompat.xrange(start, stop, step) | |
174 |
|
175 | |||
175 |
|
176 | |||
176 | def fileidlookup(store, fileid, identifier): |
|
177 | def fileidlookup(store, fileid, identifier): | |
177 | """Resolve the file node for a value. |
|
178 | """Resolve the file node for a value. | |
178 |
|
179 | |||
179 | ``store`` is an object implementing the ``ifileindex`` interface. |
|
180 | ``store`` is an object implementing the ``ifileindex`` interface. | |
180 |
|
181 | |||
181 | ``fileid`` can be: |
|
182 | ``fileid`` can be: | |
182 |
|
183 | |||
183 | * A 20 or 32 byte binary node. |
|
184 | * A 20 or 32 byte binary node. | |
184 | * An integer revision number |
|
185 | * An integer revision number | |
185 | * A 40 or 64 byte hex node. |
|
186 | * A 40 or 64 byte hex node. | |
186 | * A bytes that can be parsed as an integer representing a revision number. |
|
187 | * A bytes that can be parsed as an integer representing a revision number. | |
187 |
|
188 | |||
188 | ``identifier`` is used to populate ``error.LookupError`` with an identifier |
|
189 | ``identifier`` is used to populate ``error.LookupError`` with an identifier | |
189 | for the store. |
|
190 | for the store. | |
190 |
|
191 | |||
191 | Raises ``error.LookupError`` on failure. |
|
192 | Raises ``error.LookupError`` on failure. | |
192 | """ |
|
193 | """ | |
193 | if isinstance(fileid, int): |
|
194 | if isinstance(fileid, int): | |
194 | try: |
|
195 | try: | |
195 | return store.node(fileid) |
|
196 | return store.node(fileid) | |
196 | except IndexError: |
|
197 | except IndexError: | |
197 | raise error.LookupError( |
|
198 | raise error.LookupError( | |
198 | b'%d' % fileid, identifier, _(b'no match found') |
|
199 | b'%d' % fileid, identifier, _(b'no match found') | |
199 | ) |
|
200 | ) | |
200 |
|
201 | |||
201 | if len(fileid) in (20, 32): |
|
202 | if len(fileid) in (20, 32): | |
202 | try: |
|
203 | try: | |
203 | store.rev(fileid) |
|
204 | store.rev(fileid) | |
204 | return fileid |
|
205 | return fileid | |
205 | except error.LookupError: |
|
206 | except error.LookupError: | |
206 | pass |
|
207 | pass | |
207 |
|
208 | |||
208 | if len(fileid) in (40, 64): |
|
209 | if len(fileid) in (40, 64): | |
209 | try: |
|
210 | try: | |
210 | rawnode = bin(fileid) |
|
211 | rawnode = bin(fileid) | |
211 | store.rev(rawnode) |
|
212 | store.rev(rawnode) | |
212 | return rawnode |
|
213 | return rawnode | |
213 | except TypeError: |
|
214 | except TypeError: | |
214 | pass |
|
215 | pass | |
215 |
|
216 | |||
216 | try: |
|
217 | try: | |
217 | rev = int(fileid) |
|
218 | rev = int(fileid) | |
218 |
|
219 | |||
219 | if b'%d' % rev != fileid: |
|
220 | if b'%d' % rev != fileid: | |
220 | raise ValueError |
|
221 | raise ValueError | |
221 |
|
222 | |||
222 | try: |
|
223 | try: | |
223 | return store.node(rev) |
|
224 | return store.node(rev) | |
224 | except (IndexError, TypeError): |
|
225 | except (IndexError, TypeError): | |
225 | pass |
|
226 | pass | |
226 | except (ValueError, OverflowError): |
|
227 | except (ValueError, OverflowError): | |
227 | pass |
|
228 | pass | |
228 |
|
229 | |||
229 | raise error.LookupError(fileid, identifier, _(b'no match found')) |
|
230 | raise error.LookupError(fileid, identifier, _(b'no match found')) | |
230 |
|
231 | |||
231 |
|
232 | |||
232 | def resolvestripinfo(minlinkrev, tiprev, headrevs, linkrevfn, parentrevsfn): |
|
233 | def resolvestripinfo(minlinkrev, tiprev, headrevs, linkrevfn, parentrevsfn): | |
233 | """Resolve information needed to strip revisions. |
|
234 | """Resolve information needed to strip revisions. | |
234 |
|
235 | |||
235 | Finds the minimum revision number that must be stripped in order to |
|
236 | Finds the minimum revision number that must be stripped in order to | |
236 | strip ``minlinkrev``. |
|
237 | strip ``minlinkrev``. | |
237 |
|
238 | |||
238 | Returns a 2-tuple of the minimum revision number to do that and a set |
|
239 | Returns a 2-tuple of the minimum revision number to do that and a set | |
239 | of all revision numbers that have linkrevs that would be broken |
|
240 | of all revision numbers that have linkrevs that would be broken | |
240 | by that strip. |
|
241 | by that strip. | |
241 |
|
242 | |||
242 | ``tiprev`` is the current tip-most revision. It is ``len(store) - 1``. |
|
243 | ``tiprev`` is the current tip-most revision. It is ``len(store) - 1``. | |
243 | ``headrevs`` is an iterable of head revisions. |
|
244 | ``headrevs`` is an iterable of head revisions. | |
244 | ``linkrevfn`` is a callable that receives a revision and returns a linked |
|
245 | ``linkrevfn`` is a callable that receives a revision and returns a linked | |
245 | revision. |
|
246 | revision. | |
246 | ``parentrevsfn`` is a callable that receives a revision number and returns |
|
247 | ``parentrevsfn`` is a callable that receives a revision number and returns | |
247 | an iterable of its parent revision numbers. |
|
248 | an iterable of its parent revision numbers. | |
248 | """ |
|
249 | """ | |
249 | brokenrevs = set() |
|
250 | brokenrevs = set() | |
250 | strippoint = tiprev + 1 |
|
251 | strippoint = tiprev + 1 | |
251 |
|
252 | |||
252 | heads = {} |
|
253 | heads = {} | |
253 | futurelargelinkrevs = set() |
|
254 | futurelargelinkrevs = set() | |
254 | for head in headrevs: |
|
255 | for head in headrevs: | |
255 | headlinkrev = linkrevfn(head) |
|
256 | headlinkrev = linkrevfn(head) | |
256 | heads[head] = headlinkrev |
|
257 | heads[head] = headlinkrev | |
257 | if headlinkrev >= minlinkrev: |
|
258 | if headlinkrev >= minlinkrev: | |
258 | futurelargelinkrevs.add(headlinkrev) |
|
259 | futurelargelinkrevs.add(headlinkrev) | |
259 |
|
260 | |||
260 | # This algorithm involves walking down the rev graph, starting at the |
|
261 | # This algorithm involves walking down the rev graph, starting at the | |
261 | # heads. Since the revs are topologically sorted according to linkrev, |
|
262 | # heads. Since the revs are topologically sorted according to linkrev, | |
262 | # once all head linkrevs are below the minlink, we know there are |
|
263 | # once all head linkrevs are below the minlink, we know there are | |
263 | # no more revs that could have a linkrev greater than minlink. |
|
264 | # no more revs that could have a linkrev greater than minlink. | |
264 | # So we can stop walking. |
|
265 | # So we can stop walking. | |
265 | while futurelargelinkrevs: |
|
266 | while futurelargelinkrevs: | |
266 | strippoint -= 1 |
|
267 | strippoint -= 1 | |
267 | linkrev = heads.pop(strippoint) |
|
268 | linkrev = heads.pop(strippoint) | |
268 |
|
269 | |||
269 | if linkrev < minlinkrev: |
|
270 | if linkrev < minlinkrev: | |
270 | brokenrevs.add(strippoint) |
|
271 | brokenrevs.add(strippoint) | |
271 | else: |
|
272 | else: | |
272 | futurelargelinkrevs.remove(linkrev) |
|
273 | futurelargelinkrevs.remove(linkrev) | |
273 |
|
274 | |||
274 | for p in parentrevsfn(strippoint): |
|
275 | for p in parentrevsfn(strippoint): | |
275 | if p != nullrev: |
|
276 | if p != nullrev: | |
276 | plinkrev = linkrevfn(p) |
|
277 | plinkrev = linkrevfn(p) | |
277 | heads[p] = plinkrev |
|
278 | heads[p] = plinkrev | |
278 | if plinkrev >= minlinkrev: |
|
279 | if plinkrev >= minlinkrev: | |
279 | futurelargelinkrevs.add(plinkrev) |
|
280 | futurelargelinkrevs.add(plinkrev) | |
280 |
|
281 | |||
281 | return strippoint, brokenrevs |
|
282 | return strippoint, brokenrevs | |
282 |
|
283 | |||
283 |
|
284 | |||
284 | def emitrevisions( |
|
285 | def emitrevisions( | |
285 | store, |
|
286 | store, | |
286 | nodes, |
|
287 | nodes, | |
287 | nodesorder, |
|
288 | nodesorder, | |
288 | resultcls, |
|
289 | resultcls, | |
289 | deltaparentfn=None, |
|
290 | deltaparentfn=None, | |
290 | candeltafn=None, |
|
291 | candeltafn=None, | |
291 | rawsizefn=None, |
|
292 | rawsizefn=None, | |
292 | revdifffn=None, |
|
293 | revdifffn=None, | |
293 | flagsfn=None, |
|
294 | flagsfn=None, | |
294 | deltamode=repository.CG_DELTAMODE_STD, |
|
295 | deltamode=repository.CG_DELTAMODE_STD, | |
295 | revisiondata=False, |
|
296 | revisiondata=False, | |
296 | assumehaveparentrevisions=False, |
|
297 | assumehaveparentrevisions=False, | |
|
298 | sidedata_helpers=None, | |||
297 | ): |
|
299 | ): | |
298 | """Generic implementation of ifiledata.emitrevisions(). |
|
300 | """Generic implementation of ifiledata.emitrevisions(). | |
299 |
|
301 | |||
300 | Emitting revision data is subtly complex. This function attempts to |
|
302 | Emitting revision data is subtly complex. This function attempts to | |
301 | encapsulate all the logic for doing so in a backend-agnostic way. |
|
303 | encapsulate all the logic for doing so in a backend-agnostic way. | |
302 |
|
304 | |||
303 | ``store`` |
|
305 | ``store`` | |
304 | Object conforming to ``ifilestorage`` interface. |
|
