Show More
@@ -1,3410 +1,3478 b'' | |||||
1 | # revlog.py - storage back-end for mercurial |
|
1 | # revlog.py - storage back-end for mercurial | |
2 | # coding: utf8 |
|
2 | # coding: utf8 | |
3 | # |
|
3 | # | |
4 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> |
|
4 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> | |
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 |
|
8 | |||
9 | """Storage back-end for Mercurial. |
|
9 | """Storage back-end for Mercurial. | |
10 |
|
10 | |||
11 | This provides efficient delta storage with O(1) retrieve and append |
|
11 | This provides efficient delta storage with O(1) retrieve and append | |
12 | and O(changes) merge between branches. |
|
12 | and O(changes) merge between branches. | |
13 | """ |
|
13 | """ | |
14 |
|
14 | |||
15 |
|
15 | |||
16 | import binascii |
|
16 | import binascii | |
17 | import collections |
|
17 | import collections | |
18 | import contextlib |
|
18 | import contextlib | |
19 | import io |
|
19 | import io | |
20 | import os |
|
20 | import os | |
21 | import struct |
|
21 | import struct | |
22 | import weakref |
|
22 | import weakref | |
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 | CHANGELOGV2, |
|
38 | CHANGELOGV2, | |
39 | COMP_MODE_DEFAULT, |
|
39 | COMP_MODE_DEFAULT, | |
40 | COMP_MODE_INLINE, |
|
40 | COMP_MODE_INLINE, | |
41 | COMP_MODE_PLAIN, |
|
41 | COMP_MODE_PLAIN, | |
42 | DELTA_BASE_REUSE_NO, |
|
42 | DELTA_BASE_REUSE_NO, | |
43 | DELTA_BASE_REUSE_TRY, |
|
43 | DELTA_BASE_REUSE_TRY, | |
44 | ENTRY_RANK, |
|
44 | ENTRY_RANK, | |
45 | FEATURES_BY_VERSION, |
|
45 | FEATURES_BY_VERSION, | |
46 | FLAG_GENERALDELTA, |
|
46 | FLAG_GENERALDELTA, | |
47 | FLAG_INLINE_DATA, |
|
47 | FLAG_INLINE_DATA, | |
48 | INDEX_HEADER, |
|
48 | INDEX_HEADER, | |
49 | KIND_CHANGELOG, |
|
49 | KIND_CHANGELOG, | |
50 | KIND_FILELOG, |
|
50 | KIND_FILELOG, | |
51 | RANK_UNKNOWN, |
|
51 | RANK_UNKNOWN, | |
52 | REVLOGV0, |
|
52 | REVLOGV0, | |
53 | REVLOGV1, |
|
53 | REVLOGV1, | |
54 | REVLOGV1_FLAGS, |
|
54 | REVLOGV1_FLAGS, | |
55 | REVLOGV2, |
|
55 | REVLOGV2, | |
56 | REVLOGV2_FLAGS, |
|
56 | REVLOGV2_FLAGS, | |
57 | REVLOG_DEFAULT_FLAGS, |
|
57 | REVLOG_DEFAULT_FLAGS, | |
58 | REVLOG_DEFAULT_FORMAT, |
|
58 | REVLOG_DEFAULT_FORMAT, | |
59 | REVLOG_DEFAULT_VERSION, |
|
59 | REVLOG_DEFAULT_VERSION, | |
60 | SUPPORTED_FLAGS, |
|
60 | SUPPORTED_FLAGS, | |
61 | ) |
|
61 | ) | |
62 | from .revlogutils.flagutil import ( |
|
62 | from .revlogutils.flagutil import ( | |
63 | REVIDX_DEFAULT_FLAGS, |
|
63 | REVIDX_DEFAULT_FLAGS, | |
64 | REVIDX_ELLIPSIS, |
|
64 | REVIDX_ELLIPSIS, | |
65 | REVIDX_EXTSTORED, |
|
65 | REVIDX_EXTSTORED, | |
66 | REVIDX_FLAGS_ORDER, |
|
66 | REVIDX_FLAGS_ORDER, | |
67 | REVIDX_HASCOPIESINFO, |
|
67 | REVIDX_HASCOPIESINFO, | |
68 | REVIDX_ISCENSORED, |
|
68 | REVIDX_ISCENSORED, | |
69 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
69 | REVIDX_RAWTEXT_CHANGING_FLAGS, | |
70 | ) |
|
70 | ) | |
71 | from .thirdparty import attr |
|
71 | from .thirdparty import attr | |
72 | from . import ( |
|
72 | from . import ( | |
73 | ancestor, |
|
73 | ancestor, | |
74 | dagop, |
|
74 | dagop, | |
75 | error, |
|
75 | error, | |
76 | mdiff, |
|
76 | mdiff, | |
77 | policy, |
|
77 | policy, | |
78 | pycompat, |
|
78 | pycompat, | |
79 | revlogutils, |
|
79 | revlogutils, | |
80 | templatefilters, |
|
80 | templatefilters, | |
81 | util, |
|
81 | util, | |
82 | ) |
|
82 | ) | |
83 | from .interfaces import ( |
|
83 | from .interfaces import ( | |
84 | repository, |
|
84 | repository, | |
85 | util as interfaceutil, |
|
85 | util as interfaceutil, | |
86 | ) |
|
86 | ) | |
87 | from .revlogutils import ( |
|
87 | from .revlogutils import ( | |
88 | deltas as deltautil, |
|
88 | deltas as deltautil, | |
89 | docket as docketutil, |
|
89 | docket as docketutil, | |
90 | flagutil, |
|
90 | flagutil, | |
91 | nodemap as nodemaputil, |
|
91 | nodemap as nodemaputil, | |
92 | randomaccessfile, |
|
92 | randomaccessfile, | |
93 | revlogv0, |
|
93 | revlogv0, | |
94 | rewrite, |
|
94 | rewrite, | |
95 | sidedata as sidedatautil, |
|
95 | sidedata as sidedatautil, | |
96 | ) |
|
96 | ) | |
97 | from .utils import ( |
|
97 | from .utils import ( | |
98 | storageutil, |
|
98 | storageutil, | |
99 | stringutil, |
|
99 | stringutil, | |
100 | ) |
|
100 | ) | |
101 |
|
101 | |||
102 | # blanked usage of all the name to prevent pyflakes constraints |
|
102 | # blanked usage of all the name to prevent pyflakes constraints | |
103 | # We need these name available in the module for extensions. |
|
103 | # We need these name available in the module for extensions. | |
104 |
|
104 | |||
105 | REVLOGV0 |
|
105 | REVLOGV0 | |
106 | REVLOGV1 |
|
106 | REVLOGV1 | |
107 | REVLOGV2 |
|
107 | REVLOGV2 | |
108 | CHANGELOGV2 |
|
108 | CHANGELOGV2 | |
109 | FLAG_INLINE_DATA |
|
109 | FLAG_INLINE_DATA | |
110 | FLAG_GENERALDELTA |
|
110 | FLAG_GENERALDELTA | |
111 | REVLOG_DEFAULT_FLAGS |
|
111 | REVLOG_DEFAULT_FLAGS | |
112 | REVLOG_DEFAULT_FORMAT |
|
112 | REVLOG_DEFAULT_FORMAT | |
113 | REVLOG_DEFAULT_VERSION |
|
113 | REVLOG_DEFAULT_VERSION | |
114 | REVLOGV1_FLAGS |
|
114 | REVLOGV1_FLAGS | |
115 | REVLOGV2_FLAGS |
|
115 | REVLOGV2_FLAGS | |
116 | REVIDX_ISCENSORED |
|
116 | REVIDX_ISCENSORED | |
117 | REVIDX_ELLIPSIS |
|
117 | REVIDX_ELLIPSIS | |
118 | REVIDX_HASCOPIESINFO |
|
118 | REVIDX_HASCOPIESINFO | |
119 | REVIDX_EXTSTORED |
|
119 | REVIDX_EXTSTORED | |
120 | REVIDX_DEFAULT_FLAGS |
|
120 | REVIDX_DEFAULT_FLAGS | |
121 | REVIDX_FLAGS_ORDER |
|
121 | REVIDX_FLAGS_ORDER | |
122 | REVIDX_RAWTEXT_CHANGING_FLAGS |
|
122 | REVIDX_RAWTEXT_CHANGING_FLAGS | |
123 |
|
123 | |||
124 | parsers = policy.importmod('parsers') |
|
124 | parsers = policy.importmod('parsers') | |
125 | rustancestor = policy.importrust('ancestor') |
|
125 | rustancestor = policy.importrust('ancestor') | |
126 | rustdagop = policy.importrust('dagop') |
|
126 | rustdagop = policy.importrust('dagop') | |
127 | rustrevlog = policy.importrust('revlog') |
|
127 | rustrevlog = policy.importrust('revlog') | |
128 |
|
128 | |||
129 | # Aliased for performance. |
|
129 | # Aliased for performance. | |
130 | _zlibdecompress = zlib.decompress |
|
130 | _zlibdecompress = zlib.decompress | |
131 |
|
131 | |||
132 | # max size of inline data embedded into a revlog |
|
132 | # max size of inline data embedded into a revlog | |
133 | _maxinline = 131072 |
|
133 | _maxinline = 131072 | |
134 |
|
134 | |||
135 | # Flag processors for REVIDX_ELLIPSIS. |
|
135 | # Flag processors for REVIDX_ELLIPSIS. | |
136 | def ellipsisreadprocessor(rl, text): |
|
136 | def ellipsisreadprocessor(rl, text): | |
137 | return text, False |
|
137 | return text, False | |
138 |
|
138 | |||
139 |
|
139 | |||
140 | def ellipsiswriteprocessor(rl, text): |
|
140 | def ellipsiswriteprocessor(rl, text): | |
141 | return text, False |
|
141 | return text, False | |
142 |
|
142 | |||
143 |
|
143 | |||
144 | def ellipsisrawprocessor(rl, text): |
|
144 | def ellipsisrawprocessor(rl, text): | |
145 | return False |
|
145 | return False | |
146 |
|
146 | |||
147 |
|
147 | |||
148 | ellipsisprocessor = ( |
|
148 | ellipsisprocessor = ( | |
149 | ellipsisreadprocessor, |
|
149 | ellipsisreadprocessor, | |
150 | ellipsiswriteprocessor, |
|
150 | ellipsiswriteprocessor, | |
151 | ellipsisrawprocessor, |
|
151 | ellipsisrawprocessor, | |
152 | ) |
|
152 | ) | |
153 |
|
153 | |||
154 |
|
154 | |||
155 | def _verify_revision(rl, skipflags, state, node): |
|
155 | def _verify_revision(rl, skipflags, state, node): | |
156 | """Verify the integrity of the given revlog ``node`` while providing a hook |
|
156 | """Verify the integrity of the given revlog ``node`` while providing a hook | |
157 | point for extensions to influence the operation.""" |
|
157 | point for extensions to influence the operation.""" | |
158 | if skipflags: |
|
158 | if skipflags: | |
159 | state[b'skipread'].add(node) |
|
159 | state[b'skipread'].add(node) | |
160 | else: |
|
160 | else: | |
161 | # Side-effect: read content and verify hash. |
|
161 | # Side-effect: read content and verify hash. | |
162 | rl.revision(node) |
|
162 | rl.revision(node) | |
163 |
|
163 | |||
164 |
|
164 | |||
165 | # True if a fast implementation for persistent-nodemap is available |
|
165 | # True if a fast implementation for persistent-nodemap is available | |
166 | # |
|
166 | # | |
167 | # We also consider we have a "fast" implementation in "pure" python because |
|
167 | # We also consider we have a "fast" implementation in "pure" python because | |
168 | # people using pure don't really have performance consideration (and a |
|
168 | # people using pure don't really have performance consideration (and a | |
169 | # wheelbarrow of other slowness source) |
|
169 | # wheelbarrow of other slowness source) | |
170 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or util.safehasattr( |
|
170 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or util.safehasattr( | |
171 | parsers, 'BaseIndexObject' |
|
171 | parsers, 'BaseIndexObject' | |
172 | ) |
|
172 | ) | |
173 |
|
173 | |||
174 |
|
174 | |||
175 | @interfaceutil.implementer(repository.irevisiondelta) |
|
175 | @interfaceutil.implementer(repository.irevisiondelta) | |
176 | @attr.s(slots=True) |
|
176 | @attr.s(slots=True) | |
177 | class revlogrevisiondelta: |
|
177 | class revlogrevisiondelta: | |
178 | node = attr.ib() |
|
178 | node = attr.ib() | |
179 | p1node = attr.ib() |
|
179 | p1node = attr.ib() | |
180 | p2node = attr.ib() |
|
180 | p2node = attr.ib() | |
181 | basenode = attr.ib() |
|
181 | basenode = attr.ib() | |
182 | flags = attr.ib() |
|
182 | flags = attr.ib() | |
183 | baserevisionsize = attr.ib() |
|
183 | baserevisionsize = attr.ib() | |
184 | revision = attr.ib() |
|
184 | revision = attr.ib() | |
185 | delta = attr.ib() |
|
185 | delta = attr.ib() | |
186 | sidedata = attr.ib() |
|
186 | sidedata = attr.ib() | |
187 | protocol_flags = attr.ib() |
|
187 | protocol_flags = attr.ib() | |
188 | linknode = attr.ib(default=None) |
|
188 | linknode = attr.ib(default=None) | |
189 |
|
189 | |||
190 |
|
190 | |||
191 | @interfaceutil.implementer(repository.iverifyproblem) |
|
191 | @interfaceutil.implementer(repository.iverifyproblem) | |
192 | @attr.s(frozen=True) |
|
192 | @attr.s(frozen=True) | |
193 | class revlogproblem: |
|
193 | class revlogproblem: | |
194 | warning = attr.ib(default=None) |
|
194 | warning = attr.ib(default=None) | |
195 | error = attr.ib(default=None) |
|
195 | error = attr.ib(default=None) | |
196 | node = attr.ib(default=None) |
|
196 | node = attr.ib(default=None) | |
197 |
|
197 | |||
198 |
|
198 | |||
199 | def parse_index_v1(data, inline): |
|
199 | def parse_index_v1(data, inline): | |
200 | # call the C implementation to parse the index data |
|
200 | # call the C implementation to parse the index data | |
201 | index, cache = parsers.parse_index2(data, inline) |
|
201 | index, cache = parsers.parse_index2(data, inline) | |
202 | return index, cache |
|
202 | return index, cache | |
203 |
|
203 | |||
204 |
|
204 | |||
205 | def parse_index_v2(data, inline): |
|
205 | def parse_index_v2(data, inline): | |
206 | # call the C implementation to parse the index data |
|
206 | # call the C implementation to parse the index data | |
207 | index, cache = parsers.parse_index2(data, inline, format=REVLOGV2) |
|
207 | index, cache = parsers.parse_index2(data, inline, format=REVLOGV2) | |
208 | return index, cache |
|
208 | return index, cache | |
209 |
|
209 | |||
210 |
|
210 | |||
211 | def parse_index_cl_v2(data, inline): |
|
211 | def parse_index_cl_v2(data, inline): | |
212 | # call the C implementation to parse the index data |
|
212 | # call the C implementation to parse the index data | |
213 | index, cache = parsers.parse_index2(data, inline, format=CHANGELOGV2) |
|
213 | index, cache = parsers.parse_index2(data, inline, format=CHANGELOGV2) | |
214 | return index, cache |
|
214 | return index, cache | |
215 |
|
215 | |||
216 |
|
216 | |||
217 | if util.safehasattr(parsers, 'parse_index_devel_nodemap'): |
|
217 | if util.safehasattr(parsers, 'parse_index_devel_nodemap'): | |
218 |
|
218 | |||
219 | def parse_index_v1_nodemap(data, inline): |
|
219 | def parse_index_v1_nodemap(data, inline): | |
220 | index, cache = parsers.parse_index_devel_nodemap(data, inline) |
|
220 | index, cache = parsers.parse_index_devel_nodemap(data, inline) | |
221 | return index, cache |
|
221 | return index, cache | |
222 |
|
222 | |||
223 |
|
223 | |||
224 | else: |
|
224 | else: | |
225 | parse_index_v1_nodemap = None |
|
225 | parse_index_v1_nodemap = None | |
226 |
|
226 | |||
227 |
|
227 | |||
228 | def parse_index_v1_mixed(data, inline): |
|
228 | def parse_index_v1_mixed(data, inline): | |
229 | index, cache = parse_index_v1(data, inline) |
|
229 | index, cache = parse_index_v1(data, inline) | |
230 | return rustrevlog.MixedIndex(index), cache |
|
230 | return rustrevlog.MixedIndex(index), cache | |
231 |
|
231 | |||
232 |
|
232 | |||
233 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte |
|
233 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte | |
234 | # signed integer) |
|
234 | # signed integer) | |
235 | _maxentrysize = 0x7FFFFFFF |
|
235 | _maxentrysize = 0x7FFFFFFF | |
236 |
|
236 | |||
237 | FILE_TOO_SHORT_MSG = _( |
|
237 | FILE_TOO_SHORT_MSG = _( | |
238 | b'cannot read from revlog %s;' |
|
238 | b'cannot read from revlog %s;' | |
239 | b' expected %d bytes from offset %d, data size is %d' |
|
239 | b' expected %d bytes from offset %d, data size is %d' | |
240 | ) |
|
240 | ) | |
241 |
|
241 | |||
242 | hexdigits = b'0123456789abcdefABCDEF' |
|
242 | hexdigits = b'0123456789abcdefABCDEF' | |
243 |
|
243 | |||
244 |
|
244 | |||
245 | class revlog: |
|
245 | class revlog: | |
246 | """ |
|
246 | """ | |
247 | the underlying revision storage object |
|
247 | the underlying revision storage object | |
248 |
|
248 | |||
249 | A revlog consists of two parts, an index and the revision data. |
|
249 | A revlog consists of two parts, an index and the revision data. | |
250 |
|
250 | |||
251 | The index is a file with a fixed record size containing |
|
251 | The index is a file with a fixed record size containing | |
252 | information on each revision, including its nodeid (hash), the |
|
252 | information on each revision, including its nodeid (hash), the | |
253 | nodeids of its parents, the position and offset of its data within |
|
253 | nodeids of its parents, the position and offset of its data within | |
254 | the data file, and the revision it's based on. Finally, each entry |
|
254 | the data file, and the revision it's based on. Finally, each entry | |
255 | contains a linkrev entry that can serve as a pointer to external |
|
255 | contains a linkrev entry that can serve as a pointer to external | |
256 | data. |
|
256 | data. | |
257 |
|
257 | |||
258 | The revision data itself is a linear collection of data chunks. |
|
258 | The revision data itself is a linear collection of data chunks. | |
259 | Each chunk represents a revision and is usually represented as a |
|
259 | Each chunk represents a revision and is usually represented as a | |
260 | delta against the previous chunk. To bound lookup time, runs of |
|
260 | delta against the previous chunk. To bound lookup time, runs of | |
261 | deltas are limited to about 2 times the length of the original |
|
261 | deltas are limited to about 2 times the length of the original | |
262 | version data. This makes retrieval of a version proportional to |
|
262 | version data. This makes retrieval of a version proportional to | |
263 | its size, or O(1) relative to the number of revisions. |
|
263 | its size, or O(1) relative to the number of revisions. | |
264 |
|
264 | |||
265 | Both pieces of the revlog are written to in an append-only |
|
265 | Both pieces of the revlog are written to in an append-only | |
266 | fashion, which means we never need to rewrite a file to insert or |
|
266 | fashion, which means we never need to rewrite a file to insert or | |
267 | remove data, and can use some simple techniques to avoid the need |
|
267 | remove data, and can use some simple techniques to avoid the need | |
268 | for locking while reading. |
|
268 | for locking while reading. | |
269 |
|
269 | |||
270 | If checkambig, indexfile is opened with checkambig=True at |
|
270 | If checkambig, indexfile is opened with checkambig=True at | |
271 | writing, to avoid file stat ambiguity. |
|
271 | writing, to avoid file stat ambiguity. | |
272 |
|
272 | |||
273 | If mmaplargeindex is True, and an mmapindexthreshold is set, the |
|
273 | If mmaplargeindex is True, and an mmapindexthreshold is set, the | |
274 | index will be mmapped rather than read if it is larger than the |
|
274 | index will be mmapped rather than read if it is larger than the | |
275 | configured threshold. |
|
275 | configured threshold. | |
276 |
|
276 | |||
277 | If censorable is True, the revlog can have censored revisions. |
|
277 | If censorable is True, the revlog can have censored revisions. | |
278 |
|
278 | |||
279 | If `upperboundcomp` is not None, this is the expected maximal gain from |
|
279 | If `upperboundcomp` is not None, this is the expected maximal gain from | |
280 | compression for the data content. |
|
280 | compression for the data content. | |
281 |
|
281 | |||
282 | `concurrencychecker` is an optional function that receives 3 arguments: a |
|
282 | `concurrencychecker` is an optional function that receives 3 arguments: a | |
283 | file handle, a filename, and an expected position. It should check whether |
|
283 | file handle, a filename, and an expected position. It should check whether | |
284 | the current position in the file handle is valid, and log/warn/fail (by |
|
284 | the current position in the file handle is valid, and log/warn/fail (by | |
285 | raising). |
|
285 | raising). | |
286 |
|
286 | |||
287 | See mercurial/revlogutils/contants.py for details about the content of an |
|
287 | See mercurial/revlogutils/contants.py for details about the content of an | |
288 | index entry. |
|
288 | index entry. | |
289 | """ |
|
289 | """ | |
290 |
|
290 | |||
291 | _flagserrorclass = error.RevlogError |
|
291 | _flagserrorclass = error.RevlogError | |
292 |
|
292 | |||
293 | def __init__( |
|
293 | def __init__( | |
294 | self, |
|
294 | self, | |
295 | opener, |
|
295 | opener, | |
296 | target, |
|
296 | target, | |
297 | radix, |
|
297 | radix, | |
298 | postfix=None, # only exist for `tmpcensored` now |
|
298 | postfix=None, # only exist for `tmpcensored` now | |
299 | checkambig=False, |
|
299 | checkambig=False, | |
300 | mmaplargeindex=False, |
|
300 | mmaplargeindex=False, | |
301 | censorable=False, |
|
301 | censorable=False, | |
302 | upperboundcomp=None, |
|
302 | upperboundcomp=None, | |
303 | persistentnodemap=False, |
|
303 | persistentnodemap=False, | |
304 | concurrencychecker=None, |
|
304 | concurrencychecker=None, | |
305 | trypending=False, |
|
305 | trypending=False, | |
306 | try_split=False, |
|
306 | try_split=False, | |
307 | canonical_parent_order=True, |
|
307 | canonical_parent_order=True, | |
308 | ): |
|
308 | ): | |
309 | """ |
|
309 | """ | |
310 | create a revlog object |
|
310 | create a revlog object | |
311 |
|
311 | |||
312 | opener is a function that abstracts the file opening operation |
|
312 | opener is a function that abstracts the file opening operation | |
313 | and can be used to implement COW semantics or the like. |
|
313 | and can be used to implement COW semantics or the like. | |
314 |
|
314 | |||
315 | `target`: a (KIND, ID) tuple that identify the content stored in |
|
315 | `target`: a (KIND, ID) tuple that identify the content stored in | |
316 | this revlog. It help the rest of the code to understand what the revlog |
|
316 | this revlog. It help the rest of the code to understand what the revlog | |
317 | is about without having to resort to heuristic and index filename |
|
317 | is about without having to resort to heuristic and index filename | |
318 | analysis. Note: that this must be reliably be set by normal code, but |
|
318 | analysis. Note: that this must be reliably be set by normal code, but | |
319 | that test, debug, or performance measurement code might not set this to |
|
319 | that test, debug, or performance measurement code might not set this to | |
320 | accurate value. |
|
320 | accurate value. | |
321 | """ |
|
321 | """ | |
322 | self.upperboundcomp = upperboundcomp |
|
322 | self.upperboundcomp = upperboundcomp | |
323 |
|
323 | |||
324 | self.radix = radix |
|
324 | self.radix = radix | |
325 |
|
325 | |||
326 | self._docket_file = None |
|
326 | self._docket_file = None | |
327 | self._indexfile = None |
|
327 | self._indexfile = None | |
328 | self._datafile = None |
|
328 | self._datafile = None | |
329 | self._sidedatafile = None |
|
329 | self._sidedatafile = None | |
330 | self._nodemap_file = None |
|
330 | self._nodemap_file = None | |
331 | self.postfix = postfix |
|
331 | self.postfix = postfix | |
332 | self._trypending = trypending |
|
332 | self._trypending = trypending | |
333 | self._try_split = try_split |
|
333 | self._try_split = try_split | |
334 | self.opener = opener |
|
334 | self.opener = opener | |
335 | if persistentnodemap: |
|
335 | if persistentnodemap: | |
336 | self._nodemap_file = nodemaputil.get_nodemap_file(self) |
|
336 | self._nodemap_file = nodemaputil.get_nodemap_file(self) | |
337 |
|
337 | |||
338 | assert target[0] in ALL_KINDS |
|
338 | assert target[0] in ALL_KINDS | |
339 | assert len(target) == 2 |
|
339 | assert len(target) == 2 | |
340 | self.target = target |
|
340 | self.target = target | |
341 | # When True, indexfile is opened with checkambig=True at writing, to |
|
341 | # When True, indexfile is opened with checkambig=True at writing, to | |
342 | # avoid file stat ambiguity. |
|
342 | # avoid file stat ambiguity. | |
343 | self._checkambig = checkambig |
|
343 | self._checkambig = checkambig | |
344 | self._mmaplargeindex = mmaplargeindex |
|
344 | self._mmaplargeindex = mmaplargeindex | |
345 | self._censorable = censorable |
|
345 | self._censorable = censorable | |
346 | # 3-tuple of (node, rev, text) for a raw revision. |
|
346 | # 3-tuple of (node, rev, text) for a raw revision. | |
347 | self._revisioncache = None |
|
347 | self._revisioncache = None | |
348 | # Maps rev to chain base rev. |
|
348 | # Maps rev to chain base rev. | |
349 | self._chainbasecache = util.lrucachedict(100) |
|
349 | self._chainbasecache = util.lrucachedict(100) | |
350 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. |
|
350 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. | |
351 | self._chunkcache = (0, b'') |
|
351 | self._chunkcache = (0, b'') | |
352 | # How much data to read and cache into the raw revlog data cache. |
|
352 | # How much data to read and cache into the raw revlog data cache. | |
353 | self._chunkcachesize = 65536 |
|
353 | self._chunkcachesize = 65536 | |
354 | self._maxchainlen = None |
|
354 | self._maxchainlen = None | |
355 | self._deltabothparents = True |
|
355 | self._deltabothparents = True | |
356 | self._candidate_group_chunk_size = 0 |
|
356 | self._candidate_group_chunk_size = 0 | |
357 | self._debug_delta = False |
|
357 | self._debug_delta = False | |
358 | self.index = None |
|
358 | self.index = None | |
359 | self._docket = None |
|
359 | self._docket = None | |
360 | self._nodemap_docket = None |
|
360 | self._nodemap_docket = None | |
361 | # Mapping of partial identifiers to full nodes. |
|
361 | # Mapping of partial identifiers to full nodes. | |
362 | self._pcache = {} |
|
362 | self._pcache = {} | |
363 | # Mapping of revision integer to full node. |
|
363 | # Mapping of revision integer to full node. | |
364 | self._compengine = b'zlib' |
|
364 | self._compengine = b'zlib' | |
365 | self._compengineopts = {} |
|
365 | self._compengineopts = {} | |
366 | self._maxdeltachainspan = -1 |
|
366 | self._maxdeltachainspan = -1 | |
367 | self._withsparseread = False |
|
367 | self._withsparseread = False | |
368 | self._sparserevlog = False |
|
368 | self._sparserevlog = False | |
369 | self.hassidedata = False |
|
369 | self.hassidedata = False | |
370 | self._srdensitythreshold = 0.50 |
|
370 | self._srdensitythreshold = 0.50 | |
371 | self._srmingapsize = 262144 |
|
371 | self._srmingapsize = 262144 | |
372 |
|
372 | |||
373 | # other optionnals features |
|
373 | # other optionnals features | |
374 |
|
374 | |||
375 | # might remove rank configuration once the computation has no impact |
|
375 | # might remove rank configuration once the computation has no impact | |
376 | self._compute_rank = False |
|
376 | self._compute_rank = False | |
377 |
|
377 | |||
378 | # Make copy of flag processors so each revlog instance can support |
|
378 | # Make copy of flag processors so each revlog instance can support | |
379 | # custom flags. |
|
379 | # custom flags. | |
380 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
380 | self._flagprocessors = dict(flagutil.flagprocessors) | |
381 |
|
381 | |||
382 | # 3-tuple of file handles being used for active writing. |
|
382 | # 3-tuple of file handles being used for active writing. | |
383 | self._writinghandles = None |
|
383 | self._writinghandles = None | |
384 | # prevent nesting of addgroup |
|
384 | # prevent nesting of addgroup | |
385 | self._adding_group = None |
|
385 | self._adding_group = None | |
386 |
|
386 | |||
387 | self._loadindex() |
|
387 | self._loadindex() | |
388 |
|
388 | |||
389 | self._concurrencychecker = concurrencychecker |
|
389 | self._concurrencychecker = concurrencychecker | |
390 |
|
390 | |||
391 | # parent order is supposed to be semantically irrelevant, so we |
|
391 | # parent order is supposed to be semantically irrelevant, so we | |
392 | # normally resort parents to ensure that the first parent is non-null, |
|
392 | # normally resort parents to ensure that the first parent is non-null, | |
393 | # if there is a non-null parent at all. |
|
393 | # if there is a non-null parent at all. | |
394 | # filelog abuses the parent order as flag to mark some instances of |
|
394 | # filelog abuses the parent order as flag to mark some instances of | |
395 | # meta-encoded files, so allow it to disable this behavior. |
|
395 | # meta-encoded files, so allow it to disable this behavior. | |
396 | self.canonical_parent_order = canonical_parent_order |
|
396 | self.canonical_parent_order = canonical_parent_order | |
397 |
|
397 | |||
398 | def _init_opts(self): |
|
398 | def _init_opts(self): | |
399 | """process options (from above/config) to setup associated default revlog mode |
|
399 | """process options (from above/config) to setup associated default revlog mode | |
400 |
|
400 | |||
401 | These values might be affected when actually reading on disk information. |
|
401 | These values might be affected when actually reading on disk information. | |
402 |
|
402 | |||
403 | The relevant values are returned for use in _loadindex(). |
|
403 | The relevant values are returned for use in _loadindex(). | |
404 |
|
404 | |||
405 | * newversionflags: |
|
405 | * newversionflags: | |
406 | version header to use if we need to create a new revlog |
|
406 | version header to use if we need to create a new revlog | |
407 |
|
407 | |||
408 | * mmapindexthreshold: |
|
408 | * mmapindexthreshold: | |
409 | minimal index size for start to use mmap |
|
409 | minimal index size for start to use mmap | |
410 |
|
410 | |||
411 | * force_nodemap: |
|
411 | * force_nodemap: | |
412 | force the usage of a "development" version of the nodemap code |
|
412 | force the usage of a "development" version of the nodemap code | |
413 | """ |
|
413 | """ | |
414 | mmapindexthreshold = None |
|
414 | mmapindexthreshold = None | |
415 | opts = self.opener.options |
|
415 | opts = self.opener.options | |
416 |
|
416 | |||
417 | if b'changelogv2' in opts and self.revlog_kind == KIND_CHANGELOG: |
|
417 | if b'changelogv2' in opts and self.revlog_kind == KIND_CHANGELOG: | |
418 | new_header = CHANGELOGV2 |
|
418 | new_header = CHANGELOGV2 | |
419 | self._compute_rank = opts.get(b'changelogv2.compute-rank', True) |
|
419 | self._compute_rank = opts.get(b'changelogv2.compute-rank', True) | |
420 | elif b'revlogv2' in opts: |
|
420 | elif b'revlogv2' in opts: | |
421 | new_header = REVLOGV2 |
|
421 | new_header = REVLOGV2 | |
422 | elif b'revlogv1' in opts: |
|
422 | elif b'revlogv1' in opts: | |
423 | new_header = REVLOGV1 | FLAG_INLINE_DATA |
|
423 | new_header = REVLOGV1 | FLAG_INLINE_DATA | |
424 | if b'generaldelta' in opts: |
|
424 | if b'generaldelta' in opts: | |
425 | new_header |= FLAG_GENERALDELTA |
|
425 | new_header |= FLAG_GENERALDELTA | |
426 | elif b'revlogv0' in self.opener.options: |
|
426 | elif b'revlogv0' in self.opener.options: | |
427 | new_header = REVLOGV0 |
|
427 | new_header = REVLOGV0 | |
428 | else: |
|
428 | else: | |
429 | new_header = REVLOG_DEFAULT_VERSION |
|
429 | new_header = REVLOG_DEFAULT_VERSION | |
430 |
|
430 | |||
431 | if b'chunkcachesize' in opts: |
|
431 | if b'chunkcachesize' in opts: | |
432 | self._chunkcachesize = opts[b'chunkcachesize'] |
|
432 | self._chunkcachesize = opts[b'chunkcachesize'] | |
433 | if b'maxchainlen' in opts: |
|
433 | if b'maxchainlen' in opts: | |
434 | self._maxchainlen = opts[b'maxchainlen'] |
|
434 | self._maxchainlen = opts[b'maxchainlen'] | |
435 | if b'deltabothparents' in opts: |
|
435 | if b'deltabothparents' in opts: | |
436 | self._deltabothparents = opts[b'deltabothparents'] |
|
436 | self._deltabothparents = opts[b'deltabothparents'] | |
437 | dps_cgds = opts.get(b'delta-parent-search.candidate-group-chunk-size') |
|
437 | dps_cgds = opts.get(b'delta-parent-search.candidate-group-chunk-size') | |
438 | if dps_cgds: |
|
438 | if dps_cgds: | |
439 | self._candidate_group_chunk_size = dps_cgds |
|
439 | self._candidate_group_chunk_size = dps_cgds | |
440 | self._lazydelta = bool(opts.get(b'lazydelta', True)) |
|
440 | self._lazydelta = bool(opts.get(b'lazydelta', True)) | |
441 | self._lazydeltabase = False |
|
441 | self._lazydeltabase = False | |
442 | if self._lazydelta: |
|
442 | if self._lazydelta: | |
443 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) |
|
443 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) | |
444 | if b'debug-delta' in opts: |
|
444 | if b'debug-delta' in opts: | |
445 | self._debug_delta = opts[b'debug-delta'] |
|
445 | self._debug_delta = opts[b'debug-delta'] | |
446 | if b'compengine' in opts: |
|
446 | if b'compengine' in opts: | |
447 | self._compengine = opts[b'compengine'] |
|
447 | self._compengine = opts[b'compengine'] | |
448 | if b'zlib.level' in opts: |
|
448 | if b'zlib.level' in opts: | |
449 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] |
|
449 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] | |
450 | if b'zstd.level' in opts: |
|
450 | if b'zstd.level' in opts: | |
451 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] |
|
451 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] | |
452 | if b'maxdeltachainspan' in opts: |
|
452 | if b'maxdeltachainspan' in opts: | |
453 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] |
|
453 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] | |
454 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: |
|
454 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: | |
455 | mmapindexthreshold = opts[b'mmapindexthreshold'] |
|
455 | mmapindexthreshold = opts[b'mmapindexthreshold'] | |
456 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) |
|
456 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) | |
457 | withsparseread = bool(opts.get(b'with-sparse-read', False)) |
|
457 | withsparseread = bool(opts.get(b'with-sparse-read', False)) | |
458 | # sparse-revlog forces sparse-read |
|
458 | # sparse-revlog forces sparse-read | |
459 | self._withsparseread = self._sparserevlog or withsparseread |
|
459 | self._withsparseread = self._sparserevlog or withsparseread | |
460 | if b'sparse-read-density-threshold' in opts: |
|
460 | if b'sparse-read-density-threshold' in opts: | |
461 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] |
|
461 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] | |
462 | if b'sparse-read-min-gap-size' in opts: |
|
462 | if b'sparse-read-min-gap-size' in opts: | |
463 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] |
|
463 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] | |
464 | if opts.get(b'enableellipsis'): |
|
464 | if opts.get(b'enableellipsis'): | |
465 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor |
|
465 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor | |
466 |
|
466 | |||
467 | # revlog v0 doesn't have flag processors |
|
467 | # revlog v0 doesn't have flag processors | |
468 | for flag, processor in opts.get(b'flagprocessors', {}).items(): |
|
468 | for flag, processor in opts.get(b'flagprocessors', {}).items(): | |
469 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) |
|
469 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) | |
470 |
|
470 | |||
471 | if self._chunkcachesize <= 0: |
|
471 | if self._chunkcachesize <= 0: | |
472 | raise error.RevlogError( |
|
472 | raise error.RevlogError( | |
473 | _(b'revlog chunk cache size %r is not greater than 0') |
|
473 | _(b'revlog chunk cache size %r is not greater than 0') | |
474 | % self._chunkcachesize |
|
474 | % self._chunkcachesize | |
475 | ) |
|
475 | ) | |
476 | elif self._chunkcachesize & (self._chunkcachesize - 1): |
|
476 | elif self._chunkcachesize & (self._chunkcachesize - 1): | |
477 | raise error.RevlogError( |
|
477 | raise error.RevlogError( | |
478 | _(b'revlog chunk cache size %r is not a power of 2') |
|
478 | _(b'revlog chunk cache size %r is not a power of 2') | |
479 | % self._chunkcachesize |
|
479 | % self._chunkcachesize | |
480 | ) |
|
480 | ) | |
481 | force_nodemap = opts.get(b'devel-force-nodemap', False) |
|
481 | force_nodemap = opts.get(b'devel-force-nodemap', False) | |
482 | return new_header, mmapindexthreshold, force_nodemap |
|
482 | return new_header, mmapindexthreshold, force_nodemap | |
483 |
|
483 | |||
484 | def _get_data(self, filepath, mmap_threshold, size=None): |
|
484 | def _get_data(self, filepath, mmap_threshold, size=None): | |
485 | """return a file content with or without mmap |
|
485 | """return a file content with or without mmap | |
486 |
|
486 | |||
487 | If the file is missing return the empty string""" |
|
487 | If the file is missing return the empty string""" | |
488 | try: |
|
488 | try: | |
489 | with self.opener(filepath) as fp: |
|
489 | with self.opener(filepath) as fp: | |
490 | if mmap_threshold is not None: |
|
490 | if mmap_threshold is not None: | |
491 | file_size = self.opener.fstat(fp).st_size |
|
491 | file_size = self.opener.fstat(fp).st_size | |
492 | if file_size >= mmap_threshold: |
|
492 | if file_size >= mmap_threshold: | |
493 | if size is not None: |
|
493 | if size is not None: | |
494 | # avoid potentiel mmap crash |
|
494 | # avoid potentiel mmap crash | |
495 | size = min(file_size, size) |
|
495 | size = min(file_size, size) | |
496 | # TODO: should .close() to release resources without |
|
496 | # TODO: should .close() to release resources without | |
497 | # relying on Python GC |
|
497 | # relying on Python GC | |
498 | if size is None: |
|
498 | if size is None: | |
499 | return util.buffer(util.mmapread(fp)) |
|
499 | return util.buffer(util.mmapread(fp)) | |
500 | else: |
|
500 | else: | |
501 | return util.buffer(util.mmapread(fp, size)) |
|
501 | return util.buffer(util.mmapread(fp, size)) | |
502 | if size is None: |
|
502 | if size is None: | |
503 | return fp.read() |
|
503 | return fp.read() | |
504 | else: |
|
504 | else: | |
505 | return fp.read(size) |
|
505 | return fp.read(size) | |
506 | except FileNotFoundError: |
|
506 | except FileNotFoundError: | |
507 | return b'' |
|
507 | return b'' | |
508 |
|
508 | |||
|
509 | def get_streams(self, max_linkrev): | |||
|
510 | n = len(self) | |||
|
511 | index = self.index | |||
|
512 | while n > 0: | |||
|
513 | linkrev = index[n - 1][4] | |||
|
514 | if linkrev < max_linkrev: | |||
|
515 | break | |||
|
516 | # note: this loop will rarely go through multiple iterations, since | |||
|
517 | # it only traverses commits created during the current streaming | |||
|
518 | # pull operation. | |||
|
519 | # | |||
|
520 | # If this become a problem, using a binary search should cap the | |||
|
521 | # runtime of this. | |||
|
522 | n = n - 1 | |||
|
523 | if n == 0: | |||
|
524 | # no data to send | |||
|
525 | return [] | |||
|
526 | index_size = n * index.entry_size | |||
|
527 | data_size = self.end(n - 1) | |||
|
528 | ||||
|
529 | # XXX we might have been split (or stripped) since the object | |||
|
530 | # initialization, We need to close this race too, but having a way to | |||
|
531 | # pre-open the file we feed to the revlog and never closing them before | |||
|
532 | # we are done streaming. | |||
|
533 | ||||
|
534 | if self._inline: | |||
|
535 | ||||
|
536 | def get_stream(): | |||
|
537 | with self._indexfp() as fp: | |||
|
538 | yield None | |||
|
539 | size = index_size + data_size | |||
|
540 | if size <= 65536: | |||
|
541 | yield fp.read(size) | |||
|
542 | else: | |||
|
543 | yield from util.filechunkiter(fp, limit=size) | |||
|
544 | ||||
|
545 | inline_stream = get_stream() | |||
|
546 | next(inline_stream) | |||
|
547 | return [ | |||
|
548 | (self._indexfile, inline_stream, index_size + data_size), | |||
|
549 | ] | |||
|
550 | else: | |||
|
551 | ||||
