Show More
@@ -1,3364 +1,3384 b'' | |||||
1 | # revlog.py - storage back-end for mercurial |
|
1 | # revlog.py - storage back-end for mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> |
|
3 | # Copyright 2005-2007 Olivia Mackall <olivia@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 binascii |
|
16 | import binascii | |
17 | import collections |
|
17 | import collections | |
18 | import contextlib |
|
18 | import contextlib | |
19 | import errno |
|
19 | import errno | |
20 | import io |
|
20 | import io | |
21 | import os |
|
21 | import os | |
22 | import struct |
|
22 | import struct | |
23 | import zlib |
|
23 | import zlib | |
24 |
|
24 | |||
25 | # import stuff from node for others to import from revlog |
|
25 | # import stuff from node for others to import from revlog | |
26 | from .node import ( |
|
26 | from .node import ( | |
27 | bin, |
|
27 | bin, | |
28 | hex, |
|
28 | hex, | |
29 | nullrev, |
|
29 | nullrev, | |
30 | sha1nodeconstants, |
|
30 | sha1nodeconstants, | |
31 | short, |
|
31 | short, | |
32 | wdirrev, |
|
32 | wdirrev, | |
33 | ) |
|
33 | ) | |
34 | from .i18n import _ |
|
34 | from .i18n import _ | |
35 | from .pycompat import getattr |
|
35 | from .pycompat import getattr | |
36 | from .revlogutils.constants import ( |
|
36 | from .revlogutils.constants import ( | |
37 | ALL_KINDS, |
|
37 | ALL_KINDS, | |
|
38 | COMP_MODE_DEFAULT, | |||
38 | COMP_MODE_INLINE, |
|
39 | COMP_MODE_INLINE, | |
39 | COMP_MODE_PLAIN, |
|
40 | COMP_MODE_PLAIN, | |
40 | FEATURES_BY_VERSION, |
|
41 | FEATURES_BY_VERSION, | |
41 | FLAG_GENERALDELTA, |
|
42 | FLAG_GENERALDELTA, | |
42 | FLAG_INLINE_DATA, |
|
43 | FLAG_INLINE_DATA, | |
43 | INDEX_HEADER, |
|
44 | INDEX_HEADER, | |
44 | REVLOGV0, |
|
45 | REVLOGV0, | |
45 | REVLOGV1, |
|
46 | REVLOGV1, | |
46 | REVLOGV1_FLAGS, |
|
47 | REVLOGV1_FLAGS, | |
47 | REVLOGV2, |
|
48 | REVLOGV2, | |
48 | REVLOGV2_FLAGS, |
|
49 | REVLOGV2_FLAGS, | |
49 | REVLOG_DEFAULT_FLAGS, |
|
50 | REVLOG_DEFAULT_FLAGS, | |
50 | REVLOG_DEFAULT_FORMAT, |
|
51 | REVLOG_DEFAULT_FORMAT, | |
51 | REVLOG_DEFAULT_VERSION, |
|
52 | REVLOG_DEFAULT_VERSION, | |
52 | SUPPORTED_FLAGS, |
|
53 | SUPPORTED_FLAGS, | |
53 | ) |
|
54 | ) | |
54 | from .revlogutils.flagutil import ( |
|
55 | from .revlogutils.flagutil import ( | |
55 | REVIDX_DEFAULT_FLAGS, |
|
56 | REVIDX_DEFAULT_FLAGS, | |
56 | REVIDX_ELLIPSIS, |
|
57 | REVIDX_ELLIPSIS, | |
57 | REVIDX_EXTSTORED, |
|
58 | REVIDX_EXTSTORED, | |
58 | REVIDX_FLAGS_ORDER, |
|
59 | REVIDX_FLAGS_ORDER, | |
59 | REVIDX_HASCOPIESINFO, |
|
60 | REVIDX_HASCOPIESINFO, | |
60 | REVIDX_ISCENSORED, |
|
61 | REVIDX_ISCENSORED, | |
61 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
62 | REVIDX_RAWTEXT_CHANGING_FLAGS, | |
62 | ) |
|
63 | ) | |
63 | from .thirdparty import attr |
|
64 | from .thirdparty import attr | |
64 | from . import ( |
|
65 | from . import ( | |
65 | ancestor, |
|
66 | ancestor, | |
66 | dagop, |
|
67 | dagop, | |
67 | error, |
|
68 | error, | |
68 | mdiff, |
|
69 | mdiff, | |
69 | policy, |
|
70 | policy, | |
70 | pycompat, |
|
71 | pycompat, | |
71 | templatefilters, |
|
72 | templatefilters, | |
72 | util, |
|
73 | util, | |
73 | ) |
|
74 | ) | |
74 | from .interfaces import ( |
|
75 | from .interfaces import ( | |
75 | repository, |
|
76 | repository, | |
76 | util as interfaceutil, |
|
77 | util as interfaceutil, | |
77 | ) |
|
78 | ) | |
78 | from .revlogutils import ( |
|
79 | from .revlogutils import ( | |
79 | deltas as deltautil, |
|
80 | deltas as deltautil, | |
80 | docket as docketutil, |
|
81 | docket as docketutil, | |
81 | flagutil, |
|
82 | flagutil, | |
82 | nodemap as nodemaputil, |
|
83 | nodemap as nodemaputil, | |
83 | revlogv0, |
|
84 | revlogv0, | |
84 | sidedata as sidedatautil, |
|
85 | sidedata as sidedatautil, | |
85 | ) |
|
86 | ) | |
86 | from .utils import ( |
|
87 | from .utils import ( | |
87 | storageutil, |
|
88 | storageutil, | |
88 | stringutil, |
|
89 | stringutil, | |
89 | ) |
|
90 | ) | |
90 |
|
91 | |||
91 | # blanked usage of all the name to prevent pyflakes constraints |
|
92 | # blanked usage of all the name to prevent pyflakes constraints | |
92 | # We need these name available in the module for extensions. |
|
93 | # We need these name available in the module for extensions. | |
93 |
|
94 | |||
94 | REVLOGV0 |
|
95 | REVLOGV0 | |
95 | REVLOGV1 |
|
96 | REVLOGV1 | |
96 | REVLOGV2 |
|
97 | REVLOGV2 | |
97 | FLAG_INLINE_DATA |
|
98 | FLAG_INLINE_DATA | |
98 | FLAG_GENERALDELTA |
|
99 | FLAG_GENERALDELTA | |
99 | REVLOG_DEFAULT_FLAGS |
|
100 | REVLOG_DEFAULT_FLAGS | |
100 | REVLOG_DEFAULT_FORMAT |
|
101 | REVLOG_DEFAULT_FORMAT | |
101 | REVLOG_DEFAULT_VERSION |
|
102 | REVLOG_DEFAULT_VERSION | |
102 | REVLOGV1_FLAGS |
|
103 | REVLOGV1_FLAGS | |
103 | REVLOGV2_FLAGS |
|
104 | REVLOGV2_FLAGS | |
104 | REVIDX_ISCENSORED |
|
105 | REVIDX_ISCENSORED | |
105 | REVIDX_ELLIPSIS |
|
106 | REVIDX_ELLIPSIS | |
106 | REVIDX_HASCOPIESINFO |
|
107 | REVIDX_HASCOPIESINFO | |
107 | REVIDX_EXTSTORED |
|
108 | REVIDX_EXTSTORED | |
108 | REVIDX_DEFAULT_FLAGS |
|
109 | REVIDX_DEFAULT_FLAGS | |
109 | REVIDX_FLAGS_ORDER |
|
110 | REVIDX_FLAGS_ORDER | |
110 | REVIDX_RAWTEXT_CHANGING_FLAGS |
|
111 | REVIDX_RAWTEXT_CHANGING_FLAGS | |
111 |
|
112 | |||
112 | parsers = policy.importmod('parsers') |
|
113 | parsers = policy.importmod('parsers') | |
113 | rustancestor = policy.importrust('ancestor') |
|
114 | rustancestor = policy.importrust('ancestor') | |
114 | rustdagop = policy.importrust('dagop') |
|
115 | rustdagop = policy.importrust('dagop') | |
115 | rustrevlog = policy.importrust('revlog') |
|
116 | rustrevlog = policy.importrust('revlog') | |
116 |
|
117 | |||
117 | # Aliased for performance. |
|
118 | # Aliased for performance. | |
118 | _zlibdecompress = zlib.decompress |
|
119 | _zlibdecompress = zlib.decompress | |
119 |
|
120 | |||
120 | # max size of revlog with inline data |
|
121 | # max size of revlog with inline data | |
121 | _maxinline = 131072 |
|
122 | _maxinline = 131072 | |
122 | _chunksize = 1048576 |
|
123 | _chunksize = 1048576 | |
123 |
|
124 | |||
124 | # Flag processors for REVIDX_ELLIPSIS. |
|
125 | # Flag processors for REVIDX_ELLIPSIS. | |
125 | def ellipsisreadprocessor(rl, text): |
|
126 | def ellipsisreadprocessor(rl, text): | |
126 | return text, False |
|
127 | return text, False | |
127 |
|
128 | |||
128 |
|
129 | |||
129 | def ellipsiswriteprocessor(rl, text): |
|
130 | def ellipsiswriteprocessor(rl, text): | |
130 | return text, False |
|
131 | return text, False | |
131 |
|
132 | |||
132 |
|
133 | |||
133 | def ellipsisrawprocessor(rl, text): |
|
134 | def ellipsisrawprocessor(rl, text): | |
134 | return False |
|
135 | return False | |
135 |
|
136 | |||
136 |
|
137 | |||
137 | ellipsisprocessor = ( |
|
138 | ellipsisprocessor = ( | |
138 | ellipsisreadprocessor, |
|
139 | ellipsisreadprocessor, | |
139 | ellipsiswriteprocessor, |
|
140 | ellipsiswriteprocessor, | |
140 | ellipsisrawprocessor, |
|
141 | ellipsisrawprocessor, | |
141 | ) |
|
142 | ) | |
142 |
|
143 | |||
143 |
|
144 | |||
144 | def offset_type(offset, type): |
|
145 | def offset_type(offset, type): | |
145 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: |
|
146 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: | |
146 | raise ValueError(b'unknown revlog index flags') |
|
147 | raise ValueError(b'unknown revlog index flags') | |
147 | return int(int(offset) << 16 | type) |
|
148 | return int(int(offset) << 16 | type) | |
148 |
|
149 | |||
149 |
|
150 | |||
150 | def _verify_revision(rl, skipflags, state, node): |
|
151 | def _verify_revision(rl, skipflags, state, node): | |
151 | """Verify the integrity of the given revlog ``node`` while providing a hook |
|
152 | """Verify the integrity of the given revlog ``node`` while providing a hook | |
152 | point for extensions to influence the operation.""" |
|
153 | point for extensions to influence the operation.""" | |
153 | if skipflags: |
|
154 | if skipflags: | |
154 | state[b'skipread'].add(node) |
|
155 | state[b'skipread'].add(node) | |
155 | else: |
|
156 | else: | |
156 | # Side-effect: read content and verify hash. |
|
157 | # Side-effect: read content and verify hash. | |
157 | rl.revision(node) |
|
158 | rl.revision(node) | |
158 |
|
159 | |||
159 |
|
160 | |||
160 | # True if a fast implementation for persistent-nodemap is available |
|
161 | # True if a fast implementation for persistent-nodemap is available | |
161 | # |
|
162 | # | |
162 | # We also consider we have a "fast" implementation in "pure" python because |
|
163 | # We also consider we have a "fast" implementation in "pure" python because | |
163 | # people using pure don't really have performance consideration (and a |
|
164 | # people using pure don't really have performance consideration (and a | |
164 | # wheelbarrow of other slowness source) |
|
165 | # wheelbarrow of other slowness source) | |
165 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or util.safehasattr( |
|
166 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or util.safehasattr( | |
166 | parsers, 'BaseIndexObject' |
|
167 | parsers, 'BaseIndexObject' | |
167 | ) |
|
168 | ) | |
168 |
|
169 | |||
169 |
|
170 | |||
170 | @attr.s(slots=True, frozen=True) |
|
171 | @attr.s(slots=True, frozen=True) | |
171 | class _revisioninfo(object): |
|
172 | class _revisioninfo(object): | |
172 | """Information about a revision that allows building its fulltext |
|
173 | """Information about a revision that allows building its fulltext | |
173 | node: expected hash of the revision |
|
174 | node: expected hash of the revision | |
174 | p1, p2: parent revs of the revision |
|
175 | p1, p2: parent revs of the revision | |
175 | btext: built text cache consisting of a one-element list |
|
176 | btext: built text cache consisting of a one-element list | |
176 | cachedelta: (baserev, uncompressed_delta) or None |
|
177 | cachedelta: (baserev, uncompressed_delta) or None | |
177 | flags: flags associated to the revision storage |
|
178 | flags: flags associated to the revision storage | |
178 |
|
179 | |||
179 | One of btext[0] or cachedelta must be set. |
|
180 | One of btext[0] or cachedelta must be set. | |
180 | """ |
|
181 | """ | |
181 |
|
182 | |||
182 | node = attr.ib() |
|
183 | node = attr.ib() | |
183 | p1 = attr.ib() |
|
184 | p1 = attr.ib() | |
184 | p2 = attr.ib() |
|
185 | p2 = attr.ib() | |
185 | btext = attr.ib() |
|
186 | btext = attr.ib() | |
186 | textlen = attr.ib() |
|
187 | textlen = attr.ib() | |
187 | cachedelta = attr.ib() |
|
188 | cachedelta = attr.ib() | |
188 | flags = attr.ib() |
|
189 | flags = attr.ib() | |
189 |
|
190 | |||
190 |
|
191 | |||
191 | @interfaceutil.implementer(repository.irevisiondelta) |
|
192 | @interfaceutil.implementer(repository.irevisiondelta) | |
192 | @attr.s(slots=True) |
|
193 | @attr.s(slots=True) | |
193 | class revlogrevisiondelta(object): |
|
194 | class revlogrevisiondelta(object): | |
194 | node = attr.ib() |
|
195 | node = attr.ib() | |
195 | p1node = attr.ib() |
|
196 | p1node = attr.ib() | |
196 | p2node = attr.ib() |
|
197 | p2node = attr.ib() | |
197 | basenode = attr.ib() |
|
198 | basenode = attr.ib() | |
198 | flags = attr.ib() |
|
199 | flags = attr.ib() | |
199 | baserevisionsize = attr.ib() |
|
200 | baserevisionsize = attr.ib() | |
200 | revision = attr.ib() |
|
201 | revision = attr.ib() | |
201 | delta = attr.ib() |
|
202 | delta = attr.ib() | |
202 | sidedata = attr.ib() |
|
203 | sidedata = attr.ib() | |
203 | protocol_flags = attr.ib() |
|
204 | protocol_flags = attr.ib() | |
204 | linknode = attr.ib(default=None) |
|
205 | linknode = attr.ib(default=None) | |
205 |
|
206 | |||
206 |
|
207 | |||
207 | @interfaceutil.implementer(repository.iverifyproblem) |
|
208 | @interfaceutil.implementer(repository.iverifyproblem) | |
208 | @attr.s(frozen=True) |
|
209 | @attr.s(frozen=True) | |
209 | class revlogproblem(object): |
|
210 | class revlogproblem(object): | |
210 | warning = attr.ib(default=None) |
|
211 | warning = attr.ib(default=None) | |
211 | error = attr.ib(default=None) |
|
212 | error = attr.ib(default=None) | |
212 | node = attr.ib(default=None) |
|
213 | node = attr.ib(default=None) | |
213 |
|
214 | |||
214 |
|
215 | |||
215 | def parse_index_v1(data, inline): |
|
216 | def parse_index_v1(data, inline): | |
216 | # call the C implementation to parse the index data |
|
217 | # call the C implementation to parse the index data | |
217 | index, cache = parsers.parse_index2(data, inline) |
|
218 | index, cache = parsers.parse_index2(data, inline) | |
218 | return index, cache |
|
219 | return index, cache | |
219 |
|
220 | |||
220 |
|
221 | |||
221 | def parse_index_v2(data, inline): |
|
222 | def parse_index_v2(data, inline): | |
222 | # call the C implementation to parse the index data |
|
223 | # call the C implementation to parse the index data | |
223 | index, cache = parsers.parse_index2(data, inline, revlogv2=True) |
|
224 | index, cache = parsers.parse_index2(data, inline, revlogv2=True) | |
224 | return index, cache |
|
225 | return index, cache | |
225 |
|
226 | |||
226 |
|
227 | |||
227 | if util.safehasattr(parsers, 'parse_index_devel_nodemap'): |
|
228 | if util.safehasattr(parsers, 'parse_index_devel_nodemap'): | |
228 |
|
229 | |||
229 | def parse_index_v1_nodemap(data, inline): |
|
230 | def parse_index_v1_nodemap(data, inline): | |
230 | index, cache = parsers.parse_index_devel_nodemap(data, inline) |
|
231 | index, cache = parsers.parse_index_devel_nodemap(data, inline) | |
231 | return index, cache |
|
232 | return index, cache | |
232 |
|
233 | |||
233 |
|
234 | |||
234 | else: |
|
235 | else: | |
235 | parse_index_v1_nodemap = None |
|
236 | parse_index_v1_nodemap = None | |
236 |
|
237 | |||
237 |
|
238 | |||
238 | def parse_index_v1_mixed(data, inline): |
|
239 | def parse_index_v1_mixed(data, inline): | |
239 | index, cache = parse_index_v1(data, inline) |
|
240 | index, cache = parse_index_v1(data, inline) | |
240 | return rustrevlog.MixedIndex(index), cache |
|
241 | return rustrevlog.MixedIndex(index), cache | |
241 |
|
242 | |||
242 |
|
243 | |||
243 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte |
|
244 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte | |
244 | # signed integer) |
|
245 | # signed integer) | |
245 | _maxentrysize = 0x7FFFFFFF |
|
246 | _maxentrysize = 0x7FFFFFFF | |
246 |
|
247 | |||
247 |
|
248 | |||
248 | class revlog(object): |
|
249 | class revlog(object): | |
249 | """ |
|
250 | """ | |
250 | the underlying revision storage object |
|
251 | the underlying revision storage object | |
251 |
|
252 | |||
252 | A revlog consists of two parts, an index and the revision data. |
|
253 | A revlog consists of two parts, an index and the revision data. | |
253 |
|
254 | |||
254 | The index is a file with a fixed record size containing |
|
255 | The index is a file with a fixed record size containing | |
255 | information on each revision, including its nodeid (hash), the |
|
256 | information on each revision, including its nodeid (hash), the | |
256 | nodeids of its parents, the position and offset of its data within |
|
257 | nodeids of its parents, the position and offset of its data within | |
257 | the data file, and the revision it's based on. Finally, each entry |
|
258 | the data file, and the revision it's based on. Finally, each entry | |
258 | contains a linkrev entry that can serve as a pointer to external |
|
259 | contains a linkrev entry that can serve as a pointer to external | |
259 | data. |
|
260 | data. | |
260 |
|
261 | |||
261 | The revision data itself is a linear collection of data chunks. |
|
262 | The revision data itself is a linear collection of data chunks. | |
262 | Each chunk represents a revision and is usually represented as a |
|
263 | Each chunk represents a revision and is usually represented as a | |
263 | delta against the previous chunk. To bound lookup time, runs of |
|
264 | delta against the previous chunk. To bound lookup time, runs of | |
264 | deltas are limited to about 2 times the length of the original |
|
265 | deltas are limited to about 2 times the length of the original | |
265 | version data. This makes retrieval of a version proportional to |
|
266 | version data. This makes retrieval of a version proportional to | |
266 | its size, or O(1) relative to the number of revisions. |
|
267 | its size, or O(1) relative to the number of revisions. | |
267 |
|
268 | |||
268 | Both pieces of the revlog are written to in an append-only |
|
269 | Both pieces of the revlog are written to in an append-only | |
269 | fashion, which means we never need to rewrite a file to insert or |
|
270 | fashion, which means we never need to rewrite a file to insert or | |
270 | remove data, and can use some simple techniques to avoid the need |
|
271 | remove data, and can use some simple techniques to avoid the need | |
271 | for locking while reading. |
|
272 | for locking while reading. | |
272 |
|
273 | |||
273 | If checkambig, indexfile is opened with checkambig=True at |
|
274 | If checkambig, indexfile is opened with checkambig=True at | |
274 | writing, to avoid file stat ambiguity. |
|
275 | writing, to avoid file stat ambiguity. | |
275 |
|
276 | |||
276 | If mmaplargeindex is True, and an mmapindexthreshold is set, the |
|
277 | If mmaplargeindex is True, and an mmapindexthreshold is set, the | |
277 | index will be mmapped rather than read if it is larger than the |
|
278 | index will be mmapped rather than read if it is larger than the | |
278 | configured threshold. |
|
279 | configured threshold. | |
279 |
|
280 | |||
280 | If censorable is True, the revlog can have censored revisions. |
|
281 | If censorable is True, the revlog can have censored revisions. | |
281 |
|
282 | |||
282 | If `upperboundcomp` is not None, this is the expected maximal gain from |
|
283 | If `upperboundcomp` is not None, this is the expected maximal gain from | |
283 | compression for the data content. |
|
284 | compression for the data content. | |
284 |
|
285 | |||
285 | `concurrencychecker` is an optional function that receives 3 arguments: a |
|
286 | `concurrencychecker` is an optional function that receives 3 arguments: a | |
286 | file handle, a filename, and an expected position. It should check whether |
|
287 | file handle, a filename, and an expected position. It should check whether | |
287 | the current position in the file handle is valid, and log/warn/fail (by |
|
288 | the current position in the file handle is valid, and log/warn/fail (by | |
288 | raising). |
|
289 | raising). | |
289 |
|
290 | |||
290 |
|
291 | |||
291 | Internal details |
|
292 | Internal details | |
292 | ---------------- |
|
293 | ---------------- | |
293 |
|
294 | |||
294 | A large part of the revlog logic deals with revisions' "index entries", tuple |
|
295 | A large part of the revlog logic deals with revisions' "index entries", tuple | |
295 | objects that contains the same "items" whatever the revlog version. |
|
296 | objects that contains the same "items" whatever the revlog version. | |
296 | Different versions will have different ways of storing these items (sometimes |
|
297 | Different versions will have different ways of storing these items (sometimes | |
297 | not having them at all), but the tuple will always be the same. New fields |
|
298 | not having them at all), but the tuple will always be the same. New fields | |
298 | are usually added at the end to avoid breaking existing code that relies |
|
299 | are usually added at the end to avoid breaking existing code that relies | |
299 | on the existing order. The field are defined as follows: |
|
300 | on the existing order. The field are defined as follows: | |
300 |
|
301 | |||
301 | [0] offset: |
|
302 | [0] offset: | |
302 | The byte index of the start of revision data chunk. |
|
303 | The byte index of the start of revision data chunk. | |
303 | That value is shifted up by 16 bits. use "offset = field >> 16" to |
|
304 | That value is shifted up by 16 bits. use "offset = field >> 16" to | |
304 | retrieve it. |
|
305 | retrieve it. | |
305 |
|
306 | |||
306 | flags: |
|
307 | flags: | |
307 | A flag field that carries special information or changes the behavior |
|
308 | A flag field that carries special information or changes the behavior | |
308 | of the revision. (see `REVIDX_*` constants for details) |
|
309 | of the revision. (see `REVIDX_*` constants for details) | |
309 | The flag field only occupies the first 16 bits of this field, |
|
310 | The flag field only occupies the first 16 bits of this field, | |
310 | use "flags = field & 0xFFFF" to retrieve the value. |
|
311 | use "flags = field & 0xFFFF" to retrieve the value. | |
311 |
|
312 | |||
312 | [1] compressed length: |
|
313 | [1] compressed length: | |
313 | The size, in bytes, of the chunk on disk |
|
314 | The size, in bytes, of the chunk on disk | |
314 |
|
315 | |||
315 | [2] uncompressed length: |
|
316 | [2] uncompressed length: | |
316 | The size, in bytes, of the full revision once reconstructed. |
|
317 | The size, in bytes, of the full revision once reconstructed. | |
317 |
|
318 | |||
318 | [3] base rev: |
|
319 | [3] base rev: | |
319 | Either the base of the revision delta chain (without general |
|
320 | Either the base of the revision delta chain (without general | |
320 | delta), or the base of the delta (stored in the data chunk) |
|
321 | delta), or the base of the delta (stored in the data chunk) | |
321 | with general delta. |
|
322 | with general delta. | |
322 |
|
323 | |||
323 | [4] link rev: |
|
324 | [4] link rev: | |
324 | Changelog revision number of the changeset introducing this |
|
325 | Changelog revision number of the changeset introducing this | |
325 | revision. |
|
326 | revision. | |
326 |
|
327 | |||
327 | [5] parent 1 rev: |
|
328 | [5] parent 1 rev: | |
328 | Revision number of the first parent |
|
329 | Revision number of the first parent | |
329 |
|
330 | |||
330 | [6] parent 2 rev: |
|
331 | [6] parent 2 rev: | |
331 | Revision number of the second parent |
|
332 | Revision number of the second parent | |
332 |
|
333 | |||
333 | [7] node id: |
|
334 | [7] node id: | |
334 | The node id of the current revision |
|
335 | The node id of the current revision | |
335 |
|
336 | |||
336 | [8] sidedata offset: |
|
337 | [8] sidedata offset: | |
337 | The byte index of the start of the revision's side-data chunk. |
|
338 | The byte index of the start of the revision's side-data chunk. | |
338 |
|
339 | |||
339 | [9] sidedata chunk length: |
|
340 | [9] sidedata chunk length: | |
340 | The size, in bytes, of the revision's side-data chunk. |
|
341 | The size, in bytes, of the revision's side-data chunk. | |
341 |
|
342 | |||
342 | [10] data compression mode: |
|
343 | [10] data compression mode: | |
343 | two bits that detail the way the data chunk is compressed on disk. |
|
344 | two bits that detail the way the data chunk is compressed on disk. | |
344 | (see "COMP_MODE_*" constants for details). For revlog version 0 and |
|
345 | (see "COMP_MODE_*" constants for details). For revlog version 0 and | |
345 | 1 this will always be COMP_MODE_INLINE. |
|
346 | 1 this will always be COMP_MODE_INLINE. | |
346 |
|
347 | |||
347 | """ |
|
348 | """ | |
348 |
|
349 | |||
349 | _flagserrorclass = error.RevlogError |
|
350 | _flagserrorclass = error.RevlogError | |
350 |
|
351 | |||
351 | def __init__( |
|
352 | def __init__( | |
352 | self, |
|
353 | self, | |
353 | opener, |
|
354 | opener, | |
354 | target, |
|
355 | target, | |
355 | radix, |
|
356 | radix, | |
356 | postfix=None, # only exist for `tmpcensored` now |
|
357 | postfix=None, # only exist for `tmpcensored` now | |
357 | checkambig=False, |
|
358 | checkambig=False, | |
358 | mmaplargeindex=False, |
|
359 | mmaplargeindex=False, | |
359 | censorable=False, |
|
360 | censorable=False, | |
360 | upperboundcomp=None, |
|
361 | upperboundcomp=None, | |
361 | persistentnodemap=False, |
|
362 | persistentnodemap=False, | |
362 | concurrencychecker=None, |
|
363 | concurrencychecker=None, | |
363 | trypending=False, |
|
364 | trypending=False, | |
364 | ): |
|
365 | ): | |
365 | """ |
|
366 | """ | |
366 | create a revlog object |
|
367 | create a revlog object | |
367 |
|
368 | |||
368 | opener is a function that abstracts the file opening operation |
|
369 | opener is a function that abstracts the file opening operation | |
369 | and can be used to implement COW semantics or the like. |
|
370 | and can be used to implement COW semantics or the like. | |
370 |
|
371 | |||
371 | `target`: a (KIND, ID) tuple that identify the content stored in |
|
372 | `target`: a (KIND, ID) tuple that identify the content stored in | |
372 | this revlog. It help the rest of the code to understand what the revlog |
|
373 | this revlog. It help the rest of the code to understand what the revlog | |
373 | is about without having to resort to heuristic and index filename |
|
374 | is about without having to resort to heuristic and index filename | |
374 | analysis. Note: that this must be reliably be set by normal code, but |
|
375 | analysis. Note: that this must be reliably be set by normal code, but | |
375 | that test, debug, or performance measurement code might not set this to |
|
376 | that test, debug, or performance measurement code might not set this to | |
376 | accurate value. |
|
377 | accurate value. | |
377 | """ |
|
378 | """ | |
378 | self.upperboundcomp = upperboundcomp |
|
379 | self.upperboundcomp = upperboundcomp | |
379 |
|
380 | |||
380 | self.radix = radix |
|
381 | self.radix = radix | |
381 |
|
382 | |||
382 | self._docket_file = None |
|
383 | self._docket_file = None | |
383 | self._indexfile = None |
|
384 | self._indexfile = None | |
384 | self._datafile = None |
|
385 | self._datafile = None | |
385 | self._nodemap_file = None |
|
386 | self._nodemap_file = None | |
386 | self.postfix = postfix |
|
387 | self.postfix = postfix | |
387 | self._trypending = trypending |
|
388 | self._trypending = trypending | |
388 | self.opener = opener |
|
389 | self.opener = opener | |
389 | if persistentnodemap: |
|
390 | if persistentnodemap: | |
390 | self._nodemap_file = nodemaputil.get_nodemap_file(self) |
|
391 | self._nodemap_file = nodemaputil.get_nodemap_file(self) | |
391 |
|
392 | |||
392 | assert target[0] in ALL_KINDS |
|
393 | assert target[0] in ALL_KINDS | |
393 | assert len(target) == 2 |
|
394 | assert len(target) == 2 | |
394 | self.target = target |
|
395 | self.target = target | |
395 | # When True, indexfile is opened with checkambig=True at writing, to |
|
396 | # When True, indexfile is opened with checkambig=True at writing, to | |
396 | # avoid file stat ambiguity. |
|
397 | # avoid file stat ambiguity. | |
397 | self._checkambig = checkambig |
|
398 | self._checkambig = checkambig | |
398 | self._mmaplargeindex = mmaplargeindex |
|
399 | self._mmaplargeindex = mmaplargeindex | |
399 | self._censorable = censorable |
|
400 | self._censorable = censorable | |
400 | # 3-tuple of (node, rev, text) for a raw revision. |
|
401 | # 3-tuple of (node, rev, text) for a raw revision. | |
401 | self._revisioncache = None |
|
402 | self._revisioncache = None | |
402 | # Maps rev to chain base rev. |
|
403 | # Maps rev to chain base rev. | |
403 | self._chainbasecache = util.lrucachedict(100) |
|
404 | self._chainbasecache = util.lrucachedict(100) | |
404 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. |
|
405 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. | |
405 | self._chunkcache = (0, b'') |
|
406 | self._chunkcache = (0, b'') | |
406 | # How much data to read and cache into the raw revlog data cache. |
|
407 | # How much data to read and cache into the raw revlog data cache. | |
407 | self._chunkcachesize = 65536 |
|
408 | self._chunkcachesize = 65536 | |
408 | self._maxchainlen = None |
|
409 | self._maxchainlen = None | |
409 | self._deltabothparents = True |
|
410 | self._deltabothparents = True | |
410 | self.index = None |
|
411 | self.index = None | |
411 | self._docket = None |
|
412 | self._docket = None | |
412 | self._nodemap_docket = None |
|
413 | self._nodemap_docket = None | |
413 | # Mapping of partial identifiers to full nodes. |
|
414 | # Mapping of partial identifiers to full nodes. | |
414 | self._pcache = {} |
|
415 | self._pcache = {} | |
415 | # Mapping of revision integer to full node. |
|
416 | # Mapping of revision integer to full node. | |
416 | self._compengine = b'zlib' |
|
417 | self._compengine = b'zlib' | |
417 | self._compengineopts = {} |
|
418 | self._compengineopts = {} | |
418 | self._maxdeltachainspan = -1 |
|
419 | self._maxdeltachainspan = -1 | |
419 | self._withsparseread = False |
|
420 | self._withsparseread = False | |
420 | self._sparserevlog = False |
|
421 | self._sparserevlog = False | |
421 | self.hassidedata = False |
|
422 | self.hassidedata = False | |
422 | self._srdensitythreshold = 0.50 |
|
423 | self._srdensitythreshold = 0.50 | |
423 | self._srmingapsize = 262144 |
|
424 | self._srmingapsize = 262144 | |
424 |
|
425 | |||
425 | # Make copy of flag processors so each revlog instance can support |
|
426 | # Make copy of flag processors so each revlog instance can support | |
426 | # custom flags. |
|
427 | # custom flags. | |
427 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
428 | self._flagprocessors = dict(flagutil.flagprocessors) | |
428 |
|
429 | |||
429 | # 2-tuple of file handles being used for active writing. |
|
430 | # 2-tuple of file handles being used for active writing. | |
430 | self._writinghandles = None |
|
431 | self._writinghandles = None | |
431 | # prevent nesting of addgroup |
|
432 | # prevent nesting of addgroup | |
432 | self._adding_group = None |
|
433 | self._adding_group = None | |
433 |
|
434 | |||
434 | self._loadindex() |
|
435 | self._loadindex() | |
435 |
|
436 | |||
436 | self._concurrencychecker = concurrencychecker |
|
437 | self._concurrencychecker = concurrencychecker | |
437 |
|
438 | |||
438 | def _init_opts(self): |
|
439 | def _init_opts(self): | |
439 | """process options (from above/config) to setup associated default revlog mode |
|
440 | """process options (from above/config) to setup associated default revlog mode | |
440 |
|
441 | |||
441 | These values might be affected when actually reading on disk information. |
|
442 | These values might be affected when actually reading on disk information. | |
442 |
|
443 | |||
443 | The relevant values are returned for use in _loadindex(). |
|
444 | The relevant values are returned for use in _loadindex(). | |
444 |
|
445 | |||
445 | * newversionflags: |
|
446 | * newversionflags: | |
446 | version header to use if we need to create a new revlog |
|
447 | version header to use if we need to create a new revlog | |
447 |
|
448 | |||
448 | * mmapindexthreshold: |
|
449 | * mmapindexthreshold: | |
449 | minimal index size for start to use mmap |
|
450 | minimal index size for start to use mmap | |
450 |
|
451 | |||
451 | * force_nodemap: |
|
452 | * force_nodemap: | |
452 | force the usage of a "development" version of the nodemap code |
|
453 | force the usage of a "development" version of the nodemap code | |
453 | """ |
|
454 | """ | |
454 | mmapindexthreshold = None |
|
455 | mmapindexthreshold = None | |
455 | opts = self.opener.options |
|
456 | opts = self.opener.options | |
456 |
|
457 | |||
457 | if b'revlogv2' in opts: |
|
458 | if b'revlogv2' in opts: | |
458 | new_header = REVLOGV2 | FLAG_INLINE_DATA |
|
459 | new_header = REVLOGV2 | FLAG_INLINE_DATA | |
459 | elif b'revlogv1' in opts: |
|
460 | elif b'revlogv1' in opts: | |
460 | new_header = REVLOGV1 | FLAG_INLINE_DATA |
|
461 | new_header = REVLOGV1 | FLAG_INLINE_DATA | |
461 | if b'generaldelta' in opts: |
|
462 | if b'generaldelta' in opts: | |
462 | new_header |= FLAG_GENERALDELTA |
|
463 | new_header |= FLAG_GENERALDELTA | |
463 | elif b'revlogv0' in self.opener.options: |
|
464 | elif b'revlogv0' in self.opener.options: | |
464 | new_header = REVLOGV0 |
|
465 | new_header = REVLOGV0 | |
465 | else: |
|
466 | else: | |
466 | new_header = REVLOG_DEFAULT_VERSION |
|
467 | new_header = REVLOG_DEFAULT_VERSION | |
467 |
|
468 | |||
468 | if b'chunkcachesize' in opts: |
|
469 | if b'chunkcachesize' in opts: | |
469 | self._chunkcachesize = opts[b'chunkcachesize'] |
|
470 | self._chunkcachesize = opts[b'chunkcachesize'] | |
470 | if b'maxchainlen' in opts: |
|
471 | if b'maxchainlen' in opts: | |
471 | self._maxchainlen = opts[b'maxchainlen'] |
|
472 | self._maxchainlen = opts[b'maxchainlen'] | |
472 | if b'deltabothparents' in opts: |
|
473 | if b'deltabothparents' in opts: | |
473 | self._deltabothparents = opts[b'deltabothparents'] |
|
474 | self._deltabothparents = opts[b'deltabothparents'] | |
474 | self._lazydelta = bool(opts.get(b'lazydelta', True)) |
|
475 | self._lazydelta = bool(opts.get(b'lazydelta', True)) | |
475 | self._lazydeltabase = False |
|
476 | self._lazydeltabase = False | |
476 | if self._lazydelta: |
|
477 | if self._lazydelta: | |
477 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) |
|
478 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) | |
478 | if b'compengine' in opts: |
|
479 | if b'compengine' in opts: | |
479 | self._compengine = opts[b'compengine'] |
|
480 | self._compengine = opts[b'compengine'] | |
480 | if b'zlib.level' in opts: |
|
481 | if b'zlib.level' in opts: | |
481 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] |
|
482 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] | |
482 | if b'zstd.level' in opts: |
|
483 | if b'zstd.level' in opts: | |
483 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] |
|
484 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] | |
484 | if b'maxdeltachainspan' in opts: |
|
485 | if b'maxdeltachainspan' in opts: | |
485 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] |
|