306 | Object conforming to ``ifilestorage`` interface. | |
305 |
|
307 | |||
306 | ``nodes`` |
|
308 | ``nodes`` | |
307 | List of revision nodes whose data to emit. |
|
309 | List of revision nodes whose data to emit. | |
308 |
|
310 | |||
309 | ``resultcls`` |
|
311 | ``resultcls`` | |
310 | A type implementing the ``irevisiondelta`` interface that will be |
|
312 | A type implementing the ``irevisiondelta`` interface that will be | |
311 | constructed and returned. |
|
313 | constructed and returned. | |
312 |
|
314 | |||
313 | ``deltaparentfn`` (optional) |
|
315 | ``deltaparentfn`` (optional) | |
314 | Callable receiving a revision number and returning the revision number |
|
316 | Callable receiving a revision number and returning the revision number | |
315 | of a revision that the internal delta is stored against. This delta |
|
317 | of a revision that the internal delta is stored against. This delta | |
316 | will be preferred over computing a new arbitrary delta. |
|
318 | will be preferred over computing a new arbitrary delta. | |
317 |
|
319 | |||
318 | If not defined, a delta will always be computed from raw revision |
|
320 | If not defined, a delta will always be computed from raw revision | |
319 | data. |
|
321 | data. | |
320 |
|
322 | |||
321 | ``candeltafn`` (optional) |
|
323 | ``candeltafn`` (optional) | |
322 | Callable receiving a pair of revision numbers that returns a bool |
|
324 | Callable receiving a pair of revision numbers that returns a bool | |
323 | indicating whether a delta between them can be produced. |
|
325 | indicating whether a delta between them can be produced. | |
324 |
|
326 | |||
325 | If not defined, it is assumed that any two revisions can delta with |
|
327 | If not defined, it is assumed that any two revisions can delta with | |
326 | each other. |
|
328 | each other. | |
327 |
|
329 | |||
328 | ``rawsizefn`` (optional) |
|
330 | ``rawsizefn`` (optional) | |
329 | Callable receiving a revision number and returning the length of the |
|
331 | Callable receiving a revision number and returning the length of the | |
330 | ``store.rawdata(rev)``. |
|
332 | ``store.rawdata(rev)``. | |
331 |
|
333 | |||
332 | If not defined, ``len(store.rawdata(rev))`` will be called. |
|
334 | If not defined, ``len(store.rawdata(rev))`` will be called. | |
333 |
|
335 | |||
334 | ``revdifffn`` (optional) |
|
336 | ``revdifffn`` (optional) | |
335 | Callable receiving a pair of revision numbers that returns a delta |
|
337 | Callable receiving a pair of revision numbers that returns a delta | |
336 | between them. |
|
338 | between them. | |
337 |
|
339 | |||
338 | If not defined, a delta will be computed by invoking mdiff code |
|
340 | If not defined, a delta will be computed by invoking mdiff code | |
339 | on ``store.revision()`` results. |
|
341 | on ``store.revision()`` results. | |
340 |
|
342 | |||
341 | Defining this function allows a precomputed or stored delta to be |
|
343 | Defining this function allows a precomputed or stored delta to be | |
342 | used without having to compute on. |
|
344 | used without having to compute on. | |
343 |
|
345 | |||
344 | ``flagsfn`` (optional) |
|
346 | ``flagsfn`` (optional) | |
345 | Callable receiving a revision number and returns the integer flags |
|
347 | Callable receiving a revision number and returns the integer flags | |
346 | value for it. If not defined, flags value will be 0. |
|
348 | value for it. If not defined, flags value will be 0. | |
347 |
|
349 | |||
348 | ``deltamode`` |
|
350 | ``deltamode`` | |
349 | constaint on delta to be sent: |
|
351 | constaint on delta to be sent: | |
350 | * CG_DELTAMODE_STD - normal mode, try to reuse storage deltas, |
|
352 | * CG_DELTAMODE_STD - normal mode, try to reuse storage deltas, | |
351 | * CG_DELTAMODE_PREV - only delta against "prev", |
|
353 | * CG_DELTAMODE_PREV - only delta against "prev", | |
352 | * CG_DELTAMODE_FULL - only issue full snapshot. |
|
354 | * CG_DELTAMODE_FULL - only issue full snapshot. | |
353 |
|
355 | |||
354 | Whether to send fulltext revisions instead of deltas, if allowed. |
|
356 | Whether to send fulltext revisions instead of deltas, if allowed. | |
355 |
|
357 | |||
356 | ``nodesorder`` |
|
358 | ``nodesorder`` | |
357 | ``revisiondata`` |
|
359 | ``revisiondata`` | |
358 | ``assumehaveparentrevisions`` |
|
360 | ``assumehaveparentrevisions`` | |
|
361 | ``sidedata_helpers`` (optional) | |||
|
362 | If not None, means that sidedata should be included. | |||
|
363 | A dictionary of revlog type to tuples of `(repo, computers, removers)`: | |||
|
364 | * `repo` is used as an argument for computers | |||
|
365 | * `computers` is a list of `(category, (keys, computer)` that | |||
|
366 | compute the missing sidedata categories that were asked: | |||
|
367 | * `category` is the sidedata category | |||
|
368 | * `keys` are the sidedata keys to be affected | |||
|
369 | * `computer` is the function `(repo, store, rev, sidedata)` that | |||
|
370 | returns a new sidedata dict. | |||
|
371 | * `removers` will remove the keys corresponding to the categories | |||
|
372 | that are present, but not needed. | |||
|
373 | If both `computers` and `removers` are empty, sidedata are simply not | |||
|
374 | transformed. | |||
|
375 | Revlog types are `changelog`, `manifest` or `filelog`. | |||
359 | """ |
|
376 | """ | |
360 |
|
377 | |||
361 | fnode = store.node |
|
378 | fnode = store.node | |
362 | frev = store.rev |
|
379 | frev = store.rev | |
363 |
|
380 | |||
364 | if nodesorder == b'nodes': |
|
381 | if nodesorder == b'nodes': | |
365 | revs = [frev(n) for n in nodes] |
|
382 | revs = [frev(n) for n in nodes] | |
366 | elif nodesorder == b'linear': |
|
383 | elif nodesorder == b'linear': | |
367 | revs = {frev(n) for n in nodes} |
|
384 | revs = {frev(n) for n in nodes} | |
368 | revs = dagop.linearize(revs, store.parentrevs) |
|
385 | revs = dagop.linearize(revs, store.parentrevs) | |
369 | else: # storage and default |
|
386 | else: # storage and default | |
370 | revs = sorted(frev(n) for n in nodes) |
|
387 | revs = sorted(frev(n) for n in nodes) | |
371 |
|
388 | |||
372 | prevrev = None |
|
389 | prevrev = None | |
373 |
|
390 | |||
374 | if deltamode == repository.CG_DELTAMODE_PREV or assumehaveparentrevisions: |
|
391 | if deltamode == repository.CG_DELTAMODE_PREV or assumehaveparentrevisions: | |
375 | prevrev = store.parentrevs(revs[0])[0] |
|
392 | prevrev = store.parentrevs(revs[0])[0] | |
376 |
|
393 | |||
377 | # Set of revs available to delta against. |
|
394 | # Set of revs available to delta against. | |
378 | available = set() |
|
395 | available = set() | |
379 |
|
396 | |||
380 | for rev in revs: |
|
397 | for rev in revs: | |
381 | if rev == nullrev: |
|
398 | if rev == nullrev: | |
382 | continue |
|
399 | continue | |
383 |
|
400 | |||
384 | node = fnode(rev) |
|
401 | node = fnode(rev) | |
385 | p1rev, p2rev = store.parentrevs(rev) |
|
402 | p1rev, p2rev = store.parentrevs(rev) | |
386 |
|
403 | |||
387 | if deltaparentfn: |
|
404 | if deltaparentfn: | |
388 | deltaparentrev = deltaparentfn(rev) |
|
405 | deltaparentrev = deltaparentfn(rev) | |
389 | else: |
|
406 | else: | |
390 | deltaparentrev = nullrev |
|
407 | deltaparentrev = nullrev | |
391 |
|
408 | |||
392 | # Forced delta against previous mode. |
|
409 | # Forced delta against previous mode. | |
393 | if deltamode == repository.CG_DELTAMODE_PREV: |
|
410 | if deltamode == repository.CG_DELTAMODE_PREV: | |
394 | baserev = prevrev |
|
411 | baserev = prevrev | |
395 |
|
412 | |||
396 | # We're instructed to send fulltext. Honor that. |
|
413 | # We're instructed to send fulltext. Honor that. | |
397 | elif deltamode == repository.CG_DELTAMODE_FULL: |
|
414 | elif deltamode == repository.CG_DELTAMODE_FULL: | |
398 | baserev = nullrev |
|
415 | baserev = nullrev | |
399 | # We're instructed to use p1. Honor that |
|
416 | # We're instructed to use p1. Honor that | |
400 | elif deltamode == repository.CG_DELTAMODE_P1: |
|
417 | elif deltamode == repository.CG_DELTAMODE_P1: | |
401 | baserev = p1rev |
|
418 | baserev = p1rev | |
402 |
|
419 | |||
403 | # There is a delta in storage. We try to use that because it |
|
420 | # There is a delta in storage. We try to use that because it | |
404 | # amounts to effectively copying data from storage and is |
|
421 | # amounts to effectively copying data from storage and is | |
405 | # therefore the fastest. |
|
422 | # therefore the fastest. | |
406 | elif deltaparentrev != nullrev: |
|
423 | elif deltaparentrev != nullrev: | |
407 | # Base revision was already emitted in this group. We can |
|
424 | # Base revision was already emitted in this group. We can | |
408 | # always safely use the delta. |
|
425 | # always safely use the delta. | |
409 | if deltaparentrev in available: |
|