|
552 | def get_index_stream(): | |||
|
553 | with self._indexfp() as fp: | |||
|
554 | yield None | |||
|
555 | if index_size <= 65536: | |||
|
556 | yield fp.read(index_size) | |||
|
557 | else: | |||
|
558 | yield from util.filechunkiter(fp, limit=index_size) | |||
|
559 | ||||
|
560 | def get_data_stream(): | |||
|
561 | with self._datafp() as fp: | |||
|
562 | yield None | |||
|
563 | if data_size <= 65536: | |||
|
564 | yield fp.read(data_size) | |||
|
565 | else: | |||
|
566 | yield from util.filechunkiter(fp, limit=data_size) | |||
|
567 | ||||
|
568 | index_stream = get_index_stream() | |||
|
569 | next(index_stream) | |||
|
570 | data_stream = get_data_stream() | |||
|
571 | next(data_stream) | |||
|
572 | return [ | |||
|
573 | (self._datafile, data_stream, data_size), | |||
|
574 | (self._indexfile, index_stream, index_size), | |||
|
575 | ] | |||
|
576 | ||||
509 | def _loadindex(self, docket=None): |
|
577 | def _loadindex(self, docket=None): | |
510 |
|
578 | |||
511 | new_header, mmapindexthreshold, force_nodemap = self._init_opts() |
|
579 | new_header, mmapindexthreshold, force_nodemap = self._init_opts() | |
512 |
|
580 | |||
513 | if self.postfix is not None: |
|
581 | if self.postfix is not None: | |
514 | entry_point = b'%s.i.%s' % (self.radix, self.postfix) |
|
582 | entry_point = b'%s.i.%s' % (self.radix, self.postfix) | |
515 | elif self._trypending and self.opener.exists(b'%s.i.a' % self.radix): |
|
583 | elif self._trypending and self.opener.exists(b'%s.i.a' % self.radix): | |
516 | entry_point = b'%s.i.a' % self.radix |
|
584 | entry_point = b'%s.i.a' % self.radix | |
517 | elif self._try_split and self.opener.exists(b'%s.i.s' % self.radix): |
|
585 | elif self._try_split and self.opener.exists(b'%s.i.s' % self.radix): | |
518 | entry_point = b'%s.i.s' % self.radix |
|
586 | entry_point = b'%s.i.s' % self.radix | |
519 | else: |
|
587 | else: | |
520 | entry_point = b'%s.i' % self.radix |
|
588 | entry_point = b'%s.i' % self.radix | |
521 |
|
589 | |||
522 | if docket is not None: |
|
590 | if docket is not None: | |
523 | self._docket = docket |
|
591 | self._docket = docket | |
524 | self._docket_file = entry_point |
|
592 | self._docket_file = entry_point | |
525 | else: |
|
593 | else: | |
526 | self._initempty = True |
|
594 | self._initempty = True | |
527 | entry_data = self._get_data(entry_point, mmapindexthreshold) |
|
595 | entry_data = self._get_data(entry_point, mmapindexthreshold) | |
528 | if len(entry_data) > 0: |
|
596 | if len(entry_data) > 0: | |
529 | header = INDEX_HEADER.unpack(entry_data[:4])[0] |
|
597 | header = INDEX_HEADER.unpack(entry_data[:4])[0] | |
530 | self._initempty = False |
|
598 | self._initempty = False | |
531 | else: |
|
599 | else: | |
532 | header = new_header |
|
600 | header = new_header | |
533 |
|
601 | |||
534 | self._format_flags = header & ~0xFFFF |
|
602 | self._format_flags = header & ~0xFFFF | |
535 | self._format_version = header & 0xFFFF |
|
603 | self._format_version = header & 0xFFFF | |
536 |
|
604 | |||
537 | supported_flags = SUPPORTED_FLAGS.get(self._format_version) |
|
605 | supported_flags = SUPPORTED_FLAGS.get(self._format_version) | |
538 | if supported_flags is None: |
|
606 | if supported_flags is None: | |
539 | msg = _(b'unknown version (%d) in revlog %s') |
|
607 | msg = _(b'unknown version (%d) in revlog %s') | |
540 | msg %= (self._format_version, self.display_id) |
|
608 | msg %= (self._format_version, self.display_id) | |
541 | raise error.RevlogError(msg) |
|
609 | raise error.RevlogError(msg) | |
542 | elif self._format_flags & ~supported_flags: |
|
610 | elif self._format_flags & ~supported_flags: | |
543 | msg = _(b'unknown flags (%#04x) in version %d revlog %s') |
|
611 | msg = _(b'unknown flags (%#04x) in version %d revlog %s') | |
544 | display_flag = self._format_flags >> 16 |
|
612 | display_flag = self._format_flags >> 16 | |
545 | msg %= (display_flag, self._format_version, self.display_id) |
|
613 | msg %= (display_flag, self._format_version, self.display_id) | |
546 | raise error.RevlogError(msg) |
|
614 | raise error.RevlogError(msg) | |
547 |
|
615 | |||
548 | features = FEATURES_BY_VERSION[self._format_version] |
|
616 | features = FEATURES_BY_VERSION[self._format_version] | |
549 | self._inline = features[b'inline'](self._format_flags) |
|
617 | self._inline = features[b'inline'](self._format_flags) | |
550 | self._generaldelta = features[b'generaldelta'](self._format_flags) |
|
618 | self._generaldelta = features[b'generaldelta'](self._format_flags) | |
551 | self.hassidedata = features[b'sidedata'] |
|
619 | self.hassidedata = features[b'sidedata'] | |
552 |
|
620 | |||
553 | if not features[b'docket']: |
|
621 | if not features[b'docket']: | |
554 | self._indexfile = entry_point |
|
622 | self._indexfile = entry_point | |
555 | index_data = entry_data |
|
623 | index_data = entry_data | |
556 | else: |
|
624 | else: | |
557 | self._docket_file = entry_point |
|
625 | self._docket_file = entry_point | |
558 | if self._initempty: |
|
626 | if self._initempty: | |
559 | self._docket = docketutil.default_docket(self, header) |
|
627 | self._docket = docketutil.default_docket(self, header) | |
560 | else: |
|
628 | else: | |
561 | self._docket = docketutil.parse_docket( |
|
629 | self._docket = docketutil.parse_docket( | |
562 | self, entry_data, use_pending=self._trypending |
|
630 | self, entry_data, use_pending=self._trypending | |
563 | ) |
|
631 | ) | |
564 |
|
632 | |||
565 | if self._docket is not None: |
|
633 | if self._docket is not None: | |
566 | self._indexfile = self._docket.index_filepath() |
|
634 | self._indexfile = self._docket.index_filepath() | |
567 | index_data = b'' |
|
635 | index_data = b'' | |
568 | index_size = self._docket.index_end |
|
636 | index_size = self._docket.index_end | |
569 | if index_size > 0: |
|
637 | if index_size > 0: | |
570 | index_data = self._get_data( |
|
638 | index_data = self._get_data( | |
571 | self._indexfile, mmapindexthreshold, size=index_size |
|
639 | self._indexfile, mmapindexthreshold, size=index_size | |
572 | ) |
|
640 | ) | |
573 | if len(index_data) < index_size: |
|
641 | if len(index_data) < index_size: | |
574 | msg = _(b'too few index data for %s: got %d, expected %d') |
|
642 | msg = _(b'too few index data for %s: got %d, expected %d') | |
575 | msg %= (self.display_id, len(index_data), index_size) |
|
643 | msg %= (self.display_id, len(index_data), index_size) | |
576 | raise error.RevlogError(msg) |
|
644 | raise error.RevlogError(msg) | |
577 |
|
645 | |||
578 | self._inline = False |
|
646 | self._inline = False | |
579 | # generaldelta implied by version 2 revlogs. |
|
647 | # generaldelta implied by version 2 revlogs. | |
580 | self._generaldelta = True |
|
648 | self._generaldelta = True | |
581 | # the logic for persistent nodemap will be dealt with within the |
|
649 | # the logic for persistent nodemap will be dealt with within the | |
582 | # main docket, so disable it for now. |
|
650 | # main docket, so disable it for now. | |
583 | self._nodemap_file = None |
|
651 | self._nodemap_file = None | |
584 |
|
652 | |||
585 | if self._docket is not None: |
|
653 | if self._docket is not None: | |
586 | self._datafile = self._docket.data_filepath() |
|
654 | self._datafile = self._docket.data_filepath() | |
587 | self._sidedatafile = self._docket.sidedata_filepath() |
|
655 | self._sidedatafile = self._docket.sidedata_filepath() | |
588 | elif self.postfix is None: |
|
656 | elif self.postfix is None: | |
589 | self._datafile = b'%s.d' % self.radix |
|
657 | self._datafile = b'%s.d' % self.radix | |
590 | else: |
|
658 | else: | |
591 | self._datafile = b'%s.d.%s' % (self.radix, self.postfix) |
|
659 | self._datafile = b'%s.d.%s' % (self.radix, self.postfix) | |
592 |
|
660 | |||
593 | self.nodeconstants = sha1nodeconstants |
|
661 | self.nodeconstants = sha1nodeconstants | |
594 | self.nullid = self.nodeconstants.nullid |
|
662 | self.nullid = self.nodeconstants.nullid | |
595 |
|
663 | |||
596 | # sparse-revlog can't be on without general-delta (issue6056) |
|
664 | # sparse-revlog can't be on without general-delta (issue6056) | |
597 | if not self._generaldelta: |
|
665 | if not self._generaldelta: | |
598 | self._sparserevlog = False |
|
666 | self._sparserevlog = False | |
599 |
|
667 | |||
600 | self._storedeltachains = True |
|
668 | self._storedeltachains = True | |
601 |
|
669 | |||
602 | devel_nodemap = ( |
|
670 | devel_nodemap = ( | |
603 | self._nodemap_file |
|
671 | self._nodemap_file | |
604 | and force_nodemap |
|
672 | and force_nodemap | |
605 | and parse_index_v1_nodemap is not None |
|
673 | and parse_index_v1_nodemap is not None | |
606 | ) |
|
674 | ) | |
607 |
|
675 | |||
608 | use_rust_index = False |
|
676 | use_rust_index = False | |
609 | if rustrevlog is not None: |
|
677 | if rustrevlog is not None: | |
610 | if self._nodemap_file is not None: |
|
678 | if self._nodemap_file is not None: | |
611 | use_rust_index = True |
|
679 | use_rust_index = True | |
612 | else: |
|
680 | else: | |
613 | use_rust_index = self.opener.options.get(b'rust.index') |
|
681 | use_rust_index = self.opener.options.get(b'rust.index') | |
614 |
|
682 | |||
615 | self._parse_index = parse_index_v1 |
|
683 | self._parse_index = parse_index_v1 | |
616 | if self._format_version == REVLOGV0: |
|
684 | if self._format_version == REVLOGV0: | |
617 | self._parse_index = revlogv0.parse_index_v0 |
|
685 | self._parse_index = revlogv0.parse_index_v0 | |
618 | elif self._format_version == REVLOGV2: |
|
686 | elif self._format_version == REVLOGV2: | |
619 | self._parse_index = parse_index_v2 |
|
687 | self._parse_index = parse_index_v2 | |
620 | elif self._format_version == CHANGELOGV2: |
|
688 | elif self._format_version == CHANGELOGV2: | |
621 | self._parse_index = parse_index_cl_v2 |
|
689 | self._parse_index = parse_index_cl_v2 | |
622 | elif devel_nodemap: |
|
690 | elif devel_nodemap: | |
623 | self._parse_index = parse_index_v1_nodemap |
|
691 | self._parse_index = parse_index_v1_nodemap | |
624 | elif use_rust_index: |
|
692 | elif use_rust_index: | |
625 | self._parse_index = parse_index_v1_mixed |
|
693 | self._parse_index = parse_index_v1_mixed | |
626 | try: |
|
694 | try: | |
627 | d = self._parse_index(index_data, self._inline) |
|
695 | d = self._parse_index(index_data, self._inline) | |
628 | index, chunkcache = d |
|
696 | index, chunkcache = d | |
629 | use_nodemap = ( |
|
697 | use_nodemap = ( | |
630 | not self._inline |
|
698 | not self._inline | |
631 | and self._nodemap_file is not None |
|
699 | and self._nodemap_file is not None | |
632 | and util.safehasattr(index, 'update_nodemap_data') |
|
700 | and util.safehasattr(index, 'update_nodemap_data') | |
633 | ) |
|
701 | ) | |
634 | if use_nodemap: |
|
702 | if use_nodemap: | |
635 | nodemap_data = nodemaputil.persisted_data(self) |
|
703 | nodemap_data = nodemaputil.persisted_data(self) | |
636 | if nodemap_data is not None: |
|
704 | if nodemap_data is not None: | |
637 | docket = nodemap_data[0] |
|
705 | docket = nodemap_data[0] | |
638 | if ( |
|
706 | if ( | |
639 | len(d[0]) > docket.tip_rev |
|
707 | len(d[0]) > docket.tip_rev | |
640 | and d[0][docket.tip_rev][7] == docket.tip_node |
|
708 | and d[0][docket.tip_rev][7] == docket.tip_node | |
641 | ): |
|
709 | ): | |
642 | # no changelog tampering |
|
710 | # no changelog tampering | |
643 | self._nodemap_docket = docket |
|
711 | self._nodemap_docket = docket | |
644 | index.update_nodemap_data(*nodemap_data) |
|
712 | index.update_nodemap_data(*nodemap_data) | |
645 | except (ValueError, IndexError): |
|
713 | except (ValueError, IndexError): | |
646 | raise error.RevlogError( |
|
714 | raise error.RevlogError( | |
647 | _(b"index %s is corrupted") % self.display_id |
|
715 | _(b"index %s is corrupted") % self.display_id | |
648 | ) |
|
716 | ) | |
649 | self.index = index |
|
717 | self.index = index | |
650 | self._segmentfile = randomaccessfile.randomaccessfile( |
|
718 | self._segmentfile = randomaccessfile.randomaccessfile( | |
651 | self.opener, |
|
719 | self.opener, | |
652 | (self._indexfile if self._inline else self._datafile), |
|
720 | (self._indexfile if self._inline else self._datafile), | |
653 | self._chunkcachesize, |
|
721 | self._chunkcachesize, | |
654 | chunkcache, |
|
722 | chunkcache, | |
655 | ) |
|
723 | ) | |
656 | self._segmentfile_sidedata = randomaccessfile.randomaccessfile( |
|
724 | self._segmentfile_sidedata = randomaccessfile.randomaccessfile( | |
657 | self.opener, |
|
725 | self.opener, | |
658 | self._sidedatafile, |
|
726 | self._sidedatafile, | |
659 | self._chunkcachesize, |
|
727 | self._chunkcachesize, | |
660 | ) |
|
728 | ) | |
661 | # revnum -> (chain-length, sum-delta-length) |
|
729 | # revnum -> (chain-length, sum-delta-length) | |
662 | self._chaininfocache = util.lrucachedict(500) |
|
730 | self._chaininfocache = util.lrucachedict(500) | |
663 | # revlog header -> revlog compressor |
|
731 | # revlog header -> revlog compressor | |
664 | self._decompressors = {} |
|
732 | self._decompressors = {} | |
665 |
|
733 | |||
666 | def get_revlog(self): |
|
734 | def get_revlog(self): | |
667 | """simple function to mirror API of other not-really-revlog API""" |
|
735 | """simple function to mirror API of other not-really-revlog API""" | |
668 | return self |
|
736 | return self | |
669 |
|
737 | |||
670 | @util.propertycache |
|
738 | @util.propertycache | |
671 | def revlog_kind(self): |
|
739 | def revlog_kind(self): | |
672 | return self.target[0] |
|
740 | return self.target[0] | |
673 |
|
741 | |||
674 | @util.propertycache |
|
742 | @util.propertycache | |
675 | def display_id(self): |
|
743 | def display_id(self): | |
676 | """The public facing "ID" of the revlog that we use in message""" |
|
744 | """The public facing "ID" of the revlog that we use in message""" | |
677 | if self.revlog_kind == KIND_FILELOG: |
|
745 | if self.revlog_kind == KIND_FILELOG: | |
678 | # Reference the file without the "data/" prefix, so it is familiar |
|
746 | # Reference the file without the "data/" prefix, so it is familiar | |
679 | # to the user. |
|
747 | # to the user. | |
680 | return self.target[1] |
|
748 | return self.target[1] | |
681 | else: |
|
749 | else: | |
682 | return self.radix |
|
750 | return self.radix | |
683 |
|
751 | |||
684 | def _get_decompressor(self, t): |
|
752 | def _get_decompressor(self, t): | |
685 | try: |
|
753 | try: | |
686 | compressor = self._decompressors[t] |
|
754 | compressor = self._decompressors[t] | |
687 | except KeyError: |
|
755 | except KeyError: | |
688 | try: |
|
756 | try: | |
689 | engine = util.compengines.forrevlogheader(t) |
|
757 | engine = util.compengines.forrevlogheader(t) | |
690 | compressor = engine.revlogcompressor(self._compengineopts) |
|
758 | compressor = engine.revlogcompressor(self._compengineopts) | |
691 | self._decompressors[t] = compressor |
|
759 | self._decompressors[t] = compressor | |
692 | except KeyError: |
|
760 | except KeyError: | |
693 | raise error.RevlogError( |
|
761 | raise error.RevlogError( | |
694 | _(b'unknown compression type %s') % binascii.hexlify(t) |
|
762 | _(b'unknown compression type %s') % binascii.hexlify(t) | |
695 | ) |
|
763 | ) | |
696 | return compressor |
|
764 | return compressor | |
697 |
|
765 | |||
698 | @util.propertycache |
|
766 | @util.propertycache | |
699 | def _compressor(self): |
|
767 | def _compressor(self): | |
700 | engine = util.compengines[self._compengine] |
|
768 | engine = util.compengines[self._compengine] | |
701 | return engine.revlogcompressor(self._compengineopts) |
|
769 | return engine.revlogcompressor(self._compengineopts) | |
702 |
|
770 | |||
703 | @util.propertycache |
|
771 | @util.propertycache | |
704 | def _decompressor(self): |
|
772 | def _decompressor(self): | |
705 | """the default decompressor""" |
|
773 | """the default decompressor""" | |
706 | if self._docket is None: |
|
774 | if self._docket is None: | |
707 | return None |
|
775 | return None | |
708 | t = self._docket.default_compression_header |
|
776 | t = self._docket.default_compression_header | |
709 | c = self._get_decompressor(t) |
|
777 | c = self._get_decompressor(t) | |
710 | return c.decompress |
|
778 | return c.decompress | |
711 |
|
779 | |||
712 | def _indexfp(self): |
|
780 | def _indexfp(self): | |
713 | """file object for the revlog's index file""" |
|
781 | """file object for the revlog's index file""" | |
714 | return self.opener(self._indexfile, mode=b"r") |
|
782 | return self.opener(self._indexfile, mode=b"r") | |
715 |
|
783 | |||
716 | def __index_write_fp(self): |
|
784 | def __index_write_fp(self): | |
717 | # You should not use this directly and use `_writing` instead |
|
785 | # You should not use this directly and use `_writing` instead | |
718 | try: |
|
786 | try: | |
719 | f = self.opener( |
|
787 | f = self.opener( | |
720 | self._indexfile, mode=b"r+", checkambig=self._checkambig |
|
788 | self._indexfile, mode=b"r+", checkambig=self._checkambig | |
721 | ) |
|
789 | ) | |
722 | if self._docket is None: |
|
790 | if self._docket is None: | |
723 | f.seek(0, os.SEEK_END) |
|
791 | f.seek(0, os.SEEK_END) | |
724 | else: |
|
792 | else: | |
725 | f.seek(self._docket.index_end, os.SEEK_SET) |
|
793 | f.seek(self._docket.index_end, os.SEEK_SET) | |
726 | return f |
|
794 | return f | |
727 | except FileNotFoundError: |
|
795 | except FileNotFoundError: | |
728 | return self.opener( |
|
796 | return self.opener( | |
729 | self._indexfile, mode=b"w+", checkambig=self._checkambig |
|
797 | self._indexfile, mode=b"w+", checkambig=self._checkambig | |
730 | ) |
|
798 | ) | |
731 |
|
799 | |||
732 | def __index_new_fp(self): |
|
800 | def __index_new_fp(self): | |
733 | # You should not use this unless you are upgrading from inline revlog |
|
801 | # You should not use this unless you are upgrading from inline revlog | |
734 | return self.opener( |
|
802 | return self.opener( | |
735 | self._indexfile, |
|
803 | self._indexfile, | |
736 | mode=b"w", |
|
804 | mode=b"w", | |
737 | checkambig=self._checkambig, |
|
805 | checkambig=self._checkambig, | |
738 | atomictemp=True, |
|
806 | atomictemp=True, | |
739 | ) |
|
807 | ) | |
740 |
|
808 | |||
741 | def _datafp(self, mode=b'r'): |
|
809 | def _datafp(self, mode=b'r'): | |
742 | """file object for the revlog's data file""" |
|
810 | """file object for the revlog's data file""" | |
743 | return self.opener(self._datafile, mode=mode) |
|
811 | return self.opener(self._datafile, mode=mode) | |
744 |
|
812 | |||
745 | @contextlib.contextmanager |
|
813 | @contextlib.contextmanager | |
746 | def _sidedatareadfp(self): |
|
814 | def _sidedatareadfp(self): | |
747 | """file object suitable to read sidedata""" |
|
815 | """file object suitable to read sidedata""" | |
748 | if self._writinghandles: |
|
816 | if self._writinghandles: | |
749 | yield self._writinghandles[2] |
|
817 | yield self._writinghandles[2] | |
750 | else: |
|
818 | else: | |
751 | with self.opener(self._sidedatafile) as fp: |
|
819 | with self.opener(self._sidedatafile) as fp: | |
752 | yield fp |
|
820 | yield fp | |
753 |
|
821 | |||
754 | def tiprev(self): |
|
822 | def tiprev(self): | |
755 | return len(self.index) - 1 |
|
823 | return len(self.index) - 1 | |
756 |
|
824 | |||
757 | def tip(self): |
|
825 | def tip(self): | |
758 | return self.node(self.tiprev()) |
|
826 | return self.node(self.tiprev()) | |
759 |
|
827 | |||
760 | def __contains__(self, rev): |
|
828 | def __contains__(self, rev): | |
761 | return 0 <= rev < len(self) |
|
829 | return 0 <= rev < len(self) | |
762 |
|
830 | |||
763 | def __len__(self): |
|
831 | def __len__(self): | |
764 | return len(self.index) |
|
832 | return len(self.index) | |
765 |
|
833 | |||
766 | def __iter__(self): |
|
834 | def __iter__(self): | |
767 | return iter(range(len(self))) |
|
835 | return iter(range(len(self))) | |
768 |
|
836 | |||
769 | def revs(self, start=0, stop=None): |
|
837 | def revs(self, start=0, stop=None): | |
770 | """iterate over all rev in this revlog (from start to stop)""" |
|
838 | """iterate over all rev in this revlog (from start to stop)""" | |
771 | return storageutil.iterrevs(len(self), start=start, stop=stop) |
|
839 | return storageutil.iterrevs(len(self), start=start, stop=stop) | |
772 |
|
840 | |||
773 | def hasnode(self, node): |
|
841 | def hasnode(self, node): | |
774 | try: |
|
842 | try: | |
775 | self.rev(node) |
|
843 | self.rev(node) | |
776 | return True |
|
844 | return True | |
777 | except KeyError: |
|
845 | except KeyError: | |
778 | return False |
|
846 | return False | |
779 |
|
847 | |||
780 | def candelta(self, baserev, rev): |
|
848 | def candelta(self, baserev, rev): | |
781 | """whether two revisions (baserev, rev) can be delta-ed or not""" |
|
849 | """whether two revisions (baserev, rev) can be delta-ed or not""" | |
782 | # Disable delta if either rev requires a content-changing flag |
|
850 | # Disable delta if either rev requires a content-changing flag | |
783 | # processor (ex. LFS). This is because such flag processor can alter |
|
851 | # processor (ex. LFS). This is because such flag processor can alter | |
784 | # the rawtext content that the delta will be based on, and two clients |
|
852 | # the rawtext content that the delta will be based on, and two clients | |
785 | # could have a same revlog node with different flags (i.e. different |
|
853 | # could have a same revlog node with different flags (i.e. different | |
786 | # rawtext contents) and the delta could be incompatible. |
|
854 | # rawtext contents) and the delta could be incompatible. | |
787 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
855 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( | |
788 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS |
|
856 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS | |
789 | ): |
|
857 | ): | |
790 | return False |
|
858 | return False | |
791 | return True |
|
859 | return True | |
792 |
|
860 | |||
793 | def update_caches(self, transaction): |
|
861 | def update_caches(self, transaction): | |
794 | if self._nodemap_file is not None: |
|
862 | if self._nodemap_file is not None: | |
795 | if transaction is None: |
|
863 | if transaction is None: | |
796 | nodemaputil.update_persistent_nodemap(self) |
|
864 | nodemaputil.update_persistent_nodemap(self) | |
797 | else: |
|
865 | else: | |
798 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
866 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
799 |
|
867 | |||
800 | def clearcaches(self): |
|
868 | def clearcaches(self): | |
801 | self._revisioncache = None |
|
869 | self._revisioncache = None | |
802 | self._chainbasecache.clear() |
|
870 | self._chainbasecache.clear() | |
803 | self._segmentfile.clear_cache() |
|
871 | self._segmentfile.clear_cache() | |
804 | self._segmentfile_sidedata.clear_cache() |
|
872 | self._segmentfile_sidedata.clear_cache() | |
805 | self._pcache = {} |
|
873 | self._pcache = {} | |
806 | self._nodemap_docket = None |
|
874 | self._nodemap_docket = None | |
807 | self.index.clearcaches() |
|
875 | self.index.clearcaches() | |
808 | # The python code is the one responsible for validating the docket, we |
|
876 | # The python code is the one responsible for validating the docket, we | |
809 | # end up having to refresh it here. |
|
877 | # end up having to refresh it here. | |
810 | use_nodemap = ( |
|
878 | use_nodemap = ( | |
811 | not self._inline |
|
879 | not self._inline | |
812 | and self._nodemap_file is not None |
|
880 | and self._nodemap_file is not None | |
813 | and util.safehasattr(self.index, 'update_nodemap_data') |
|
881 | and util.safehasattr(self.index, 'update_nodemap_data') | |
814 | ) |
|
882 | ) | |
815 | if use_nodemap: |
|
883 | if use_nodemap: | |
816 | nodemap_data = nodemaputil.persisted_data(self) |
|
884 | nodemap_data = nodemaputil.persisted_data(self) | |
817 | if nodemap_data is not None: |
|
885 | if nodemap_data is not None: | |
818 | self._nodemap_docket = nodemap_data[0] |
|
886 | self._nodemap_docket = nodemap_data[0] | |
819 | self.index.update_nodemap_data(*nodemap_data) |
|
887 | self.index.update_nodemap_data(*nodemap_data) | |
820 |
|
888 | |||
821 | def rev(self, node): |
|
889 | def rev(self, node): | |
822 | try: |
|
890 | try: | |
823 | return self.index.rev(node) |
|
891 | return self.index.rev(node) | |
824 | except TypeError: |
|
892 | except TypeError: | |
825 | raise |
|
893 | raise | |
826 | except error.RevlogError: |
|
894 | except error.RevlogError: | |
827 | # parsers.c radix tree lookup failed |
|
895 | # parsers.c radix tree lookup failed | |
828 | if ( |
|
896 | if ( | |
829 | node == self.nodeconstants.wdirid |
|
897 | node == self.nodeconstants.wdirid | |
830 | or node in self.nodeconstants.wdirfilenodeids |
|
898 | or node in self.nodeconstants.wdirfilenodeids | |
831 | ): |
|
899 | ): | |
832 | raise error.WdirUnsupported |
|
900 | raise error.WdirUnsupported | |
833 | raise error.LookupError(node, self.display_id, _(b'no node')) |
|
901 | raise error.LookupError(node, self.display_id, _(b'no node')) | |
834 |
|
902 | |||
835 | # Accessors for index entries. |
|
903 | # Accessors for index entries. | |
836 |
|
904 | |||
837 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes |
|
905 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes | |
838 | # are flags. |
|
906 | # are flags. | |
839 | def start(self, rev): |
|
907 | def start(self, rev): | |
840 | return int(self.index[rev][0] >> 16) |
|
908 | return int(self.index[rev][0] >> 16) | |
841 |
|
909 | |||
842 | def sidedata_cut_off(self, rev): |
|
910 | def sidedata_cut_off(self, rev): | |
843 | sd_cut_off = self.index[rev][8] |
|
911 | sd_cut_off = self.index[rev][8] | |
844 | if sd_cut_off != 0: |
|
912 | if sd_cut_off != 0: | |
845 | return sd_cut_off |
|
913 | return sd_cut_off | |
846 | # This is some annoying dance, because entries without sidedata |
|
914 | # This is some annoying dance, because entries without sidedata | |
847 | # currently use 0 as their ofsset. (instead of previous-offset + |
|
915 | # currently use 0 as their ofsset. (instead of previous-offset + | |
848 | # previous-size) |
|
916 | # previous-size) | |
849 | # |
|
917 | # | |
850 | # We should reconsider this sidedata β 0 sidata_offset policy. |
|
918 | # We should reconsider this sidedata β 0 sidata_offset policy. | |
851 | # In the meantime, we need this. |
|
919 | # In the meantime, we need this. | |
852 | while 0 <= rev: |
|
920 | while 0 <= rev: | |
853 | e = self.index[rev] |
|
921 | e = self.index[rev] | |
854 | if e[9] != 0: |
|
922 | if e[9] != 0: | |
855 | return e[8] + e[9] |
|
923 | return e[8] + e[9] | |
856 | rev -= 1 |
|
924 | rev -= 1 | |
857 | return 0 |
|
925 | return 0 | |
858 |
|
926 | |||
859 | def flags(self, rev): |
|
927 | def flags(self, rev): | |
860 | return self.index[rev][0] & 0xFFFF |
|
928 | return self.index[rev][0] & 0xFFFF | |
861 |
|
929 | |||
862 | def length(self, rev): |
|
930 | def length(self, rev): | |
863 | return self.index[rev][1] |
|
931 | return self.index[rev][1] | |
864 |
|
932 | |||
865 | def sidedata_length(self, rev): |
|
933 | def sidedata_length(self, rev): | |
866 | if not self.hassidedata: |
|
934 | if not self.hassidedata: | |
867 | return 0 |
|
935 | return 0 | |
868 | return self.index[rev][9] |
|
936 | return self.index[rev][9] | |
869 |
|
937 | |||
870 | def rawsize(self, rev): |
|
938 | def rawsize(self, rev): | |
871 | """return the length of the uncompressed text for a given revision""" |
|
939 | """return the length of the uncompressed text for a given revision""" | |
872 | l = self.index[rev][2] |
|
940 | l = self.index[rev][2] | |
873 | if l >= 0: |
|
941 | if l >= 0: | |
874 | return l |
|
942 | return l | |
875 |
|
943 | |||
876 | t = self.rawdata(rev) |
|
944 | t = self.rawdata(rev) | |
877 | return len(t) |
|
945 | return len(t) | |
878 |
|
946 | |||
879 | def size(self, rev): |
|
947 | def size(self, rev): | |
880 | """length of non-raw text (processed by a "read" flag processor)""" |
|
948 | """length of non-raw text (processed by a "read" flag processor)""" | |
881 | # fast path: if no "read" flag processor could change the content, |
|
949 | # fast path: if no "read" flag processor could change the content, | |
882 | # size is rawsize. note: ELLIPSIS is known to not change the content. |
|
950 | # size is rawsize. note: ELLIPSIS is known to not change the content. | |
883 | flags = self.flags(rev) |
|
951 | flags = self.flags(rev) | |
884 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: |
|
952 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: | |
885 | return self.rawsize(rev) |
|
953 | return self.rawsize(rev) | |
886 |
|
954 | |||
887 | return len(self.revision(rev)) |
|
955 | return len(self.revision(rev)) | |
888 |
|
956 | |||
889 | def fast_rank(self, rev): |
|
957 | def fast_rank(self, rev): | |
890 | """Return the rank of a revision if already known, or None otherwise. |
|
958 | """Return the rank of a revision if already known, or None otherwise. | |
891 |
|
959 | |||
892 | The rank of a revision is the size of the sub-graph it defines as a |
|
960 | The rank of a revision is the size of the sub-graph it defines as a | |
893 | head. Equivalently, the rank of a revision `r` is the size of the set |
|
961 | head. Equivalently, the rank of a revision `r` is the size of the set | |
894 | `ancestors(r)`, `r` included. |
|
962 | `ancestors(r)`, `r` included. | |
895 |
|
963 | |||
896 | This method returns the rank retrieved from the revlog in constant |
|
964 | This method returns the rank retrieved from the revlog in constant | |
897 | time. It makes no attempt at computing unknown values for versions of |
|
965 | time. It makes no attempt at computing unknown values for versions of | |
898 | the revlog which do not persist the rank. |
|
966 | the revlog which do not persist the rank. | |
899 | """ |
|
967 | """ | |
900 | rank = self.index[rev][ENTRY_RANK] |
|
968 | rank = self.index[rev][ENTRY_RANK] | |
901 | if self._format_version != CHANGELOGV2 or rank == RANK_UNKNOWN: |
|
969 | if self._format_version != CHANGELOGV2 or rank == RANK_UNKNOWN: | |
902 | return None |
|
970 | return None | |
903 | if rev == nullrev: |
|
971 | if rev == nullrev: | |
904 | return 0 # convention |
|
972 | return 0 # convention | |
905 | return rank |
|
973 | return rank | |
906 |
|
974 | |||
907 | def chainbase(self, rev): |
|
975 | def chainbase(self, rev): | |
908 | base = self._chainbasecache.get(rev) |
|
976 | base = self._chainbasecache.get(rev) | |
909 | if base is not None: |
|
977 | if base is not None: | |
910 | return base |
|
978 | return base | |
911 |
|
979 | |||
912 | index = self.index |
|
980 | index = self.index | |
913 | iterrev = rev |
|
981 | iterrev = rev | |
914 | base = index[iterrev][3] |
|
982 | base = index[iterrev][3] | |
915 | while base != iterrev: |
|
983 | while base != iterrev: | |
916 | iterrev = base |
|
984 | iterrev = base | |
917 | base = index[iterrev][3] |
|
985 | base = index[iterrev][3] | |
918 |
|
986 | |||
919 | self._chainbasecache[rev] = base |
|
987 | self._chainbasecache[rev] = base | |
920 | return base |
|
988 | return base | |
921 |
|
989 | |||
922 | def linkrev(self, rev): |
|
990 | def linkrev(self, rev): | |
923 | return self.index[rev][4] |
|
991 | return self.index[rev][4] | |
924 |
|
992 | |||
925 | def parentrevs(self, rev): |
|
993 | def parentrevs(self, rev): | |
926 | try: |
|
994 | try: | |
927 | entry = self.index[rev] |
|
995 | entry = self.index[rev] | |
928 | except IndexError: |
|
996 | except IndexError: | |
929 | if rev == wdirrev: |
|
997 | if rev == wdirrev: | |
930 | raise error.WdirUnsupported |
|
998 | raise error.WdirUnsupported | |
931 | raise |
|
999 | raise | |
932 |
|
1000 | |||
933 | if self.canonical_parent_order and entry[5] == nullrev: |
|
1001 | if self.canonical_parent_order and entry[5] == nullrev: | |
934 | return entry[6], entry[5] |
|
1002 | return entry[6], entry[5] | |
935 | else: |
|
1003 | else: | |
936 | return entry[5], entry[6] |
|
1004 | return entry[5], entry[6] | |
937 |
|
1005 | |||
938 | # fast parentrevs(rev) where rev isn't filtered |
|
1006 | # fast parentrevs(rev) where rev isn't filtered | |
939 | _uncheckedparentrevs = parentrevs |
|
1007 | _uncheckedparentrevs = parentrevs | |
940 |
|
1008 | |||
941 | def node(self, rev): |
|
1009 | def node(self, rev): | |
942 | try: |
|
1010 | try: | |
943 | return self.index[rev][7] |
|
1011 | return self.index[rev][7] | |
944 | except IndexError: |
|
1012 | except IndexError: | |
945 | if rev == wdirrev: |
|
1013 | if rev == wdirrev: | |
946 | raise error.WdirUnsupported |
|
1014 | raise error.WdirUnsupported | |
947 | raise |
|
1015 | raise | |
948 |
|
1016 | |||
949 | # Derived from index values. |
|
1017 | # Derived from index values. | |
950 |
|
1018 | |||
951 | def end(self, rev): |
|
1019 | def end(self, rev): | |
952 | return self.start(rev) + self.length(rev) |
|
1020 | return self.start(rev) + self.length(rev) | |
953 |
|
1021 | |||
954 | def parents(self, node): |
|
1022 | def parents(self, node): | |
955 | i = self.index |
|
1023 | i = self.index | |
956 | d = i[self.rev(node)] |
|
1024 | d = i[self.rev(node)] | |
957 | # inline node() to avoid function call overhead |
|
1025 | # inline node() to avoid function call overhead | |
958 | if self.canonical_parent_order and d[5] == self.nullid: |
|
1026 | if self.canonical_parent_order and d[5] == self.nullid: | |
959 | return i[d[6]][7], i[d[5]][7] |
|
1027 | return i[d[6]][7], i[d[5]][7] | |
960 | else: |
|
1028 | else: | |
961 | return i[d[5]][7], i[d[6]][7] |
|
1029 | return i[d[5]][7], i[d[6]][7] | |
962 |
|
1030 | |||
963 | def chainlen(self, rev): |
|
1031 | def chainlen(self, rev): | |
964 | return self._chaininfo(rev)[0] |
|
1032 | return self._chaininfo(rev)[0] | |
965 |
|
1033 | |||
966 | def _chaininfo(self, rev): |
|
1034 | def _chaininfo(self, rev): | |
967 | chaininfocache = self._chaininfocache |
|
1035 | chaininfocache = self._chaininfocache | |
968 | if rev in chaininfocache: |
|
1036 | if rev in chaininfocache: | |
969 | return chaininfocache[rev] |
|
1037 | return chaininfocache[rev] | |
970 | index = self.index |
|
1038 | index = self.index | |
971 | generaldelta = self._generaldelta |
|
1039 | generaldelta = self._generaldelta | |
972 | iterrev = rev |
|
1040 | iterrev = rev | |
973 | e = index[iterrev] |
|
1041 | e = index[iterrev] | |
974 | clen = 0 |
|
1042 | clen = 0 | |
975 | compresseddeltalen = 0 |
|
1043 | compresseddeltalen = 0 | |
976 | while iterrev != e[3]: |
|
1044 | while iterrev != e[3]: | |
977 | clen += 1 |
|
1045 | clen += 1 | |
978 | compresseddeltalen += e[1] |
|
1046 | compresseddeltalen += e[1] | |
979 | if generaldelta: |
|
1047 | if generaldelta: | |
980 | iterrev = e[3] |
|
1048 | iterrev = e[3] | |
981 | else: |
|
1049 | else: | |
982 | iterrev -= 1 |
|
1050 | iterrev -= 1 | |
983 | if iterrev in chaininfocache: |
|
1051 | if iterrev in chaininfocache: | |
984 | t = chaininfocache[iterrev] |
|
1052 | t = chaininfocache[iterrev] | |
985 | clen += t[0] |
|
1053 | clen += t[0] | |
986 | compresseddeltalen += t[1] |
|
1054 | compresseddeltalen += t[1] | |
987 | break |
|
1055 | break | |
988 | e = index[iterrev] |
|
1056 | e = index[iterrev] | |
989 | else: |
|
1057 | else: | |
990 | # Add text length of base since decompressing that also takes |
|
1058 | # Add text length of base since decompressing that also takes | |
991 | # work. For cache hits the length is already included. |
|
1059 | # work. For cache hits the length is already included. | |
992 | compresseddeltalen += e[1] |
|
1060 | compresseddeltalen += e[1] | |
993 | r = (clen, compresseddeltalen) |
|
1061 | r = (clen, compresseddeltalen) | |
994 | chaininfocache[rev] = r |
|
1062 | chaininfocache[rev] = r | |
995 | return r |
|
1063 | return r | |
996 |
|
1064 | |||
997 | def _deltachain(self, rev, stoprev=None): |
|
1065 | def _deltachain(self, rev, stoprev=None): | |
998 | """Obtain the delta chain for a revision. |
|
1066 | """Obtain the delta chain for a revision. | |
999 |
|
1067 | |||
1000 | ``stoprev`` specifies a revision to stop at. If not specified, we |
|
1068 | ``stoprev`` specifies a revision to stop at. If not specified, we | |
1001 | stop at the base of the chain. |
|
1069 | stop at the base of the chain. | |
1002 |
|
1070 | |||
1003 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of |
|
1071 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of | |
1004 | revs in ascending order and ``stopped`` is a bool indicating whether |
|
1072 | revs in ascending order and ``stopped`` is a bool indicating whether | |
1005 | ``stoprev`` was hit. |
|
1073 | ``stoprev`` was hit. | |
1006 | """ |
|
1074 | """ | |
1007 | # Try C implementation. |
|
1075 | # Try C implementation. | |
1008 | try: |
|
1076 | try: | |
1009 | return self.index.deltachain(rev, stoprev, self._generaldelta) |
|
1077 | return self.index.deltachain(rev, stoprev, self._generaldelta) | |
1010 | except AttributeError: |
|
1078 | except AttributeError: | |
1011 | pass |
|
1079 | pass | |
1012 |
|
1080 | |||
1013 | chain = [] |
|
1081 | chain = [] | |
1014 |
|
1082 | |||
1015 | # Alias to prevent attribute lookup in tight loop. |
|
1083 | # Alias to prevent attribute lookup in tight loop. | |
1016 | index = self.index |
|
1084 | index = self.index | |
1017 | generaldelta = self._generaldelta |
|
1085 | generaldelta = self._generaldelta | |
1018 |
|
1086 | |||
1019 | iterrev = rev |
|
1087 | iterrev = rev | |
1020 | e = index[iterrev] |
|
1088 | e = index[iterrev] | |