486 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] | |
486 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: |
|
487 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: | |
487 | mmapindexthreshold = opts[b'mmapindexthreshold'] |
|
488 | mmapindexthreshold = opts[b'mmapindexthreshold'] | |
488 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) |
|
489 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) | |
489 | withsparseread = bool(opts.get(b'with-sparse-read', False)) |
|
490 | withsparseread = bool(opts.get(b'with-sparse-read', False)) | |
490 | # sparse-revlog forces sparse-read |
|
491 | # sparse-revlog forces sparse-read | |
491 | self._withsparseread = self._sparserevlog or withsparseread |
|
492 | self._withsparseread = self._sparserevlog or withsparseread | |
492 | if b'sparse-read-density-threshold' in opts: |
|
493 | if b'sparse-read-density-threshold' in opts: | |
493 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] |
|
494 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] | |
494 | if b'sparse-read-min-gap-size' in opts: |
|
495 | if b'sparse-read-min-gap-size' in opts: | |
495 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] |
|
496 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] | |
496 | if opts.get(b'enableellipsis'): |
|
497 | if opts.get(b'enableellipsis'): | |
497 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor |
|
498 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor | |
498 |
|
499 | |||
499 | # revlog v0 doesn't have flag processors |
|
500 | # revlog v0 doesn't have flag processors | |
500 | for flag, processor in pycompat.iteritems( |
|
501 | for flag, processor in pycompat.iteritems( | |
501 | opts.get(b'flagprocessors', {}) |
|
502 | opts.get(b'flagprocessors', {}) | |
502 | ): |
|
503 | ): | |
503 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) |
|
504 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) | |
504 |
|
505 | |||
505 | if self._chunkcachesize <= 0: |
|
506 | if self._chunkcachesize <= 0: | |
506 | raise error.RevlogError( |
|
507 | raise error.RevlogError( | |
507 | _(b'revlog chunk cache size %r is not greater than 0') |
|
508 | _(b'revlog chunk cache size %r is not greater than 0') | |
508 | % self._chunkcachesize |
|
509 | % self._chunkcachesize | |
509 | ) |
|
510 | ) | |
510 | elif self._chunkcachesize & (self._chunkcachesize - 1): |
|
511 | elif self._chunkcachesize & (self._chunkcachesize - 1): | |
511 | raise error.RevlogError( |
|
512 | raise error.RevlogError( | |
512 | _(b'revlog chunk cache size %r is not a power of 2') |
|
513 | _(b'revlog chunk cache size %r is not a power of 2') | |
513 | % self._chunkcachesize |
|
514 | % self._chunkcachesize | |
514 | ) |
|
515 | ) | |
515 | force_nodemap = opts.get(b'devel-force-nodemap', False) |
|
516 | force_nodemap = opts.get(b'devel-force-nodemap', False) | |
516 | return new_header, mmapindexthreshold, force_nodemap |
|
517 | return new_header, mmapindexthreshold, force_nodemap | |
517 |
|
518 | |||
518 | def _get_data(self, filepath, mmap_threshold, size=None): |
|
519 | def _get_data(self, filepath, mmap_threshold, size=None): | |
519 | """return a file content with or without mmap |
|
520 | """return a file content with or without mmap | |
520 |
|
521 | |||
521 | If the file is missing return the empty string""" |
|
522 | If the file is missing return the empty string""" | |
522 | try: |
|
523 | try: | |
523 | with self.opener(filepath) as fp: |
|
524 | with self.opener(filepath) as fp: | |
524 | if mmap_threshold is not None: |
|
525 | if mmap_threshold is not None: | |
525 | file_size = self.opener.fstat(fp).st_size |
|
526 | file_size = self.opener.fstat(fp).st_size | |
526 | if file_size >= mmap_threshold: |
|
527 | if file_size >= mmap_threshold: | |
527 | if size is not None: |
|
528 | if size is not None: | |
528 | # avoid potentiel mmap crash |
|
529 | # avoid potentiel mmap crash | |
529 | size = min(file_size, size) |
|
530 | size = min(file_size, size) | |
530 | # TODO: should .close() to release resources without |
|
531 | # TODO: should .close() to release resources without | |
531 | # relying on Python GC |
|
532 | # relying on Python GC | |
532 | if size is None: |
|
533 | if size is None: | |
533 | return util.buffer(util.mmapread(fp)) |
|
534 | return util.buffer(util.mmapread(fp)) | |
534 | else: |
|
535 | else: | |
535 | return util.buffer(util.mmapread(fp, size)) |
|
536 | return util.buffer(util.mmapread(fp, size)) | |
536 | if size is None: |
|
537 | if size is None: | |
537 | return fp.read() |
|
538 | return fp.read() | |
538 | else: |
|
539 | else: | |
539 | return fp.read(size) |
|
540 | return fp.read(size) | |
540 | except IOError as inst: |
|
541 | except IOError as inst: | |
541 | if inst.errno != errno.ENOENT: |
|
542 | if inst.errno != errno.ENOENT: | |
542 | raise |
|
543 | raise | |
543 | return b'' |
|
544 | return b'' | |
544 |
|
545 | |||
545 | def _loadindex(self): |
|
546 | def _loadindex(self): | |
546 |
|
547 | |||
547 | new_header, mmapindexthreshold, force_nodemap = self._init_opts() |
|
548 | new_header, mmapindexthreshold, force_nodemap = self._init_opts() | |
548 |
|
549 | |||
549 | if self.postfix is not None: |
|
550 | if self.postfix is not None: | |
550 | entry_point = b'%s.i.%s' % (self.radix, self.postfix) |
|
551 | entry_point = b'%s.i.%s' % (self.radix, self.postfix) | |
551 | elif self._trypending and self.opener.exists(b'%s.i.a' % self.radix): |
|
552 | elif self._trypending and self.opener.exists(b'%s.i.a' % self.radix): | |
552 | entry_point = b'%s.i.a' % self.radix |
|
553 | entry_point = b'%s.i.a' % self.radix | |
553 | else: |
|
554 | else: | |
554 | entry_point = b'%s.i' % self.radix |
|
555 | entry_point = b'%s.i' % self.radix | |
555 |
|
556 | |||
556 | entry_data = b'' |
|
557 | entry_data = b'' | |
557 | self._initempty = True |
|
558 | self._initempty = True | |
558 | entry_data = self._get_data(entry_point, mmapindexthreshold) |
|
559 | entry_data = self._get_data(entry_point, mmapindexthreshold) | |
559 | if len(entry_data) > 0: |
|
560 | if len(entry_data) > 0: | |
560 | header = INDEX_HEADER.unpack(entry_data[:4])[0] |
|
561 | header = INDEX_HEADER.unpack(entry_data[:4])[0] | |
561 | self._initempty = False |
|
562 | self._initempty = False | |
562 | else: |
|
563 | else: | |
563 | header = new_header |
|
564 | header = new_header | |
564 |
|
565 | |||
565 | self._format_flags = header & ~0xFFFF |
|
566 | self._format_flags = header & ~0xFFFF | |
566 | self._format_version = header & 0xFFFF |
|
567 | self._format_version = header & 0xFFFF | |
567 |
|
568 | |||
568 | supported_flags = SUPPORTED_FLAGS.get(self._format_version) |
|
569 | supported_flags = SUPPORTED_FLAGS.get(self._format_version) | |
569 | if supported_flags is None: |
|
570 | if supported_flags is None: | |
570 | msg = _(b'unknown version (%d) in revlog %s') |
|
571 | msg = _(b'unknown version (%d) in revlog %s') | |
571 | msg %= (self._format_version, self.display_id) |
|
572 | msg %= (self._format_version, self.display_id) | |
572 | raise error.RevlogError(msg) |
|
573 | raise error.RevlogError(msg) | |
573 | elif self._format_flags & ~supported_flags: |
|
574 | elif self._format_flags & ~supported_flags: | |
574 | msg = _(b'unknown flags (%#04x) in version %d revlog %s') |
|
575 | msg = _(b'unknown flags (%#04x) in version %d revlog %s') | |
575 | display_flag = self._format_flags >> 16 |
|
576 | display_flag = self._format_flags >> 16 | |
576 | msg %= (display_flag, self._format_version, self.display_id) |
|
577 | msg %= (display_flag, self._format_version, self.display_id) | |
577 | raise error.RevlogError(msg) |
|
578 | raise error.RevlogError(msg) | |
578 |
|
579 | |||
579 | features = FEATURES_BY_VERSION[self._format_version] |
|
580 | features = FEATURES_BY_VERSION[self._format_version] | |
580 | self._inline = features[b'inline'](self._format_flags) |
|
581 | self._inline = features[b'inline'](self._format_flags) | |
581 | self._generaldelta = features[b'generaldelta'](self._format_flags) |
|
582 | self._generaldelta = features[b'generaldelta'](self._format_flags) | |
582 | self.hassidedata = features[b'sidedata'] |
|
583 | self.hassidedata = features[b'sidedata'] | |
583 |
|
584 | |||
584 | if not features[b'docket']: |
|
585 | if not features[b'docket']: | |
585 | self._indexfile = entry_point |
|
586 | self._indexfile = entry_point | |
586 | index_data = entry_data |
|
587 | index_data = entry_data | |
587 | else: |
|
588 | else: | |
588 | self._docket_file = entry_point |
|
589 | self._docket_file = entry_point | |
589 | if self._initempty: |
|
590 | if self._initempty: | |
590 | self._docket = docketutil.default_docket(self, header) |
|
591 | self._docket = docketutil.default_docket(self, header) | |
591 | else: |
|
592 | else: | |
592 | self._docket = docketutil.parse_docket( |
|
593 | self._docket = docketutil.parse_docket( | |
593 | self, entry_data, use_pending=self._trypending |
|
594 | self, entry_data, use_pending=self._trypending | |
594 | ) |
|
595 | ) | |
595 | self._indexfile = self._docket.index_filepath() |
|
596 | self._indexfile = self._docket.index_filepath() | |
596 | index_data = b'' |
|
597 | index_data = b'' | |
597 | index_size = self._docket.index_end |
|
598 | index_size = self._docket.index_end | |
598 | if index_size > 0: |
|
599 | if index_size > 0: | |
599 | index_data = self._get_data( |
|
600 | index_data = self._get_data( | |
600 | self._indexfile, mmapindexthreshold, size=index_size |
|
601 | self._indexfile, mmapindexthreshold, size=index_size | |
601 | ) |
|
602 | ) | |
602 | if len(index_data) < index_size: |
|
603 | if len(index_data) < index_size: | |
603 | msg = _(b'too few index data for %s: got %d, expected %d') |
|
604 | msg = _(b'too few index data for %s: got %d, expected %d') | |
604 | msg %= (self.display_id, len(index_data), index_size) |
|
605 | msg %= (self.display_id, len(index_data), index_size) | |
605 | raise error.RevlogError(msg) |
|
606 | raise error.RevlogError(msg) | |
606 |
|
607 | |||
607 | self._inline = False |
|
608 | self._inline = False | |
608 | # generaldelta implied by version 2 revlogs. |
|
609 | # generaldelta implied by version 2 revlogs. | |
609 | self._generaldelta = True |
|
610 | self._generaldelta = True | |
610 | # the logic for persistent nodemap will be dealt with within the |
|
611 | # the logic for persistent nodemap will be dealt with within the | |
611 | # main docket, so disable it for now. |
|
612 | # main docket, so disable it for now. | |
612 | self._nodemap_file = None |
|
613 | self._nodemap_file = None | |
613 |
|
614 | |||
614 | if self.postfix is None: |
|
615 | if self.postfix is None: | |
615 | self._datafile = b'%s.d' % self.radix |
|
616 | self._datafile = b'%s.d' % self.radix | |
616 | else: |
|
617 | else: | |
617 | self._datafile = b'%s.d.%s' % (self.radix, self.postfix) |
|
618 | self._datafile = b'%s.d.%s' % (self.radix, self.postfix) | |
618 |
|
619 | |||
619 | self.nodeconstants = sha1nodeconstants |
|
620 | self.nodeconstants = sha1nodeconstants | |
620 | self.nullid = self.nodeconstants.nullid |
|
621 | self.nullid = self.nodeconstants.nullid | |
621 |
|
622 | |||
622 | # sparse-revlog can't be on without general-delta (issue6056) |
|
623 | # sparse-revlog can't be on without general-delta (issue6056) | |
623 | if not self._generaldelta: |
|
624 | if not self._generaldelta: | |
624 | self._sparserevlog = False |
|
625 | self._sparserevlog = False | |
625 |
|
626 | |||
626 | self._storedeltachains = True |
|
627 | self._storedeltachains = True | |
627 |
|
628 | |||
628 | devel_nodemap = ( |
|
629 | devel_nodemap = ( | |
629 | self._nodemap_file |
|
630 | self._nodemap_file | |
630 | and force_nodemap |
|
631 | and force_nodemap | |
631 | and parse_index_v1_nodemap is not None |
|
632 | and parse_index_v1_nodemap is not None | |
632 | ) |
|
633 | ) | |
633 |
|
634 | |||
634 | use_rust_index = False |
|
635 | use_rust_index = False | |
635 | if rustrevlog is not None: |
|
636 | if rustrevlog is not None: | |
636 | if self._nodemap_file is not None: |
|
637 | if self._nodemap_file is not None: | |
637 | use_rust_index = True |
|
638 | use_rust_index = True | |
638 | else: |
|
639 | else: | |
639 | use_rust_index = self.opener.options.get(b'rust.index') |
|
640 | use_rust_index = self.opener.options.get(b'rust.index') | |
640 |
|
641 | |||
641 | self._parse_index = parse_index_v1 |
|
642 | self._parse_index = parse_index_v1 | |
642 | if self._format_version == REVLOGV0: |
|
643 | if self._format_version == REVLOGV0: | |
643 | self._parse_index = revlogv0.parse_index_v0 |
|
644 | self._parse_index = revlogv0.parse_index_v0 | |
644 | elif self._format_version == REVLOGV2: |
|
645 | elif self._format_version == REVLOGV2: | |
645 | self._parse_index = parse_index_v2 |
|
646 | self._parse_index = parse_index_v2 | |
646 | elif devel_nodemap: |
|
647 | elif devel_nodemap: | |
647 | self._parse_index = parse_index_v1_nodemap |
|
648 | self._parse_index = parse_index_v1_nodemap | |
648 | elif use_rust_index: |
|
649 | elif use_rust_index: | |
649 | self._parse_index = parse_index_v1_mixed |
|
650 | self._parse_index = parse_index_v1_mixed | |
650 | try: |
|
651 | try: | |
651 | d = self._parse_index(index_data, self._inline) |
|
652 | d = self._parse_index(index_data, self._inline) | |
652 | index, _chunkcache = d |
|
653 | index, _chunkcache = d | |
653 | use_nodemap = ( |
|
654 | use_nodemap = ( | |
654 | not self._inline |
|
655 | not self._inline | |
655 | and self._nodemap_file is not None |
|
656 | and self._nodemap_file is not None | |
656 | and util.safehasattr(index, 'update_nodemap_data') |
|
657 | and util.safehasattr(index, 'update_nodemap_data') | |
657 | ) |
|
658 | ) | |
658 | if use_nodemap: |
|
659 | if use_nodemap: | |
659 | nodemap_data = nodemaputil.persisted_data(self) |
|
660 | nodemap_data = nodemaputil.persisted_data(self) | |
660 | if nodemap_data is not None: |
|
661 | if nodemap_data is not None: | |
661 | docket = nodemap_data[0] |
|
662 | docket = nodemap_data[0] | |
662 | if ( |
|
663 | if ( | |
663 | len(d[0]) > docket.tip_rev |
|
664 | len(d[0]) > docket.tip_rev | |
664 | and d[0][docket.tip_rev][7] == docket.tip_node |
|
665 | and d[0][docket.tip_rev][7] == docket.tip_node | |
665 | ): |
|
666 | ): | |
666 | # no changelog tampering |
|
667 | # no changelog tampering | |
667 | self._nodemap_docket = docket |
|
668 | self._nodemap_docket = docket | |
668 | index.update_nodemap_data(*nodemap_data) |
|
669 | index.update_nodemap_data(*nodemap_data) | |
669 | except (ValueError, IndexError): |
|
670 | except (ValueError, IndexError): | |
670 | raise error.RevlogError( |
|
671 | raise error.RevlogError( | |
671 | _(b"index %s is corrupted") % self.display_id |
|
672 | _(b"index %s is corrupted") % self.display_id | |
672 | ) |
|
673 | ) | |
673 | self.index, self._chunkcache = d |
|
674 | self.index, self._chunkcache = d | |
674 | if not self._chunkcache: |
|
675 | if not self._chunkcache: | |
675 | self._chunkclear() |
|
676 | self._chunkclear() | |
676 | # revnum -> (chain-length, sum-delta-length) |
|
677 | # revnum -> (chain-length, sum-delta-length) | |
677 | self._chaininfocache = util.lrucachedict(500) |
|
678 | self._chaininfocache = util.lrucachedict(500) | |
678 | # revlog header -> revlog compressor |
|
679 | # revlog header -> revlog compressor | |
679 | self._decompressors = {} |
|
680 | self._decompressors = {} | |
680 |
|
681 | |||
681 | @util.propertycache |
|
682 | @util.propertycache | |
682 | def revlog_kind(self): |
|
683 | def revlog_kind(self): | |
683 | return self.target[0] |
|
684 | return self.target[0] | |
684 |
|
685 | |||
685 | @util.propertycache |
|
686 | @util.propertycache | |
686 | def display_id(self): |
|
687 | def display_id(self): | |
687 | """The public facing "ID" of the revlog that we use in message""" |
|
688 | """The public facing "ID" of the revlog that we use in message""" | |
688 | # Maybe we should build a user facing representation of |
|
689 | # Maybe we should build a user facing representation of | |
689 | # revlog.target instead of using `self.radix` |
|
690 | # revlog.target instead of using `self.radix` | |
690 | return self.radix |
|
691 | return self.radix | |
691 |
|
692 | |||
692 | def _get_decompressor(self, t): |
|
693 | def _get_decompressor(self, t): | |
693 | try: |
|
694 | try: | |
694 | compressor = self._decompressors[t] |
|
695 | compressor = self._decompressors[t] | |
695 | except KeyError: |
|
696 | except KeyError: | |
696 | try: |
|
697 | try: | |
697 | engine = util.compengines.forrevlogheader(t) |
|
698 | engine = util.compengines.forrevlogheader(t) | |
698 | compressor = engine.revlogcompressor(self._compengineopts) |
|
699 | compressor = engine.revlogcompressor(self._compengineopts) | |
699 | self._decompressors[t] = compressor |
|
700 | self._decompressors[t] = compressor | |
700 | except KeyError: |
|
701 | except KeyError: | |
701 | raise error.RevlogError( |
|
702 | raise error.RevlogError( | |
702 | _(b'unknown compression type %s') % binascii.hexlify(t) |
|
703 | _(b'unknown compression type %s') % binascii.hexlify(t) | |
703 | ) |
|
704 | ) | |
704 | return compressor |
|
705 | return compressor | |
705 |
|
706 | |||
706 | @util.propertycache |
|
707 | @util.propertycache | |
707 | def _compressor(self): |
|
708 | def _compressor(self): | |
708 | engine = util.compengines[self._compengine] |
|
709 | engine = util.compengines[self._compengine] | |
709 | return engine.revlogcompressor(self._compengineopts) |
|
710 | return engine.revlogcompressor(self._compengineopts) | |
710 |
|
711 | |||
|
712 | @util.propertycache | |||
|
713 | def _decompressor(self): | |||
|
714 | """the default decompressor""" | |||
|
715 | if self._docket is None: | |||
|
716 | return None | |||
|
717 | t = self._docket.default_compression_header | |||
|
718 | c = self._get_decompressor(t) | |||
|
719 | return c.decompress | |||
|
720 | ||||
711 | def _indexfp(self): |
|
721 | def _indexfp(self): | |
712 | """file object for the revlog's index file""" |
|
722 | """file object for the revlog's index file""" | |
713 | return self.opener(self._indexfile, mode=b"r") |
|
723 | return self.opener(self._indexfile, mode=b"r") | |
714 |
|
724 | |||
715 | def __index_write_fp(self): |
|
725 | def __index_write_fp(self): | |
716 | # You should not use this directly and use `_writing` instead |
|
726 | # You should not use this directly and use `_writing` instead | |
717 | try: |
|
727 | try: | |
718 | f = self.opener( |
|
728 | f = self.opener( | |
719 | self._indexfile, mode=b"r+", checkambig=self._checkambig |
|
729 | self._indexfile, mode=b"r+", checkambig=self._checkambig | |
720 | ) |
|
730 | ) | |
721 | if self._docket is None: |
|
731 | if self._docket is None: | |
722 | f.seek(0, os.SEEK_END) |
|
732 | f.seek(0, os.SEEK_END) | |
723 | else: |
|
733 | else: | |
724 | f.seek(self._docket.index_end, os.SEEK_SET) |
|
734 | f.seek(self._docket.index_end, os.SEEK_SET) | |
725 | return f |
|
735 | return f | |
726 | except IOError as inst: |
|
736 | except IOError as inst: | |
727 | if inst.errno != errno.ENOENT: |
|
737 | if inst.errno != errno.ENOENT: | |
728 | raise |
|
738 | raise | |
729 | return self.opener( |
|
739 | return self.opener( | |
730 | self._indexfile, mode=b"w+", checkambig=self._checkambig |
|
740 | self._indexfile, mode=b"w+", checkambig=self._checkambig | |
731 | ) |
|
741 | ) | |
732 |
|
742 | |||
733 | def __index_new_fp(self): |
|
743 | def __index_new_fp(self): | |
734 | # You should not use this unless you are upgrading from inline revlog |
|
744 | # You should not use this unless you are upgrading from inline revlog | |
735 | return self.opener( |
|
745 | return self.opener( | |
736 | self._indexfile, |
|
746 | self._indexfile, | |
737 | mode=b"w", |
|
747 | mode=b"w", | |
738 | checkambig=self._checkambig, |
|
748 | checkambig=self._checkambig, | |
739 | atomictemp=True, |
|
749 | atomictemp=True, | |
740 | ) |
|
750 | ) | |
741 |
|
751 | |||
742 | def _datafp(self, mode=b'r'): |
|
752 | def _datafp(self, mode=b'r'): | |
743 | """file object for the revlog's data file""" |
|
753 | """file object for the revlog's data file""" | |
744 | return self.opener(self._datafile, mode=mode) |
|
754 | return self.opener(self._datafile, mode=mode) | |
745 |
|
755 | |||
746 | @contextlib.contextmanager |
|
756 | @contextlib.contextmanager | |
747 | def _datareadfp(self, existingfp=None): |
|
757 | def _datareadfp(self, existingfp=None): | |
748 | """file object suitable to read data""" |
|
758 | """file object suitable to read data""" | |
749 | # Use explicit file handle, if given. |
|
759 | # Use explicit file handle, if given. | |
750 | if existingfp is not None: |
|
760 | if existingfp is not None: | |
751 | yield existingfp |
|
761 | yield existingfp | |
752 |
|
762 | |||
753 | # Use a file handle being actively used for writes, if available. |
|
763 | # Use a file handle being actively used for writes, if available. | |
754 | # There is some danger to doing this because reads will seek the |
|
764 | # There is some danger to doing this because reads will seek the | |
755 | # file. However, _writeentry() performs a SEEK_END before all writes, |
|
765 | # file. However, _writeentry() performs a SEEK_END before all writes, | |
756 | # so we should be safe. |
|
766 | # so we should be safe. | |
757 | elif self._writinghandles: |
|
767 | elif self._writinghandles: | |
758 | if self._inline: |
|
768 | if self._inline: | |
759 | yield self._writinghandles[0] |
|
769 | yield self._writinghandles[0] | |
760 | else: |
|
770 | else: | |
761 | yield self._writinghandles[1] |
|
771 | yield self._writinghandles[1] | |
762 |
|
772 | |||
763 | # Otherwise open a new file handle. |
|
773 | # Otherwise open a new file handle. | |
764 | else: |
|
774 | else: | |
765 | if self._inline: |
|
775 | if self._inline: | |
766 | func = self._indexfp |
|
776 | func = self._indexfp | |
767 | else: |
|
777 | else: | |
768 | func = self._datafp |
|
778 | func = self._datafp | |
769 | with func() as fp: |
|
779 | with func() as fp: | |
770 | yield fp |
|
780 | yield fp | |
771 |
|
781 | |||
772 | def tiprev(self): |
|
782 | def tiprev(self): | |
773 | return len(self.index) - 1 |
|
783 | return len(self.index) - 1 | |
774 |
|
784 | |||
775 | def tip(self): |
|
785 | def tip(self): | |
776 | return self.node(self.tiprev()) |
|
786 | return self.node(self.tiprev()) | |
777 |
|
787 | |||
778 | def __contains__(self, rev): |
|
788 | def __contains__(self, rev): | |
779 | return 0 <= rev < len(self) |
|
789 | return 0 <= rev < len(self) | |
780 |
|
790 | |||
781 | def __len__(self): |
|
791 | def __len__(self): | |
782 | return len(self.index) |
|
792 | return len(self.index) | |
783 |
|
793 | |||
784 | def __iter__(self): |
|
794 | def __iter__(self): | |
785 | return iter(pycompat.xrange(len(self))) |
|
795 | return iter(pycompat.xrange(len(self))) | |
786 |
|
796 | |||
787 | def revs(self, start=0, stop=None): |
|
797 | def revs(self, start=0, stop=None): | |
788 | """iterate over all rev in this revlog (from start to stop)""" |
|
798 | """iterate over all rev in this revlog (from start to stop)""" | |
789 | return storageutil.iterrevs(len(self), start=start, stop=stop) |
|
799 | return storageutil.iterrevs(len(self), start=start, stop=stop) | |
790 |
|
800 | |||
791 | @property |
|
801 | @property | |
792 | def nodemap(self): |
|
802 | def nodemap(self): | |
793 | msg = ( |
|
803 | msg = ( | |
794 | b"revlog.nodemap is deprecated, " |
|
804 | b"revlog.nodemap is deprecated, " | |
795 | b"use revlog.index.[has_node|rev|get_rev]" |
|
805 | b"use revlog.index.[has_node|rev|get_rev]" | |
796 | ) |
|
806 | ) | |
797 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
807 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) | |
798 | return self.index.nodemap |
|
808 | return self.index.nodemap | |
799 |
|
809 | |||
800 | @property |
|
810 | @property | |
801 | def _nodecache(self): |
|
811 | def _nodecache(self): | |
802 | msg = b"revlog._nodecache is deprecated, use revlog.index.nodemap" |
|
812 | msg = b"revlog._nodecache is deprecated, use revlog.index.nodemap" | |
803 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
813 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) | |
804 | return self.index.nodemap |
|
814 | return self.index.nodemap | |
805 |
|
815 | |||
806 | def hasnode(self, node): |
|
816 | def hasnode(self, node): | |
807 | try: |
|
817 | try: | |
808 | self.rev(node) |
|
818 | self.rev(node) | |
809 | return True |
|
819 | return True | |
810 | except KeyError: |
|
820 | except KeyError: | |
811 | return False |
|
821 | return False | |
812 |
|
822 | |||
813 | def candelta(self, baserev, rev): |
|
823 | def candelta(self, baserev, rev): | |
814 | """whether two revisions (baserev, rev) can be delta-ed or not""" |
|
824 | """whether two revisions (baserev, rev) can be delta-ed or not""" | |
815 | # Disable delta if either rev requires a content-changing flag |
|
825 | # Disable delta if either rev requires a content-changing flag | |
816 | # processor (ex. LFS). This is because such flag processor can alter |
|
826 | # processor (ex. LFS). This is because such flag processor can alter | |
817 | # the rawtext content that the delta will be based on, and two clients |
|
827 | # the rawtext content that the delta will be based on, and two clients | |
818 | # could have a same revlog node with different flags (i.e. different |
|
828 | # could have a same revlog node with different flags (i.e. different | |
819 | # rawtext contents) and the delta could be incompatible. |
|
829 | # rawtext contents) and the delta could be incompatible. | |
820 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
830 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( | |
821 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS |
|
831 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS | |
822 | ): |
|
832 | ): | |
823 | return False |
|
833 | return False | |
824 | return True |
|
834 | return True | |
825 |
|
835 | |||
826 | def update_caches(self, transaction): |
|
836 | def update_caches(self, transaction): | |
827 | if self._nodemap_file is not None: |
|
837 | if self._nodemap_file is not None: | |
828 | if transaction is None: |
|
838 | if transaction is None: | |
829 | nodemaputil.update_persistent_nodemap(self) |
|
839 | nodemaputil.update_persistent_nodemap(self) | |
830 | else: |
|
840 | else: | |
831 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
841 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
832 |
|
842 | |||
833 | def clearcaches(self): |
|
843 | def clearcaches(self): | |
834 | self._revisioncache = None |
|
844 | self._revisioncache = None | |
835 | self._chainbasecache.clear() |
|
845 | self._chainbasecache.clear() | |
836 | self._chunkcache = (0, b'') |
|
846 | self._chunkcache = (0, b'') | |
837 | self._pcache = {} |
|
847 | self._pcache = {} | |
838 | self._nodemap_docket = None |
|
848 | self._nodemap_docket = None | |
839 | self.index.clearcaches() |
|
849 | self.index.clearcaches() | |
840 | # The python code is the one responsible for validating the docket, we |
|
850 | # The python code is the one responsible for validating the docket, we | |
841 | # end up having to refresh it here. |
|
851 | # end up having to refresh it here. | |
842 | use_nodemap = ( |
|
852 | use_nodemap = ( | |
843 | not self._inline |
|
853 | not self._inline | |
844 | and self._nodemap_file is not None |
|
854 | and self._nodemap_file is not None | |
845 | and util.safehasattr(self.index, 'update_nodemap_data') |
|
855 | and util.safehasattr(self.index, 'update_nodemap_data') | |
846 | ) |
|
856 | ) | |
847 | if use_nodemap: |
|
857 | if use_nodemap: | |
848 | nodemap_data = nodemaputil.persisted_data(self) |
|
858 | nodemap_data = nodemaputil.persisted_data(self) | |
849 | if nodemap_data is not None: |
|
859 | if nodemap_data is not None: | |
850 | self._nodemap_docket = nodemap_data[0] |
|
860 | self._nodemap_docket = nodemap_data[0] | |
851 | self.index.update_nodemap_data(*nodemap_data) |
|
861 | self.index.update_nodemap_data(*nodemap_data) | |
852 |
|
862 | |||
853 | def rev(self, node): |
|
863 | def rev(self, node): | |
854 | try: |
|
864 | try: | |
855 | return self.index.rev(node) |
|
865 | return self.index.rev(node) | |
856 | except TypeError: |
|
866 | except TypeError: | |
857 | raise |
|
867 | raise | |
858 | except error.RevlogError: |
|
868 | except error.RevlogError: | |
859 | # parsers.c radix tree lookup failed |
|
869 | # parsers.c radix tree lookup failed | |
860 | if ( |
|
870 | if ( | |
861 | node == self.nodeconstants.wdirid |
|
871 | node == self.nodeconstants.wdirid | |
862 | or node in self.nodeconstants.wdirfilenodeids |
|
872 | or node in self.nodeconstants.wdirfilenodeids | |
863 | ): |
|
873 | ): | |
864 | raise error.WdirUnsupported |
|
874 | raise error.WdirUnsupported | |
865 | raise error.LookupError(node, self.display_id, _(b'no node')) |
|
875 | raise error.LookupError(node, self.display_id, _(b'no node')) | |
866 |
|
876 | |||
867 | # Accessors for index entries. |
|
877 | # Accessors for index entries. | |
868 |
|
878 | |||
869 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes |
|
879 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes | |
870 | # are flags. |
|
880 | # are flags. | |
871 | def start(self, rev): |
|
881 | def start(self, rev): | |
872 | return int(self.index[rev][0] >> 16) |
|
882 | return int(self.index[rev][0] >> 16) | |
873 |
|
883 | |||
874 | def flags(self, rev): |
|
884 | def flags(self, rev): | |
875 | return self.index[rev][0] & 0xFFFF |
|
885 | return self.index[rev][0] & 0xFFFF | |
876 |
|
886 | |||
877 | def length(self, rev): |
|
887 | def length(self, rev): | |
878 | return self.index[rev][1] |
|
888 | return self.index[rev][1] | |
879 |
|
889 | |||
880 | def sidedata_length(self, rev): |
|
890 | def sidedata_length(self, rev): | |
881 | if not self.hassidedata: |
|
891 | if not self.hassidedata: | |
882 | return 0 |
|
892 | return 0 | |
883 | return self.index[rev][9] |
|
893 | return self.index[rev][9] | |
884 |
|
894 | |||
885 | def rawsize(self, rev): |
|
895 | def rawsize(self, rev): | |
886 | """return the length of the uncompressed text for a given revision""" |
|
896 | """return the length of the uncompressed text for a given revision""" | |
887 | l = self.index[rev][2] |
|
897 | l = self.index[rev][2] | |
888 | if l >= 0: |
|
898 | if l >= 0: | |
889 | return l |
|
899 | return l | |
890 |
|
900 | |||
891 | t = self.rawdata(rev) |
|
901 | t = self.rawdata(rev) | |
892 | return len(t) |
|
902 | return len(t) | |
893 |
|
903 | |||
894 | def size(self, rev): |
|
904 | def size(self, rev): | |
895 | """length of non-raw text (processed by a "read" flag processor)""" |
|
905 | """length of non-raw text (processed by a "read" flag processor)""" | |
896 | # fast path: if no "read" flag processor could change the content, |
|
906 | # fast path: if no "read" flag processor could change the content, | |
897 | # size is rawsize. note: ELLIPSIS is known to not change the content. |
|
907 | # size is rawsize. note: ELLIPSIS is known to not change the content. | |
898 | flags = self.flags(rev) |
|
908 | flags = self.flags(rev) | |
899 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: |
|
909 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: | |
900 | return self.rawsize(rev) |
|
910 | return self.rawsize(rev) | |
901 |
|
911 | |||
902 | return len(self.revision(rev, raw=False)) |
|
912 | return len(self.revision(rev, raw=False)) | |
903 |
|
913 | |||
904 | def chainbase(self, rev): |
|
914 | def chainbase(self, rev): | |
905 | base = self._chainbasecache.get(rev) |
|
915 | base = self._chainbasecache.get(rev) | |
906 | if base is not None: |
|
916 | if base is not None: | |
907 | return base |
|
917 | return base | |
908 |
|
918 | |||
909 | index = self.index |
|
919 | index = self.index | |
910 | iterrev = rev |
|
920 | iterrev = rev | |
911 | base = index[iterrev][3] |
|
921 | base = index[iterrev][3] | |
912 | while base != iterrev: |
|
922 | while base != iterrev: | |
913 | iterrev = base |
|
923 | iterrev = base | |
914 | base = index[iterrev][3] |
|
924 | base = index[iterrev][3] | |
915 |
|
925 | |||
916 | self._chainbasecache[rev] = base |
|
926 | self._chainbasecache[rev] = base | |
917 | return base |
|
927 | return base | |
918 |
|
928 | |||
919 | def linkrev(self, rev): |
|
929 | def linkrev(self, rev): | |
920 | return self.index[rev][4] |
|
930 | return self.index[rev][4] | |
921 |
|
931 | |||
922 | def parentrevs(self, rev): |
|
932 | def parentrevs(self, rev): | |
923 | try: |
|
933 | try: | |
924 | entry = self.index[rev] |
|
934 | entry = self.index[rev] | |
925 | except IndexError: |
|
935 | except IndexError: | |
926 | if rev == wdirrev: |
|
936 | if rev == wdirrev: | |
927 | raise error.WdirUnsupported |
|
937 | raise error.WdirUnsupported | |
928 | raise |
|
938 | raise | |
929 | if entry[5] == nullrev: |
|
939 | if entry[5] == nullrev: | |
930 | return entry[6], entry[5] |
|
940 | return entry[6], entry[5] | |
931 | else: |
|
941 | else: | |
932 | return entry[5], entry[6] |
|
942 | return entry[5], entry[6] | |
933 |
|
943 | |||
934 | # fast parentrevs(rev) where rev isn't filtered |
|
944 | # fast parentrevs(rev) where rev isn't filtered | |
935 | _uncheckedparentrevs = parentrevs |
|
945 | _uncheckedparentrevs = parentrevs | |
936 |
|
946 | |||
937 | def node(self, rev): |
|
947 | def node(self, rev): | |
938 | try: |
|
948 | try: | |
939 | return self.index[rev][7] |
|
949 | return self.index[rev][7] | |
940 | except IndexError: |
|
950 | except IndexError: | |
941 | if rev == wdirrev: |
|
951 | if rev == wdirrev: | |
942 | raise error.WdirUnsupported |
|
952 | raise error.WdirUnsupported | |
943 | raise |
|
953 | raise | |
944 |
|
954 | |||
945 | # Derived from index values. |
|