426 | if deltaparentrev in available: | |
410 | baserev = deltaparentrev |
|
427 | baserev = deltaparentrev | |
411 |
|
428 | |||
412 | # Base revision is a parent that hasn't been emitted already. |
|
429 | # Base revision is a parent that hasn't been emitted already. | |
413 | # Use it if we can assume the receiver has the parent revision. |
|
430 | # Use it if we can assume the receiver has the parent revision. | |
414 | elif assumehaveparentrevisions and deltaparentrev in (p1rev, p2rev): |
|
431 | elif assumehaveparentrevisions and deltaparentrev in (p1rev, p2rev): | |
415 | baserev = deltaparentrev |
|
432 | baserev = deltaparentrev | |
416 |
|
433 | |||
417 | # No guarantee the receiver has the delta parent. Send delta |
|
434 | # No guarantee the receiver has the delta parent. Send delta | |
418 | # against last revision (if possible), which in the common case |
|
435 | # against last revision (if possible), which in the common case | |
419 | # should be similar enough to this revision that the delta is |
|
436 | # should be similar enough to this revision that the delta is | |
420 | # reasonable. |
|
437 | # reasonable. | |
421 | elif prevrev is not None: |
|
438 | elif prevrev is not None: | |
422 | baserev = prevrev |
|
439 | baserev = prevrev | |
423 | else: |
|
440 | else: | |
424 | baserev = nullrev |
|
441 | baserev = nullrev | |
425 |
|
442 | |||
426 | # Storage has a fulltext revision. |
|
443 | # Storage has a fulltext revision. | |
427 |
|
444 | |||
428 | # Let's use the previous revision, which is as good a guess as any. |
|
445 | # Let's use the previous revision, which is as good a guess as any. | |
429 | # There is definitely room to improve this logic. |
|
446 | # There is definitely room to improve this logic. | |
430 | elif prevrev is not None: |
|
447 | elif prevrev is not None: | |
431 | baserev = prevrev |
|
448 | baserev = prevrev | |
432 | else: |
|
449 | else: | |
433 | baserev = nullrev |
|
450 | baserev = nullrev | |
434 |
|
451 | |||
435 | # But we can't actually use our chosen delta base for whatever |
|
452 | # But we can't actually use our chosen delta base for whatever | |
436 | # reason. Reset to fulltext. |
|
453 | # reason. Reset to fulltext. | |
437 | if baserev != nullrev and (candeltafn and not candeltafn(baserev, rev)): |
|
454 | if baserev != nullrev and (candeltafn and not candeltafn(baserev, rev)): | |
438 | baserev = nullrev |
|
455 | baserev = nullrev | |
439 |
|
456 | |||
440 | revision = None |
|
457 | revision = None | |
441 | delta = None |
|
458 | delta = None | |
442 | baserevisionsize = None |
|
459 | baserevisionsize = None | |
443 |
|
460 | |||
444 | if revisiondata: |
|
461 | if revisiondata: | |
445 | if store.iscensored(baserev) or store.iscensored(rev): |
|
462 | if store.iscensored(baserev) or store.iscensored(rev): | |
446 | try: |
|
463 | try: | |
447 | revision = store.rawdata(node) |
|
464 | revision = store.rawdata(node) | |
448 | except error.CensoredNodeError as e: |
|
465 | except error.CensoredNodeError as e: | |
449 | revision = e.tombstone |
|
466 | revision = e.tombstone | |
450 |
|
467 | |||
451 | if baserev != nullrev: |
|
468 | if baserev != nullrev: | |
452 | if rawsizefn: |
|
469 | if rawsizefn: | |
453 | baserevisionsize = rawsizefn(baserev) |
|
470 | baserevisionsize = rawsizefn(baserev) | |
454 | else: |
|
471 | else: | |
455 | baserevisionsize = len(store.rawdata(baserev)) |
|
472 | baserevisionsize = len(store.rawdata(baserev)) | |
456 |
|
473 | |||
457 | elif ( |
|
474 | elif ( | |
458 | baserev == nullrev and deltamode != repository.CG_DELTAMODE_PREV |
|
475 | baserev == nullrev and deltamode != repository.CG_DELTAMODE_PREV | |
459 | ): |
|
476 | ): | |
460 | revision = store.rawdata(node) |
|
477 | revision = store.rawdata(node) | |
461 | available.add(rev) |
|
478 | available.add(rev) | |
462 | else: |
|
479 | else: | |
463 | if revdifffn: |
|
480 | if revdifffn: | |
464 | delta = revdifffn(baserev, rev) |
|
481 | delta = revdifffn(baserev, rev) | |
465 | else: |
|
482 | else: | |
466 | delta = mdiff.textdiff( |
|
483 | delta = mdiff.textdiff( | |
467 | store.rawdata(baserev), store.rawdata(rev) |
|
484 | store.rawdata(baserev), store.rawdata(rev) | |
468 | ) |
|
485 | ) | |
469 |
|
486 | |||
470 | available.add(rev) |
|
487 | available.add(rev) | |
471 |
|
488 | |||
|
489 | sidedata = None | |||
|
490 | if sidedata_helpers: | |||
|
491 | sidedata = store.sidedata(rev) | |||
|
492 | sidedata = run_sidedata_helpers( | |||
|
493 | store=store, | |||
|
494 | sidedata_helpers=sidedata_helpers, | |||
|
495 | sidedata=sidedata, | |||
|
496 | rev=rev, | |||
|
497 | ) | |||
|
498 | sidedata = sidedatamod.serialize_sidedata(sidedata) | |||
|
499 | ||||
472 | yield resultcls( |
|
500 | yield resultcls( | |
473 | node=node, |
|
501 | node=node, | |
474 | p1node=fnode(p1rev), |
|
502 | p1node=fnode(p1rev), | |
475 | p2node=fnode(p2rev), |
|
503 | p2node=fnode(p2rev), | |
476 | basenode=fnode(baserev), |
|
504 | basenode=fnode(baserev), | |
477 | flags=flagsfn(rev) if flagsfn else 0, |
|
505 | flags=flagsfn(rev) if flagsfn else 0, | |
478 | baserevisionsize=baserevisionsize, |
|
506 | baserevisionsize=baserevisionsize, | |
479 | revision=revision, |
|
507 | revision=revision, | |
480 | delta=delta, |
|
508 | delta=delta, | |
481 | sidedata=sidedata, |
|
509 | sidedata=sidedata, | |
482 | ) |
|
510 | ) | |
483 |
|
511 | |||
484 | prevrev = rev |
|
512 | prevrev = rev | |
485 |
|
513 | |||
486 |
|
514 | |||
|
515 | def run_sidedata_helpers(store, sidedata_helpers, sidedata, rev): | |||
|
516 | """Returns the sidedata for the given revision after running through | |||
|
517 | the given helpers. | |||
|
518 | - `store`: the revlog this applies to (changelog, manifest, or filelog | |||
|
519 | instance) | |||
|
520 | - `sidedata_helpers`: see `storageutil.emitrevisions` | |||
|
521 | - `sidedata`: previous sidedata at the given rev, if any | |||
|
522 | - `rev`: affected rev of `store` | |||
|
523 | """ | |||
|
524 | repo, sd_computers, sd_removers = sidedata_helpers | |||
|
525 | kind = store.revlog_kind | |||
|
526 | for _keys, sd_computer in sd_computers.get(kind, []): | |||
|
527 | sidedata = sd_computer(repo, store, rev, sidedata) | |||
|
528 | for keys, _computer in sd_removers.get(kind, []): | |||
|
529 | for key in keys: | |||
|
530 | sidedata.pop(key, None) | |||
|
531 | return sidedata | |||
|
532 | ||||
|
533 | ||||
487 | def deltaiscensored(delta, baserev, baselenfn): |
|
534 | def deltaiscensored(delta, baserev, baselenfn): | |
488 | """Determine if a delta represents censored revision data. |
|
535 | """Determine if a delta represents censored revision data. | |
489 |
|
536 | |||
490 | ``baserev`` is the base revision this delta is encoded against. |
|
537 | ``baserev`` is the base revision this delta is encoded against. | |
491 | ``baselenfn`` is a callable receiving a revision number that resolves the |
|
538 | ``baselenfn`` is a callable receiving a revision number that resolves the | |
492 | length of the revision fulltext. |
|
539 | length of the revision fulltext. | |
493 |
|
540 | |||
494 | Returns a bool indicating if the result of the delta represents a censored |
|
541 | Returns a bool indicating if the result of the delta represents a censored | |
495 | revision. |
|
542 | revision. | |
496 | """ |
|
543 | """ | |
497 | # Fragile heuristic: unless new file meta keys are added alphabetically |
|
544 | # Fragile heuristic: unless new file meta keys are added alphabetically | |
498 | # preceding "censored", all censored revisions are prefixed by |
|
545 | # preceding "censored", all censored revisions are prefixed by | |
499 | # "\1\ncensored:". A delta producing such a censored revision must be a |
|
546 | # "\1\ncensored:". A delta producing such a censored revision must be a | |
500 | # full-replacement delta, so we inspect the first and only patch in the |
|
547 | # full-replacement delta, so we inspect the first and only patch in the | |
501 | # delta for this prefix. |
|
548 | # delta for this prefix. | |
502 | hlen = struct.calcsize(b">lll") |
|
549 | hlen = struct.calcsize(b">lll") | |
503 | if len(delta) <= hlen: |
|
550 | if len(delta) <= hlen: | |
504 | return False |
|
551 | return False | |
505 |
|
552 | |||
506 | oldlen = baselenfn(baserev) |
|
553 | oldlen = baselenfn(baserev) | |
507 | newlen = len(delta) - hlen |
|
554 | newlen = len(delta) - hlen | |
508 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): |
|
555 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): | |
509 | return False |
|
556 | return False | |
510 |
|
557 | |||
511 | add = b"\1\ncensored:" |
|
558 | add = b"\1\ncensored:" | |
512 | addlen = len(add) |
|
559 | addlen = len(add) | |
513 | return newlen >= addlen and delta[hlen : hlen + addlen] == add |
|