1021 | while iterrev != e[3] and iterrev != stoprev: |
|
1089 | while iterrev != e[3] and iterrev != stoprev: | |
1022 | chain.append(iterrev) |
|
1090 | chain.append(iterrev) | |
1023 | if generaldelta: |
|
1091 | if generaldelta: | |
1024 | iterrev = e[3] |
|
1092 | iterrev = e[3] | |
1025 | else: |
|
1093 | else: | |
1026 | iterrev -= 1 |
|
1094 | iterrev -= 1 | |
1027 | e = index[iterrev] |
|
1095 | e = index[iterrev] | |
1028 |
|
1096 | |||
1029 | if iterrev == stoprev: |
|
1097 | if iterrev == stoprev: | |
1030 | stopped = True |
|
1098 | stopped = True | |
1031 | else: |
|
1099 | else: | |
1032 | chain.append(iterrev) |
|
1100 | chain.append(iterrev) | |
1033 | stopped = False |
|
1101 | stopped = False | |
1034 |
|
1102 | |||
1035 | chain.reverse() |
|
1103 | chain.reverse() | |
1036 | return chain, stopped |
|
1104 | return chain, stopped | |
1037 |
|
1105 | |||
1038 | def ancestors(self, revs, stoprev=0, inclusive=False): |
|
1106 | def ancestors(self, revs, stoprev=0, inclusive=False): | |
1039 | """Generate the ancestors of 'revs' in reverse revision order. |
|
1107 | """Generate the ancestors of 'revs' in reverse revision order. | |
1040 | Does not generate revs lower than stoprev. |
|
1108 | Does not generate revs lower than stoprev. | |
1041 |
|
1109 | |||
1042 | See the documentation for ancestor.lazyancestors for more details.""" |
|
1110 | See the documentation for ancestor.lazyancestors for more details.""" | |
1043 |
|
1111 | |||
1044 | # first, make sure start revisions aren't filtered |
|
1112 | # first, make sure start revisions aren't filtered | |
1045 | revs = list(revs) |
|
1113 | revs = list(revs) | |
1046 | checkrev = self.node |
|
1114 | checkrev = self.node | |
1047 | for r in revs: |
|
1115 | for r in revs: | |
1048 | checkrev(r) |
|
1116 | checkrev(r) | |
1049 | # and we're sure ancestors aren't filtered as well |
|
1117 | # and we're sure ancestors aren't filtered as well | |
1050 |
|
1118 | |||
1051 | if rustancestor is not None and self.index.rust_ext_compat: |
|
1119 | if rustancestor is not None and self.index.rust_ext_compat: | |
1052 | lazyancestors = rustancestor.LazyAncestors |
|
1120 | lazyancestors = rustancestor.LazyAncestors | |
1053 | arg = self.index |
|
1121 | arg = self.index | |
1054 | else: |
|
1122 | else: | |
1055 | lazyancestors = ancestor.lazyancestors |
|
1123 | lazyancestors = ancestor.lazyancestors | |
1056 | arg = self._uncheckedparentrevs |
|
1124 | arg = self._uncheckedparentrevs | |
1057 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) |
|
1125 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) | |
1058 |
|
1126 | |||
1059 | def descendants(self, revs): |
|
1127 | def descendants(self, revs): | |
1060 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
1128 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) | |
1061 |
|
1129 | |||
1062 | def findcommonmissing(self, common=None, heads=None): |
|
1130 | def findcommonmissing(self, common=None, heads=None): | |
1063 | """Return a tuple of the ancestors of common and the ancestors of heads |
|
1131 | """Return a tuple of the ancestors of common and the ancestors of heads | |
1064 | that are not ancestors of common. In revset terminology, we return the |
|
1132 | that are not ancestors of common. In revset terminology, we return the | |
1065 | tuple: |
|
1133 | tuple: | |
1066 |
|
1134 | |||
1067 | ::common, (::heads) - (::common) |
|
1135 | ::common, (::heads) - (::common) | |
1068 |
|
1136 | |||
1069 | The list is sorted by revision number, meaning it is |
|
1137 | The list is sorted by revision number, meaning it is | |
1070 | topologically sorted. |
|
1138 | topologically sorted. | |
1071 |
|
1139 | |||
1072 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1140 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1073 | not supplied, uses all of the revlog's heads. If common is not |
|
1141 | not supplied, uses all of the revlog's heads. If common is not | |
1074 | supplied, uses nullid.""" |
|
1142 | supplied, uses nullid.""" | |
1075 | if common is None: |
|
1143 | if common is None: | |
1076 | common = [self.nullid] |
|
1144 | common = [self.nullid] | |
1077 | if heads is None: |
|
1145 | if heads is None: | |
1078 | heads = self.heads() |
|
1146 | heads = self.heads() | |
1079 |
|
1147 | |||
1080 | common = [self.rev(n) for n in common] |
|
1148 | common = [self.rev(n) for n in common] | |
1081 | heads = [self.rev(n) for n in heads] |
|
1149 | heads = [self.rev(n) for n in heads] | |
1082 |
|
1150 | |||
1083 | # we want the ancestors, but inclusive |
|
1151 | # we want the ancestors, but inclusive | |
1084 | class lazyset: |
|
1152 | class lazyset: | |
1085 | def __init__(self, lazyvalues): |
|
1153 | def __init__(self, lazyvalues): | |
1086 | self.addedvalues = set() |
|
1154 | self.addedvalues = set() | |
1087 | self.lazyvalues = lazyvalues |
|
1155 | self.lazyvalues = lazyvalues | |
1088 |
|
1156 | |||
1089 | def __contains__(self, value): |
|
1157 | def __contains__(self, value): | |
1090 | return value in self.addedvalues or value in self.lazyvalues |
|
1158 | return value in self.addedvalues or value in self.lazyvalues | |
1091 |
|
1159 | |||
1092 | def __iter__(self): |
|
1160 | def __iter__(self): | |
1093 | added = self.addedvalues |
|
1161 | added = self.addedvalues | |
1094 | for r in added: |
|
1162 | for r in added: | |
1095 | yield r |
|
1163 | yield r | |
1096 | for r in self.lazyvalues: |
|
1164 | for r in self.lazyvalues: | |
1097 | if not r in added: |
|
1165 | if not r in added: | |
1098 | yield r |
|
1166 | yield r | |
1099 |
|
1167 | |||
1100 | def add(self, value): |
|
1168 | def add(self, value): | |
1101 | self.addedvalues.add(value) |
|
1169 | self.addedvalues.add(value) | |
1102 |
|
1170 | |||
1103 | def update(self, values): |
|
1171 | def update(self, values): | |
1104 | self.addedvalues.update(values) |
|
1172 | self.addedvalues.update(values) | |
1105 |
|
1173 | |||
1106 | has = lazyset(self.ancestors(common)) |
|
1174 | has = lazyset(self.ancestors(common)) | |
1107 | has.add(nullrev) |
|
1175 | has.add(nullrev) | |
1108 | has.update(common) |
|
1176 | has.update(common) | |
1109 |
|
1177 | |||
1110 | # take all ancestors from heads that aren't in has |
|
1178 | # take all ancestors from heads that aren't in has | |
1111 | missing = set() |
|
1179 | missing = set() | |
1112 | visit = collections.deque(r for r in heads if r not in has) |
|
1180 | visit = collections.deque(r for r in heads if r not in has) | |
1113 | while visit: |
|
1181 | while visit: | |
1114 | r = visit.popleft() |
|
1182 | r = visit.popleft() | |
1115 | if r in missing: |
|
1183 | if r in missing: | |
1116 | continue |
|
1184 | continue | |
1117 | else: |
|
1185 | else: | |
1118 | missing.add(r) |
|
1186 | missing.add(r) | |
1119 | for p in self.parentrevs(r): |
|
1187 | for p in self.parentrevs(r): | |
1120 | if p not in has: |
|
1188 | if p not in has: | |
1121 | visit.append(p) |
|
1189 | visit.append(p) | |
1122 | missing = list(missing) |
|
1190 | missing = list(missing) | |
1123 | missing.sort() |
|
1191 | missing.sort() | |
1124 | return has, [self.node(miss) for miss in missing] |
|
1192 | return has, [self.node(miss) for miss in missing] | |
1125 |
|
1193 | |||
1126 | def incrementalmissingrevs(self, common=None): |
|
1194 | def incrementalmissingrevs(self, common=None): | |
1127 | """Return an object that can be used to incrementally compute the |
|
1195 | """Return an object that can be used to incrementally compute the | |
1128 | revision numbers of the ancestors of arbitrary sets that are not |
|
1196 | revision numbers of the ancestors of arbitrary sets that are not | |
1129 | ancestors of common. This is an ancestor.incrementalmissingancestors |
|
1197 | ancestors of common. This is an ancestor.incrementalmissingancestors | |
1130 | object. |
|
1198 | object. | |
1131 |
|
1199 | |||
1132 | 'common' is a list of revision numbers. If common is not supplied, uses |
|
1200 | 'common' is a list of revision numbers. If common is not supplied, uses | |
1133 | nullrev. |
|
1201 | nullrev. | |
1134 | """ |
|
1202 | """ | |
1135 | if common is None: |
|
1203 | if common is None: | |
1136 | common = [nullrev] |
|
1204 | common = [nullrev] | |
1137 |
|
1205 | |||
1138 | if rustancestor is not None and self.index.rust_ext_compat: |
|
1206 | if rustancestor is not None and self.index.rust_ext_compat: | |
1139 | return rustancestor.MissingAncestors(self.index, common) |
|
1207 | return rustancestor.MissingAncestors(self.index, common) | |
1140 | return ancestor.incrementalmissingancestors(self.parentrevs, common) |
|
1208 | return ancestor.incrementalmissingancestors(self.parentrevs, common) | |
1141 |
|
1209 | |||
1142 | def findmissingrevs(self, common=None, heads=None): |
|
1210 | def findmissingrevs(self, common=None, heads=None): | |
1143 | """Return the revision numbers of the ancestors of heads that |
|
1211 | """Return the revision numbers of the ancestors of heads that | |
1144 | are not ancestors of common. |
|
1212 | are not ancestors of common. | |
1145 |
|
1213 | |||
1146 | More specifically, return a list of revision numbers corresponding to |
|
1214 | More specifically, return a list of revision numbers corresponding to | |
1147 | nodes N such that every N satisfies the following constraints: |
|
1215 | nodes N such that every N satisfies the following constraints: | |
1148 |
|
1216 | |||
1149 | 1. N is an ancestor of some node in 'heads' |
|
1217 | 1. N is an ancestor of some node in 'heads' | |
1150 | 2. N is not an ancestor of any node in 'common' |
|
1218 | 2. N is not an ancestor of any node in 'common' | |
1151 |
|
1219 | |||
1152 | The list is sorted by revision number, meaning it is |
|
1220 | The list is sorted by revision number, meaning it is | |
1153 | topologically sorted. |
|
1221 | topologically sorted. | |
1154 |
|
1222 | |||
1155 | 'heads' and 'common' are both lists of revision numbers. If heads is |
|
1223 | 'heads' and 'common' are both lists of revision numbers. If heads is | |
1156 | not supplied, uses all of the revlog's heads. If common is not |
|
1224 | not supplied, uses all of the revlog's heads. If common is not | |
1157 | supplied, uses nullid.""" |
|
1225 | supplied, uses nullid.""" | |
1158 | if common is None: |
|
1226 | if common is None: | |
1159 | common = [nullrev] |
|
1227 | common = [nullrev] | |
1160 | if heads is None: |
|
1228 | if heads is None: | |
1161 | heads = self.headrevs() |
|
1229 | heads = self.headrevs() | |
1162 |
|
1230 | |||
1163 | inc = self.incrementalmissingrevs(common=common) |
|
1231 | inc = self.incrementalmissingrevs(common=common) | |
1164 | return inc.missingancestors(heads) |
|
1232 | return inc.missingancestors(heads) | |
1165 |
|
1233 | |||
1166 | def findmissing(self, common=None, heads=None): |
|
1234 | def findmissing(self, common=None, heads=None): | |
1167 | """Return the ancestors of heads that are not ancestors of common. |
|
1235 | """Return the ancestors of heads that are not ancestors of common. | |
1168 |
|
1236 | |||
1169 | More specifically, return a list of nodes N such that every N |
|
1237 | More specifically, return a list of nodes N such that every N | |
1170 | satisfies the following constraints: |
|
1238 | satisfies the following constraints: | |
1171 |
|
1239 | |||
1172 | 1. N is an ancestor of some node in 'heads' |
|
1240 | 1. N is an ancestor of some node in 'heads' | |
1173 | 2. N is not an ancestor of any node in 'common' |
|
1241 | 2. N is not an ancestor of any node in 'common' | |
1174 |
|
1242 | |||
1175 | The list is sorted by revision number, meaning it is |
|
1243 | The list is sorted by revision number, meaning it is | |
1176 | topologically sorted. |
|
1244 | topologically sorted. | |
1177 |
|
1245 | |||
1178 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1246 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1179 | not supplied, uses all of the revlog's heads. If common is not |
|
1247 | not supplied, uses all of the revlog's heads. If common is not | |
1180 | supplied, uses nullid.""" |
|
1248 | supplied, uses nullid.""" | |
1181 | if common is None: |
|
1249 | if common is None: | |
1182 | common = [self.nullid] |
|
1250 | common = [self.nullid] | |
1183 | if heads is None: |
|
1251 | if heads is None: | |
1184 | heads = self.heads() |
|
1252 | heads = self.heads() | |
1185 |
|
1253 | |||
1186 | common = [self.rev(n) for n in common] |
|
1254 | common = [self.rev(n) for n in common] | |
1187 | heads = [self.rev(n) for n in heads] |
|
1255 | heads = [self.rev(n) for n in heads] | |
1188 |
|
1256 | |||
1189 | inc = self.incrementalmissingrevs(common=common) |
|
1257 | inc = self.incrementalmissingrevs(common=common) | |
1190 | return [self.node(r) for r in inc.missingancestors(heads)] |
|
1258 | return [self.node(r) for r in inc.missingancestors(heads)] | |
1191 |
|
1259 | |||
1192 | def nodesbetween(self, roots=None, heads=None): |
|
1260 | def nodesbetween(self, roots=None, heads=None): | |
1193 | """Return a topological path from 'roots' to 'heads'. |
|
1261 | """Return a topological path from 'roots' to 'heads'. | |
1194 |
|
1262 | |||
1195 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a |
|
1263 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a | |
1196 | topologically sorted list of all nodes N that satisfy both of |
|
1264 | topologically sorted list of all nodes N that satisfy both of | |
1197 | these constraints: |
|
1265 | these constraints: | |
1198 |
|
1266 | |||
1199 | 1. N is a descendant of some node in 'roots' |
|
1267 | 1. N is a descendant of some node in 'roots' | |
1200 | 2. N is an ancestor of some node in 'heads' |
|
1268 | 2. N is an ancestor of some node in 'heads' | |
1201 |
|
1269 | |||
1202 | Every node is considered to be both a descendant and an ancestor |
|
1270 | Every node is considered to be both a descendant and an ancestor | |
1203 | of itself, so every reachable node in 'roots' and 'heads' will be |
|
1271 | of itself, so every reachable node in 'roots' and 'heads' will be | |
1204 | included in 'nodes'. |
|
1272 | included in 'nodes'. | |
1205 |
|
1273 | |||
1206 | 'outroots' is the list of reachable nodes in 'roots', i.e., the |
|
1274 | 'outroots' is the list of reachable nodes in 'roots', i.e., the | |
1207 | subset of 'roots' that is returned in 'nodes'. Likewise, |
|
1275 | subset of 'roots' that is returned in 'nodes'. Likewise, | |
1208 | 'outheads' is the subset of 'heads' that is also in 'nodes'. |
|
1276 | 'outheads' is the subset of 'heads' that is also in 'nodes'. | |
1209 |
|
1277 | |||
1210 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is |
|
1278 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is | |
1211 | unspecified, uses nullid as the only root. If 'heads' is |
|
1279 | unspecified, uses nullid as the only root. If 'heads' is | |
1212 | unspecified, uses list of all of the revlog's heads.""" |
|
1280 | unspecified, uses list of all of the revlog's heads.""" | |
1213 | nonodes = ([], [], []) |
|
1281 | nonodes = ([], [], []) | |
1214 | if roots is not None: |
|
1282 | if roots is not None: | |
1215 | roots = list(roots) |
|
1283 | roots = list(roots) | |
1216 | if not roots: |
|
1284 | if not roots: | |
1217 | return nonodes |
|
1285 | return nonodes | |
1218 | lowestrev = min([self.rev(n) for n in roots]) |
|
1286 | lowestrev = min([self.rev(n) for n in roots]) | |
1219 | else: |
|
1287 | else: | |
1220 | roots = [self.nullid] # Everybody's a descendant of nullid |
|
1288 | roots = [self.nullid] # Everybody's a descendant of nullid | |
1221 | lowestrev = nullrev |
|
1289 | lowestrev = nullrev | |
1222 | if (lowestrev == nullrev) and (heads is None): |
|
1290 | if (lowestrev == nullrev) and (heads is None): | |
1223 | # We want _all_ the nodes! |
|
1291 | # We want _all_ the nodes! | |
1224 | return ( |
|
1292 | return ( | |
1225 | [self.node(r) for r in self], |
|
1293 | [self.node(r) for r in self], | |
1226 | [self.nullid], |
|
1294 | [self.nullid], | |
1227 | list(self.heads()), |
|
1295 | list(self.heads()), | |
1228 | ) |
|
1296 | ) | |
1229 | if heads is None: |
|
1297 | if heads is None: | |
1230 | # All nodes are ancestors, so the latest ancestor is the last |
|
1298 | # All nodes are ancestors, so the latest ancestor is the last | |
1231 | # node. |
|
1299 | # node. | |
1232 | highestrev = len(self) - 1 |
|
1300 | highestrev = len(self) - 1 | |
1233 | # Set ancestors to None to signal that every node is an ancestor. |
|
1301 | # Set ancestors to None to signal that every node is an ancestor. | |
1234 | ancestors = None |
|
1302 | ancestors = None | |
1235 | # Set heads to an empty dictionary for later discovery of heads |
|
1303 | # Set heads to an empty dictionary for later discovery of heads | |
1236 | heads = {} |
|
1304 | heads = {} | |
1237 | else: |
|
1305 | else: | |
1238 | heads = list(heads) |
|
1306 | heads = list(heads) | |
1239 | if not heads: |
|
1307 | if not heads: | |
1240 | return nonodes |
|
1308 | return nonodes | |
1241 | ancestors = set() |
|
1309 | ancestors = set() | |
1242 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
1310 | # Turn heads into a dictionary so we can remove 'fake' heads. | |
1243 | # Also, later we will be using it to filter out the heads we can't |
|
1311 | # Also, later we will be using it to filter out the heads we can't | |
1244 | # find from roots. |
|
1312 | # find from roots. | |
1245 | heads = dict.fromkeys(heads, False) |
|
1313 | heads = dict.fromkeys(heads, False) | |
1246 | # Start at the top and keep marking parents until we're done. |
|
1314 | # Start at the top and keep marking parents until we're done. | |
1247 | nodestotag = set(heads) |
|
1315 | nodestotag = set(heads) | |
1248 | # Remember where the top was so we can use it as a limit later. |
|
1316 | # Remember where the top was so we can use it as a limit later. | |
1249 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
1317 | highestrev = max([self.rev(n) for n in nodestotag]) | |
1250 | while nodestotag: |
|
1318 | while nodestotag: | |
1251 | # grab a node to tag |
|
1319 | # grab a node to tag | |
1252 | n = nodestotag.pop() |
|
1320 | n = nodestotag.pop() | |
1253 | # Never tag nullid |
|
1321 | # Never tag nullid | |
1254 | if n == self.nullid: |
|
1322 | if n == self.nullid: | |
1255 | continue |
|
1323 | continue | |
1256 | # A node's revision number represents its place in a |
|
1324 | # A node's revision number represents its place in a | |
1257 | # topologically sorted list of nodes. |
|
1325 | # topologically sorted list of nodes. | |
1258 | r = self.rev(n) |
|
1326 | r = self.rev(n) | |
1259 | if r >= lowestrev: |
|
1327 | if r >= lowestrev: | |
1260 | if n not in ancestors: |
|
1328 | if n not in ancestors: | |
1261 | # If we are possibly a descendant of one of the roots |
|
1329 | # If we are possibly a descendant of one of the roots | |
1262 | # and we haven't already been marked as an ancestor |
|
1330 | # and we haven't already been marked as an ancestor | |
1263 | ancestors.add(n) # Mark as ancestor |
|
1331 | ancestors.add(n) # Mark as ancestor | |
1264 | # Add non-nullid parents to list of nodes to tag. |
|
1332 | # Add non-nullid parents to list of nodes to tag. | |
1265 | nodestotag.update( |
|
1333 | nodestotag.update( | |
1266 | [p for p in self.parents(n) if p != self.nullid] |
|
1334 | [p for p in self.parents(n) if p != self.nullid] | |
1267 | ) |
|
1335 | ) | |
1268 | elif n in heads: # We've seen it before, is it a fake head? |
|
1336 | elif n in heads: # We've seen it before, is it a fake head? | |
1269 | # So it is, real heads should not be the ancestors of |
|
1337 | # So it is, real heads should not be the ancestors of | |
1270 | # any other heads. |
|
1338 | # any other heads. | |
1271 | heads.pop(n) |
|
1339 | heads.pop(n) | |
1272 | if not ancestors: |
|
1340 | if not ancestors: | |
1273 | return nonodes |
|
1341 | return nonodes | |
1274 | # Now that we have our set of ancestors, we want to remove any |
|
1342 | # Now that we have our set of ancestors, we want to remove any | |
1275 | # roots that are not ancestors. |
|
1343 | # roots that are not ancestors. | |
1276 |
|
1344 | |||
1277 | # If one of the roots was nullid, everything is included anyway. |
|
1345 | # If one of the roots was nullid, everything is included anyway. | |
1278 | if lowestrev > nullrev: |
|
1346 | if lowestrev > nullrev: | |
1279 | # But, since we weren't, let's recompute the lowest rev to not |
|
1347 | # But, since we weren't, let's recompute the lowest rev to not | |
1280 | # include roots that aren't ancestors. |
|
1348 | # include roots that aren't ancestors. | |
1281 |
|
1349 | |||
1282 | # Filter out roots that aren't ancestors of heads |
|
1350 | # Filter out roots that aren't ancestors of heads | |
1283 | roots = [root for root in roots if root in ancestors] |
|
1351 | roots = [root for root in roots if root in ancestors] | |
1284 | # Recompute the lowest revision |
|
1352 | # Recompute the lowest revision | |
1285 | if roots: |
|
1353 | if roots: | |
1286 | lowestrev = min([self.rev(root) for root in roots]) |
|
1354 | lowestrev = min([self.rev(root) for root in roots]) | |
1287 | else: |
|
1355 | else: | |
1288 | # No more roots? Return empty list |
|
1356 | # No more roots? Return empty list | |
1289 | return nonodes |
|
1357 | return nonodes | |
1290 | else: |
|
1358 | else: | |
1291 | # We are descending from nullid, and don't need to care about |
|
1359 | # We are descending from nullid, and don't need to care about | |
1292 | # any other roots. |
|
1360 | # any other roots. | |
1293 | lowestrev = nullrev |
|
1361 | lowestrev = nullrev | |
1294 | roots = [self.nullid] |
|
1362 | roots = [self.nullid] | |
1295 | # Transform our roots list into a set. |
|
1363 | # Transform our roots list into a set. | |
1296 | descendants = set(roots) |
|
1364 | descendants = set(roots) | |
1297 | # Also, keep the original roots so we can filter out roots that aren't |
|
1365 | # Also, keep the original roots so we can filter out roots that aren't | |
1298 | # 'real' roots (i.e. are descended from other roots). |
|
1366 | # 'real' roots (i.e. are descended from other roots). | |
1299 | roots = descendants.copy() |
|
1367 | roots = descendants.copy() | |
1300 | # Our topologically sorted list of output nodes. |
|
1368 | # Our topologically sorted list of output nodes. | |
1301 | orderedout = [] |
|
1369 | orderedout = [] | |
1302 | # Don't start at nullid since we don't want nullid in our output list, |
|
1370 | # Don't start at nullid since we don't want nullid in our output list, | |
1303 | # and if nullid shows up in descendants, empty parents will look like |
|
1371 | # and if nullid shows up in descendants, empty parents will look like | |
1304 | # they're descendants. |
|
1372 | # they're descendants. | |
1305 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): |
|
1373 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): | |
1306 | n = self.node(r) |
|
1374 | n = self.node(r) | |
1307 | isdescendant = False |
|
1375 | isdescendant = False | |
1308 | if lowestrev == nullrev: # Everybody is a descendant of nullid |
|
1376 | if lowestrev == nullrev: # Everybody is a descendant of nullid | |
1309 | isdescendant = True |
|
1377 | isdescendant = True | |
1310 | elif n in descendants: |
|
1378 | elif n in descendants: | |
1311 | # n is already a descendant |
|
1379 | # n is already a descendant | |
1312 | isdescendant = True |
|
1380 | isdescendant = True | |
1313 | # This check only needs to be done here because all the roots |
|
1381 | # This check only needs to be done here because all the roots | |
1314 | # will start being marked is descendants before the loop. |
|
1382 | # will start being marked is descendants before the loop. | |
1315 | if n in roots: |
|
1383 | if n in roots: | |
1316 | # If n was a root, check if it's a 'real' root. |
|
1384 | # If n was a root, check if it's a 'real' root. | |
1317 | p = tuple(self.parents(n)) |
|
1385 | p = tuple(self.parents(n)) | |
1318 | # If any of its parents are descendants, it's not a root. |
|
1386 | # If any of its parents are descendants, it's not a root. | |
1319 | if (p[0] in descendants) or (p[1] in descendants): |
|
1387 | if (p[0] in descendants) or (p[1] in descendants): | |
1320 | roots.remove(n) |
|
1388 | roots.remove(n) | |
1321 | else: |
|
1389 | else: | |
1322 | p = tuple(self.parents(n)) |
|
1390 | p = tuple(self.parents(n)) | |
1323 | # A node is a descendant if either of its parents are |
|
1391 | # A node is a descendant if either of its parents are | |
1324 | # descendants. (We seeded the dependents list with the roots |
|
1392 | # descendants. (We seeded the dependents list with the roots | |
1325 | # up there, remember?) |
|
1393 | # up there, remember?) | |
1326 | if (p[0] in descendants) or (p[1] in descendants): |
|
1394 | if (p[0] in descendants) or (p[1] in descendants): | |
1327 | descendants.add(n) |
|
1395 | descendants.add(n) | |
1328 | isdescendant = True |
|
1396 | isdescendant = True | |
1329 | if isdescendant and ((ancestors is None) or (n in ancestors)): |
|
1397 | if isdescendant and ((ancestors is None) or (n in ancestors)): | |
1330 | # Only include nodes that are both descendants and ancestors. |
|
1398 | # Only include nodes that are both descendants and ancestors. | |
1331 | orderedout.append(n) |
|
1399 | orderedout.append(n) | |
1332 | if (ancestors is not None) and (n in heads): |
|
1400 | if (ancestors is not None) and (n in heads): | |
1333 | # We're trying to figure out which heads are reachable |
|
1401 | # We're trying to figure out which heads are reachable | |
1334 | # from roots. |
|
1402 | # from roots. | |
1335 | # Mark this head as having been reached |
|
1403 | # Mark this head as having been reached | |
1336 | heads[n] = True |
|
1404 | heads[n] = True | |
1337 | elif ancestors is None: |
|
1405 | elif ancestors is None: | |
1338 | # Otherwise, we're trying to discover the heads. |
|
1406 | # Otherwise, we're trying to discover the heads. | |
1339 | # Assume this is a head because if it isn't, the next step |
|
1407 | # Assume this is a head because if it isn't, the next step | |
1340 | # will eventually remove it. |
|
1408 | # will eventually remove it. | |
1341 | heads[n] = True |
|
1409 | heads[n] = True | |
1342 | # But, obviously its parents aren't. |
|
1410 | # But, obviously its parents aren't. | |
1343 | for p in self.parents(n): |
|
1411 | for p in self.parents(n): | |
1344 | heads.pop(p, None) |
|
1412 | heads.pop(p, None) | |
1345 | heads = [head for head, flag in heads.items() if flag] |
|
1413 | heads = [head for head, flag in heads.items() if flag] | |
1346 | roots = list(roots) |
|
1414 | roots = list(roots) | |
1347 | assert orderedout |
|
1415 | assert orderedout | |
1348 | assert roots |
|
1416 | assert roots | |
1349 | assert heads |
|
1417 | assert heads | |
1350 | return (orderedout, roots, heads) |
|
1418 | return (orderedout, roots, heads) | |
1351 |
|
1419 | |||
1352 | def headrevs(self, revs=None): |
|
1420 | def headrevs(self, revs=None): | |
1353 | if revs is None: |
|
1421 | if revs is None: | |
1354 | try: |
|
1422 | try: | |
1355 | return self.index.headrevs() |
|
1423 | return self.index.headrevs() | |
1356 | except AttributeError: |
|
1424 | except AttributeError: | |
1357 | return self._headrevs() |
|
1425 | return self._headrevs() | |
1358 | if rustdagop is not None and self.index.rust_ext_compat: |
|
1426 | if rustdagop is not None and self.index.rust_ext_compat: | |
1359 | return rustdagop.headrevs(self.index, revs) |
|
1427 | return rustdagop.headrevs(self.index, revs) | |
1360 | return dagop.headrevs(revs, self._uncheckedparentrevs) |
|
1428 | return dagop.headrevs(revs, self._uncheckedparentrevs) | |
1361 |
|
1429 | |||
1362 | def computephases(self, roots): |
|
1430 | def computephases(self, roots): | |
1363 | return self.index.computephasesmapsets(roots) |
|
1431 | return self.index.computephasesmapsets(roots) | |
1364 |
|
1432 | |||
1365 | def _headrevs(self): |
|
1433 | def _headrevs(self): | |
1366 | count = len(self) |
|
1434 | count = len(self) | |
1367 | if not count: |
|
1435 | if not count: | |
1368 | return [nullrev] |
|
1436 | return [nullrev] | |
1369 | # we won't iter over filtered rev so nobody is a head at start |
|
1437 | # we won't iter over filtered rev so nobody is a head at start | |
1370 | ishead = [0] * (count + 1) |
|
1438 | ishead = [0] * (count + 1) | |
1371 | index = self.index |
|
1439 | index = self.index | |
1372 | for r in self: |
|
1440 | for r in self: | |
1373 | ishead[r] = 1 # I may be an head |
|
1441 | ishead[r] = 1 # I may be an head | |
1374 | e = index[r] |
|
1442 | e = index[r] | |
1375 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not |
|
1443 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not | |
1376 | return [r for r, val in enumerate(ishead) if val] |
|
1444 | return [r for r, val in enumerate(ishead) if val] | |
1377 |
|
1445 | |||
1378 | def heads(self, start=None, stop=None): |
|
1446 | def heads(self, start=None, stop=None): | |
1379 | """return the list of all nodes that have no children |
|
1447 | """return the list of all nodes that have no children | |
1380 |
|
1448 | |||
1381 | if start is specified, only heads that are descendants of |
|
1449 | if start is specified, only heads that are descendants of | |
1382 | start will be returned |
|
1450 | start will be returned | |
1383 | if stop is specified, it will consider all the revs from stop |
|
1451 | if stop is specified, it will consider all the revs from stop | |
1384 | as if they had no children |
|
1452 | as if they had no children | |
1385 | """ |
|
1453 | """ | |
1386 | if start is None and stop is None: |
|
1454 | if start is None and stop is None: | |
1387 | if not len(self): |
|
1455 | if not len(self): | |
1388 | return [self.nullid] |
|
1456 | return [self.nullid] | |
1389 | return [self.node(r) for r in self.headrevs()] |
|
1457 | return [self.node(r) for r in self.headrevs()] | |
1390 |
|
1458 | |||
1391 | if start is None: |
|
1459 | if start is None: | |
1392 | start = nullrev |
|
1460 | start = nullrev | |
1393 | else: |
|
1461 | else: | |
1394 | start = self.rev(start) |
|
1462 | start = self.rev(start) | |
1395 |
|
1463 | |||
1396 | stoprevs = {self.rev(n) for n in stop or []} |
|
1464 | stoprevs = {self.rev(n) for n in stop or []} | |
1397 |
|
1465 | |||
1398 | revs = dagop.headrevssubset( |
|
1466 | revs = dagop.headrevssubset( | |
1399 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs |
|
1467 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs | |
1400 | ) |
|
1468 | ) | |
1401 |
|
1469 | |||
1402 | return [self.node(rev) for rev in revs] |
|
1470 | return [self.node(rev) for rev in revs] | |
1403 |
|
1471 | |||
1404 | def children(self, node): |
|
1472 | def children(self, node): | |
1405 | """find the children of a given node""" |
|
1473 | """find the children of a given node""" | |
1406 | c = [] |
|
1474 | c = [] | |
1407 | p = self.rev(node) |
|
1475 | p = self.rev(node) | |
1408 | for r in self.revs(start=p + 1): |
|
1476 | for r in self.revs(start=p + 1): | |
1409 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
1477 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
1410 | if prevs: |
|
1478 | if prevs: | |
1411 | for pr in prevs: |
|
1479 | for pr in prevs: | |
1412 | if pr == p: |
|
1480 | if pr == p: | |
1413 | c.append(self.node(r)) |
|
1481 | c.append(self.node(r)) | |
1414 | elif p == nullrev: |
|
1482 | elif p == nullrev: | |
1415 | c.append(self.node(r)) |
|
1483 | c.append(self.node(r)) | |
1416 | return c |
|
1484 | return c | |
1417 |
|
1485 | |||
1418 | def commonancestorsheads(self, a, b): |
|
1486 | def commonancestorsheads(self, a, b): | |
1419 | """calculate all the heads of the common ancestors of nodes a and b""" |
|
1487 | """calculate all the heads of the common ancestors of nodes a and b""" | |
1420 | a, b = self.rev(a), self.rev(b) |
|
1488 | a, b = self.rev(a), self.rev(b) | |
1421 | ancs = self._commonancestorsheads(a, b) |
|
1489 | ancs = self._commonancestorsheads(a, b) | |
1422 | return pycompat.maplist(self.node, ancs) |
|
1490 | return pycompat.maplist(self.node, ancs) | |
1423 |
|
1491 | |||
1424 | def _commonancestorsheads(self, *revs): |
|
1492 | def _commonancestorsheads(self, *revs): | |
1425 | """calculate all the heads of the common ancestors of revs""" |
|
1493 | """calculate all the heads of the common ancestors of revs""" | |
1426 | try: |
|
1494 | try: | |
1427 | ancs = self.index.commonancestorsheads(*revs) |
|
1495 | ancs = self.index.commonancestorsheads(*revs) | |
1428 | except (AttributeError, OverflowError): # C implementation failed |
|
1496 | except (AttributeError, OverflowError): # C implementation failed | |
1429 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) |
|
1497 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) | |
1430 | return ancs |
|
1498 | return ancs | |
1431 |
|
1499 | |||
1432 | def isancestor(self, a, b): |
|
1500 | def isancestor(self, a, b): | |
1433 | """return True if node a is an ancestor of node b |
|
1501 | """return True if node a is an ancestor of node b | |
1434 |
|
1502 | |||
1435 | A revision is considered an ancestor of itself.""" |
|
1503 | A revision is considered an ancestor of itself.""" | |
1436 | a, b = self.rev(a), self.rev(b) |
|
1504 | a, b = self.rev(a), self.rev(b) | |
1437 | return self.isancestorrev(a, b) |
|
1505 | return self.isancestorrev(a, b) | |
1438 |
|
1506 | |||
1439 | def isancestorrev(self, a, b): |
|
1507 | def isancestorrev(self, a, b): | |
1440 | """return True if revision a is an ancestor of revision b |
|
1508 | """return True if revision a is an ancestor of revision b | |
1441 |
|
1509 | |||
1442 | A revision is considered an ancestor of itself. |
|
1510 | A revision is considered an ancestor of itself. | |
1443 |
|
1511 | |||
1444 | The implementation of this is trivial but the use of |
|
1512 | The implementation of this is trivial but the use of | |
1445 | reachableroots is not.""" |
|
1513 | reachableroots is not.""" | |
1446 | if a == nullrev: |
|
1514 | if a == nullrev: | |
1447 | return True |
|
1515 | return True | |
1448 | elif a == b: |
|
1516 | elif a == b: | |
1449 | return True |
|
1517 | return True | |
1450 | elif a > b: |
|
1518 | elif a > b: | |
1451 | return False |
|
1519 | return False | |
1452 | return bool(self.reachableroots(a, [b], [a], includepath=False)) |
|
1520 | return bool(self.reachableroots(a, [b], [a], includepath=False)) | |
1453 |
|
1521 | |||
1454 | def reachableroots(self, minroot, heads, roots, includepath=False): |
|
1522 | def reachableroots(self, minroot, heads, roots, includepath=False): | |
1455 | """return (heads(::(<roots> and <roots>::<heads>))) |
|
1523 | """return (heads(::(<roots> and <roots>::<heads>))) | |
1456 |
|
1524 | |||
1457 | If includepath is True, return (<roots>::<heads>).""" |
|
1525 | If includepath is True, return (<roots>::<heads>).""" | |
1458 | try: |
|
1526 | try: | |
1459 | return self.index.reachableroots2( |
|
1527 | return self.index.reachableroots2( | |
1460 | minroot, heads, roots, includepath |
|
1528 | minroot, heads, roots, includepath | |
1461 | ) |
|
1529 | ) | |
1462 | except AttributeError: |
|
1530 | except AttributeError: | |
1463 | return dagop._reachablerootspure( |
|
1531 | return dagop._reachablerootspure( | |
1464 | self.parentrevs, minroot, roots, heads, includepath |
|
1532 | self.parentrevs, minroot, roots, heads, includepath | |
1465 | ) |
|
1533 | ) | |
1466 |
|
1534 | |||
1467 | def ancestor(self, a, b): |
|
1535 | def ancestor(self, a, b): | |
1468 | """calculate the "best" common ancestor of nodes a and b""" |
|
1536 | """calculate the "best" common ancestor of nodes a and b""" | |
1469 |
|
1537 | |||
1470 | a, b = self.rev(a), self.rev(b) |
|
1538 | a, b = self.rev(a), self.rev(b) | |
1471 | try: |
|
1539 | try: | |
1472 | ancs = self.index.ancestors(a, b) |
|
1540 | ancs = self.index.ancestors(a, b) | |
1473 | except (AttributeError, OverflowError): |
|
1541 | except (AttributeError, OverflowError): | |
1474 | ancs = ancestor.ancestors(self.parentrevs, a, b) |
|
1542 | ancs = ancestor.ancestors(self.parentrevs, a, b) | |
1475 | if ancs: |
|
1543 | if ancs: | |
1476 | # choose a consistent winner when there's a tie |
|
1544 | # choose a consistent winner when there's a tie | |
1477 | return min(map(self.node, ancs)) |
|
1545 | return min(map(self.node, ancs)) | |
1478 | return self.nullid |
|
1546 | return self.nullid | |
1479 |
|
1547 | |||
1480 | def _match(self, id): |
|
1548 | def _match(self, id): | |
1481 | if isinstance(id, int): |
|
1549 | if isinstance(id, int): | |
1482 | # rev |
|
1550 | # rev | |
1483 | return self.node(id) |
|
1551 | return self.node(id) | |
1484 | if len(id) == self.nodeconstants.nodelen: |
|
1552 | if len(id) == self.nodeconstants.nodelen: | |
1485 | # possibly a binary node |
|
1553 | # possibly a binary node | |
1486 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
1554 | # odds of a binary node being all hex in ASCII are 1 in 10**25 | |
1487 | try: |
|
1555 | try: | |
1488 | node = id |
|
1556 | node = id | |
1489 | self.rev(node) # quick search the index |
|
1557 | self.rev(node) # quick search the index | |
1490 | return node |
|
1558 | return node | |
1491 | except error.LookupError: |
|
1559 | except error.LookupError: | |
1492 | pass # may be partial hex id |
|
1560 | pass # may be partial hex id | |
1493 | try: |
|
1561 | try: | |
1494 | # str(rev) |
|
1562 | # str(rev) | |
1495 | rev = int(id) |
|
1563 | rev = int(id) | |
1496 | if b"%d" % rev != id: |
|