955 | # Derived from index values. | |
946 |
|
956 | |||
947 | def end(self, rev): |
|
957 | def end(self, rev): | |
948 | return self.start(rev) + self.length(rev) |
|
958 | return self.start(rev) + self.length(rev) | |
949 |
|
959 | |||
950 | def parents(self, node): |
|
960 | def parents(self, node): | |
951 | i = self.index |
|
961 | i = self.index | |
952 | d = i[self.rev(node)] |
|
962 | d = i[self.rev(node)] | |
953 | # inline node() to avoid function call overhead |
|
963 | # inline node() to avoid function call overhead | |
954 | if d[5] == self.nullid: |
|
964 | if d[5] == self.nullid: | |
955 | return i[d[6]][7], i[d[5]][7] |
|
965 | return i[d[6]][7], i[d[5]][7] | |
956 | else: |
|
966 | else: | |
957 | return i[d[5]][7], i[d[6]][7] |
|
967 | return i[d[5]][7], i[d[6]][7] | |
958 |
|
968 | |||
959 | def chainlen(self, rev): |
|
969 | def chainlen(self, rev): | |
960 | return self._chaininfo(rev)[0] |
|
970 | return self._chaininfo(rev)[0] | |
961 |
|
971 | |||
962 | def _chaininfo(self, rev): |
|
972 | def _chaininfo(self, rev): | |
963 | chaininfocache = self._chaininfocache |
|
973 | chaininfocache = self._chaininfocache | |
964 | if rev in chaininfocache: |
|
974 | if rev in chaininfocache: | |
965 | return chaininfocache[rev] |
|
975 | return chaininfocache[rev] | |
966 | index = self.index |
|
976 | index = self.index | |
967 | generaldelta = self._generaldelta |
|
977 | generaldelta = self._generaldelta | |
968 | iterrev = rev |
|
978 | iterrev = rev | |
969 | e = index[iterrev] |
|
979 | e = index[iterrev] | |
970 | clen = 0 |
|
980 | clen = 0 | |
971 | compresseddeltalen = 0 |
|
981 | compresseddeltalen = 0 | |
972 | while iterrev != e[3]: |
|
982 | while iterrev != e[3]: | |
973 | clen += 1 |
|
983 | clen += 1 | |
974 | compresseddeltalen += e[1] |
|
984 | compresseddeltalen += e[1] | |
975 | if generaldelta: |
|
985 | if generaldelta: | |
976 | iterrev = e[3] |
|
986 | iterrev = e[3] | |
977 | else: |
|
987 | else: | |
978 | iterrev -= 1 |
|
988 | iterrev -= 1 | |
979 | if iterrev in chaininfocache: |
|
989 | if iterrev in chaininfocache: | |
980 | t = chaininfocache[iterrev] |
|
990 | t = chaininfocache[iterrev] | |
981 | clen += t[0] |
|
991 | clen += t[0] | |
982 | compresseddeltalen += t[1] |
|
992 | compresseddeltalen += t[1] | |
983 | break |
|
993 | break | |
984 | e = index[iterrev] |
|
994 | e = index[iterrev] | |
985 | else: |
|
995 | else: | |
986 | # Add text length of base since decompressing that also takes |
|
996 | # Add text length of base since decompressing that also takes | |
987 | # work. For cache hits the length is already included. |
|
997 | # work. For cache hits the length is already included. | |
988 | compresseddeltalen += e[1] |
|
998 | compresseddeltalen += e[1] | |
989 | r = (clen, compresseddeltalen) |
|
999 | r = (clen, compresseddeltalen) | |
990 | chaininfocache[rev] = r |
|
1000 | chaininfocache[rev] = r | |
991 | return r |
|
1001 | return r | |
992 |
|
1002 | |||
993 | def _deltachain(self, rev, stoprev=None): |
|
1003 | def _deltachain(self, rev, stoprev=None): | |
994 | """Obtain the delta chain for a revision. |
|
1004 | """Obtain the delta chain for a revision. | |
995 |
|
1005 | |||
996 | ``stoprev`` specifies a revision to stop at. If not specified, we |
|
1006 | ``stoprev`` specifies a revision to stop at. If not specified, we | |
997 | stop at the base of the chain. |
|
1007 | stop at the base of the chain. | |
998 |
|
1008 | |||
999 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of |
|
1009 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of | |
1000 | revs in ascending order and ``stopped`` is a bool indicating whether |
|
1010 | revs in ascending order and ``stopped`` is a bool indicating whether | |
1001 | ``stoprev`` was hit. |
|
1011 | ``stoprev`` was hit. | |
1002 | """ |
|
1012 | """ | |
1003 | # Try C implementation. |
|
1013 | # Try C implementation. | |
1004 | try: |
|
1014 | try: | |
1005 | return self.index.deltachain(rev, stoprev, self._generaldelta) |
|
1015 | return self.index.deltachain(rev, stoprev, self._generaldelta) | |
1006 | except AttributeError: |
|
1016 | except AttributeError: | |
1007 | pass |
|
1017 | pass | |
1008 |
|
1018 | |||
1009 | chain = [] |
|
1019 | chain = [] | |
1010 |
|
1020 | |||
1011 | # Alias to prevent attribute lookup in tight loop. |
|
1021 | # Alias to prevent attribute lookup in tight loop. | |
1012 | index = self.index |
|
1022 | index = self.index | |
1013 | generaldelta = self._generaldelta |
|
1023 | generaldelta = self._generaldelta | |
1014 |
|
1024 | |||
1015 | iterrev = rev |
|
1025 | iterrev = rev | |
1016 | e = index[iterrev] |
|
1026 | e = index[iterrev] | |
1017 | while iterrev != e[3] and iterrev != stoprev: |
|
1027 | while iterrev != e[3] and iterrev != stoprev: | |
1018 | chain.append(iterrev) |
|
1028 | chain.append(iterrev) | |
1019 | if generaldelta: |
|
1029 | if generaldelta: | |
1020 | iterrev = e[3] |
|
1030 | iterrev = e[3] | |
1021 | else: |
|
1031 | else: | |
1022 | iterrev -= 1 |
|
1032 | iterrev -= 1 | |
1023 | e = index[iterrev] |
|
1033 | e = index[iterrev] | |
1024 |
|
1034 | |||
1025 | if iterrev == stoprev: |
|
1035 | if iterrev == stoprev: | |
1026 | stopped = True |
|
1036 | stopped = True | |
1027 | else: |
|
1037 | else: | |
1028 | chain.append(iterrev) |
|
1038 | chain.append(iterrev) | |
1029 | stopped = False |
|
1039 | stopped = False | |
1030 |
|
1040 | |||
1031 | chain.reverse() |
|
1041 | chain.reverse() | |
1032 | return chain, stopped |
|
1042 | return chain, stopped | |
1033 |
|
1043 | |||
1034 | def ancestors(self, revs, stoprev=0, inclusive=False): |
|
1044 | def ancestors(self, revs, stoprev=0, inclusive=False): | |
1035 | """Generate the ancestors of 'revs' in reverse revision order. |
|
1045 | """Generate the ancestors of 'revs' in reverse revision order. | |
1036 | Does not generate revs lower than stoprev. |
|
1046 | Does not generate revs lower than stoprev. | |
1037 |
|
1047 | |||
1038 | See the documentation for ancestor.lazyancestors for more details.""" |
|
1048 | See the documentation for ancestor.lazyancestors for more details.""" | |
1039 |
|
1049 | |||
1040 | # first, make sure start revisions aren't filtered |
|
1050 | # first, make sure start revisions aren't filtered | |
1041 | revs = list(revs) |
|
1051 | revs = list(revs) | |
1042 | checkrev = self.node |
|
1052 | checkrev = self.node | |
1043 | for r in revs: |
|
1053 | for r in revs: | |
1044 | checkrev(r) |
|
1054 | checkrev(r) | |
1045 | # and we're sure ancestors aren't filtered as well |
|
1055 | # and we're sure ancestors aren't filtered as well | |
1046 |
|
1056 | |||
1047 | if rustancestor is not None: |
|
1057 | if rustancestor is not None: | |
1048 | lazyancestors = rustancestor.LazyAncestors |
|
1058 | lazyancestors = rustancestor.LazyAncestors | |
1049 | arg = self.index |
|
1059 | arg = self.index | |
1050 | else: |
|
1060 | else: | |
1051 | lazyancestors = ancestor.lazyancestors |
|
1061 | lazyancestors = ancestor.lazyancestors | |
1052 | arg = self._uncheckedparentrevs |
|
1062 | arg = self._uncheckedparentrevs | |
1053 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) |
|
1063 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) | |
1054 |
|
1064 | |||
1055 | def descendants(self, revs): |
|
1065 | def descendants(self, revs): | |
1056 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
1066 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) | |
1057 |
|
1067 | |||
1058 | def findcommonmissing(self, common=None, heads=None): |
|
1068 | def findcommonmissing(self, common=None, heads=None): | |
1059 | """Return a tuple of the ancestors of common and the ancestors of heads |
|
1069 | """Return a tuple of the ancestors of common and the ancestors of heads | |
1060 | that are not ancestors of common. In revset terminology, we return the |
|
1070 | that are not ancestors of common. In revset terminology, we return the | |
1061 | tuple: |
|
1071 | tuple: | |
1062 |
|
1072 | |||
1063 | ::common, (::heads) - (::common) |
|
1073 | ::common, (::heads) - (::common) | |
1064 |
|
1074 | |||
1065 | The list is sorted by revision number, meaning it is |
|
1075 | The list is sorted by revision number, meaning it is | |
1066 | topologically sorted. |
|
1076 | topologically sorted. | |
1067 |
|
1077 | |||
1068 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1078 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1069 | not supplied, uses all of the revlog's heads. If common is not |
|
1079 | not supplied, uses all of the revlog's heads. If common is not | |
1070 | supplied, uses nullid.""" |
|
1080 | supplied, uses nullid.""" | |
1071 | if common is None: |
|
1081 | if common is None: | |
1072 | common = [self.nullid] |
|
1082 | common = [self.nullid] | |
1073 | if heads is None: |
|
1083 | if heads is None: | |
1074 | heads = self.heads() |
|
1084 | heads = self.heads() | |
1075 |
|
1085 | |||
1076 | common = [self.rev(n) for n in common] |
|
1086 | common = [self.rev(n) for n in common] | |
1077 | heads = [self.rev(n) for n in heads] |
|
1087 | heads = [self.rev(n) for n in heads] | |
1078 |
|
1088 | |||
1079 | # we want the ancestors, but inclusive |
|
1089 | # we want the ancestors, but inclusive | |
1080 | class lazyset(object): |
|
1090 | class lazyset(object): | |
1081 | def __init__(self, lazyvalues): |
|
1091 | def __init__(self, lazyvalues): | |
1082 | self.addedvalues = set() |
|
1092 | self.addedvalues = set() | |
1083 | self.lazyvalues = lazyvalues |
|
1093 | self.lazyvalues = lazyvalues | |
1084 |
|
1094 | |||
1085 | def __contains__(self, value): |
|
1095 | def __contains__(self, value): | |
1086 | return value in self.addedvalues or value in self.lazyvalues |
|
1096 | return value in self.addedvalues or value in self.lazyvalues | |
1087 |
|
1097 | |||
1088 | def __iter__(self): |
|
1098 | def __iter__(self): | |
1089 | added = self.addedvalues |
|
1099 | added = self.addedvalues | |
1090 | for r in added: |
|
1100 | for r in added: | |
1091 | yield r |
|
1101 | yield r | |
1092 | for r in self.lazyvalues: |
|
1102 | for r in self.lazyvalues: | |
1093 | if not r in added: |
|
1103 | if not r in added: | |
1094 | yield r |
|
1104 | yield r | |
1095 |
|
1105 | |||
1096 | def add(self, value): |
|
1106 | def add(self, value): | |
1097 | self.addedvalues.add(value) |
|
1107 | self.addedvalues.add(value) | |
1098 |
|
1108 | |||
1099 | def update(self, values): |
|
1109 | def update(self, values): | |
1100 | self.addedvalues.update(values) |
|
1110 | self.addedvalues.update(values) | |
1101 |
|
1111 | |||
1102 | has = lazyset(self.ancestors(common)) |
|
1112 | has = lazyset(self.ancestors(common)) | |
1103 | has.add(nullrev) |
|
1113 | has.add(nullrev) | |
1104 | has.update(common) |
|
1114 | has.update(common) | |
1105 |
|
1115 | |||
1106 | # take all ancestors from heads that aren't in has |
|
1116 | # take all ancestors from heads that aren't in has | |
1107 | missing = set() |
|
1117 | missing = set() | |
1108 | visit = collections.deque(r for r in heads if r not in has) |
|
1118 | visit = collections.deque(r for r in heads if r not in has) | |
1109 | while visit: |
|
1119 | while visit: | |
1110 | r = visit.popleft() |
|
1120 | r = visit.popleft() | |
1111 | if r in missing: |
|
1121 | if r in missing: | |
1112 | continue |
|
1122 | continue | |
1113 | else: |
|
1123 | else: | |
1114 | missing.add(r) |
|
1124 | missing.add(r) | |
1115 | for p in self.parentrevs(r): |
|
1125 | for p in self.parentrevs(r): | |
1116 | if p not in has: |
|
1126 | if p not in has: | |
1117 | visit.append(p) |
|
1127 | visit.append(p) | |
1118 | missing = list(missing) |
|
1128 | missing = list(missing) | |
1119 | missing.sort() |
|
1129 | missing.sort() | |
1120 | return has, [self.node(miss) for miss in missing] |
|
1130 | return has, [self.node(miss) for miss in missing] | |
1121 |
|
1131 | |||
1122 | def incrementalmissingrevs(self, common=None): |
|
1132 | def incrementalmissingrevs(self, common=None): | |
1123 | """Return an object that can be used to incrementally compute the |
|
1133 | """Return an object that can be used to incrementally compute the | |
1124 | revision numbers of the ancestors of arbitrary sets that are not |
|
1134 | revision numbers of the ancestors of arbitrary sets that are not | |
1125 | ancestors of common. This is an ancestor.incrementalmissingancestors |
|
1135 | ancestors of common. This is an ancestor.incrementalmissingancestors | |
1126 | object. |
|
1136 | object. | |
1127 |
|
1137 | |||
1128 | 'common' is a list of revision numbers. If common is not supplied, uses |
|
1138 | 'common' is a list of revision numbers. If common is not supplied, uses | |
1129 | nullrev. |
|
1139 | nullrev. | |
1130 | """ |
|
1140 | """ | |
1131 | if common is None: |
|
1141 | if common is None: | |
1132 | common = [nullrev] |
|
1142 | common = [nullrev] | |
1133 |
|
1143 | |||
1134 | if rustancestor is not None: |
|
1144 | if rustancestor is not None: | |
1135 | return rustancestor.MissingAncestors(self.index, common) |
|
1145 | return rustancestor.MissingAncestors(self.index, common) | |
1136 | return ancestor.incrementalmissingancestors(self.parentrevs, common) |
|
1146 | return ancestor.incrementalmissingancestors(self.parentrevs, common) | |
1137 |
|
1147 | |||
1138 | def findmissingrevs(self, common=None, heads=None): |
|
1148 | def findmissingrevs(self, common=None, heads=None): | |
1139 | """Return the revision numbers of the ancestors of heads that |
|
1149 | """Return the revision numbers of the ancestors of heads that | |
1140 | are not ancestors of common. |
|
1150 | are not ancestors of common. | |
1141 |
|
1151 | |||
1142 | More specifically, return a list of revision numbers corresponding to |
|
1152 | More specifically, return a list of revision numbers corresponding to | |
1143 | nodes N such that every N satisfies the following constraints: |
|
1153 | nodes N such that every N satisfies the following constraints: | |
1144 |
|
1154 | |||
1145 | 1. N is an ancestor of some node in 'heads' |
|
1155 | 1. N is an ancestor of some node in 'heads' | |
1146 | 2. N is not an ancestor of any node in 'common' |
|
1156 | 2. N is not an ancestor of any node in 'common' | |
1147 |
|
1157 | |||
1148 | The list is sorted by revision number, meaning it is |
|
1158 | The list is sorted by revision number, meaning it is | |
1149 | topologically sorted. |
|
1159 | topologically sorted. | |
1150 |
|
1160 | |||
1151 | 'heads' and 'common' are both lists of revision numbers. If heads is |
|
1161 | 'heads' and 'common' are both lists of revision numbers. If heads is | |
1152 | not supplied, uses all of the revlog's heads. If common is not |
|
1162 | not supplied, uses all of the revlog's heads. If common is not | |
1153 | supplied, uses nullid.""" |
|
1163 | supplied, uses nullid.""" | |
1154 | if common is None: |
|
1164 | if common is None: | |
1155 | common = [nullrev] |
|
1165 | common = [nullrev] | |
1156 | if heads is None: |
|
1166 | if heads is None: | |
1157 | heads = self.headrevs() |
|
1167 | heads = self.headrevs() | |
1158 |
|
1168 | |||
1159 | inc = self.incrementalmissingrevs(common=common) |
|
1169 | inc = self.incrementalmissingrevs(common=common) | |
1160 | return inc.missingancestors(heads) |
|
1170 | return inc.missingancestors(heads) | |
1161 |
|
1171 | |||
1162 | def findmissing(self, common=None, heads=None): |
|
1172 | def findmissing(self, common=None, heads=None): | |
1163 | """Return the ancestors of heads that are not ancestors of common. |
|
1173 | """Return the ancestors of heads that are not ancestors of common. | |
1164 |
|
1174 | |||
1165 | More specifically, return a list of nodes N such that every N |
|
1175 | More specifically, return a list of nodes N such that every N | |
1166 | satisfies the following constraints: |
|
1176 | satisfies the following constraints: | |
1167 |
|
1177 | |||
1168 | 1. N is an ancestor of some node in 'heads' |
|
1178 | 1. N is an ancestor of some node in 'heads' | |
1169 | 2. N is not an ancestor of any node in 'common' |
|
1179 | 2. N is not an ancestor of any node in 'common' | |
1170 |
|
1180 | |||
1171 | The list is sorted by revision number, meaning it is |
|
1181 | The list is sorted by revision number, meaning it is | |
1172 | topologically sorted. |
|
1182 | topologically sorted. | |
1173 |
|
1183 | |||
1174 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1184 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1175 | not supplied, uses all of the revlog's heads. If common is not |
|
1185 | not supplied, uses all of the revlog's heads. If common is not | |
1176 | supplied, uses nullid.""" |
|
1186 | supplied, uses nullid.""" | |
1177 | if common is None: |
|
1187 | if common is None: | |
1178 | common = [self.nullid] |
|
1188 | common = [self.nullid] | |
1179 | if heads is None: |
|
1189 | if heads is None: | |
1180 | heads = self.heads() |
|
1190 | heads = self.heads() | |
1181 |
|
1191 | |||
1182 | common = [self.rev(n) for n in common] |
|
1192 | common = [self.rev(n) for n in common] | |
1183 | heads = [self.rev(n) for n in heads] |
|
1193 | heads = [self.rev(n) for n in heads] | |
1184 |
|
1194 | |||
1185 | inc = self.incrementalmissingrevs(common=common) |
|
1195 | inc = self.incrementalmissingrevs(common=common) | |
1186 | return [self.node(r) for r in inc.missingancestors(heads)] |
|
1196 | return [self.node(r) for r in inc.missingancestors(heads)] | |
1187 |
|
1197 | |||
1188 | def nodesbetween(self, roots=None, heads=None): |
|
1198 | def nodesbetween(self, roots=None, heads=None): | |
1189 | """Return a topological path from 'roots' to 'heads'. |
|
1199 | """Return a topological path from 'roots' to 'heads'. | |
1190 |
|
1200 | |||
1191 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a |
|
1201 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a | |
1192 | topologically sorted list of all nodes N that satisfy both of |
|
1202 | topologically sorted list of all nodes N that satisfy both of | |
1193 | these constraints: |
|
1203 | these constraints: | |
1194 |
|
1204 | |||
1195 | 1. N is a descendant of some node in 'roots' |
|
1205 | 1. N is a descendant of some node in 'roots' | |
1196 | 2. N is an ancestor of some node in 'heads' |
|
1206 | 2. N is an ancestor of some node in 'heads' | |
1197 |
|
1207 | |||
1198 | Every node is considered to be both a descendant and an ancestor |
|
1208 | Every node is considered to be both a descendant and an ancestor | |
1199 | of itself, so every reachable node in 'roots' and 'heads' will be |
|
1209 | of itself, so every reachable node in 'roots' and 'heads' will be | |
1200 | included in 'nodes'. |
|
1210 | included in 'nodes'. | |
1201 |
|
1211 | |||
1202 | 'outroots' is the list of reachable nodes in 'roots', i.e., the |
|
1212 | 'outroots' is the list of reachable nodes in 'roots', i.e., the | |
1203 | subset of 'roots' that is returned in 'nodes'. Likewise, |
|
1213 | subset of 'roots' that is returned in 'nodes'. Likewise, | |
1204 | 'outheads' is the subset of 'heads' that is also in 'nodes'. |
|
1214 | 'outheads' is the subset of 'heads' that is also in 'nodes'. | |
1205 |
|
1215 | |||
1206 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is |
|
1216 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is | |
1207 | unspecified, uses nullid as the only root. If 'heads' is |
|
1217 | unspecified, uses nullid as the only root. If 'heads' is | |
1208 | unspecified, uses list of all of the revlog's heads.""" |
|
1218 | unspecified, uses list of all of the revlog's heads.""" | |
1209 | nonodes = ([], [], []) |
|
1219 | nonodes = ([], [], []) | |
1210 | if roots is not None: |
|
1220 | if roots is not None: | |
1211 | roots = list(roots) |
|
1221 | roots = list(roots) | |
1212 | if not roots: |
|
1222 | if not roots: | |
1213 | return nonodes |
|
1223 | return nonodes | |
1214 | lowestrev = min([self.rev(n) for n in roots]) |
|
1224 | lowestrev = min([self.rev(n) for n in roots]) | |
1215 | else: |
|
1225 | else: | |
1216 | roots = [self.nullid] # Everybody's a descendant of nullid |
|
1226 | roots = [self.nullid] # Everybody's a descendant of nullid | |
1217 | lowestrev = nullrev |
|
1227 | lowestrev = nullrev | |
1218 | if (lowestrev == nullrev) and (heads is None): |
|
1228 | if (lowestrev == nullrev) and (heads is None): | |
1219 | # We want _all_ the nodes! |
|
1229 | # We want _all_ the nodes! | |
1220 | return ( |
|
1230 | return ( | |
1221 | [self.node(r) for r in self], |
|
1231 | [self.node(r) for r in self], | |
1222 | [self.nullid], |
|
1232 | [self.nullid], | |
1223 | list(self.heads()), |
|
1233 | list(self.heads()), | |
1224 | ) |
|
1234 | ) | |
1225 | if heads is None: |
|
1235 | if heads is None: | |
1226 | # All nodes are ancestors, so the latest ancestor is the last |
|
1236 | # All nodes are ancestors, so the latest ancestor is the last | |
1227 | # node. |
|
1237 | # node. | |
1228 | highestrev = len(self) - 1 |
|
1238 | highestrev = len(self) - 1 | |
1229 | # Set ancestors to None to signal that every node is an ancestor. |
|
1239 | # Set ancestors to None to signal that every node is an ancestor. | |
1230 | ancestors = None |
|
1240 | ancestors = None | |
1231 | # Set heads to an empty dictionary for later discovery of heads |
|
1241 | # Set heads to an empty dictionary for later discovery of heads | |
1232 | heads = {} |
|
1242 | heads = {} | |
1233 | else: |
|
1243 | else: | |
1234 | heads = list(heads) |
|
1244 | heads = list(heads) | |
1235 | if not heads: |
|
1245 | if not heads: | |
1236 | return nonodes |
|
1246 | return nonodes | |
1237 | ancestors = set() |
|
1247 | ancestors = set() | |
1238 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
1248 | # Turn heads into a dictionary so we can remove 'fake' heads. | |
1239 | # Also, later we will be using it to filter out the heads we can't |
|
1249 | # Also, later we will be using it to filter out the heads we can't | |
1240 | # find from roots. |
|
1250 | # find from roots. | |
1241 | heads = dict.fromkeys(heads, False) |
|
1251 | heads = dict.fromkeys(heads, False) | |
1242 | # Start at the top and keep marking parents until we're done. |
|
1252 | # Start at the top and keep marking parents until we're done. | |
1243 | nodestotag = set(heads) |
|
1253 | nodestotag = set(heads) | |
1244 | # Remember where the top was so we can use it as a limit later. |
|
1254 | # Remember where the top was so we can use it as a limit later. | |
1245 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
1255 | highestrev = max([self.rev(n) for n in nodestotag]) | |
1246 | while nodestotag: |
|
1256 | while nodestotag: | |
1247 | # grab a node to tag |
|
1257 | # grab a node to tag | |
1248 | n = nodestotag.pop() |
|
1258 | n = nodestotag.pop() | |
1249 | # Never tag nullid |
|
1259 | # Never tag nullid | |
1250 | if n == self.nullid: |
|
1260 | if n == self.nullid: | |
1251 | continue |
|
1261 | continue | |
1252 | # A node's revision number represents its place in a |
|
1262 | # A node's revision number represents its place in a | |
1253 | # topologically sorted list of nodes. |
|
1263 | # topologically sorted list of nodes. | |
1254 | r = self.rev(n) |
|
1264 | r = self.rev(n) | |
1255 | if r >= lowestrev: |
|
1265 | if r >= lowestrev: | |
1256 | if n not in ancestors: |
|
1266 | if n not in ancestors: | |
1257 | # If we are possibly a descendant of one of the roots |
|
1267 | # If we are possibly a descendant of one of the roots | |
1258 | # and we haven't already been marked as an ancestor |
|
1268 | # and we haven't already been marked as an ancestor | |
1259 | ancestors.add(n) # Mark as ancestor |
|
1269 | ancestors.add(n) # Mark as ancestor | |
1260 | # Add non-nullid parents to list of nodes to tag. |
|
1270 | # Add non-nullid parents to list of nodes to tag. | |
1261 | nodestotag.update( |
|
1271 | nodestotag.update( | |
1262 | [p for p in self.parents(n) if p != self.nullid] |
|
1272 | [p for p in self.parents(n) if p != self.nullid] | |
1263 | ) |
|
1273 | ) | |
1264 | elif n in heads: # We've seen it before, is it a fake head? |
|
1274 | elif n in heads: # We've seen it before, is it a fake head? | |
1265 | # So it is, real heads should not be the ancestors of |
|
1275 | # So it is, real heads should not be the ancestors of | |
1266 | # any other heads. |
|
1276 | # any other heads. | |
1267 | heads.pop(n) |
|
1277 | heads.pop(n) | |
1268 | if not ancestors: |
|
1278 | if not ancestors: | |
1269 | return nonodes |
|
1279 | return nonodes | |
1270 | # Now that we have our set of ancestors, we want to remove any |
|
1280 | # Now that we have our set of ancestors, we want to remove any | |
1271 | # roots that are not ancestors. |
|
1281 | # roots that are not ancestors. | |
1272 |
|
1282 | |||
1273 | # If one of the roots was nullid, everything is included anyway. |
|
1283 | # If one of the roots was nullid, everything is included anyway. | |
1274 | if lowestrev > nullrev: |
|
1284 | if lowestrev > nullrev: | |
1275 | # But, since we weren't, let's recompute the lowest rev to not |
|
1285 | # But, since we weren't, let's recompute the lowest rev to not | |
1276 | # include roots that aren't ancestors. |
|
1286 | # include roots that aren't ancestors. | |
1277 |
|
1287 | |||
1278 | # Filter out roots that aren't ancestors of heads |
|
1288 | # Filter out roots that aren't ancestors of heads | |
1279 | roots = [root for root in roots if root in ancestors] |
|
1289 | roots = [root for root in roots if root in ancestors] | |
1280 | # Recompute the lowest revision |
|
1290 | # Recompute the lowest revision | |
1281 | if roots: |
|
1291 | if roots: | |
1282 | lowestrev = min([self.rev(root) for root in roots]) |
|
1292 | lowestrev = min([self.rev(root) for root in roots]) | |
1283 | else: |
|
1293 | else: | |
1284 | # No more roots? Return empty list |
|
1294 | # No more roots? Return empty list | |
1285 | return nonodes |
|
1295 | return nonodes | |
1286 | else: |
|
1296 | else: | |
1287 | # We are descending from nullid, and don't need to care about |
|
1297 | # We are descending from nullid, and don't need to care about | |
1288 | # any other roots. |
|
1298 | # any other roots. | |
1289 | lowestrev = nullrev |
|
1299 | lowestrev = nullrev | |
1290 | roots = [self.nullid] |
|
1300 | roots = [self.nullid] | |
1291 | # Transform our roots list into a set. |
|
1301 | # Transform our roots list into a set. | |
1292 | descendants = set(roots) |
|
1302 | descendants = set(roots) | |
1293 | # Also, keep the original roots so we can filter out roots that aren't |
|
1303 | # Also, keep the original roots so we can filter out roots that aren't | |
1294 | # 'real' roots (i.e. are descended from other roots). |
|
1304 | # 'real' roots (i.e. are descended from other roots). | |
1295 | roots = descendants.copy() |
|
1305 | roots = descendants.copy() | |
1296 | # Our topologically sorted list of output nodes. |
|
1306 | # Our topologically sorted list of output nodes. | |
1297 | orderedout = [] |
|
1307 | orderedout = [] | |
1298 | # Don't start at nullid since we don't want nullid in our output list, |
|
1308 | # Don't start at nullid since we don't want nullid in our output list, | |
1299 | # and if nullid shows up in descendants, empty parents will look like |
|
1309 | # and if nullid shows up in descendants, empty parents will look like | |
1300 | # they're descendants. |
|
1310 | # they're descendants. | |
1301 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): |
|
1311 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): | |
1302 | n = self.node(r) |
|
1312 | n = self.node(r) | |
1303 | isdescendant = False |
|
1313 | isdescendant = False | |
1304 | if lowestrev == nullrev: # Everybody is a descendant of nullid |
|
1314 | if lowestrev == nullrev: # Everybody is a descendant of nullid | |
1305 | isdescendant = True |
|
1315 | isdescendant = True | |
1306 | elif n in descendants: |
|
1316 | elif n in descendants: | |
1307 | # n is already a descendant |
|
1317 | # n is already a descendant | |
1308 | isdescendant = True |
|
1318 | isdescendant = True | |
1309 | # This check only needs to be done here because all the roots |
|
1319 | # This check only needs to be done here because all the roots | |
1310 | # will start being marked is descendants before the loop. |
|
1320 | # will start being marked is descendants before the loop. | |
1311 | if n in roots: |
|
1321 | if n in roots: | |
1312 | # If n was a root, check if it's a 'real' root. |
|
1322 | # If n was a root, check if it's a 'real' root. | |
1313 | p = tuple(self.parents(n)) |
|
1323 | p = tuple(self.parents(n)) | |
1314 | # If any of its parents are descendants, it's not a root. |
|
1324 | # If any of its parents are descendants, it's not a root. | |
1315 | if (p[0] in descendants) or (p[1] in descendants): |
|
1325 | if (p[0] in descendants) or (p[1] in descendants): | |
1316 | roots.remove(n) |
|
1326 | roots.remove(n) | |
1317 | else: |
|
1327 | else: | |
1318 | p = tuple(self.parents(n)) |
|
1328 | p = tuple(self.parents(n)) | |
1319 | # A node is a descendant if either of its parents are |
|
1329 | # A node is a descendant if either of its parents are | |
1320 | # descendants. (We seeded the dependents list with the roots |
|
1330 | # descendants. (We seeded the dependents list with the roots | |
1321 | # up there, remember?) |
|
1331 | # up there, remember?) | |
1322 | if (p[0] in descendants) or (p[1] in descendants): |
|
1332 | if (p[0] in descendants) or (p[1] in descendants): | |
1323 | descendants.add(n) |
|
1333 | descendants.add(n) | |
1324 | isdescendant = True |
|
1334 | isdescendant = True | |
1325 | if isdescendant and ((ancestors is None) or (n in ancestors)): |
|
1335 | if isdescendant and ((ancestors is None) or (n in ancestors)): | |
1326 | # Only include nodes that are both descendants and ancestors. |
|
1336 | # Only include nodes that are both descendants and ancestors. | |
1327 | orderedout.append(n) |
|
1337 | orderedout.append(n) | |
1328 | if (ancestors is not None) and (n in heads): |
|
1338 | if (ancestors is not None) and (n in heads): | |
1329 | # We're trying to figure out which heads are reachable |
|
1339 | # We're trying to figure out which heads are reachable | |
1330 | # from roots. |
|
1340 | # from roots. | |
1331 | # Mark this head as having been reached |
|
1341 | # Mark this head as having been reached | |
1332 | heads[n] = True |
|
1342 | heads[n] = True | |
1333 | elif ancestors is None: |
|
1343 | elif ancestors is None: | |
1334 | # Otherwise, we're trying to discover the heads. |
|
1344 | # Otherwise, we're trying to discover the heads. | |
1335 | # Assume this is a head because if it isn't, the next step |
|
1345 | # Assume this is a head because if it isn't, the next step | |
1336 | # will eventually remove it. |
|
1346 | # will eventually remove it. | |
1337 | heads[n] = True |
|
1347 | heads[n] = True | |
1338 | # But, obviously its parents aren't. |
|
1348 | # But, obviously its parents aren't. | |
1339 | for p in self.parents(n): |
|
1349 | for p in self.parents(n): | |
1340 | heads.pop(p, None) |
|
1350 | heads.pop(p, None) | |
1341 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] |
|
1351 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] | |
1342 | roots = list(roots) |
|
1352 | roots = list(roots) | |
1343 | assert orderedout |
|
1353 | assert orderedout | |
1344 | assert roots |
|
1354 | assert roots | |
1345 | assert heads |
|
1355 | assert heads | |
1346 | return (orderedout, roots, heads) |
|
1356 | return (orderedout, roots, heads) | |
1347 |
|
1357 | |||
1348 | def headrevs(self, revs=None): |
|
1358 | def headrevs(self, revs=None): | |
1349 | if revs is None: |
|
1359 | if revs is None: | |
1350 | try: |
|
1360 | try: | |
1351 | return self.index.headrevs() |
|
1361 | return self.index.headrevs() | |
1352 | except AttributeError: |
|
1362 | except AttributeError: | |
1353 | return self._headrevs() |
|
1363 | return self._headrevs() | |
1354 | if rustdagop is not None: |
|
1364 | if rustdagop is not None: | |
1355 | return rustdagop.headrevs(self.index, revs) |
|
1365 | return rustdagop.headrevs(self.index, revs) | |
1356 | return dagop.headrevs(revs, self._uncheckedparentrevs) |
|
1366 | return dagop.headrevs(revs, self._uncheckedparentrevs) | |
1357 |
|
1367 | |||
1358 | def computephases(self, roots): |
|
1368 | def computephases(self, roots): | |
1359 | return self.index.computephasesmapsets(roots) |
|
1369 | return self.index.computephasesmapsets(roots) | |
1360 |
|
1370 | |||
1361 | def _headrevs(self): |
|
1371 | def _headrevs(self): | |
1362 | count = len(self) |
|
1372 | count = len(self) | |
1363 | if not count: |
|
1373 | if not count: | |
1364 | return [nullrev] |
|
1374 | return [nullrev] | |
1365 | # we won't iter over filtered rev so nobody is a head at start |
|
1375 | # we won't iter over filtered rev so nobody is a head at start | |
1366 | ishead = [0] * (count + 1) |
|
1376 | ishead = [0] * (count + 1) | |
1367 | index = self.index |
|
1377 | index = self.index | |
1368 | for r in self: |
|
1378 | for r in self: | |
1369 | ishead[r] = 1 # I may be an head |
|
1379 | ishead[r] = 1 # I may be an head | |
1370 | e = index[r] |
|
1380 | e = index[r] | |
1371 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not |
|
1381 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not | |
1372 | return [r for r, val in enumerate(ishead) if val] |
|
1382 | return [r for r, val in enumerate(ishead) if val] | |
1373 |
|
1383 | |||
1374 | def heads(self, start=None, stop=None): |
|
1384 | def heads(self, start=None, stop=None): | |
1375 | """return the list of all nodes that have no children |
|
1385 | """return the list of all nodes that have no children | |
1376 |
|
1386 | |||
1377 | if start is specified, only heads that are descendants of |
|
1387 | if start is specified, only heads that are descendants of | |
1378 | start will be returned |
|
1388 | start will be returned | |