560 | return newlen >= addlen and delta[hlen : hlen + addlen] == add |
@@ -1,738 +1,740 b'' | |||||
1 | # simplestorerepo.py - Extension that swaps in alternate repository storage. |
|
1 | # simplestorerepo.py - Extension that swaps in alternate repository storage. | |
2 | # |
|
2 | # | |
3 | # Copyright 2018 Gregory Szorc <gregory.szorc@gmail.com> |
|
3 | # Copyright 2018 Gregory Szorc <gregory.szorc@gmail.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | # To use this with the test suite: |
|
8 | # To use this with the test suite: | |
9 | # |
|
9 | # | |
10 | # $ HGREPOFEATURES="simplestore" ./run-tests.py \ |
|
10 | # $ HGREPOFEATURES="simplestore" ./run-tests.py \ | |
11 | # --extra-config-opt extensions.simplestore=`pwd`/simplestorerepo.py |
|
11 | # --extra-config-opt extensions.simplestore=`pwd`/simplestorerepo.py | |
12 |
|
12 | |||
13 | from __future__ import absolute_import |
|
13 | from __future__ import absolute_import | |
14 |
|
14 | |||
15 | import stat |
|
15 | import stat | |
16 |
|
16 | |||
17 | from mercurial.i18n import _ |
|
17 | from mercurial.i18n import _ | |
18 | from mercurial.node import ( |
|
18 | from mercurial.node import ( | |
19 | bin, |
|
19 | bin, | |
20 | hex, |
|
20 | hex, | |
21 | nullid, |
|
21 | nullid, | |
22 | nullrev, |
|
22 | nullrev, | |
23 | ) |
|
23 | ) | |
24 | from mercurial.thirdparty import attr |
|
24 | from mercurial.thirdparty import attr | |
25 | from mercurial import ( |
|
25 | from mercurial import ( | |
26 | ancestor, |
|
26 | ancestor, | |
27 | bundlerepo, |
|
27 | bundlerepo, | |
28 | error, |
|
28 | error, | |
29 | extensions, |
|
29 | extensions, | |
30 | localrepo, |
|
30 | localrepo, | |
31 | mdiff, |
|
31 | mdiff, | |
32 | pycompat, |
|
32 | pycompat, | |
33 | revlog, |
|
33 | revlog, | |
34 | store, |
|
34 | store, | |
35 | verify, |
|
35 | verify, | |
36 | ) |
|
36 | ) | |
37 | from mercurial.interfaces import ( |
|
37 | from mercurial.interfaces import ( | |
38 | repository, |
|
38 | repository, | |
39 | util as interfaceutil, |
|
39 | util as interfaceutil, | |
40 | ) |
|
40 | ) | |
41 | from mercurial.utils import ( |
|
41 | from mercurial.utils import ( | |
42 | cborutil, |
|
42 | cborutil, | |
43 | storageutil, |
|
43 | storageutil, | |
44 | ) |
|
44 | ) | |
45 | from mercurial.revlogutils import flagutil |
|
45 | from mercurial.revlogutils import flagutil | |
46 |
|
46 | |||
47 | # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for |
|
47 | # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for | |
48 | # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should |
|
48 | # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should | |
49 | # be specifying the version(s) of Mercurial they are tested with, or |
|
49 | # be specifying the version(s) of Mercurial they are tested with, or | |
50 | # leave the attribute unspecified. |
|
50 | # leave the attribute unspecified. | |
51 | testedwith = b'ships-with-hg-core' |
|
51 | testedwith = b'ships-with-hg-core' | |
52 |
|
52 | |||
53 | REQUIREMENT = b'testonly-simplestore' |
|
53 | REQUIREMENT = b'testonly-simplestore' | |
54 |
|
54 | |||
55 |
|
55 | |||
56 | def validatenode(node): |
|
56 | def validatenode(node): | |
57 | if isinstance(node, int): |
|
57 | if isinstance(node, int): | |
58 | raise ValueError('expected node; got int') |
|
58 | raise ValueError('expected node; got int') | |
59 |
|
59 | |||
60 | if len(node) != 20: |
|
60 | if len(node) != 20: | |
61 | raise ValueError('expected 20 byte node') |
|
61 | raise ValueError('expected 20 byte node') | |
62 |
|
62 | |||
63 |
|
63 | |||
64 | def validaterev(rev): |
|
64 | def validaterev(rev): | |
65 | if not isinstance(rev, int): |
|
65 | if not isinstance(rev, int): | |
66 | raise ValueError('expected int') |
|
66 | raise ValueError('expected int') | |
67 |
|
67 | |||
68 |
|
68 | |||
69 | class simplestoreerror(error.StorageError): |
|
69 | class simplestoreerror(error.StorageError): | |
70 | pass |
|
70 | pass | |
71 |
|
71 | |||
72 |
|
72 | |||
73 | @interfaceutil.implementer(repository.irevisiondelta) |
|
73 | @interfaceutil.implementer(repository.irevisiondelta) | |
74 | @attr.s(slots=True) |
|
74 | @attr.s(slots=True) | |
75 | class simplestorerevisiondelta(object): |
|
75 | class simplestorerevisiondelta(object): | |
76 | node = attr.ib() |
|
76 | node = attr.ib() | |
77 | p1node = attr.ib() |
|
77 | p1node = attr.ib() | |
78 | p2node = attr.ib() |
|
78 | p2node = attr.ib() | |
79 | basenode = attr.ib() |
|
79 | basenode = attr.ib() | |
80 | flags = attr.ib() |
|
80 | flags = attr.ib() | |
81 | baserevisionsize = attr.ib() |
|
81 | baserevisionsize = attr.ib() | |
82 | revision = attr.ib() |
|
82 | revision = attr.ib() | |
83 | delta = attr.ib() |
|
83 | delta = attr.ib() | |
84 | linknode = attr.ib(default=None) |
|
84 | linknode = attr.ib(default=None) | |
85 |
|
85 | |||
86 |
|
86 | |||
87 | @interfaceutil.implementer(repository.iverifyproblem) |
|
87 | @interfaceutil.implementer(repository.iverifyproblem) | |
88 | @attr.s(frozen=True) |
|
88 | @attr.s(frozen=True) | |
89 | class simplefilestoreproblem(object): |
|
89 | class simplefilestoreproblem(object): | |
90 | warning = attr.ib(default=None) |
|
90 | warning = attr.ib(default=None) | |
91 | error = attr.ib(default=None) |
|
91 | error = attr.ib(default=None) | |
92 | node = attr.ib(default=None) |
|
92 | node = attr.ib(default=None) | |
93 |
|
93 | |||
94 |
|
94 | |||
95 | @interfaceutil.implementer(repository.ifilestorage) |
|
95 | @interfaceutil.implementer(repository.ifilestorage) | |
96 | class filestorage(object): |
|
96 | class filestorage(object): | |
97 | """Implements storage for a tracked path. |
|
97 | """Implements storage for a tracked path. | |
98 |
|
98 | |||
99 | Data is stored in the VFS in a directory corresponding to the tracked |
|
99 | Data is stored in the VFS in a directory corresponding to the tracked | |
100 | path. |
|
100 | path. | |
101 |
|
101 | |||
102 | Index data is stored in an ``index`` file using CBOR. |
|
102 | Index data is stored in an ``index`` file using CBOR. | |
103 |
|
103 | |||
104 | Fulltext data is stored in files having names of the node. |
|
104 | Fulltext data is stored in files having names of the node. | |
105 | """ |
|
105 | """ | |
106 |
|
106 | |||
107 | _flagserrorclass = simplestoreerror |
|
107 | _flagserrorclass = simplestoreerror | |
108 |
|
108 | |||
109 | def __init__(self, svfs, path): |
|
109 | def __init__(self, svfs, path): | |
110 | self._svfs = svfs |
|
110 | self._svfs = svfs | |
111 | self._path = path |
|
111 | self._path = path | |
112 |
|
112 | |||
113 | self._storepath = b'/'.join([b'data', path]) |
|
113 | self._storepath = b'/'.join([b'data', path]) | |
114 | self._indexpath = b'/'.join([self._storepath, b'index']) |
|
114 | self._indexpath = b'/'.join([self._storepath, b'index']) | |
115 |
|
115 | |||
116 | indexdata = self._svfs.tryread(self._indexpath) |
|
116 | indexdata = self._svfs.tryread(self._indexpath) | |
117 | if indexdata: |
|
117 | if indexdata: | |
118 | indexdata = cborutil.decodeall(indexdata) |
|
118 | indexdata = cborutil.decodeall(indexdata) | |
119 |
|
119 | |||
120 | self._indexdata = indexdata or [] |
|
120 | self._indexdata = indexdata or [] | |
121 | self._indexbynode = {} |
|
121 | self._indexbynode = {} | |
122 | self._indexbyrev = {} |
|
122 | self._indexbyrev = {} | |
123 | self._index = [] |
|
123 | self._index = [] | |
124 | self._refreshindex() |
|
124 | self._refreshindex() | |
125 |
|
125 | |||
126 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
126 | self._flagprocessors = dict(flagutil.flagprocessors) | |
127 |
|
127 | |||
128 | def _refreshindex(self): |
|
128 | def _refreshindex(self): | |
129 | self._indexbynode.clear() |
|
129 | self._indexbynode.clear() | |
130 | self._indexbyrev.clear() |
|
130 | self._indexbyrev.clear() | |
131 | self._index = [] |
|
131 | self._index = [] | |
132 |
|
132 | |||
133 | for i, entry in enumerate(self._indexdata): |
|
133 | for i, entry in enumerate(self._indexdata): | |
134 | self._indexbynode[entry[b'node']] = entry |
|
134 | self._indexbynode[entry[b'node']] = entry | |
135 | self._indexbyrev[i] = entry |
|
135 | self._indexbyrev[i] = entry | |
136 |
|
136 | |||
137 | self._indexbynode[nullid] = { |
|
137 | self._indexbynode[nullid] = { | |
138 | b'node': nullid, |
|
138 | b'node': nullid, | |
139 | b'p1': nullid, |
|
139 | b'p1': nullid, | |
140 | b'p2': nullid, |
|
140 | b'p2': nullid, | |
141 | b'linkrev': nullrev, |
|
141 | b'linkrev': nullrev, | |
142 | b'flags': 0, |
|
142 | b'flags': 0, | |
143 | } |
|
143 | } | |
144 |
|
144 | |||
145 | self._indexbyrev[nullrev] = { |
|
145 | self._indexbyrev[nullrev] = { | |
146 | b'node': nullid, |
|
146 | b'node': nullid, | |
147 | b'p1': nullid, |
|
147 | b'p1': nullid, | |
148 | b'p2': nullid, |
|
148 | b'p2': nullid, | |