1564 | if b"%d" % rev != id: | |
1497 | raise ValueError |
|
1565 | raise ValueError | |
1498 | if rev < 0: |
|
1566 | if rev < 0: | |
1499 | rev = len(self) + rev |
|
1567 | rev = len(self) + rev | |
1500 | if rev < 0 or rev >= len(self): |
|
1568 | if rev < 0 or rev >= len(self): | |
1501 | raise ValueError |
|
1569 | raise ValueError | |
1502 | return self.node(rev) |
|
1570 | return self.node(rev) | |
1503 | except (ValueError, OverflowError): |
|
1571 | except (ValueError, OverflowError): | |
1504 | pass |
|
1572 | pass | |
1505 | if len(id) == 2 * self.nodeconstants.nodelen: |
|
1573 | if len(id) == 2 * self.nodeconstants.nodelen: | |
1506 | try: |
|
1574 | try: | |
1507 | # a full hex nodeid? |
|
1575 | # a full hex nodeid? | |
1508 | node = bin(id) |
|
1576 | node = bin(id) | |
1509 | self.rev(node) |
|
1577 | self.rev(node) | |
1510 | return node |
|
1578 | return node | |
1511 | except (binascii.Error, error.LookupError): |
|
1579 | except (binascii.Error, error.LookupError): | |
1512 | pass |
|
1580 | pass | |
1513 |
|
1581 | |||
1514 | def _partialmatch(self, id): |
|
1582 | def _partialmatch(self, id): | |
1515 | # we don't care wdirfilenodeids as they should be always full hash |
|
1583 | # we don't care wdirfilenodeids as they should be always full hash | |
1516 | maybewdir = self.nodeconstants.wdirhex.startswith(id) |
|
1584 | maybewdir = self.nodeconstants.wdirhex.startswith(id) | |
1517 | ambiguous = False |
|
1585 | ambiguous = False | |
1518 | try: |
|
1586 | try: | |
1519 | partial = self.index.partialmatch(id) |
|
1587 | partial = self.index.partialmatch(id) | |
1520 | if partial and self.hasnode(partial): |
|
1588 | if partial and self.hasnode(partial): | |
1521 | if maybewdir: |
|
1589 | if maybewdir: | |
1522 | # single 'ff...' match in radix tree, ambiguous with wdir |
|
1590 | # single 'ff...' match in radix tree, ambiguous with wdir | |
1523 | ambiguous = True |
|
1591 | ambiguous = True | |
1524 | else: |
|
1592 | else: | |
1525 | return partial |
|
1593 | return partial | |
1526 | elif maybewdir: |
|
1594 | elif maybewdir: | |
1527 | # no 'ff...' match in radix tree, wdir identified |
|
1595 | # no 'ff...' match in radix tree, wdir identified | |
1528 | raise error.WdirUnsupported |
|
1596 | raise error.WdirUnsupported | |
1529 | else: |
|
1597 | else: | |
1530 | return None |
|
1598 | return None | |
1531 | except error.RevlogError: |
|
1599 | except error.RevlogError: | |
1532 | # parsers.c radix tree lookup gave multiple matches |
|
1600 | # parsers.c radix tree lookup gave multiple matches | |
1533 | # fast path: for unfiltered changelog, radix tree is accurate |
|
1601 | # fast path: for unfiltered changelog, radix tree is accurate | |
1534 | if not getattr(self, 'filteredrevs', None): |
|
1602 | if not getattr(self, 'filteredrevs', None): | |
1535 | ambiguous = True |
|
1603 | ambiguous = True | |
1536 | # fall through to slow path that filters hidden revisions |
|
1604 | # fall through to slow path that filters hidden revisions | |
1537 | except (AttributeError, ValueError): |
|
1605 | except (AttributeError, ValueError): | |
1538 | # we are pure python, or key is not hex |
|
1606 | # we are pure python, or key is not hex | |
1539 | pass |
|
1607 | pass | |
1540 | if ambiguous: |
|
1608 | if ambiguous: | |
1541 | raise error.AmbiguousPrefixLookupError( |
|
1609 | raise error.AmbiguousPrefixLookupError( | |
1542 | id, self.display_id, _(b'ambiguous identifier') |
|
1610 | id, self.display_id, _(b'ambiguous identifier') | |
1543 | ) |
|
1611 | ) | |
1544 |
|
1612 | |||
1545 | if id in self._pcache: |
|
1613 | if id in self._pcache: | |
1546 | return self._pcache[id] |
|
1614 | return self._pcache[id] | |
1547 |
|
1615 | |||
1548 | if len(id) <= 40: |
|
1616 | if len(id) <= 40: | |
1549 | # hex(node)[:...] |
|
1617 | # hex(node)[:...] | |
1550 | l = len(id) // 2 * 2 # grab an even number of digits |
|
1618 | l = len(id) // 2 * 2 # grab an even number of digits | |
1551 | try: |
|
1619 | try: | |
1552 | # we're dropping the last digit, so let's check that it's hex, |
|
1620 | # we're dropping the last digit, so let's check that it's hex, | |
1553 | # to avoid the expensive computation below if it's not |
|
1621 | # to avoid the expensive computation below if it's not | |
1554 | if len(id) % 2 > 0: |
|
1622 | if len(id) % 2 > 0: | |
1555 | if not (id[-1] in hexdigits): |
|
1623 | if not (id[-1] in hexdigits): | |
1556 | return None |
|
1624 | return None | |
1557 | prefix = bin(id[:l]) |
|
1625 | prefix = bin(id[:l]) | |
1558 | except binascii.Error: |
|
1626 | except binascii.Error: | |
1559 | pass |
|
1627 | pass | |
1560 | else: |
|
1628 | else: | |
1561 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] |
|
1629 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] | |
1562 | nl = [ |
|
1630 | nl = [ | |
1563 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) |
|
1631 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) | |
1564 | ] |
|
1632 | ] | |
1565 | if self.nodeconstants.nullhex.startswith(id): |
|
1633 | if self.nodeconstants.nullhex.startswith(id): | |
1566 | nl.append(self.nullid) |
|
1634 | nl.append(self.nullid) | |
1567 | if len(nl) > 0: |
|
1635 | if len(nl) > 0: | |
1568 | if len(nl) == 1 and not maybewdir: |
|
1636 | if len(nl) == 1 and not maybewdir: | |
1569 | self._pcache[id] = nl[0] |
|
1637 | self._pcache[id] = nl[0] | |
1570 | return nl[0] |
|
1638 | return nl[0] | |
1571 | raise error.AmbiguousPrefixLookupError( |
|
1639 | raise error.AmbiguousPrefixLookupError( | |
1572 | id, self.display_id, _(b'ambiguous identifier') |
|
1640 | id, self.display_id, _(b'ambiguous identifier') | |
1573 | ) |
|
1641 | ) | |
1574 | if maybewdir: |
|
1642 | if maybewdir: | |
1575 | raise error.WdirUnsupported |
|
1643 | raise error.WdirUnsupported | |
1576 | return None |
|
1644 | return None | |
1577 |
|
1645 | |||
1578 | def lookup(self, id): |
|
1646 | def lookup(self, id): | |
1579 | """locate a node based on: |
|
1647 | """locate a node based on: | |
1580 | - revision number or str(revision number) |
|
1648 | - revision number or str(revision number) | |
1581 | - nodeid or subset of hex nodeid |
|
1649 | - nodeid or subset of hex nodeid | |
1582 | """ |
|
1650 | """ | |
1583 | n = self._match(id) |
|
1651 | n = self._match(id) | |
1584 | if n is not None: |
|
1652 | if n is not None: | |
1585 | return n |
|
1653 | return n | |
1586 | n = self._partialmatch(id) |
|
1654 | n = self._partialmatch(id) | |
1587 | if n: |
|
1655 | if n: | |
1588 | return n |
|
1656 | return n | |
1589 |
|
1657 | |||
1590 | raise error.LookupError(id, self.display_id, _(b'no match found')) |
|
1658 | raise error.LookupError(id, self.display_id, _(b'no match found')) | |
1591 |
|
1659 | |||
1592 | def shortest(self, node, minlength=1): |
|
1660 | def shortest(self, node, minlength=1): | |
1593 | """Find the shortest unambiguous prefix that matches node.""" |
|
1661 | """Find the shortest unambiguous prefix that matches node.""" | |
1594 |
|
1662 | |||
1595 | def isvalid(prefix): |
|
1663 | def isvalid(prefix): | |
1596 | try: |
|
1664 | try: | |
1597 | matchednode = self._partialmatch(prefix) |
|
1665 | matchednode = self._partialmatch(prefix) | |
1598 | except error.AmbiguousPrefixLookupError: |
|
1666 | except error.AmbiguousPrefixLookupError: | |
1599 | return False |
|
1667 | return False | |
1600 | except error.WdirUnsupported: |
|
1668 | except error.WdirUnsupported: | |
1601 | # single 'ff...' match |
|
1669 | # single 'ff...' match | |
1602 | return True |
|
1670 | return True | |
1603 | if matchednode is None: |
|
1671 | if matchednode is None: | |
1604 | raise error.LookupError(node, self.display_id, _(b'no node')) |
|
1672 | raise error.LookupError(node, self.display_id, _(b'no node')) | |
1605 | return True |
|
1673 | return True | |
1606 |
|
1674 | |||
1607 | def maybewdir(prefix): |
|
1675 | def maybewdir(prefix): | |
1608 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) |
|
1676 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) | |
1609 |
|
1677 | |||
1610 | hexnode = hex(node) |
|
1678 | hexnode = hex(node) | |
1611 |
|
1679 | |||
1612 | def disambiguate(hexnode, minlength): |
|
1680 | def disambiguate(hexnode, minlength): | |
1613 | """Disambiguate against wdirid.""" |
|
1681 | """Disambiguate against wdirid.""" | |
1614 | for length in range(minlength, len(hexnode) + 1): |
|
1682 | for length in range(minlength, len(hexnode) + 1): | |
1615 | prefix = hexnode[:length] |
|
1683 | prefix = hexnode[:length] | |
1616 | if not maybewdir(prefix): |
|
1684 | if not maybewdir(prefix): | |
1617 | return prefix |
|
1685 | return prefix | |
1618 |
|
1686 | |||
1619 | if not getattr(self, 'filteredrevs', None): |
|
1687 | if not getattr(self, 'filteredrevs', None): | |
1620 | try: |
|
1688 | try: | |
1621 | length = max(self.index.shortest(node), minlength) |
|
1689 | length = max(self.index.shortest(node), minlength) | |
1622 | return disambiguate(hexnode, length) |
|
1690 | return disambiguate(hexnode, length) | |
1623 | except error.RevlogError: |
|
1691 | except error.RevlogError: | |
1624 | if node != self.nodeconstants.wdirid: |
|
1692 | if node != self.nodeconstants.wdirid: | |
1625 | raise error.LookupError( |
|
1693 | raise error.LookupError( | |
1626 | node, self.display_id, _(b'no node') |
|
1694 | node, self.display_id, _(b'no node') | |
1627 | ) |
|
1695 | ) | |
1628 | except AttributeError: |
|
1696 | except AttributeError: | |
1629 | # Fall through to pure code |
|
1697 | # Fall through to pure code | |
1630 | pass |
|
1698 | pass | |
1631 |
|
1699 | |||
1632 | if node == self.nodeconstants.wdirid: |
|
1700 | if node == self.nodeconstants.wdirid: | |
1633 | for length in range(minlength, len(hexnode) + 1): |
|
1701 | for length in range(minlength, len(hexnode) + 1): | |
1634 | prefix = hexnode[:length] |
|
1702 | prefix = hexnode[:length] | |
1635 | if isvalid(prefix): |
|
1703 | if isvalid(prefix): | |
1636 | return prefix |
|
1704 | return prefix | |
1637 |
|
1705 | |||
1638 | for length in range(minlength, len(hexnode) + 1): |
|
1706 | for length in range(minlength, len(hexnode) + 1): | |
1639 | prefix = hexnode[:length] |
|
1707 | prefix = hexnode[:length] | |
1640 | if isvalid(prefix): |
|
1708 | if isvalid(prefix): | |
1641 | return disambiguate(hexnode, length) |
|
1709 | return disambiguate(hexnode, length) | |
1642 |
|
1710 | |||
1643 | def cmp(self, node, text): |
|
1711 | def cmp(self, node, text): | |
1644 | """compare text with a given file revision |
|
1712 | """compare text with a given file revision | |
1645 |
|
1713 | |||
1646 | returns True if text is different than what is stored. |
|
1714 | returns True if text is different than what is stored. | |
1647 | """ |
|
1715 | """ | |
1648 | p1, p2 = self.parents(node) |
|
1716 | p1, p2 = self.parents(node) | |
1649 | return storageutil.hashrevisionsha1(text, p1, p2) != node |
|
1717 | return storageutil.hashrevisionsha1(text, p1, p2) != node | |
1650 |
|
1718 | |||
1651 | def _getsegmentforrevs(self, startrev, endrev, df=None): |
|
1719 | def _getsegmentforrevs(self, startrev, endrev, df=None): | |
1652 | """Obtain a segment of raw data corresponding to a range of revisions. |
|
1720 | """Obtain a segment of raw data corresponding to a range of revisions. | |
1653 |
|
1721 | |||
1654 | Accepts the start and end revisions and an optional already-open |
|
1722 | Accepts the start and end revisions and an optional already-open | |
1655 | file handle to be used for reading. If the file handle is read, its |
|
1723 | file handle to be used for reading. If the file handle is read, its | |
1656 | seek position will not be preserved. |
|
1724 | seek position will not be preserved. | |
1657 |
|
1725 | |||
1658 | Requests for data may be satisfied by a cache. |
|
1726 | Requests for data may be satisfied by a cache. | |
1659 |
|
1727 | |||
1660 | Returns a 2-tuple of (offset, data) for the requested range of |
|
1728 | Returns a 2-tuple of (offset, data) for the requested range of | |
1661 | revisions. Offset is the integer offset from the beginning of the |
|
1729 | revisions. Offset is the integer offset from the beginning of the | |
1662 | revlog and data is a str or buffer of the raw byte data. |
|
1730 | revlog and data is a str or buffer of the raw byte data. | |
1663 |
|
1731 | |||
1664 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` |
|
1732 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` | |
1665 | to determine where each revision's data begins and ends. |
|
1733 | to determine where each revision's data begins and ends. | |
1666 | """ |
|
1734 | """ | |
1667 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons |
|
1735 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons | |
1668 | # (functions are expensive). |
|
1736 | # (functions are expensive). | |
1669 | index = self.index |
|
1737 | index = self.index | |
1670 | istart = index[startrev] |
|
1738 | istart = index[startrev] | |
1671 | start = int(istart[0] >> 16) |
|
1739 | start = int(istart[0] >> 16) | |
1672 | if startrev == endrev: |
|
1740 | if startrev == endrev: | |
1673 | end = start + istart[1] |
|
1741 | end = start + istart[1] | |
1674 | else: |
|
1742 | else: | |
1675 | iend = index[endrev] |
|
1743 | iend = index[endrev] | |
1676 | end = int(iend[0] >> 16) + iend[1] |
|
1744 | end = int(iend[0] >> 16) + iend[1] | |
1677 |
|
1745 | |||
1678 | if self._inline: |
|
1746 | if self._inline: | |
1679 | start += (startrev + 1) * self.index.entry_size |
|
1747 | start += (startrev + 1) * self.index.entry_size | |
1680 | end += (endrev + 1) * self.index.entry_size |
|
1748 | end += (endrev + 1) * self.index.entry_size | |
1681 | length = end - start |
|
1749 | length = end - start | |
1682 |
|
1750 | |||
1683 | return start, self._segmentfile.read_chunk(start, length, df) |
|
1751 | return start, self._segmentfile.read_chunk(start, length, df) | |
1684 |
|
1752 | |||
1685 | def _chunk(self, rev, df=None): |
|
1753 | def _chunk(self, rev, df=None): | |
1686 | """Obtain a single decompressed chunk for a revision. |
|
1754 | """Obtain a single decompressed chunk for a revision. | |
1687 |
|
1755 | |||
1688 | Accepts an integer revision and an optional already-open file handle |
|
1756 | Accepts an integer revision and an optional already-open file handle | |
1689 | to be used for reading. If used, the seek position of the file will not |
|
1757 | to be used for reading. If used, the seek position of the file will not | |
1690 | be preserved. |
|
1758 | be preserved. | |
1691 |
|
1759 | |||
1692 | Returns a str holding uncompressed data for the requested revision. |
|
1760 | Returns a str holding uncompressed data for the requested revision. | |
1693 | """ |
|
1761 | """ | |
1694 | compression_mode = self.index[rev][10] |
|
1762 | compression_mode = self.index[rev][10] | |
1695 | data = self._getsegmentforrevs(rev, rev, df=df)[1] |
|
1763 | data = self._getsegmentforrevs(rev, rev, df=df)[1] | |
1696 | if compression_mode == COMP_MODE_PLAIN: |
|
1764 | if compression_mode == COMP_MODE_PLAIN: | |
1697 | return data |
|
1765 | return data | |
1698 | elif compression_mode == COMP_MODE_DEFAULT: |
|
1766 | elif compression_mode == COMP_MODE_DEFAULT: | |
1699 | return self._decompressor(data) |
|
1767 | return self._decompressor(data) | |
1700 | elif compression_mode == COMP_MODE_INLINE: |
|
1768 | elif compression_mode == COMP_MODE_INLINE: | |
1701 | return self.decompress(data) |
|
1769 | return self.decompress(data) | |
1702 | else: |
|
1770 | else: | |
1703 | msg = b'unknown compression mode %d' |
|
1771 | msg = b'unknown compression mode %d' | |
1704 | msg %= compression_mode |
|
1772 | msg %= compression_mode | |
1705 | raise error.RevlogError(msg) |
|
1773 | raise error.RevlogError(msg) | |
1706 |
|
1774 | |||
1707 | def _chunks(self, revs, df=None, targetsize=None): |
|
1775 | def _chunks(self, revs, df=None, targetsize=None): | |
1708 | """Obtain decompressed chunks for the specified revisions. |
|
1776 | """Obtain decompressed chunks for the specified revisions. | |
1709 |
|
1777 | |||
1710 | Accepts an iterable of numeric revisions that are assumed to be in |
|
1778 | Accepts an iterable of numeric revisions that are assumed to be in | |
1711 | ascending order. Also accepts an optional already-open file handle |
|
1779 | ascending order. Also accepts an optional already-open file handle | |
1712 | to be used for reading. If used, the seek position of the file will |
|
1780 | to be used for reading. If used, the seek position of the file will | |
1713 | not be preserved. |
|
1781 | not be preserved. | |
1714 |
|
1782 | |||
1715 | This function is similar to calling ``self._chunk()`` multiple times, |
|
1783 | This function is similar to calling ``self._chunk()`` multiple times, | |
1716 | but is faster. |
|
1784 | but is faster. | |
1717 |
|
1785 | |||
1718 | Returns a list with decompressed data for each requested revision. |
|
1786 | Returns a list with decompressed data for each requested revision. | |
1719 | """ |
|
1787 | """ | |
1720 | if not revs: |
|
1788 | if not revs: | |
1721 | return [] |
|
1789 | return [] | |
1722 | start = self.start |
|
1790 | start = self.start | |
1723 | length = self.length |
|
1791 | length = self.length | |
1724 | inline = self._inline |
|
1792 | inline = self._inline | |
1725 | iosize = self.index.entry_size |
|
1793 | iosize = self.index.entry_size | |
1726 | buffer = util.buffer |
|
1794 | buffer = util.buffer | |
1727 |
|
1795 | |||
1728 | l = [] |
|
1796 | l = [] | |
1729 | ladd = l.append |
|
1797 | ladd = l.append | |
1730 |
|
1798 | |||
1731 | if not self._withsparseread: |
|
1799 | if not self._withsparseread: | |
1732 | slicedchunks = (revs,) |
|
1800 | slicedchunks = (revs,) | |
1733 | else: |
|
1801 | else: | |
1734 | slicedchunks = deltautil.slicechunk( |
|
1802 | slicedchunks = deltautil.slicechunk( | |
1735 | self, revs, targetsize=targetsize |
|
1803 | self, revs, targetsize=targetsize | |
1736 | ) |
|
1804 | ) | |
1737 |
|
1805 | |||
1738 | for revschunk in slicedchunks: |
|
1806 | for revschunk in slicedchunks: | |
1739 | firstrev = revschunk[0] |
|
1807 | firstrev = revschunk[0] | |
1740 | # Skip trailing revisions with empty diff |
|
1808 | # Skip trailing revisions with empty diff | |
1741 | for lastrev in revschunk[::-1]: |
|
1809 | for lastrev in revschunk[::-1]: | |
1742 | if length(lastrev) != 0: |
|
1810 | if length(lastrev) != 0: | |
1743 | break |
|
1811 | break | |
1744 |
|
1812 | |||
1745 | try: |
|
1813 | try: | |
1746 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) |
|
1814 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) | |
1747 | except OverflowError: |
|
1815 | except OverflowError: | |
1748 | # issue4215 - we can't cache a run of chunks greater than |
|
1816 | # issue4215 - we can't cache a run of chunks greater than | |
1749 | # 2G on Windows |
|
1817 | # 2G on Windows | |
1750 | return [self._chunk(rev, df=df) for rev in revschunk] |
|
1818 | return [self._chunk(rev, df=df) for rev in revschunk] | |
1751 |
|
1819 | |||
1752 | decomp = self.decompress |
|
1820 | decomp = self.decompress | |
1753 | # self._decompressor might be None, but will not be used in that case |
|
1821 | # self._decompressor might be None, but will not be used in that case | |
1754 | def_decomp = self._decompressor |
|
1822 | def_decomp = self._decompressor | |
1755 | for rev in revschunk: |
|
1823 | for rev in revschunk: | |
1756 | chunkstart = start(rev) |
|
1824 | chunkstart = start(rev) | |
1757 | if inline: |
|
1825 | if inline: | |
1758 | chunkstart += (rev + 1) * iosize |
|
1826 | chunkstart += (rev + 1) * iosize | |
1759 | chunklength = length(rev) |
|
1827 | chunklength = length(rev) | |
1760 | comp_mode = self.index[rev][10] |
|
1828 | comp_mode = self.index[rev][10] | |
1761 | c = buffer(data, chunkstart - offset, chunklength) |
|
1829 | c = buffer(data, chunkstart - offset, chunklength) | |
1762 | if comp_mode == COMP_MODE_PLAIN: |
|
1830 | if comp_mode == COMP_MODE_PLAIN: | |
1763 | ladd(c) |
|
1831 | ladd(c) | |
1764 | elif comp_mode == COMP_MODE_INLINE: |
|
1832 | elif comp_mode == COMP_MODE_INLINE: | |
1765 | ladd(decomp(c)) |
|
1833 | ladd(decomp(c)) | |
1766 | elif comp_mode == COMP_MODE_DEFAULT: |
|
1834 | elif comp_mode == COMP_MODE_DEFAULT: | |
1767 | ladd(def_decomp(c)) |
|
1835 | ladd(def_decomp(c)) | |
1768 | else: |
|
1836 | else: | |
1769 | msg = b'unknown compression mode %d' |
|
1837 | msg = b'unknown compression mode %d' | |
1770 | msg %= comp_mode |
|
1838 | msg %= comp_mode | |
1771 | raise error.RevlogError(msg) |
|
1839 | raise error.RevlogError(msg) | |
1772 |
|
1840 | |||
1773 | return l |
|
1841 | return l | |
1774 |
|
1842 | |||
1775 | def deltaparent(self, rev): |
|
1843 | def deltaparent(self, rev): | |
1776 | """return deltaparent of the given revision""" |
|
1844 | """return deltaparent of the given revision""" | |
1777 | base = self.index[rev][3] |
|
1845 | base = self.index[rev][3] | |
1778 | if base == rev: |
|
1846 | if base == rev: | |
1779 | return nullrev |
|
1847 | return nullrev | |
1780 | elif self._generaldelta: |
|
1848 | elif self._generaldelta: | |
1781 | return base |
|
1849 | return base | |
1782 | else: |
|
1850 | else: | |
1783 | return rev - 1 |
|
1851 | return rev - 1 | |
1784 |
|
1852 | |||
1785 | def issnapshot(self, rev): |
|
1853 | def issnapshot(self, rev): | |
1786 | """tells whether rev is a snapshot""" |
|
1854 | """tells whether rev is a snapshot""" | |
1787 | if not self._sparserevlog: |
|
1855 | if not self._sparserevlog: | |
1788 | return self.deltaparent(rev) == nullrev |
|
1856 | return self.deltaparent(rev) == nullrev | |
1789 | elif util.safehasattr(self.index, 'issnapshot'): |
|
1857 | elif util.safehasattr(self.index, 'issnapshot'): | |
1790 | # directly assign the method to cache the testing and access |
|
1858 | # directly assign the method to cache the testing and access | |
1791 | self.issnapshot = self.index.issnapshot |
|
1859 | self.issnapshot = self.index.issnapshot | |
1792 | return self.issnapshot(rev) |
|
1860 | return self.issnapshot(rev) | |
1793 | if rev == nullrev: |
|
1861 | if rev == nullrev: | |
1794 | return True |
|
1862 | return True | |
1795 | entry = self.index[rev] |
|
1863 | entry = self.index[rev] | |
1796 | base = entry[3] |
|
1864 | base = entry[3] | |
1797 | if base == rev: |
|
1865 | if base == rev: | |
1798 | return True |
|
1866 | return True | |
1799 | if base == nullrev: |
|
1867 | if base == nullrev: | |
1800 | return True |
|
1868 | return True | |
1801 | p1 = entry[5] |
|
1869 | p1 = entry[5] | |
1802 | while self.length(p1) == 0: |
|
1870 | while self.length(p1) == 0: | |
1803 | b = self.deltaparent(p1) |
|
1871 | b = self.deltaparent(p1) | |
1804 | if b == p1: |
|
1872 | if b == p1: | |
1805 | break |
|
1873 | break | |
1806 | p1 = b |
|
1874 | p1 = b | |
1807 | p2 = entry[6] |
|
1875 | p2 = entry[6] | |
1808 | while self.length(p2) == 0: |
|
1876 | while self.length(p2) == 0: | |
1809 | b = self.deltaparent(p2) |
|
1877 | b = self.deltaparent(p2) | |
1810 | if b == p2: |
|
1878 | if b == p2: | |
1811 | break |
|
1879 | break | |
1812 | p2 = b |
|
1880 | p2 = b | |
1813 | if base == p1 or base == p2: |
|
1881 | if base == p1 or base == p2: | |
1814 | return False |
|
1882 | return False | |
1815 | return self.issnapshot(base) |
|
1883 | return self.issnapshot(base) | |
1816 |
|
1884 | |||
1817 | def snapshotdepth(self, rev): |
|
1885 | def snapshotdepth(self, rev): | |
1818 | """number of snapshot in the chain before this one""" |
|
1886 | """number of snapshot in the chain before this one""" | |
1819 | if not self.issnapshot(rev): |
|
1887 | if not self.issnapshot(rev): | |
1820 | raise error.ProgrammingError(b'revision %d not a snapshot') |
|
1888 | raise error.ProgrammingError(b'revision %d not a snapshot') | |
1821 | return len(self._deltachain(rev)[0]) - 1 |
|
1889 | return len(self._deltachain(rev)[0]) - 1 | |
1822 |
|
1890 | |||
1823 | def revdiff(self, rev1, rev2): |
|
1891 | def revdiff(self, rev1, rev2): | |
1824 | """return or calculate a delta between two revisions |
|
1892 | """return or calculate a delta between two revisions | |
1825 |
|
1893 | |||
1826 | The delta calculated is in binary form and is intended to be written to |
|
1894 | The delta calculated is in binary form and is intended to be written to | |
1827 | revlog data directly. So this function needs raw revision data. |
|
1895 | revlog data directly. So this function needs raw revision data. | |
1828 | """ |
|
1896 | """ | |
1829 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: |
|
1897 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: | |
1830 | return bytes(self._chunk(rev2)) |
|
1898 | return bytes(self._chunk(rev2)) | |
1831 |
|
1899 | |||
1832 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) |
|
1900 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) | |
1833 |
|
1901 | |||
1834 | def revision(self, nodeorrev, _df=None): |
|
1902 | def revision(self, nodeorrev, _df=None): | |
1835 | """return an uncompressed revision of a given node or revision |
|
1903 | """return an uncompressed revision of a given node or revision | |
1836 | number. |
|
1904 | number. | |
1837 |
|
1905 | |||
1838 | _df - an existing file handle to read from. (internal-only) |
|
1906 | _df - an existing file handle to read from. (internal-only) | |
1839 | """ |
|
1907 | """ | |
1840 | return self._revisiondata(nodeorrev, _df) |
|
1908 | return self._revisiondata(nodeorrev, _df) | |
1841 |
|
1909 | |||
1842 | def sidedata(self, nodeorrev, _df=None): |
|
1910 | def sidedata(self, nodeorrev, _df=None): | |
1843 | """a map of extra data related to the changeset but not part of the hash |
|
1911 | """a map of extra data related to the changeset but not part of the hash | |
1844 |
|
1912 | |||
1845 | This function currently return a dictionary. However, more advanced |
|
1913 | This function currently return a dictionary. However, more advanced | |
1846 | mapping object will likely be used in the future for a more |
|
1914 | mapping object will likely be used in the future for a more | |
1847 | efficient/lazy code. |
|
1915 | efficient/lazy code. | |
1848 | """ |
|
1916 | """ | |
1849 | # deal with <nodeorrev> argument type |
|
1917 | # deal with <nodeorrev> argument type | |
1850 | if isinstance(nodeorrev, int): |
|
1918 | if isinstance(nodeorrev, int): | |
1851 | rev = nodeorrev |
|
1919 | rev = nodeorrev | |
1852 | else: |
|
1920 | else: | |
1853 | rev = self.rev(nodeorrev) |
|
1921 | rev = self.rev(nodeorrev) | |
1854 | return self._sidedata(rev) |
|
1922 | return self._sidedata(rev) | |
1855 |
|
1923 | |||
1856 | def _revisiondata(self, nodeorrev, _df=None, raw=False): |
|
1924 | def _revisiondata(self, nodeorrev, _df=None, raw=False): | |
1857 | # deal with <nodeorrev> argument type |
|
1925 | # deal with <nodeorrev> argument type | |
1858 | if isinstance(nodeorrev, int): |
|
1926 | if isinstance(nodeorrev, int): | |
1859 | rev = nodeorrev |
|
1927 | rev = nodeorrev | |
1860 | node = self.node(rev) |
|
1928 | node = self.node(rev) | |
1861 | else: |
|
1929 | else: | |
1862 | node = nodeorrev |
|
1930 | node = nodeorrev | |
1863 | rev = None |
|
1931 | rev = None | |
1864 |
|
1932 | |||
1865 | # fast path the special `nullid` rev |
|
1933 | # fast path the special `nullid` rev | |
1866 | if node == self.nullid: |
|
1934 | if node == self.nullid: | |
1867 | return b"" |
|
1935 | return b"" | |
1868 |
|
1936 | |||
1869 | # ``rawtext`` is the text as stored inside the revlog. Might be the |
|
1937 | # ``rawtext`` is the text as stored inside the revlog. Might be the | |
1870 | # revision or might need to be processed to retrieve the revision. |
|
1938 | # revision or might need to be processed to retrieve the revision. | |
1871 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) |
|
1939 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) | |
1872 |
|
1940 | |||
1873 | if raw and validated: |
|
1941 | if raw and validated: | |
1874 | # if we don't want to process the raw text and that raw |
|
1942 | # if we don't want to process the raw text and that raw | |
1875 | # text is cached, we can exit early. |
|
1943 | # text is cached, we can exit early. | |
1876 | return rawtext |
|
1944 | return rawtext | |
1877 | if rev is None: |
|
1945 | if rev is None: | |
1878 | rev = self.rev(node) |
|
1946 | rev = self.rev(node) | |
1879 | # the revlog's flag for this revision |
|
1947 | # the revlog's flag for this revision | |
1880 | # (usually alter its state or content) |
|
1948 | # (usually alter its state or content) | |
1881 | flags = self.flags(rev) |
|
1949 | flags = self.flags(rev) | |
1882 |
|
1950 | |||
1883 | if validated and flags == REVIDX_DEFAULT_FLAGS: |
|
1951 | if validated and flags == REVIDX_DEFAULT_FLAGS: | |
1884 | # no extra flags set, no flag processor runs, text = rawtext |
|
1952 | # no extra flags set, no flag processor runs, text = rawtext | |
1885 | return rawtext |
|
1953 | return rawtext | |
1886 |
|
1954 | |||
1887 | if raw: |
|
1955 | if raw: | |
1888 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
1956 | validatehash = flagutil.processflagsraw(self, rawtext, flags) | |
1889 | text = rawtext |
|
1957 | text = rawtext | |
1890 | else: |
|
1958 | else: | |
1891 | r = flagutil.processflagsread(self, rawtext, flags) |
|
1959 | r = flagutil.processflagsread(self, rawtext, flags) | |
1892 | text, validatehash = r |
|
1960 | text, validatehash = r | |
1893 | if validatehash: |
|
1961 | if validatehash: | |
1894 | self.checkhash(text, node, rev=rev) |
|
1962 | self.checkhash(text, node, rev=rev) | |
1895 | if not validated: |
|
1963 | if not validated: | |
1896 | self._revisioncache = (node, rev, rawtext) |
|
1964 | self._revisioncache = (node, rev, rawtext) | |
1897 |
|
1965 | |||
1898 | return text |
|
1966 | return text | |
1899 |
|
1967 | |||
1900 | def _rawtext(self, node, rev, _df=None): |
|
1968 | def _rawtext(self, node, rev, _df=None): | |
1901 | """return the possibly unvalidated rawtext for a revision |
|
1969 | """return the possibly unvalidated rawtext for a revision | |
1902 |
|
1970 | |||
1903 | returns (rev, rawtext, validated) |
|
1971 | returns (rev, rawtext, validated) | |
1904 | """ |
|
1972 | """ | |
1905 |
|
1973 | |||
1906 | # revision in the cache (could be useful to apply delta) |
|
1974 | # revision in the cache (could be useful to apply delta) | |
1907 | cachedrev = None |
|
1975 | cachedrev = None | |
1908 | # An intermediate text to apply deltas to |
|
1976 | # An intermediate text to apply deltas to | |
1909 | basetext = None |
|
1977 | basetext = None | |
1910 |
|
1978 | |||
1911 | # Check if we have the entry in cache |
|
1979 | # Check if we have the entry in cache | |
1912 | # The cache entry looks like (node, rev, rawtext) |
|
1980 | # The cache entry looks like (node, rev, rawtext) | |
1913 | if self._revisioncache: |
|
1981 | if self._revisioncache: | |
1914 | if self._revisioncache[0] == node: |
|
1982 | if self._revisioncache[0] == node: | |
1915 | return (rev, self._revisioncache[2], True) |
|
1983 | return (rev, self._revisioncache[2], True) | |
1916 | cachedrev = self._revisioncache[1] |
|
1984 | cachedrev = self._revisioncache[1] | |
1917 |
|
1985 | |||
1918 | if rev is None: |
|
1986 | if rev is None: | |
1919 | rev = self.rev(node) |
|
1987 | rev = self.rev(node) | |
1920 |
|
1988 | |||
1921 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) |
|
1989 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) | |
1922 | if stopped: |
|
1990 | if stopped: | |
1923 | basetext = self._revisioncache[2] |
|
1991 | basetext = self._revisioncache[2] | |
1924 |
|
1992 | |||
1925 | # drop cache to save memory, the caller is expected to |
|
1993 | # drop cache to save memory, the caller is expected to | |
1926 | # update self._revisioncache after validating the text |
|
1994 | # update self._revisioncache after validating the text | |
1927 | self._revisioncache = None |
|
1995 | self._revisioncache = None | |
1928 |
|
1996 | |||
1929 | targetsize = None |
|
1997 | targetsize = None | |
1930 | rawsize = self.index[rev][2] |
|
1998 | rawsize = self.index[rev][2] | |
1931 | if 0 <= rawsize: |
|
1999 | if 0 <= rawsize: | |
1932 | targetsize = 4 * rawsize |
|
2000 | targetsize = 4 * rawsize | |
1933 |
|
2001 | |||
1934 | bins = self._chunks(chain, df=_df, targetsize=targetsize) |
|
2002 | bins = self._chunks(chain, df=_df, targetsize=targetsize) | |
1935 | if basetext is None: |
|
2003 | if basetext is None: | |
1936 | basetext = bytes(bins[0]) |
|
2004 | basetext = bytes(bins[0]) | |
1937 | bins = bins[1:] |
|
2005 | bins = bins[1:] | |
1938 |
|
2006 | |||
1939 | rawtext = mdiff.patches(basetext, bins) |
|
2007 | rawtext = mdiff.patches(basetext, bins) | |
1940 | del basetext # let us have a chance to free memory early |
|
2008 | del basetext # let us have a chance to free memory early | |
1941 | return (rev, rawtext, False) |
|
2009 | return (rev, rawtext, False) | |
1942 |
|
2010 | |||
1943 | def _sidedata(self, rev): |
|
2011 | def _sidedata(self, rev): | |
1944 | """Return the sidedata for a given revision number.""" |
|
2012 | """Return the sidedata for a given revision number.""" | |
1945 | index_entry = self.index[rev] |
|
2013 | index_entry = self.index[rev] | |
1946 | sidedata_offset = index_entry[8] |
|
2014 | sidedata_offset = index_entry[8] | |
1947 | sidedata_size = index_entry[9] |
|
2015 | sidedata_size = index_entry[9] | |
1948 |
|
2016 | |||
1949 | if self._inline: |
|
2017 | if self._inline: | |
1950 | sidedata_offset += self.index.entry_size * (1 + rev) |
|
2018 | sidedata_offset += self.index.entry_size * (1 + rev) | |
1951 | if sidedata_size == 0: |
|
2019 | if sidedata_size == 0: | |
1952 | return {} |
|
2020 | return {} | |
1953 |
|
2021 | |||
1954 | if self._docket.sidedata_end < sidedata_offset + sidedata_size: |
|
2022 | if self._docket.sidedata_end < sidedata_offset + sidedata_size: | |
1955 | filename = self._sidedatafile |
|
2023 | filename = self._sidedatafile | |
1956 | end = self._docket.sidedata_end |
|
2024 | end = self._docket.sidedata_end | |
1957 | offset = sidedata_offset |
|
2025 | offset = sidedata_offset | |
1958 | length = sidedata_size |
|
2026 | length = sidedata_size | |
1959 | m = FILE_TOO_SHORT_MSG % (filename, length, offset, end) |
|
2027 | m = FILE_TOO_SHORT_MSG % (filename, length, offset, end) | |
1960 | raise error.RevlogError(m) |
|
2028 | raise error.RevlogError(m) | |
1961 |
|
2029 | |||
1962 | comp_segment = self._segmentfile_sidedata.read_chunk( |
|
2030 | comp_segment = self._segmentfile_sidedata.read_chunk( | |
1963 | sidedata_offset, sidedata_size |
|
2031 | sidedata_offset, sidedata_size | |
1964 | ) |
|
2032 | ) | |
1965 |
|
2033 | |||
1966 | comp = self.index[rev][11] |
|
2034 | comp = self.index[rev][11] | |
1967 | if comp == COMP_MODE_PLAIN: |
|
2035 | if comp == COMP_MODE_PLAIN: | |
1968 | segment = comp_segment |
|
2036 | segment = comp_segment | |
1969 | elif comp == COMP_MODE_DEFAULT: |
|
2037 | elif comp == COMP_MODE_DEFAULT: | |
1970 | segment = self._decompressor(comp_segment) |
|
2038 | segment = self._decompressor(comp_segment) | |
1971 | elif comp == COMP_MODE_INLINE: |
|
2039 | elif comp == COMP_MODE_INLINE: | |
1972 | segment = self.decompress(comp_segment) |
|
2040 | segment = self.decompress(comp_segment) | |
1973 | else: |
|
2041 | else: | |
1974 | msg = b'unknown compression mode %d' |
|
2042 | msg = b'unknown compression mode %d' | |
1975 | msg %= comp |
|
2043 | msg %= comp | |
1976 | raise error.RevlogError(msg) |
|
2044 | raise error.RevlogError(msg) | |
1977 |
|
2045 | |||
1978 | sidedata = sidedatautil.deserialize_sidedata(segment) |
|
2046 | sidedata = sidedatautil.deserialize_sidedata(segment) | |
1979 | return sidedata |
|
2047 | return sidedata | |
1980 |
|
2048 | |||
1981 | def rawdata(self, nodeorrev, _df=None): |
|
2049 | def rawdata(self, nodeorrev, _df=None): | |
1982 | """return an uncompressed raw data of a given node or revision number. |
|
2050 | """return an uncompressed raw data of a given node or revision number. | |
1983 |
|
2051 | |||
1984 | _df - an existing file handle to read from. (internal-only) |
|
2052 | _df - an existing file handle to read from. (internal-only) | |
1985 | """ |
|
2053 | """ | |
1986 | return self._revisiondata(nodeorrev, _df, raw=True) |
|
2054 | return self._revisiondata(nodeorrev, _df, raw=True) | |
1987 |
|
2055 | |||
1988 | def hash(self, text, p1, p2): |
|
2056 | def hash(self, text, p1, p2): | |
1989 | """Compute a node hash. |
|
2057 | """Compute a node hash. | |
1990 |
|
2058 | |||
1991 | Available as a function so that subclasses can replace the hash |
|
2059 | Available as a function so that subclasses can replace the hash | |
1992 | as needed. |
|
2060 | as needed. | |
1993 | """ |
|
2061 | """ | |
1994 | return storageutil.hashrevisionsha1(text, p1, p2) |
|
2062 | return storageutil.hashrevisionsha1(text, p1, p2) | |
1995 |
|
2063 | |||
1996 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
2064 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
1997 | """Check node hash integrity. |
|
2065 | """Check node hash integrity. | |
1998 |
|
2066 | |||
1999 | Available as a function so that subclasses can extend hash mismatch |
|
2067 | Available as a function so that subclasses can extend hash mismatch | |
2000 | behaviors as needed. |
|
2068 | behaviors as needed. | |
2001 | """ |
|
2069 | """ | |
2002 | try: |
|
2070 | try: | |
2003 | if p1 is None and p2 is None: |
|
2071 | if p1 is None and p2 is None: | |
2004 | p1, p2 = self.parents(node) |
|
2072 | p1, p2 = self.parents(node) | |