1379 | if stop is specified, it will consider all the revs from stop |
|
1389 | if stop is specified, it will consider all the revs from stop | |
1380 | as if they had no children |
|
1390 | as if they had no children | |
1381 | """ |
|
1391 | """ | |
1382 | if start is None and stop is None: |
|
1392 | if start is None and stop is None: | |
1383 | if not len(self): |
|
1393 | if not len(self): | |
1384 | return [self.nullid] |
|
1394 | return [self.nullid] | |
1385 | return [self.node(r) for r in self.headrevs()] |
|
1395 | return [self.node(r) for r in self.headrevs()] | |
1386 |
|
1396 | |||
1387 | if start is None: |
|
1397 | if start is None: | |
1388 | start = nullrev |
|
1398 | start = nullrev | |
1389 | else: |
|
1399 | else: | |
1390 | start = self.rev(start) |
|
1400 | start = self.rev(start) | |
1391 |
|
1401 | |||
1392 | stoprevs = {self.rev(n) for n in stop or []} |
|
1402 | stoprevs = {self.rev(n) for n in stop or []} | |
1393 |
|
1403 | |||
1394 | revs = dagop.headrevssubset( |
|
1404 | revs = dagop.headrevssubset( | |
1395 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs |
|
1405 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs | |
1396 | ) |
|
1406 | ) | |
1397 |
|
1407 | |||
1398 | return [self.node(rev) for rev in revs] |
|
1408 | return [self.node(rev) for rev in revs] | |
1399 |
|
1409 | |||
1400 | def children(self, node): |
|
1410 | def children(self, node): | |
1401 | """find the children of a given node""" |
|
1411 | """find the children of a given node""" | |
1402 | c = [] |
|
1412 | c = [] | |
1403 | p = self.rev(node) |
|
1413 | p = self.rev(node) | |
1404 | for r in self.revs(start=p + 1): |
|
1414 | for r in self.revs(start=p + 1): | |
1405 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
1415 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
1406 | if prevs: |
|
1416 | if prevs: | |
1407 | for pr in prevs: |
|
1417 | for pr in prevs: | |
1408 | if pr == p: |
|
1418 | if pr == p: | |
1409 | c.append(self.node(r)) |
|
1419 | c.append(self.node(r)) | |
1410 | elif p == nullrev: |
|
1420 | elif p == nullrev: | |
1411 | c.append(self.node(r)) |
|
1421 | c.append(self.node(r)) | |
1412 | return c |
|
1422 | return c | |
1413 |
|
1423 | |||
1414 | def commonancestorsheads(self, a, b): |
|
1424 | def commonancestorsheads(self, a, b): | |
1415 | """calculate all the heads of the common ancestors of nodes a and b""" |
|
1425 | """calculate all the heads of the common ancestors of nodes a and b""" | |
1416 | a, b = self.rev(a), self.rev(b) |
|
1426 | a, b = self.rev(a), self.rev(b) | |
1417 | ancs = self._commonancestorsheads(a, b) |
|
1427 | ancs = self._commonancestorsheads(a, b) | |
1418 | return pycompat.maplist(self.node, ancs) |
|
1428 | return pycompat.maplist(self.node, ancs) | |
1419 |
|
1429 | |||
1420 | def _commonancestorsheads(self, *revs): |
|
1430 | def _commonancestorsheads(self, *revs): | |
1421 | """calculate all the heads of the common ancestors of revs""" |
|
1431 | """calculate all the heads of the common ancestors of revs""" | |
1422 | try: |
|
1432 | try: | |
1423 | ancs = self.index.commonancestorsheads(*revs) |
|
1433 | ancs = self.index.commonancestorsheads(*revs) | |
1424 | except (AttributeError, OverflowError): # C implementation failed |
|
1434 | except (AttributeError, OverflowError): # C implementation failed | |
1425 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) |
|
1435 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) | |
1426 | return ancs |
|
1436 | return ancs | |
1427 |
|
1437 | |||
1428 | def isancestor(self, a, b): |
|
1438 | def isancestor(self, a, b): | |
1429 | """return True if node a is an ancestor of node b |
|
1439 | """return True if node a is an ancestor of node b | |
1430 |
|
1440 | |||
1431 | A revision is considered an ancestor of itself.""" |
|
1441 | A revision is considered an ancestor of itself.""" | |
1432 | a, b = self.rev(a), self.rev(b) |
|
1442 | a, b = self.rev(a), self.rev(b) | |
1433 | return self.isancestorrev(a, b) |
|
1443 | return self.isancestorrev(a, b) | |
1434 |
|
1444 | |||
1435 | def isancestorrev(self, a, b): |
|
1445 | def isancestorrev(self, a, b): | |
1436 | """return True if revision a is an ancestor of revision b |
|
1446 | """return True if revision a is an ancestor of revision b | |
1437 |
|
1447 | |||
1438 | A revision is considered an ancestor of itself. |
|
1448 | A revision is considered an ancestor of itself. | |
1439 |
|
1449 | |||
1440 | The implementation of this is trivial but the use of |
|
1450 | The implementation of this is trivial but the use of | |
1441 | reachableroots is not.""" |
|
1451 | reachableroots is not.""" | |
1442 | if a == nullrev: |
|
1452 | if a == nullrev: | |
1443 | return True |
|
1453 | return True | |
1444 | elif a == b: |
|
1454 | elif a == b: | |
1445 | return True |
|
1455 | return True | |
1446 | elif a > b: |
|
1456 | elif a > b: | |
1447 | return False |
|
1457 | return False | |
1448 | return bool(self.reachableroots(a, [b], [a], includepath=False)) |
|
1458 | return bool(self.reachableroots(a, [b], [a], includepath=False)) | |
1449 |
|
1459 | |||
1450 | def reachableroots(self, minroot, heads, roots, includepath=False): |
|
1460 | def reachableroots(self, minroot, heads, roots, includepath=False): | |
1451 | """return (heads(::(<roots> and <roots>::<heads>))) |
|
1461 | """return (heads(::(<roots> and <roots>::<heads>))) | |
1452 |
|
1462 | |||
1453 | If includepath is True, return (<roots>::<heads>).""" |
|
1463 | If includepath is True, return (<roots>::<heads>).""" | |
1454 | try: |
|
1464 | try: | |
1455 | return self.index.reachableroots2( |
|
1465 | return self.index.reachableroots2( | |
1456 | minroot, heads, roots, includepath |
|
1466 | minroot, heads, roots, includepath | |
1457 | ) |
|
1467 | ) | |
1458 | except AttributeError: |
|
1468 | except AttributeError: | |
1459 | return dagop._reachablerootspure( |
|
1469 | return dagop._reachablerootspure( | |
1460 | self.parentrevs, minroot, roots, heads, includepath |
|
1470 | self.parentrevs, minroot, roots, heads, includepath | |
1461 | ) |
|
1471 | ) | |
1462 |
|
1472 | |||
1463 | def ancestor(self, a, b): |
|
1473 | def ancestor(self, a, b): | |
1464 | """calculate the "best" common ancestor of nodes a and b""" |
|
1474 | """calculate the "best" common ancestor of nodes a and b""" | |
1465 |
|
1475 | |||
1466 | a, b = self.rev(a), self.rev(b) |
|
1476 | a, b = self.rev(a), self.rev(b) | |
1467 | try: |
|
1477 | try: | |
1468 | ancs = self.index.ancestors(a, b) |
|
1478 | ancs = self.index.ancestors(a, b) | |
1469 | except (AttributeError, OverflowError): |
|
1479 | except (AttributeError, OverflowError): | |
1470 | ancs = ancestor.ancestors(self.parentrevs, a, b) |
|
1480 | ancs = ancestor.ancestors(self.parentrevs, a, b) | |
1471 | if ancs: |
|
1481 | if ancs: | |
1472 | # choose a consistent winner when there's a tie |
|
1482 | # choose a consistent winner when there's a tie | |
1473 | return min(map(self.node, ancs)) |
|
1483 | return min(map(self.node, ancs)) | |
1474 | return self.nullid |
|
1484 | return self.nullid | |
1475 |
|
1485 | |||
1476 | def _match(self, id): |
|
1486 | def _match(self, id): | |
1477 | if isinstance(id, int): |
|
1487 | if isinstance(id, int): | |
1478 | # rev |
|
1488 | # rev | |
1479 | return self.node(id) |
|
1489 | return self.node(id) | |
1480 | if len(id) == self.nodeconstants.nodelen: |
|
1490 | if len(id) == self.nodeconstants.nodelen: | |
1481 | # possibly a binary node |
|
1491 | # possibly a binary node | |
1482 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
1492 | # odds of a binary node being all hex in ASCII are 1 in 10**25 | |
1483 | try: |
|
1493 | try: | |
1484 | node = id |
|
1494 | node = id | |
1485 | self.rev(node) # quick search the index |
|
1495 | self.rev(node) # quick search the index | |
1486 | return node |
|
1496 | return node | |
1487 | except error.LookupError: |
|
1497 | except error.LookupError: | |
1488 | pass # may be partial hex id |
|
1498 | pass # may be partial hex id | |
1489 | try: |
|
1499 | try: | |
1490 | # str(rev) |
|
1500 | # str(rev) | |
1491 | rev = int(id) |
|
1501 | rev = int(id) | |
1492 | if b"%d" % rev != id: |
|
1502 | if b"%d" % rev != id: | |
1493 | raise ValueError |
|
1503 | raise ValueError | |
1494 | if rev < 0: |
|
1504 | if rev < 0: | |
1495 | rev = len(self) + rev |
|
1505 | rev = len(self) + rev | |
1496 | if rev < 0 or rev >= len(self): |
|
1506 | if rev < 0 or rev >= len(self): | |
1497 | raise ValueError |
|
1507 | raise ValueError | |
1498 | return self.node(rev) |
|
1508 | return self.node(rev) | |
1499 | except (ValueError, OverflowError): |
|
1509 | except (ValueError, OverflowError): | |
1500 | pass |
|
1510 | pass | |
1501 | if len(id) == 2 * self.nodeconstants.nodelen: |
|
1511 | if len(id) == 2 * self.nodeconstants.nodelen: | |
1502 | try: |
|
1512 | try: | |
1503 | # a full hex nodeid? |
|
1513 | # a full hex nodeid? | |
1504 | node = bin(id) |
|
1514 | node = bin(id) | |
1505 | self.rev(node) |
|
1515 | self.rev(node) | |
1506 | return node |
|
1516 | return node | |
1507 | except (TypeError, error.LookupError): |
|
1517 | except (TypeError, error.LookupError): | |
1508 | pass |
|
1518 | pass | |
1509 |
|
1519 | |||
1510 | def _partialmatch(self, id): |
|
1520 | def _partialmatch(self, id): | |
1511 | # we don't care wdirfilenodeids as they should be always full hash |
|
1521 | # we don't care wdirfilenodeids as they should be always full hash | |
1512 | maybewdir = self.nodeconstants.wdirhex.startswith(id) |
|
1522 | maybewdir = self.nodeconstants.wdirhex.startswith(id) | |
1513 | try: |
|
1523 | try: | |
1514 | partial = self.index.partialmatch(id) |
|
1524 | partial = self.index.partialmatch(id) | |
1515 | if partial and self.hasnode(partial): |
|
1525 | if partial and self.hasnode(partial): | |
1516 | if maybewdir: |
|
1526 | if maybewdir: | |
1517 | # single 'ff...' match in radix tree, ambiguous with wdir |
|
1527 | # single 'ff...' match in radix tree, ambiguous with wdir | |
1518 | raise error.RevlogError |
|
1528 | raise error.RevlogError | |
1519 | return partial |
|
1529 | return partial | |
1520 | if maybewdir: |
|
1530 | if maybewdir: | |
1521 | # no 'ff...' match in radix tree, wdir identified |
|
1531 | # no 'ff...' match in radix tree, wdir identified | |
1522 | raise error.WdirUnsupported |
|
1532 | raise error.WdirUnsupported | |
1523 | return None |
|
1533 | return None | |
1524 | except error.RevlogError: |
|
1534 | except error.RevlogError: | |
1525 | # parsers.c radix tree lookup gave multiple matches |
|
1535 | # parsers.c radix tree lookup gave multiple matches | |
1526 | # fast path: for unfiltered changelog, radix tree is accurate |
|
1536 | # fast path: for unfiltered changelog, radix tree is accurate | |
1527 | if not getattr(self, 'filteredrevs', None): |
|
1537 | if not getattr(self, 'filteredrevs', None): | |
1528 | raise error.AmbiguousPrefixLookupError( |
|
1538 | raise error.AmbiguousPrefixLookupError( | |
1529 | id, self.display_id, _(b'ambiguous identifier') |
|
1539 | id, self.display_id, _(b'ambiguous identifier') | |
1530 | ) |
|
1540 | ) | |
1531 | # fall through to slow path that filters hidden revisions |
|
1541 | # fall through to slow path that filters hidden revisions | |
1532 | except (AttributeError, ValueError): |
|
1542 | except (AttributeError, ValueError): | |
1533 | # we are pure python, or key was too short to search radix tree |
|
1543 | # we are pure python, or key was too short to search radix tree | |
1534 | pass |
|
1544 | pass | |
1535 |
|
1545 | |||
1536 | if id in self._pcache: |
|
1546 | if id in self._pcache: | |
1537 | return self._pcache[id] |
|
1547 | return self._pcache[id] | |
1538 |
|
1548 | |||
1539 | if len(id) <= 40: |
|
1549 | if len(id) <= 40: | |
1540 | try: |
|
1550 | try: | |
1541 | # hex(node)[:...] |
|
1551 | # hex(node)[:...] | |
1542 | l = len(id) // 2 # grab an even number of digits |
|
1552 | l = len(id) // 2 # grab an even number of digits | |
1543 | prefix = bin(id[: l * 2]) |
|
1553 | prefix = bin(id[: l * 2]) | |
1544 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] |
|
1554 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] | |
1545 | nl = [ |
|
1555 | nl = [ | |
1546 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) |
|
1556 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) | |
1547 | ] |
|
1557 | ] | |
1548 | if self.nodeconstants.nullhex.startswith(id): |
|
1558 | if self.nodeconstants.nullhex.startswith(id): | |
1549 | nl.append(self.nullid) |
|
1559 | nl.append(self.nullid) | |
1550 | if len(nl) > 0: |
|
1560 | if len(nl) > 0: | |
1551 | if len(nl) == 1 and not maybewdir: |
|
1561 | if len(nl) == 1 and not maybewdir: | |
1552 | self._pcache[id] = nl[0] |
|
1562 | self._pcache[id] = nl[0] | |
1553 | return nl[0] |
|
1563 | return nl[0] | |
1554 | raise error.AmbiguousPrefixLookupError( |
|
1564 | raise error.AmbiguousPrefixLookupError( | |
1555 | id, self.display_id, _(b'ambiguous identifier') |
|
1565 | id, self.display_id, _(b'ambiguous identifier') | |
1556 | ) |
|
1566 | ) | |
1557 | if maybewdir: |
|
1567 | if maybewdir: | |
1558 | raise error.WdirUnsupported |
|
1568 | raise error.WdirUnsupported | |
1559 | return None |
|
1569 | return None | |
1560 | except TypeError: |
|
1570 | except TypeError: | |
1561 | pass |
|
1571 | pass | |
1562 |
|
1572 | |||
1563 | def lookup(self, id): |
|
1573 | def lookup(self, id): | |
1564 | """locate a node based on: |
|
1574 | """locate a node based on: | |
1565 | - revision number or str(revision number) |
|
1575 | - revision number or str(revision number) | |
1566 | - nodeid or subset of hex nodeid |
|
1576 | - nodeid or subset of hex nodeid | |
1567 | """ |
|
1577 | """ | |
1568 | n = self._match(id) |
|
1578 | n = self._match(id) | |
1569 | if n is not None: |
|
1579 | if n is not None: | |
1570 | return n |
|
1580 | return n | |
1571 | n = self._partialmatch(id) |
|
1581 | n = self._partialmatch(id) | |
1572 | if n: |
|
1582 | if n: | |
1573 | return n |
|
1583 | return n | |
1574 |
|
1584 | |||
1575 | raise error.LookupError(id, self.display_id, _(b'no match found')) |
|
1585 | raise error.LookupError(id, self.display_id, _(b'no match found')) | |
1576 |
|
1586 | |||
1577 | def shortest(self, node, minlength=1): |
|
1587 | def shortest(self, node, minlength=1): | |
1578 | """Find the shortest unambiguous prefix that matches node.""" |
|
1588 | """Find the shortest unambiguous prefix that matches node.""" | |
1579 |
|
1589 | |||
1580 | def isvalid(prefix): |
|
1590 | def isvalid(prefix): | |
1581 | try: |
|
1591 | try: | |
1582 | matchednode = self._partialmatch(prefix) |
|
1592 | matchednode = self._partialmatch(prefix) | |
1583 | except error.AmbiguousPrefixLookupError: |
|
1593 | except error.AmbiguousPrefixLookupError: | |
1584 | return False |
|
1594 | return False | |
1585 | except error.WdirUnsupported: |
|
1595 | except error.WdirUnsupported: | |
1586 | # single 'ff...' match |
|
1596 | # single 'ff...' match | |
1587 | return True |
|
1597 | return True | |
1588 | if matchednode is None: |
|
1598 | if matchednode is None: | |
1589 | raise error.LookupError(node, self.display_id, _(b'no node')) |
|
1599 | raise error.LookupError(node, self.display_id, _(b'no node')) | |
1590 | return True |
|
1600 | return True | |
1591 |
|
1601 | |||
1592 | def maybewdir(prefix): |
|
1602 | def maybewdir(prefix): | |
1593 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) |
|
1603 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) | |
1594 |
|
1604 | |||
1595 | hexnode = hex(node) |
|
1605 | hexnode = hex(node) | |
1596 |
|
1606 | |||
1597 | def disambiguate(hexnode, minlength): |
|
1607 | def disambiguate(hexnode, minlength): | |
1598 | """Disambiguate against wdirid.""" |
|
1608 | """Disambiguate against wdirid.""" | |
1599 | for length in range(minlength, len(hexnode) + 1): |
|
1609 | for length in range(minlength, len(hexnode) + 1): | |
1600 | prefix = hexnode[:length] |
|
1610 | prefix = hexnode[:length] | |
1601 | if not maybewdir(prefix): |
|
1611 | if not maybewdir(prefix): | |
1602 | return prefix |
|
1612 | return prefix | |
1603 |
|
1613 | |||
1604 | if not getattr(self, 'filteredrevs', None): |
|
1614 | if not getattr(self, 'filteredrevs', None): | |
1605 | try: |
|
1615 | try: | |
1606 | length = max(self.index.shortest(node), minlength) |
|
1616 | length = max(self.index.shortest(node), minlength) | |
1607 | return disambiguate(hexnode, length) |
|
1617 | return disambiguate(hexnode, length) | |
1608 | except error.RevlogError: |
|
1618 | except error.RevlogError: | |
1609 | if node != self.nodeconstants.wdirid: |
|
1619 | if node != self.nodeconstants.wdirid: | |
1610 | raise error.LookupError( |
|
1620 | raise error.LookupError( | |
1611 | node, self.display_id, _(b'no node') |
|
1621 | node, self.display_id, _(b'no node') | |
1612 | ) |
|
1622 | ) | |
1613 | except AttributeError: |
|
1623 | except AttributeError: | |
1614 | # Fall through to pure code |
|
1624 | # Fall through to pure code | |
1615 | pass |
|
1625 | pass | |
1616 |
|
1626 | |||
1617 | if node == self.nodeconstants.wdirid: |
|
1627 | if node == self.nodeconstants.wdirid: | |
1618 | for length in range(minlength, len(hexnode) + 1): |
|
1628 | for length in range(minlength, len(hexnode) + 1): | |
1619 | prefix = hexnode[:length] |
|
1629 | prefix = hexnode[:length] | |
1620 | if isvalid(prefix): |
|
1630 | if isvalid(prefix): | |
1621 | return prefix |
|
1631 | return prefix | |
1622 |
|
1632 | |||
1623 | for length in range(minlength, len(hexnode) + 1): |
|
1633 | for length in range(minlength, len(hexnode) + 1): | |
1624 | prefix = hexnode[:length] |
|
1634 | prefix = hexnode[:length] | |
1625 | if isvalid(prefix): |
|
1635 | if isvalid(prefix): | |
1626 | return disambiguate(hexnode, length) |
|
1636 | return disambiguate(hexnode, length) | |
1627 |
|
1637 | |||
1628 | def cmp(self, node, text): |
|
1638 | def cmp(self, node, text): | |
1629 | """compare text with a given file revision |
|
1639 | """compare text with a given file revision | |
1630 |
|
1640 | |||
1631 | returns True if text is different than what is stored. |
|
1641 | returns True if text is different than what is stored. | |
1632 | """ |
|
1642 | """ | |
1633 | p1, p2 = self.parents(node) |
|
1643 | p1, p2 = self.parents(node) | |
1634 | return storageutil.hashrevisionsha1(text, p1, p2) != node |
|
1644 | return storageutil.hashrevisionsha1(text, p1, p2) != node | |
1635 |
|
1645 | |||
1636 | def _cachesegment(self, offset, data): |
|
1646 | def _cachesegment(self, offset, data): | |
1637 | """Add a segment to the revlog cache. |
|
1647 | """Add a segment to the revlog cache. | |
1638 |
|
1648 | |||
1639 | Accepts an absolute offset and the data that is at that location. |
|
1649 | Accepts an absolute offset and the data that is at that location. | |
1640 | """ |
|
1650 | """ | |
1641 | o, d = self._chunkcache |
|
1651 | o, d = self._chunkcache | |
1642 | # try to add to existing cache |
|
1652 | # try to add to existing cache | |
1643 | if o + len(d) == offset and len(d) + len(data) < _chunksize: |
|
1653 | if o + len(d) == offset and len(d) + len(data) < _chunksize: | |
1644 | self._chunkcache = o, d + data |
|
1654 | self._chunkcache = o, d + data | |
1645 | else: |
|
1655 | else: | |
1646 | self._chunkcache = offset, data |
|
1656 | self._chunkcache = offset, data | |
1647 |
|
1657 | |||
1648 | def _readsegment(self, offset, length, df=None): |
|
1658 | def _readsegment(self, offset, length, df=None): | |
1649 | """Load a segment of raw data from the revlog. |
|
1659 | """Load a segment of raw data from the revlog. | |
1650 |
|
1660 | |||
1651 | Accepts an absolute offset, length to read, and an optional existing |
|
1661 | Accepts an absolute offset, length to read, and an optional existing | |
1652 | file handle to read from. |
|
1662 | file handle to read from. | |
1653 |
|
1663 | |||
1654 | If an existing file handle is passed, it will be seeked and the |
|
1664 | If an existing file handle is passed, it will be seeked and the | |
1655 | original seek position will NOT be restored. |
|
1665 | original seek position will NOT be restored. | |
1656 |
|
1666 | |||
1657 | Returns a str or buffer of raw byte data. |
|
1667 | Returns a str or buffer of raw byte data. | |
1658 |
|
1668 | |||
1659 | Raises if the requested number of bytes could not be read. |
|
1669 | Raises if the requested number of bytes could not be read. | |
1660 | """ |
|
1670 | """ | |
1661 | # Cache data both forward and backward around the requested |
|
1671 | # Cache data both forward and backward around the requested | |
1662 | # data, in a fixed size window. This helps speed up operations |
|
1672 | # data, in a fixed size window. This helps speed up operations | |
1663 | # involving reading the revlog backwards. |
|
1673 | # involving reading the revlog backwards. | |
1664 | cachesize = self._chunkcachesize |
|
1674 | cachesize = self._chunkcachesize | |
1665 | realoffset = offset & ~(cachesize - 1) |
|
1675 | realoffset = offset & ~(cachesize - 1) | |
1666 | reallength = ( |
|
1676 | reallength = ( | |
1667 | (offset + length + cachesize) & ~(cachesize - 1) |
|
1677 | (offset + length + cachesize) & ~(cachesize - 1) | |
1668 | ) - realoffset |
|
1678 | ) - realoffset | |
1669 | with self._datareadfp(df) as df: |
|
1679 | with self._datareadfp(df) as df: | |
1670 | df.seek(realoffset) |
|
1680 | df.seek(realoffset) | |
1671 | d = df.read(reallength) |
|
1681 | d = df.read(reallength) | |
1672 |
|
1682 | |||
1673 | self._cachesegment(realoffset, d) |
|
1683 | self._cachesegment(realoffset, d) | |
1674 | if offset != realoffset or reallength != length: |
|
1684 | if offset != realoffset or reallength != length: | |
1675 | startoffset = offset - realoffset |
|
1685 | startoffset = offset - realoffset | |
1676 | if len(d) - startoffset < length: |
|
1686 | if len(d) - startoffset < length: | |
1677 | raise error.RevlogError( |
|
1687 | raise error.RevlogError( | |
1678 | _( |
|
1688 | _( | |
1679 | b'partial read of revlog %s; expected %d bytes from ' |
|
1689 | b'partial read of revlog %s; expected %d bytes from ' | |
1680 | b'offset %d, got %d' |
|
1690 | b'offset %d, got %d' | |
1681 | ) |
|
1691 | ) | |
1682 | % ( |
|
1692 | % ( | |
1683 | self._indexfile if self._inline else self._datafile, |
|
1693 | self._indexfile if self._inline else self._datafile, | |
1684 | length, |
|
1694 | length, | |
1685 | offset, |
|
1695 | offset, | |
1686 | len(d) - startoffset, |
|
1696 | len(d) - startoffset, | |
1687 | ) |
|
1697 | ) | |
1688 | ) |
|
1698 | ) | |
1689 |
|
1699 | |||
1690 | return util.buffer(d, startoffset, length) |
|
1700 | return util.buffer(d, startoffset, length) | |
1691 |
|
1701 | |||
1692 | if len(d) < length: |
|
1702 | if len(d) < length: | |
1693 | raise error.RevlogError( |
|
1703 | raise error.RevlogError( | |
1694 | _( |
|
1704 | _( | |
1695 | b'partial read of revlog %s; expected %d bytes from offset ' |
|
1705 | b'partial read of revlog %s; expected %d bytes from offset ' | |
1696 | b'%d, got %d' |
|
1706 | b'%d, got %d' | |
1697 | ) |
|
1707 | ) | |
1698 | % ( |
|
1708 | % ( | |
1699 | self._indexfile if self._inline else self._datafile, |
|
1709 | self._indexfile if self._inline else self._datafile, | |
1700 | length, |
|
1710 | length, | |
1701 | offset, |
|
1711 | offset, | |
1702 | len(d), |
|
1712 | len(d), | |
1703 | ) |
|
1713 | ) | |
1704 | ) |
|
1714 | ) | |
1705 |
|
1715 | |||
1706 | return d |
|
1716 | return d | |
1707 |
|
1717 | |||
1708 | def _getsegment(self, offset, length, df=None): |
|
1718 | def _getsegment(self, offset, length, df=None): | |
1709 | """Obtain a segment of raw data from the revlog. |
|
1719 | """Obtain a segment of raw data from the revlog. | |
1710 |
|
1720 | |||
1711 | Accepts an absolute offset, length of bytes to obtain, and an |
|
1721 | Accepts an absolute offset, length of bytes to obtain, and an | |
1712 | optional file handle to the already-opened revlog. If the file |
|
1722 | optional file handle to the already-opened revlog. If the file | |
1713 | handle is used, it's original seek position will not be preserved. |
|
1723 | handle is used, it's original seek position will not be preserved. | |
1714 |
|
1724 | |||
1715 | Requests for data may be returned from a cache. |
|
1725 | Requests for data may be returned from a cache. | |
1716 |
|
1726 | |||
1717 | Returns a str or a buffer instance of raw byte data. |
|
1727 | Returns a str or a buffer instance of raw byte data. | |
1718 | """ |
|
1728 | """ | |
1719 | o, d = self._chunkcache |
|
1729 | o, d = self._chunkcache | |
1720 | l = len(d) |
|
1730 | l = len(d) | |
1721 |
|
1731 | |||
1722 | # is it in the cache? |
|
1732 | # is it in the cache? | |
1723 | cachestart = offset - o |
|
1733 | cachestart = offset - o | |
1724 | cacheend = cachestart + length |
|
1734 | cacheend = cachestart + length | |
1725 | if cachestart >= 0 and cacheend <= l: |
|
1735 | if cachestart >= 0 and cacheend <= l: | |
1726 | if cachestart == 0 and cacheend == l: |
|
1736 | if cachestart == 0 and cacheend == l: | |
1727 | return d # avoid a copy |
|
1737 | return d # avoid a copy | |
1728 | return util.buffer(d, cachestart, cacheend - cachestart) |
|
1738 | return util.buffer(d, cachestart, cacheend - cachestart) | |
1729 |
|
1739 | |||
1730 | return self._readsegment(offset, length, df=df) |
|
1740 | return self._readsegment(offset, length, df=df) | |
1731 |
|
1741 | |||
1732 | def _getsegmentforrevs(self, startrev, endrev, df=None): |
|
1742 | def _getsegmentforrevs(self, startrev, endrev, df=None): | |
1733 | """Obtain a segment of raw data corresponding to a range of revisions. |
|
1743 | """Obtain a segment of raw data corresponding to a range of revisions. | |
1734 |
|
1744 | |||
1735 | Accepts the start and end revisions and an optional already-open |
|
1745 | Accepts the start and end revisions and an optional already-open | |
1736 | file handle to be used for reading. If the file handle is read, its |
|
1746 | file handle to be used for reading. If the file handle is read, its | |
1737 | seek position will not be preserved. |
|
1747 | seek position will not be preserved. | |
1738 |
|
1748 | |||
1739 | Requests for data may be satisfied by a cache. |
|
1749 | Requests for data may be satisfied by a cache. | |
1740 |
|
1750 | |||
1741 | Returns a 2-tuple of (offset, data) for the requested range of |
|
1751 | Returns a 2-tuple of (offset, data) for the requested range of | |
1742 | revisions. Offset is the integer offset from the beginning of the |
|
1752 | revisions. Offset is the integer offset from the beginning of the | |
1743 | revlog and data is a str or buffer of the raw byte data. |
|
1753 | revlog and data is a str or buffer of the raw byte data. | |
1744 |
|
1754 | |||
1745 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` |
|
1755 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` | |
1746 | to determine where each revision's data begins and ends. |
|
1756 | to determine where each revision's data begins and ends. | |
1747 | """ |
|
1757 | """ | |
1748 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons |
|
1758 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons | |
1749 | # (functions are expensive). |
|
1759 | # (functions are expensive). | |
1750 | index = self.index |
|
1760 | index = self.index | |
1751 | istart = index[startrev] |
|
1761 | istart = index[startrev] | |
1752 | start = int(istart[0] >> 16) |
|
1762 | start = int(istart[0] >> 16) | |
1753 | if startrev == endrev: |
|
1763 | if startrev == endrev: | |
1754 | end = start + istart[1] |
|
1764 | end = start + istart[1] | |
1755 | else: |
|
1765 | else: | |
1756 | iend = index[endrev] |
|
1766 | iend = index[endrev] | |
1757 | end = int(iend[0] >> 16) + iend[1] |
|
1767 | end = int(iend[0] >> 16) + iend[1] | |
1758 |
|
1768 | |||
1759 | if self._inline: |
|
1769 | if self._inline: | |
1760 | start += (startrev + 1) * self.index.entry_size |
|
1770 | start += (startrev + 1) * self.index.entry_size | |
1761 | end += (endrev + 1) * self.index.entry_size |
|
1771 | end += (endrev + 1) * self.index.entry_size | |
1762 | length = end - start |
|
1772 | length = end - start | |
1763 |
|
1773 | |||
1764 | return start, self._getsegment(start, length, df=df) |
|
1774 | return start, self._getsegment(start, length, df=df) | |
1765 |
|
1775 | |||
1766 | def _chunk(self, rev, df=None): |
|
1776 | def _chunk(self, rev, df=None): | |
1767 | """Obtain a single decompressed chunk for a revision. |
|
1777 | """Obtain a single decompressed chunk for a revision. | |
1768 |
|
1778 | |||
1769 | Accepts an integer revision and an optional already-open file handle |
|
1779 | Accepts an integer revision and an optional already-open file handle | |
1770 | to be used for reading. If used, the seek position of the file will not |
|
1780 | to be used for reading. If used, the seek position of the file will not | |
1771 | be preserved. |
|
1781 | be preserved. | |
1772 |
|
1782 | |||
1773 | Returns a str holding uncompressed data for the requested revision. |
|
1783 | Returns a str holding uncompressed data for the requested revision. | |
1774 | """ |
|
1784 | """ | |
1775 | compression_mode = self.index[rev][10] |
|
1785 | compression_mode = self.index[rev][10] | |
1776 | data = self._getsegmentforrevs(rev, rev, df=df)[1] |
|
1786 | data = self._getsegmentforrevs(rev, rev, df=df)[1] | |
1777 | if compression_mode == COMP_MODE_PLAIN: |
|
1787 | if compression_mode == COMP_MODE_PLAIN: | |
1778 | return data |
|
1788 | return data | |
|
1789 | elif compression_mode == COMP_MODE_DEFAULT: | |||
|
1790 | return self._decompressor(data) | |||
1779 | elif compression_mode == COMP_MODE_INLINE: |
|
1791 | elif compression_mode == COMP_MODE_INLINE: | |
1780 | return self.decompress(data) |
|
1792 | return self.decompress(data) | |
1781 | else: |
|
1793 | else: | |
1782 | msg = 'unknown compression mode %d' |
|
1794 | msg = 'unknown compression mode %d' | |
1783 | msg %= compression_mode |
|
1795 | msg %= compression_mode | |
1784 | raise error.RevlogError(msg) |
|
1796 | raise error.RevlogError(msg) | |
1785 |
|
1797 | |||
1786 | def _chunks(self, revs, df=None, targetsize=None): |
|
1798 | def _chunks(self, revs, df=None, targetsize=None): | |
1787 | """Obtain decompressed chunks for the specified revisions. |
|
1799 | """Obtain decompressed chunks for the specified revisions. | |
1788 |
|
1800 | |||
1789 | Accepts an iterable of numeric revisions that are assumed to be in |
|
1801 | Accepts an iterable of numeric revisions that are assumed to be in | |
1790 | ascending order. Also accepts an optional already-open file handle |
|
1802 | ascending order. Also accepts an optional already-open file handle | |
1791 | to be used for reading. If used, the seek position of the file will |
|
1803 | to be used for reading. If used, the seek position of the file will | |
1792 | not be preserved. |
|
1804 | not be preserved. | |
1793 |
|
1805 | |||
1794 | This function is similar to calling ``self._chunk()`` multiple times, |
|
1806 | This function is similar to calling ``self._chunk()`` multiple times, | |
1795 | but is faster. |
|
1807 | but is faster. | |
1796 |
|
1808 | |||
1797 | Returns a list with decompressed data for each requested revision. |
|
1809 | Returns a list with decompressed data for each requested revision. | |
1798 | """ |
|
1810 | """ | |
1799 | if not revs: |
|
1811 | if not revs: | |
1800 | return [] |
|
1812 | return [] | |
1801 | start = self.start |
|
1813 | start = self.start | |
1802 | length = self.length |
|
1814 | length = self.length | |
1803 | inline = self._inline |
|
1815 | inline = self._inline | |
1804 | iosize = self.index.entry_size |
|
1816 | iosize = self.index.entry_size | |
1805 | buffer = util.buffer |
|
1817 | buffer = util.buffer | |
1806 |
|
1818 | |||
1807 | l = [] |
|
1819 | l = [] | |
1808 | ladd = l.append |
|
1820 | ladd = l.append | |
1809 |
|
1821 | |||
1810 | if not self._withsparseread: |
|
1822 | if not self._withsparseread: | |
1811 | slicedchunks = (revs,) |
|
1823 | slicedchunks = (revs,) | |
1812 | else: |
|
1824 | else: | |
1813 | slicedchunks = deltautil.slicechunk( |
|
1825 | slicedchunks = deltautil.slicechunk( | |
1814 | self, revs, targetsize=targetsize |
|
1826 | self, revs, targetsize=targetsize | |
1815 | ) |
|
1827 | ) | |
1816 |
|
1828 | |||
1817 | for revschunk in slicedchunks: |
|
1829 | for revschunk in slicedchunks: | |
1818 | firstrev = revschunk[0] |
|
1830 | firstrev = revschunk[0] | |
1819 | # Skip trailing revisions with empty diff |
|
1831 | # Skip trailing revisions with empty diff | |
1820 | for lastrev in revschunk[::-1]: |
|
1832 | for lastrev in revschunk[::-1]: | |
1821 | if length(lastrev) != 0: |
|
1833 | if length(lastrev) != 0: | |
1822 | break |
|
1834 | break | |
1823 |
|
1835 | |||
1824 | try: |
|
1836 | try: | |
1825 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) |
|
1837 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) | |
1826 | except OverflowError: |
|
1838 | except OverflowError: | |
1827 | # issue4215 - we can't cache a run of chunks greater than |
|
1839 | # issue4215 - we can't cache a run of chunks greater than | |
1828 | # 2G on Windows |
|
1840 | # 2G on Windows | |
1829 | return [self._chunk(rev, df=df) for rev in revschunk] |
|
1841 | return [self._chunk(rev, df=df) for rev in revschunk] | |
1830 |
|
1842 | |||
1831 | decomp = self.decompress |
|
1843 | decomp = self.decompress | |
|
1844 | # self._decompressor might be None, but will not be used in that case | |||
|
1845 | def_decomp = self._decompressor | |||
1832 | for rev in revschunk: |
|
1846 | for rev in revschunk: | |
1833 | chunkstart = start(rev) |
|
1847 | chunkstart = start(rev) | |
1834 | if inline: |
|
1848 | if inline: | |
1835 | chunkstart += (rev + 1) * iosize |
|
1849 | chunkstart += (rev + 1) * iosize | |