149 | b'linkrev': nullrev, |
|
149 | b'linkrev': nullrev, | |
150 | b'flags': 0, |
|
150 | b'flags': 0, | |
151 | } |
|
151 | } | |
152 |
|
152 | |||
153 | for i, entry in enumerate(self._indexdata): |
|
153 | for i, entry in enumerate(self._indexdata): | |
154 | p1rev, p2rev = self.parentrevs(self.rev(entry[b'node'])) |
|
154 | p1rev, p2rev = self.parentrevs(self.rev(entry[b'node'])) | |
155 |
|
155 | |||
156 | # start, length, rawsize, chainbase, linkrev, p1, p2, node |
|
156 | # start, length, rawsize, chainbase, linkrev, p1, p2, node | |
157 | self._index.append( |
|
157 | self._index.append( | |
158 | (0, 0, 0, -1, entry[b'linkrev'], p1rev, p2rev, entry[b'node']) |
|
158 | (0, 0, 0, -1, entry[b'linkrev'], p1rev, p2rev, entry[b'node']) | |
159 | ) |
|
159 | ) | |
160 |
|
160 | |||
161 | self._index.append((0, 0, 0, -1, -1, -1, -1, nullid)) |
|
161 | self._index.append((0, 0, 0, -1, -1, -1, -1, nullid)) | |
162 |
|
162 | |||
163 | def __len__(self): |
|
163 | def __len__(self): | |
164 | return len(self._indexdata) |
|
164 | return len(self._indexdata) | |
165 |
|
165 | |||
166 | def __iter__(self): |
|
166 | def __iter__(self): | |
167 | return iter(range(len(self))) |
|
167 | return iter(range(len(self))) | |
168 |
|
168 | |||
169 | def revs(self, start=0, stop=None): |
|
169 | def revs(self, start=0, stop=None): | |
170 | step = 1 |
|
170 | step = 1 | |
171 | if stop is not None: |
|
171 | if stop is not None: | |
172 | if start > stop: |
|
172 | if start > stop: | |
173 | step = -1 |
|
173 | step = -1 | |
174 |
|
174 | |||
175 | stop += step |
|
175 | stop += step | |
176 | else: |
|
176 | else: | |
177 | stop = len(self) |
|
177 | stop = len(self) | |
178 |
|
178 | |||
179 | return range(start, stop, step) |
|
179 | return range(start, stop, step) | |
180 |
|
180 | |||
181 | def parents(self, node): |
|
181 | def parents(self, node): | |
182 | validatenode(node) |
|
182 | validatenode(node) | |
183 |
|
183 | |||
184 | if node not in self._indexbynode: |
|
184 | if node not in self._indexbynode: | |
185 | raise KeyError('unknown node') |
|
185 | raise KeyError('unknown node') | |
186 |
|
186 | |||
187 | entry = self._indexbynode[node] |
|
187 | entry = self._indexbynode[node] | |
188 |
|
188 | |||
189 | return entry[b'p1'], entry[b'p2'] |
|
189 | return entry[b'p1'], entry[b'p2'] | |
190 |
|
190 | |||
191 | def parentrevs(self, rev): |
|
191 | def parentrevs(self, rev): | |
192 | p1, p2 = self.parents(self._indexbyrev[rev][b'node']) |
|
192 | p1, p2 = self.parents(self._indexbyrev[rev][b'node']) | |
193 | return self.rev(p1), self.rev(p2) |
|
193 | return self.rev(p1), self.rev(p2) | |
194 |
|
194 | |||
195 | def rev(self, node): |
|
195 | def rev(self, node): | |
196 | validatenode(node) |
|
196 | validatenode(node) | |
197 |
|
197 | |||
198 | try: |
|
198 | try: | |
199 | self._indexbynode[node] |
|
199 | self._indexbynode[node] | |
200 | except KeyError: |
|
200 | except KeyError: | |
201 | raise error.LookupError(node, self._indexpath, _('no node')) |
|
201 | raise error.LookupError(node, self._indexpath, _('no node')) | |
202 |
|
202 | |||
203 | for rev, entry in self._indexbyrev.items(): |
|
203 | for rev, entry in self._indexbyrev.items(): | |
204 | if entry[b'node'] == node: |
|
204 | if entry[b'node'] == node: | |
205 | return rev |
|
205 | return rev | |
206 |
|
206 | |||
207 | raise error.ProgrammingError(b'this should not occur') |
|
207 | raise error.ProgrammingError(b'this should not occur') | |
208 |
|
208 | |||
209 | def node(self, rev): |
|
209 | def node(self, rev): | |
210 | validaterev(rev) |
|
210 | validaterev(rev) | |
211 |
|
211 | |||
212 | return self._indexbyrev[rev][b'node'] |
|
212 | return self._indexbyrev[rev][b'node'] | |
213 |
|
213 | |||
214 | def hasnode(self, node): |
|
214 | def hasnode(self, node): | |
215 | validatenode(node) |
|
215 | validatenode(node) | |
216 | return node in self._indexbynode |
|
216 | return node in self._indexbynode | |
217 |
|
217 | |||
218 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
218 | def censorrevision(self, tr, censornode, tombstone=b''): | |
219 | raise NotImplementedError('TODO') |
|
219 | raise NotImplementedError('TODO') | |
220 |
|
220 | |||
221 | def lookup(self, node): |
|
221 | def lookup(self, node): | |
222 | if isinstance(node, int): |
|
222 | if isinstance(node, int): | |
223 | return self.node(node) |
|
223 | return self.node(node) | |
224 |
|
224 | |||
225 | if len(node) == 20: |
|
225 | if len(node) == 20: | |
226 | self.rev(node) |
|
226 | self.rev(node) | |
227 | return node |
|
227 | return node | |
228 |
|
228 | |||
229 | try: |
|
229 | try: | |
230 | rev = int(node) |
|
230 | rev = int(node) | |
231 | if '%d' % rev != node: |
|
231 | if '%d' % rev != node: | |
232 | raise ValueError |
|
232 | raise ValueError | |
233 |
|
233 | |||
234 | if rev < 0: |
|
234 | if rev < 0: | |
235 | rev = len(self) + rev |
|
235 | rev = len(self) + rev | |
236 | if rev < 0 or rev >= len(self): |
|
236 | if rev < 0 or rev >= len(self): | |
237 | raise ValueError |
|
237 | raise ValueError | |
238 |
|
238 | |||
239 | return self.node(rev) |
|
239 | return self.node(rev) | |
240 | except (ValueError, OverflowError): |
|
240 | except (ValueError, OverflowError): | |
241 | pass |
|
241 | pass | |
242 |
|
242 | |||
243 | if len(node) == 40: |
|
243 | if len(node) == 40: | |
244 | try: |
|
244 | try: | |
245 | rawnode = bin(node) |
|
245 | rawnode = bin(node) | |
246 | self.rev(rawnode) |
|
246 | self.rev(rawnode) | |
247 | return rawnode |
|
247 | return rawnode | |
248 | except TypeError: |
|
248 | except TypeError: | |
249 | pass |
|
249 | pass | |
250 |
|
250 | |||
251 | raise error.LookupError(node, self._path, _('invalid lookup input')) |
|
251 | raise error.LookupError(node, self._path, _('invalid lookup input')) | |
252 |
|
252 | |||
253 | def linkrev(self, rev): |
|
253 | def linkrev(self, rev): | |
254 | validaterev(rev) |
|
254 | validaterev(rev) | |
255 |
|
255 | |||
256 | return self._indexbyrev[rev][b'linkrev'] |
|
256 | return self._indexbyrev[rev][b'linkrev'] | |
257 |
|
257 | |||
258 | def _flags(self, rev): |
|
258 | def _flags(self, rev): | |
259 | validaterev(rev) |
|
259 | validaterev(rev) | |
260 |
|
260 | |||
261 | return self._indexbyrev[rev][b'flags'] |
|
261 | return self._indexbyrev[rev][b'flags'] | |
262 |
|
262 | |||
263 | def _candelta(self, baserev, rev): |
|
263 | def _candelta(self, baserev, rev): | |
264 | validaterev(baserev) |
|
264 | validaterev(baserev) | |
265 | validaterev(rev) |
|
265 | validaterev(rev) | |
266 |
|
266 | |||
267 | if (self._flags(baserev) & revlog.REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
267 | if (self._flags(baserev) & revlog.REVIDX_RAWTEXT_CHANGING_FLAGS) or ( | |
268 | self._flags(rev) & revlog.REVIDX_RAWTEXT_CHANGING_FLAGS |
|
268 | self._flags(rev) & revlog.REVIDX_RAWTEXT_CHANGING_FLAGS | |
269 | ): |
|
269 | ): | |
270 | return False |
|
270 | return False | |
271 |
|
271 | |||
272 | return True |
|
272 | return True | |
273 |
|
273 | |||
274 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
274 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
275 | if p1 is None and p2 is None: |
|
275 | if p1 is None and p2 is None: | |
276 | p1, p2 = self.parents(node) |
|
276 | p1, p2 = self.parents(node) | |
277 | if node != storageutil.hashrevisionsha1(text, p1, p2): |
|
277 | if node != storageutil.hashrevisionsha1(text, p1, p2): | |
278 | raise simplestoreerror( |
|
278 | raise simplestoreerror( | |
279 | _("integrity check failed on %s") % self._path |
|
279 | _("integrity check failed on %s") % self._path | |
280 | ) |
|
280 | ) | |
281 |
|
281 | |||
282 | def revision(self, nodeorrev, raw=False): |
|
282 | def revision(self, nodeorrev, raw=False): | |
283 | if isinstance(nodeorrev, int): |
|
283 | if isinstance(nodeorrev, int): | |
284 | node = self.node(nodeorrev) |
|
284 | node = self.node(nodeorrev) | |
285 | else: |
|
285 | else: | |
286 | node = nodeorrev |
|
286 | node = nodeorrev | |
287 | validatenode(node) |
|
287 | validatenode(node) | |
288 |
|
288 | |||
289 | if node == nullid: |
|
289 | if node == nullid: | |
290 | return b'' |
|
290 | return b'' | |
291 |
|
291 | |||
292 | rev = self.rev(node) |
|
292 | rev = self.rev(node) | |
293 | flags = self._flags(rev) |
|
293 | flags = self._flags(rev) | |
294 |
|
294 | |||
295 | path = b'/'.join([self._storepath, hex(node)]) |
|
295 | path = b'/'.join([self._storepath, hex(node)]) | |
296 | rawtext = self._svfs.read(path) |
|
296 | rawtext = self._svfs.read(path) | |
297 |
|
297 | |||
298 | if raw: |
|
298 | if raw: | |
299 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
299 | validatehash = flagutil.processflagsraw(self, rawtext, flags) | |
300 | text = rawtext |
|
300 | text = rawtext | |
301 | else: |
|
301 | else: | |