2005 | if node != self.hash(text, p1, p2): |
|
2073 | if node != self.hash(text, p1, p2): | |
2006 | # Clear the revision cache on hash failure. The revision cache |
|
2074 | # Clear the revision cache on hash failure. The revision cache | |
2007 | # only stores the raw revision and clearing the cache does have |
|
2075 | # only stores the raw revision and clearing the cache does have | |
2008 | # the side-effect that we won't have a cache hit when the raw |
|
2076 | # the side-effect that we won't have a cache hit when the raw | |
2009 | # revision data is accessed. But this case should be rare and |
|
2077 | # revision data is accessed. But this case should be rare and | |
2010 | # it is extra work to teach the cache about the hash |
|
2078 | # it is extra work to teach the cache about the hash | |
2011 | # verification state. |
|
2079 | # verification state. | |
2012 | if self._revisioncache and self._revisioncache[0] == node: |
|
2080 | if self._revisioncache and self._revisioncache[0] == node: | |
2013 | self._revisioncache = None |
|
2081 | self._revisioncache = None | |
2014 |
|
2082 | |||
2015 | revornode = rev |
|
2083 | revornode = rev | |
2016 | if revornode is None: |
|
2084 | if revornode is None: | |
2017 | revornode = templatefilters.short(hex(node)) |
|
2085 | revornode = templatefilters.short(hex(node)) | |
2018 | raise error.RevlogError( |
|
2086 | raise error.RevlogError( | |
2019 | _(b"integrity check failed on %s:%s") |
|
2087 | _(b"integrity check failed on %s:%s") | |
2020 | % (self.display_id, pycompat.bytestr(revornode)) |
|
2088 | % (self.display_id, pycompat.bytestr(revornode)) | |
2021 | ) |
|
2089 | ) | |
2022 | except error.RevlogError: |
|
2090 | except error.RevlogError: | |
2023 | if self._censorable and storageutil.iscensoredtext(text): |
|
2091 | if self._censorable and storageutil.iscensoredtext(text): | |
2024 | raise error.CensoredNodeError(self.display_id, node, text) |
|
2092 | raise error.CensoredNodeError(self.display_id, node, text) | |
2025 | raise |
|
2093 | raise | |
2026 |
|
2094 | |||
2027 | def _enforceinlinesize(self, tr, side_write=True): |
|
2095 | def _enforceinlinesize(self, tr, side_write=True): | |
2028 | """Check if the revlog is too big for inline and convert if so. |
|
2096 | """Check if the revlog is too big for inline and convert if so. | |
2029 |
|
2097 | |||
2030 | This should be called after revisions are added to the revlog. If the |
|
2098 | This should be called after revisions are added to the revlog. If the | |
2031 | revlog has grown too large to be an inline revlog, it will convert it |
|
2099 | revlog has grown too large to be an inline revlog, it will convert it | |
2032 | to use multiple index and data files. |
|
2100 | to use multiple index and data files. | |
2033 | """ |
|
2101 | """ | |
2034 | tiprev = len(self) - 1 |
|
2102 | tiprev = len(self) - 1 | |
2035 | total_size = self.start(tiprev) + self.length(tiprev) |
|
2103 | total_size = self.start(tiprev) + self.length(tiprev) | |
2036 | if not self._inline or total_size < _maxinline: |
|
2104 | if not self._inline or total_size < _maxinline: | |
2037 | return |
|
2105 | return | |
2038 |
|
2106 | |||
2039 | troffset = tr.findoffset(self._indexfile) |
|
2107 | troffset = tr.findoffset(self._indexfile) | |
2040 | if troffset is None: |
|
2108 | if troffset is None: | |
2041 | raise error.RevlogError( |
|
2109 | raise error.RevlogError( | |
2042 | _(b"%s not found in the transaction") % self._indexfile |
|
2110 | _(b"%s not found in the transaction") % self._indexfile | |
2043 | ) |
|
2111 | ) | |
2044 | if troffset: |
|
2112 | if troffset: | |
2045 | tr.addbackup(self._indexfile, for_offset=True) |
|
2113 | tr.addbackup(self._indexfile, for_offset=True) | |
2046 | tr.add(self._datafile, 0) |
|
2114 | tr.add(self._datafile, 0) | |
2047 |
|
2115 | |||
2048 | existing_handles = False |
|
2116 | existing_handles = False | |
2049 | if self._writinghandles is not None: |
|
2117 | if self._writinghandles is not None: | |
2050 | existing_handles = True |
|
2118 | existing_handles = True | |
2051 | fp = self._writinghandles[0] |
|
2119 | fp = self._writinghandles[0] | |
2052 | fp.flush() |
|
2120 | fp.flush() | |
2053 | fp.close() |
|
2121 | fp.close() | |
2054 | # We can't use the cached file handle after close(). So prevent |
|
2122 | # We can't use the cached file handle after close(). So prevent | |
2055 | # its usage. |
|
2123 | # its usage. | |
2056 | self._writinghandles = None |
|
2124 | self._writinghandles = None | |
2057 | self._segmentfile.writing_handle = None |
|
2125 | self._segmentfile.writing_handle = None | |
2058 | # No need to deal with sidedata writing handle as it is only |
|
2126 | # No need to deal with sidedata writing handle as it is only | |
2059 | # relevant with revlog-v2 which is never inline, not reaching |
|
2127 | # relevant with revlog-v2 which is never inline, not reaching | |
2060 | # this code |
|
2128 | # this code | |
2061 | if side_write: |
|
2129 | if side_write: | |
2062 | old_index_file_path = self._indexfile |
|
2130 | old_index_file_path = self._indexfile | |
2063 | new_index_file_path = self._indexfile + b'.s' |
|
2131 | new_index_file_path = self._indexfile + b'.s' | |
2064 | opener = self.opener |
|
2132 | opener = self.opener | |
2065 | weak_self = weakref.ref(self) |
|
2133 | weak_self = weakref.ref(self) | |
2066 |
|
2134 | |||
2067 | # the "split" index replace the real index when the transaction is finalized |
|
2135 | # the "split" index replace the real index when the transaction is finalized | |
2068 | def finalize_callback(tr): |
|
2136 | def finalize_callback(tr): | |
2069 | opener.rename( |
|
2137 | opener.rename( | |
2070 | new_index_file_path, |
|
2138 | new_index_file_path, | |
2071 | old_index_file_path, |
|
2139 | old_index_file_path, | |
2072 | checkambig=True, |
|
2140 | checkambig=True, | |
2073 | ) |
|
2141 | ) | |
2074 | maybe_self = weak_self() |
|
2142 | maybe_self = weak_self() | |
2075 | if maybe_self is not None: |
|
2143 | if maybe_self is not None: | |
2076 | maybe_self._indexfile = old_index_file_path |
|
2144 | maybe_self._indexfile = old_index_file_path | |
2077 |
|
2145 | |||
2078 | def abort_callback(tr): |
|
2146 | def abort_callback(tr): | |
2079 | maybe_self = weak_self() |
|
2147 | maybe_self = weak_self() | |
2080 | if maybe_self is not None: |
|
2148 | if maybe_self is not None: | |
2081 | maybe_self._indexfile = old_index_file_path |
|
2149 | maybe_self._indexfile = old_index_file_path | |
2082 |
|
2150 | |||
2083 | tr.registertmp(new_index_file_path) |
|
2151 | tr.registertmp(new_index_file_path) | |
2084 | if self.target[1] is not None: |
|
2152 | if self.target[1] is not None: | |
2085 | callback_id = b'000-revlog-split-%d-%s' % self.target |
|
2153 | callback_id = b'000-revlog-split-%d-%s' % self.target | |
2086 | else: |
|
2154 | else: | |
2087 | callback_id = b'000-revlog-split-%d' % self.target[0] |
|
2155 | callback_id = b'000-revlog-split-%d' % self.target[0] | |
2088 | tr.addfinalize(callback_id, finalize_callback) |
|
2156 | tr.addfinalize(callback_id, finalize_callback) | |
2089 | tr.addabort(callback_id, abort_callback) |
|
2157 | tr.addabort(callback_id, abort_callback) | |
2090 |
|
2158 | |||
2091 | new_dfh = self._datafp(b'w+') |
|
2159 | new_dfh = self._datafp(b'w+') | |
2092 | new_dfh.truncate(0) # drop any potentially existing data |
|
2160 | new_dfh.truncate(0) # drop any potentially existing data | |
2093 | try: |
|
2161 | try: | |
2094 | with self._indexfp() as read_ifh: |
|
2162 | with self._indexfp() as read_ifh: | |
2095 | for r in self: |
|
2163 | for r in self: | |
2096 | new_dfh.write(self._getsegmentforrevs(r, r, df=read_ifh)[1]) |
|
2164 | new_dfh.write(self._getsegmentforrevs(r, r, df=read_ifh)[1]) | |
2097 | new_dfh.flush() |
|
2165 | new_dfh.flush() | |
2098 |
|
2166 | |||
2099 | if side_write: |
|
2167 | if side_write: | |
2100 | self._indexfile = new_index_file_path |
|
2168 | self._indexfile = new_index_file_path | |
2101 | with self.__index_new_fp() as fp: |
|
2169 | with self.__index_new_fp() as fp: | |
2102 | self._format_flags &= ~FLAG_INLINE_DATA |
|
2170 | self._format_flags &= ~FLAG_INLINE_DATA | |
2103 | self._inline = False |
|
2171 | self._inline = False | |
2104 | for i in self: |
|
2172 | for i in self: | |
2105 | e = self.index.entry_binary(i) |
|
2173 | e = self.index.entry_binary(i) | |
2106 | if i == 0 and self._docket is None: |
|
2174 | if i == 0 and self._docket is None: | |
2107 | header = self._format_flags | self._format_version |
|
2175 | header = self._format_flags | self._format_version | |
2108 | header = self.index.pack_header(header) |
|
2176 | header = self.index.pack_header(header) | |
2109 | e = header + e |
|
2177 | e = header + e | |
2110 | fp.write(e) |
|
2178 | fp.write(e) | |
2111 | if self._docket is not None: |
|
2179 | if self._docket is not None: | |
2112 | self._docket.index_end = fp.tell() |
|
2180 | self._docket.index_end = fp.tell() | |
2113 |
|
2181 | |||
2114 | # If we don't use side-write, the temp file replace the real |
|
2182 | # If we don't use side-write, the temp file replace the real | |
2115 | # index when we exit the context manager |
|
2183 | # index when we exit the context manager | |
2116 |
|
2184 | |||
2117 | nodemaputil.setup_persistent_nodemap(tr, self) |
|
2185 | nodemaputil.setup_persistent_nodemap(tr, self) | |
2118 | self._segmentfile = randomaccessfile.randomaccessfile( |
|
2186 | self._segmentfile = randomaccessfile.randomaccessfile( | |
2119 | self.opener, |
|
2187 | self.opener, | |
2120 | self._datafile, |
|
2188 | self._datafile, | |
2121 | self._chunkcachesize, |
|
2189 | self._chunkcachesize, | |
2122 | ) |
|
2190 | ) | |
2123 |
|
2191 | |||
2124 | if existing_handles: |
|
2192 | if existing_handles: | |
2125 | # switched from inline to conventional reopen the index |
|
2193 | # switched from inline to conventional reopen the index | |
2126 | ifh = self.__index_write_fp() |
|
2194 | ifh = self.__index_write_fp() | |
2127 | self._writinghandles = (ifh, new_dfh, None) |
|
2195 | self._writinghandles = (ifh, new_dfh, None) | |
2128 | self._segmentfile.writing_handle = new_dfh |
|
2196 | self._segmentfile.writing_handle = new_dfh | |
2129 | new_dfh = None |
|
2197 | new_dfh = None | |
2130 | # No need to deal with sidedata writing handle as it is only |
|
2198 | # No need to deal with sidedata writing handle as it is only | |
2131 | # relevant with revlog-v2 which is never inline, not reaching |
|
2199 | # relevant with revlog-v2 which is never inline, not reaching | |
2132 | # this code |
|
2200 | # this code | |
2133 | finally: |
|
2201 | finally: | |
2134 | if new_dfh is not None: |
|
2202 | if new_dfh is not None: | |
2135 | new_dfh.close() |
|
2203 | new_dfh.close() | |
2136 |
|
2204 | |||
2137 | def _nodeduplicatecallback(self, transaction, node): |
|
2205 | def _nodeduplicatecallback(self, transaction, node): | |
2138 | """called when trying to add a node already stored.""" |
|
2206 | """called when trying to add a node already stored.""" | |
2139 |
|
2207 | |||
2140 | @contextlib.contextmanager |
|
2208 | @contextlib.contextmanager | |
2141 | def reading(self): |
|
2209 | def reading(self): | |
2142 | """Context manager that keeps data and sidedata files open for reading""" |
|
2210 | """Context manager that keeps data and sidedata files open for reading""" | |
2143 | with self._segmentfile.reading(): |
|
2211 | with self._segmentfile.reading(): | |
2144 | with self._segmentfile_sidedata.reading(): |
|
2212 | with self._segmentfile_sidedata.reading(): | |
2145 | yield |
|
2213 | yield | |
2146 |
|
2214 | |||
2147 | @contextlib.contextmanager |
|
2215 | @contextlib.contextmanager | |
2148 | def _writing(self, transaction): |
|
2216 | def _writing(self, transaction): | |
2149 | if self._trypending: |
|
2217 | if self._trypending: | |
2150 | msg = b'try to write in a `trypending` revlog: %s' |
|
2218 | msg = b'try to write in a `trypending` revlog: %s' | |
2151 | msg %= self.display_id |
|
2219 | msg %= self.display_id | |
2152 | raise error.ProgrammingError(msg) |
|
2220 | raise error.ProgrammingError(msg) | |
2153 | if self._writinghandles is not None: |
|
2221 | if self._writinghandles is not None: | |
2154 | yield |
|
2222 | yield | |
2155 | else: |
|
2223 | else: | |
2156 | ifh = dfh = sdfh = None |
|
2224 | ifh = dfh = sdfh = None | |
2157 | try: |
|
2225 | try: | |
2158 | r = len(self) |
|
2226 | r = len(self) | |
2159 | # opening the data file. |
|
2227 | # opening the data file. | |
2160 | dsize = 0 |
|
2228 | dsize = 0 | |
2161 | if r: |
|
2229 | if r: | |
2162 | dsize = self.end(r - 1) |
|
2230 | dsize = self.end(r - 1) | |
2163 | dfh = None |
|
2231 | dfh = None | |
2164 | if not self._inline: |
|
2232 | if not self._inline: | |
2165 | try: |
|
2233 | try: | |
2166 | dfh = self._datafp(b"r+") |
|
2234 | dfh = self._datafp(b"r+") | |
2167 | if self._docket is None: |
|
2235 | if self._docket is None: | |
2168 | dfh.seek(0, os.SEEK_END) |
|
2236 | dfh.seek(0, os.SEEK_END) | |
2169 | else: |
|
2237 | else: | |
2170 | dfh.seek(self._docket.data_end, os.SEEK_SET) |
|
2238 | dfh.seek(self._docket.data_end, os.SEEK_SET) | |
2171 | except FileNotFoundError: |
|
2239 | except FileNotFoundError: | |
2172 | dfh = self._datafp(b"w+") |
|
2240 | dfh = self._datafp(b"w+") | |
2173 | transaction.add(self._datafile, dsize) |
|
2241 | transaction.add(self._datafile, dsize) | |
2174 | if self._sidedatafile is not None: |
|
2242 | if self._sidedatafile is not None: | |
2175 | # revlog-v2 does not inline, help Pytype |
|
2243 | # revlog-v2 does not inline, help Pytype | |
2176 | assert dfh is not None |
|
2244 | assert dfh is not None | |
2177 | try: |
|
2245 | try: | |
2178 | sdfh = self.opener(self._sidedatafile, mode=b"r+") |
|
2246 | sdfh = self.opener(self._sidedatafile, mode=b"r+") | |
2179 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) |
|
2247 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) | |
2180 | except FileNotFoundError: |
|
2248 | except FileNotFoundError: | |
2181 | sdfh = self.opener(self._sidedatafile, mode=b"w+") |
|
2249 | sdfh = self.opener(self._sidedatafile, mode=b"w+") | |
2182 | transaction.add( |
|
2250 | transaction.add( | |
2183 | self._sidedatafile, self._docket.sidedata_end |
|
2251 | self._sidedatafile, self._docket.sidedata_end | |
2184 | ) |
|
2252 | ) | |
2185 |
|
2253 | |||
2186 | # opening the index file. |
|
2254 | # opening the index file. | |
2187 | isize = r * self.index.entry_size |
|
2255 | isize = r * self.index.entry_size | |
2188 | ifh = self.__index_write_fp() |
|
2256 | ifh = self.__index_write_fp() | |
2189 | if self._inline: |
|
2257 | if self._inline: | |
2190 | transaction.add(self._indexfile, dsize + isize) |
|
2258 | transaction.add(self._indexfile, dsize + isize) | |
2191 | else: |
|
2259 | else: | |
2192 | transaction.add(self._indexfile, isize) |
|
2260 | transaction.add(self._indexfile, isize) | |
2193 | # exposing all file handle for writing. |
|
2261 | # exposing all file handle for writing. | |
2194 | self._writinghandles = (ifh, dfh, sdfh) |
|
2262 | self._writinghandles = (ifh, dfh, sdfh) | |
2195 | self._segmentfile.writing_handle = ifh if self._inline else dfh |
|
2263 | self._segmentfile.writing_handle = ifh if self._inline else dfh | |
2196 | self._segmentfile_sidedata.writing_handle = sdfh |
|
2264 | self._segmentfile_sidedata.writing_handle = sdfh | |
2197 | yield |
|
2265 | yield | |
2198 | if self._docket is not None: |
|
2266 | if self._docket is not None: | |
2199 | self._write_docket(transaction) |
|
2267 | self._write_docket(transaction) | |
2200 | finally: |
|
2268 | finally: | |
2201 | self._writinghandles = None |
|
2269 | self._writinghandles = None | |
2202 | self._segmentfile.writing_handle = None |
|
2270 | self._segmentfile.writing_handle = None | |
2203 | self._segmentfile_sidedata.writing_handle = None |
|
2271 | self._segmentfile_sidedata.writing_handle = None | |
2204 | if dfh is not None: |
|
2272 | if dfh is not None: | |
2205 | dfh.close() |
|
2273 | dfh.close() | |
2206 | if sdfh is not None: |
|
2274 | if sdfh is not None: | |
2207 | sdfh.close() |
|
2275 | sdfh.close() | |
2208 | # closing the index file last to avoid exposing referent to |
|
2276 | # closing the index file last to avoid exposing referent to | |
2209 | # potential unflushed data content. |
|
2277 | # potential unflushed data content. | |
2210 | if ifh is not None: |
|
2278 | if ifh is not None: | |
2211 | ifh.close() |
|
2279 | ifh.close() | |
2212 |
|
2280 | |||
2213 | def _write_docket(self, transaction): |
|
2281 | def _write_docket(self, transaction): | |
2214 | """write the current docket on disk |
|
2282 | """write the current docket on disk | |
2215 |
|
2283 | |||
2216 | Exist as a method to help changelog to implement transaction logic |
|
2284 | Exist as a method to help changelog to implement transaction logic | |
2217 |
|
2285 | |||
2218 | We could also imagine using the same transaction logic for all revlog |
|
2286 | We could also imagine using the same transaction logic for all revlog | |
2219 | since docket are cheap.""" |
|
2287 | since docket are cheap.""" | |
2220 | self._docket.write(transaction) |
|
2288 | self._docket.write(transaction) | |
2221 |
|
2289 | |||
2222 | def addrevision( |
|
2290 | def addrevision( | |
2223 | self, |
|
2291 | self, | |
2224 | text, |
|
2292 | text, | |
2225 | transaction, |
|
2293 | transaction, | |
2226 | link, |
|
2294 | link, | |
2227 | p1, |
|
2295 | p1, | |
2228 | p2, |
|
2296 | p2, | |
2229 | cachedelta=None, |
|
2297 | cachedelta=None, | |
2230 | node=None, |
|
2298 | node=None, | |
2231 | flags=REVIDX_DEFAULT_FLAGS, |
|
2299 | flags=REVIDX_DEFAULT_FLAGS, | |
2232 | deltacomputer=None, |
|
2300 | deltacomputer=None, | |
2233 | sidedata=None, |
|
2301 | sidedata=None, | |
2234 | ): |
|
2302 | ): | |
2235 | """add a revision to the log |
|
2303 | """add a revision to the log | |
2236 |
|
2304 | |||
2237 | text - the revision data to add |
|
2305 | text - the revision data to add | |
2238 | transaction - the transaction object used for rollback |
|
2306 | transaction - the transaction object used for rollback | |
2239 | link - the linkrev data to add |
|
2307 | link - the linkrev data to add | |
2240 | p1, p2 - the parent nodeids of the revision |
|
2308 | p1, p2 - the parent nodeids of the revision | |
2241 | cachedelta - an optional precomputed delta |
|
2309 | cachedelta - an optional precomputed delta | |
2242 | node - nodeid of revision; typically node is not specified, and it is |
|
2310 | node - nodeid of revision; typically node is not specified, and it is | |
2243 | computed by default as hash(text, p1, p2), however subclasses might |
|
2311 | computed by default as hash(text, p1, p2), however subclasses might | |
2244 | use different hashing method (and override checkhash() in such case) |
|
2312 | use different hashing method (and override checkhash() in such case) | |
2245 | flags - the known flags to set on the revision |
|
2313 | flags - the known flags to set on the revision | |
2246 | deltacomputer - an optional deltacomputer instance shared between |
|
2314 | deltacomputer - an optional deltacomputer instance shared between | |
2247 | multiple calls |
|
2315 | multiple calls | |
2248 | """ |
|
2316 | """ | |
2249 | if link == nullrev: |
|
2317 | if link == nullrev: | |
2250 | raise error.RevlogError( |
|
2318 | raise error.RevlogError( | |
2251 | _(b"attempted to add linkrev -1 to %s") % self.display_id |
|
2319 | _(b"attempted to add linkrev -1 to %s") % self.display_id | |
2252 | ) |
|
2320 | ) | |
2253 |
|
2321 | |||
2254 | if sidedata is None: |
|
2322 | if sidedata is None: | |
2255 | sidedata = {} |
|
2323 | sidedata = {} | |
2256 | elif sidedata and not self.hassidedata: |
|
2324 | elif sidedata and not self.hassidedata: | |
2257 | raise error.ProgrammingError( |
|
2325 | raise error.ProgrammingError( | |
2258 | _(b"trying to add sidedata to a revlog who don't support them") |
|
2326 | _(b"trying to add sidedata to a revlog who don't support them") | |
2259 | ) |
|
2327 | ) | |
2260 |
|
2328 | |||
2261 | if flags: |
|
2329 | if flags: | |
2262 | node = node or self.hash(text, p1, p2) |
|
2330 | node = node or self.hash(text, p1, p2) | |
2263 |
|
2331 | |||
2264 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) |
|
2332 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) | |
2265 |
|
2333 | |||
2266 | # If the flag processor modifies the revision data, ignore any provided |
|
2334 | # If the flag processor modifies the revision data, ignore any provided | |
2267 | # cachedelta. |
|
2335 | # cachedelta. | |
2268 | if rawtext != text: |
|
2336 | if rawtext != text: | |
2269 | cachedelta = None |
|
2337 | cachedelta = None | |
2270 |
|
2338 | |||
2271 | if len(rawtext) > _maxentrysize: |
|
2339 | if len(rawtext) > _maxentrysize: | |
2272 | raise error.RevlogError( |
|
2340 | raise error.RevlogError( | |
2273 | _( |
|
2341 | _( | |
2274 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" |
|
2342 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" | |
2275 | ) |
|
2343 | ) | |
2276 | % (self.display_id, len(rawtext)) |
|
2344 | % (self.display_id, len(rawtext)) | |
2277 | ) |
|
2345 | ) | |
2278 |
|
2346 | |||
2279 | node = node or self.hash(rawtext, p1, p2) |
|
2347 | node = node or self.hash(rawtext, p1, p2) | |
2280 | rev = self.index.get_rev(node) |
|
2348 | rev = self.index.get_rev(node) | |
2281 | if rev is not None: |
|
2349 | if rev is not None: | |
2282 | return rev |
|
2350 | return rev | |
2283 |
|
2351 | |||
2284 | if validatehash: |
|
2352 | if validatehash: | |
2285 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
2353 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
2286 |
|
2354 | |||
2287 | return self.addrawrevision( |
|
2355 | return self.addrawrevision( | |
2288 | rawtext, |
|
2356 | rawtext, | |
2289 | transaction, |
|
2357 | transaction, | |
2290 | link, |
|
2358 | link, | |
2291 | p1, |
|
2359 | p1, | |
2292 | p2, |
|
2360 | p2, | |
2293 | node, |
|
2361 | node, | |
2294 | flags, |
|
2362 | flags, | |
2295 | cachedelta=cachedelta, |
|
2363 | cachedelta=cachedelta, | |
2296 | deltacomputer=deltacomputer, |
|
2364 | deltacomputer=deltacomputer, | |
2297 | sidedata=sidedata, |
|
2365 | sidedata=sidedata, | |
2298 | ) |
|
2366 | ) | |
2299 |
|
2367 | |||
2300 | def addrawrevision( |
|
2368 | def addrawrevision( | |
2301 | self, |
|
2369 | self, | |
2302 | rawtext, |
|
2370 | rawtext, | |
2303 | transaction, |
|
2371 | transaction, | |
2304 | link, |
|
2372 | link, | |
2305 | p1, |
|
2373 | p1, | |
2306 | p2, |
|
2374 | p2, | |
2307 | node, |
|
2375 | node, | |
2308 | flags, |
|
2376 | flags, | |
2309 | cachedelta=None, |
|
2377 | cachedelta=None, | |
2310 | deltacomputer=None, |
|
2378 | deltacomputer=None, | |
2311 | sidedata=None, |
|
2379 | sidedata=None, | |
2312 | ): |
|
2380 | ): | |
2313 | """add a raw revision with known flags, node and parents |
|
2381 | """add a raw revision with known flags, node and parents | |
2314 | useful when reusing a revision not stored in this revlog (ex: received |
|
2382 | useful when reusing a revision not stored in this revlog (ex: received | |
2315 | over wire, or read from an external bundle). |
|
2383 | over wire, or read from an external bundle). | |
2316 | """ |
|
2384 | """ | |
2317 | with self._writing(transaction): |
|
2385 | with self._writing(transaction): | |
2318 | return self._addrevision( |
|
2386 | return self._addrevision( | |
2319 | node, |
|
2387 | node, | |
2320 | rawtext, |
|
2388 | rawtext, | |
2321 | transaction, |
|
2389 | transaction, | |
2322 | link, |
|
2390 | link, | |
2323 | p1, |
|
2391 | p1, | |
2324 | p2, |
|
2392 | p2, | |
2325 | flags, |
|
2393 | flags, | |
2326 | cachedelta, |
|
2394 | cachedelta, | |
2327 | deltacomputer=deltacomputer, |
|
2395 | deltacomputer=deltacomputer, | |
2328 | sidedata=sidedata, |
|
2396 | sidedata=sidedata, | |
2329 | ) |
|
2397 | ) | |
2330 |
|
2398 | |||
2331 | def compress(self, data): |
|
2399 | def compress(self, data): | |
2332 | """Generate a possibly-compressed representation of data.""" |
|
2400 | """Generate a possibly-compressed representation of data.""" | |
2333 | if not data: |
|
2401 | if not data: | |
2334 | return b'', data |
|
2402 | return b'', data | |
2335 |
|
2403 | |||
2336 | compressed = self._compressor.compress(data) |
|
2404 | compressed = self._compressor.compress(data) | |
2337 |
|
2405 | |||
2338 | if compressed: |
|
2406 | if compressed: | |
2339 | # The revlog compressor added the header in the returned data. |
|
2407 | # The revlog compressor added the header in the returned data. | |
2340 | return b'', compressed |
|
2408 | return b'', compressed | |
2341 |
|
2409 | |||
2342 | if data[0:1] == b'\0': |
|
2410 | if data[0:1] == b'\0': | |
2343 | return b'', data |
|
2411 | return b'', data | |
2344 | return b'u', data |
|
2412 | return b'u', data | |
2345 |
|
2413 | |||
2346 | def decompress(self, data): |
|
2414 | def decompress(self, data): | |
2347 | """Decompress a revlog chunk. |
|
2415 | """Decompress a revlog chunk. | |
2348 |
|
2416 | |||
2349 | The chunk is expected to begin with a header identifying the |
|
2417 | The chunk is expected to begin with a header identifying the | |
2350 | format type so it can be routed to an appropriate decompressor. |
|
2418 | format type so it can be routed to an appropriate decompressor. | |
2351 | """ |
|
2419 | """ | |
2352 | if not data: |
|
2420 | if not data: | |
2353 | return data |
|
2421 | return data | |
2354 |
|
2422 | |||
2355 | # Revlogs are read much more frequently than they are written and many |
|
2423 | # Revlogs are read much more frequently than they are written and many | |
2356 | # chunks only take microseconds to decompress, so performance is |
|
2424 | # chunks only take microseconds to decompress, so performance is | |
2357 | # important here. |
|
2425 | # important here. | |
2358 | # |
|
2426 | # | |
2359 | # We can make a few assumptions about revlogs: |
|
2427 | # We can make a few assumptions about revlogs: | |
2360 | # |
|
2428 | # | |
2361 | # 1) the majority of chunks will be compressed (as opposed to inline |
|
2429 | # 1) the majority of chunks will be compressed (as opposed to inline | |
2362 | # raw data). |
|
2430 | # raw data). | |
2363 | # 2) decompressing *any* data will likely by at least 10x slower than |
|
2431 | # 2) decompressing *any* data will likely by at least 10x slower than | |
2364 | # returning raw inline data. |
|
2432 | # returning raw inline data. | |
2365 | # 3) we want to prioritize common and officially supported compression |
|
2433 | # 3) we want to prioritize common and officially supported compression | |
2366 | # engines |
|
2434 | # engines | |
2367 | # |
|
2435 | # | |
2368 | # It follows that we want to optimize for "decompress compressed data |
|
2436 | # It follows that we want to optimize for "decompress compressed data | |
2369 | # when encoded with common and officially supported compression engines" |
|
2437 | # when encoded with common and officially supported compression engines" | |
2370 | # case over "raw data" and "data encoded by less common or non-official |
|
2438 | # 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 |
|
2439 | # compression engines." That is why we have the inline lookup first | |
2372 | # followed by the compengines lookup. |
|
2440 | # followed by the compengines lookup. | |
2373 | # |
|
2441 | # | |
2374 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib |
|
2442 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib | |
2375 | # compressed chunks. And this matters for changelog and manifest reads. |
|
2443 | # compressed chunks. And this matters for changelog and manifest reads. | |
2376 | t = data[0:1] |
|
2444 | t = data[0:1] | |
2377 |
|
2445 | |||
2378 | if t == b'x': |
|
2446 | if t == b'x': | |
2379 | try: |
|
2447 | try: | |
2380 | return _zlibdecompress(data) |
|
2448 | return _zlibdecompress(data) | |
2381 | except zlib.error as e: |
|
2449 | except zlib.error as e: | |
2382 | raise error.RevlogError( |
|
2450 | raise error.RevlogError( | |
2383 | _(b'revlog decompress error: %s') |
|
2451 | _(b'revlog decompress error: %s') | |
2384 | % stringutil.forcebytestr(e) |
|
2452 | % stringutil.forcebytestr(e) | |
2385 | ) |
|
2453 | ) | |
2386 | # '\0' is more common than 'u' so it goes first. |
|
2454 | # '\0' is more common than 'u' so it goes first. | |
2387 | elif t == b'\0': |
|
2455 | elif t == b'\0': | |
2388 | return data |
|
2456 | return data | |
2389 | elif t == b'u': |
|
2457 | elif t == b'u': | |
2390 | return util.buffer(data, 1) |
|
2458 | return util.buffer(data, 1) | |
2391 |
|
2459 | |||
2392 | compressor = self._get_decompressor(t) |
|
2460 | compressor = self._get_decompressor(t) | |
2393 |
|
2461 | |||
2394 | return compressor.decompress(data) |
|
2462 | return compressor.decompress(data) | |
2395 |
|
2463 | |||
2396 | def _addrevision( |
|
2464 | def _addrevision( | |
2397 | self, |
|
2465 | self, | |
2398 | node, |
|
2466 | node, | |
2399 | rawtext, |
|
2467 | rawtext, | |
2400 | transaction, |
|
2468 | transaction, | |
2401 | link, |
|
2469 | link, | |
2402 | p1, |
|
2470 | p1, | |
2403 | p2, |
|
2471 | p2, | |
2404 | flags, |
|
2472 | flags, | |
2405 | cachedelta, |
|
2473 | cachedelta, | |
2406 | alwayscache=False, |
|
2474 | alwayscache=False, | |
2407 | deltacomputer=None, |
|
2475 | deltacomputer=None, | |
2408 | sidedata=None, |
|
2476 | sidedata=None, | |
2409 | ): |
|
2477 | ): | |
2410 | """internal function to add revisions to the log |
|
2478 | """internal function to add revisions to the log | |
2411 |
|
2479 | |||
2412 | see addrevision for argument descriptions. |
|
2480 | see addrevision for argument descriptions. | |
2413 |
|
2481 | |||
2414 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. |
|
2482 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. | |
2415 |
|
2483 | |||
2416 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will |
|
2484 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will | |
2417 | be used. |
|
2485 | be used. | |
2418 |
|
2486 | |||
2419 | invariants: |
|
2487 | invariants: | |
2420 | - rawtext is optional (can be None); if not set, cachedelta must be set. |
|
2488 | - rawtext is optional (can be None); if not set, cachedelta must be set. | |
2421 | if both are set, they must correspond to each other. |
|
2489 | if both are set, they must correspond to each other. | |
2422 | """ |
|
2490 | """ | |
2423 | if node == self.nullid: |
|
2491 | if node == self.nullid: | |
2424 | raise error.RevlogError( |
|
2492 | raise error.RevlogError( | |
2425 | _(b"%s: attempt to add null revision") % self.display_id |
|
2493 | _(b"%s: attempt to add null revision") % self.display_id | |
2426 | ) |
|
2494 | ) | |
2427 | if ( |
|
2495 | if ( | |
2428 | node == self.nodeconstants.wdirid |
|
2496 | node == self.nodeconstants.wdirid | |
2429 | or node in self.nodeconstants.wdirfilenodeids |
|
2497 | or node in self.nodeconstants.wdirfilenodeids | |
2430 | ): |
|
2498 | ): | |
2431 | raise error.RevlogError( |
|
2499 | raise error.RevlogError( | |
2432 | _(b"%s: attempt to add wdir revision") % self.display_id |
|
2500 | _(b"%s: attempt to add wdir revision") % self.display_id | |
2433 | ) |
|
2501 | ) | |
2434 | if self._writinghandles is None: |
|
2502 | if self._writinghandles is None: | |
2435 | msg = b'adding revision outside `revlog._writing` context' |
|
2503 | msg = b'adding revision outside `revlog._writing` context' | |
2436 | raise error.ProgrammingError(msg) |
|
2504 | raise error.ProgrammingError(msg) | |
2437 |
|
2505 | |||
2438 | if self._inline: |
|
2506 | if self._inline: | |
2439 | fh = self._writinghandles[0] |
|
2507 | fh = self._writinghandles[0] | |
2440 | else: |
|
2508 | else: | |
2441 | fh = self._writinghandles[1] |
|
2509 | fh = self._writinghandles[1] | |
2442 |
|
2510 | |||
2443 | btext = [rawtext] |
|
2511 | btext = [rawtext] | |
2444 |
|
2512 | |||
2445 | curr = len(self) |
|
2513 | curr = len(self) | |
2446 | prev = curr - 1 |
|
2514 | prev = curr - 1 | |
2447 |
|
2515 | |||
2448 | offset = self._get_data_offset(prev) |
|
2516 | offset = self._get_data_offset(prev) | |
2449 |
|
2517 | |||
2450 | if self._concurrencychecker: |
|
2518 | if self._concurrencychecker: | |
2451 | ifh, dfh, sdfh = self._writinghandles |
|
2519 | ifh, dfh, sdfh = self._writinghandles | |
2452 | # XXX no checking for the sidedata file |
|
2520 | # XXX no checking for the sidedata file | |
2453 | if self._inline: |
|
2521 | if self._inline: | |
2454 | # offset is "as if" it were in the .d file, so we need to add on |
|
2522 | # offset is "as if" it were in the .d file, so we need to add on | |
2455 | # the size of the entry metadata. |
|
2523 | # the size of the entry metadata. | |
2456 | self._concurrencychecker( |
|
2524 | self._concurrencychecker( | |
2457 | ifh, self._indexfile, offset + curr * self.index.entry_size |
|
2525 | ifh, self._indexfile, offset + curr * self.index.entry_size | |
2458 | ) |
|
2526 | ) | |
2459 | else: |
|
2527 | else: | |
2460 | # Entries in the .i are a consistent size. |
|
2528 | # Entries in the .i are a consistent size. | |
2461 | self._concurrencychecker( |
|
2529 | self._concurrencychecker( | |
2462 | ifh, self._indexfile, curr * self.index.entry_size |
|
2530 | ifh, self._indexfile, curr * self.index.entry_size | |
2463 | ) |
|
2531 | ) | |
2464 | self._concurrencychecker(dfh, self._datafile, offset) |
|
2532 | self._concurrencychecker(dfh, self._datafile, offset) | |
2465 |
|
2533 | |||
2466 | p1r, p2r = self.rev(p1), self.rev(p2) |
|
2534 | p1r, p2r = self.rev(p1), self.rev(p2) | |
2467 |
|
2535 | |||
2468 | # full versions are inserted when the needed deltas |
|
2536 | # full versions are inserted when the needed deltas | |
2469 | # become comparable to the uncompressed text |
|
2537 | # become comparable to the uncompressed text | |
2470 | if rawtext is None: |
|
2538 | if rawtext is None: | |
2471 | # need rawtext size, before changed by flag processors, which is |
|
2539 | # need rawtext size, before changed by flag processors, which is | |
2472 | # the non-raw size. use revlog explicitly to avoid filelog's extra |
|
2540 | # the non-raw size. use revlog explicitly to avoid filelog's extra | |
2473 | # logic that might remove metadata size. |
|
2541 | # logic that might remove metadata size. | |
2474 | textlen = mdiff.patchedsize( |
|
2542 | textlen = mdiff.patchedsize( | |
2475 | revlog.size(self, cachedelta[0]), cachedelta[1] |
|
2543 | revlog.size(self, cachedelta[0]), cachedelta[1] | |
2476 | ) |
|
2544 | ) | |
2477 | else: |
|
2545 | else: | |
2478 | textlen = len(rawtext) |
|
2546 | textlen = len(rawtext) | |
2479 |
|
2547 | |||
2480 | if deltacomputer is None: |
|
2548 | if deltacomputer is None: | |
2481 | write_debug = None |
|
2549 | write_debug = None | |
2482 | if self._debug_delta: |
|
2550 | if self._debug_delta: | |
2483 | write_debug = transaction._report |
|
2551 | write_debug = transaction._report | |
2484 | deltacomputer = deltautil.deltacomputer( |
|
2552 | deltacomputer = deltautil.deltacomputer( | |
2485 | self, write_debug=write_debug |
|
2553 | self, write_debug=write_debug | |
2486 | ) |
|
2554 | ) | |
2487 |
|
2555 | |||
2488 | if cachedelta is not None and len(cachedelta) == 2: |
|
2556 | if cachedelta is not None and len(cachedelta) == 2: | |
2489 | # If the cached delta has no information about how it should be |
|
2557 | # If the cached delta has no information about how it should be | |
2490 | # reused, add the default reuse instruction according to the |
|
2558 | # reused, add the default reuse instruction according to the | |
2491 | # revlog's configuration. |
|
2559 | # revlog's configuration. | |
2492 | if self._generaldelta and self._lazydeltabase: |
|
2560 | if self._generaldelta and self._lazydeltabase: | |
2493 | delta_base_reuse = DELTA_BASE_REUSE_TRY |
|
2561 | delta_base_reuse = DELTA_BASE_REUSE_TRY | |
2494 | else: |
|
2562 | else: | |
2495 | delta_base_reuse = DELTA_BASE_REUSE_NO |
|
2563 | delta_base_reuse = DELTA_BASE_REUSE_NO | |
2496 | cachedelta = (cachedelta[0], cachedelta[1], delta_base_reuse) |
|
2564 | cachedelta = (cachedelta[0], cachedelta[1], delta_base_reuse) | |
2497 |
|
2565 | |||
2498 | revinfo = revlogutils.revisioninfo( |
|
2566 | revinfo = revlogutils.revisioninfo( | |
2499 | node, |
|
2567 | node, | |
2500 | p1, |
|
2568 | p1, | |
2501 | p2, |
|
2569 | p2, | |
2502 | btext, |
|
2570 | btext, | |
2503 | textlen, |
|
2571 | textlen, | |
2504 | cachedelta, |
|
2572 | cachedelta, | |
2505 | flags, |
|
2573 | flags, | |
2506 | ) |
|
2574 | ) | |
2507 |
|
2575 | |||
2508 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) |
|
2576 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) | |
2509 |
|
2577 | |||
2510 | compression_mode = COMP_MODE_INLINE |
|
2578 | compression_mode = COMP_MODE_INLINE | |
2511 | if self._docket is not None: |
|
2579 | if self._docket is not None: | |
2512 | default_comp = self._docket.default_compression_header |
|
2580 | default_comp = self._docket.default_compression_header | |
2513 | r = deltautil.delta_compression(default_comp, deltainfo) |
|
2581 | r = deltautil.delta_compression(default_comp, deltainfo) | |
2514 | compression_mode, deltainfo = r |
|
2582 | compression_mode, deltainfo = r | |
2515 |
|
2583 | |||
2516 | sidedata_compression_mode = COMP_MODE_INLINE |
|
2584 | sidedata_compression_mode = COMP_MODE_INLINE | |
2517 | if sidedata and self.hassidedata: |
|
2585 | if sidedata and self.hassidedata: | |
2518 | sidedata_compression_mode = COMP_MODE_PLAIN |