1836 | chunklength = length(rev) |
|
1850 | chunklength = length(rev) | |
1837 | comp_mode = self.index[rev][10] |
|
1851 | comp_mode = self.index[rev][10] | |
1838 | c = buffer(data, chunkstart - offset, chunklength) |
|
1852 | c = buffer(data, chunkstart - offset, chunklength) | |
1839 | if comp_mode == COMP_MODE_PLAIN: |
|
1853 | if comp_mode == COMP_MODE_PLAIN: | |
1840 | ladd(c) |
|
1854 | ladd(c) | |
1841 | elif comp_mode == COMP_MODE_INLINE: |
|
1855 | elif comp_mode == COMP_MODE_INLINE: | |
1842 | ladd(decomp(c)) |
|
1856 | ladd(decomp(c)) | |
|
1857 | elif comp_mode == COMP_MODE_DEFAULT: | |||
|
1858 | ladd(def_decomp(c)) | |||
1843 | else: |
|
1859 | else: | |
1844 | msg = 'unknown compression mode %d' |
|
1860 | msg = 'unknown compression mode %d' | |
1845 | msg %= comp_mode |
|
1861 | msg %= comp_mode | |
1846 | raise error.RevlogError(msg) |
|
1862 | raise error.RevlogError(msg) | |
1847 |
|
1863 | |||
1848 | return l |
|
1864 | return l | |
1849 |
|
1865 | |||
1850 | def _chunkclear(self): |
|
1866 | def _chunkclear(self): | |
1851 | """Clear the raw chunk cache.""" |
|
1867 | """Clear the raw chunk cache.""" | |
1852 | self._chunkcache = (0, b'') |
|
1868 | self._chunkcache = (0, b'') | |
1853 |
|
1869 | |||
1854 | def deltaparent(self, rev): |
|
1870 | def deltaparent(self, rev): | |
1855 | """return deltaparent of the given revision""" |
|
1871 | """return deltaparent of the given revision""" | |
1856 | base = self.index[rev][3] |
|
1872 | base = self.index[rev][3] | |
1857 | if base == rev: |
|
1873 | if base == rev: | |
1858 | return nullrev |
|
1874 | return nullrev | |
1859 | elif self._generaldelta: |
|
1875 | elif self._generaldelta: | |
1860 | return base |
|
1876 | return base | |
1861 | else: |
|
1877 | else: | |
1862 | return rev - 1 |
|
1878 | return rev - 1 | |
1863 |
|
1879 | |||
1864 | def issnapshot(self, rev): |
|
1880 | def issnapshot(self, rev): | |
1865 | """tells whether rev is a snapshot""" |
|
1881 | """tells whether rev is a snapshot""" | |
1866 | if not self._sparserevlog: |
|
1882 | if not self._sparserevlog: | |
1867 | return self.deltaparent(rev) == nullrev |
|
1883 | return self.deltaparent(rev) == nullrev | |
1868 | elif util.safehasattr(self.index, b'issnapshot'): |
|
1884 | elif util.safehasattr(self.index, b'issnapshot'): | |
1869 | # directly assign the method to cache the testing and access |
|
1885 | # directly assign the method to cache the testing and access | |
1870 | self.issnapshot = self.index.issnapshot |
|
1886 | self.issnapshot = self.index.issnapshot | |
1871 | return self.issnapshot(rev) |
|
1887 | return self.issnapshot(rev) | |
1872 | if rev == nullrev: |
|
1888 | if rev == nullrev: | |
1873 | return True |
|
1889 | return True | |
1874 | entry = self.index[rev] |
|
1890 | entry = self.index[rev] | |
1875 | base = entry[3] |
|
1891 | base = entry[3] | |
1876 | if base == rev: |
|
1892 | if base == rev: | |
1877 | return True |
|
1893 | return True | |
1878 | if base == nullrev: |
|
1894 | if base == nullrev: | |
1879 | return True |
|
1895 | return True | |
1880 | p1 = entry[5] |
|
1896 | p1 = entry[5] | |
1881 | p2 = entry[6] |
|
1897 | p2 = entry[6] | |
1882 | if base == p1 or base == p2: |
|
1898 | if base == p1 or base == p2: | |
1883 | return False |
|
1899 | return False | |
1884 | return self.issnapshot(base) |
|
1900 | return self.issnapshot(base) | |
1885 |
|
1901 | |||
1886 | def snapshotdepth(self, rev): |
|
1902 | def snapshotdepth(self, rev): | |
1887 | """number of snapshot in the chain before this one""" |
|
1903 | """number of snapshot in the chain before this one""" | |
1888 | if not self.issnapshot(rev): |
|
1904 | if not self.issnapshot(rev): | |
1889 | raise error.ProgrammingError(b'revision %d not a snapshot') |
|
1905 | raise error.ProgrammingError(b'revision %d not a snapshot') | |
1890 | return len(self._deltachain(rev)[0]) - 1 |
|
1906 | return len(self._deltachain(rev)[0]) - 1 | |
1891 |
|
1907 | |||
1892 | def revdiff(self, rev1, rev2): |
|
1908 | def revdiff(self, rev1, rev2): | |
1893 | """return or calculate a delta between two revisions |
|
1909 | """return or calculate a delta between two revisions | |
1894 |
|
1910 | |||
1895 | The delta calculated is in binary form and is intended to be written to |
|
1911 | The delta calculated is in binary form and is intended to be written to | |
1896 | revlog data directly. So this function needs raw revision data. |
|
1912 | revlog data directly. So this function needs raw revision data. | |
1897 | """ |
|
1913 | """ | |
1898 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: |
|
1914 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: | |
1899 | return bytes(self._chunk(rev2)) |
|
1915 | return bytes(self._chunk(rev2)) | |
1900 |
|
1916 | |||
1901 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) |
|
1917 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) | |
1902 |
|
1918 | |||
1903 | def _processflags(self, text, flags, operation, raw=False): |
|
1919 | def _processflags(self, text, flags, operation, raw=False): | |
1904 | """deprecated entry point to access flag processors""" |
|
1920 | """deprecated entry point to access flag processors""" | |
1905 | msg = b'_processflag(...) use the specialized variant' |
|
1921 | msg = b'_processflag(...) use the specialized variant' | |
1906 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1922 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) | |
1907 | if raw: |
|
1923 | if raw: | |
1908 | return text, flagutil.processflagsraw(self, text, flags) |
|
1924 | return text, flagutil.processflagsraw(self, text, flags) | |
1909 | elif operation == b'read': |
|
1925 | elif operation == b'read': | |
1910 | return flagutil.processflagsread(self, text, flags) |
|
1926 | return flagutil.processflagsread(self, text, flags) | |
1911 | else: # write operation |
|
1927 | else: # write operation | |
1912 | return flagutil.processflagswrite(self, text, flags) |
|
1928 | return flagutil.processflagswrite(self, text, flags) | |
1913 |
|
1929 | |||
1914 | def revision(self, nodeorrev, _df=None, raw=False): |
|
1930 | def revision(self, nodeorrev, _df=None, raw=False): | |
1915 | """return an uncompressed revision of a given node or revision |
|
1931 | """return an uncompressed revision of a given node or revision | |
1916 | number. |
|
1932 | number. | |
1917 |
|
1933 | |||
1918 | _df - an existing file handle to read from. (internal-only) |
|
1934 | _df - an existing file handle to read from. (internal-only) | |
1919 | raw - an optional argument specifying if the revision data is to be |
|
1935 | raw - an optional argument specifying if the revision data is to be | |
1920 | treated as raw data when applying flag transforms. 'raw' should be set |
|
1936 | treated as raw data when applying flag transforms. 'raw' should be set | |
1921 | to True when generating changegroups or in debug commands. |
|
1937 | to True when generating changegroups or in debug commands. | |
1922 | """ |
|
1938 | """ | |
1923 | if raw: |
|
1939 | if raw: | |
1924 | msg = ( |
|
1940 | msg = ( | |
1925 | b'revlog.revision(..., raw=True) is deprecated, ' |
|
1941 | b'revlog.revision(..., raw=True) is deprecated, ' | |
1926 | b'use revlog.rawdata(...)' |
|
1942 | b'use revlog.rawdata(...)' | |
1927 | ) |
|
1943 | ) | |
1928 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1944 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) | |
1929 | return self._revisiondata(nodeorrev, _df, raw=raw)[0] |
|
1945 | return self._revisiondata(nodeorrev, _df, raw=raw)[0] | |
1930 |
|
1946 | |||
1931 | def sidedata(self, nodeorrev, _df=None): |
|
1947 | def sidedata(self, nodeorrev, _df=None): | |
1932 | """a map of extra data related to the changeset but not part of the hash |
|
1948 | """a map of extra data related to the changeset but not part of the hash | |
1933 |
|
1949 | |||
1934 | This function currently return a dictionary. However, more advanced |
|
1950 | This function currently return a dictionary. However, more advanced | |
1935 | mapping object will likely be used in the future for a more |
|
1951 | mapping object will likely be used in the future for a more | |
1936 | efficient/lazy code. |
|
1952 | efficient/lazy code. | |
1937 | """ |
|
1953 | """ | |
1938 | return self._revisiondata(nodeorrev, _df)[1] |
|
1954 | return self._revisiondata(nodeorrev, _df)[1] | |
1939 |
|
1955 | |||
1940 | def _revisiondata(self, nodeorrev, _df=None, raw=False): |
|
1956 | def _revisiondata(self, nodeorrev, _df=None, raw=False): | |
1941 | # deal with <nodeorrev> argument type |
|
1957 | # deal with <nodeorrev> argument type | |
1942 | if isinstance(nodeorrev, int): |
|
1958 | if isinstance(nodeorrev, int): | |
1943 | rev = nodeorrev |
|
1959 | rev = nodeorrev | |
1944 | node = self.node(rev) |
|
1960 | node = self.node(rev) | |
1945 | else: |
|
1961 | else: | |
1946 | node = nodeorrev |
|
1962 | node = nodeorrev | |
1947 | rev = None |
|
1963 | rev = None | |
1948 |
|
1964 | |||
1949 | # fast path the special `nullid` rev |
|
1965 | # fast path the special `nullid` rev | |
1950 | if node == self.nullid: |
|
1966 | if node == self.nullid: | |
1951 | return b"", {} |
|
1967 | return b"", {} | |
1952 |
|
1968 | |||
1953 | # ``rawtext`` is the text as stored inside the revlog. Might be the |
|
1969 | # ``rawtext`` is the text as stored inside the revlog. Might be the | |
1954 | # revision or might need to be processed to retrieve the revision. |
|
1970 | # revision or might need to be processed to retrieve the revision. | |
1955 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) |
|
1971 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) | |
1956 |
|
1972 | |||
1957 | if self.hassidedata: |
|
1973 | if self.hassidedata: | |
1958 | if rev is None: |
|
1974 | if rev is None: | |
1959 | rev = self.rev(node) |
|
1975 | rev = self.rev(node) | |
1960 | sidedata = self._sidedata(rev) |
|
1976 | sidedata = self._sidedata(rev) | |
1961 | else: |
|
1977 | else: | |
1962 | sidedata = {} |
|
1978 | sidedata = {} | |
1963 |
|
1979 | |||
1964 | if raw and validated: |
|
1980 | if raw and validated: | |
1965 | # if we don't want to process the raw text and that raw |
|
1981 | # if we don't want to process the raw text and that raw | |
1966 | # text is cached, we can exit early. |
|
1982 | # text is cached, we can exit early. | |
1967 | return rawtext, sidedata |
|
1983 | return rawtext, sidedata | |
1968 | if rev is None: |
|
1984 | if rev is None: | |
1969 | rev = self.rev(node) |
|
1985 | rev = self.rev(node) | |
1970 | # the revlog's flag for this revision |
|
1986 | # the revlog's flag for this revision | |
1971 | # (usually alter its state or content) |
|
1987 | # (usually alter its state or content) | |
1972 | flags = self.flags(rev) |
|
1988 | flags = self.flags(rev) | |
1973 |
|
1989 | |||
1974 | if validated and flags == REVIDX_DEFAULT_FLAGS: |
|
1990 | if validated and flags == REVIDX_DEFAULT_FLAGS: | |
1975 | # no extra flags set, no flag processor runs, text = rawtext |
|
1991 | # no extra flags set, no flag processor runs, text = rawtext | |
1976 | return rawtext, sidedata |
|
1992 | return rawtext, sidedata | |
1977 |
|
1993 | |||
1978 | if raw: |
|
1994 | if raw: | |
1979 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
1995 | validatehash = flagutil.processflagsraw(self, rawtext, flags) | |
1980 | text = rawtext |
|
1996 | text = rawtext | |
1981 | else: |
|
1997 | else: | |
1982 | r = flagutil.processflagsread(self, rawtext, flags) |
|
1998 | r = flagutil.processflagsread(self, rawtext, flags) | |
1983 | text, validatehash = r |
|
1999 | text, validatehash = r | |
1984 | if validatehash: |
|
2000 | if validatehash: | |
1985 | self.checkhash(text, node, rev=rev) |
|
2001 | self.checkhash(text, node, rev=rev) | |
1986 | if not validated: |
|
2002 | if not validated: | |
1987 | self._revisioncache = (node, rev, rawtext) |
|
2003 | self._revisioncache = (node, rev, rawtext) | |
1988 |
|
2004 | |||
1989 | return text, sidedata |
|
2005 | return text, sidedata | |
1990 |
|
2006 | |||
1991 | def _rawtext(self, node, rev, _df=None): |
|
2007 | def _rawtext(self, node, rev, _df=None): | |
1992 | """return the possibly unvalidated rawtext for a revision |
|
2008 | """return the possibly unvalidated rawtext for a revision | |
1993 |
|
2009 | |||
1994 | returns (rev, rawtext, validated) |
|
2010 | returns (rev, rawtext, validated) | |
1995 | """ |
|
2011 | """ | |
1996 |
|
2012 | |||
1997 | # revision in the cache (could be useful to apply delta) |
|
2013 | # revision in the cache (could be useful to apply delta) | |
1998 | cachedrev = None |
|
2014 | cachedrev = None | |
1999 | # An intermediate text to apply deltas to |
|
2015 | # An intermediate text to apply deltas to | |
2000 | basetext = None |
|
2016 | basetext = None | |
2001 |
|
2017 | |||
2002 | # Check if we have the entry in cache |
|
2018 | # Check if we have the entry in cache | |
2003 | # The cache entry looks like (node, rev, rawtext) |
|
2019 | # The cache entry looks like (node, rev, rawtext) | |
2004 | if self._revisioncache: |
|
2020 | if self._revisioncache: | |
2005 | if self._revisioncache[0] == node: |
|
2021 | if self._revisioncache[0] == node: | |
2006 | return (rev, self._revisioncache[2], True) |
|
2022 | return (rev, self._revisioncache[2], True) | |
2007 | cachedrev = self._revisioncache[1] |
|
2023 | cachedrev = self._revisioncache[1] | |
2008 |
|
2024 | |||
2009 | if rev is None: |
|
2025 | if rev is None: | |
2010 | rev = self.rev(node) |
|
2026 | rev = self.rev(node) | |
2011 |
|
2027 | |||
2012 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) |
|
2028 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) | |
2013 | if stopped: |
|
2029 | if stopped: | |
2014 | basetext = self._revisioncache[2] |
|
2030 | basetext = self._revisioncache[2] | |
2015 |
|
2031 | |||
2016 | # drop cache to save memory, the caller is expected to |
|
2032 | # drop cache to save memory, the caller is expected to | |
2017 | # update self._revisioncache after validating the text |
|
2033 | # update self._revisioncache after validating the text | |
2018 | self._revisioncache = None |
|
2034 | self._revisioncache = None | |
2019 |
|
2035 | |||
2020 | targetsize = None |
|
2036 | targetsize = None | |
2021 | rawsize = self.index[rev][2] |
|
2037 | rawsize = self.index[rev][2] | |
2022 | if 0 <= rawsize: |
|
2038 | if 0 <= rawsize: | |
2023 | targetsize = 4 * rawsize |
|
2039 | targetsize = 4 * rawsize | |
2024 |
|
2040 | |||
2025 | bins = self._chunks(chain, df=_df, targetsize=targetsize) |
|
2041 | bins = self._chunks(chain, df=_df, targetsize=targetsize) | |
2026 | if basetext is None: |
|
2042 | if basetext is None: | |
2027 | basetext = bytes(bins[0]) |
|
2043 | basetext = bytes(bins[0]) | |
2028 | bins = bins[1:] |
|
2044 | bins = bins[1:] | |
2029 |
|
2045 | |||
2030 | rawtext = mdiff.patches(basetext, bins) |
|
2046 | rawtext = mdiff.patches(basetext, bins) | |
2031 | del basetext # let us have a chance to free memory early |
|
2047 | del basetext # let us have a chance to free memory early | |
2032 | return (rev, rawtext, False) |
|
2048 | return (rev, rawtext, False) | |
2033 |
|
2049 | |||
2034 | def _sidedata(self, rev): |
|
2050 | def _sidedata(self, rev): | |
2035 | """Return the sidedata for a given revision number.""" |
|
2051 | """Return the sidedata for a given revision number.""" | |
2036 | index_entry = self.index[rev] |
|
2052 | index_entry = self.index[rev] | |
2037 | sidedata_offset = index_entry[8] |
|
2053 | sidedata_offset = index_entry[8] | |
2038 | sidedata_size = index_entry[9] |
|
2054 | sidedata_size = index_entry[9] | |
2039 |
|
2055 | |||
2040 | if self._inline: |
|
2056 | if self._inline: | |
2041 | sidedata_offset += self.index.entry_size * (1 + rev) |
|
2057 | sidedata_offset += self.index.entry_size * (1 + rev) | |
2042 | if sidedata_size == 0: |
|
2058 | if sidedata_size == 0: | |
2043 | return {} |
|
2059 | return {} | |
2044 |
|
2060 | |||
2045 | segment = self._getsegment(sidedata_offset, sidedata_size) |
|
2061 | segment = self._getsegment(sidedata_offset, sidedata_size) | |
2046 | sidedata = sidedatautil.deserialize_sidedata(segment) |
|
2062 | sidedata = sidedatautil.deserialize_sidedata(segment) | |
2047 | return sidedata |
|
2063 | return sidedata | |
2048 |
|
2064 | |||
2049 | def rawdata(self, nodeorrev, _df=None): |
|
2065 | def rawdata(self, nodeorrev, _df=None): | |
2050 | """return an uncompressed raw data of a given node or revision number. |
|
2066 | """return an uncompressed raw data of a given node or revision number. | |
2051 |
|
2067 | |||
2052 | _df - an existing file handle to read from. (internal-only) |
|
2068 | _df - an existing file handle to read from. (internal-only) | |
2053 | """ |
|
2069 | """ | |
2054 | return self._revisiondata(nodeorrev, _df, raw=True)[0] |
|
2070 | return self._revisiondata(nodeorrev, _df, raw=True)[0] | |
2055 |
|
2071 | |||
2056 | def hash(self, text, p1, p2): |
|
2072 | def hash(self, text, p1, p2): | |
2057 | """Compute a node hash. |
|
2073 | """Compute a node hash. | |
2058 |
|
2074 | |||
2059 | Available as a function so that subclasses can replace the hash |
|
2075 | Available as a function so that subclasses can replace the hash | |
2060 | as needed. |
|
2076 | as needed. | |
2061 | """ |
|
2077 | """ | |
2062 | return storageutil.hashrevisionsha1(text, p1, p2) |
|
2078 | return storageutil.hashrevisionsha1(text, p1, p2) | |
2063 |
|
2079 | |||
2064 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
2080 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
2065 | """Check node hash integrity. |
|
2081 | """Check node hash integrity. | |
2066 |
|
2082 | |||
2067 | Available as a function so that subclasses can extend hash mismatch |
|
2083 | Available as a function so that subclasses can extend hash mismatch | |
2068 | behaviors as needed. |
|
2084 | behaviors as needed. | |
2069 | """ |
|
2085 | """ | |
2070 | try: |
|
2086 | try: | |
2071 | if p1 is None and p2 is None: |
|
2087 | if p1 is None and p2 is None: | |
2072 | p1, p2 = self.parents(node) |
|
2088 | p1, p2 = self.parents(node) | |
2073 | if node != self.hash(text, p1, p2): |
|
2089 | if node != self.hash(text, p1, p2): | |
2074 | # Clear the revision cache on hash failure. The revision cache |
|
2090 | # Clear the revision cache on hash failure. The revision cache | |
2075 | # only stores the raw revision and clearing the cache does have |
|
2091 | # only stores the raw revision and clearing the cache does have | |
2076 | # the side-effect that we won't have a cache hit when the raw |
|
2092 | # the side-effect that we won't have a cache hit when the raw | |
2077 | # revision data is accessed. But this case should be rare and |
|
2093 | # revision data is accessed. But this case should be rare and | |
2078 | # it is extra work to teach the cache about the hash |
|
2094 | # it is extra work to teach the cache about the hash | |
2079 | # verification state. |
|
2095 | # verification state. | |
2080 | if self._revisioncache and self._revisioncache[0] == node: |
|
2096 | if self._revisioncache and self._revisioncache[0] == node: | |
2081 | self._revisioncache = None |
|
2097 | self._revisioncache = None | |
2082 |
|
2098 | |||
2083 | revornode = rev |
|
2099 | revornode = rev | |
2084 | if revornode is None: |
|
2100 | if revornode is None: | |
2085 | revornode = templatefilters.short(hex(node)) |
|
2101 | revornode = templatefilters.short(hex(node)) | |
2086 | raise error.RevlogError( |
|
2102 | raise error.RevlogError( | |
2087 | _(b"integrity check failed on %s:%s") |
|
2103 | _(b"integrity check failed on %s:%s") | |
2088 | % (self.display_id, pycompat.bytestr(revornode)) |
|
2104 | % (self.display_id, pycompat.bytestr(revornode)) | |
2089 | ) |
|
2105 | ) | |
2090 | except error.RevlogError: |
|
2106 | except error.RevlogError: | |
2091 | if self._censorable and storageutil.iscensoredtext(text): |
|
2107 | if self._censorable and storageutil.iscensoredtext(text): | |
2092 | raise error.CensoredNodeError(self.display_id, node, text) |
|
2108 | raise error.CensoredNodeError(self.display_id, node, text) | |
2093 | raise |
|
2109 | raise | |
2094 |
|
2110 | |||
2095 | def _enforceinlinesize(self, tr): |
|
2111 | def _enforceinlinesize(self, tr): | |
2096 | """Check if the revlog is too big for inline and convert if so. |
|
2112 | """Check if the revlog is too big for inline and convert if so. | |
2097 |
|
2113 | |||
2098 | This should be called after revisions are added to the revlog. If the |
|
2114 | This should be called after revisions are added to the revlog. If the | |
2099 | revlog has grown too large to be an inline revlog, it will convert it |
|
2115 | revlog has grown too large to be an inline revlog, it will convert it | |
2100 | to use multiple index and data files. |
|
2116 | to use multiple index and data files. | |
2101 | """ |
|
2117 | """ | |
2102 | tiprev = len(self) - 1 |
|
2118 | tiprev = len(self) - 1 | |
2103 | total_size = self.start(tiprev) + self.length(tiprev) |
|
2119 | total_size = self.start(tiprev) + self.length(tiprev) | |
2104 | if not self._inline or total_size < _maxinline: |
|
2120 | if not self._inline or total_size < _maxinline: | |
2105 | return |
|
2121 | return | |
2106 |
|
2122 | |||
2107 | troffset = tr.findoffset(self._indexfile) |
|
2123 | troffset = tr.findoffset(self._indexfile) | |
2108 | if troffset is None: |
|
2124 | if troffset is None: | |
2109 | raise error.RevlogError( |
|
2125 | raise error.RevlogError( | |
2110 | _(b"%s not found in the transaction") % self._indexfile |
|
2126 | _(b"%s not found in the transaction") % self._indexfile | |
2111 | ) |
|
2127 | ) | |
2112 | trindex = 0 |
|
2128 | trindex = 0 | |
2113 | tr.add(self._datafile, 0) |
|
2129 | tr.add(self._datafile, 0) | |
2114 |
|
2130 | |||
2115 | existing_handles = False |
|
2131 | existing_handles = False | |
2116 | if self._writinghandles is not None: |
|
2132 | if self._writinghandles is not None: | |
2117 | existing_handles = True |
|
2133 | existing_handles = True | |
2118 | fp = self._writinghandles[0] |
|
2134 | fp = self._writinghandles[0] | |
2119 | fp.flush() |
|
2135 | fp.flush() | |
2120 | fp.close() |
|
2136 | fp.close() | |
2121 | # We can't use the cached file handle after close(). So prevent |
|
2137 | # We can't use the cached file handle after close(). So prevent | |
2122 | # its usage. |
|
2138 | # its usage. | |
2123 | self._writinghandles = None |
|
2139 | self._writinghandles = None | |
2124 |
|
2140 | |||
2125 | new_dfh = self._datafp(b'w+') |
|
2141 | new_dfh = self._datafp(b'w+') | |
2126 | new_dfh.truncate(0) # drop any potentially existing data |
|
2142 | new_dfh.truncate(0) # drop any potentially existing data | |
2127 | try: |
|
2143 | try: | |
2128 | with self._indexfp() as read_ifh: |
|
2144 | with self._indexfp() as read_ifh: | |
2129 | for r in self: |
|
2145 | for r in self: | |
2130 | new_dfh.write(self._getsegmentforrevs(r, r, df=read_ifh)[1]) |
|
2146 | new_dfh.write(self._getsegmentforrevs(r, r, df=read_ifh)[1]) | |
2131 | if troffset <= self.start(r): |
|
2147 | if troffset <= self.start(r): | |
2132 | trindex = r |
|
2148 | trindex = r | |
2133 | new_dfh.flush() |
|
2149 | new_dfh.flush() | |
2134 |
|
2150 | |||
2135 | with self.__index_new_fp() as fp: |
|
2151 | with self.__index_new_fp() as fp: | |
2136 | self._format_flags &= ~FLAG_INLINE_DATA |
|
2152 | self._format_flags &= ~FLAG_INLINE_DATA | |
2137 | self._inline = False |
|
2153 | self._inline = False | |
2138 | for i in self: |
|
2154 | for i in self: | |
2139 | e = self.index.entry_binary(i) |
|
2155 | e = self.index.entry_binary(i) | |
2140 | if i == 0 and self._docket is None: |
|
2156 | if i == 0 and self._docket is None: | |
2141 | header = self._format_flags | self._format_version |
|
2157 | header = self._format_flags | self._format_version | |
2142 | header = self.index.pack_header(header) |
|
2158 | header = self.index.pack_header(header) | |
2143 | e = header + e |
|
2159 | e = header + e | |
2144 | fp.write(e) |
|
2160 | fp.write(e) | |
2145 | if self._docket is not None: |
|
2161 | if self._docket is not None: | |
2146 | self._docket.index_end = fp.tell() |
|
2162 | self._docket.index_end = fp.tell() | |
2147 | # the temp file replace the real index when we exit the context |
|
2163 | # the temp file replace the real index when we exit the context | |
2148 | # manager |
|
2164 | # manager | |
2149 |
|
2165 | |||
2150 | tr.replace(self._indexfile, trindex * self.index.entry_size) |
|
2166 | tr.replace(self._indexfile, trindex * self.index.entry_size) | |
2151 | nodemaputil.setup_persistent_nodemap(tr, self) |
|
2167 | nodemaputil.setup_persistent_nodemap(tr, self) | |
2152 | self._chunkclear() |
|
2168 | self._chunkclear() | |
2153 |
|
2169 | |||
2154 | if existing_handles: |
|
2170 | if existing_handles: | |
2155 | # switched from inline to conventional reopen the index |
|
2171 | # switched from inline to conventional reopen the index | |
2156 | ifh = self.__index_write_fp() |
|
2172 | ifh = self.__index_write_fp() | |
2157 | self._writinghandles = (ifh, new_dfh) |
|
2173 | self._writinghandles = (ifh, new_dfh) | |
2158 | new_dfh = None |
|
2174 | new_dfh = None | |
2159 | finally: |
|
2175 | finally: | |
2160 | if new_dfh is not None: |
|
2176 | if new_dfh is not None: | |
2161 | new_dfh.close() |
|
2177 | new_dfh.close() | |
2162 |
|
2178 | |||
2163 | def _nodeduplicatecallback(self, transaction, node): |
|
2179 | def _nodeduplicatecallback(self, transaction, node): | |
2164 | """called when trying to add a node already stored.""" |
|
2180 | """called when trying to add a node already stored.""" | |
2165 |
|
2181 | |||
2166 | @contextlib.contextmanager |
|
2182 | @contextlib.contextmanager | |
2167 | def _writing(self, transaction): |
|
2183 | def _writing(self, transaction): | |
2168 | if self._trypending: |
|
2184 | if self._trypending: | |
2169 | msg = b'try to write in a `trypending` revlog: %s' |
|
2185 | msg = b'try to write in a `trypending` revlog: %s' | |
2170 | msg %= self.display_id |
|
2186 | msg %= self.display_id | |
2171 | raise error.ProgrammingError(msg) |
|
2187 | raise error.ProgrammingError(msg) | |
2172 | if self._writinghandles is not None: |
|
2188 | if self._writinghandles is not None: | |
2173 | yield |
|
2189 | yield | |
2174 | else: |
|
2190 | else: | |
2175 | r = len(self) |
|
2191 | r = len(self) | |
2176 | dsize = 0 |
|
2192 | dsize = 0 | |
2177 | if r: |
|
2193 | if r: | |
2178 | dsize = self.end(r - 1) |
|
2194 | dsize = self.end(r - 1) | |
2179 | dfh = None |
|
2195 | dfh = None | |
2180 | if not self._inline: |
|
2196 | if not self._inline: | |
2181 | try: |
|
2197 | try: | |
2182 | dfh = self._datafp(b"r+") |
|
2198 | dfh = self._datafp(b"r+") | |
2183 | if self._docket is None: |
|
2199 | if self._docket is None: | |
2184 | dfh.seek(0, os.SEEK_END) |
|
2200 | dfh.seek(0, os.SEEK_END) | |
2185 | else: |
|
2201 | else: | |
2186 | dfh.seek(self._docket.data_end, os.SEEK_SET) |
|
2202 | dfh.seek(self._docket.data_end, os.SEEK_SET) | |
2187 | except IOError as inst: |
|
2203 | except IOError as inst: | |
2188 | if inst.errno != errno.ENOENT: |
|
2204 | if inst.errno != errno.ENOENT: | |
2189 | raise |
|
2205 | raise | |
2190 | dfh = self._datafp(b"w+") |
|
2206 | dfh = self._datafp(b"w+") | |
2191 | transaction.add(self._datafile, dsize) |
|
2207 | transaction.add(self._datafile, dsize) | |
2192 | try: |
|
2208 | try: | |
2193 | isize = r * self.index.entry_size |
|
2209 | isize = r * self.index.entry_size | |
2194 | ifh = self.__index_write_fp() |
|
2210 | ifh = self.__index_write_fp() | |
2195 | if self._inline: |
|
2211 | if self._inline: | |
2196 | transaction.add(self._indexfile, dsize + isize) |
|
2212 | transaction.add(self._indexfile, dsize + isize) | |
2197 | else: |
|
2213 | else: | |
2198 | transaction.add(self._indexfile, isize) |
|
2214 | transaction.add(self._indexfile, isize) | |
2199 | try: |
|
2215 | try: | |
2200 | self._writinghandles = (ifh, dfh) |
|
2216 | self._writinghandles = (ifh, dfh) | |
2201 | try: |
|
2217 | try: | |
2202 | yield |
|
2218 | yield | |
2203 | if self._docket is not None: |
|
2219 | if self._docket is not None: | |
2204 | self._write_docket(transaction) |
|
2220 | self._write_docket(transaction) | |
2205 | finally: |
|
2221 | finally: | |
2206 | self._writinghandles = None |
|
2222 | self._writinghandles = None | |
2207 | finally: |
|
2223 | finally: | |
2208 | ifh.close() |
|
2224 | ifh.close() | |
2209 | finally: |
|
2225 | finally: | |
2210 | if dfh is not None: |
|
2226 | if dfh is not None: | |
2211 | dfh.close() |
|
2227 | dfh.close() | |
2212 |
|
2228 | |||
2213 | def _write_docket(self, transaction): |
|
2229 | def _write_docket(self, transaction): | |
2214 | """write the current docket on disk |
|
2230 | """write the current docket on disk | |
2215 |
|
2231 | |||
2216 | Exist as a method to help changelog to implement transaction logic |
|
2232 | Exist as a method to help changelog to implement transaction logic | |
2217 |
|
2233 | |||
2218 | We could also imagine using the same transaction logic for all revlog |
|
2234 | We could also imagine using the same transaction logic for all revlog | |
2219 | since docket are cheap.""" |
|
2235 | since docket are cheap.""" | |
2220 | self._docket.write(transaction) |
|
2236 | self._docket.write(transaction) | |
2221 |
|
2237 | |||
2222 | def addrevision( |
|
2238 | def addrevision( | |
2223 | self, |
|
2239 | self, | |
2224 | text, |
|
2240 | text, | |
2225 | transaction, |
|
2241 | transaction, | |
2226 | link, |
|
2242 | link, | |
2227 | p1, |
|
2243 | p1, | |
2228 | p2, |
|
2244 | p2, | |
2229 | cachedelta=None, |
|
2245 | cachedelta=None, | |
2230 | node=None, |
|
2246 | node=None, | |
2231 | flags=REVIDX_DEFAULT_FLAGS, |
|
2247 | flags=REVIDX_DEFAULT_FLAGS, | |
2232 | deltacomputer=None, |
|
2248 | deltacomputer=None, | |
2233 | sidedata=None, |
|
2249 | sidedata=None, | |
2234 | ): |
|
2250 | ): | |
2235 | """add a revision to the log |
|
2251 | """add a revision to the log | |
2236 |
|
2252 | |||
2237 | text - the revision data to add |
|
2253 | text - the revision data to add | |
2238 | transaction - the transaction object used for rollback |
|
2254 | transaction - the transaction object used for rollback | |
2239 | link - the linkrev data to add |
|
2255 | link - the linkrev data to add | |
2240 | p1, p2 - the parent nodeids of the revision |
|
2256 | p1, p2 - the parent nodeids of the revision | |
2241 | cachedelta - an optional precomputed delta |
|
2257 | cachedelta - an optional precomputed delta | |
2242 | node - nodeid of revision; typically node is not specified, and it is |
|
2258 | node - nodeid of revision; typically node is not specified, and it is | |
2243 | computed by default as hash(text, p1, p2), however subclasses might |
|
2259 | computed by default as hash(text, p1, p2), however subclasses might | |
2244 | use different hashing method (and override checkhash() in such case) |
|
2260 | use different hashing method (and override checkhash() in such case) | |
2245 | flags - the known flags to set on the revision |
|
2261 | flags - the known flags to set on the revision | |
2246 | deltacomputer - an optional deltacomputer instance shared between |
|
2262 | deltacomputer - an optional deltacomputer instance shared between | |
2247 | multiple calls |
|
2263 | multiple calls | |
2248 | """ |
|
2264 | """ | |
2249 | if link == nullrev: |
|
2265 | if link == nullrev: | |
2250 | raise error.RevlogError( |
|
2266 | raise error.RevlogError( | |
2251 | _(b"attempted to add linkrev -1 to %s") % self.display_id |
|
2267 | _(b"attempted to add linkrev -1 to %s") % self.display_id | |
2252 | ) |
|
2268 | ) | |
2253 |
|
2269 | |||
2254 | if sidedata is None: |
|
2270 | if sidedata is None: | |
2255 | sidedata = {} |
|
2271 | sidedata = {} | |
2256 | elif sidedata and not self.hassidedata: |
|
2272 | elif sidedata and not self.hassidedata: | |
2257 | raise error.ProgrammingError( |
|
2273 | raise error.ProgrammingError( | |
2258 | _(b"trying to add sidedata to a revlog who don't support them") |
|
2274 | _(b"trying to add sidedata to a revlog who don't support them") | |
2259 | ) |
|
2275 | ) | |
2260 |
|
2276 | |||
2261 | if flags: |
|
2277 | if flags: | |
2262 | node = node or self.hash(text, p1, p2) |
|
2278 | node = node or self.hash(text, p1, p2) | |
2263 |
|
2279 | |||
2264 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) |
|
2280 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) | |
2265 |
|
2281 | |||
2266 | # If the flag processor modifies the revision data, ignore any provided |
|
2282 | # If the flag processor modifies the revision data, ignore any provided | |
2267 | # cachedelta. |
|
2283 | # cachedelta. | |
2268 | if rawtext != text: |
|
2284 | if rawtext != text: | |
2269 | cachedelta = None |
|
2285 | cachedelta = None | |
2270 |
|
2286 | |||
2271 | if len(rawtext) > _maxentrysize: |
|
2287 | if len(rawtext) > _maxentrysize: | |
2272 | raise error.RevlogError( |
|
2288 | raise error.RevlogError( | |
2273 | _( |
|
2289 | _( | |
2274 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" |
|
2290 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" | |
2275 | ) |
|
2291 | ) | |
2276 | % (self.display_id, len(rawtext)) |
|
2292 | % (self.display_id, len(rawtext)) | |
2277 | ) |
|
2293 | ) | |
2278 |
|
2294 | |||
2279 | node = node or self.hash(rawtext, p1, p2) |
|
2295 | node = node or self.hash(rawtext, p1, p2) | |
2280 | rev = self.index.get_rev(node) |
|
2296 | rev = self.index.get_rev(node) | |
2281 | if rev is not None: |
|
2297 | if rev is not None: | |
2282 | return rev |
|
2298 | return rev | |
2283 |
|
2299 | |||
2284 | if validatehash: |
|
2300 | if validatehash: | |
2285 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
2301 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
2286 |
|
2302 | |||