302 | r = flagutil.processflagsread(self, rawtext, flags) |
|
302 | r = flagutil.processflagsread(self, rawtext, flags) | |
303 | text, validatehash = r |
|
303 | text, validatehash = r | |
304 | if validatehash: |
|
304 | if validatehash: | |
305 | self.checkhash(text, node, rev=rev) |
|
305 | self.checkhash(text, node, rev=rev) | |
306 |
|
306 | |||
307 | return text |
|
307 | return text | |
308 |
|
308 | |||
309 | def rawdata(self, nodeorrev): |
|
309 | def rawdata(self, nodeorrev): | |
310 | return self.revision(raw=True) |
|
310 | return self.revision(raw=True) | |
311 |
|
311 | |||
312 | def read(self, node): |
|
312 | def read(self, node): | |
313 | validatenode(node) |
|
313 | validatenode(node) | |
314 |
|
314 | |||
315 | revision = self.revision(node) |
|
315 | revision = self.revision(node) | |
316 |
|
316 | |||
317 | if not revision.startswith(b'\1\n'): |
|
317 | if not revision.startswith(b'\1\n'): | |
318 | return revision |
|
318 | return revision | |
319 |
|
319 | |||
320 | start = revision.index(b'\1\n', 2) |
|
320 | start = revision.index(b'\1\n', 2) | |
321 | return revision[start + 2 :] |
|
321 | return revision[start + 2 :] | |
322 |
|
322 | |||
323 | def renamed(self, node): |
|
323 | def renamed(self, node): | |
324 | validatenode(node) |
|
324 | validatenode(node) | |
325 |
|
325 | |||
326 | if self.parents(node)[0] != nullid: |
|
326 | if self.parents(node)[0] != nullid: | |
327 | return False |
|
327 | return False | |
328 |
|
328 | |||
329 | fulltext = self.revision(node) |
|
329 | fulltext = self.revision(node) | |
330 | m = storageutil.parsemeta(fulltext)[0] |
|
330 | m = storageutil.parsemeta(fulltext)[0] | |
331 |
|
331 | |||
332 | if m and 'copy' in m: |
|
332 | if m and 'copy' in m: | |
333 | return m['copy'], bin(m['copyrev']) |
|
333 | return m['copy'], bin(m['copyrev']) | |
334 |
|
334 | |||
335 | return False |
|
335 | return False | |
336 |
|
336 | |||
337 | def cmp(self, node, text): |
|
337 | def cmp(self, node, text): | |
338 | validatenode(node) |
|
338 | validatenode(node) | |
339 |
|
339 | |||
340 | t = text |
|
340 | t = text | |
341 |
|
341 | |||
342 | if text.startswith(b'\1\n'): |
|
342 | if text.startswith(b'\1\n'): | |
343 | t = b'\1\n\1\n' + text |
|
343 | t = b'\1\n\1\n' + text | |
344 |
|
344 | |||
345 | p1, p2 = self.parents(node) |
|
345 | p1, p2 = self.parents(node) | |
346 |
|
346 | |||
347 | if storageutil.hashrevisionsha1(t, p1, p2) == node: |
|
347 | if storageutil.hashrevisionsha1(t, p1, p2) == node: | |
348 | return False |
|
348 | return False | |
349 |
|
349 | |||
350 | if self.iscensored(self.rev(node)): |
|
350 | if self.iscensored(self.rev(node)): | |
351 | return text != b'' |
|
351 | return text != b'' | |
352 |
|
352 | |||
353 | if self.renamed(node): |
|
353 | if self.renamed(node): | |
354 | t2 = self.read(node) |
|
354 | t2 = self.read(node) | |
355 | return t2 != text |
|
355 | return t2 != text | |
356 |
|
356 | |||
357 | return True |
|
357 | return True | |
358 |
|
358 | |||
359 | def size(self, rev): |
|
359 | def size(self, rev): | |
360 | validaterev(rev) |
|
360 | validaterev(rev) | |
361 |
|
361 | |||
362 | node = self._indexbyrev[rev][b'node'] |
|
362 | node = self._indexbyrev[rev][b'node'] | |
363 |
|
363 | |||
364 | if self.renamed(node): |
|
364 | if self.renamed(node): | |
365 | return len(self.read(node)) |
|
365 | return len(self.read(node)) | |
366 |
|
366 | |||
367 | if self.iscensored(rev): |
|
367 | if self.iscensored(rev): | |
368 | return 0 |
|
368 | return 0 | |
369 |
|
369 | |||
370 | return len(self.revision(node)) |
|
370 | return len(self.revision(node)) | |
371 |
|
371 | |||
372 | def iscensored(self, rev): |
|
372 | def iscensored(self, rev): | |
373 | validaterev(rev) |
|
373 | validaterev(rev) | |
374 |
|
374 | |||
375 | return self._flags(rev) & repository.REVISION_FLAG_CENSORED |
|
375 | return self._flags(rev) & repository.REVISION_FLAG_CENSORED | |
376 |
|
376 | |||
377 | def commonancestorsheads(self, a, b): |
|
377 | def commonancestorsheads(self, a, b): | |
378 | validatenode(a) |
|
378 | validatenode(a) | |
379 | validatenode(b) |
|
379 | validatenode(b) | |
380 |
|
380 | |||
381 | a = self.rev(a) |
|
381 | a = self.rev(a) | |
382 | b = self.rev(b) |
|
382 | b = self.rev(b) | |
383 |
|
383 | |||
384 | ancestors = ancestor.commonancestorsheads(self.parentrevs, a, b) |
|
384 | ancestors = ancestor.commonancestorsheads(self.parentrevs, a, b) | |
385 | return pycompat.maplist(self.node, ancestors) |
|
385 | return pycompat.maplist(self.node, ancestors) | |
386 |
|
386 | |||
387 | def descendants(self, revs): |
|
387 | def descendants(self, revs): | |
388 | # This is a copy of revlog.descendants() |
|
388 | # This is a copy of revlog.descendants() | |
389 | first = min(revs) |
|
389 | first = min(revs) | |
390 | if first == nullrev: |
|
390 | if first == nullrev: | |
391 | for i in self: |
|
391 | for i in self: | |
392 | yield i |
|
392 | yield i | |
393 | return |
|
393 | return | |
394 |
|
394 | |||
395 | seen = set(revs) |
|
395 | seen = set(revs) | |
396 | for i in self.revs(start=first + 1): |
|
396 | for i in self.revs(start=first + 1): | |
397 | for x in self.parentrevs(i): |
|
397 | for x in self.parentrevs(i): | |
398 | if x != nullrev and x in seen: |
|
398 | if x != nullrev and x in seen: | |
399 | seen.add(i) |
|
399 | seen.add(i) | |
400 | yield i |
|
400 | yield i | |
401 | break |
|
401 | break | |
402 |
|
402 | |||
403 | # Required by verify. |
|
403 | # Required by verify. | |
404 | def files(self): |
|
404 | def files(self): | |
405 | entries = self._svfs.listdir(self._storepath) |
|
405 | entries = self._svfs.listdir(self._storepath) | |
406 |
|
406 | |||
407 | # Strip out undo.backup.* files created as part of transaction |
|
407 | # Strip out undo.backup.* files created as part of transaction | |
408 | # recording. |
|
408 | # recording. | |
409 | entries = [f for f in entries if not f.startswith('undo.backup.')] |
|
409 | entries = [f for f in entries if not f.startswith('undo.backup.')] | |
410 |
|
410 | |||
411 | return [b'/'.join((self._storepath, f)) for f in entries] |
|
411 | return [b'/'.join((self._storepath, f)) for f in entries] | |
412 |
|
412 | |||
413 | def storageinfo( |
|
413 | def storageinfo( | |
414 | self, |
|
414 | self, | |
415 | exclusivefiles=False, |
|
415 | exclusivefiles=False, | |
416 | sharedfiles=False, |
|
416 | sharedfiles=False, | |
417 | revisionscount=False, |
|
417 | revisionscount=False, | |
418 | trackedsize=False, |
|
418 | trackedsize=False, | |
419 | storedsize=False, |
|
419 | storedsize=False, | |
420 | ): |
|
420 | ): | |
421 | # TODO do a real implementation of this |
|
421 | # TODO do a real implementation of this | |
422 | return { |
|
422 | return { | |
423 | 'exclusivefiles': [], |
|
423 | 'exclusivefiles': [], | |
424 | 'sharedfiles': [], |
|
424 | 'sharedfiles': [], | |
425 | 'revisionscount': len(self), |
|
425 | 'revisionscount': len(self), | |
426 | 'trackedsize': 0, |
|
426 | 'trackedsize': 0, | |
427 | 'storedsize': None, |
|
427 | 'storedsize': None, | |
428 | } |
|
428 | } | |
429 |
|
429 | |||
430 | def verifyintegrity(self, state): |
|
430 | def verifyintegrity(self, state): | |
431 | state['skipread'] = set() |
|
431 | state['skipread'] = set() | |
432 | for rev in self: |
|
432 | for rev in self: | |
433 | node = self.node(rev) |
|
433 | node = self.node(rev) | |
434 | try: |
|
434 | try: | |
435 | self.revision(node) |
|
435 | self.revision(node) | |
436 | except Exception as e: |
|
436 | except Exception as e: | |
437 | yield simplefilestoreproblem( |
|
437 | yield simplefilestoreproblem( | |
438 | error='unpacking %s: %s' % (node, e), node=node |
|
438 | error='unpacking %s: %s' % (node, e), node=node | |
439 | ) |
|
439 | ) | |
440 | state['skipread'].add(node) |
|
440 | state['skipread'].add(node) | |
441 |
|
441 | |||
442 | def emitrevisions( |
|
442 | def emitrevisions( | |
443 | self, |
|
443 | self, | |
444 | nodes, |
|
444 | nodes, | |
445 | nodesorder=None, |
|
445 | nodesorder=None, | |
446 | revisiondata=False, |
|
446 | revisiondata=False, | |
447 | assumehaveparentrevisions=False, |
|
447 | assumehaveparentrevisions=False, | |
448 | deltamode=repository.CG_DELTAMODE_STD, |
|
448 | deltamode=repository.CG_DELTAMODE_STD, | |
|
449 | sidedata_helpers=None, | |||
449 | ): |
|
450 | ): | |
450 | # TODO this will probably break on some ordering options. |
|
451 | # TODO this will probably break on some ordering options. | |
451 | nodes = [n for n in nodes if n != nullid] |
|
452 | nodes = [n for n in nodes if n != nullid] | |
452 | if not nodes: |
|
453 | if not nodes: | |
453 | return |
|
454 | return | |
454 | for delta in storageutil.emitrevisions( |
|
455 | for delta in storageutil.emitrevisions( | |
455 | self, |
|
456 | self, | |
456 | nodes, |
|
457 | nodes, | |
457 | nodesorder, |