|
2586 | sidedata_compression_mode = COMP_MODE_PLAIN | |
2519 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) |
|
2587 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) | |
2520 | sidedata_offset = self._docket.sidedata_end |
|
2588 | sidedata_offset = self._docket.sidedata_end | |
2521 | h, comp_sidedata = self.compress(serialized_sidedata) |
|
2589 | h, comp_sidedata = self.compress(serialized_sidedata) | |
2522 | if ( |
|
2590 | if ( | |
2523 | h != b'u' |
|
2591 | h != b'u' | |
2524 | and comp_sidedata[0:1] != b'\0' |
|
2592 | and comp_sidedata[0:1] != b'\0' | |
2525 | and len(comp_sidedata) < len(serialized_sidedata) |
|
2593 | and len(comp_sidedata) < len(serialized_sidedata) | |
2526 | ): |
|
2594 | ): | |
2527 | assert not h |
|
2595 | assert not h | |
2528 | if ( |
|
2596 | if ( | |
2529 | comp_sidedata[0:1] |
|
2597 | comp_sidedata[0:1] | |
2530 | == self._docket.default_compression_header |
|
2598 | == self._docket.default_compression_header | |
2531 | ): |
|
2599 | ): | |
2532 | sidedata_compression_mode = COMP_MODE_DEFAULT |
|
2600 | sidedata_compression_mode = COMP_MODE_DEFAULT | |
2533 | serialized_sidedata = comp_sidedata |
|
2601 | serialized_sidedata = comp_sidedata | |
2534 | else: |
|
2602 | else: | |
2535 | sidedata_compression_mode = COMP_MODE_INLINE |
|
2603 | sidedata_compression_mode = COMP_MODE_INLINE | |
2536 | serialized_sidedata = comp_sidedata |
|
2604 | serialized_sidedata = comp_sidedata | |
2537 | else: |
|
2605 | else: | |
2538 | serialized_sidedata = b"" |
|
2606 | serialized_sidedata = b"" | |
2539 | # Don't store the offset if the sidedata is empty, that way |
|
2607 | # Don't store the offset if the sidedata is empty, that way | |
2540 | # we can easily detect empty sidedata and they will be no different |
|
2608 | # we can easily detect empty sidedata and they will be no different | |
2541 | # than ones we manually add. |
|
2609 | # than ones we manually add. | |
2542 | sidedata_offset = 0 |
|
2610 | sidedata_offset = 0 | |
2543 |
|
2611 | |||
2544 | rank = RANK_UNKNOWN |
|
2612 | rank = RANK_UNKNOWN | |
2545 | if self._compute_rank: |
|
2613 | if self._compute_rank: | |
2546 | if (p1r, p2r) == (nullrev, nullrev): |
|
2614 | if (p1r, p2r) == (nullrev, nullrev): | |
2547 | rank = 1 |
|
2615 | rank = 1 | |
2548 | elif p1r != nullrev and p2r == nullrev: |
|
2616 | elif p1r != nullrev and p2r == nullrev: | |
2549 | rank = 1 + self.fast_rank(p1r) |
|
2617 | rank = 1 + self.fast_rank(p1r) | |
2550 | elif p1r == nullrev and p2r != nullrev: |
|
2618 | elif p1r == nullrev and p2r != nullrev: | |
2551 | rank = 1 + self.fast_rank(p2r) |
|
2619 | rank = 1 + self.fast_rank(p2r) | |
2552 | else: # merge node |
|
2620 | else: # merge node | |
2553 | if rustdagop is not None and self.index.rust_ext_compat: |
|
2621 | if rustdagop is not None and self.index.rust_ext_compat: | |
2554 | rank = rustdagop.rank(self.index, p1r, p2r) |
|
2622 | rank = rustdagop.rank(self.index, p1r, p2r) | |
2555 | else: |
|
2623 | else: | |
2556 | pmin, pmax = sorted((p1r, p2r)) |
|
2624 | pmin, pmax = sorted((p1r, p2r)) | |
2557 | rank = 1 + self.fast_rank(pmax) |
|
2625 | rank = 1 + self.fast_rank(pmax) | |
2558 | rank += sum(1 for _ in self.findmissingrevs([pmax], [pmin])) |
|
2626 | rank += sum(1 for _ in self.findmissingrevs([pmax], [pmin])) | |
2559 |
|
2627 | |||
2560 | e = revlogutils.entry( |
|
2628 | e = revlogutils.entry( | |
2561 | flags=flags, |
|
2629 | flags=flags, | |
2562 | data_offset=offset, |
|
2630 | data_offset=offset, | |
2563 | data_compressed_length=deltainfo.deltalen, |
|
2631 | data_compressed_length=deltainfo.deltalen, | |
2564 | data_uncompressed_length=textlen, |
|
2632 | data_uncompressed_length=textlen, | |
2565 | data_compression_mode=compression_mode, |
|
2633 | data_compression_mode=compression_mode, | |
2566 | data_delta_base=deltainfo.base, |
|
2634 | data_delta_base=deltainfo.base, | |
2567 | link_rev=link, |
|
2635 | link_rev=link, | |
2568 | parent_rev_1=p1r, |
|
2636 | parent_rev_1=p1r, | |
2569 | parent_rev_2=p2r, |
|
2637 | parent_rev_2=p2r, | |
2570 | node_id=node, |
|
2638 | node_id=node, | |
2571 | sidedata_offset=sidedata_offset, |
|
2639 | sidedata_offset=sidedata_offset, | |
2572 | sidedata_compressed_length=len(serialized_sidedata), |
|
2640 | sidedata_compressed_length=len(serialized_sidedata), | |
2573 | sidedata_compression_mode=sidedata_compression_mode, |
|
2641 | sidedata_compression_mode=sidedata_compression_mode, | |
2574 | rank=rank, |
|
2642 | rank=rank, | |
2575 | ) |
|
2643 | ) | |
2576 |
|
2644 | |||
2577 | self.index.append(e) |
|
2645 | self.index.append(e) | |
2578 | entry = self.index.entry_binary(curr) |
|
2646 | entry = self.index.entry_binary(curr) | |
2579 | if curr == 0 and self._docket is None: |
|
2647 | if curr == 0 and self._docket is None: | |
2580 | header = self._format_flags | self._format_version |
|
2648 | header = self._format_flags | self._format_version | |
2581 | header = self.index.pack_header(header) |
|
2649 | header = self.index.pack_header(header) | |
2582 | entry = header + entry |
|
2650 | entry = header + entry | |
2583 | self._writeentry( |
|
2651 | self._writeentry( | |
2584 | transaction, |
|
2652 | transaction, | |
2585 | entry, |
|
2653 | entry, | |
2586 | deltainfo.data, |
|
2654 | deltainfo.data, | |
2587 | link, |
|
2655 | link, | |
2588 | offset, |
|
2656 | offset, | |
2589 | serialized_sidedata, |
|
2657 | serialized_sidedata, | |
2590 | sidedata_offset, |
|
2658 | sidedata_offset, | |
2591 | ) |
|
2659 | ) | |
2592 |
|
2660 | |||
2593 | rawtext = btext[0] |
|
2661 | rawtext = btext[0] | |
2594 |
|
2662 | |||
2595 | if alwayscache and rawtext is None: |
|
2663 | if alwayscache and rawtext is None: | |
2596 | rawtext = deltacomputer.buildtext(revinfo, fh) |
|
2664 | rawtext = deltacomputer.buildtext(revinfo, fh) | |
2597 |
|
2665 | |||
2598 | if type(rawtext) == bytes: # only accept immutable objects |
|
2666 | if type(rawtext) == bytes: # only accept immutable objects | |
2599 | self._revisioncache = (node, curr, rawtext) |
|
2667 | self._revisioncache = (node, curr, rawtext) | |
2600 | self._chainbasecache[curr] = deltainfo.chainbase |
|
2668 | self._chainbasecache[curr] = deltainfo.chainbase | |
2601 | return curr |
|
2669 | return curr | |
2602 |
|
2670 | |||
2603 | def _get_data_offset(self, prev): |
|
2671 | def _get_data_offset(self, prev): | |
2604 | """Returns the current offset in the (in-transaction) data file. |
|
2672 | """Returns the current offset in the (in-transaction) data file. | |
2605 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket |
|
2673 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket | |
2606 | file to store that information: since sidedata can be rewritten to the |
|
2674 | file to store that information: since sidedata can be rewritten to the | |
2607 | end of the data file within a transaction, you can have cases where, for |
|
2675 | end of the data file within a transaction, you can have cases where, for | |
2608 | example, rev `n` does not have sidedata while rev `n - 1` does, leading |
|
2676 | example, rev `n` does not have sidedata while rev `n - 1` does, leading | |
2609 | to `n - 1`'s sidedata being written after `n`'s data. |
|
2677 | to `n - 1`'s sidedata being written after `n`'s data. | |
2610 |
|
2678 | |||
2611 | TODO cache this in a docket file before getting out of experimental.""" |
|
2679 | TODO cache this in a docket file before getting out of experimental.""" | |
2612 | if self._docket is None: |
|
2680 | if self._docket is None: | |
2613 | return self.end(prev) |
|
2681 | return self.end(prev) | |
2614 | else: |
|
2682 | else: | |
2615 | return self._docket.data_end |
|
2683 | return self._docket.data_end | |
2616 |
|
2684 | |||
2617 | def _writeentry( |
|
2685 | def _writeentry( | |
2618 | self, transaction, entry, data, link, offset, sidedata, sidedata_offset |
|
2686 | self, transaction, entry, data, link, offset, sidedata, sidedata_offset | |
2619 | ): |
|
2687 | ): | |
2620 | # Files opened in a+ mode have inconsistent behavior on various |
|
2688 | # Files opened in a+ mode have inconsistent behavior on various | |
2621 | # platforms. Windows requires that a file positioning call be made |
|
2689 | # platforms. Windows requires that a file positioning call be made | |
2622 | # when the file handle transitions between reads and writes. See |
|
2690 | # when the file handle transitions between reads and writes. See | |
2623 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
2691 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other | |
2624 | # platforms, Python or the platform itself can be buggy. Some versions |
|
2692 | # platforms, Python or the platform itself can be buggy. Some versions | |
2625 | # of Solaris have been observed to not append at the end of the file |
|
2693 | # of Solaris have been observed to not append at the end of the file | |
2626 | # if the file was seeked to before the end. See issue4943 for more. |
|
2694 | # if the file was seeked to before the end. See issue4943 for more. | |
2627 | # |
|
2695 | # | |
2628 | # We work around this issue by inserting a seek() before writing. |
|
2696 | # We work around this issue by inserting a seek() before writing. | |
2629 | # Note: This is likely not necessary on Python 3. However, because |
|
2697 | # Note: This is likely not necessary on Python 3. However, because | |
2630 | # the file handle is reused for reads and may be seeked there, we need |
|
2698 | # the file handle is reused for reads and may be seeked there, we need | |
2631 | # to be careful before changing this. |
|
2699 | # to be careful before changing this. | |
2632 | if self._writinghandles is None: |
|
2700 | if self._writinghandles is None: | |
2633 | msg = b'adding revision outside `revlog._writing` context' |
|
2701 | msg = b'adding revision outside `revlog._writing` context' | |
2634 | raise error.ProgrammingError(msg) |
|
2702 | raise error.ProgrammingError(msg) | |
2635 | ifh, dfh, sdfh = self._writinghandles |
|
2703 | ifh, dfh, sdfh = self._writinghandles | |
2636 | if self._docket is None: |
|
2704 | if self._docket is None: | |
2637 | ifh.seek(0, os.SEEK_END) |
|
2705 | ifh.seek(0, os.SEEK_END) | |
2638 | else: |
|
2706 | else: | |
2639 | ifh.seek(self._docket.index_end, os.SEEK_SET) |
|
2707 | ifh.seek(self._docket.index_end, os.SEEK_SET) | |
2640 | if dfh: |
|
2708 | if dfh: | |
2641 | if self._docket is None: |
|
2709 | if self._docket is None: | |
2642 | dfh.seek(0, os.SEEK_END) |
|
2710 | dfh.seek(0, os.SEEK_END) | |
2643 | else: |
|
2711 | else: | |
2644 | dfh.seek(self._docket.data_end, os.SEEK_SET) |
|
2712 | dfh.seek(self._docket.data_end, os.SEEK_SET) | |
2645 | if sdfh: |
|
2713 | if sdfh: | |
2646 | sdfh.seek(self._docket.sidedata_end, os.SEEK_SET) |
|
2714 | sdfh.seek(self._docket.sidedata_end, os.SEEK_SET) | |
2647 |
|
2715 | |||
2648 | curr = len(self) - 1 |
|
2716 | curr = len(self) - 1 | |
2649 | if not self._inline: |
|
2717 | if not self._inline: | |
2650 | transaction.add(self._datafile, offset) |
|
2718 | transaction.add(self._datafile, offset) | |
2651 | if self._sidedatafile: |
|
2719 | if self._sidedatafile: | |
2652 | transaction.add(self._sidedatafile, sidedata_offset) |
|
2720 | transaction.add(self._sidedatafile, sidedata_offset) | |
2653 | transaction.add(self._indexfile, curr * len(entry)) |
|
2721 | transaction.add(self._indexfile, curr * len(entry)) | |
2654 | if data[0]: |
|
2722 | if data[0]: | |
2655 | dfh.write(data[0]) |
|
2723 | dfh.write(data[0]) | |
2656 | dfh.write(data[1]) |
|
2724 | dfh.write(data[1]) | |
2657 | if sidedata: |
|
2725 | if sidedata: | |
2658 | sdfh.write(sidedata) |
|
2726 | sdfh.write(sidedata) | |
2659 | ifh.write(entry) |
|
2727 | ifh.write(entry) | |
2660 | else: |
|
2728 | else: | |
2661 | offset += curr * self.index.entry_size |
|
2729 | offset += curr * self.index.entry_size | |
2662 | transaction.add(self._indexfile, offset) |
|
2730 | transaction.add(self._indexfile, offset) | |
2663 | ifh.write(entry) |
|
2731 | ifh.write(entry) | |
2664 | ifh.write(data[0]) |
|
2732 | ifh.write(data[0]) | |
2665 | ifh.write(data[1]) |
|
2733 | ifh.write(data[1]) | |
2666 | assert not sidedata |
|
2734 | assert not sidedata | |
2667 | self._enforceinlinesize(transaction) |
|
2735 | self._enforceinlinesize(transaction) | |
2668 | if self._docket is not None: |
|
2736 | if self._docket is not None: | |
2669 | # revlog-v2 always has 3 writing handles, help Pytype |
|
2737 | # revlog-v2 always has 3 writing handles, help Pytype | |
2670 | wh1 = self._writinghandles[0] |
|
2738 | wh1 = self._writinghandles[0] | |
2671 | wh2 = self._writinghandles[1] |
|
2739 | wh2 = self._writinghandles[1] | |
2672 | wh3 = self._writinghandles[2] |
|
2740 | wh3 = self._writinghandles[2] | |
2673 | assert wh1 is not None |
|
2741 | assert wh1 is not None | |
2674 | assert wh2 is not None |
|
2742 | assert wh2 is not None | |
2675 | assert wh3 is not None |
|
2743 | assert wh3 is not None | |
2676 | self._docket.index_end = wh1.tell() |
|
2744 | self._docket.index_end = wh1.tell() | |
2677 | self._docket.data_end = wh2.tell() |
|
2745 | self._docket.data_end = wh2.tell() | |
2678 | self._docket.sidedata_end = wh3.tell() |
|
2746 | self._docket.sidedata_end = wh3.tell() | |
2679 |
|
2747 | |||
2680 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
2748 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
2681 |
|
2749 | |||
2682 | def addgroup( |
|
2750 | def addgroup( | |
2683 | self, |
|
2751 | self, | |
2684 | deltas, |
|
2752 | deltas, | |
2685 | linkmapper, |
|
2753 | linkmapper, | |
2686 | transaction, |
|
2754 | transaction, | |
2687 | alwayscache=False, |
|
2755 | alwayscache=False, | |
2688 | addrevisioncb=None, |
|
2756 | addrevisioncb=None, | |
2689 | duplicaterevisioncb=None, |
|
2757 | duplicaterevisioncb=None, | |
2690 | debug_info=None, |
|
2758 | debug_info=None, | |
2691 | delta_base_reuse_policy=None, |
|
2759 | delta_base_reuse_policy=None, | |
2692 | ): |
|
2760 | ): | |
2693 | """ |
|
2761 | """ | |
2694 | add a delta group |
|
2762 | add a delta group | |
2695 |
|
2763 | |||
2696 | given a set of deltas, add them to the revision log. the |
|
2764 | given a set of deltas, add them to the revision log. the | |
2697 | first delta is against its parent, which should be in our |
|
2765 | first delta is against its parent, which should be in our | |
2698 | log, the rest are against the previous delta. |
|
2766 | log, the rest are against the previous delta. | |
2699 |
|
2767 | |||
2700 | If ``addrevisioncb`` is defined, it will be called with arguments of |
|
2768 | If ``addrevisioncb`` is defined, it will be called with arguments of | |
2701 | this revlog and the node that was added. |
|
2769 | this revlog and the node that was added. | |
2702 | """ |
|
2770 | """ | |
2703 |
|
2771 | |||
2704 | if self._adding_group: |
|
2772 | if self._adding_group: | |
2705 | raise error.ProgrammingError(b'cannot nest addgroup() calls') |
|
2773 | raise error.ProgrammingError(b'cannot nest addgroup() calls') | |
2706 |
|
2774 | |||
2707 | # read the default delta-base reuse policy from revlog config if the |
|
2775 | # read the default delta-base reuse policy from revlog config if the | |
2708 | # group did not specify one. |
|
2776 | # group did not specify one. | |
2709 | if delta_base_reuse_policy is None: |
|
2777 | if delta_base_reuse_policy is None: | |
2710 | if self._generaldelta and self._lazydeltabase: |
|
2778 | if self._generaldelta and self._lazydeltabase: | |
2711 | delta_base_reuse_policy = DELTA_BASE_REUSE_TRY |
|
2779 | delta_base_reuse_policy = DELTA_BASE_REUSE_TRY | |
2712 | else: |
|
2780 | else: | |
2713 | delta_base_reuse_policy = DELTA_BASE_REUSE_NO |
|
2781 | delta_base_reuse_policy = DELTA_BASE_REUSE_NO | |
2714 |
|
2782 | |||
2715 | self._adding_group = True |
|
2783 | self._adding_group = True | |
2716 | empty = True |
|
2784 | empty = True | |
2717 | try: |
|
2785 | try: | |
2718 | with self._writing(transaction): |
|
2786 | with self._writing(transaction): | |
2719 | write_debug = None |
|
2787 | write_debug = None | |
2720 | if self._debug_delta: |
|
2788 | if self._debug_delta: | |
2721 | write_debug = transaction._report |
|
2789 | write_debug = transaction._report | |
2722 | deltacomputer = deltautil.deltacomputer( |
|
2790 | deltacomputer = deltautil.deltacomputer( | |
2723 | self, |
|
2791 | self, | |
2724 | write_debug=write_debug, |
|
2792 | write_debug=write_debug, | |
2725 | debug_info=debug_info, |
|
2793 | debug_info=debug_info, | |
2726 | ) |
|
2794 | ) | |
2727 | # loop through our set of deltas |
|
2795 | # loop through our set of deltas | |
2728 | for data in deltas: |
|
2796 | for data in deltas: | |
2729 | ( |
|
2797 | ( | |
2730 | node, |
|
2798 | node, | |
2731 | p1, |
|
2799 | p1, | |
2732 | p2, |
|
2800 | p2, | |
2733 | linknode, |
|
2801 | linknode, | |
2734 | deltabase, |
|
2802 | deltabase, | |
2735 | delta, |
|
2803 | delta, | |
2736 | flags, |
|
2804 | flags, | |
2737 | sidedata, |
|
2805 | sidedata, | |
2738 | ) = data |
|
2806 | ) = data | |
2739 | link = linkmapper(linknode) |
|
2807 | link = linkmapper(linknode) | |
2740 | flags = flags or REVIDX_DEFAULT_FLAGS |
|
2808 | flags = flags or REVIDX_DEFAULT_FLAGS | |
2741 |
|
2809 | |||
2742 | rev = self.index.get_rev(node) |
|
2810 | rev = self.index.get_rev(node) | |
2743 | if rev is not None: |
|
2811 | if rev is not None: | |
2744 | # this can happen if two branches make the same change |
|
2812 | # this can happen if two branches make the same change | |
2745 | self._nodeduplicatecallback(transaction, rev) |
|
2813 | self._nodeduplicatecallback(transaction, rev) | |
2746 | if duplicaterevisioncb: |
|
2814 | if duplicaterevisioncb: | |
2747 | duplicaterevisioncb(self, rev) |
|
2815 | duplicaterevisioncb(self, rev) | |
2748 | empty = False |
|
2816 | empty = False | |
2749 | continue |
|
2817 | continue | |
2750 |
|
2818 | |||
2751 | for p in (p1, p2): |
|
2819 | for p in (p1, p2): | |
2752 | if not self.index.has_node(p): |
|
2820 | if not self.index.has_node(p): | |
2753 | raise error.LookupError( |
|
2821 | raise error.LookupError( | |
2754 | p, self.radix, _(b'unknown parent') |
|
2822 | p, self.radix, _(b'unknown parent') | |
2755 | ) |
|
2823 | ) | |
2756 |
|
2824 | |||
2757 | if not self.index.has_node(deltabase): |
|
2825 | if not self.index.has_node(deltabase): | |
2758 | raise error.LookupError( |
|
2826 | raise error.LookupError( | |
2759 | deltabase, self.display_id, _(b'unknown delta base') |
|
2827 | deltabase, self.display_id, _(b'unknown delta base') | |
2760 | ) |
|
2828 | ) | |
2761 |
|
2829 | |||
2762 | baserev = self.rev(deltabase) |
|
2830 | baserev = self.rev(deltabase) | |
2763 |
|
2831 | |||
2764 | if baserev != nullrev and self.iscensored(baserev): |
|
2832 | if baserev != nullrev and self.iscensored(baserev): | |
2765 | # if base is censored, delta must be full replacement in a |
|
2833 | # if base is censored, delta must be full replacement in a | |
2766 | # single patch operation |
|
2834 | # single patch operation | |
2767 | hlen = struct.calcsize(b">lll") |
|
2835 | hlen = struct.calcsize(b">lll") | |
2768 | oldlen = self.rawsize(baserev) |
|
2836 | oldlen = self.rawsize(baserev) | |
2769 | newlen = len(delta) - hlen |
|
2837 | newlen = len(delta) - hlen | |
2770 | if delta[:hlen] != mdiff.replacediffheader( |
|
2838 | if delta[:hlen] != mdiff.replacediffheader( | |
2771 | oldlen, newlen |
|
2839 | oldlen, newlen | |
2772 | ): |
|
2840 | ): | |
2773 | raise error.CensoredBaseError( |
|
2841 | raise error.CensoredBaseError( | |
2774 | self.display_id, self.node(baserev) |
|
2842 | self.display_id, self.node(baserev) | |
2775 | ) |
|
2843 | ) | |
2776 |
|
2844 | |||
2777 | if not flags and self._peek_iscensored(baserev, delta): |
|
2845 | if not flags and self._peek_iscensored(baserev, delta): | |
2778 | flags |= REVIDX_ISCENSORED |
|
2846 | flags |= REVIDX_ISCENSORED | |
2779 |
|
2847 | |||
2780 | # We assume consumers of addrevisioncb will want to retrieve |
|
2848 | # We assume consumers of addrevisioncb will want to retrieve | |
2781 | # the added revision, which will require a call to |
|
2849 | # the added revision, which will require a call to | |
2782 | # revision(). revision() will fast path if there is a cache |
|
2850 | # revision(). revision() will fast path if there is a cache | |
2783 | # hit. So, we tell _addrevision() to always cache in this case. |
|
2851 | # hit. So, we tell _addrevision() to always cache in this case. | |
2784 | # We're only using addgroup() in the context of changegroup |
|
2852 | # We're only using addgroup() in the context of changegroup | |
2785 | # generation so the revision data can always be handled as raw |
|
2853 | # generation so the revision data can always be handled as raw | |
2786 | # by the flagprocessor. |
|
2854 | # by the flagprocessor. | |
2787 | rev = self._addrevision( |
|
2855 | rev = self._addrevision( | |
2788 | node, |
|
2856 | node, | |
2789 | None, |
|
2857 | None, | |
2790 | transaction, |
|
2858 | transaction, | |
2791 | link, |
|
2859 | link, | |
2792 | p1, |
|
2860 | p1, | |
2793 | p2, |
|
2861 | p2, | |
2794 | flags, |
|
2862 | flags, | |
2795 | (baserev, delta, delta_base_reuse_policy), |
|
2863 | (baserev, delta, delta_base_reuse_policy), | |
2796 | alwayscache=alwayscache, |
|
2864 | alwayscache=alwayscache, | |
2797 | deltacomputer=deltacomputer, |
|
2865 | deltacomputer=deltacomputer, | |
2798 | sidedata=sidedata, |
|
2866 | sidedata=sidedata, | |
2799 | ) |
|
2867 | ) | |
2800 |
|
2868 | |||
2801 | if addrevisioncb: |
|
2869 | if addrevisioncb: | |
2802 | addrevisioncb(self, rev) |
|
2870 | addrevisioncb(self, rev) | |
2803 | empty = False |
|
2871 | empty = False | |
2804 | finally: |
|
2872 | finally: | |
2805 | self._adding_group = False |
|
2873 | self._adding_group = False | |
2806 | return not empty |
|
2874 | return not empty | |
2807 |
|
2875 | |||
2808 | def iscensored(self, rev): |
|
2876 | def iscensored(self, rev): | |
2809 | """Check if a file revision is censored.""" |
|
2877 | """Check if a file revision is censored.""" | |
2810 | if not self._censorable: |
|
2878 | if not self._censorable: | |
2811 | return False |
|
2879 | return False | |
2812 |
|
2880 | |||
2813 | return self.flags(rev) & REVIDX_ISCENSORED |
|
2881 | return self.flags(rev) & REVIDX_ISCENSORED | |
2814 |
|
2882 | |||
2815 | def _peek_iscensored(self, baserev, delta): |
|
2883 | def _peek_iscensored(self, baserev, delta): | |
2816 | """Quickly check if a delta produces a censored revision.""" |
|
2884 | """Quickly check if a delta produces a censored revision.""" | |
2817 | if not self._censorable: |
|
2885 | if not self._censorable: | |
2818 | return False |
|
2886 | return False | |
2819 |
|
2887 | |||
2820 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) |
|
2888 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) | |
2821 |
|
2889 | |||
2822 | def getstrippoint(self, minlink): |
|
2890 | def getstrippoint(self, minlink): | |
2823 | """find the minimum rev that must be stripped to strip the linkrev |
|
2891 | """find the minimum rev that must be stripped to strip the linkrev | |
2824 |
|
2892 | |||
2825 | Returns a tuple containing the minimum rev and a set of all revs that |
|
2893 | Returns a tuple containing the minimum rev and a set of all revs that | |
2826 | have linkrevs that will be broken by this strip. |
|
2894 | have linkrevs that will be broken by this strip. | |
2827 | """ |
|
2895 | """ | |
2828 | return storageutil.resolvestripinfo( |
|
2896 | return storageutil.resolvestripinfo( | |
2829 | minlink, |
|
2897 | minlink, | |
2830 | len(self) - 1, |
|
2898 | len(self) - 1, | |
2831 | self.headrevs(), |
|
2899 | self.headrevs(), | |
2832 | self.linkrev, |
|
2900 | self.linkrev, | |
2833 | self.parentrevs, |
|
2901 | self.parentrevs, | |
2834 | ) |
|
2902 | ) | |
2835 |
|
2903 | |||
2836 | def strip(self, minlink, transaction): |
|
2904 | def strip(self, minlink, transaction): | |
2837 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
2905 | """truncate the revlog on the first revision with a linkrev >= minlink | |
2838 |
|
2906 | |||
2839 | This function is called when we're stripping revision minlink and |
|
2907 | This function is called when we're stripping revision minlink and | |
2840 | its descendants from the repository. |
|
2908 | its descendants from the repository. | |
2841 |
|
2909 | |||
2842 | We have to remove all revisions with linkrev >= minlink, because |
|
2910 | We have to remove all revisions with linkrev >= minlink, because | |
2843 | the equivalent changelog revisions will be renumbered after the |
|
2911 | the equivalent changelog revisions will be renumbered after the | |
2844 | strip. |
|
2912 | strip. | |
2845 |
|
2913 | |||
2846 | So we truncate the revlog on the first of these revisions, and |
|
2914 | So we truncate the revlog on the first of these revisions, and | |
2847 | trust that the caller has saved the revisions that shouldn't be |
|
2915 | trust that the caller has saved the revisions that shouldn't be | |
2848 | removed and that it'll re-add them after this truncation. |
|
2916 | removed and that it'll re-add them after this truncation. | |
2849 | """ |
|
2917 | """ | |
2850 | if len(self) == 0: |
|
2918 | if len(self) == 0: | |
2851 | return |
|
2919 | return | |
2852 |
|
2920 | |||
2853 | rev, _ = self.getstrippoint(minlink) |
|
2921 | rev, _ = self.getstrippoint(minlink) | |
2854 | if rev == len(self): |
|
2922 | if rev == len(self): | |
2855 | return |
|
2923 | return | |
2856 |
|
2924 | |||
2857 | # first truncate the files on disk |
|
2925 | # first truncate the files on disk | |
2858 | data_end = self.start(rev) |
|
2926 | data_end = self.start(rev) | |
2859 | if not self._inline: |
|
2927 | if not self._inline: | |
2860 | transaction.add(self._datafile, data_end) |
|
2928 | transaction.add(self._datafile, data_end) | |
2861 | end = rev * self.index.entry_size |
|
2929 | end = rev * self.index.entry_size | |
2862 | else: |
|
2930 | else: | |
2863 | end = data_end + (rev * self.index.entry_size) |
|
2931 | end = data_end + (rev * self.index.entry_size) | |
2864 |
|
2932 | |||
2865 | if self._sidedatafile: |
|
2933 | if self._sidedatafile: | |
2866 | sidedata_end = self.sidedata_cut_off(rev) |
|
2934 | sidedata_end = self.sidedata_cut_off(rev) | |
2867 | transaction.add(self._sidedatafile, sidedata_end) |
|
2935 | transaction.add(self._sidedatafile, sidedata_end) | |
2868 |
|
2936 | |||
2869 | transaction.add(self._indexfile, end) |
|
2937 | transaction.add(self._indexfile, end) | |
2870 | if self._docket is not None: |
|
2938 | if self._docket is not None: | |
2871 | # XXX we could, leverage the docket while stripping. However it is |
|
2939 | # XXX we could, leverage the docket while stripping. However it is | |
2872 | # not powerfull enough at the time of this comment |
|
2940 | # not powerfull enough at the time of this comment | |
2873 | self._docket.index_end = end |
|
2941 | self._docket.index_end = end | |
2874 | self._docket.data_end = data_end |
|
2942 | self._docket.data_end = data_end | |
2875 | self._docket.sidedata_end = sidedata_end |
|
2943 | self._docket.sidedata_end = sidedata_end | |
2876 | self._docket.write(transaction, stripping=True) |
|
2944 | self._docket.write(transaction, stripping=True) | |
2877 |
|
2945 | |||
2878 | # then reset internal state in memory to forget those revisions |
|
2946 | # then reset internal state in memory to forget those revisions | |
2879 | self._revisioncache = None |
|
2947 | self._revisioncache = None | |
2880 | self._chaininfocache = util.lrucachedict(500) |
|
2948 | self._chaininfocache = util.lrucachedict(500) | |
2881 | self._segmentfile.clear_cache() |
|
2949 | self._segmentfile.clear_cache() | |
2882 | self._segmentfile_sidedata.clear_cache() |
|
2950 | self._segmentfile_sidedata.clear_cache() | |
2883 |
|
2951 | |||
2884 | del self.index[rev:-1] |
|
2952 | del self.index[rev:-1] | |
2885 |
|
2953 | |||
2886 | def checksize(self): |
|
2954 | def checksize(self): | |
2887 | """Check size of index and data files |
|
2955 | """Check size of index and data files | |
2888 |
|
2956 | |||
2889 | return a (dd, di) tuple. |
|
2957 | return a (dd, di) tuple. | |
2890 | - dd: extra bytes for the "data" file |
|
2958 | - dd: extra bytes for the "data" file | |
2891 | - di: extra bytes for the "index" file |
|
2959 | - di: extra bytes for the "index" file | |
2892 |
|
2960 | |||
2893 | A healthy revlog will return (0, 0). |
|
2961 | A healthy revlog will return (0, 0). | |
2894 | """ |
|
2962 | """ | |
2895 | expected = 0 |
|
2963 | expected = 0 | |
2896 | if len(self): |
|
2964 | if len(self): | |
2897 | expected = max(0, self.end(len(self) - 1)) |
|
2965 | expected = max(0, self.end(len(self) - 1)) | |
2898 |
|
2966 | |||
2899 | try: |
|
2967 | try: | |
2900 | with self._datafp() as f: |
|
2968 | with self._datafp() as f: | |
2901 | f.seek(0, io.SEEK_END) |
|
2969 | f.seek(0, io.SEEK_END) | |
2902 | actual = f.tell() |
|
2970 | actual = f.tell() | |
2903 | dd = actual - expected |
|
2971 | dd = actual - expected | |
2904 | except FileNotFoundError: |
|
2972 | except FileNotFoundError: | |
2905 | dd = 0 |
|
2973 | dd = 0 | |
2906 |
|
2974 | |||
2907 | try: |
|
2975 | try: | |
2908 | f = self.opener(self._indexfile) |
|
2976 | f = self.opener(self._indexfile) | |
2909 | f.seek(0, io.SEEK_END) |
|
2977 | f.seek(0, io.SEEK_END) | |
2910 | actual = f.tell() |
|
2978 | actual = f.tell() | |
2911 | f.close() |
|
2979 | f.close() | |
2912 | s = self.index.entry_size |
|
2980 | s = self.index.entry_size | |
2913 | i = max(0, actual // s) |
|
2981 | i = max(0, actual // s) | |
2914 | di = actual - (i * s) |
|
2982 | di = actual - (i * s) | |
2915 | if self._inline: |
|
2983 | if self._inline: | |
2916 | databytes = 0 |
|
2984 | databytes = 0 | |
2917 | for r in self: |
|
2985 | for r in self: | |
2918 | databytes += max(0, self.length(r)) |
|
2986 | databytes += max(0, self.length(r)) | |
2919 | dd = 0 |
|
2987 | dd = 0 | |
2920 | di = actual - len(self) * s - databytes |
|
2988 | di = actual - len(self) * s - databytes | |
2921 | except FileNotFoundError: |
|
2989 | except FileNotFoundError: | |
2922 | di = 0 |
|
2990 | di = 0 | |
2923 |
|
2991 | |||
2924 | return (dd, di) |
|
2992 | return (dd, di) | |
2925 |
|
2993 | |||
2926 | def files(self): |
|
2994 | def files(self): | |
2927 | res = [self._indexfile] |
|
2995 | res = [self._indexfile] | |
2928 | if self._docket_file is None: |
|
2996 | if self._docket_file is None: | |
2929 | if not self._inline: |
|
2997 | if not self._inline: | |
2930 | res.append(self._datafile) |
|
2998 | res.append(self._datafile) | |
2931 | else: |
|
2999 | else: | |
2932 | res.append(self._docket_file) |
|
3000 | res.append(self._docket_file) | |
2933 | res.extend(self._docket.old_index_filepaths(include_empty=False)) |
|
3001 | res.extend(self._docket.old_index_filepaths(include_empty=False)) | |
2934 | if self._docket.data_end: |
|
3002 | if self._docket.data_end: | |
2935 | res.append(self._datafile) |
|
3003 | res.append(self._datafile) | |
2936 | res.extend(self._docket.old_data_filepaths(include_empty=False)) |
|
3004 | res.extend(self._docket.old_data_filepaths(include_empty=False)) | |
2937 | if self._docket.sidedata_end: |
|
3005 | if self._docket.sidedata_end: | |
2938 | res.append(self._sidedatafile) |
|
3006 | res.append(self._sidedatafile) | |
2939 | res.extend(self._docket.old_sidedata_filepaths(include_empty=False)) |
|
3007 | res.extend(self._docket.old_sidedata_filepaths(include_empty=False)) | |
2940 | return res |
|
3008 | return res | |
2941 |
|
3009 | |||
2942 | def emitrevisions( |
|
3010 | def emitrevisions( | |
2943 | self, |
|
3011 | self, | |
2944 | nodes, |
|
3012 | nodes, | |
2945 | nodesorder=None, |
|
3013 | nodesorder=None, | |
2946 | revisiondata=False, |
|
3014 | revisiondata=False, | |
2947 | assumehaveparentrevisions=False, |
|
3015 | assumehaveparentrevisions=False, | |
2948 | deltamode=repository.CG_DELTAMODE_STD, |
|
3016 | deltamode=repository.CG_DELTAMODE_STD, | |
2949 | sidedata_helpers=None, |
|
3017 | sidedata_helpers=None, | |
2950 | debug_info=None, |
|
3018 | debug_info=None, | |
2951 | ): |
|
3019 | ): | |
2952 | if nodesorder not in (b'nodes', b'storage', b'linear', None): |
|
3020 | if nodesorder not in (b'nodes', b'storage', b'linear', None): | |
2953 | raise error.ProgrammingError( |
|
3021 | raise error.ProgrammingError( | |
2954 | b'unhandled value for nodesorder: %s' % nodesorder |
|
3022 | b'unhandled value for nodesorder: %s' % nodesorder | |
2955 | ) |
|
3023 | ) | |
2956 |
|
3024 | |||
2957 | if nodesorder is None and not self._generaldelta: |
|
3025 | if nodesorder is None and not self._generaldelta: | |
2958 | nodesorder = b'storage' |
|
3026 | nodesorder = b'storage' | |
2959 |
|
3027 | |||
2960 | if ( |
|
3028 | if ( | |
2961 | not self._storedeltachains |
|
3029 | not self._storedeltachains | |
2962 | and deltamode != repository.CG_DELTAMODE_PREV |
|
3030 | and deltamode != repository.CG_DELTAMODE_PREV | |
2963 | ): |
|
3031 | ): | |
2964 | deltamode = repository.CG_DELTAMODE_FULL |
|
3032 | deltamode = repository.CG_DELTAMODE_FULL | |
2965 |
|
3033 | |||
2966 | return storageutil.emitrevisions( |
|
3034 | return storageutil.emitrevisions( | |
2967 | self, |
|
3035 | self, | |
2968 | nodes, |
|
3036 | nodes, | |
2969 | nodesorder, |
|
3037 | nodesorder, | |
2970 | revlogrevisiondelta, |
|
3038 | revlogrevisiondelta, | |
2971 | deltaparentfn=self.deltaparent, |
|
3039 | deltaparentfn=self.deltaparent, | |
2972 | candeltafn=self.candelta, |
|
3040 | candeltafn=self.candelta, | |
2973 | rawsizefn=self.rawsize, |
|
3041 | rawsizefn=self.rawsize, | |
2974 | revdifffn=self.revdiff, |
|
3042 | revdifffn=self.revdiff, | |
2975 | flagsfn=self.flags, |
|
3043 | flagsfn=self.flags, | |
2976 | deltamode=deltamode, |
|
3044 | deltamode=deltamode, | |
2977 | revisiondata=revisiondata, |
|
3045 | revisiondata=revisiondata, | |
2978 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
3046 | assumehaveparentrevisions=assumehaveparentrevisions, | |
2979 | sidedata_helpers=sidedata_helpers, |
|
3047 | sidedata_helpers=sidedata_helpers, | |
2980 | debug_info=debug_info, |
|
3048 | debug_info=debug_info, | |
2981 | ) |
|
3049 | ) | |
2982 |
|
3050 | |||
2983 | DELTAREUSEALWAYS = b'always' |
|
3051 | DELTAREUSEALWAYS = b'always' | |
2984 | DELTAREUSESAMEREVS = b'samerevs' |
|
3052 | DELTAREUSESAMEREVS = b'samerevs' | |
2985 | DELTAREUSENEVER = b'never' |
|
3053 | DELTAREUSENEVER = b'never' | |
2986 |
|
3054 | |||
2987 | DELTAREUSEFULLADD = b'fulladd' |
|
3055 | DELTAREUSEFULLADD = b'fulladd' | |
2988 |
|
3056 | |||
2989 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} |
|
3057 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} | |
2990 |
|
3058 | |||
2991 | def clone( |
|
3059 | def clone( | |
2992 | self, |
|
3060 | self, | |
2993 | tr, |
|
3061 | tr, | |
2994 | destrevlog, |
|
3062 | destrevlog, | |
2995 | addrevisioncb=None, |
|
3063 | addrevisioncb=None, | |
2996 | deltareuse=DELTAREUSESAMEREVS, |
|
3064 | deltareuse=DELTAREUSESAMEREVS, | |
2997 | forcedeltabothparents=None, |
|
3065 | forcedeltabothparents=None, | |
2998 | sidedata_helpers=None, |
|
3066 | sidedata_helpers=None, | |
2999 | ): |
|
3067 | ): | |
3000 | """Copy this revlog to another, possibly with format changes. |
|
3068 | """Copy this revlog to another, possibly with format changes. | |
3001 |
|
3069 | |||
3002 | The destination revlog will contain the same revisions and nodes. |
|
3070 | The destination revlog will contain the same revisions and nodes. | |
3003 | However, it may not be bit-for-bit identical due to e.g. delta encoding |
|
3071 | However, it may not be bit-for-bit identical due to e.g. delta encoding | |
3004 | differences. |
|
3072 | differences. | |
3005 |
|
3073 | |||
3006 | The ``deltareuse`` argument control how deltas from the existing revlog |
|
3074 | The ``deltareuse`` argument control how deltas from the existing revlog | |
3007 | are preserved in the destination revlog. The argument can have the |
|
3075 | are preserved in the destination revlog. The argument can have the | |
3008 | following values: |
|
3076 | following values: | |
3009 |
|
3077 | |||
3010 | DELTAREUSEALWAYS |
|
3078 | DELTAREUSEALWAYS | |
3011 | Deltas will always be reused (if possible), even if the destination |
|
3079 | Deltas will always be reused (if possible), even if the destination | |
3012 | revlog would not select the same revisions for the delta. This is the |
|
3080 | revlog would not select the same revisions for the delta. This is the | |
3013 | fastest mode of operation. |
|
3081 | fastest mode of operation. | |
3014 | DELTAREUSESAMEREVS |
|
3082 | DELTAREUSESAMEREVS | |
3015 | Deltas will be reused if the destination revlog would pick the same |
|
3083 | Deltas will be reused if the destination revlog would pick the same | |
3016 | revisions for the delta. This mode strikes a balance between speed |
|
3084 | revisions for the delta. This mode strikes a balance between speed | |
3017 | and optimization. |
|
3085 | and optimization. | |
3018 | DELTAREUSENEVER |
|
3086 | DELTAREUSENEVER | |
3019 | Deltas will never be reused. This is the slowest mode of execution. |
|
3087 | Deltas will never be reused. This is the slowest mode of execution. | |
3020 | This mode can be used to recompute deltas (e.g. if the diff/delta |
|
3088 | This mode can be used to recompute deltas (e.g. if the diff/delta | |
3021 | algorithm changes). |
|
3089 | algorithm changes). | |
3022 | DELTAREUSEFULLADD |