2287 | return self.addrawrevision( |
|
2303 | return self.addrawrevision( | |
2288 | rawtext, |
|
2304 | rawtext, | |
2289 | transaction, |
|
2305 | transaction, | |
2290 | link, |
|
2306 | link, | |
2291 | p1, |
|
2307 | p1, | |
2292 | p2, |
|
2308 | p2, | |
2293 | node, |
|
2309 | node, | |
2294 | flags, |
|
2310 | flags, | |
2295 | cachedelta=cachedelta, |
|
2311 | cachedelta=cachedelta, | |
2296 | deltacomputer=deltacomputer, |
|
2312 | deltacomputer=deltacomputer, | |
2297 | sidedata=sidedata, |
|
2313 | sidedata=sidedata, | |
2298 | ) |
|
2314 | ) | |
2299 |
|
2315 | |||
2300 | def addrawrevision( |
|
2316 | def addrawrevision( | |
2301 | self, |
|
2317 | self, | |
2302 | rawtext, |
|
2318 | rawtext, | |
2303 | transaction, |
|
2319 | transaction, | |
2304 | link, |
|
2320 | link, | |
2305 | p1, |
|
2321 | p1, | |
2306 | p2, |
|
2322 | p2, | |
2307 | node, |
|
2323 | node, | |
2308 | flags, |
|
2324 | flags, | |
2309 | cachedelta=None, |
|
2325 | cachedelta=None, | |
2310 | deltacomputer=None, |
|
2326 | deltacomputer=None, | |
2311 | sidedata=None, |
|
2327 | sidedata=None, | |
2312 | ): |
|
2328 | ): | |
2313 | """add a raw revision with known flags, node and parents |
|
2329 | """add a raw revision with known flags, node and parents | |
2314 | useful when reusing a revision not stored in this revlog (ex: received |
|
2330 | useful when reusing a revision not stored in this revlog (ex: received | |
2315 | over wire, or read from an external bundle). |
|
2331 | over wire, or read from an external bundle). | |
2316 | """ |
|
2332 | """ | |
2317 | with self._writing(transaction): |
|
2333 | with self._writing(transaction): | |
2318 | return self._addrevision( |
|
2334 | return self._addrevision( | |
2319 | node, |
|
2335 | node, | |
2320 | rawtext, |
|
2336 | rawtext, | |
2321 | transaction, |
|
2337 | transaction, | |
2322 | link, |
|
2338 | link, | |
2323 | p1, |
|
2339 | p1, | |
2324 | p2, |
|
2340 | p2, | |
2325 | flags, |
|
2341 | flags, | |
2326 | cachedelta, |
|
2342 | cachedelta, | |
2327 | deltacomputer=deltacomputer, |
|
2343 | deltacomputer=deltacomputer, | |
2328 | sidedata=sidedata, |
|
2344 | sidedata=sidedata, | |
2329 | ) |
|
2345 | ) | |
2330 |
|
2346 | |||
2331 | def compress(self, data): |
|
2347 | def compress(self, data): | |
2332 | """Generate a possibly-compressed representation of data.""" |
|
2348 | """Generate a possibly-compressed representation of data.""" | |
2333 | if not data: |
|
2349 | if not data: | |
2334 | return b'', data |
|
2350 | return b'', data | |
2335 |
|
2351 | |||
2336 | compressed = self._compressor.compress(data) |
|
2352 | compressed = self._compressor.compress(data) | |
2337 |
|
2353 | |||
2338 | if compressed: |
|
2354 | if compressed: | |
2339 | # The revlog compressor added the header in the returned data. |
|
2355 | # The revlog compressor added the header in the returned data. | |
2340 | return b'', compressed |
|
2356 | return b'', compressed | |
2341 |
|
2357 | |||
2342 | if data[0:1] == b'\0': |
|
2358 | if data[0:1] == b'\0': | |
2343 | return b'', data |
|
2359 | return b'', data | |
2344 | return b'u', data |
|
2360 | return b'u', data | |
2345 |
|
2361 | |||
2346 | def decompress(self, data): |
|
2362 | def decompress(self, data): | |
2347 | """Decompress a revlog chunk. |
|
2363 | """Decompress a revlog chunk. | |
2348 |
|
2364 | |||
2349 | The chunk is expected to begin with a header identifying the |
|
2365 | The chunk is expected to begin with a header identifying the | |
2350 | format type so it can be routed to an appropriate decompressor. |
|
2366 | format type so it can be routed to an appropriate decompressor. | |
2351 | """ |
|
2367 | """ | |
2352 | if not data: |
|
2368 | if not data: | |
2353 | return data |
|
2369 | return data | |
2354 |
|
2370 | |||
2355 | # Revlogs are read much more frequently than they are written and many |
|
2371 | # Revlogs are read much more frequently than they are written and many | |
2356 | # chunks only take microseconds to decompress, so performance is |
|
2372 | # chunks only take microseconds to decompress, so performance is | |
2357 | # important here. |
|
2373 | # important here. | |
2358 | # |
|
2374 | # | |
2359 | # We can make a few assumptions about revlogs: |
|
2375 | # We can make a few assumptions about revlogs: | |
2360 | # |
|
2376 | # | |
2361 | # 1) the majority of chunks will be compressed (as opposed to inline |
|
2377 | # 1) the majority of chunks will be compressed (as opposed to inline | |
2362 | # raw data). |
|
2378 | # raw data). | |
2363 | # 2) decompressing *any* data will likely by at least 10x slower than |
|
2379 | # 2) decompressing *any* data will likely by at least 10x slower than | |
2364 | # returning raw inline data. |
|
2380 | # returning raw inline data. | |
2365 | # 3) we want to prioritize common and officially supported compression |
|
2381 | # 3) we want to prioritize common and officially supported compression | |
2366 | # engines |
|
2382 | # engines | |
2367 | # |
|
2383 | # | |
2368 | # It follows that we want to optimize for "decompress compressed data |
|
2384 | # It follows that we want to optimize for "decompress compressed data | |
2369 | # when encoded with common and officially supported compression engines" |
|
2385 | # when encoded with common and officially supported compression engines" | |
2370 | # case over "raw data" and "data encoded by less common or non-official |
|
2386 | # case over "raw data" and "data encoded by less common or non-official | |
2371 | # compression engines." That is why we have the inline lookup first |
|
2387 | # compression engines." That is why we have the inline lookup first | |
2372 | # followed by the compengines lookup. |
|
2388 | # followed by the compengines lookup. | |
2373 | # |
|
2389 | # | |
2374 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib |
|
2390 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib | |
2375 | # compressed chunks. And this matters for changelog and manifest reads. |
|
2391 | # compressed chunks. And this matters for changelog and manifest reads. | |
2376 | t = data[0:1] |
|
2392 | t = data[0:1] | |
2377 |
|
2393 | |||
2378 | if t == b'x': |
|
2394 | if t == b'x': | |
2379 | try: |
|
2395 | try: | |
2380 | return _zlibdecompress(data) |
|
2396 | return _zlibdecompress(data) | |
2381 | except zlib.error as e: |
|
2397 | except zlib.error as e: | |
2382 | raise error.RevlogError( |
|
2398 | raise error.RevlogError( | |
2383 | _(b'revlog decompress error: %s') |
|
2399 | _(b'revlog decompress error: %s') | |
2384 | % stringutil.forcebytestr(e) |
|
2400 | % stringutil.forcebytestr(e) | |
2385 | ) |
|
2401 | ) | |
2386 | # '\0' is more common than 'u' so it goes first. |
|
2402 | # '\0' is more common than 'u' so it goes first. | |
2387 | elif t == b'\0': |
|
2403 | elif t == b'\0': | |
2388 | return data |
|
2404 | return data | |
2389 | elif t == b'u': |
|
2405 | elif t == b'u': | |
2390 | return util.buffer(data, 1) |
|
2406 | return util.buffer(data, 1) | |
2391 |
|
2407 | |||
2392 | compressor = self._get_decompressor(t) |
|
2408 | compressor = self._get_decompressor(t) | |
2393 |
|
2409 | |||
2394 | return compressor.decompress(data) |
|
2410 | return compressor.decompress(data) | |
2395 |
|
2411 | |||
2396 | def _addrevision( |
|
2412 | def _addrevision( | |
2397 | self, |
|
2413 | self, | |
2398 | node, |
|
2414 | node, | |
2399 | rawtext, |
|
2415 | rawtext, | |
2400 | transaction, |
|
2416 | transaction, | |
2401 | link, |
|
2417 | link, | |
2402 | p1, |
|
2418 | p1, | |
2403 | p2, |
|
2419 | p2, | |
2404 | flags, |
|
2420 | flags, | |
2405 | cachedelta, |
|
2421 | cachedelta, | |
2406 | alwayscache=False, |
|
2422 | alwayscache=False, | |
2407 | deltacomputer=None, |
|
2423 | deltacomputer=None, | |
2408 | sidedata=None, |
|
2424 | sidedata=None, | |
2409 | ): |
|
2425 | ): | |
2410 | """internal function to add revisions to the log |
|
2426 | """internal function to add revisions to the log | |
2411 |
|
2427 | |||
2412 | see addrevision for argument descriptions. |
|
2428 | see addrevision for argument descriptions. | |
2413 |
|
2429 | |||
2414 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. |
|
2430 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. | |
2415 |
|
2431 | |||
2416 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will |
|
2432 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will | |
2417 | be used. |
|
2433 | be used. | |
2418 |
|
2434 | |||
2419 | invariants: |
|
2435 | invariants: | |
2420 | - rawtext is optional (can be None); if not set, cachedelta must be set. |
|
2436 | - rawtext is optional (can be None); if not set, cachedelta must be set. | |
2421 | if both are set, they must correspond to each other. |
|
2437 | if both are set, they must correspond to each other. | |
2422 | """ |
|
2438 | """ | |
2423 | if node == self.nullid: |
|
2439 | if node == self.nullid: | |
2424 | raise error.RevlogError( |
|
2440 | raise error.RevlogError( | |
2425 | _(b"%s: attempt to add null revision") % self.display_id |
|
2441 | _(b"%s: attempt to add null revision") % self.display_id | |
2426 | ) |
|
2442 | ) | |
2427 | if ( |
|
2443 | if ( | |
2428 | node == self.nodeconstants.wdirid |
|
2444 | node == self.nodeconstants.wdirid | |
2429 | or node in self.nodeconstants.wdirfilenodeids |
|
2445 | or node in self.nodeconstants.wdirfilenodeids | |
2430 | ): |
|
2446 | ): | |
2431 | raise error.RevlogError( |
|
2447 | raise error.RevlogError( | |
2432 | _(b"%s: attempt to add wdir revision") % self.display_id |
|
2448 | _(b"%s: attempt to add wdir revision") % self.display_id | |
2433 | ) |
|
2449 | ) | |
2434 | if self._writinghandles is None: |
|
2450 | if self._writinghandles is None: | |
2435 | msg = b'adding revision outside `revlog._writing` context' |
|
2451 | msg = b'adding revision outside `revlog._writing` context' | |
2436 | raise error.ProgrammingError(msg) |
|
2452 | raise error.ProgrammingError(msg) | |
2437 |
|
2453 | |||
2438 | if self._inline: |
|
2454 | if self._inline: | |
2439 | fh = self._writinghandles[0] |
|
2455 | fh = self._writinghandles[0] | |
2440 | else: |
|
2456 | else: | |
2441 | fh = self._writinghandles[1] |
|
2457 | fh = self._writinghandles[1] | |
2442 |
|
2458 | |||
2443 | btext = [rawtext] |
|
2459 | btext = [rawtext] | |
2444 |
|
2460 | |||
2445 | curr = len(self) |
|
2461 | curr = len(self) | |
2446 | prev = curr - 1 |
|
2462 | prev = curr - 1 | |
2447 |
|
2463 | |||
2448 | offset = self._get_data_offset(prev) |
|
2464 | offset = self._get_data_offset(prev) | |
2449 |
|
2465 | |||
2450 | if self._concurrencychecker: |
|
2466 | if self._concurrencychecker: | |
2451 | ifh, dfh = self._writinghandles |
|
2467 | ifh, dfh = self._writinghandles | |
2452 | if self._inline: |
|
2468 | if self._inline: | |
2453 | # offset is "as if" it were in the .d file, so we need to add on |
|
2469 | # offset is "as if" it were in the .d file, so we need to add on | |
2454 | # the size of the entry metadata. |
|
2470 | # the size of the entry metadata. | |
2455 | self._concurrencychecker( |
|
2471 | self._concurrencychecker( | |
2456 | ifh, self._indexfile, offset + curr * self.index.entry_size |
|
2472 | ifh, self._indexfile, offset + curr * self.index.entry_size | |
2457 | ) |
|
2473 | ) | |
2458 | else: |
|
2474 | else: | |
2459 | # Entries in the .i are a consistent size. |
|
2475 | # Entries in the .i are a consistent size. | |
2460 | self._concurrencychecker( |
|
2476 | self._concurrencychecker( | |
2461 | ifh, self._indexfile, curr * self.index.entry_size |
|
2477 | ifh, self._indexfile, curr * self.index.entry_size | |
2462 | ) |
|
2478 | ) | |
2463 | self._concurrencychecker(dfh, self._datafile, offset) |
|
2479 | self._concurrencychecker(dfh, self._datafile, offset) | |
2464 |
|
2480 | |||
2465 | p1r, p2r = self.rev(p1), self.rev(p2) |
|
2481 | p1r, p2r = self.rev(p1), self.rev(p2) | |
2466 |
|
2482 | |||
2467 | # full versions are inserted when the needed deltas |
|
2483 | # full versions are inserted when the needed deltas | |
2468 | # become comparable to the uncompressed text |
|
2484 | # become comparable to the uncompressed text | |
2469 | if rawtext is None: |
|
2485 | if rawtext is None: | |
2470 | # need rawtext size, before changed by flag processors, which is |
|
2486 | # need rawtext size, before changed by flag processors, which is | |
2471 | # the non-raw size. use revlog explicitly to avoid filelog's extra |
|
2487 | # the non-raw size. use revlog explicitly to avoid filelog's extra | |
2472 | # logic that might remove metadata size. |
|
2488 | # logic that might remove metadata size. | |
2473 | textlen = mdiff.patchedsize( |
|
2489 | textlen = mdiff.patchedsize( | |
2474 | revlog.size(self, cachedelta[0]), cachedelta[1] |
|
2490 | revlog.size(self, cachedelta[0]), cachedelta[1] | |
2475 | ) |
|
2491 | ) | |
2476 | else: |
|
2492 | else: | |
2477 | textlen = len(rawtext) |
|
2493 | textlen = len(rawtext) | |
2478 |
|
2494 | |||
2479 | if deltacomputer is None: |
|
2495 | if deltacomputer is None: | |
2480 | deltacomputer = deltautil.deltacomputer(self) |
|
2496 | deltacomputer = deltautil.deltacomputer(self) | |
2481 |
|
2497 | |||
2482 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) |
|
2498 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) | |
2483 |
|
2499 | |||
2484 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) |
|
2500 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) | |
2485 |
|
2501 | |||
2486 | compression_mode = COMP_MODE_INLINE |
|
2502 | compression_mode = COMP_MODE_INLINE | |
2487 | if self._docket is not None: |
|
2503 | if self._docket is not None: | |
2488 | h, d = deltainfo.data |
|
2504 | h, d = deltainfo.data | |
2489 | if not h and not d: |
|
2505 | if not h and not d: | |
2490 | # not data to store at all... declare them uncompressed |
|
2506 | # not data to store at all... declare them uncompressed | |
2491 | compression_mode = COMP_MODE_PLAIN |
|
2507 | compression_mode = COMP_MODE_PLAIN | |
2492 |
elif not h |
|
2508 | elif not h: | |
2493 | compression_mode = COMP_MODE_PLAIN |
|
2509 | t = d[0:1] | |
|
2510 | if t == b'\0': | |||
|
2511 | compression_mode = COMP_MODE_PLAIN | |||
|
2512 | elif t == self._docket.default_compression_header: | |||
|
2513 | compression_mode = COMP_MODE_DEFAULT | |||
2494 | elif h == b'u': |
|
2514 | elif h == b'u': | |
2495 | # we have a more efficient way to declare uncompressed |
|
2515 | # we have a more efficient way to declare uncompressed | |
2496 | h = b'' |
|
2516 | h = b'' | |
2497 | compression_mode = COMP_MODE_PLAIN |
|
2517 | compression_mode = COMP_MODE_PLAIN | |
2498 | deltainfo = deltautil.drop_u_compression(deltainfo) |
|
2518 | deltainfo = deltautil.drop_u_compression(deltainfo) | |
2499 |
|
2519 | |||
2500 | if sidedata and self.hassidedata: |
|
2520 | if sidedata and self.hassidedata: | |
2501 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) |
|
2521 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) | |
2502 | sidedata_offset = offset + deltainfo.deltalen |
|
2522 | sidedata_offset = offset + deltainfo.deltalen | |
2503 | else: |
|
2523 | else: | |
2504 | serialized_sidedata = b"" |
|
2524 | serialized_sidedata = b"" | |
2505 | # Don't store the offset if the sidedata is empty, that way |
|
2525 | # Don't store the offset if the sidedata is empty, that way | |
2506 | # we can easily detect empty sidedata and they will be no different |
|
2526 | # we can easily detect empty sidedata and they will be no different | |
2507 | # than ones we manually add. |
|
2527 | # than ones we manually add. | |
2508 | sidedata_offset = 0 |
|
2528 | sidedata_offset = 0 | |
2509 |
|
2529 | |||
2510 | e = ( |
|
2530 | e = ( | |
2511 | offset_type(offset, flags), |
|
2531 | offset_type(offset, flags), | |
2512 | deltainfo.deltalen, |
|
2532 | deltainfo.deltalen, | |
2513 | textlen, |
|
2533 | textlen, | |
2514 | deltainfo.base, |
|
2534 | deltainfo.base, | |
2515 | link, |
|
2535 | link, | |
2516 | p1r, |
|
2536 | p1r, | |
2517 | p2r, |
|
2537 | p2r, | |
2518 | node, |
|
2538 | node, | |
2519 | sidedata_offset, |
|
2539 | sidedata_offset, | |
2520 | len(serialized_sidedata), |
|
2540 | len(serialized_sidedata), | |
2521 | compression_mode, |
|
2541 | compression_mode, | |
2522 | ) |
|
2542 | ) | |
2523 |
|
2543 | |||
2524 | self.index.append(e) |
|
2544 | self.index.append(e) | |
2525 | entry = self.index.entry_binary(curr) |
|
2545 | entry = self.index.entry_binary(curr) | |
2526 | if curr == 0 and self._docket is None: |
|
2546 | if curr == 0 and self._docket is None: | |
2527 | header = self._format_flags | self._format_version |
|
2547 | header = self._format_flags | self._format_version | |
2528 | header = self.index.pack_header(header) |
|
2548 | header = self.index.pack_header(header) | |
2529 | entry = header + entry |
|
2549 | entry = header + entry | |
2530 | self._writeentry( |
|
2550 | self._writeentry( | |
2531 | transaction, |
|
2551 | transaction, | |
2532 | entry, |
|
2552 | entry, | |
2533 | deltainfo.data, |
|
2553 | deltainfo.data, | |
2534 | link, |
|
2554 | link, | |
2535 | offset, |
|
2555 | offset, | |
2536 | serialized_sidedata, |
|
2556 | serialized_sidedata, | |
2537 | ) |
|
2557 | ) | |
2538 |
|
2558 | |||
2539 | rawtext = btext[0] |
|
2559 | rawtext = btext[0] | |
2540 |
|
2560 | |||
2541 | if alwayscache and rawtext is None: |
|
2561 | if alwayscache and rawtext is None: | |
2542 | rawtext = deltacomputer.buildtext(revinfo, fh) |
|
2562 | rawtext = deltacomputer.buildtext(revinfo, fh) | |
2543 |
|
2563 | |||
2544 | if type(rawtext) == bytes: # only accept immutable objects |
|
2564 | if type(rawtext) == bytes: # only accept immutable objects | |
2545 | self._revisioncache = (node, curr, rawtext) |
|
2565 | self._revisioncache = (node, curr, rawtext) | |
2546 | self._chainbasecache[curr] = deltainfo.chainbase |
|
2566 | self._chainbasecache[curr] = deltainfo.chainbase | |
2547 | return curr |
|
2567 | return curr | |
2548 |
|
2568 | |||
2549 | def _get_data_offset(self, prev): |
|
2569 | def _get_data_offset(self, prev): | |
2550 | """Returns the current offset in the (in-transaction) data file. |
|
2570 | """Returns the current offset in the (in-transaction) data file. | |
2551 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket |
|
2571 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket | |
2552 | file to store that information: since sidedata can be rewritten to the |
|
2572 | file to store that information: since sidedata can be rewritten to the | |
2553 | end of the data file within a transaction, you can have cases where, for |
|
2573 | end of the data file within a transaction, you can have cases where, for | |
2554 | example, rev `n` does not have sidedata while rev `n - 1` does, leading |
|
2574 | example, rev `n` does not have sidedata while rev `n - 1` does, leading | |
2555 | to `n - 1`'s sidedata being written after `n`'s data. |
|
2575 | to `n - 1`'s sidedata being written after `n`'s data. | |
2556 |
|
2576 | |||
2557 | TODO cache this in a docket file before getting out of experimental.""" |
|
2577 | TODO cache this in a docket file before getting out of experimental.""" | |
2558 | if self._docket is None: |
|
2578 | if self._docket is None: | |
2559 | return self.end(prev) |
|
2579 | return self.end(prev) | |
2560 | else: |
|
2580 | else: | |
2561 | return self._docket.data_end |
|
2581 | return self._docket.data_end | |
2562 |
|
2582 | |||
2563 | def _writeentry(self, transaction, entry, data, link, offset, sidedata): |
|
2583 | def _writeentry(self, transaction, entry, data, link, offset, sidedata): | |
2564 | # Files opened in a+ mode have inconsistent behavior on various |
|
2584 | # Files opened in a+ mode have inconsistent behavior on various | |
2565 | # platforms. Windows requires that a file positioning call be made |
|
2585 | # platforms. Windows requires that a file positioning call be made | |
2566 | # when the file handle transitions between reads and writes. See |
|
2586 | # when the file handle transitions between reads and writes. See | |
2567 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
2587 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other | |
2568 | # platforms, Python or the platform itself can be buggy. Some versions |
|
2588 | # platforms, Python or the platform itself can be buggy. Some versions | |
2569 | # of Solaris have been observed to not append at the end of the file |
|
2589 | # of Solaris have been observed to not append at the end of the file | |
2570 | # if the file was seeked to before the end. See issue4943 for more. |
|
2590 | # if the file was seeked to before the end. See issue4943 for more. | |
2571 | # |
|
2591 | # | |
2572 | # We work around this issue by inserting a seek() before writing. |
|
2592 | # We work around this issue by inserting a seek() before writing. | |
2573 | # Note: This is likely not necessary on Python 3. However, because |
|
2593 | # Note: This is likely not necessary on Python 3. However, because | |
2574 | # the file handle is reused for reads and may be seeked there, we need |
|
2594 | # the file handle is reused for reads and may be seeked there, we need | |
2575 | # to be careful before changing this. |
|
2595 | # to be careful before changing this. | |
2576 | if self._writinghandles is None: |
|
2596 | if self._writinghandles is None: | |
2577 | msg = b'adding revision outside `revlog._writing` context' |
|
2597 | msg = b'adding revision outside `revlog._writing` context' | |
2578 | raise error.ProgrammingError(msg) |
|
2598 | raise error.ProgrammingError(msg) | |
2579 | ifh, dfh = self._writinghandles |
|
2599 | ifh, dfh = self._writinghandles | |
2580 | if self._docket is None: |
|
2600 | if self._docket is None: | |
2581 | ifh.seek(0, os.SEEK_END) |
|
2601 | ifh.seek(0, os.SEEK_END) | |
2582 | else: |
|
2602 | else: | |
2583 | ifh.seek(self._docket.index_end, os.SEEK_SET) |
|
2603 | ifh.seek(self._docket.index_end, os.SEEK_SET) | |
2584 | if dfh: |
|
2604 | if dfh: | |
2585 | if self._docket is None: |
|
2605 | if self._docket is None: | |
2586 | dfh.seek(0, os.SEEK_END) |
|
2606 | dfh.seek(0, os.SEEK_END) | |
2587 | else: |
|
2607 | else: | |
2588 | dfh.seek(self._docket.data_end, os.SEEK_SET) |
|
2608 | dfh.seek(self._docket.data_end, os.SEEK_SET) | |
2589 |
|
2609 | |||
2590 | curr = len(self) - 1 |
|
2610 | curr = len(self) - 1 | |
2591 | if not self._inline: |
|
2611 | if not self._inline: | |
2592 | transaction.add(self._datafile, offset) |
|
2612 | transaction.add(self._datafile, offset) | |
2593 | transaction.add(self._indexfile, curr * len(entry)) |
|
2613 | transaction.add(self._indexfile, curr * len(entry)) | |
2594 | if data[0]: |
|
2614 | if data[0]: | |
2595 | dfh.write(data[0]) |
|
2615 | dfh.write(data[0]) | |
2596 | dfh.write(data[1]) |
|
2616 | dfh.write(data[1]) | |
2597 | if sidedata: |
|
2617 | if sidedata: | |
2598 | dfh.write(sidedata) |
|
2618 | dfh.write(sidedata) | |
2599 | ifh.write(entry) |
|
2619 | ifh.write(entry) | |
2600 | else: |
|
2620 | else: | |
2601 | offset += curr * self.index.entry_size |
|
2621 | offset += curr * self.index.entry_size | |
2602 | transaction.add(self._indexfile, offset) |
|
2622 | transaction.add(self._indexfile, offset) | |
2603 | ifh.write(entry) |
|
2623 | ifh.write(entry) | |
2604 | ifh.write(data[0]) |
|
2624 | ifh.write(data[0]) | |
2605 | ifh.write(data[1]) |
|
2625 | ifh.write(data[1]) | |
2606 | if sidedata: |
|
2626 | if sidedata: | |
2607 | ifh.write(sidedata) |
|
2627 | ifh.write(sidedata) | |
2608 | self._enforceinlinesize(transaction) |
|
2628 | self._enforceinlinesize(transaction) | |
2609 | if self._docket is not None: |
|
2629 | if self._docket is not None: | |
2610 | self._docket.index_end = self._writinghandles[0].tell() |
|
2630 | self._docket.index_end = self._writinghandles[0].tell() | |
2611 | self._docket.data_end = self._writinghandles[1].tell() |
|
2631 | self._docket.data_end = self._writinghandles[1].tell() | |
2612 |
|
2632 | |||
2613 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
2633 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
2614 |
|
2634 | |||
2615 | def addgroup( |
|
2635 | def addgroup( | |
2616 | self, |
|
2636 | self, | |
2617 | deltas, |
|
2637 | deltas, | |
2618 | linkmapper, |
|
2638 | linkmapper, | |
2619 | transaction, |
|
2639 | transaction, | |
2620 | alwayscache=False, |
|
2640 | alwayscache=False, | |
2621 | addrevisioncb=None, |
|
2641 | addrevisioncb=None, | |
2622 | duplicaterevisioncb=None, |
|
2642 | duplicaterevisioncb=None, | |
2623 | ): |
|
2643 | ): | |
2624 | """ |
|
2644 | """ | |
2625 | add a delta group |
|
2645 | add a delta group | |
2626 |
|
2646 | |||
2627 | given a set of deltas, add them to the revision log. the |
|
2647 | given a set of deltas, add them to the revision log. the | |
2628 | first delta is against its parent, which should be in our |
|
2648 | first delta is against its parent, which should be in our | |
2629 | log, the rest are against the previous delta. |
|
2649 | log, the rest are against the previous delta. | |
2630 |
|
2650 | |||
2631 | If ``addrevisioncb`` is defined, it will be called with arguments of |
|
2651 | If ``addrevisioncb`` is defined, it will be called with arguments of | |
2632 | this revlog and the node that was added. |
|
2652 | this revlog and the node that was added. | |
2633 | """ |
|
2653 | """ | |
2634 |
|
2654 | |||
2635 | if self._adding_group: |
|
2655 | if self._adding_group: | |
2636 | raise error.ProgrammingError(b'cannot nest addgroup() calls') |
|
2656 | raise error.ProgrammingError(b'cannot nest addgroup() calls') | |
2637 |
|
2657 | |||
2638 | self._adding_group = True |
|
2658 | self._adding_group = True | |
2639 | empty = True |
|
2659 | empty = True | |
2640 | try: |
|
2660 | try: | |
2641 | with self._writing(transaction): |
|
2661 | with self._writing(transaction): | |
2642 | deltacomputer = deltautil.deltacomputer(self) |
|
2662 | deltacomputer = deltautil.deltacomputer(self) | |
2643 | # loop through our set of deltas |
|
2663 | # loop through our set of deltas | |
2644 | for data in deltas: |
|
2664 | for data in deltas: | |
2645 | ( |
|
2665 | ( | |
2646 | node, |
|
2666 | node, | |
2647 | p1, |
|
2667 | p1, | |
2648 | p2, |
|
2668 | p2, | |
2649 | linknode, |
|
2669 | linknode, | |
2650 | deltabase, |
|
2670 | deltabase, | |
2651 | delta, |
|
2671 | delta, | |
2652 | flags, |
|
2672 | flags, | |
2653 | sidedata, |
|
2673 | sidedata, | |
2654 | ) = data |
|
2674 | ) = data | |
2655 | link = linkmapper(linknode) |
|
2675 | link = linkmapper(linknode) | |
2656 | flags = flags or REVIDX_DEFAULT_FLAGS |
|
2676 | flags = flags or REVIDX_DEFAULT_FLAGS | |
2657 |
|
2677 | |||
2658 | rev = self.index.get_rev(node) |
|
2678 | rev = self.index.get_rev(node) | |
2659 | if rev is not None: |
|
2679 | if rev is not None: | |
2660 | # this can happen if two branches make the same change |
|
2680 | # this can happen if two branches make the same change | |
2661 | self._nodeduplicatecallback(transaction, rev) |
|
2681 | self._nodeduplicatecallback(transaction, rev) | |
2662 | if duplicaterevisioncb: |
|
2682 | if duplicaterevisioncb: | |
2663 | duplicaterevisioncb(self, rev) |
|
2683 | duplicaterevisioncb(self, rev) | |
2664 | empty = False |
|
2684 | empty = False | |
2665 | continue |
|
2685 | continue | |
2666 |
|
2686 | |||
2667 | for p in (p1, p2): |
|
2687 | for p in (p1, p2): | |
2668 | if not self.index.has_node(p): |
|
2688 | if not self.index.has_node(p): | |
2669 | raise error.LookupError( |
|
2689 | raise error.LookupError( | |
2670 | p, self.radix, _(b'unknown parent') |
|
2690 | p, self.radix, _(b'unknown parent') | |
2671 | ) |
|
2691 | ) | |
2672 |
|
2692 | |||
2673 | if not self.index.has_node(deltabase): |
|
2693 | if not self.index.has_node(deltabase): | |
2674 | raise error.LookupError( |
|
2694 | raise error.LookupError( | |
2675 | deltabase, self.display_id, _(b'unknown delta base') |
|
2695 | deltabase, self.display_id, _(b'unknown delta base') | |
2676 | ) |
|
2696 | ) | |
2677 |
|
2697 | |||
2678 | baserev = self.rev(deltabase) |
|
2698 | baserev = self.rev(deltabase) | |
2679 |
|
2699 | |||
2680 | if baserev != nullrev and self.iscensored(baserev): |
|
2700 | if baserev != nullrev and self.iscensored(baserev): | |
2681 | # if base is censored, delta must be full replacement in a |
|
2701 | # if base is censored, delta must be full replacement in a | |
2682 | # single patch operation |
|
2702 | # single patch operation | |
2683 | hlen = struct.calcsize(b">lll") |
|
2703 | hlen = struct.calcsize(b">lll") | |
2684 | oldlen = self.rawsize(baserev) |
|
2704 | oldlen = self.rawsize(baserev) | |
2685 | newlen = len(delta) - hlen |
|
2705 | newlen = len(delta) - hlen | |
2686 | if delta[:hlen] != mdiff.replacediffheader( |
|
2706 | if delta[:hlen] != mdiff.replacediffheader( | |
2687 | oldlen, newlen |
|
2707 | oldlen, newlen | |
2688 | ): |
|
2708 | ): | |
2689 | raise error.CensoredBaseError( |
|
2709 | raise error.CensoredBaseError( | |
2690 | self.display_id, self.node(baserev) |
|
2710 | self.display_id, self.node(baserev) | |
2691 | ) |
|
2711 | ) | |
2692 |
|
2712 | |||
2693 | if not flags and self._peek_iscensored(baserev, delta): |
|
2713 | if not flags and self._peek_iscensored(baserev, delta): | |
2694 | flags |= REVIDX_ISCENSORED |
|
2714 | flags |= REVIDX_ISCENSORED | |
2695 |
|
2715 | |||
2696 | # We assume consumers of addrevisioncb will want to retrieve |
|
2716 | # We assume consumers of addrevisioncb will want to retrieve | |
2697 | # the added revision, which will require a call to |
|
2717 | # the added revision, which will require a call to | |
2698 | # revision(). revision() will fast path if there is a cache |
|
2718 | # revision(). revision() will fast path if there is a cache | |
2699 | # hit. So, we tell _addrevision() to always cache in this case. |
|
2719 | # hit. So, we tell _addrevision() to always cache in this case. | |
2700 | # We're only using addgroup() in the context of changegroup |
|
2720 | # We're only using addgroup() in the context of changegroup | |
2701 | # generation so the revision data can always be handled as raw |
|
2721 | # generation so the revision data can always be handled as raw | |
2702 | # by the flagprocessor. |
|
2722 | # by the flagprocessor. | |
2703 | rev = self._addrevision( |
|
2723 | rev = self._addrevision( | |
2704 | node, |
|
2724 | node, | |
2705 | None, |
|
2725 | None, | |
2706 | transaction, |
|
2726 | transaction, | |
2707 | link, |
|
2727 | link, | |
2708 | p1, |
|
2728 | p1, | |
2709 | p2, |
|
2729 | p2, | |
2710 | flags, |
|
2730 | flags, | |
2711 | (baserev, delta), |
|
2731 | (baserev, delta), | |
2712 | alwayscache=alwayscache, |
|
2732 | alwayscache=alwayscache, | |
2713 | deltacomputer=deltacomputer, |
|
2733 | deltacomputer=deltacomputer, | |
2714 | sidedata=sidedata, |
|
2734 | sidedata=sidedata, | |
2715 | ) |
|
2735 | ) | |
2716 |
|
2736 | |||
2717 | if addrevisioncb: |
|
2737 | if addrevisioncb: | |
2718 | addrevisioncb(self, rev) |
|
2738 | addrevisioncb(self, rev) | |
2719 | empty = False |
|
2739 | empty = False | |
2720 | finally: |
|
2740 | finally: | |
2721 | self._adding_group = False |
|
2741 | self._adding_group = False | |
2722 | return not empty |
|
2742 | return not empty | |
2723 |
|
2743 | |||
2724 | def iscensored(self, rev): |
|
2744 | def iscensored(self, rev): | |
2725 | """Check if a file revision is censored.""" |
|
2745 | """Check if a file revision is censored.""" | |
2726 | if not self._censorable: |
|
2746 | if not self._censorable: | |
2727 | return False |
|
2747 | return False | |
2728 |
|
2748 | |||
2729 | return self.flags(rev) & REVIDX_ISCENSORED |
|
2749 | return self.flags(rev) & REVIDX_ISCENSORED | |
2730 |
|
2750 | |||
2731 | def _peek_iscensored(self, baserev, delta): |
|
2751 | def _peek_iscensored(self, baserev, delta): | |
2732 | """Quickly check if a delta produces a censored revision.""" |
|
2752 | """Quickly check if a delta produces a censored revision.""" | |
2733 | if not self._censorable: |
|
2753 | if not self._censorable: | |
2734 | return False |
|
2754 | return False | |
2735 |
|
2755 | |||
2736 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) |
|
2756 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) | |
2737 |
|
2757 | |||
2738 | def getstrippoint(self, minlink): |
|
2758 | def getstrippoint(self, minlink): | |
2739 | """find the minimum rev that must be stripped to strip the linkrev |
|
2759 | """find the minimum rev that must be stripped to strip the linkrev | |
2740 |
|
2760 | |||
2741 | Returns a tuple containing the minimum rev and a set of all revs that |
|
2761 | Returns a tuple containing the minimum rev and a set of all revs that | |
2742 | have linkrevs that will be broken by this strip. |
|
2762 | have linkrevs that will be broken by this strip. | |
2743 | """ |
|
2763 | """ | |
2744 | return storageutil.resolvestripinfo( |
|
2764 | return storageutil.resolvestripinfo( | |
2745 | minlink, |
|
2765 | minlink, | |
2746 | len(self) - 1, |
|
2766 | len(self) - 1, | |
2747 | self.headrevs(), |
|
2767 | self.headrevs(), | |
2748 | self.linkrev, |
|
2768 | self.linkrev, | |
2749 | self.parentrevs, |
|
2769 | self.parentrevs, | |
2750 | ) |
|
2770 | ) | |
2751 |
|
2771 | |||
2752 | def strip(self, minlink, transaction): |
|
2772 | def strip(self, minlink, transaction): | |
2753 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
2773 | """truncate the revlog on the first revision with a linkrev >= minlink | |
2754 |
|
2774 | |||
2755 | This function is called when we're stripping revision minlink and |
|
2775 | This function is called when we're stripping revision minlink and | |
2756 | its descendants from the repository. |
|
2776 | its descendants from the repository. | |
2757 |
|
2777 | |||
2758 | We have to remove all revisions with linkrev >= minlink, because |
|
2778 | We have to remove all revisions with linkrev >= minlink, because | |