|
458 | nodesorder, | |
458 | simplestorerevisiondelta, |
|
459 | simplestorerevisiondelta, | |
459 | revisiondata=revisiondata, |
|
460 | revisiondata=revisiondata, | |
460 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
461 | assumehaveparentrevisions=assumehaveparentrevisions, | |
461 | deltamode=deltamode, |
|
462 | deltamode=deltamode, | |
|
463 | sidedata_helpers=sidedata_helpers, | |||
462 | ): |
|
464 | ): | |
463 | yield delta |
|
465 | yield delta | |
464 |
|
466 | |||
465 | def add(self, text, meta, transaction, linkrev, p1, p2): |
|
467 | def add(self, text, meta, transaction, linkrev, p1, p2): | |
466 | if meta or text.startswith(b'\1\n'): |
|
468 | if meta or text.startswith(b'\1\n'): | |
467 | text = storageutil.packmeta(meta, text) |
|
469 | text = storageutil.packmeta(meta, text) | |
468 |
|
470 | |||
469 | return self.addrevision(text, transaction, linkrev, p1, p2) |
|
471 | return self.addrevision(text, transaction, linkrev, p1, p2) | |
470 |
|
472 | |||
471 | def addrevision( |
|
473 | def addrevision( | |
472 | self, |
|
474 | self, | |
473 | text, |
|
475 | text, | |
474 | transaction, |
|
476 | transaction, | |
475 | linkrev, |
|
477 | linkrev, | |
476 | p1, |
|
478 | p1, | |
477 | p2, |
|
479 | p2, | |
478 | node=None, |
|
480 | node=None, | |
479 | flags=revlog.REVIDX_DEFAULT_FLAGS, |
|
481 | flags=revlog.REVIDX_DEFAULT_FLAGS, | |
480 | cachedelta=None, |
|
482 | cachedelta=None, | |
481 | ): |
|
483 | ): | |
482 | validatenode(p1) |
|
484 | validatenode(p1) | |
483 | validatenode(p2) |
|
485 | validatenode(p2) | |
484 |
|
486 | |||
485 | if flags: |
|
487 | if flags: | |
486 | node = node or storageutil.hashrevisionsha1(text, p1, p2) |
|
488 | node = node or storageutil.hashrevisionsha1(text, p1, p2) | |
487 |
|
489 | |||
488 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) |
|
490 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) | |
489 |
|
491 | |||
490 | node = node or storageutil.hashrevisionsha1(text, p1, p2) |
|
492 | node = node or storageutil.hashrevisionsha1(text, p1, p2) | |
491 |
|
493 | |||
492 | if node in self._indexbynode: |
|
494 | if node in self._indexbynode: | |
493 | return node |
|
495 | return node | |
494 |
|
496 | |||
495 | if validatehash: |
|
497 | if validatehash: | |
496 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
498 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
497 |
|
499 | |||
498 | return self._addrawrevision( |
|
500 | return self._addrawrevision( | |
499 | node, rawtext, transaction, linkrev, p1, p2, flags |
|
501 | node, rawtext, transaction, linkrev, p1, p2, flags | |
500 | ) |
|
502 | ) | |
501 |
|
503 | |||
502 | def _addrawrevision(self, node, rawtext, transaction, link, p1, p2, flags): |
|
504 | def _addrawrevision(self, node, rawtext, transaction, link, p1, p2, flags): | |
503 | transaction.addbackup(self._indexpath) |
|
505 | transaction.addbackup(self._indexpath) | |
504 |
|
506 | |||
505 | path = b'/'.join([self._storepath, hex(node)]) |
|
507 | path = b'/'.join([self._storepath, hex(node)]) | |
506 |
|
508 | |||
507 | self._svfs.write(path, rawtext) |
|
509 | self._svfs.write(path, rawtext) | |
508 |
|
510 | |||
509 | self._indexdata.append( |
|
511 | self._indexdata.append( | |
510 | { |
|
512 | { | |
511 | b'node': node, |
|
513 | b'node': node, | |
512 | b'p1': p1, |
|
514 | b'p1': p1, | |
513 | b'p2': p2, |
|
515 | b'p2': p2, | |
514 | b'linkrev': link, |
|
516 | b'linkrev': link, | |
515 | b'flags': flags, |
|
517 | b'flags': flags, | |
516 | } |
|
518 | } | |
517 | ) |
|
519 | ) | |
518 |
|
520 | |||
519 | self._reflectindexupdate() |
|
521 | self._reflectindexupdate() | |
520 |
|
522 | |||
521 | return node |
|
523 | return node | |
522 |
|
524 | |||
523 | def _reflectindexupdate(self): |
|
525 | def _reflectindexupdate(self): | |
524 | self._refreshindex() |
|
526 | self._refreshindex() | |
525 | self._svfs.write( |
|
527 | self._svfs.write( | |
526 | self._indexpath, ''.join(cborutil.streamencode(self._indexdata)) |
|
528 | self._indexpath, ''.join(cborutil.streamencode(self._indexdata)) | |
527 | ) |
|
529 | ) | |
528 |
|
530 | |||
529 | def addgroup( |
|
531 | def addgroup( | |
530 | self, |
|
532 | self, | |
531 | deltas, |
|
533 | deltas, | |
532 | linkmapper, |
|
534 | linkmapper, | |
533 | transaction, |
|
535 | transaction, | |
534 | addrevisioncb=None, |
|
536 | addrevisioncb=None, | |
535 | duplicaterevisioncb=None, |
|
537 | duplicaterevisioncb=None, | |
536 | maybemissingparents=False, |
|
538 | maybemissingparents=False, | |
537 | ): |
|
539 | ): | |
538 | if maybemissingparents: |
|
540 | if maybemissingparents: | |
539 | raise error.Abort( |
|
541 | raise error.Abort( | |
540 | _('simple store does not support missing parents ' 'write mode') |
|
542 | _('simple store does not support missing parents ' 'write mode') | |
541 | ) |
|
543 | ) | |
542 |
|
544 | |||
543 | empty = True |
|
545 | empty = True | |
544 |
|
546 | |||
545 | transaction.addbackup(self._indexpath) |
|
547 | transaction.addbackup(self._indexpath) | |
546 |
|
548 | |||
547 | for node, p1, p2, linknode, deltabase, delta, flags in deltas: |
|
549 | for node, p1, p2, linknode, deltabase, delta, flags in deltas: | |
548 | linkrev = linkmapper(linknode) |
|
550 | linkrev = linkmapper(linknode) | |
549 | flags = flags or revlog.REVIDX_DEFAULT_FLAGS |
|
551 | flags = flags or revlog.REVIDX_DEFAULT_FLAGS | |
550 |
|
552 | |||
551 | if node in self._indexbynode: |
|
553 | if node in self._indexbynode: | |
552 | if duplicaterevisioncb: |
|
554 | if duplicaterevisioncb: | |
553 | duplicaterevisioncb(self, self.rev(node)) |
|
555 | duplicaterevisioncb(self, self.rev(node)) | |
554 | empty = False |
|
556 | empty = False | |
555 | continue |
|
557 | continue | |
556 |
|
558 | |||
557 | # Need to resolve the fulltext from the delta base. |
|
559 | # Need to resolve the fulltext from the delta base. | |
558 | if deltabase == nullid: |
|
560 | if deltabase == nullid: | |
559 | text = mdiff.patch(b'', delta) |
|
561 | text = mdiff.patch(b'', delta) | |
560 | else: |
|
562 | else: | |
561 | text = mdiff.patch(self.revision(deltabase), delta) |
|
563 | text = mdiff.patch(self.revision(deltabase), delta) | |
562 |
|
564 | |||
563 | rev = self._addrawrevision( |
|
565 | rev = self._addrawrevision( | |
564 | node, text, transaction, linkrev, p1, p2, flags |
|
566 | node, text, transaction, linkrev, p1, p2, flags | |
565 | ) |
|
567 | ) | |
566 |
|
568 | |||
567 | if addrevisioncb: |
|
569 | if addrevisioncb: | |
568 | addrevisioncb(self, rev) |
|
570 | addrevisioncb(self, rev) | |
569 | empty = False |
|
571 | empty = False | |
570 | return not empty |
|
572 | return not empty | |
571 |
|
573 | |||
572 | def _headrevs(self): |
|
574 | def _headrevs(self): | |
573 | # Assume all revisions are heads by default. |
|
575 | # Assume all revisions are heads by default. | |
574 | revishead = {rev: True for rev in self._indexbyrev} |
|
576 | revishead = {rev: True for rev in self._indexbyrev} | |
575 |
|
577 | |||
576 | for rev, entry in self._indexbyrev.items(): |
|
578 | for rev, entry in self._indexbyrev.items(): | |
577 | # Unset head flag for all seen parents. |
|
579 | # Unset head flag for all seen parents. | |
578 | revishead[self.rev(entry[b'p1'])] = False |
|
580 | revishead[self.rev(entry[b'p1'])] = False | |
579 | revishead[self.rev(entry[b'p2'])] = False |
|
581 | revishead[self.rev(entry[b'p2'])] = False | |
580 |
|
582 | |||
581 | return [rev for rev, ishead in sorted(revishead.items()) if ishead] |
|
583 | return [rev for rev, ishead in sorted(revishead.items()) if ishead] | |
582 |
|
584 | |||
583 | def heads(self, start=None, stop=None): |
|
585 | def heads(self, start=None, stop=None): | |
584 | # This is copied from revlog.py. |
|
586 | # This is copied from revlog.py. | |
585 | if start is None and stop is None: |
|
587 | if start is None and stop is None: | |
586 | if not len(self): |
|
588 | if not len(self): | |
587 | return [nullid] |
|
589 | return [nullid] | |
588 | return [self.node(r) for r in self._headrevs()] |
|
590 | return [self.node(r) for r in self._headrevs()] | |
589 |
|
591 | |||
590 | if start is None: |
|
592 | if start is None: | |
591 | start = nullid |
|
593 | start = nullid | |
592 | if stop is None: |
|
594 | if stop is None: | |
593 | stop = [] |
|
595 | stop = [] | |
594 | stoprevs = {self.rev(n) for n in stop} |
|
596 | stoprevs = {self.rev(n) for n in stop} | |
595 | startrev = self.rev(start) |
|
597 | startrev = self.rev(start) | |
596 | reachable = {startrev} |
|
598 | reachable = {startrev} | |
597 | heads = {startrev} |
|
599 | heads = {startrev} | |
598 |
|
600 | |||
599 | parentrevs = self.parentrevs |
|
601 | parentrevs = self.parentrevs | |
600 | for r in self.revs(start=startrev + 1): |
|
602 | for r in self.revs(start=startrev + 1): | |