|
3090 | DELTAREUSEFULLADD | |
3023 | Revision will be re-added as if their were new content. This is |
|
3091 | Revision will be re-added as if their were new content. This is | |
3024 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. |
|
3092 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. | |
3025 | eg: large file detection and handling. |
|
3093 | eg: large file detection and handling. | |
3026 |
|
3094 | |||
3027 | Delta computation can be slow, so the choice of delta reuse policy can |
|
3095 | Delta computation can be slow, so the choice of delta reuse policy can | |
3028 | significantly affect run time. |
|
3096 | significantly affect run time. | |
3029 |
|
3097 | |||
3030 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between |
|
3098 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between | |
3031 | two extremes. Deltas will be reused if they are appropriate. But if the |
|
3099 | two extremes. Deltas will be reused if they are appropriate. But if the | |
3032 | delta could choose a better revision, it will do so. This means if you |
|
3100 | delta could choose a better revision, it will do so. This means if you | |
3033 | are converting a non-generaldelta revlog to a generaldelta revlog, |
|
3101 | are converting a non-generaldelta revlog to a generaldelta revlog, | |
3034 | deltas will be recomputed if the delta's parent isn't a parent of the |
|
3102 | deltas will be recomputed if the delta's parent isn't a parent of the | |
3035 | revision. |
|
3103 | revision. | |
3036 |
|
3104 | |||
3037 | In addition to the delta policy, the ``forcedeltabothparents`` |
|
3105 | In addition to the delta policy, the ``forcedeltabothparents`` | |
3038 | argument controls whether to force compute deltas against both parents |
|
3106 | argument controls whether to force compute deltas against both parents | |
3039 | for merges. By default, the current default is used. |
|
3107 | for merges. By default, the current default is used. | |
3040 |
|
3108 | |||
3041 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on |
|
3109 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on | |
3042 | `sidedata_helpers`. |
|
3110 | `sidedata_helpers`. | |
3043 | """ |
|
3111 | """ | |
3044 | if deltareuse not in self.DELTAREUSEALL: |
|
3112 | if deltareuse not in self.DELTAREUSEALL: | |
3045 | raise ValueError( |
|
3113 | raise ValueError( | |
3046 | _(b'value for deltareuse invalid: %s') % deltareuse |
|
3114 | _(b'value for deltareuse invalid: %s') % deltareuse | |
3047 | ) |
|
3115 | ) | |
3048 |
|
3116 | |||
3049 | if len(destrevlog): |
|
3117 | if len(destrevlog): | |
3050 | raise ValueError(_(b'destination revlog is not empty')) |
|
3118 | raise ValueError(_(b'destination revlog is not empty')) | |
3051 |
|
3119 | |||
3052 | if getattr(self, 'filteredrevs', None): |
|
3120 | if getattr(self, 'filteredrevs', None): | |
3053 | raise ValueError(_(b'source revlog has filtered revisions')) |
|
3121 | raise ValueError(_(b'source revlog has filtered revisions')) | |
3054 | if getattr(destrevlog, 'filteredrevs', None): |
|
3122 | if getattr(destrevlog, 'filteredrevs', None): | |
3055 | raise ValueError(_(b'destination revlog has filtered revisions')) |
|
3123 | raise ValueError(_(b'destination revlog has filtered revisions')) | |
3056 |
|
3124 | |||
3057 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, |
|
3125 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, | |
3058 | # if possible. |
|
3126 | # if possible. | |
3059 | oldlazydelta = destrevlog._lazydelta |
|
3127 | oldlazydelta = destrevlog._lazydelta | |
3060 | oldlazydeltabase = destrevlog._lazydeltabase |
|
3128 | oldlazydeltabase = destrevlog._lazydeltabase | |
3061 | oldamd = destrevlog._deltabothparents |
|
3129 | oldamd = destrevlog._deltabothparents | |
3062 |
|
3130 | |||
3063 | try: |
|
3131 | try: | |
3064 | if deltareuse == self.DELTAREUSEALWAYS: |
|
3132 | if deltareuse == self.DELTAREUSEALWAYS: | |
3065 | destrevlog._lazydeltabase = True |
|
3133 | destrevlog._lazydeltabase = True | |
3066 | destrevlog._lazydelta = True |
|
3134 | destrevlog._lazydelta = True | |
3067 | elif deltareuse == self.DELTAREUSESAMEREVS: |
|
3135 | elif deltareuse == self.DELTAREUSESAMEREVS: | |
3068 | destrevlog._lazydeltabase = False |
|
3136 | destrevlog._lazydeltabase = False | |
3069 | destrevlog._lazydelta = True |
|
3137 | destrevlog._lazydelta = True | |
3070 | elif deltareuse == self.DELTAREUSENEVER: |
|
3138 | elif deltareuse == self.DELTAREUSENEVER: | |
3071 | destrevlog._lazydeltabase = False |
|
3139 | destrevlog._lazydeltabase = False | |
3072 | destrevlog._lazydelta = False |
|
3140 | destrevlog._lazydelta = False | |
3073 |
|
3141 | |||
3074 | destrevlog._deltabothparents = forcedeltabothparents or oldamd |
|
3142 | destrevlog._deltabothparents = forcedeltabothparents or oldamd | |
3075 |
|
3143 | |||
3076 | self._clone( |
|
3144 | self._clone( | |
3077 | tr, |
|
3145 | tr, | |
3078 | destrevlog, |
|
3146 | destrevlog, | |
3079 | addrevisioncb, |
|
3147 | addrevisioncb, | |
3080 | deltareuse, |
|
3148 | deltareuse, | |
3081 | forcedeltabothparents, |
|
3149 | forcedeltabothparents, | |
3082 | sidedata_helpers, |
|
3150 | sidedata_helpers, | |
3083 | ) |
|
3151 | ) | |
3084 |
|
3152 | |||
3085 | finally: |
|
3153 | finally: | |
3086 | destrevlog._lazydelta = oldlazydelta |
|
3154 | destrevlog._lazydelta = oldlazydelta | |
3087 | destrevlog._lazydeltabase = oldlazydeltabase |
|
3155 | destrevlog._lazydeltabase = oldlazydeltabase | |
3088 | destrevlog._deltabothparents = oldamd |
|
3156 | destrevlog._deltabothparents = oldamd | |
3089 |
|
3157 | |||
3090 | def _clone( |
|
3158 | def _clone( | |
3091 | self, |
|
3159 | self, | |
3092 | tr, |
|
3160 | tr, | |
3093 | destrevlog, |
|
3161 | destrevlog, | |
3094 | addrevisioncb, |
|
3162 | addrevisioncb, | |
3095 | deltareuse, |
|
3163 | deltareuse, | |
3096 | forcedeltabothparents, |
|
3164 | forcedeltabothparents, | |
3097 | sidedata_helpers, |
|
3165 | sidedata_helpers, | |
3098 | ): |
|
3166 | ): | |
3099 | """perform the core duty of `revlog.clone` after parameter processing""" |
|
3167 | """perform the core duty of `revlog.clone` after parameter processing""" | |
3100 | write_debug = None |
|
3168 | write_debug = None | |
3101 | if self._debug_delta: |
|
3169 | if self._debug_delta: | |
3102 | write_debug = tr._report |
|
3170 | write_debug = tr._report | |
3103 | deltacomputer = deltautil.deltacomputer( |
|
3171 | deltacomputer = deltautil.deltacomputer( | |
3104 | destrevlog, |
|
3172 | destrevlog, | |
3105 | write_debug=write_debug, |
|
3173 | write_debug=write_debug, | |
3106 | ) |
|
3174 | ) | |
3107 | index = self.index |
|
3175 | index = self.index | |
3108 | for rev in self: |
|
3176 | for rev in self: | |
3109 | entry = index[rev] |
|
3177 | entry = index[rev] | |
3110 |
|
3178 | |||
3111 | # Some classes override linkrev to take filtered revs into |
|
3179 | # Some classes override linkrev to take filtered revs into | |
3112 | # account. Use raw entry from index. |
|
3180 | # account. Use raw entry from index. | |
3113 | flags = entry[0] & 0xFFFF |
|
3181 | flags = entry[0] & 0xFFFF | |
3114 | linkrev = entry[4] |
|
3182 | linkrev = entry[4] | |
3115 | p1 = index[entry[5]][7] |
|
3183 | p1 = index[entry[5]][7] | |
3116 | p2 = index[entry[6]][7] |
|
3184 | p2 = index[entry[6]][7] | |
3117 | node = entry[7] |
|
3185 | node = entry[7] | |
3118 |
|
3186 | |||
3119 | # (Possibly) reuse the delta from the revlog if allowed and |
|
3187 | # (Possibly) reuse the delta from the revlog if allowed and | |
3120 | # the revlog chunk is a delta. |
|
3188 | # the revlog chunk is a delta. | |
3121 | cachedelta = None |
|
3189 | cachedelta = None | |
3122 | rawtext = None |
|
3190 | rawtext = None | |
3123 | if deltareuse == self.DELTAREUSEFULLADD: |
|
3191 | if deltareuse == self.DELTAREUSEFULLADD: | |
3124 | text = self._revisiondata(rev) |
|
3192 | text = self._revisiondata(rev) | |
3125 | sidedata = self.sidedata(rev) |
|
3193 | sidedata = self.sidedata(rev) | |
3126 |
|
3194 | |||
3127 | if sidedata_helpers is not None: |
|
3195 | if sidedata_helpers is not None: | |
3128 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
3196 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
3129 | self, sidedata_helpers, sidedata, rev |
|
3197 | self, sidedata_helpers, sidedata, rev | |
3130 | ) |
|
3198 | ) | |
3131 | flags = flags | new_flags[0] & ~new_flags[1] |
|
3199 | flags = flags | new_flags[0] & ~new_flags[1] | |
3132 |
|
3200 | |||
3133 | destrevlog.addrevision( |
|
3201 | destrevlog.addrevision( | |
3134 | text, |
|
3202 | text, | |
3135 | tr, |
|
3203 | tr, | |
3136 | linkrev, |
|
3204 | linkrev, | |
3137 | p1, |
|
3205 | p1, | |
3138 | p2, |
|
3206 | p2, | |
3139 | cachedelta=cachedelta, |
|
3207 | cachedelta=cachedelta, | |
3140 | node=node, |
|
3208 | node=node, | |
3141 | flags=flags, |
|
3209 | flags=flags, | |
3142 | deltacomputer=deltacomputer, |
|
3210 | deltacomputer=deltacomputer, | |
3143 | sidedata=sidedata, |
|
3211 | sidedata=sidedata, | |
3144 | ) |
|
3212 | ) | |
3145 | else: |
|
3213 | else: | |
3146 | if destrevlog._lazydelta: |
|
3214 | if destrevlog._lazydelta: | |
3147 | dp = self.deltaparent(rev) |
|
3215 | dp = self.deltaparent(rev) | |
3148 | if dp != nullrev: |
|
3216 | if dp != nullrev: | |
3149 | cachedelta = (dp, bytes(self._chunk(rev))) |
|
3217 | cachedelta = (dp, bytes(self._chunk(rev))) | |
3150 |
|
3218 | |||
3151 | sidedata = None |
|
3219 | sidedata = None | |
3152 | if not cachedelta: |
|
3220 | if not cachedelta: | |
3153 | rawtext = self._revisiondata(rev) |
|
3221 | rawtext = self._revisiondata(rev) | |
3154 | sidedata = self.sidedata(rev) |
|
3222 | sidedata = self.sidedata(rev) | |
3155 | if sidedata is None: |
|
3223 | if sidedata is None: | |
3156 | sidedata = self.sidedata(rev) |
|
3224 | sidedata = self.sidedata(rev) | |
3157 |
|
3225 | |||
3158 | if sidedata_helpers is not None: |
|
3226 | if sidedata_helpers is not None: | |
3159 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
3227 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
3160 | self, sidedata_helpers, sidedata, rev |
|
3228 | self, sidedata_helpers, sidedata, rev | |
3161 | ) |
|
3229 | ) | |
3162 | flags = flags | new_flags[0] & ~new_flags[1] |
|
3230 | flags = flags | new_flags[0] & ~new_flags[1] | |
3163 |
|
3231 | |||
3164 | with destrevlog._writing(tr): |
|
3232 | with destrevlog._writing(tr): | |
3165 | destrevlog._addrevision( |
|
3233 | destrevlog._addrevision( | |
3166 | node, |
|
3234 | node, | |
3167 | rawtext, |
|
3235 | rawtext, | |
3168 | tr, |
|
3236 | tr, | |
3169 | linkrev, |
|
3237 | linkrev, | |
3170 | p1, |
|
3238 | p1, | |
3171 | p2, |
|
3239 | p2, | |
3172 | flags, |
|
3240 | flags, | |
3173 | cachedelta, |
|
3241 | cachedelta, | |
3174 | deltacomputer=deltacomputer, |
|
3242 | deltacomputer=deltacomputer, | |
3175 | sidedata=sidedata, |
|
3243 | sidedata=sidedata, | |
3176 | ) |
|
3244 | ) | |
3177 |
|
3245 | |||
3178 | if addrevisioncb: |
|
3246 | if addrevisioncb: | |
3179 | addrevisioncb(self, rev, node) |
|
3247 | addrevisioncb(self, rev, node) | |
3180 |
|
3248 | |||
3181 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
3249 | def censorrevision(self, tr, censornode, tombstone=b''): | |
3182 | if self._format_version == REVLOGV0: |
|
3250 | if self._format_version == REVLOGV0: | |
3183 | raise error.RevlogError( |
|
3251 | raise error.RevlogError( | |
3184 | _(b'cannot censor with version %d revlogs') |
|
3252 | _(b'cannot censor with version %d revlogs') | |
3185 | % self._format_version |
|
3253 | % self._format_version | |
3186 | ) |
|
3254 | ) | |
3187 | elif self._format_version == REVLOGV1: |
|
3255 | elif self._format_version == REVLOGV1: | |
3188 | rewrite.v1_censor(self, tr, censornode, tombstone) |
|
3256 | rewrite.v1_censor(self, tr, censornode, tombstone) | |
3189 | else: |
|
3257 | else: | |
3190 | rewrite.v2_censor(self, tr, censornode, tombstone) |
|
3258 | rewrite.v2_censor(self, tr, censornode, tombstone) | |
3191 |
|
3259 | |||
3192 | def verifyintegrity(self, state): |
|
3260 | def verifyintegrity(self, state): | |
3193 | """Verifies the integrity of the revlog. |
|
3261 | """Verifies the integrity of the revlog. | |
3194 |
|
3262 | |||
3195 | Yields ``revlogproblem`` instances describing problems that are |
|
3263 | Yields ``revlogproblem`` instances describing problems that are | |
3196 | found. |
|
3264 | found. | |
3197 | """ |
|
3265 | """ | |
3198 | dd, di = self.checksize() |
|
3266 | dd, di = self.checksize() | |
3199 | if dd: |
|
3267 | if dd: | |
3200 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) |
|
3268 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) | |
3201 | if di: |
|
3269 | if di: | |
3202 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) |
|
3270 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) | |
3203 |
|
3271 | |||
3204 | version = self._format_version |
|
3272 | version = self._format_version | |
3205 |
|
3273 | |||
3206 | # The verifier tells us what version revlog we should be. |
|
3274 | # The verifier tells us what version revlog we should be. | |
3207 | if version != state[b'expectedversion']: |
|
3275 | if version != state[b'expectedversion']: | |
3208 | yield revlogproblem( |
|
3276 | yield revlogproblem( | |
3209 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") |
|
3277 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") | |
3210 | % (self.display_id, version, state[b'expectedversion']) |
|
3278 | % (self.display_id, version, state[b'expectedversion']) | |
3211 | ) |
|
3279 | ) | |
3212 |
|
3280 | |||
3213 | state[b'skipread'] = set() |
|
3281 | state[b'skipread'] = set() | |
3214 | state[b'safe_renamed'] = set() |
|
3282 | state[b'safe_renamed'] = set() | |
3215 |
|
3283 | |||
3216 | for rev in self: |
|
3284 | for rev in self: | |
3217 | node = self.node(rev) |
|
3285 | node = self.node(rev) | |
3218 |
|
3286 | |||
3219 | # Verify contents. 4 cases to care about: |
|
3287 | # Verify contents. 4 cases to care about: | |
3220 | # |
|
3288 | # | |
3221 | # common: the most common case |
|
3289 | # common: the most common case | |
3222 | # rename: with a rename |
|
3290 | # rename: with a rename | |
3223 | # meta: file content starts with b'\1\n', the metadata |
|
3291 | # meta: file content starts with b'\1\n', the metadata | |
3224 | # header defined in filelog.py, but without a rename |
|
3292 | # header defined in filelog.py, but without a rename | |
3225 | # ext: content stored externally |
|
3293 | # ext: content stored externally | |
3226 | # |
|
3294 | # | |
3227 | # More formally, their differences are shown below: |
|
3295 | # More formally, their differences are shown below: | |
3228 | # |
|
3296 | # | |
3229 | # | common | rename | meta | ext |
|
3297 | # | common | rename | meta | ext | |
3230 | # ------------------------------------------------------- |
|
3298 | # ------------------------------------------------------- | |
3231 | # flags() | 0 | 0 | 0 | not 0 |
|
3299 | # flags() | 0 | 0 | 0 | not 0 | |
3232 | # renamed() | False | True | False | ? |
|
3300 | # renamed() | False | True | False | ? | |
3233 | # rawtext[0:2]=='\1\n'| False | True | True | ? |
|
3301 | # rawtext[0:2]=='\1\n'| False | True | True | ? | |
3234 | # |
|
3302 | # | |
3235 | # "rawtext" means the raw text stored in revlog data, which |
|
3303 | # "rawtext" means the raw text stored in revlog data, which | |
3236 | # could be retrieved by "rawdata(rev)". "text" |
|
3304 | # could be retrieved by "rawdata(rev)". "text" | |
3237 | # mentioned below is "revision(rev)". |
|
3305 | # mentioned below is "revision(rev)". | |
3238 | # |
|
3306 | # | |
3239 | # There are 3 different lengths stored physically: |
|
3307 | # There are 3 different lengths stored physically: | |
3240 | # 1. L1: rawsize, stored in revlog index |
|
3308 | # 1. L1: rawsize, stored in revlog index | |
3241 | # 2. L2: len(rawtext), stored in revlog data |
|
3309 | # 2. L2: len(rawtext), stored in revlog data | |
3242 | # 3. L3: len(text), stored in revlog data if flags==0, or |
|
3310 | # 3. L3: len(text), stored in revlog data if flags==0, or | |
3243 | # possibly somewhere else if flags!=0 |
|
3311 | # possibly somewhere else if flags!=0 | |
3244 | # |
|
3312 | # | |
3245 | # L1 should be equal to L2. L3 could be different from them. |
|
3313 | # L1 should be equal to L2. L3 could be different from them. | |
3246 | # "text" may or may not affect commit hash depending on flag |
|
3314 | # "text" may or may not affect commit hash depending on flag | |
3247 | # processors (see flagutil.addflagprocessor). |
|
3315 | # processors (see flagutil.addflagprocessor). | |
3248 | # |
|
3316 | # | |
3249 | # | common | rename | meta | ext |
|
3317 | # | common | rename | meta | ext | |
3250 | # ------------------------------------------------- |
|
3318 | # ------------------------------------------------- | |
3251 | # rawsize() | L1 | L1 | L1 | L1 |
|
3319 | # rawsize() | L1 | L1 | L1 | L1 | |
3252 | # size() | L1 | L2-LM | L1(*) | L1 (?) |
|
3320 | # size() | L1 | L2-LM | L1(*) | L1 (?) | |
3253 | # len(rawtext) | L2 | L2 | L2 | L2 |
|
3321 | # len(rawtext) | L2 | L2 | L2 | L2 | |
3254 | # len(text) | L2 | L2 | L2 | L3 |
|
3322 | # len(text) | L2 | L2 | L2 | L3 | |
3255 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) |
|
3323 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) | |
3256 | # |
|
3324 | # | |
3257 | # LM: length of metadata, depending on rawtext |
|
3325 | # LM: length of metadata, depending on rawtext | |
3258 | # (*): not ideal, see comment in filelog.size |
|
3326 | # (*): not ideal, see comment in filelog.size | |
3259 | # (?): could be "- len(meta)" if the resolved content has |
|
3327 | # (?): could be "- len(meta)" if the resolved content has | |
3260 | # rename metadata |
|
3328 | # rename metadata | |
3261 | # |
|
3329 | # | |
3262 | # Checks needed to be done: |
|
3330 | # Checks needed to be done: | |
3263 | # 1. length check: L1 == L2, in all cases. |
|
3331 | # 1. length check: L1 == L2, in all cases. | |
3264 | # 2. hash check: depending on flag processor, we may need to |
|
3332 | # 2. hash check: depending on flag processor, we may need to | |
3265 | # use either "text" (external), or "rawtext" (in revlog). |
|
3333 | # use either "text" (external), or "rawtext" (in revlog). | |
3266 |
|
3334 | |||
3267 | try: |
|
3335 | try: | |
3268 | skipflags = state.get(b'skipflags', 0) |
|
3336 | skipflags = state.get(b'skipflags', 0) | |
3269 | if skipflags: |
|
3337 | if skipflags: | |
3270 | skipflags &= self.flags(rev) |
|
3338 | skipflags &= self.flags(rev) | |
3271 |
|
3339 | |||
3272 | _verify_revision(self, skipflags, state, node) |
|
3340 | _verify_revision(self, skipflags, state, node) | |
3273 |
|
3341 | |||
3274 | l1 = self.rawsize(rev) |
|
3342 | l1 = self.rawsize(rev) | |
3275 | l2 = len(self.rawdata(node)) |
|
3343 | l2 = len(self.rawdata(node)) | |
3276 |
|
3344 | |||
3277 | if l1 != l2: |
|
3345 | if l1 != l2: | |
3278 | yield revlogproblem( |
|
3346 | yield revlogproblem( | |
3279 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), |
|
3347 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), | |
3280 | node=node, |
|
3348 | node=node, | |
3281 | ) |
|
3349 | ) | |
3282 |
|
3350 | |||
3283 | except error.CensoredNodeError: |
|
3351 | except error.CensoredNodeError: | |
3284 | if state[b'erroroncensored']: |
|
3352 | if state[b'erroroncensored']: | |
3285 | yield revlogproblem( |
|
3353 | yield revlogproblem( | |
3286 | error=_(b'censored file data'), node=node |
|
3354 | error=_(b'censored file data'), node=node | |
3287 | ) |
|
3355 | ) | |
3288 | state[b'skipread'].add(node) |
|
3356 | state[b'skipread'].add(node) | |
3289 | except Exception as e: |
|
3357 | except Exception as e: | |
3290 | yield revlogproblem( |
|
3358 | yield revlogproblem( | |
3291 | error=_(b'unpacking %s: %s') |
|
3359 | error=_(b'unpacking %s: %s') | |
3292 | % (short(node), stringutil.forcebytestr(e)), |
|
3360 | % (short(node), stringutil.forcebytestr(e)), | |
3293 | node=node, |
|
3361 | node=node, | |
3294 | ) |
|
3362 | ) | |
3295 | state[b'skipread'].add(node) |
|
3363 | state[b'skipread'].add(node) | |
3296 |
|
3364 | |||
3297 | def storageinfo( |
|
3365 | def storageinfo( | |
3298 | self, |
|
3366 | self, | |
3299 | exclusivefiles=False, |
|
3367 | exclusivefiles=False, | |
3300 | sharedfiles=False, |
|
3368 | sharedfiles=False, | |
3301 | revisionscount=False, |
|
3369 | revisionscount=False, | |
3302 | trackedsize=False, |
|
3370 | trackedsize=False, | |
3303 | storedsize=False, |
|
3371 | storedsize=False, | |
3304 | ): |
|
3372 | ): | |
3305 | d = {} |
|
3373 | d = {} | |
3306 |
|
3374 | |||
3307 | if exclusivefiles: |
|
3375 | if exclusivefiles: | |
3308 | d[b'exclusivefiles'] = [(self.opener, self._indexfile)] |
|
3376 | d[b'exclusivefiles'] = [(self.opener, self._indexfile)] | |
3309 | if not self._inline: |
|
3377 | if not self._inline: | |
3310 | d[b'exclusivefiles'].append((self.opener, self._datafile)) |
|
3378 | d[b'exclusivefiles'].append((self.opener, self._datafile)) | |
3311 |
|
3379 | |||
3312 | if sharedfiles: |
|
3380 | if sharedfiles: | |
3313 | d[b'sharedfiles'] = [] |
|
3381 | d[b'sharedfiles'] = [] | |
3314 |
|
3382 | |||
3315 | if revisionscount: |
|
3383 | if revisionscount: | |
3316 | d[b'revisionscount'] = len(self) |
|
3384 | d[b'revisionscount'] = len(self) | |
3317 |
|
3385 | |||
3318 | if trackedsize: |
|
3386 | if trackedsize: | |
3319 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) |
|
3387 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) | |
3320 |
|
3388 | |||
3321 | if storedsize: |
|
3389 | if storedsize: | |
3322 | d[b'storedsize'] = sum( |
|
3390 | d[b'storedsize'] = sum( | |
3323 | self.opener.stat(path).st_size for path in self.files() |
|
3391 | self.opener.stat(path).st_size for path in self.files() | |
3324 | ) |
|
3392 | ) | |
3325 |
|
3393 | |||
3326 | return d |
|
3394 | return d | |
3327 |
|
3395 | |||
3328 | def rewrite_sidedata(self, transaction, helpers, startrev, endrev): |
|
3396 | def rewrite_sidedata(self, transaction, helpers, startrev, endrev): | |
3329 | if not self.hassidedata: |
|
3397 | if not self.hassidedata: | |
3330 | return |
|
3398 | return | |
3331 | # revlog formats with sidedata support does not support inline |
|
3399 | # revlog formats with sidedata support does not support inline | |
3332 | assert not self._inline |
|
3400 | assert not self._inline | |
3333 | if not helpers[1] and not helpers[2]: |
|
3401 | if not helpers[1] and not helpers[2]: | |
3334 | # Nothing to generate or remove |
|
3402 | # Nothing to generate or remove | |
3335 | return |
|
3403 | return | |
3336 |
|
3404 | |||
3337 | new_entries = [] |
|
3405 | new_entries = [] | |
3338 | # append the new sidedata |
|
3406 | # append the new sidedata | |
3339 | with self._writing(transaction): |
|
3407 | with self._writing(transaction): | |
3340 | ifh, dfh, sdfh = self._writinghandles |
|
3408 | ifh, dfh, sdfh = self._writinghandles | |
3341 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) |
|
3409 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) | |
3342 |
|
3410 | |||
3343 | current_offset = sdfh.tell() |
|
3411 | current_offset = sdfh.tell() | |
3344 | for rev in range(startrev, endrev + 1): |
|
3412 | for rev in range(startrev, endrev + 1): | |
3345 | entry = self.index[rev] |
|
3413 | entry = self.index[rev] | |
3346 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( |
|
3414 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( | |
3347 | store=self, |
|
3415 | store=self, | |
3348 | sidedata_helpers=helpers, |
|
3416 | sidedata_helpers=helpers, | |
3349 | sidedata={}, |
|
3417 | sidedata={}, | |
3350 | rev=rev, |
|
3418 | rev=rev, | |
3351 | ) |
|
3419 | ) | |
3352 |
|
3420 | |||
3353 | serialized_sidedata = sidedatautil.serialize_sidedata( |
|
3421 | serialized_sidedata = sidedatautil.serialize_sidedata( | |
3354 | new_sidedata |
|
3422 | new_sidedata | |
3355 | ) |
|
3423 | ) | |
3356 |
|
3424 | |||
3357 | sidedata_compression_mode = COMP_MODE_INLINE |
|
3425 | sidedata_compression_mode = COMP_MODE_INLINE | |
3358 | if serialized_sidedata and self.hassidedata: |
|
3426 | if serialized_sidedata and self.hassidedata: | |
3359 | sidedata_compression_mode = COMP_MODE_PLAIN |
|
3427 | sidedata_compression_mode = COMP_MODE_PLAIN | |
3360 | h, comp_sidedata = self.compress(serialized_sidedata) |
|
3428 | h, comp_sidedata = self.compress(serialized_sidedata) | |
3361 | if ( |
|
3429 | if ( | |
3362 | h != b'u' |
|
3430 | h != b'u' | |
3363 | and comp_sidedata[0] != b'\0' |
|
3431 | and comp_sidedata[0] != b'\0' | |
3364 | and len(comp_sidedata) < len(serialized_sidedata) |
|
3432 | and len(comp_sidedata) < len(serialized_sidedata) | |
3365 | ): |
|
3433 | ): | |
3366 | assert not h |
|
3434 | assert not h | |
3367 | if ( |
|
3435 | if ( | |
3368 | comp_sidedata[0] |
|
3436 | comp_sidedata[0] | |
3369 | == self._docket.default_compression_header |
|
3437 | == self._docket.default_compression_header | |
3370 | ): |
|
3438 | ): | |
3371 | sidedata_compression_mode = COMP_MODE_DEFAULT |
|
3439 | sidedata_compression_mode = COMP_MODE_DEFAULT | |
3372 | serialized_sidedata = comp_sidedata |
|
3440 | serialized_sidedata = comp_sidedata | |
3373 | else: |
|
3441 | else: | |
3374 | sidedata_compression_mode = COMP_MODE_INLINE |
|
3442 | sidedata_compression_mode = COMP_MODE_INLINE | |
3375 | serialized_sidedata = comp_sidedata |
|
3443 | serialized_sidedata = comp_sidedata | |
3376 | if entry[8] != 0 or entry[9] != 0: |
|
3444 | if entry[8] != 0 or entry[9] != 0: | |
3377 | # rewriting entries that already have sidedata is not |
|
3445 | # rewriting entries that already have sidedata is not | |
3378 | # supported yet, because it introduces garbage data in the |
|
3446 | # supported yet, because it introduces garbage data in the | |
3379 | # revlog. |
|
3447 | # revlog. | |
3380 | msg = b"rewriting existing sidedata is not supported yet" |
|
3448 | msg = b"rewriting existing sidedata is not supported yet" | |
3381 | raise error.Abort(msg) |
|
3449 | raise error.Abort(msg) | |
3382 |
|
3450 | |||
3383 | # Apply (potential) flags to add and to remove after running |
|
3451 | # Apply (potential) flags to add and to remove after running | |
3384 | # the sidedata helpers |
|
3452 | # the sidedata helpers | |
3385 | new_offset_flags = entry[0] | flags[0] & ~flags[1] |
|
3453 | new_offset_flags = entry[0] | flags[0] & ~flags[1] | |
3386 | entry_update = ( |
|
3454 | entry_update = ( | |
3387 | current_offset, |
|
3455 | current_offset, | |
3388 | len(serialized_sidedata), |
|
3456 | len(serialized_sidedata), | |
3389 | new_offset_flags, |
|
3457 | new_offset_flags, | |
3390 | sidedata_compression_mode, |
|
3458 | sidedata_compression_mode, | |
3391 | ) |
|
3459 | ) | |
3392 |
|
3460 | |||
3393 | # the sidedata computation might have move the file cursors around |
|
3461 | # the sidedata computation might have move the file cursors around | |
3394 | sdfh.seek(current_offset, os.SEEK_SET) |
|
3462 | sdfh.seek(current_offset, os.SEEK_SET) | |
3395 | sdfh.write(serialized_sidedata) |
|
3463 | sdfh.write(serialized_sidedata) | |
3396 | new_entries.append(entry_update) |
|
3464 | new_entries.append(entry_update) | |
3397 | current_offset += len(serialized_sidedata) |
|
3465 | current_offset += len(serialized_sidedata) | |
3398 | self._docket.sidedata_end = sdfh.tell() |
|
3466 | self._docket.sidedata_end = sdfh.tell() | |
3399 |
|
3467 | |||
3400 | # rewrite the new index entries |
|
3468 | # rewrite the new index entries | |
3401 | ifh.seek(startrev * self.index.entry_size) |
|
3469 | ifh.seek(startrev * self.index.entry_size) | |
3402 | for i, e in enumerate(new_entries): |
|
3470 | for i, e in enumerate(new_entries): | |
3403 | rev = startrev + i |
|
3471 | rev = startrev + i | |
3404 | self.index.replace_sidedata_info(rev, *e) |
|
3472 | self.index.replace_sidedata_info(rev, *e) | |
3405 | packed = self.index.entry_binary(rev) |
|
3473 | packed = self.index.entry_binary(rev) | |
3406 | if rev == 0 and self._docket is None: |
|
3474 | if rev == 0 and self._docket is None: | |
3407 | header = self._format_flags | self._format_version |
|
3475 | header = self._format_flags | self._format_version | |
3408 | header = self.index.pack_header(header) |
|
3476 | header = self.index.pack_header(header) | |
3409 | packed = header + packed |
|
3477 | packed = header + packed | |
3410 | ifh.write(packed) |
|
3478 | ifh.write(packed) |
@@ -1,1159 +1,1216 b'' | |||||
1 | # store.py - repository store handling for Mercurial |
|
1 | # store.py - repository store handling for Mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2008 Olivia Mackall <olivia@selenic.com> |
|
3 | # Copyright 2008 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 | import collections |
|
8 | import collections | |
9 | import functools |
|
9 | import functools | |
10 | import os |
|
10 | import os | |
11 | import re |
|
11 | import re | |
12 | import stat |
|
12 | import stat | |
13 | from typing import Generator, List |
|
13 | from typing import Generator, List | |
14 |
|
14 | |||
15 | from .i18n import _ |
|
15 | from .i18n import _ | |
16 | from .pycompat import getattr |
|
16 | from .pycompat import getattr | |
17 | from .thirdparty import attr |
|
17 | from .thirdparty import attr | |
18 | from .node import hex |
|
18 | from .node import hex | |
19 | from . import ( |
|
19 | from . import ( | |
20 | changelog, |
|
20 | changelog, | |
21 | error, |
|
21 | error, | |
22 | filelog, |
|
22 | filelog, | |
23 | manifest, |
|
23 | manifest, | |
24 | policy, |
|
24 | policy, | |
25 | pycompat, |
|
25 | pycompat, | |
26 | util, |
|
26 | util, | |
27 | vfs as vfsmod, |
|
27 | vfs as vfsmod, | |
28 | ) |
|
28 | ) | |
29 | from .utils import hashutil |
|
29 | from .utils import hashutil | |
30 |
|
30 | |||
31 | parsers = policy.importmod('parsers') |
|
31 | parsers = policy.importmod('parsers') | |
32 | # how much bytes should be read from fncache in one read |
|
32 | # how much bytes should be read from fncache in one read | |
33 | # It is done to prevent loading large fncache files into memory |
|
33 | # It is done to prevent loading large fncache files into memory | |
34 | fncache_chunksize = 10 ** 6 |
|
34 | fncache_chunksize = 10 ** 6 | |
35 |
|
35 | |||
36 |
|
36 | |||
37 | def _match_tracked_entry(entry, matcher): |
|
37 | def _match_tracked_entry(entry, matcher): | |
38 | """parses a fncache entry and returns whether the entry is tracking a path |
|
38 | """parses a fncache entry and returns whether the entry is tracking a path | |
39 | matched by matcher or not. |
|
39 | matched by matcher or not. | |
40 |
|
40 | |||
41 | If matcher is None, returns True""" |
|
41 | If matcher is None, returns True""" | |
42 |
|
42 | |||
43 | if matcher is None: |
|
43 | if matcher is None: | |
44 | return True |
|
44 | return True | |
45 | if entry.is_filelog: |
|
45 | if entry.is_filelog: | |
46 | return matcher(entry.target_id) |
|
46 | return matcher(entry.target_id) | |
47 | elif entry.is_manifestlog: |
|
47 | elif entry.is_manifestlog: | |
48 | return matcher.visitdir(entry.target_id.rstrip(b'/')) |
|
48 | return matcher.visitdir(entry.target_id.rstrip(b'/')) | |
49 | raise error.ProgrammingError(b"cannot process entry %r" % entry) |
|
49 | raise error.ProgrammingError(b"cannot process entry %r" % entry) | |
50 |
|
50 | |||
51 |
|
51 | |||
52 | # This avoids a collision between a file named foo and a dir named |
|
52 | # This avoids a collision between a file named foo and a dir named | |
53 | # foo.i or foo.d |
|
53 | # foo.i or foo.d | |
54 | def _encodedir(path): |
|
54 | def _encodedir(path): | |
55 | """ |
|
55 | """ | |
56 | >>> _encodedir(b'data/foo.i') |
|
56 | >>> _encodedir(b'data/foo.i') | |
57 | 'data/foo.i' |
|
57 | 'data/foo.i' | |
58 | >>> _encodedir(b'data/foo.i/bla.i') |
|
58 | >>> _encodedir(b'data/foo.i/bla.i') | |
59 | 'data/foo.i.hg/bla.i' |
|
59 | 'data/foo.i.hg/bla.i' | |
60 | >>> _encodedir(b'data/foo.i.hg/bla.i') |
|
60 | >>> _encodedir(b'data/foo.i.hg/bla.i') | |
61 | 'data/foo.i.hg.hg/bla.i' |
|
61 | 'data/foo.i.hg.hg/bla.i' | |
62 | >>> _encodedir(b'data/foo.i\\ndata/foo.i/bla.i\\ndata/foo.i.hg/bla.i\\n') |
|
62 | >>> _encodedir(b'data/foo.i\\ndata/foo.i/bla.i\\ndata/foo.i.hg/bla.i\\n') | |
63 | 'data/foo.i\\ndata/foo.i.hg/bla.i\\ndata/foo.i.hg.hg/bla.i\\n' |
|
63 | 'data/foo.i\\ndata/foo.i.hg/bla.i\\ndata/foo.i.hg.hg/bla.i\\n' | |
64 | """ |
|
64 | """ | |
65 | return ( |
|
65 | return ( | |
66 | path.replace(b".hg/", b".hg.hg/") |
|
66 | path.replace(b".hg/", b".hg.hg/") | |
67 | .replace(b".i/", b".i.hg/") |
|
67 | .replace(b".i/", b".i.hg/") | |
68 | .replace(b".d/", b".d.hg/") |
|
68 | .replace(b".d/", b".d.hg/") | |
69 | ) |
|
69 | ) | |
70 |
|
70 | |||
71 |
|
71 | |||
72 | encodedir = getattr(parsers, 'encodedir', _encodedir) |
|
72 | encodedir = getattr(parsers, 'encodedir', _encodedir) | |
73 |
|
73 | |||
74 |
|
74 | |||
75 | def decodedir(path): |
|
75 | def decodedir(path): | |
76 | """ |
|
76 | """ | |
77 | >>> decodedir(b'data/foo.i') |
|
77 | >>> decodedir(b'data/foo.i') | |
78 | 'data/foo.i' |
|
78 | 'data/foo.i' | |
79 | >>> decodedir(b'data/foo.i.hg/bla.i') |
|
79 | >>> decodedir(b'data/foo.i.hg/bla.i') | |
80 | 'data/foo.i/bla.i' |
|
80 | 'data/foo.i/bla.i' | |
81 | >>> decodedir(b'data/foo.i.hg.hg/bla.i') |
|
81 | >>> decodedir(b'data/foo.i.hg.hg/bla.i') | |
82 | 'data/foo.i.hg/bla.i' |
|
82 | 'data/foo.i.hg/bla.i' | |
83 | """ |
|
83 | """ | |
84 | if b".hg/" not in path: |
|
84 | if b".hg/" not in path: | |
85 | return path |
|
85 | return path | |
86 | return ( |
|
86 | return ( | |
87 | path.replace(b".d.hg/", b".d/") |
|
87 | path.replace(b".d.hg/", b".d/") | |
88 | .replace(b".i.hg/", b".i/") |
|
88 | .replace(b".i.hg/", b".i/") | |
89 | .replace(b".hg.hg/", b".hg/") |
|
89 | .replace(b".hg.hg/", b".hg/") | |
90 | ) |
|
90 | ) | |
91 |
|
91 | |||
92 |
|
92 | |||
93 | def _reserved(): |
|
93 | def _reserved(): | |
94 | """characters that are problematic for filesystems |
|
94 | """characters that are problematic for filesystems | |
95 |
|
95 | |||
96 | * ascii escapes (0..31) |
|
96 | * ascii escapes (0..31) | |
97 | * ascii hi (126..255) |
|
97 | * ascii hi (126..255) | |
98 | * windows specials |
|
98 | * windows specials | |
99 |
|
99 | |||
100 | these characters will be escaped by encodefunctions |
|
100 | these characters will be escaped by encodefunctions | |
101 | """ |
|
101 | """ | |
102 | winreserved = [ord(x) for x in u'\\:*?"<>|'] |
|
102 | winreserved = [ord(x) for x in u'\\:*?"<>|'] | |
103 | for x in range(32): |
|
103 | for x in range(32): | |
104 | yield x |
|
104 | yield x | |
105 | for x in range(126, 256): |
|
105 | for x in range(126, 256): | |
106 | yield x |
|
106 | yield x | |
107 | for x in winreserved: |
|
107 | for x in winreserved: | |
108 | yield x |
|
108 | yield x | |
109 |
|
109 | |||
110 |
|
110 | |||
111 | def _buildencodefun(): |
|
111 | def _buildencodefun(): | |
112 | """ |
|
112 | """ | |
113 | >>> enc, dec = _buildencodefun() |
|
113 | >>> enc, dec = _buildencodefun() | |
114 |
|
114 | |||
115 | >>> enc(b'nothing/special.txt') |
|
115 | >>> enc(b'nothing/special.txt') | |
116 | 'nothing/special.txt' |
|
116 | 'nothing/special.txt' | |
117 | >>> dec(b'nothing/special.txt') |
|
117 | >>> dec(b'nothing/special.txt') | |
118 | 'nothing/special.txt' |
|
118 | 'nothing/special.txt' | |
119 |
|
119 | |||
120 | >>> enc(b'HELLO') |
|
120 | >>> enc(b'HELLO') | |
121 | '_h_e_l_l_o' |
|
121 | '_h_e_l_l_o' | |
122 | >>> dec(b'_h_e_l_l_o') |
|
122 | >>> dec(b'_h_e_l_l_o') | |
123 | 'HELLO' |
|
123 | 'HELLO' | |
124 |
|
124 | |||
125 | >>> enc(b'hello:world?') |
|
125 | >>> enc(b'hello:world?') | |
126 | 'hello~3aworld~3f' |
|
126 | 'hello~3aworld~3f' | |
127 | >>> dec(b'hello~3aworld~3f') |
|
127 | >>> dec(b'hello~3aworld~3f') | |
128 | 'hello:world?' |
|
128 | 'hello:world?' | |
129 |
|
129 | |||
130 | >>> enc(b'the\\x07quick\\xADshot') |
|
130 | >>> enc(b'the\\x07quick\\xADshot') | |
131 | 'the~07quick~adshot' |
|
131 | 'the~07quick~adshot' | |
132 | >>> dec(b'the~07quick~adshot') |
|
132 | >>> dec(b'the~07quick~adshot') | |
133 | 'the\\x07quick\\xadshot' |
|
133 | 'the\\x07quick\\xadshot' | |
134 | """ |
|
134 | """ | |
135 | e = b'_' |
|
135 | e = b'_' | |
136 | xchr = pycompat.bytechr |
|
136 | xchr = pycompat.bytechr | |
137 | asciistr = list(map(xchr, range(127))) |
|
137 | asciistr = list(map(xchr, range(127))) | |
138 | capitals = list(range(ord(b"A"), ord(b"Z") + 1)) |
|
138 | capitals = list(range(ord(b"A"), ord(b"Z") + 1)) | |
139 |
|
139 | |||
140 | cmap = {x: x for x in asciistr} |
|
140 | cmap = {x: x for x in asciistr} | |
141 | for x in _reserved(): |
|
141 | for x in _reserved(): | |
142 | cmap[xchr(x)] = b"~%02x" % x |
|
142 | cmap[xchr(x)] = b"~%02x" % x | |
143 | for x in capitals + [ord(e)]: |
|
143 | for x in capitals + [ord(e)]: | |
144 | cmap[xchr(x)] = e + xchr(x).lower() |
|
144 | cmap[xchr(x)] = e + xchr(x).lower() | |
145 |
|
145 | |||
146 | dmap = {} |
|
146 | dmap = {} | |
147 | for k, v in cmap.items(): |
|
147 | for k, v in cmap.items(): | |
148 | dmap[v] = k |
|
148 | dmap[v] = k | |
149 |
|
149 | |||
150 | def decode(s): |
|
150 | def decode(s): | |
151 | i = 0 |
|
151 | i = 0 | |
152 | while i < len(s): |
|
152 | while i < len(s): | |
153 | for l in range(1, 4): |
|
153 | for l in range(1, 4): | |
154 | try: |
|
154 | try: | |
155 | yield dmap[s[i : i + l]] |
|
155 | yield dmap[s[i : i + l]] | |
156 | i += l |
|
156 | i += l | |
157 | break |
|
157 | break | |
158 | except KeyError: |