2759 | the equivalent changelog revisions will be renumbered after the |
|
2779 | the equivalent changelog revisions will be renumbered after the | |
2760 | strip. |
|
2780 | strip. | |
2761 |
|
2781 | |||
2762 | So we truncate the revlog on the first of these revisions, and |
|
2782 | So we truncate the revlog on the first of these revisions, and | |
2763 | trust that the caller has saved the revisions that shouldn't be |
|
2783 | trust that the caller has saved the revisions that shouldn't be | |
2764 | removed and that it'll re-add them after this truncation. |
|
2784 | removed and that it'll re-add them after this truncation. | |
2765 | """ |
|
2785 | """ | |
2766 | if len(self) == 0: |
|
2786 | if len(self) == 0: | |
2767 | return |
|
2787 | return | |
2768 |
|
2788 | |||
2769 | rev, _ = self.getstrippoint(minlink) |
|
2789 | rev, _ = self.getstrippoint(minlink) | |
2770 | if rev == len(self): |
|
2790 | if rev == len(self): | |
2771 | return |
|
2791 | return | |
2772 |
|
2792 | |||
2773 | # first truncate the files on disk |
|
2793 | # first truncate the files on disk | |
2774 | data_end = self.start(rev) |
|
2794 | data_end = self.start(rev) | |
2775 | if not self._inline: |
|
2795 | if not self._inline: | |
2776 | transaction.add(self._datafile, data_end) |
|
2796 | transaction.add(self._datafile, data_end) | |
2777 | end = rev * self.index.entry_size |
|
2797 | end = rev * self.index.entry_size | |
2778 | else: |
|
2798 | else: | |
2779 | end = data_end + (rev * self.index.entry_size) |
|
2799 | end = data_end + (rev * self.index.entry_size) | |
2780 |
|
2800 | |||
2781 | transaction.add(self._indexfile, end) |
|
2801 | transaction.add(self._indexfile, end) | |
2782 | if self._docket is not None: |
|
2802 | if self._docket is not None: | |
2783 | # XXX we could, leverage the docket while stripping. However it is |
|
2803 | # XXX we could, leverage the docket while stripping. However it is | |
2784 | # not powerfull enough at the time of this comment |
|
2804 | # not powerfull enough at the time of this comment | |
2785 | self._docket.index_end = end |
|
2805 | self._docket.index_end = end | |
2786 | self._docket.data_end = data_end |
|
2806 | self._docket.data_end = data_end | |
2787 | self._docket.write(transaction, stripping=True) |
|
2807 | self._docket.write(transaction, stripping=True) | |
2788 |
|
2808 | |||
2789 | # then reset internal state in memory to forget those revisions |
|
2809 | # then reset internal state in memory to forget those revisions | |
2790 | self._revisioncache = None |
|
2810 | self._revisioncache = None | |
2791 | self._chaininfocache = util.lrucachedict(500) |
|
2811 | self._chaininfocache = util.lrucachedict(500) | |
2792 | self._chunkclear() |
|
2812 | self._chunkclear() | |
2793 |
|
2813 | |||
2794 | del self.index[rev:-1] |
|
2814 | del self.index[rev:-1] | |
2795 |
|
2815 | |||
2796 | def checksize(self): |
|
2816 | def checksize(self): | |
2797 | """Check size of index and data files |
|
2817 | """Check size of index and data files | |
2798 |
|
2818 | |||
2799 | return a (dd, di) tuple. |
|
2819 | return a (dd, di) tuple. | |
2800 | - dd: extra bytes for the "data" file |
|
2820 | - dd: extra bytes for the "data" file | |
2801 | - di: extra bytes for the "index" file |
|
2821 | - di: extra bytes for the "index" file | |
2802 |
|
2822 | |||
2803 | A healthy revlog will return (0, 0). |
|
2823 | A healthy revlog will return (0, 0). | |
2804 | """ |
|
2824 | """ | |
2805 | expected = 0 |
|
2825 | expected = 0 | |
2806 | if len(self): |
|
2826 | if len(self): | |
2807 | expected = max(0, self.end(len(self) - 1)) |
|
2827 | expected = max(0, self.end(len(self) - 1)) | |
2808 |
|
2828 | |||
2809 | try: |
|
2829 | try: | |
2810 | with self._datafp() as f: |
|
2830 | with self._datafp() as f: | |
2811 | f.seek(0, io.SEEK_END) |
|
2831 | f.seek(0, io.SEEK_END) | |
2812 | actual = f.tell() |
|
2832 | actual = f.tell() | |
2813 | dd = actual - expected |
|
2833 | dd = actual - expected | |
2814 | except IOError as inst: |
|
2834 | except IOError as inst: | |
2815 | if inst.errno != errno.ENOENT: |
|
2835 | if inst.errno != errno.ENOENT: | |
2816 | raise |
|
2836 | raise | |
2817 | dd = 0 |
|
2837 | dd = 0 | |
2818 |
|
2838 | |||
2819 | try: |
|
2839 | try: | |
2820 | f = self.opener(self._indexfile) |
|
2840 | f = self.opener(self._indexfile) | |
2821 | f.seek(0, io.SEEK_END) |
|
2841 | f.seek(0, io.SEEK_END) | |
2822 | actual = f.tell() |
|
2842 | actual = f.tell() | |
2823 | f.close() |
|
2843 | f.close() | |
2824 | s = self.index.entry_size |
|
2844 | s = self.index.entry_size | |
2825 | i = max(0, actual // s) |
|
2845 | i = max(0, actual // s) | |
2826 | di = actual - (i * s) |
|
2846 | di = actual - (i * s) | |
2827 | if self._inline: |
|
2847 | if self._inline: | |
2828 | databytes = 0 |
|
2848 | databytes = 0 | |
2829 | for r in self: |
|
2849 | for r in self: | |
2830 | databytes += max(0, self.length(r)) |
|
2850 | databytes += max(0, self.length(r)) | |
2831 | dd = 0 |
|
2851 | dd = 0 | |
2832 | di = actual - len(self) * s - databytes |
|
2852 | di = actual - len(self) * s - databytes | |
2833 | except IOError as inst: |
|
2853 | except IOError as inst: | |
2834 | if inst.errno != errno.ENOENT: |
|
2854 | if inst.errno != errno.ENOENT: | |
2835 | raise |
|
2855 | raise | |
2836 | di = 0 |
|
2856 | di = 0 | |
2837 |
|
2857 | |||
2838 | return (dd, di) |
|
2858 | return (dd, di) | |
2839 |
|
2859 | |||
2840 | def files(self): |
|
2860 | def files(self): | |
2841 | res = [self._indexfile] |
|
2861 | res = [self._indexfile] | |
2842 | if not self._inline: |
|
2862 | if not self._inline: | |
2843 | res.append(self._datafile) |
|
2863 | res.append(self._datafile) | |
2844 | return res |
|
2864 | return res | |
2845 |
|
2865 | |||
2846 | def emitrevisions( |
|
2866 | def emitrevisions( | |
2847 | self, |
|
2867 | self, | |
2848 | nodes, |
|
2868 | nodes, | |
2849 | nodesorder=None, |
|
2869 | nodesorder=None, | |
2850 | revisiondata=False, |
|
2870 | revisiondata=False, | |
2851 | assumehaveparentrevisions=False, |
|
2871 | assumehaveparentrevisions=False, | |
2852 | deltamode=repository.CG_DELTAMODE_STD, |
|
2872 | deltamode=repository.CG_DELTAMODE_STD, | |
2853 | sidedata_helpers=None, |
|
2873 | sidedata_helpers=None, | |
2854 | ): |
|
2874 | ): | |
2855 | if nodesorder not in (b'nodes', b'storage', b'linear', None): |
|
2875 | if nodesorder not in (b'nodes', b'storage', b'linear', None): | |
2856 | raise error.ProgrammingError( |
|
2876 | raise error.ProgrammingError( | |
2857 | b'unhandled value for nodesorder: %s' % nodesorder |
|
2877 | b'unhandled value for nodesorder: %s' % nodesorder | |
2858 | ) |
|
2878 | ) | |
2859 |
|
2879 | |||
2860 | if nodesorder is None and not self._generaldelta: |
|
2880 | if nodesorder is None and not self._generaldelta: | |
2861 | nodesorder = b'storage' |
|
2881 | nodesorder = b'storage' | |
2862 |
|
2882 | |||
2863 | if ( |
|
2883 | if ( | |
2864 | not self._storedeltachains |
|
2884 | not self._storedeltachains | |
2865 | and deltamode != repository.CG_DELTAMODE_PREV |
|
2885 | and deltamode != repository.CG_DELTAMODE_PREV | |
2866 | ): |
|
2886 | ): | |
2867 | deltamode = repository.CG_DELTAMODE_FULL |
|
2887 | deltamode = repository.CG_DELTAMODE_FULL | |
2868 |
|
2888 | |||
2869 | return storageutil.emitrevisions( |
|
2889 | return storageutil.emitrevisions( | |
2870 | self, |
|
2890 | self, | |
2871 | nodes, |
|
2891 | nodes, | |
2872 | nodesorder, |
|
2892 | nodesorder, | |
2873 | revlogrevisiondelta, |
|
2893 | revlogrevisiondelta, | |
2874 | deltaparentfn=self.deltaparent, |
|
2894 | deltaparentfn=self.deltaparent, | |
2875 | candeltafn=self.candelta, |
|
2895 | candeltafn=self.candelta, | |
2876 | rawsizefn=self.rawsize, |
|
2896 | rawsizefn=self.rawsize, | |
2877 | revdifffn=self.revdiff, |
|
2897 | revdifffn=self.revdiff, | |
2878 | flagsfn=self.flags, |
|
2898 | flagsfn=self.flags, | |
2879 | deltamode=deltamode, |
|
2899 | deltamode=deltamode, | |
2880 | revisiondata=revisiondata, |
|
2900 | revisiondata=revisiondata, | |
2881 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
2901 | assumehaveparentrevisions=assumehaveparentrevisions, | |
2882 | sidedata_helpers=sidedata_helpers, |
|
2902 | sidedata_helpers=sidedata_helpers, | |
2883 | ) |
|
2903 | ) | |
2884 |
|
2904 | |||
2885 | DELTAREUSEALWAYS = b'always' |
|
2905 | DELTAREUSEALWAYS = b'always' | |
2886 | DELTAREUSESAMEREVS = b'samerevs' |
|
2906 | DELTAREUSESAMEREVS = b'samerevs' | |
2887 | DELTAREUSENEVER = b'never' |
|
2907 | DELTAREUSENEVER = b'never' | |
2888 |
|
2908 | |||
2889 | DELTAREUSEFULLADD = b'fulladd' |
|
2909 | DELTAREUSEFULLADD = b'fulladd' | |
2890 |
|
2910 | |||
2891 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} |
|
2911 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} | |
2892 |
|
2912 | |||
2893 | def clone( |
|
2913 | def clone( | |
2894 | self, |
|
2914 | self, | |
2895 | tr, |
|
2915 | tr, | |
2896 | destrevlog, |
|
2916 | destrevlog, | |
2897 | addrevisioncb=None, |
|
2917 | addrevisioncb=None, | |
2898 | deltareuse=DELTAREUSESAMEREVS, |
|
2918 | deltareuse=DELTAREUSESAMEREVS, | |
2899 | forcedeltabothparents=None, |
|
2919 | forcedeltabothparents=None, | |
2900 | sidedata_helpers=None, |
|
2920 | sidedata_helpers=None, | |
2901 | ): |
|
2921 | ): | |
2902 | """Copy this revlog to another, possibly with format changes. |
|
2922 | """Copy this revlog to another, possibly with format changes. | |
2903 |
|
2923 | |||
2904 | The destination revlog will contain the same revisions and nodes. |
|
2924 | The destination revlog will contain the same revisions and nodes. | |
2905 | However, it may not be bit-for-bit identical due to e.g. delta encoding |
|
2925 | However, it may not be bit-for-bit identical due to e.g. delta encoding | |
2906 | differences. |
|
2926 | differences. | |
2907 |
|
2927 | |||
2908 | The ``deltareuse`` argument control how deltas from the existing revlog |
|
2928 | The ``deltareuse`` argument control how deltas from the existing revlog | |
2909 | are preserved in the destination revlog. The argument can have the |
|
2929 | are preserved in the destination revlog. The argument can have the | |
2910 | following values: |
|
2930 | following values: | |
2911 |
|
2931 | |||
2912 | DELTAREUSEALWAYS |
|
2932 | DELTAREUSEALWAYS | |
2913 | Deltas will always be reused (if possible), even if the destination |
|
2933 | Deltas will always be reused (if possible), even if the destination | |
2914 | revlog would not select the same revisions for the delta. This is the |
|
2934 | revlog would not select the same revisions for the delta. This is the | |
2915 | fastest mode of operation. |
|
2935 | fastest mode of operation. | |
2916 | DELTAREUSESAMEREVS |
|
2936 | DELTAREUSESAMEREVS | |
2917 | Deltas will be reused if the destination revlog would pick the same |
|
2937 | Deltas will be reused if the destination revlog would pick the same | |
2918 | revisions for the delta. This mode strikes a balance between speed |
|
2938 | revisions for the delta. This mode strikes a balance between speed | |
2919 | and optimization. |
|
2939 | and optimization. | |
2920 | DELTAREUSENEVER |
|
2940 | DELTAREUSENEVER | |
2921 | Deltas will never be reused. This is the slowest mode of execution. |
|
2941 | Deltas will never be reused. This is the slowest mode of execution. | |
2922 | This mode can be used to recompute deltas (e.g. if the diff/delta |
|
2942 | This mode can be used to recompute deltas (e.g. if the diff/delta | |
2923 | algorithm changes). |
|
2943 | algorithm changes). | |
2924 | DELTAREUSEFULLADD |
|
2944 | DELTAREUSEFULLADD | |
2925 | Revision will be re-added as if their were new content. This is |
|
2945 | Revision will be re-added as if their were new content. This is | |
2926 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. |
|
2946 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. | |
2927 | eg: large file detection and handling. |
|
2947 | eg: large file detection and handling. | |
2928 |
|
2948 | |||
2929 | Delta computation can be slow, so the choice of delta reuse policy can |
|
2949 | Delta computation can be slow, so the choice of delta reuse policy can | |
2930 | significantly affect run time. |
|
2950 | significantly affect run time. | |
2931 |
|
2951 | |||
2932 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between |
|
2952 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between | |
2933 | two extremes. Deltas will be reused if they are appropriate. But if the |
|
2953 | two extremes. Deltas will be reused if they are appropriate. But if the | |
2934 | delta could choose a better revision, it will do so. This means if you |
|
2954 | delta could choose a better revision, it will do so. This means if you | |
2935 | are converting a non-generaldelta revlog to a generaldelta revlog, |
|
2955 | are converting a non-generaldelta revlog to a generaldelta revlog, | |
2936 | deltas will be recomputed if the delta's parent isn't a parent of the |
|
2956 | deltas will be recomputed if the delta's parent isn't a parent of the | |
2937 | revision. |
|
2957 | revision. | |
2938 |
|
2958 | |||
2939 | In addition to the delta policy, the ``forcedeltabothparents`` |
|
2959 | In addition to the delta policy, the ``forcedeltabothparents`` | |
2940 | argument controls whether to force compute deltas against both parents |
|
2960 | argument controls whether to force compute deltas against both parents | |
2941 | for merges. By default, the current default is used. |
|
2961 | for merges. By default, the current default is used. | |
2942 |
|
2962 | |||
2943 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on |
|
2963 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on | |
2944 | `sidedata_helpers`. |
|
2964 | `sidedata_helpers`. | |
2945 | """ |
|
2965 | """ | |
2946 | if deltareuse not in self.DELTAREUSEALL: |
|
2966 | if deltareuse not in self.DELTAREUSEALL: | |
2947 | raise ValueError( |
|
2967 | raise ValueError( | |
2948 | _(b'value for deltareuse invalid: %s') % deltareuse |
|
2968 | _(b'value for deltareuse invalid: %s') % deltareuse | |
2949 | ) |
|
2969 | ) | |
2950 |
|
2970 | |||
2951 | if len(destrevlog): |
|
2971 | if len(destrevlog): | |
2952 | raise ValueError(_(b'destination revlog is not empty')) |
|
2972 | raise ValueError(_(b'destination revlog is not empty')) | |
2953 |
|
2973 | |||
2954 | if getattr(self, 'filteredrevs', None): |
|
2974 | if getattr(self, 'filteredrevs', None): | |
2955 | raise ValueError(_(b'source revlog has filtered revisions')) |
|
2975 | raise ValueError(_(b'source revlog has filtered revisions')) | |
2956 | if getattr(destrevlog, 'filteredrevs', None): |
|
2976 | if getattr(destrevlog, 'filteredrevs', None): | |
2957 | raise ValueError(_(b'destination revlog has filtered revisions')) |
|
2977 | raise ValueError(_(b'destination revlog has filtered revisions')) | |
2958 |
|
2978 | |||
2959 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, |
|
2979 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, | |
2960 | # if possible. |
|
2980 | # if possible. | |
2961 | oldlazydelta = destrevlog._lazydelta |
|
2981 | oldlazydelta = destrevlog._lazydelta | |
2962 | oldlazydeltabase = destrevlog._lazydeltabase |
|
2982 | oldlazydeltabase = destrevlog._lazydeltabase | |
2963 | oldamd = destrevlog._deltabothparents |
|
2983 | oldamd = destrevlog._deltabothparents | |
2964 |
|
2984 | |||
2965 | try: |
|
2985 | try: | |
2966 | if deltareuse == self.DELTAREUSEALWAYS: |
|
2986 | if deltareuse == self.DELTAREUSEALWAYS: | |
2967 | destrevlog._lazydeltabase = True |
|
2987 | destrevlog._lazydeltabase = True | |
2968 | destrevlog._lazydelta = True |
|
2988 | destrevlog._lazydelta = True | |
2969 | elif deltareuse == self.DELTAREUSESAMEREVS: |
|
2989 | elif deltareuse == self.DELTAREUSESAMEREVS: | |
2970 | destrevlog._lazydeltabase = False |
|
2990 | destrevlog._lazydeltabase = False | |
2971 | destrevlog._lazydelta = True |
|
2991 | destrevlog._lazydelta = True | |
2972 | elif deltareuse == self.DELTAREUSENEVER: |
|
2992 | elif deltareuse == self.DELTAREUSENEVER: | |
2973 | destrevlog._lazydeltabase = False |
|
2993 | destrevlog._lazydeltabase = False | |
2974 | destrevlog._lazydelta = False |
|
2994 | destrevlog._lazydelta = False | |
2975 |
|
2995 | |||
2976 | destrevlog._deltabothparents = forcedeltabothparents or oldamd |
|
2996 | destrevlog._deltabothparents = forcedeltabothparents or oldamd | |
2977 |
|
2997 | |||
2978 | self._clone( |
|
2998 | self._clone( | |
2979 | tr, |
|
2999 | tr, | |
2980 | destrevlog, |
|
3000 | destrevlog, | |
2981 | addrevisioncb, |
|
3001 | addrevisioncb, | |
2982 | deltareuse, |
|
3002 | deltareuse, | |
2983 | forcedeltabothparents, |
|
3003 | forcedeltabothparents, | |
2984 | sidedata_helpers, |
|
3004 | sidedata_helpers, | |
2985 | ) |
|
3005 | ) | |
2986 |
|
3006 | |||
2987 | finally: |
|
3007 | finally: | |
2988 | destrevlog._lazydelta = oldlazydelta |
|
3008 | destrevlog._lazydelta = oldlazydelta | |
2989 | destrevlog._lazydeltabase = oldlazydeltabase |
|
3009 | destrevlog._lazydeltabase = oldlazydeltabase | |
2990 | destrevlog._deltabothparents = oldamd |
|
3010 | destrevlog._deltabothparents = oldamd | |
2991 |
|
3011 | |||
2992 | def _clone( |
|
3012 | def _clone( | |
2993 | self, |
|
3013 | self, | |
2994 | tr, |
|
3014 | tr, | |
2995 | destrevlog, |
|
3015 | destrevlog, | |
2996 | addrevisioncb, |
|
3016 | addrevisioncb, | |
2997 | deltareuse, |
|
3017 | deltareuse, | |
2998 | forcedeltabothparents, |
|
3018 | forcedeltabothparents, | |
2999 | sidedata_helpers, |
|
3019 | sidedata_helpers, | |
3000 | ): |
|
3020 | ): | |
3001 | """perform the core duty of `revlog.clone` after parameter processing""" |
|
3021 | """perform the core duty of `revlog.clone` after parameter processing""" | |
3002 | deltacomputer = deltautil.deltacomputer(destrevlog) |
|
3022 | deltacomputer = deltautil.deltacomputer(destrevlog) | |
3003 | index = self.index |
|
3023 | index = self.index | |
3004 | for rev in self: |
|
3024 | for rev in self: | |
3005 | entry = index[rev] |
|
3025 | entry = index[rev] | |
3006 |
|
3026 | |||
3007 | # Some classes override linkrev to take filtered revs into |
|
3027 | # Some classes override linkrev to take filtered revs into | |
3008 | # account. Use raw entry from index. |
|
3028 | # account. Use raw entry from index. | |
3009 | flags = entry[0] & 0xFFFF |
|
3029 | flags = entry[0] & 0xFFFF | |
3010 | linkrev = entry[4] |
|
3030 | linkrev = entry[4] | |
3011 | p1 = index[entry[5]][7] |
|
3031 | p1 = index[entry[5]][7] | |
3012 | p2 = index[entry[6]][7] |
|
3032 | p2 = index[entry[6]][7] | |
3013 | node = entry[7] |
|
3033 | node = entry[7] | |
3014 |
|
3034 | |||
3015 | # (Possibly) reuse the delta from the revlog if allowed and |
|
3035 | # (Possibly) reuse the delta from the revlog if allowed and | |
3016 | # the revlog chunk is a delta. |
|
3036 | # the revlog chunk is a delta. | |
3017 | cachedelta = None |
|
3037 | cachedelta = None | |
3018 | rawtext = None |
|
3038 | rawtext = None | |
3019 | if deltareuse == self.DELTAREUSEFULLADD: |
|
3039 | if deltareuse == self.DELTAREUSEFULLADD: | |
3020 | text, sidedata = self._revisiondata(rev) |
|
3040 | text, sidedata = self._revisiondata(rev) | |
3021 |
|
3041 | |||
3022 | if sidedata_helpers is not None: |
|
3042 | if sidedata_helpers is not None: | |
3023 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
3043 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
3024 | self, sidedata_helpers, sidedata, rev |
|
3044 | self, sidedata_helpers, sidedata, rev | |
3025 | ) |
|
3045 | ) | |
3026 | flags = flags | new_flags[0] & ~new_flags[1] |
|
3046 | flags = flags | new_flags[0] & ~new_flags[1] | |
3027 |
|
3047 | |||
3028 | destrevlog.addrevision( |
|
3048 | destrevlog.addrevision( | |
3029 | text, |
|
3049 | text, | |
3030 | tr, |
|
3050 | tr, | |
3031 | linkrev, |
|
3051 | linkrev, | |
3032 | p1, |
|
3052 | p1, | |
3033 | p2, |
|
3053 | p2, | |
3034 | cachedelta=cachedelta, |
|
3054 | cachedelta=cachedelta, | |
3035 | node=node, |
|
3055 | node=node, | |
3036 | flags=flags, |
|
3056 | flags=flags, | |
3037 | deltacomputer=deltacomputer, |
|
3057 | deltacomputer=deltacomputer, | |
3038 | sidedata=sidedata, |
|
3058 | sidedata=sidedata, | |
3039 | ) |
|
3059 | ) | |
3040 | else: |
|
3060 | else: | |
3041 | if destrevlog._lazydelta: |
|
3061 | if destrevlog._lazydelta: | |
3042 | dp = self.deltaparent(rev) |
|
3062 | dp = self.deltaparent(rev) | |
3043 | if dp != nullrev: |
|
3063 | if dp != nullrev: | |
3044 | cachedelta = (dp, bytes(self._chunk(rev))) |
|
3064 | cachedelta = (dp, bytes(self._chunk(rev))) | |
3045 |
|
3065 | |||
3046 | sidedata = None |
|
3066 | sidedata = None | |
3047 | if not cachedelta: |
|
3067 | if not cachedelta: | |
3048 | rawtext, sidedata = self._revisiondata(rev) |
|
3068 | rawtext, sidedata = self._revisiondata(rev) | |
3049 | if sidedata is None: |
|
3069 | if sidedata is None: | |
3050 | sidedata = self.sidedata(rev) |
|
3070 | sidedata = self.sidedata(rev) | |
3051 |
|
3071 | |||
3052 | if sidedata_helpers is not None: |
|
3072 | if sidedata_helpers is not None: | |
3053 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
3073 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
3054 | self, sidedata_helpers, sidedata, rev |
|
3074 | self, sidedata_helpers, sidedata, rev | |
3055 | ) |
|
3075 | ) | |
3056 | flags = flags | new_flags[0] & ~new_flags[1] |
|
3076 | flags = flags | new_flags[0] & ~new_flags[1] | |
3057 |
|
3077 | |||
3058 | with destrevlog._writing(tr): |
|
3078 | with destrevlog._writing(tr): | |
3059 | destrevlog._addrevision( |
|
3079 | destrevlog._addrevision( | |
3060 | node, |
|
3080 | node, | |
3061 | rawtext, |
|
3081 | rawtext, | |
3062 | tr, |
|
3082 | tr, | |
3063 | linkrev, |
|
3083 | linkrev, | |
3064 | p1, |
|
3084 | p1, | |
3065 | p2, |
|
3085 | p2, | |
3066 | flags, |
|
3086 | flags, | |
3067 | cachedelta, |
|
3087 | cachedelta, | |
3068 | deltacomputer=deltacomputer, |
|
3088 | deltacomputer=deltacomputer, | |
3069 | sidedata=sidedata, |
|
3089 | sidedata=sidedata, | |
3070 | ) |
|
3090 | ) | |
3071 |
|
3091 | |||
3072 | if addrevisioncb: |
|
3092 | if addrevisioncb: | |
3073 | addrevisioncb(self, rev, node) |
|
3093 | addrevisioncb(self, rev, node) | |
3074 |
|
3094 | |||
3075 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
3095 | def censorrevision(self, tr, censornode, tombstone=b''): | |
3076 | if self._format_version == REVLOGV0: |
|
3096 | if self._format_version == REVLOGV0: | |
3077 | raise error.RevlogError( |
|
3097 | raise error.RevlogError( | |
3078 | _(b'cannot censor with version %d revlogs') |
|
3098 | _(b'cannot censor with version %d revlogs') | |
3079 | % self._format_version |
|
3099 | % self._format_version | |
3080 | ) |
|
3100 | ) | |
3081 |
|
3101 | |||
3082 | censorrev = self.rev(censornode) |
|
3102 | censorrev = self.rev(censornode) | |
3083 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') |
|
3103 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') | |
3084 |
|
3104 | |||
3085 | if len(tombstone) > self.rawsize(censorrev): |
|
3105 | if len(tombstone) > self.rawsize(censorrev): | |
3086 | raise error.Abort( |
|
3106 | raise error.Abort( | |
3087 | _(b'censor tombstone must be no longer than censored data') |
|
3107 | _(b'censor tombstone must be no longer than censored data') | |
3088 | ) |
|
3108 | ) | |
3089 |
|
3109 | |||
3090 | # Rewriting the revlog in place is hard. Our strategy for censoring is |
|
3110 | # Rewriting the revlog in place is hard. Our strategy for censoring is | |
3091 | # to create a new revlog, copy all revisions to it, then replace the |
|
3111 | # to create a new revlog, copy all revisions to it, then replace the | |
3092 | # revlogs on transaction close. |
|
3112 | # revlogs on transaction close. | |
3093 | # |
|
3113 | # | |
3094 | # This is a bit dangerous. We could easily have a mismatch of state. |
|
3114 | # This is a bit dangerous. We could easily have a mismatch of state. | |
3095 | newrl = revlog( |
|
3115 | newrl = revlog( | |
3096 | self.opener, |
|
3116 | self.opener, | |
3097 | target=self.target, |
|
3117 | target=self.target, | |
3098 | radix=self.radix, |
|
3118 | radix=self.radix, | |
3099 | postfix=b'tmpcensored', |
|
3119 | postfix=b'tmpcensored', | |
3100 | censorable=True, |
|
3120 | censorable=True, | |
3101 | ) |
|
3121 | ) | |
3102 | newrl._format_version = self._format_version |
|
3122 | newrl._format_version = self._format_version | |
3103 | newrl._format_flags = self._format_flags |
|
3123 | newrl._format_flags = self._format_flags | |
3104 | newrl._generaldelta = self._generaldelta |
|
3124 | newrl._generaldelta = self._generaldelta | |
3105 | newrl._parse_index = self._parse_index |
|
3125 | newrl._parse_index = self._parse_index | |
3106 |
|
3126 | |||
3107 | for rev in self.revs(): |
|
3127 | for rev in self.revs(): | |
3108 | node = self.node(rev) |
|
3128 | node = self.node(rev) | |
3109 | p1, p2 = self.parents(node) |
|
3129 | p1, p2 = self.parents(node) | |
3110 |
|
3130 | |||
3111 | if rev == censorrev: |
|
3131 | if rev == censorrev: | |
3112 | newrl.addrawrevision( |
|
3132 | newrl.addrawrevision( | |
3113 | tombstone, |
|
3133 | tombstone, | |
3114 | tr, |
|
3134 | tr, | |
3115 | self.linkrev(censorrev), |
|
3135 | self.linkrev(censorrev), | |
3116 | p1, |
|
3136 | p1, | |
3117 | p2, |
|
3137 | p2, | |
3118 | censornode, |
|
3138 | censornode, | |
3119 | REVIDX_ISCENSORED, |
|
3139 | REVIDX_ISCENSORED, | |
3120 | ) |
|
3140 | ) | |
3121 |
|
3141 | |||
3122 | if newrl.deltaparent(rev) != nullrev: |
|
3142 | if newrl.deltaparent(rev) != nullrev: | |
3123 | raise error.Abort( |
|
3143 | raise error.Abort( | |
3124 | _( |
|
3144 | _( | |
3125 | b'censored revision stored as delta; ' |
|
3145 | b'censored revision stored as delta; ' | |
3126 | b'cannot censor' |
|
3146 | b'cannot censor' | |
3127 | ), |
|
3147 | ), | |
3128 | hint=_( |
|
3148 | hint=_( | |
3129 | b'censoring of revlogs is not ' |
|
3149 | b'censoring of revlogs is not ' | |
3130 | b'fully implemented; please report ' |
|
3150 | b'fully implemented; please report ' | |
3131 | b'this bug' |
|
3151 | b'this bug' | |
3132 | ), |
|
3152 | ), | |
3133 | ) |
|
3153 | ) | |
3134 | continue |
|
3154 | continue | |
3135 |
|
3155 | |||
3136 | if self.iscensored(rev): |
|
3156 | if self.iscensored(rev): | |
3137 | if self.deltaparent(rev) != nullrev: |
|
3157 | if self.deltaparent(rev) != nullrev: | |
3138 | raise error.Abort( |
|
3158 | raise error.Abort( | |
3139 | _( |
|
3159 | _( | |
3140 | b'cannot censor due to censored ' |
|
3160 | b'cannot censor due to censored ' | |
3141 | b'revision having delta stored' |
|
3161 | b'revision having delta stored' | |
3142 | ) |
|
3162 | ) | |
3143 | ) |
|
3163 | ) | |
3144 | rawtext = self._chunk(rev) |
|
3164 | rawtext = self._chunk(rev) | |
3145 | else: |
|
3165 | else: | |
3146 | rawtext = self.rawdata(rev) |
|
3166 | rawtext = self.rawdata(rev) | |
3147 |
|
3167 | |||
3148 | newrl.addrawrevision( |
|
3168 | newrl.addrawrevision( | |
3149 | rawtext, tr, self.linkrev(rev), p1, p2, node, self.flags(rev) |
|
3169 | rawtext, tr, self.linkrev(rev), p1, p2, node, self.flags(rev) | |
3150 | ) |
|
3170 | ) | |
3151 |
|
3171 | |||
3152 | tr.addbackup(self._indexfile, location=b'store') |
|
3172 | tr.addbackup(self._indexfile, location=b'store') | |
3153 | if not self._inline: |
|
3173 | if not self._inline: | |
3154 | tr.addbackup(self._datafile, location=b'store') |
|
3174 | tr.addbackup(self._datafile, location=b'store') | |
3155 |
|
3175 | |||
3156 | self.opener.rename(newrl._indexfile, self._indexfile) |
|
3176 | self.opener.rename(newrl._indexfile, self._indexfile) | |
3157 | if not self._inline: |
|
3177 | if not self._inline: | |
3158 | self.opener.rename(newrl._datafile, self._datafile) |
|
3178 | self.opener.rename(newrl._datafile, self._datafile) | |
3159 |
|
3179 | |||
3160 | self.clearcaches() |
|
3180 | self.clearcaches() | |
3161 | self._loadindex() |
|
3181 | self._loadindex() | |
3162 |
|
3182 | |||
3163 | def verifyintegrity(self, state): |
|
3183 | def verifyintegrity(self, state): | |
3164 | """Verifies the integrity of the revlog. |
|
3184 | """Verifies the integrity of the revlog. | |
3165 |
|
3185 | |||
3166 | Yields ``revlogproblem`` instances describing problems that are |
|
3186 | Yields ``revlogproblem`` instances describing problems that are | |
3167 | found. |
|
3187 | found. | |
3168 | """ |
|
3188 | """ | |
3169 | dd, di = self.checksize() |
|
3189 | dd, di = self.checksize() | |
3170 | if dd: |
|
3190 | if dd: | |
3171 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) |
|
3191 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) | |
3172 | if di: |
|
3192 | if di: | |
3173 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) |
|
3193 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) | |
3174 |
|
3194 | |||
3175 | version = self._format_version |
|
3195 | version = self._format_version | |
3176 |
|
3196 | |||
3177 | # The verifier tells us what version revlog we should be. |
|
3197 | # The verifier tells us what version revlog we should be. | |
3178 | if version != state[b'expectedversion']: |
|
3198 | if version != state[b'expectedversion']: | |
3179 | yield revlogproblem( |
|
3199 | yield revlogproblem( | |
3180 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") |
|
3200 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") | |
3181 | % (self.display_id, version, state[b'expectedversion']) |
|
3201 | % (self.display_id, version, state[b'expectedversion']) | |
3182 | ) |
|
3202 | ) | |
3183 |
|
3203 | |||
3184 | state[b'skipread'] = set() |
|
3204 | state[b'skipread'] = set() | |
3185 | state[b'safe_renamed'] = set() |
|
3205 | state[b'safe_renamed'] = set() | |
3186 |
|
3206 | |||
3187 | for rev in self: |
|
3207 | for rev in self: | |
3188 | node = self.node(rev) |
|
3208 | node = self.node(rev) | |
3189 |
|
3209 | |||
3190 | # Verify contents. 4 cases to care about: |
|
3210 | # Verify contents. 4 cases to care about: | |
3191 | # |
|
3211 | # | |
3192 | # common: the most common case |
|
3212 | # common: the most common case | |
3193 | # rename: with a rename |
|
3213 | # rename: with a rename | |
3194 | # meta: file content starts with b'\1\n', the metadata |
|
3214 | # meta: file content starts with b'\1\n', the metadata | |
3195 | # header defined in filelog.py, but without a rename |
|
3215 | # header defined in filelog.py, but without a rename | |
3196 | # ext: content stored externally |
|
3216 | # ext: content stored externally | |
3197 | # |
|
3217 | # | |
3198 | # More formally, their differences are shown below: |
|
3218 | # More formally, their differences are shown below: | |
3199 | # |
|
3219 | # | |
3200 | # | common | rename | meta | ext |
|
3220 | # | common | rename | meta | ext | |
3201 | # ------------------------------------------------------- |
|
3221 | # ------------------------------------------------------- | |
3202 | # flags() | 0 | 0 | 0 | not 0 |
|
3222 | # flags() | 0 | 0 | 0 | not 0 | |
3203 | # renamed() | False | True | False | ? |
|
3223 | # renamed() | False | True | False | ? | |
3204 | # rawtext[0:2]=='\1\n'| False | True | True | ? |
|
3224 | # rawtext[0:2]=='\1\n'| False | True | True | ? | |
3205 | # |
|
3225 | # | |
3206 | # "rawtext" means the raw text stored in revlog data, which |
|
3226 | # "rawtext" means the raw text stored in revlog data, which | |
3207 | # could be retrieved by "rawdata(rev)". "text" |
|
3227 | # could be retrieved by "rawdata(rev)". "text" | |
3208 | # mentioned below is "revision(rev)". |
|
3228 | # mentioned below is "revision(rev)". | |
3209 | # |
|
3229 | # | |
3210 | # There are 3 different lengths stored physically: |
|
3230 | # There are 3 different lengths stored physically: | |
3211 | # 1. L1: rawsize, stored in revlog index |
|
3231 | # 1. L1: rawsize, stored in revlog index | |
3212 | # 2. L2: len(rawtext), stored in revlog data |
|
3232 | # 2. L2: len(rawtext), stored in revlog data | |
3213 | # 3. L3: len(text), stored in revlog data if flags==0, or |
|
3233 | # 3. L3: len(text), stored in revlog data if flags==0, or | |
3214 | # possibly somewhere else if flags!=0 |
|
3234 | # possibly somewhere else if flags!=0 | |
3215 | # |
|
3235 | # | |
3216 | # L1 should be equal to L2. L3 could be different from them. |
|
3236 | # L1 should be equal to L2. L3 could be different from them. | |
3217 | # "text" may or may not affect commit hash depending on flag |
|
3237 | # "text" may or may not affect commit hash depending on flag | |
3218 | # processors (see flagutil.addflagprocessor). |
|
3238 | # processors (see flagutil.addflagprocessor). | |
3219 | # |
|
3239 | # | |
3220 | # | common | rename | meta | ext |
|
3240 | # | common | rename | meta | ext | |
3221 | # ------------------------------------------------- |
|
3241 | # ------------------------------------------------- | |
3222 | # rawsize() | L1 | L1 | L1 | L1 |
|
3242 | # rawsize() | L1 | L1 | L1 | L1 | |
3223 | # size() | L1 | L2-LM | L1(*) | L1 (?) |
|
3243 | # size() | L1 | L2-LM | L1(*) | L1 (?) | |
3224 | # len(rawtext) | L2 | L2 | L2 | L2 |
|
3244 | # len(rawtext) | L2 | L2 | L2 | L2 | |
3225 | # len(text) | L2 | L2 | L2 | L3 |
|
3245 | # len(text) | L2 | L2 | L2 | L3 | |
3226 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) |
|
3246 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) | |
3227 | # |
|
3247 | # | |
3228 | # LM: length of metadata, depending on rawtext |
|
3248 | # LM: length of metadata, depending on rawtext | |
3229 | # (*): not ideal, see comment in filelog.size |
|
3249 | # (*): not ideal, see comment in filelog.size | |
3230 | # (?): could be "- len(meta)" if the resolved content has |
|
3250 | # (?): could be "- len(meta)" if the resolved content has | |
3231 | # rename metadata |
|
3251 | # rename metadata | |
3232 | # |
|