601 | for p in parentrevs(r): |
|
603 | for p in parentrevs(r): | |
602 | if p in reachable: |
|
604 | if p in reachable: | |
603 | if r not in stoprevs: |
|
605 | if r not in stoprevs: | |
604 | reachable.add(r) |
|
606 | reachable.add(r) | |
605 | heads.add(r) |
|
607 | heads.add(r) | |
606 | if p in heads and p not in stoprevs: |
|
608 | if p in heads and p not in stoprevs: | |
607 | heads.remove(p) |
|
609 | heads.remove(p) | |
608 |
|
610 | |||
609 | return [self.node(r) for r in heads] |
|
611 | return [self.node(r) for r in heads] | |
610 |
|
612 | |||
611 | def children(self, node): |
|
613 | def children(self, node): | |
612 | validatenode(node) |
|
614 | validatenode(node) | |
613 |
|
615 | |||
614 | # This is a copy of revlog.children(). |
|
616 | # This is a copy of revlog.children(). | |
615 | c = [] |
|
617 | c = [] | |
616 | p = self.rev(node) |
|
618 | p = self.rev(node) | |
617 | for r in self.revs(start=p + 1): |
|
619 | for r in self.revs(start=p + 1): | |
618 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
620 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
619 | if prevs: |
|
621 | if prevs: | |
620 | for pr in prevs: |
|
622 | for pr in prevs: | |
621 | if pr == p: |
|
623 | if pr == p: | |
622 | c.append(self.node(r)) |
|
624 | c.append(self.node(r)) | |
623 | elif p == nullrev: |
|
625 | elif p == nullrev: | |
624 | c.append(self.node(r)) |
|
626 | c.append(self.node(r)) | |
625 | return c |
|
627 | return c | |
626 |
|
628 | |||
627 | def getstrippoint(self, minlink): |
|
629 | def getstrippoint(self, minlink): | |
628 | return storageutil.resolvestripinfo( |
|
630 | return storageutil.resolvestripinfo( | |
629 | minlink, |
|
631 | minlink, | |
630 | len(self) - 1, |
|
632 | len(self) - 1, | |
631 | self._headrevs(), |
|
633 | self._headrevs(), | |
632 | self.linkrev, |
|
634 | self.linkrev, | |
633 | self.parentrevs, |
|
635 | self.parentrevs, | |
634 | ) |
|
636 | ) | |
635 |
|
637 | |||
636 | def strip(self, minlink, transaction): |
|
638 | def strip(self, minlink, transaction): | |
637 | if not len(self): |
|
639 | if not len(self): | |
638 | return |
|
640 | return | |
639 |
|
641 | |||
640 | rev, _ignored = self.getstrippoint(minlink) |
|
642 | rev, _ignored = self.getstrippoint(minlink) | |
641 | if rev == len(self): |
|
643 | if rev == len(self): | |
642 | return |
|
644 | return | |
643 |
|
645 | |||
644 | # Purge index data starting at the requested revision. |
|
646 | # Purge index data starting at the requested revision. | |
645 | self._indexdata[rev:] = [] |
|
647 | self._indexdata[rev:] = [] | |
646 | self._reflectindexupdate() |
|
648 | self._reflectindexupdate() | |
647 |
|
649 | |||
648 |
|
650 | |||
649 | def issimplestorefile(f, kind, st): |
|
651 | def issimplestorefile(f, kind, st): | |
650 | if kind != stat.S_IFREG: |
|
652 | if kind != stat.S_IFREG: | |
651 | return False |
|
653 | return False | |
652 |
|
654 | |||
653 | if store.isrevlog(f, kind, st): |
|
655 | if store.isrevlog(f, kind, st): | |
654 | return False |
|
656 | return False | |
655 |
|
657 | |||
656 | # Ignore transaction undo files. |
|
658 | # Ignore transaction undo files. | |
657 | if f.startswith('undo.'): |
|
659 | if f.startswith('undo.'): | |
658 | return False |
|
660 | return False | |
659 |
|
661 | |||
660 | # Otherwise assume it belongs to the simple store. |
|
662 | # Otherwise assume it belongs to the simple store. | |
661 | return True |
|
663 | return True | |
662 |
|
664 | |||
663 |
|
665 | |||
664 | class simplestore(store.encodedstore): |
|
666 | class simplestore(store.encodedstore): | |
665 | def datafiles(self): |
|
667 | def datafiles(self): | |
666 | for x in super(simplestore, self).datafiles(): |
|
668 | for x in super(simplestore, self).datafiles(): | |
667 | yield x |
|
669 | yield x | |
668 |
|
670 | |||
669 | # Supplement with non-revlog files. |
|
671 | # Supplement with non-revlog files. | |
670 | extrafiles = self._walk('data', True, filefilter=issimplestorefile) |
|
672 | extrafiles = self._walk('data', True, filefilter=issimplestorefile) | |
671 |
|
673 | |||
672 | for unencoded, encoded, size in extrafiles: |
|
674 | for unencoded, encoded, size in extrafiles: | |
673 | try: |
|
675 | try: | |
674 | unencoded = store.decodefilename(unencoded) |
|
676 | unencoded = store.decodefilename(unencoded) | |
675 | except KeyError: |
|
677 | except KeyError: | |
676 | unencoded = None |
|
678 | unencoded = None | |
677 |
|
679 | |||
678 | yield unencoded, encoded, size |
|
680 | yield unencoded, encoded, size | |
679 |
|
681 | |||
680 |
|
682 | |||
681 | def reposetup(ui, repo): |
|
683 | def reposetup(ui, repo): | |
682 | if not repo.local(): |
|
684 | if not repo.local(): | |
683 | return |
|
685 | return | |
684 |
|
686 | |||
685 | if isinstance(repo, bundlerepo.bundlerepository): |
|
687 | if isinstance(repo, bundlerepo.bundlerepository): | |
686 | raise error.Abort(_('cannot use simple store with bundlerepo')) |
|
688 | raise error.Abort(_('cannot use simple store with bundlerepo')) | |
687 |
|
689 | |||
688 | class simplestorerepo(repo.__class__): |
|
690 | class simplestorerepo(repo.__class__): | |
689 | def file(self, f): |
|
691 | def file(self, f): | |
690 | return filestorage(self.svfs, f) |
|
692 | return filestorage(self.svfs, f) | |
691 |
|
693 | |||
692 | repo.__class__ = simplestorerepo |
|
694 | repo.__class__ = simplestorerepo | |
693 |
|
695 | |||
694 |
|
696 | |||
695 | def featuresetup(ui, supported): |
|
697 | def featuresetup(ui, supported): | |
696 | supported.add(REQUIREMENT) |
|
698 | supported.add(REQUIREMENT) | |
697 |
|
699 | |||
698 |
|
700 | |||
699 | def newreporequirements(orig, ui, createopts): |
|
701 | def newreporequirements(orig, ui, createopts): | |
700 | """Modifies default requirements for new repos to use the simple store.""" |
|
702 | """Modifies default requirements for new repos to use the simple store.""" | |
701 | requirements = orig(ui, createopts) |
|
703 | requirements = orig(ui, createopts) | |
702 |
|
704 | |||
703 | # These requirements are only used to affect creation of the store |
|
705 | # These requirements are only used to affect creation of the store | |
704 | # object. We have our own store. So we can remove them. |
|
706 | # object. We have our own store. So we can remove them. | |
705 | # TODO do this once we feel like taking the test hit. |
|
707 | # TODO do this once we feel like taking the test hit. | |
706 | # if 'fncache' in requirements: |
|
708 | # if 'fncache' in requirements: | |
707 | # requirements.remove('fncache') |
|
709 | # requirements.remove('fncache') | |
708 | # if 'dotencode' in requirements: |
|
710 | # if 'dotencode' in requirements: | |
709 | # requirements.remove('dotencode') |
|
711 | # requirements.remove('dotencode') | |
710 |
|
712 | |||
711 | requirements.add(REQUIREMENT) |
|
713 | requirements.add(REQUIREMENT) | |
712 |
|
714 | |||
713 | return requirements |
|
715 | return requirements | |
714 |
|
716 | |||
715 |
|
717 | |||
716 | def makestore(orig, requirements, path, vfstype): |
|
718 | def makestore(orig, requirements, path, vfstype): | |
717 | if REQUIREMENT not in requirements: |
|
719 | if REQUIREMENT not in requirements: | |
718 | return orig(requirements, path, vfstype) |
|
720 | return orig(requirements, path, vfstype) | |
719 |
|
721 | |||
720 | return simplestore(path, vfstype) |
|
722 | return simplestore(path, vfstype) | |
721 |
|
723 | |||
722 |
|
724 | |||
723 | def verifierinit(orig, self, *args, **kwargs): |
|
725 | def verifierinit(orig, self, *args, **kwargs): | |
724 | orig(self, *args, **kwargs) |
|
726 | orig(self, *args, **kwargs) | |
725 |
|
727 | |||
726 | # We don't care that files in the store don't align with what is |
|
728 | # We don't care that files in the store don't align with what is | |
727 | # advertised. So suppress these warnings. |
|
729 | # advertised. So suppress these warnings. | |
728 | self.warnorphanstorefiles = False |
|
730 | self.warnorphanstorefiles = False | |
729 |
|
731 | |||
730 |
|
732 | |||
731 | def extsetup(ui): |
|
733 | def extsetup(ui): | |
732 | localrepo.featuresetupfuncs.add(featuresetup) |
|
734 | localrepo.featuresetupfuncs.add(featuresetup) | |
733 |
|
735 | |||
734 | extensions.wrapfunction( |
|
736 | extensions.wrapfunction( | |
735 | localrepo, 'newreporequirements', newreporequirements |
|
737 | localrepo, 'newreporequirements', newreporequirements | |
736 | ) |
|
738 | ) | |
737 | extensions.wrapfunction(localrepo, 'makestore', makestore) |
|
739 | extensions.wrapfunction(localrepo, 'makestore', makestore) | |
738 | extensions.wrapfunction(verify.verifier, '__init__', verifierinit) |
|
740 | extensions.wrapfunction(verify.verifier, '__init__', verifierinit) |
General Comments 0
You need to be logged in to leave comments.
Login now