3252 | # | |
3233 | # Checks needed to be done: |
|
3253 | # Checks needed to be done: | |
3234 | # 1. length check: L1 == L2, in all cases. |
|
3254 | # 1. length check: L1 == L2, in all cases. | |
3235 | # 2. hash check: depending on flag processor, we may need to |
|
3255 | # 2. hash check: depending on flag processor, we may need to | |
3236 | # use either "text" (external), or "rawtext" (in revlog). |
|
3256 | # use either "text" (external), or "rawtext" (in revlog). | |
3237 |
|
3257 | |||
3238 | try: |
|
3258 | try: | |
3239 | skipflags = state.get(b'skipflags', 0) |
|
3259 | skipflags = state.get(b'skipflags', 0) | |
3240 | if skipflags: |
|
3260 | if skipflags: | |
3241 | skipflags &= self.flags(rev) |
|
3261 | skipflags &= self.flags(rev) | |
3242 |
|
3262 | |||
3243 | _verify_revision(self, skipflags, state, node) |
|
3263 | _verify_revision(self, skipflags, state, node) | |
3244 |
|
3264 | |||
3245 | l1 = self.rawsize(rev) |
|
3265 | l1 = self.rawsize(rev) | |
3246 | l2 = len(self.rawdata(node)) |
|
3266 | l2 = len(self.rawdata(node)) | |
3247 |
|
3267 | |||
3248 | if l1 != l2: |
|
3268 | if l1 != l2: | |
3249 | yield revlogproblem( |
|
3269 | yield revlogproblem( | |
3250 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), |
|
3270 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), | |
3251 | node=node, |
|
3271 | node=node, | |
3252 | ) |
|
3272 | ) | |
3253 |
|
3273 | |||
3254 | except error.CensoredNodeError: |
|
3274 | except error.CensoredNodeError: | |
3255 | if state[b'erroroncensored']: |
|
3275 | if state[b'erroroncensored']: | |
3256 | yield revlogproblem( |
|
3276 | yield revlogproblem( | |
3257 | error=_(b'censored file data'), node=node |
|
3277 | error=_(b'censored file data'), node=node | |
3258 | ) |
|
3278 | ) | |
3259 | state[b'skipread'].add(node) |
|
3279 | state[b'skipread'].add(node) | |
3260 | except Exception as e: |
|
3280 | except Exception as e: | |
3261 | yield revlogproblem( |
|
3281 | yield revlogproblem( | |
3262 | error=_(b'unpacking %s: %s') |
|
3282 | error=_(b'unpacking %s: %s') | |
3263 | % (short(node), stringutil.forcebytestr(e)), |
|
3283 | % (short(node), stringutil.forcebytestr(e)), | |
3264 | node=node, |
|
3284 | node=node, | |
3265 | ) |
|
3285 | ) | |
3266 | state[b'skipread'].add(node) |
|
3286 | state[b'skipread'].add(node) | |
3267 |
|
3287 | |||
3268 | def storageinfo( |
|
3288 | def storageinfo( | |
3269 | self, |
|
3289 | self, | |
3270 | exclusivefiles=False, |
|
3290 | exclusivefiles=False, | |
3271 | sharedfiles=False, |
|
3291 | sharedfiles=False, | |
3272 | revisionscount=False, |
|
3292 | revisionscount=False, | |
3273 | trackedsize=False, |
|
3293 | trackedsize=False, | |
3274 | storedsize=False, |
|
3294 | storedsize=False, | |
3275 | ): |
|
3295 | ): | |
3276 | d = {} |
|
3296 | d = {} | |
3277 |
|
3297 | |||
3278 | if exclusivefiles: |
|
3298 | if exclusivefiles: | |
3279 | d[b'exclusivefiles'] = [(self.opener, self._indexfile)] |
|
3299 | d[b'exclusivefiles'] = [(self.opener, self._indexfile)] | |
3280 | if not self._inline: |
|
3300 | if not self._inline: | |
3281 | d[b'exclusivefiles'].append((self.opener, self._datafile)) |
|
3301 | d[b'exclusivefiles'].append((self.opener, self._datafile)) | |
3282 |
|
3302 | |||
3283 | if sharedfiles: |
|
3303 | if sharedfiles: | |
3284 | d[b'sharedfiles'] = [] |
|
3304 | d[b'sharedfiles'] = [] | |
3285 |
|
3305 | |||
3286 | if revisionscount: |
|
3306 | if revisionscount: | |
3287 | d[b'revisionscount'] = len(self) |
|
3307 | d[b'revisionscount'] = len(self) | |
3288 |
|
3308 | |||
3289 | if trackedsize: |
|
3309 | if trackedsize: | |
3290 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) |
|
3310 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) | |
3291 |
|
3311 | |||
3292 | if storedsize: |
|
3312 | if storedsize: | |
3293 | d[b'storedsize'] = sum( |
|
3313 | d[b'storedsize'] = sum( | |
3294 | self.opener.stat(path).st_size for path in self.files() |
|
3314 | self.opener.stat(path).st_size for path in self.files() | |
3295 | ) |
|
3315 | ) | |
3296 |
|
3316 | |||
3297 | return d |
|
3317 | return d | |
3298 |
|
3318 | |||
3299 | def rewrite_sidedata(self, transaction, helpers, startrev, endrev): |
|
3319 | def rewrite_sidedata(self, transaction, helpers, startrev, endrev): | |
3300 | if not self.hassidedata: |
|
3320 | if not self.hassidedata: | |
3301 | return |
|
3321 | return | |
3302 | # revlog formats with sidedata support does not support inline |
|
3322 | # revlog formats with sidedata support does not support inline | |
3303 | assert not self._inline |
|
3323 | assert not self._inline | |
3304 | if not helpers[1] and not helpers[2]: |
|
3324 | if not helpers[1] and not helpers[2]: | |
3305 | # Nothing to generate or remove |
|
3325 | # Nothing to generate or remove | |
3306 | return |
|
3326 | return | |
3307 |
|
3327 | |||
3308 | new_entries = [] |
|
3328 | new_entries = [] | |
3309 | # append the new sidedata |
|
3329 | # append the new sidedata | |
3310 | with self._writing(transaction): |
|
3330 | with self._writing(transaction): | |
3311 | ifh, dfh = self._writinghandles |
|
3331 | ifh, dfh = self._writinghandles | |
3312 | if self._docket is not None: |
|
3332 | if self._docket is not None: | |
3313 | dfh.seek(self._docket.data_end, os.SEEK_SET) |
|
3333 | dfh.seek(self._docket.data_end, os.SEEK_SET) | |
3314 | else: |
|
3334 | else: | |
3315 | dfh.seek(0, os.SEEK_END) |
|
3335 | dfh.seek(0, os.SEEK_END) | |
3316 |
|
3336 | |||
3317 | current_offset = dfh.tell() |
|
3337 | current_offset = dfh.tell() | |
3318 | for rev in range(startrev, endrev + 1): |
|
3338 | for rev in range(startrev, endrev + 1): | |
3319 | entry = self.index[rev] |
|
3339 | entry = self.index[rev] | |
3320 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( |
|
3340 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( | |
3321 | store=self, |
|
3341 | store=self, | |
3322 | sidedata_helpers=helpers, |
|
3342 | sidedata_helpers=helpers, | |
3323 | sidedata={}, |
|
3343 | sidedata={}, | |
3324 | rev=rev, |
|
3344 | rev=rev, | |
3325 | ) |
|
3345 | ) | |
3326 |
|
3346 | |||
3327 | serialized_sidedata = sidedatautil.serialize_sidedata( |
|
3347 | serialized_sidedata = sidedatautil.serialize_sidedata( | |
3328 | new_sidedata |
|
3348 | new_sidedata | |
3329 | ) |
|
3349 | ) | |
3330 | if entry[8] != 0 or entry[9] != 0: |
|
3350 | if entry[8] != 0 or entry[9] != 0: | |
3331 | # rewriting entries that already have sidedata is not |
|
3351 | # rewriting entries that already have sidedata is not | |
3332 | # supported yet, because it introduces garbage data in the |
|
3352 | # supported yet, because it introduces garbage data in the | |
3333 | # revlog. |
|
3353 | # revlog. | |
3334 | msg = b"rewriting existing sidedata is not supported yet" |
|
3354 | msg = b"rewriting existing sidedata is not supported yet" | |
3335 | raise error.Abort(msg) |
|
3355 | raise error.Abort(msg) | |
3336 |
|
3356 | |||
3337 | # Apply (potential) flags to add and to remove after running |
|
3357 | # Apply (potential) flags to add and to remove after running | |
3338 | # the sidedata helpers |
|
3358 | # the sidedata helpers | |
3339 | new_offset_flags = entry[0] | flags[0] & ~flags[1] |
|
3359 | new_offset_flags = entry[0] | flags[0] & ~flags[1] | |
3340 | entry_update = ( |
|
3360 | entry_update = ( | |
3341 | current_offset, |
|
3361 | current_offset, | |
3342 | len(serialized_sidedata), |
|
3362 | len(serialized_sidedata), | |
3343 | new_offset_flags, |
|
3363 | new_offset_flags, | |
3344 | ) |
|
3364 | ) | |
3345 |
|
3365 | |||
3346 | # the sidedata computation might have move the file cursors around |
|
3366 | # the sidedata computation might have move the file cursors around | |
3347 | dfh.seek(current_offset, os.SEEK_SET) |
|
3367 | dfh.seek(current_offset, os.SEEK_SET) | |
3348 | dfh.write(serialized_sidedata) |
|
3368 | dfh.write(serialized_sidedata) | |
3349 | new_entries.append(entry_update) |
|
3369 | new_entries.append(entry_update) | |
3350 | current_offset += len(serialized_sidedata) |
|
3370 | current_offset += len(serialized_sidedata) | |
3351 | if self._docket is not None: |
|
3371 | if self._docket is not None: | |
3352 | self._docket.data_end = dfh.tell() |
|
3372 | self._docket.data_end = dfh.tell() | |
3353 |
|
3373 | |||
3354 | # rewrite the new index entries |
|
3374 | # rewrite the new index entries | |
3355 | ifh.seek(startrev * self.index.entry_size) |
|
3375 | ifh.seek(startrev * self.index.entry_size) | |
3356 | for i, e in enumerate(new_entries): |
|
3376 | for i, e in enumerate(new_entries): | |
3357 | rev = startrev + i |
|
3377 | rev = startrev + i | |
3358 | self.index.replace_sidedata_info(rev, *e) |
|
3378 | self.index.replace_sidedata_info(rev, *e) | |
3359 | packed = self.index.entry_binary(rev) |
|
3379 | packed = self.index.entry_binary(rev) | |
3360 | if rev == 0 and self._docket is None: |
|
3380 | if rev == 0 and self._docket is None: | |
3361 | header = self._format_flags | self._format_version |
|
3381 | header = self._format_flags | self._format_version | |
3362 | header = self.index.pack_header(header) |
|
3382 | header = self.index.pack_header(header) | |
3363 | packed = header + packed |
|
3383 | packed = header + packed | |
3364 | ifh.write(packed) |
|
3384 | ifh.write(packed) |
@@ -1,169 +1,179 b'' | |||||
1 | # revlogdeltas.py - constant used for revlog logic. |
|
1 | # revlogdeltas.py - constant used for revlog logic. | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> |
|
3 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> | |
4 | # Copyright 2018 Octobus <contact@octobus.net> |
|
4 | # Copyright 2018 Octobus <contact@octobus.net> | |
5 | # |
|
5 | # | |
6 | # This software may be used and distributed according to the terms of the |
|
6 | # This software may be used and distributed according to the terms of the | |
7 | # GNU General Public License version 2 or any later version. |
|
7 | # GNU General Public License version 2 or any later version. | |
8 | """Helper class to compute deltas stored inside revlogs""" |
|
8 | """Helper class to compute deltas stored inside revlogs""" | |
9 |
|
9 | |||
10 | from __future__ import absolute_import |
|
10 | from __future__ import absolute_import | |
11 |
|
11 | |||
12 | import struct |
|
12 | import struct | |
13 |
|
13 | |||
14 | from ..interfaces import repository |
|
14 | from ..interfaces import repository | |
15 |
|
15 | |||
16 | ### Internal utily constants |
|
16 | ### Internal utily constants | |
17 |
|
17 | |||
18 | KIND_CHANGELOG = 1001 # over 256 to not be comparable with a bytes |
|
18 | KIND_CHANGELOG = 1001 # over 256 to not be comparable with a bytes | |
19 | KIND_MANIFESTLOG = 1002 |
|
19 | KIND_MANIFESTLOG = 1002 | |
20 | KIND_FILELOG = 1003 |
|
20 | KIND_FILELOG = 1003 | |
21 | KIND_OTHER = 1004 |
|
21 | KIND_OTHER = 1004 | |
22 |
|
22 | |||
23 | ALL_KINDS = { |
|
23 | ALL_KINDS = { | |
24 | KIND_CHANGELOG, |
|
24 | KIND_CHANGELOG, | |
25 | KIND_MANIFESTLOG, |
|
25 | KIND_MANIFESTLOG, | |
26 | KIND_FILELOG, |
|
26 | KIND_FILELOG, | |
27 | KIND_OTHER, |
|
27 | KIND_OTHER, | |
28 | } |
|
28 | } | |
29 |
|
29 | |||
30 | ### main revlog header |
|
30 | ### main revlog header | |
31 |
|
31 | |||
32 | INDEX_HEADER = struct.Struct(b">I") |
|
32 | INDEX_HEADER = struct.Struct(b">I") | |
33 |
|
33 | |||
34 | ## revlog version |
|
34 | ## revlog version | |
35 | REVLOGV0 = 0 |
|
35 | REVLOGV0 = 0 | |
36 | REVLOGV1 = 1 |
|
36 | REVLOGV1 = 1 | |
37 | # Dummy value until file format is finalized. |
|
37 | # Dummy value until file format is finalized. | |
38 | REVLOGV2 = 0xDEAD |
|
38 | REVLOGV2 = 0xDEAD | |
39 |
|
39 | |||
40 | ## global revlog header flags |
|
40 | ## global revlog header flags | |
41 | # Shared across v1 and v2. |
|
41 | # Shared across v1 and v2. | |
42 | FLAG_INLINE_DATA = 1 << 16 |
|
42 | FLAG_INLINE_DATA = 1 << 16 | |
43 | # Only used by v1, implied by v2. |
|
43 | # Only used by v1, implied by v2. | |
44 | FLAG_GENERALDELTA = 1 << 17 |
|
44 | FLAG_GENERALDELTA = 1 << 17 | |
45 | REVLOG_DEFAULT_FLAGS = FLAG_INLINE_DATA |
|
45 | REVLOG_DEFAULT_FLAGS = FLAG_INLINE_DATA | |
46 | REVLOG_DEFAULT_FORMAT = REVLOGV1 |
|
46 | REVLOG_DEFAULT_FORMAT = REVLOGV1 | |
47 | REVLOG_DEFAULT_VERSION = REVLOG_DEFAULT_FORMAT | REVLOG_DEFAULT_FLAGS |
|
47 | REVLOG_DEFAULT_VERSION = REVLOG_DEFAULT_FORMAT | REVLOG_DEFAULT_FLAGS | |
48 | REVLOGV0_FLAGS = 0 |
|
48 | REVLOGV0_FLAGS = 0 | |
49 | REVLOGV1_FLAGS = FLAG_INLINE_DATA | FLAG_GENERALDELTA |
|
49 | REVLOGV1_FLAGS = FLAG_INLINE_DATA | FLAG_GENERALDELTA | |
50 | REVLOGV2_FLAGS = FLAG_INLINE_DATA |
|
50 | REVLOGV2_FLAGS = FLAG_INLINE_DATA | |
51 |
|
51 | |||
52 | ### individual entry |
|
52 | ### individual entry | |
53 |
|
53 | |||
54 | ## index v0: |
|
54 | ## index v0: | |
55 | # 4 bytes: offset |
|
55 | # 4 bytes: offset | |
56 | # 4 bytes: compressed length |
|
56 | # 4 bytes: compressed length | |
57 | # 4 bytes: base rev |
|
57 | # 4 bytes: base rev | |
58 | # 4 bytes: link rev |
|
58 | # 4 bytes: link rev | |
59 | # 20 bytes: parent 1 nodeid |
|
59 | # 20 bytes: parent 1 nodeid | |
60 | # 20 bytes: parent 2 nodeid |
|
60 | # 20 bytes: parent 2 nodeid | |
61 | # 20 bytes: nodeid |
|
61 | # 20 bytes: nodeid | |
62 | INDEX_ENTRY_V0 = struct.Struct(b">4l20s20s20s") |
|
62 | INDEX_ENTRY_V0 = struct.Struct(b">4l20s20s20s") | |
63 |
|
63 | |||
64 | ## index v1 |
|
64 | ## index v1 | |
65 | # 6 bytes: offset |
|
65 | # 6 bytes: offset | |
66 | # 2 bytes: flags |
|
66 | # 2 bytes: flags | |
67 | # 4 bytes: compressed length |
|
67 | # 4 bytes: compressed length | |
68 | # 4 bytes: uncompressed length |
|
68 | # 4 bytes: uncompressed length | |
69 | # 4 bytes: base rev |
|
69 | # 4 bytes: base rev | |
70 | # 4 bytes: link rev |
|
70 | # 4 bytes: link rev | |
71 | # 4 bytes: parent 1 rev |
|
71 | # 4 bytes: parent 1 rev | |
72 | # 4 bytes: parent 2 rev |
|
72 | # 4 bytes: parent 2 rev | |
73 | # 32 bytes: nodeid |
|
73 | # 32 bytes: nodeid | |
74 | INDEX_ENTRY_V1 = struct.Struct(b">Qiiiiii20s12x") |
|
74 | INDEX_ENTRY_V1 = struct.Struct(b">Qiiiiii20s12x") | |
75 | assert INDEX_ENTRY_V1.size == 32 * 2 |
|
75 | assert INDEX_ENTRY_V1.size == 32 * 2 | |
76 |
|
76 | |||
77 | # 6 bytes: offset |
|
77 | # 6 bytes: offset | |
78 | # 2 bytes: flags |
|
78 | # 2 bytes: flags | |
79 | # 4 bytes: compressed length |
|
79 | # 4 bytes: compressed length | |
80 | # 4 bytes: uncompressed length |
|
80 | # 4 bytes: uncompressed length | |
81 | # 4 bytes: base rev |
|
81 | # 4 bytes: base rev | |
82 | # 4 bytes: link rev |
|
82 | # 4 bytes: link rev | |
83 | # 4 bytes: parent 1 rev |
|
83 | # 4 bytes: parent 1 rev | |
84 | # 4 bytes: parent 2 rev |
|
84 | # 4 bytes: parent 2 rev | |
85 | # 32 bytes: nodeid |
|
85 | # 32 bytes: nodeid | |
86 | # 8 bytes: sidedata offset |
|
86 | # 8 bytes: sidedata offset | |
87 | # 4 bytes: sidedata compressed length |
|
87 | # 4 bytes: sidedata compressed length | |
88 | # 1 bytes: compression mode (2 lower bit are data_compression_mode) |
|
88 | # 1 bytes: compression mode (2 lower bit are data_compression_mode) | |
89 | # 19 bytes: Padding to align to 96 bytes (see RevlogV2Plan wiki page) |
|
89 | # 19 bytes: Padding to align to 96 bytes (see RevlogV2Plan wiki page) | |
90 | INDEX_ENTRY_V2 = struct.Struct(b">Qiiiiii20s12xQiB19x") |
|
90 | INDEX_ENTRY_V2 = struct.Struct(b">Qiiiiii20s12xQiB19x") | |
91 | assert INDEX_ENTRY_V2.size == 32 * 3, INDEX_ENTRY_V2.size |
|
91 | assert INDEX_ENTRY_V2.size == 32 * 3, INDEX_ENTRY_V2.size | |
92 |
|
92 | |||
93 | # revlog index flags |
|
93 | # revlog index flags | |
94 |
|
94 | |||
95 | # For historical reasons, revlog's internal flags were exposed via the |
|
95 | # For historical reasons, revlog's internal flags were exposed via the | |
96 | # wire protocol and are even exposed in parts of the storage APIs. |
|
96 | # wire protocol and are even exposed in parts of the storage APIs. | |
97 |
|
97 | |||
98 | # revision has censor metadata, must be verified |
|
98 | # revision has censor metadata, must be verified | |
99 | REVIDX_ISCENSORED = repository.REVISION_FLAG_CENSORED |
|
99 | REVIDX_ISCENSORED = repository.REVISION_FLAG_CENSORED | |
100 | # revision hash does not match data (narrowhg) |
|
100 | # revision hash does not match data (narrowhg) | |
101 | REVIDX_ELLIPSIS = repository.REVISION_FLAG_ELLIPSIS |
|
101 | REVIDX_ELLIPSIS = repository.REVISION_FLAG_ELLIPSIS | |
102 | # revision data is stored externally |
|
102 | # revision data is stored externally | |
103 | REVIDX_EXTSTORED = repository.REVISION_FLAG_EXTSTORED |
|
103 | REVIDX_EXTSTORED = repository.REVISION_FLAG_EXTSTORED | |
104 | # revision changes files in a way that could affect copy tracing. |
|
104 | # revision changes files in a way that could affect copy tracing. | |
105 | REVIDX_HASCOPIESINFO = repository.REVISION_FLAG_HASCOPIESINFO |
|
105 | REVIDX_HASCOPIESINFO = repository.REVISION_FLAG_HASCOPIESINFO | |
106 | REVIDX_DEFAULT_FLAGS = 0 |
|
106 | REVIDX_DEFAULT_FLAGS = 0 | |
107 | # stable order in which flags need to be processed and their processors applied |
|
107 | # stable order in which flags need to be processed and their processors applied | |
108 | REVIDX_FLAGS_ORDER = [ |
|
108 | REVIDX_FLAGS_ORDER = [ | |
109 | REVIDX_ISCENSORED, |
|
109 | REVIDX_ISCENSORED, | |
110 | REVIDX_ELLIPSIS, |
|
110 | REVIDX_ELLIPSIS, | |
111 | REVIDX_EXTSTORED, |
|
111 | REVIDX_EXTSTORED, | |
112 | REVIDX_HASCOPIESINFO, |
|
112 | REVIDX_HASCOPIESINFO, | |
113 | ] |
|
113 | ] | |
114 |
|
114 | |||
115 | # bitmark for flags that could cause rawdata content change |
|
115 | # bitmark for flags that could cause rawdata content change | |
116 | REVIDX_RAWTEXT_CHANGING_FLAGS = REVIDX_ISCENSORED | REVIDX_EXTSTORED |
|
116 | REVIDX_RAWTEXT_CHANGING_FLAGS = REVIDX_ISCENSORED | REVIDX_EXTSTORED | |
117 |
|
117 | |||
118 | ## chunk compression mode constants: |
|
118 | ## chunk compression mode constants: | |
119 | # These constants are used in revlog version >=2 to denote the compression used |
|
119 | # These constants are used in revlog version >=2 to denote the compression used | |
120 | # for a chunk. |
|
120 | # for a chunk. | |
121 |
|
121 | |||
122 | # Chunk use no compression, the data stored on disk can be directly use as |
|
122 | # Chunk use no compression, the data stored on disk can be directly use as | |
123 | # chunk value. Without any header information prefixed. |
|
123 | # chunk value. Without any header information prefixed. | |
124 | COMP_MODE_PLAIN = 0 |
|
124 | COMP_MODE_PLAIN = 0 | |
125 |
|
125 | |||
|
126 | # Chunk use the "default compression" for the revlog (usually defined in the | |||
|
127 | # revlog docket). A header is still used. | |||
|
128 | # | |||
|
129 | # XXX: keeping a header is probably not useful and we should probably drop it. | |||
|
130 | # | |||
|
131 | # XXX: The value of allow mixed type of compression in the revlog is unclear | |||
|
132 | # and we should consider making PLAIN/DEFAULT the only available mode for | |||
|
133 | # revlog v2, disallowing INLINE mode. | |||
|
134 | COMP_MODE_DEFAULT = 1 | |||
|
135 | ||||
126 | # Chunk use a compression mode stored "inline" at the start of the chunk |
|
136 | # Chunk use a compression mode stored "inline" at the start of the chunk | |
127 | # itself. This is the mode always used for revlog version "0" and "1" |
|
137 | # itself. This is the mode always used for revlog version "0" and "1" | |
128 | COMP_MODE_INLINE = 2 |
|
138 | COMP_MODE_INLINE = 2 | |
129 |
|
139 | |||
130 | SUPPORTED_FLAGS = { |
|
140 | SUPPORTED_FLAGS = { | |
131 | REVLOGV0: REVLOGV0_FLAGS, |
|
141 | REVLOGV0: REVLOGV0_FLAGS, | |
132 | REVLOGV1: REVLOGV1_FLAGS, |
|
142 | REVLOGV1: REVLOGV1_FLAGS, | |
133 | REVLOGV2: REVLOGV2_FLAGS, |
|
143 | REVLOGV2: REVLOGV2_FLAGS, | |
134 | } |
|
144 | } | |
135 |
|
145 | |||
136 | _no = lambda flags: False |
|
146 | _no = lambda flags: False | |
137 | _yes = lambda flags: True |
|
147 | _yes = lambda flags: True | |
138 |
|
148 | |||
139 |
|
149 | |||
140 | def _from_flag(flag): |
|
150 | def _from_flag(flag): | |
141 | return lambda flags: bool(flags & flag) |
|
151 | return lambda flags: bool(flags & flag) | |
142 |
|
152 | |||
143 |
|
153 | |||
144 | FEATURES_BY_VERSION = { |
|
154 | FEATURES_BY_VERSION = { | |
145 | REVLOGV0: { |
|
155 | REVLOGV0: { | |
146 | b'inline': _no, |
|
156 | b'inline': _no, | |
147 | b'generaldelta': _no, |
|
157 | b'generaldelta': _no, | |
148 | b'sidedata': False, |
|
158 | b'sidedata': False, | |
149 | b'docket': False, |
|
159 | b'docket': False, | |
150 | }, |
|
160 | }, | |
151 | REVLOGV1: { |
|
161 | REVLOGV1: { | |
152 | b'inline': _from_flag(FLAG_INLINE_DATA), |
|
162 | b'inline': _from_flag(FLAG_INLINE_DATA), | |
153 | b'generaldelta': _from_flag(FLAG_GENERALDELTA), |
|
163 | b'generaldelta': _from_flag(FLAG_GENERALDELTA), | |
154 | b'sidedata': False, |
|
164 | b'sidedata': False, | |
155 | b'docket': False, |
|
165 | b'docket': False, | |
156 | }, |
|
166 | }, | |
157 | REVLOGV2: { |
|
167 | REVLOGV2: { | |
158 | # The point of inline-revlog is to reduce the number of files used in |
|
168 | # The point of inline-revlog is to reduce the number of files used in | |
159 | # the store. Using a docket defeat this purpose. So we needs other |
|
169 | # the store. Using a docket defeat this purpose. So we needs other | |
160 | # means to reduce the number of files for revlogv2. |
|
170 | # means to reduce the number of files for revlogv2. | |
161 | b'inline': _no, |
|
171 | b'inline': _no, | |
162 | b'generaldelta': _yes, |
|
172 | b'generaldelta': _yes, | |
163 | b'sidedata': True, |
|
173 | b'sidedata': True, | |
164 | b'docket': True, |
|
174 | b'docket': True, | |
165 | }, |
|
175 | }, | |
166 | } |
|
176 | } | |
167 |
|
177 | |||
168 |
|
178 | |||
169 | SPARSE_REVLOG_MAX_CHAIN_LENGTH = 1000 |
|
179 | SPARSE_REVLOG_MAX_CHAIN_LENGTH = 1000 |
@@ -1,167 +1,179 b'' | |||||
1 | # docket - code related to revlog "docket" |
|
1 | # docket - code related to revlog "docket" | |
2 | # |
|
2 | # | |
3 | # Copyright 2021 Pierre-Yves David <pierre-yves.david@octobus.net> |
|
3 | # Copyright 2021 Pierre-Yves David <pierre-yves.david@octobus.net> | |
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 | ### Revlog docket file |
|
8 | ### Revlog docket file | |
9 | # |
|
9 | # | |
10 | # The revlog is stored on disk using multiple files: |
|
10 | # The revlog is stored on disk using multiple files: | |
11 | # |
|
11 | # | |
12 | # * a small docket file, containing metadata and a pointer, |
|
12 | # * a small docket file, containing metadata and a pointer, | |
13 | # |
|
13 | # | |
14 | # * an index file, containing fixed width information about revisions, |
|
14 | # * an index file, containing fixed width information about revisions, | |
15 | # |
|
15 | # | |
16 | # * a data file, containing variable width data for these revisions, |
|
16 | # * a data file, containing variable width data for these revisions, | |
17 |
|
17 | |||
18 | from __future__ import absolute_import |
|
18 | from __future__ import absolute_import | |
19 |
|
19 | |||
20 | import struct |
|
20 | import struct | |
21 |
|
21 | |||
22 | from .. import ( |
|
22 | from .. import ( | |
23 | error, |
|
23 | error, | |
|
24 | util, | |||
24 | ) |
|
25 | ) | |
25 |
|
26 | |||
26 | from . import ( |
|
27 | from . import ( | |
27 | constants, |
|
28 | constants, | |
28 | ) |
|
29 | ) | |
29 |
|
30 | |||
30 | # Docket format |
|
31 | # Docket format | |
31 | # |
|
32 | # | |
32 | # * 4 bytes: revlog version |
|
33 | # * 4 bytes: revlog version | |
33 | # | This is mandatory as docket must be compatible with the previous |
|
34 | # | This is mandatory as docket must be compatible with the previous | |
34 | # | revlog index header. |
|
35 | # | revlog index header. | |
35 | # * 8 bytes: size of index-data |
|
36 | # * 8 bytes: size of index-data | |
36 | # * 8 bytes: pending size of index-data |
|
37 | # * 8 bytes: pending size of index-data | |
37 | # * 8 bytes: size of data |
|
38 | # * 8 bytes: size of data | |
38 | # * 8 bytes: pending size of data |
|
39 | # * 8 bytes: pending size of data | |
39 | S_HEADER = struct.Struct(constants.INDEX_HEADER.format + 'LLLL') |
|
40 | # * 1 bytes: default compression header | |
|
41 | S_HEADER = struct.Struct(constants.INDEX_HEADER.format + 'LLLLc') | |||
40 |
|
42 | |||
41 |
|
43 | |||
42 | class RevlogDocket(object): |
|
44 | class RevlogDocket(object): | |
43 | """metadata associated with revlog""" |
|
45 | """metadata associated with revlog""" | |
44 |
|
46 | |||
45 | def __init__( |
|
47 | def __init__( | |
46 | self, |
|
48 | self, | |
47 | revlog, |
|
49 | revlog, | |
48 | use_pending=False, |
|
50 | use_pending=False, | |
49 | version_header=None, |
|
51 | version_header=None, | |
50 | index_end=0, |
|
52 | index_end=0, | |
51 | pending_index_end=0, |
|
53 | pending_index_end=0, | |
52 | data_end=0, |
|
54 | data_end=0, | |
53 | pending_data_end=0, |
|
55 | pending_data_end=0, | |
|
56 | default_compression_header=None, | |||
54 | ): |
|
57 | ): | |
55 | self._version_header = version_header |
|
58 | self._version_header = version_header | |
56 | self._read_only = bool(use_pending) |
|
59 | self._read_only = bool(use_pending) | |
57 | self._dirty = False |
|
60 | self._dirty = False | |
58 | self._radix = revlog.radix |
|
61 | self._radix = revlog.radix | |
59 | self._path = revlog._docket_file |
|
62 | self._path = revlog._docket_file | |
60 | self._opener = revlog.opener |
|
63 | self._opener = revlog.opener | |
61 | # thes asserts should be True as long as we have a single index filename |
|
64 | # thes asserts should be True as long as we have a single index filename | |
62 | assert index_end <= pending_index_end |
|
65 | assert index_end <= pending_index_end | |
63 | assert data_end <= pending_data_end |
|
66 | assert data_end <= pending_data_end | |
64 | self._initial_index_end = index_end |
|
67 | self._initial_index_end = index_end | |
65 | self._pending_index_end = pending_index_end |
|
68 | self._pending_index_end = pending_index_end | |
66 | self._initial_data_end = data_end |
|
69 | self._initial_data_end = data_end | |
67 | self._pending_data_end = pending_data_end |
|
70 | self._pending_data_end = pending_data_end | |
68 | if use_pending: |
|
71 | if use_pending: | |
69 | self._index_end = self._pending_index_end |
|
72 | self._index_end = self._pending_index_end | |
70 | self._data_end = self._pending_data_end |
|
73 | self._data_end = self._pending_data_end | |
71 | else: |
|
74 | else: | |
72 | self._index_end = self._initial_index_end |
|
75 | self._index_end = self._initial_index_end | |
73 | self._data_end = self._initial_data_end |
|
76 | self._data_end = self._initial_data_end | |
|
77 | self.default_compression_header = default_compression_header | |||
74 |
|
78 | |||
75 | def index_filepath(self): |
|
79 | def index_filepath(self): | |
76 | """file path to the current index file associated to this docket""" |
|
80 | """file path to the current index file associated to this docket""" | |
77 | # very simplistic version at first |
|
81 | # very simplistic version at first | |
78 | return b"%s.idx" % self._radix |
|
82 | return b"%s.idx" % self._radix | |
79 |
|
83 | |||
80 | @property |
|
84 | @property | |
81 | def index_end(self): |
|
85 | def index_end(self): | |
82 | return self._index_end |
|
86 | return self._index_end | |
83 |
|
87 | |||
84 | @index_end.setter |
|
88 | @index_end.setter | |
85 | def index_end(self, new_size): |
|
89 | def index_end(self, new_size): | |
86 | if new_size != self._index_end: |
|
90 | if new_size != self._index_end: | |
87 | self._index_end = new_size |
|
91 | self._index_end = new_size | |
88 | self._dirty = True |
|
92 | self._dirty = True | |
89 |
|
93 | |||
90 | @property |
|
94 | @property | |
91 | def data_end(self): |
|
95 | def data_end(self): | |
92 | return self._data_end |
|
96 | return self._data_end | |
93 |
|
97 | |||
94 | @data_end.setter |
|
98 | @data_end.setter | |
95 | def data_end(self, new_size): |
|
99 | def data_end(self, new_size): | |
96 | if new_size != self._data_end: |
|
100 | if new_size != self._data_end: | |
97 | self._data_end = new_size |
|
101 | self._data_end = new_size | |
98 | self._dirty = True |
|
102 | self._dirty = True | |
99 |
|
103 | |||
100 | def write(self, transaction, pending=False, stripping=False): |
|
104 | def write(self, transaction, pending=False, stripping=False): | |
101 | """write the modification of disk if any |
|
105 | """write the modification of disk if any | |
102 |
|
106 | |||
103 | This make the new content visible to all process""" |
|
107 | This make the new content visible to all process""" | |
104 | if not self._dirty: |
|
108 | if not self._dirty: | |
105 | return False |
|
109 | return False | |
106 | else: |
|
110 | else: | |
107 | if self._read_only: |
|
111 | if self._read_only: | |
108 | msg = b'writing read-only docket: %s' |
|
112 | msg = b'writing read-only docket: %s' | |
109 | msg %= self._path |
|
113 | msg %= self._path | |
110 | raise error.ProgrammingError(msg) |
|
114 | raise error.ProgrammingError(msg) | |
111 | if not stripping: |
|
115 | if not stripping: | |
112 | # XXX we could, leverage the docket while stripping. However it |
|
116 | # XXX we could, leverage the docket while stripping. However it | |
113 | # is not powerfull enough at the time of this comment |
|
117 | # is not powerfull enough at the time of this comment | |
114 | transaction.addbackup(self._path, location=b'store') |
|
118 | transaction.addbackup(self._path, location=b'store') | |
115 | with self._opener(self._path, mode=b'w', atomictemp=True) as f: |
|
119 | with self._opener(self._path, mode=b'w', atomictemp=True) as f: | |
116 | f.write(self._serialize(pending=pending)) |
|
120 | f.write(self._serialize(pending=pending)) | |
117 | # if pending we still need to the write final data eventually |
|
121 | # if pending we still need to the write final data eventually | |
118 | self._dirty = pending |
|
122 | self._dirty = pending | |
119 | return True |
|
123 | return True | |
120 |
|
124 | |||
121 | def _serialize(self, pending=False): |
|
125 | def _serialize(self, pending=False): | |
122 | if pending: |
|
126 | if pending: | |
123 | official_index_end = self._initial_index_end |
|
127 | official_index_end = self._initial_index_end | |
124 | official_data_end = self._initial_data_end |
|
128 | official_data_end = self._initial_data_end | |
125 | else: |
|
129 | else: | |
126 | official_index_end = self._index_end |
|
130 | official_index_end = self._index_end | |
127 | official_data_end = self._data_end |
|
131 | official_data_end = self._data_end | |
128 |
|
132 | |||
129 | # this assert should be True as long as we have a single index filename |
|
133 | # this assert should be True as long as we have a single index filename | |
130 | assert official_data_end <= self._data_end |
|
134 | assert official_data_end <= self._data_end | |
131 | data = ( |
|
135 | data = ( | |
132 | self._version_header, |
|
136 | self._version_header, | |
133 | official_index_end, |
|
137 | official_index_end, | |
134 | self._index_end, |
|
138 | self._index_end, | |
135 | official_data_end, |
|
139 | official_data_end, | |
136 | self._data_end, |
|
140 | self._data_end, | |
|
141 | self.default_compression_header, | |||
137 | ) |
|
142 | ) | |
138 | return S_HEADER.pack(*data) |
|
143 | return S_HEADER.pack(*data) | |
139 |
|
144 | |||
140 |
|
145 | |||
141 | def default_docket(revlog, version_header): |
|
146 | def default_docket(revlog, version_header): | |
142 | """given a revlog version a new docket object for the given revlog""" |
|
147 | """given a revlog version a new docket object for the given revlog""" | |
143 | if (version_header & 0xFFFF) != constants.REVLOGV2: |
|
148 | if (version_header & 0xFFFF) != constants.REVLOGV2: | |
144 | return None |
|
149 | return None | |
145 | docket = RevlogDocket(revlog, version_header=version_header) |
|
150 | comp = util.compengines[revlog._compengine].revlogheader() | |
|
151 | docket = RevlogDocket( | |||
|
152 | revlog, | |||
|
153 | version_header=version_header, | |||
|
154 | default_compression_header=comp, | |||
|
155 | ) | |||
146 | docket._dirty = True |
|
156 | docket._dirty = True | |
147 | return docket |
|
157 | return docket | |
148 |
|
158 | |||
149 |
|
159 | |||
150 | def parse_docket(revlog, data, use_pending=False): |
|
160 | def parse_docket(revlog, data, use_pending=False): | |
151 | """given some docket data return a docket object for the given revlog""" |
|
161 | """given some docket data return a docket object for the given revlog""" | |
152 | header = S_HEADER.unpack(data[: S_HEADER.size]) |
|
162 | header = S_HEADER.unpack(data[: S_HEADER.size]) | |
153 | version_header = header[0] |
|
163 | version_header = header[0] | |
154 | index_size = header[1] |
|
164 | index_size = header[1] | |
155 | pending_index_size = header[2] |
|
165 | pending_index_size = header[2] | |
156 | data_size = header[3] |
|
166 | data_size = header[3] | |
157 | pending_data_size = header[4] |
|
167 | pending_data_size = header[4] | |
|
168 | default_compression_header = header[5] | |||
158 | docket = RevlogDocket( |
|
169 | docket = RevlogDocket( | |
159 | revlog, |
|
170 | revlog, | |
160 | use_pending=use_pending, |
|
171 | use_pending=use_pending, | |
161 | version_header=version_header, |
|
172 | version_header=version_header, | |
162 | index_end=index_size, |
|
173 | index_end=index_size, | |
163 | pending_index_end=pending_index_size, |
|
174 | pending_index_end=pending_index_size, | |
164 | data_end=data_size, |
|
175 | data_end=data_size, | |
165 | pending_data_end=pending_data_size, |
|
176 | pending_data_end=pending_data_size, | |
|
177 | default_compression_header=default_compression_header, | |||
166 | ) |
|
178 | ) | |
167 | return docket |
|
179 | return docket |
General Comments 0
You need to be logged in to leave comments.
Login now