|
158 | except KeyError: | |
159 | pass |
|
159 | pass | |
160 | else: |
|
160 | else: | |
161 | raise KeyError |
|
161 | raise KeyError | |
162 |
|
162 | |||
163 | return ( |
|
163 | return ( | |
164 | lambda s: b''.join([cmap[s[c : c + 1]] for c in range(len(s))]), |
|
164 | lambda s: b''.join([cmap[s[c : c + 1]] for c in range(len(s))]), | |
165 | lambda s: b''.join(list(decode(s))), |
|
165 | lambda s: b''.join(list(decode(s))), | |
166 | ) |
|
166 | ) | |
167 |
|
167 | |||
168 |
|
168 | |||
169 | _encodefname, _decodefname = _buildencodefun() |
|
169 | _encodefname, _decodefname = _buildencodefun() | |
170 |
|
170 | |||
171 |
|
171 | |||
172 | def encodefilename(s): |
|
172 | def encodefilename(s): | |
173 | """ |
|
173 | """ | |
174 | >>> encodefilename(b'foo.i/bar.d/bla.hg/hi:world?/HELLO') |
|
174 | >>> encodefilename(b'foo.i/bar.d/bla.hg/hi:world?/HELLO') | |
175 | 'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o' |
|
175 | 'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o' | |
176 | """ |
|
176 | """ | |
177 | return _encodefname(encodedir(s)) |
|
177 | return _encodefname(encodedir(s)) | |
178 |
|
178 | |||
179 |
|
179 | |||
180 | def decodefilename(s): |
|
180 | def decodefilename(s): | |
181 | """ |
|
181 | """ | |
182 | >>> decodefilename(b'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o') |
|
182 | >>> decodefilename(b'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o') | |
183 | 'foo.i/bar.d/bla.hg/hi:world?/HELLO' |
|
183 | 'foo.i/bar.d/bla.hg/hi:world?/HELLO' | |
184 | """ |
|
184 | """ | |
185 | return decodedir(_decodefname(s)) |
|
185 | return decodedir(_decodefname(s)) | |
186 |
|
186 | |||
187 |
|
187 | |||
188 | def _buildlowerencodefun(): |
|
188 | def _buildlowerencodefun(): | |
189 | """ |
|
189 | """ | |
190 | >>> f = _buildlowerencodefun() |
|
190 | >>> f = _buildlowerencodefun() | |
191 | >>> f(b'nothing/special.txt') |
|
191 | >>> f(b'nothing/special.txt') | |
192 | 'nothing/special.txt' |
|
192 | 'nothing/special.txt' | |
193 | >>> f(b'HELLO') |
|
193 | >>> f(b'HELLO') | |
194 | 'hello' |
|
194 | 'hello' | |
195 | >>> f(b'hello:world?') |
|
195 | >>> f(b'hello:world?') | |
196 | 'hello~3aworld~3f' |
|
196 | 'hello~3aworld~3f' | |
197 | >>> f(b'the\\x07quick\\xADshot') |
|
197 | >>> f(b'the\\x07quick\\xADshot') | |
198 | 'the~07quick~adshot' |
|
198 | 'the~07quick~adshot' | |
199 | """ |
|
199 | """ | |
200 | xchr = pycompat.bytechr |
|
200 | xchr = pycompat.bytechr | |
201 | cmap = {xchr(x): xchr(x) for x in range(127)} |
|
201 | cmap = {xchr(x): xchr(x) for x in range(127)} | |
202 | for x in _reserved(): |
|
202 | for x in _reserved(): | |
203 | cmap[xchr(x)] = b"~%02x" % x |
|
203 | cmap[xchr(x)] = b"~%02x" % x | |
204 | for x in range(ord(b"A"), ord(b"Z") + 1): |
|
204 | for x in range(ord(b"A"), ord(b"Z") + 1): | |
205 | cmap[xchr(x)] = xchr(x).lower() |
|
205 | cmap[xchr(x)] = xchr(x).lower() | |
206 |
|
206 | |||
207 | def lowerencode(s): |
|
207 | def lowerencode(s): | |
208 | return b"".join([cmap[c] for c in pycompat.iterbytestr(s)]) |
|
208 | return b"".join([cmap[c] for c in pycompat.iterbytestr(s)]) | |
209 |
|
209 | |||
210 | return lowerencode |
|
210 | return lowerencode | |
211 |
|
211 | |||
212 |
|
212 | |||
213 | lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun() |
|
213 | lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun() | |
214 |
|
214 | |||
215 | # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9 |
|
215 | # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9 | |
216 | _winres3 = (b'aux', b'con', b'prn', b'nul') # length 3 |
|
216 | _winres3 = (b'aux', b'con', b'prn', b'nul') # length 3 | |
217 | _winres4 = (b'com', b'lpt') # length 4 (with trailing 1..9) |
|
217 | _winres4 = (b'com', b'lpt') # length 4 (with trailing 1..9) | |
218 |
|
218 | |||
219 |
|
219 | |||
220 | def _auxencode(path, dotencode): |
|
220 | def _auxencode(path, dotencode): | |
221 | """ |
|
221 | """ | |
222 | Encodes filenames containing names reserved by Windows or which end in |
|
222 | Encodes filenames containing names reserved by Windows or which end in | |
223 | period or space. Does not touch other single reserved characters c. |
|
223 | period or space. Does not touch other single reserved characters c. | |
224 | Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here. |
|
224 | Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here. | |
225 | Additionally encodes space or period at the beginning, if dotencode is |
|
225 | Additionally encodes space or period at the beginning, if dotencode is | |
226 | True. Parameter path is assumed to be all lowercase. |
|
226 | True. Parameter path is assumed to be all lowercase. | |
227 | A segment only needs encoding if a reserved name appears as a |
|
227 | A segment only needs encoding if a reserved name appears as a | |
228 | basename (e.g. "aux", "aux.foo"). A directory or file named "foo.aux" |
|
228 | basename (e.g. "aux", "aux.foo"). A directory or file named "foo.aux" | |
229 | doesn't need encoding. |
|
229 | doesn't need encoding. | |
230 |
|
230 | |||
231 | >>> s = b'.foo/aux.txt/txt.aux/con/prn/nul/foo.' |
|
231 | >>> s = b'.foo/aux.txt/txt.aux/con/prn/nul/foo.' | |
232 | >>> _auxencode(s.split(b'/'), True) |
|
232 | >>> _auxencode(s.split(b'/'), True) | |
233 | ['~2efoo', 'au~78.txt', 'txt.aux', 'co~6e', 'pr~6e', 'nu~6c', 'foo~2e'] |
|
233 | ['~2efoo', 'au~78.txt', 'txt.aux', 'co~6e', 'pr~6e', 'nu~6c', 'foo~2e'] | |
234 | >>> s = b'.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.' |
|
234 | >>> s = b'.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.' | |
235 | >>> _auxencode(s.split(b'/'), False) |
|
235 | >>> _auxencode(s.split(b'/'), False) | |
236 | ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e'] |
|
236 | ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e'] | |
237 | >>> _auxencode([b'foo. '], True) |
|
237 | >>> _auxencode([b'foo. '], True) | |
238 | ['foo.~20'] |
|
238 | ['foo.~20'] | |
239 | >>> _auxencode([b' .foo'], True) |
|
239 | >>> _auxencode([b' .foo'], True) | |
240 | ['~20.foo'] |
|
240 | ['~20.foo'] | |
241 | """ |
|
241 | """ | |
242 | for i, n in enumerate(path): |
|
242 | for i, n in enumerate(path): | |
243 | if not n: |
|
243 | if not n: | |
244 | continue |
|
244 | continue | |
245 | if dotencode and n[0] in b'. ': |
|
245 | if dotencode and n[0] in b'. ': | |
246 | n = b"~%02x" % ord(n[0:1]) + n[1:] |
|
246 | n = b"~%02x" % ord(n[0:1]) + n[1:] | |
247 | path[i] = n |
|
247 | path[i] = n | |
248 | else: |
|
248 | else: | |
249 | l = n.find(b'.') |
|
249 | l = n.find(b'.') | |
250 | if l == -1: |
|
250 | if l == -1: | |
251 | l = len(n) |
|
251 | l = len(n) | |
252 | if (l == 3 and n[:3] in _winres3) or ( |
|
252 | if (l == 3 and n[:3] in _winres3) or ( | |
253 | l == 4 |
|
253 | l == 4 | |
254 | and n[3:4] <= b'9' |
|
254 | and n[3:4] <= b'9' | |
255 | and n[3:4] >= b'1' |
|
255 | and n[3:4] >= b'1' | |
256 | and n[:3] in _winres4 |
|
256 | and n[:3] in _winres4 | |
257 | ): |
|
257 | ): | |
258 | # encode third letter ('aux' -> 'au~78') |
|
258 | # encode third letter ('aux' -> 'au~78') | |
259 | ec = b"~%02x" % ord(n[2:3]) |
|
259 | ec = b"~%02x" % ord(n[2:3]) | |
260 | n = n[0:2] + ec + n[3:] |
|
260 | n = n[0:2] + ec + n[3:] | |
261 | path[i] = n |
|
261 | path[i] = n | |
262 | if n[-1] in b'. ': |
|
262 | if n[-1] in b'. ': | |
263 | # encode last period or space ('foo...' -> 'foo..~2e') |
|
263 | # encode last period or space ('foo...' -> 'foo..~2e') | |
264 | path[i] = n[:-1] + b"~%02x" % ord(n[-1:]) |
|
264 | path[i] = n[:-1] + b"~%02x" % ord(n[-1:]) | |
265 | return path |
|
265 | return path | |
266 |
|
266 | |||
267 |
|
267 | |||
268 | _maxstorepathlen = 120 |
|
268 | _maxstorepathlen = 120 | |
269 | _dirprefixlen = 8 |
|
269 | _dirprefixlen = 8 | |
270 | _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4 |
|
270 | _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4 | |
271 |
|
271 | |||
272 |
|
272 | |||
273 | def _hashencode(path, dotencode): |
|
273 | def _hashencode(path, dotencode): | |
274 | digest = hex(hashutil.sha1(path).digest()) |
|
274 | digest = hex(hashutil.sha1(path).digest()) | |
275 | le = lowerencode(path[5:]).split(b'/') # skips prefix 'data/' or 'meta/' |
|
275 | le = lowerencode(path[5:]).split(b'/') # skips prefix 'data/' or 'meta/' | |
276 | parts = _auxencode(le, dotencode) |
|
276 | parts = _auxencode(le, dotencode) | |
277 | basename = parts[-1] |
|
277 | basename = parts[-1] | |
278 | _root, ext = os.path.splitext(basename) |
|
278 | _root, ext = os.path.splitext(basename) | |
279 | sdirs = [] |
|
279 | sdirs = [] | |
280 | sdirslen = 0 |
|
280 | sdirslen = 0 | |
281 | for p in parts[:-1]: |
|
281 | for p in parts[:-1]: | |
282 | d = p[:_dirprefixlen] |
|
282 | d = p[:_dirprefixlen] | |
283 | if d[-1] in b'. ': |
|
283 | if d[-1] in b'. ': | |
284 | # Windows can't access dirs ending in period or space |
|
284 | # Windows can't access dirs ending in period or space | |
285 | d = d[:-1] + b'_' |
|
285 | d = d[:-1] + b'_' | |
286 | if sdirslen == 0: |
|
286 | if sdirslen == 0: | |
287 | t = len(d) |
|
287 | t = len(d) | |
288 | else: |
|
288 | else: | |
289 | t = sdirslen + 1 + len(d) |
|
289 | t = sdirslen + 1 + len(d) | |
290 | if t > _maxshortdirslen: |
|
290 | if t > _maxshortdirslen: | |
291 | break |
|
291 | break | |
292 | sdirs.append(d) |
|
292 | sdirs.append(d) | |
293 | sdirslen = t |
|
293 | sdirslen = t | |
294 | dirs = b'/'.join(sdirs) |
|
294 | dirs = b'/'.join(sdirs) | |
295 | if len(dirs) > 0: |
|
295 | if len(dirs) > 0: | |
296 | dirs += b'/' |
|
296 | dirs += b'/' | |
297 | res = b'dh/' + dirs + digest + ext |
|
297 | res = b'dh/' + dirs + digest + ext | |
298 | spaceleft = _maxstorepathlen - len(res) |
|
298 | spaceleft = _maxstorepathlen - len(res) | |
299 | if spaceleft > 0: |
|
299 | if spaceleft > 0: | |
300 | filler = basename[:spaceleft] |
|
300 | filler = basename[:spaceleft] | |
301 | res = b'dh/' + dirs + filler + digest + ext |
|
301 | res = b'dh/' + dirs + filler + digest + ext | |
302 | return res |
|
302 | return res | |
303 |
|
303 | |||
304 |
|
304 | |||
305 | def _hybridencode(path, dotencode): |
|
305 | def _hybridencode(path, dotencode): | |
306 | """encodes path with a length limit |
|
306 | """encodes path with a length limit | |
307 |
|
307 | |||
308 | Encodes all paths that begin with 'data/', according to the following. |
|
308 | Encodes all paths that begin with 'data/', according to the following. | |
309 |
|
309 | |||
310 | Default encoding (reversible): |
|
310 | Default encoding (reversible): | |
311 |
|
311 | |||
312 | Encodes all uppercase letters 'X' as '_x'. All reserved or illegal |
|
312 | Encodes all uppercase letters 'X' as '_x'. All reserved or illegal | |
313 | characters are encoded as '~xx', where xx is the two digit hex code |
|
313 | characters are encoded as '~xx', where xx is the two digit hex code | |
314 | of the character (see encodefilename). |
|
314 | of the character (see encodefilename). | |
315 | Relevant path components consisting of Windows reserved filenames are |
|
315 | Relevant path components consisting of Windows reserved filenames are | |
316 | masked by encoding the third character ('aux' -> 'au~78', see _auxencode). |
|
316 | masked by encoding the third character ('aux' -> 'au~78', see _auxencode). | |
317 |
|
317 | |||
318 | Hashed encoding (not reversible): |
|
318 | Hashed encoding (not reversible): | |
319 |
|
319 | |||
320 | If the default-encoded path is longer than _maxstorepathlen, a |
|
320 | If the default-encoded path is longer than _maxstorepathlen, a | |
321 | non-reversible hybrid hashing of the path is done instead. |
|
321 | non-reversible hybrid hashing of the path is done instead. | |
322 | This encoding uses up to _dirprefixlen characters of all directory |
|
322 | This encoding uses up to _dirprefixlen characters of all directory | |
323 | levels of the lowerencoded path, but not more levels than can fit into |
|
323 | levels of the lowerencoded path, but not more levels than can fit into | |
324 | _maxshortdirslen. |
|
324 | _maxshortdirslen. | |
325 | Then follows the filler followed by the sha digest of the full path. |
|
325 | Then follows the filler followed by the sha digest of the full path. | |
326 | The filler is the beginning of the basename of the lowerencoded path |
|
326 | The filler is the beginning of the basename of the lowerencoded path | |
327 | (the basename is everything after the last path separator). The filler |
|
327 | (the basename is everything after the last path separator). The filler | |
328 | is as long as possible, filling in characters from the basename until |
|
328 | is as long as possible, filling in characters from the basename until | |
329 | the encoded path has _maxstorepathlen characters (or all chars of the |
|
329 | the encoded path has _maxstorepathlen characters (or all chars of the | |
330 | basename have been taken). |
|
330 | basename have been taken). | |
331 | The extension (e.g. '.i' or '.d') is preserved. |
|
331 | The extension (e.g. '.i' or '.d') is preserved. | |
332 |
|
332 | |||
333 | The string 'data/' at the beginning is replaced with 'dh/', if the hashed |
|
333 | The string 'data/' at the beginning is replaced with 'dh/', if the hashed | |
334 | encoding was used. |
|
334 | encoding was used. | |
335 | """ |
|
335 | """ | |
336 | path = encodedir(path) |
|
336 | path = encodedir(path) | |
337 | ef = _encodefname(path).split(b'/') |
|
337 | ef = _encodefname(path).split(b'/') | |
338 | res = b'/'.join(_auxencode(ef, dotencode)) |
|
338 | res = b'/'.join(_auxencode(ef, dotencode)) | |
339 | if len(res) > _maxstorepathlen: |
|
339 | if len(res) > _maxstorepathlen: | |
340 | res = _hashencode(path, dotencode) |
|
340 | res = _hashencode(path, dotencode) | |
341 | return res |
|
341 | return res | |
342 |
|
342 | |||
343 |
|
343 | |||
344 | def _pathencode(path): |
|
344 | def _pathencode(path): | |
345 | de = encodedir(path) |
|
345 | de = encodedir(path) | |
346 | if len(path) > _maxstorepathlen: |
|
346 | if len(path) > _maxstorepathlen: | |
347 | return _hashencode(de, True) |
|
347 | return _hashencode(de, True) | |
348 | ef = _encodefname(de).split(b'/') |
|
348 | ef = _encodefname(de).split(b'/') | |
349 | res = b'/'.join(_auxencode(ef, True)) |
|
349 | res = b'/'.join(_auxencode(ef, True)) | |
350 | if len(res) > _maxstorepathlen: |
|
350 | if len(res) > _maxstorepathlen: | |
351 | return _hashencode(de, True) |
|
351 | return _hashencode(de, True) | |
352 | return res |
|
352 | return res | |
353 |
|
353 | |||
354 |
|
354 | |||
355 | _pathencode = getattr(parsers, 'pathencode', _pathencode) |
|
355 | _pathencode = getattr(parsers, 'pathencode', _pathencode) | |
356 |
|
356 | |||
357 |
|
357 | |||
358 | def _plainhybridencode(f): |
|
358 | def _plainhybridencode(f): | |
359 | return _hybridencode(f, False) |
|
359 | return _hybridencode(f, False) | |
360 |
|
360 | |||
361 |
|
361 | |||
362 | def _calcmode(vfs): |
|
362 | def _calcmode(vfs): | |
363 | try: |
|
363 | try: | |
364 | # files in .hg/ will be created using this mode |
|
364 | # files in .hg/ will be created using this mode | |
365 | mode = vfs.stat().st_mode |
|
365 | mode = vfs.stat().st_mode | |
366 | # avoid some useless chmods |
|
366 | # avoid some useless chmods | |
367 | if (0o777 & ~util.umask) == (0o777 & mode): |
|
367 | if (0o777 & ~util.umask) == (0o777 & mode): | |
368 | mode = None |
|
368 | mode = None | |
369 | except OSError: |
|
369 | except OSError: | |
370 | mode = None |
|
370 | mode = None | |
371 | return mode |
|
371 | return mode | |
372 |
|
372 | |||
373 |
|
373 | |||
374 | _data = [ |
|
374 | _data = [ | |
375 | b'bookmarks', |
|
375 | b'bookmarks', | |
376 | b'narrowspec', |
|
376 | b'narrowspec', | |
377 | b'data', |
|
377 | b'data', | |
378 | b'meta', |
|
378 | b'meta', | |
379 | b'00manifest.d', |
|
379 | b'00manifest.d', | |
380 | b'00manifest.i', |
|
380 | b'00manifest.i', | |
381 | b'00changelog.d', |
|
381 | b'00changelog.d', | |
382 | b'00changelog.i', |
|
382 | b'00changelog.i', | |
383 | b'phaseroots', |
|
383 | b'phaseroots', | |
384 | b'obsstore', |
|
384 | b'obsstore', | |
385 | b'requires', |
|
385 | b'requires', | |
386 | ] |
|
386 | ] | |
387 |
|
387 | |||
388 | REVLOG_FILES_MAIN_EXT = (b'.i',) |
|
388 | REVLOG_FILES_MAIN_EXT = (b'.i',) | |
389 | REVLOG_FILES_OTHER_EXT = ( |
|
389 | REVLOG_FILES_OTHER_EXT = ( | |
390 | b'.idx', |
|
390 | b'.idx', | |
391 | b'.d', |
|
391 | b'.d', | |
392 | b'.dat', |
|
392 | b'.dat', | |
393 | b'.n', |
|
393 | b'.n', | |
394 | b'.nd', |
|
394 | b'.nd', | |
395 | b'.sda', |
|
395 | b'.sda', | |
396 | ) |
|
396 | ) | |
397 | # file extension that also use a `-SOMELONGIDHASH.ext` form |
|
397 | # file extension that also use a `-SOMELONGIDHASH.ext` form | |
398 | REVLOG_FILES_LONG_EXT = ( |
|
398 | REVLOG_FILES_LONG_EXT = ( | |
399 | b'.nd', |
|
399 | b'.nd', | |
400 | b'.idx', |
|
400 | b'.idx', | |
401 | b'.dat', |
|
401 | b'.dat', | |
402 | b'.sda', |
|
402 | b'.sda', | |
403 | ) |
|
403 | ) | |
404 | # files that are "volatile" and might change between listing and streaming |
|
404 | # files that are "volatile" and might change between listing and streaming | |
405 | # |
|
405 | # | |
406 | # note: the ".nd" file are nodemap data and won't "change" but they might be |
|
406 | # note: the ".nd" file are nodemap data and won't "change" but they might be | |
407 | # deleted. |
|
407 | # deleted. | |
408 | REVLOG_FILES_VOLATILE_EXT = (b'.n', b'.nd') |
|
408 | REVLOG_FILES_VOLATILE_EXT = (b'.n', b'.nd') | |
409 |
|
409 | |||
410 | # some exception to the above matching |
|
410 | # some exception to the above matching | |
411 | # |
|
411 | # | |
412 | # XXX This is currently not in use because of issue6542 |
|
412 | # XXX This is currently not in use because of issue6542 | |
413 | EXCLUDED = re.compile(br'.*undo\.[^/]+\.(nd?|i)$') |
|
413 | EXCLUDED = re.compile(br'.*undo\.[^/]+\.(nd?|i)$') | |
414 |
|
414 | |||
415 |
|
415 | |||
416 | def is_revlog(f, kind, st): |
|
416 | def is_revlog(f, kind, st): | |
417 | if kind != stat.S_IFREG: |
|
417 | if kind != stat.S_IFREG: | |
418 | return None |
|
418 | return None | |
419 | return revlog_type(f) |
|
419 | return revlog_type(f) | |
420 |
|
420 | |||
421 |
|
421 | |||
422 | def revlog_type(f): |
|
422 | def revlog_type(f): | |
423 | # XXX we need to filter `undo.` created by the transaction here, however |
|
423 | # XXX we need to filter `undo.` created by the transaction here, however | |
424 | # being naive about it also filter revlog for `undo.*` files, leading to |
|
424 | # being naive about it also filter revlog for `undo.*` files, leading to | |
425 | # issue6542. So we no longer use EXCLUDED. |
|
425 | # issue6542. So we no longer use EXCLUDED. | |
426 | if f.endswith(REVLOG_FILES_MAIN_EXT): |
|
426 | if f.endswith(REVLOG_FILES_MAIN_EXT): | |
427 | return FILEFLAGS_REVLOG_MAIN |
|
427 | return FILEFLAGS_REVLOG_MAIN | |
428 | elif f.endswith(REVLOG_FILES_OTHER_EXT): |
|
428 | elif f.endswith(REVLOG_FILES_OTHER_EXT): | |
429 | t = FILETYPE_FILELOG_OTHER |
|
429 | t = FILETYPE_FILELOG_OTHER | |
430 | if f.endswith(REVLOG_FILES_VOLATILE_EXT): |
|
430 | if f.endswith(REVLOG_FILES_VOLATILE_EXT): | |
431 | t |= FILEFLAGS_VOLATILE |
|
431 | t |= FILEFLAGS_VOLATILE | |
432 | return t |
|
432 | return t | |
433 | return None |
|
433 | return None | |
434 |
|
434 | |||
435 |
|
435 | |||
436 | # the file is part of changelog data |
|
436 | # the file is part of changelog data | |
437 | FILEFLAGS_CHANGELOG = 1 << 13 |
|
437 | FILEFLAGS_CHANGELOG = 1 << 13 | |
438 | # the file is part of manifest data |
|
438 | # the file is part of manifest data | |
439 | FILEFLAGS_MANIFESTLOG = 1 << 12 |
|
439 | FILEFLAGS_MANIFESTLOG = 1 << 12 | |
440 | # the file is part of filelog data |
|
440 | # the file is part of filelog data | |
441 | FILEFLAGS_FILELOG = 1 << 11 |
|
441 | FILEFLAGS_FILELOG = 1 << 11 | |
442 | # file that are not directly part of a revlog |
|
442 | # file that are not directly part of a revlog | |
443 | FILEFLAGS_OTHER = 1 << 10 |
|
443 | FILEFLAGS_OTHER = 1 << 10 | |
444 |
|
444 | |||
445 | # the main entry point for a revlog |
|
445 | # the main entry point for a revlog | |
446 | FILEFLAGS_REVLOG_MAIN = 1 << 1 |
|
446 | FILEFLAGS_REVLOG_MAIN = 1 << 1 | |
447 | # a secondary file for a revlog |
|
447 | # a secondary file for a revlog | |
448 | FILEFLAGS_REVLOG_OTHER = 1 << 0 |
|
448 | FILEFLAGS_REVLOG_OTHER = 1 << 0 | |
449 |
|
449 | |||
450 | # files that are "volatile" and might change between listing and streaming |
|
450 | # files that are "volatile" and might change between listing and streaming | |
451 | FILEFLAGS_VOLATILE = 1 << 20 |
|
451 | FILEFLAGS_VOLATILE = 1 << 20 | |
452 |
|
452 | |||
453 | FILETYPE_CHANGELOG_MAIN = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_MAIN |
|
453 | FILETYPE_CHANGELOG_MAIN = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_MAIN | |
454 | FILETYPE_CHANGELOG_OTHER = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_OTHER |
|
454 | FILETYPE_CHANGELOG_OTHER = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_OTHER | |
455 | FILETYPE_MANIFESTLOG_MAIN = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_MAIN |
|
455 | FILETYPE_MANIFESTLOG_MAIN = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_MAIN | |
456 | FILETYPE_MANIFESTLOG_OTHER = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_OTHER |
|
456 | FILETYPE_MANIFESTLOG_OTHER = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_OTHER | |
457 | FILETYPE_FILELOG_MAIN = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_MAIN |
|
457 | FILETYPE_FILELOG_MAIN = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_MAIN | |
458 | FILETYPE_FILELOG_OTHER = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_OTHER |
|
458 | FILETYPE_FILELOG_OTHER = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_OTHER | |
459 | FILETYPE_OTHER = FILEFLAGS_OTHER |
|
459 | FILETYPE_OTHER = FILEFLAGS_OTHER | |
460 |
|
460 | |||
461 |
|
461 | |||
462 | @attr.s(slots=True) |
|
462 | @attr.s(slots=True) | |
463 | class StoreFile: |
|
463 | class StoreFile: | |
464 | """a file matching a store entry""" |
|
464 | """a file matching a store entry""" | |
465 |
|
465 | |||
466 | unencoded_path = attr.ib() |
|
466 | unencoded_path = attr.ib() | |
467 | _file_size = attr.ib(default=None) |
|
467 | _file_size = attr.ib(default=None) | |
468 | is_volatile = attr.ib(default=False) |
|
468 | is_volatile = attr.ib(default=False) | |
469 |
|
469 | |||
470 | def file_size(self, vfs): |
|
470 | def file_size(self, vfs): | |
471 | if self._file_size is None: |
|
471 | if self._file_size is None: | |
472 | if vfs is None: |
|
472 | if vfs is None: | |
473 | msg = b"calling vfs-less file_size without prior call: %s" |
|
473 | msg = b"calling vfs-less file_size without prior call: %s" | |
474 | msg %= self.unencoded_path |
|
474 | msg %= self.unencoded_path | |
475 | raise error.ProgrammingError(msg) |
|
475 | raise error.ProgrammingError(msg) | |
476 | try: |
|
476 | try: | |
477 | self._file_size = vfs.stat(self.unencoded_path).st_size |
|
477 | self._file_size = vfs.stat(self.unencoded_path).st_size | |
478 | except FileNotFoundError: |
|
478 | except FileNotFoundError: | |
479 | self._file_size = 0 |
|
479 | self._file_size = 0 | |
480 | return self._file_size |
|
480 | return self._file_size | |
481 |
|
481 | |||
482 | def get_stream(self, vfs, copies): |
|
482 | def get_stream(self, vfs, copies): | |
483 | """return data "stream" information for this file |
|
483 | """return data "stream" information for this file | |
484 |
|
484 | |||
485 | (unencoded_file_path, content_iterator, content_size) |
|
485 | (unencoded_file_path, content_iterator, content_size) | |
486 | """ |
|
486 | """ | |
487 | size = self.file_size(None) |
|
487 | size = self.file_size(None) | |
488 |
|
488 | |||
489 | def get_stream(): |
|
489 | def get_stream(): | |
490 | actual_path = copies[vfs.join(self.unencoded_path)] |
|
490 | actual_path = copies[vfs.join(self.unencoded_path)] | |
491 | with open(actual_path, 'rb') as fp: |
|
491 | with open(actual_path, 'rb') as fp: | |
492 | yield None # ready to stream |
|
492 | yield None # ready to stream | |
493 | if size <= 65536: |
|
493 | if size <= 65536: | |
494 | yield fp.read(size) |
|
494 | yield fp.read(size) | |
495 | else: |
|
495 | else: | |
496 | yield from util.filechunkiter(fp, limit=size) |
|
496 | yield from util.filechunkiter(fp, limit=size) | |
497 |
|
497 | |||
498 | s = get_stream() |
|
498 | s = get_stream() | |
499 | next(s) |
|
499 | next(s) | |
500 | return (self.unencoded_path, s, size) |
|
500 | return (self.unencoded_path, s, size) | |
501 |
|
501 | |||
502 |
|
502 | |||
503 | @attr.s(slots=True, init=False) |
|
503 | @attr.s(slots=True, init=False) | |
504 | class BaseStoreEntry: |
|
504 | class BaseStoreEntry: | |
505 | """An entry in the store |
|
505 | """An entry in the store | |
506 |
|
506 | |||
507 | This is returned by `store.walk` and represent some data in the store.""" |
|
507 | This is returned by `store.walk` and represent some data in the store.""" | |
508 |
|
508 | |||
509 | def files(self) -> List[StoreFile]: |
|
509 | def files(self) -> List[StoreFile]: | |
510 | raise NotImplementedError |
|
510 | raise NotImplementedError | |
511 |
|
511 | |||
512 |
def get_streams( |
|
512 | def get_streams( | |
|
513 | self, | |||
|
514 | repo=None, | |||
|
515 | vfs=None, | |||
|
516 | copies=None, | |||
|
517 | max_changeset=None, | |||
|
518 | ): | |||
513 | """return a list of data stream associated to files for this entry |
|
519 | """return a list of data stream associated to files for this entry | |
514 |
|
520 | |||
515 | return [(unencoded_file_path, content_iterator, content_size), β¦] |
|
521 | return [(unencoded_file_path, content_iterator, content_size), β¦] | |
516 | """ |
|
522 | """ | |
517 | assert vfs is not None |
|
523 | assert vfs is not None | |
518 | return [f.get_stream(vfs, copies) for f in self.files()] |
|
524 | return [f.get_stream(vfs, copies) for f in self.files()] | |
519 |
|
525 | |||
520 |
|
526 | |||
521 | @attr.s(slots=True, init=False) |
|
527 | @attr.s(slots=True, init=False) | |
522 | class SimpleStoreEntry(BaseStoreEntry): |
|
528 | class SimpleStoreEntry(BaseStoreEntry): | |
523 | """A generic entry in the store""" |
|
529 | """A generic entry in the store""" | |
524 |
|
530 | |||
525 | is_revlog = False |
|
531 | is_revlog = False | |
526 |
|
532 | |||
527 | _entry_path = attr.ib() |
|
533 | _entry_path = attr.ib() | |
528 | _is_volatile = attr.ib(default=False) |
|
534 | _is_volatile = attr.ib(default=False) | |
529 | _file_size = attr.ib(default=None) |
|
535 | _file_size = attr.ib(default=None) | |
530 | _files = attr.ib(default=None) |
|
536 | _files = attr.ib(default=None) | |
531 |
|
537 | |||
532 | def __init__( |
|
538 | def __init__( | |
533 | self, |
|
539 | self, | |
534 | entry_path, |
|
540 | entry_path, | |
535 | is_volatile=False, |
|
541 | is_volatile=False, | |
536 | file_size=None, |
|
542 | file_size=None, | |
537 | ): |
|
543 | ): | |
538 | super().__init__() |
|
544 | super().__init__() | |
539 | self._entry_path = entry_path |
|
545 | self._entry_path = entry_path | |
540 | self._is_volatile = is_volatile |
|
546 | self._is_volatile = is_volatile | |
541 | self._file_size = file_size |
|
547 | self._file_size = file_size | |
542 | self._files = None |
|
548 | self._files = None | |
543 |
|
549 | |||
544 | def files(self) -> List[StoreFile]: |
|
550 | def files(self) -> List[StoreFile]: | |
545 | if self._files is None: |
|
551 | if self._files is None: | |
546 | self._files = [ |
|
552 | self._files = [ | |
547 | StoreFile( |
|
553 | StoreFile( | |
548 | unencoded_path=self._entry_path, |
|
554 | unencoded_path=self._entry_path, | |
549 | file_size=self._file_size, |
|
555 | file_size=self._file_size, | |
550 | is_volatile=self._is_volatile, |
|
556 | is_volatile=self._is_volatile, | |
551 | ) |
|
557 | ) | |
552 | ] |
|
558 | ] | |
553 | return self._files |
|
559 | return self._files | |
554 |
|
560 | |||
555 |
|
561 | |||
556 | @attr.s(slots=True, init=False) |
|
562 | @attr.s(slots=True, init=False) | |
557 | class RevlogStoreEntry(BaseStoreEntry): |
|
563 | class RevlogStoreEntry(BaseStoreEntry): | |
558 | """A revlog entry in the store""" |
|
564 | """A revlog entry in the store""" | |
559 |
|
565 | |||
560 | is_revlog = True |
|
566 | is_revlog = True | |
561 |
|
567 | |||
562 | revlog_type = attr.ib(default=None) |
|
568 | revlog_type = attr.ib(default=None) | |
563 | target_id = attr.ib(default=None) |
|
569 | target_id = attr.ib(default=None) | |
564 | _path_prefix = attr.ib(default=None) |
|
570 | _path_prefix = attr.ib(default=None) | |
565 | _details = attr.ib(default=None) |
|
571 | _details = attr.ib(default=None) | |
566 | _files = attr.ib(default=None) |
|
572 | _files = attr.ib(default=None) | |
567 |
|
573 | |||
568 | def __init__( |
|
574 | def __init__( | |
569 | self, |
|
575 | self, | |
570 | revlog_type, |
|
576 | revlog_type, | |
571 | path_prefix, |
|
577 | path_prefix, | |
572 | target_id, |
|
578 | target_id, | |
573 | details, |
|
579 | details, | |
574 | ): |
|
580 | ): | |
575 | super().__init__() |
|
581 | super().__init__() | |
576 | self.revlog_type = revlog_type |
|
582 | self.revlog_type = revlog_type | |
577 | self.target_id = target_id |
|
583 | self.target_id = target_id | |
578 | self._path_prefix = path_prefix |
|
584 | self._path_prefix = path_prefix | |
579 | assert b'.i' in details, (path_prefix, details) |
|
585 | assert b'.i' in details, (path_prefix, details) | |
580 | self._details = details |
|
586 | self._details = details | |
581 | self._files = None |
|
587 | self._files = None | |
582 |
|
588 | |||
583 | @property |
|
589 | @property | |
584 | def is_changelog(self): |
|
590 | def is_changelog(self): | |
585 | return self.revlog_type & FILEFLAGS_CHANGELOG |
|
591 | return self.revlog_type & FILEFLAGS_CHANGELOG | |
586 |
|
592 | |||
587 | @property |
|
593 | @property | |
588 | def is_manifestlog(self): |
|
594 | def is_manifestlog(self): | |
589 | return self.revlog_type & FILEFLAGS_MANIFESTLOG |
|
595 | return self.revlog_type & FILEFLAGS_MANIFESTLOG | |
590 |
|
596 | |||
591 | @property |
|
597 | @property | |
592 | def is_filelog(self): |
|
598 | def is_filelog(self): | |
593 | return self.revlog_type & FILEFLAGS_FILELOG |
|
599 | return self.revlog_type & FILEFLAGS_FILELOG | |
594 |
|
600 | |||
595 | def main_file_path(self): |
|
601 | def main_file_path(self): | |
596 | """unencoded path of the main revlog file""" |
|
602 | """unencoded path of the main revlog file""" | |
597 | return self._path_prefix + b'.i' |
|
603 | return self._path_prefix + b'.i' | |
598 |
|
604 | |||
599 | def files(self) -> List[StoreFile]: |
|
605 | def files(self) -> List[StoreFile]: | |
600 | if self._files is None: |
|
606 | if self._files is None: | |
601 | self._files = [] |
|
607 | self._files = [] | |
602 | for ext in sorted(self._details, key=_ext_key): |
|
608 | for ext in sorted(self._details, key=_ext_key): | |
603 | path = self._path_prefix + ext |
|
609 | path = self._path_prefix + ext | |
604 | data = self._details[ext] |
|
610 | data = self._details[ext] | |
605 | self._files.append(StoreFile(unencoded_path=path, **data)) |
|
611 | self._files.append(StoreFile(unencoded_path=path, **data)) | |
606 | return self._files |
|
612 | return self._files | |
607 |
|
613 | |||
|
614 | def get_streams( | |||
|
615 | self, | |||
|
616 | repo=None, | |||
|
617 | vfs=None, | |||
|
618 | copies=None, | |||
|
619 | max_changeset=None, | |||
|
620 | ): | |||
|
621 | if repo is None or max_changeset is None: | |||
|
622 | return super().get_streams( | |||
|
623 | repo=repo, | |||
|
624 | vfs=vfs, | |||
|
625 | copies=copies, | |||
|
626 | max_changeset=max_changeset, | |||
|
627 | ) | |||
|
628 | if any(k.endswith(b'.idx') for k in self._details.keys()): | |||
|
629 | # This use revlog-v2, ignore for now | |||
|
630 | return super().get_streams( | |||
|
631 | repo=repo, | |||
|
632 | vfs=vfs, | |||
|
633 | copies=copies, | |||
|
634 | max_changeset=max_changeset, | |||
|
635 | ) | |||
|
636 | name_to_ext = {} | |||
|
637 | for ext in self._details.keys(): | |||
|
638 | name_to_ext[self._path_prefix + ext] = ext | |||
|
639 | name_to_size = {} | |||
|
640 | for f in self.files(): | |||
|
641 | name_to_size[f.unencoded_path] = f.file_size(None) | |||
|
642 | stream = [ | |||
|
643 | f.get_stream(vfs, copies) | |||
|
644 | for f in self.files() | |||
|
645 | if name_to_ext[f.unencoded_path] not in (b'.d', b'.i') | |||
|
646 | ] | |||
|
647 | ||||
|
648 | rl = self.get_revlog_instance(repo).get_revlog() | |||
|
649 | rl_stream = rl.get_streams(max_changeset) | |||
|
650 | for name, s, size in rl_stream: | |||
|
651 | if name_to_size.get(name, 0) != size: | |||
|
652 | msg = _(b"expected %d bytes but %d provided for %s") | |||
|
653 | msg %= name_to_size.get(name, 0), size, name | |||
|
654 | raise error.Abort(msg) | |||
|
655 | stream.extend(rl_stream) | |||
|
656 | files = self.files() | |||
|
657 | assert len(stream) == len(files), ( | |||
|
658 | stream, | |||
|
659 | files, | |||
|
660 | self._path_prefix, | |||
|
661 | self.target_id, | |||
|
662 | ) | |||
|
663 | return stream | |||
|
664 | ||||
608 | def get_revlog_instance(self, repo): |
|
665 | def get_revlog_instance(self, repo): | |
609 | """Obtain a revlog instance from this store entry |
|
666 | """Obtain a revlog instance from this store entry | |
610 |
|
667 | |||
611 | An instance of the appropriate class is returned. |
|
668 | An instance of the appropriate class is returned. | |
612 | """ |
|
669 | """ | |
613 | if self.is_changelog: |
|
670 | if self.is_changelog: | |
614 | return changelog.changelog(repo.svfs) |
|
671 | return changelog.changelog(repo.svfs) | |
615 | elif self.is_manifestlog: |
|
672 | elif self.is_manifestlog: | |
616 | mandir = self.target_id |
|
673 | mandir = self.target_id | |
617 | return manifest.manifestrevlog( |
|
674 | return manifest.manifestrevlog( | |
618 | repo.nodeconstants, repo.svfs, tree=mandir |
|
675 | repo.nodeconstants, repo.svfs, tree=mandir | |
619 | ) |
|
676 | ) | |
620 | else: |
|
677 | else: | |
621 | return filelog.filelog(repo.svfs, self.target_id) |
|
678 | return filelog.filelog(repo.svfs, self.target_id) | |
622 |
|
679 | |||
623 |
|
680 | |||
624 | def _gather_revlog(files_data): |
|
681 | def _gather_revlog(files_data): | |
625 | """group files per revlog prefix |
|
682 | """group files per revlog prefix | |
626 |
|
683 | |||
627 | The returns a two level nested dict. The top level key is the revlog prefix |
|
684 | The returns a two level nested dict. The top level key is the revlog prefix | |
628 | without extension, the second level is all the file "suffix" that were |
|
685 | without extension, the second level is all the file "suffix" that were | |
629 | seen for this revlog and arbitrary file data as value. |
|
686 | seen for this revlog and arbitrary file data as value. | |
630 | """ |
|
687 | """ | |
631 | revlogs = collections.defaultdict(dict) |
|
688 | revlogs = collections.defaultdict(dict) | |
632 | for u, value in files_data: |
|
689 | for u, value in files_data: | |
633 | name, ext = _split_revlog_ext(u) |
|
690 | name, ext = _split_revlog_ext(u) | |
634 | revlogs[name][ext] = value |
|
691 | revlogs[name][ext] = value | |
635 | return sorted(revlogs.items()) |
|
692 | return sorted(revlogs.items()) | |
636 |
|
693 | |||
637 |
|
694 | |||
638 | def _split_revlog_ext(filename): |
|
695 | def _split_revlog_ext(filename): | |
639 | """split the revlog file prefix from the variable extension""" |
|
696 | """split the revlog file prefix from the variable extension""" | |
640 | if filename.endswith(REVLOG_FILES_LONG_EXT): |
|
697 | if filename.endswith(REVLOG_FILES_LONG_EXT): | |
641 | char = b'-' |
|
698 | char = b'-' | |
642 | else: |
|
699 | else: | |
643 | char = b'.' |
|
700 | char = b'.' | |
644 | idx = filename.rfind(char) |
|
701 | idx = filename.rfind(char) | |
645 | return filename[:idx], filename[idx:] |
|
702 | return filename[:idx], filename[idx:] | |
646 |
|
703 | |||
647 |
|
704 | |||
648 | def _ext_key(ext): |
|
705 | def _ext_key(ext): | |
649 | """a key to order revlog suffix |
|
706 | """a key to order revlog suffix | |
650 |
|
707 | |||
651 | important to issue .i after other entry.""" |
|
708 | important to issue .i after other entry.""" | |
652 | # the only important part of this order is to keep the `.i` last. |
|
709 | # the only important part of this order is to keep the `.i` last. | |
653 | if ext.endswith(b'.n'): |
|
710 | if ext.endswith(b'.n'): | |
654 | return (0, ext) |
|
711 | return (0, ext) | |
655 | elif ext.endswith(b'.nd'): |
|
712 | elif ext.endswith(b'.nd'): | |
656 | return (10, ext) |
|
713 | return (10, ext) | |
657 | elif ext.endswith(b'.d'): |
|
714 | elif ext.endswith(b'.d'): | |
658 | return (20, ext) |
|
715 | return (20, ext) | |
659 | elif ext.endswith(b'.i'): |
|
716 | elif ext.endswith(b'.i'): | |
660 | return (50, ext) |
|
717 | return (50, ext) | |
661 | else: |
|
718 | else: | |
662 | return (40, ext) |
|
719 | return (40, ext) | |
663 |
|
720 | |||
664 |
|
721 | |||
665 | class basicstore: |
|
722 | class basicstore: | |
666 | '''base class for local repository stores''' |
|
723 | '''base class for local repository stores''' | |
667 |
|
724 | |||
668 | def __init__(self, path, vfstype): |
|
725 | def __init__(self, path, vfstype): | |
669 | vfs = vfstype(path) |
|
726 | vfs = vfstype(path) | |
670 | self.path = vfs.base |
|
727 | self.path = vfs.base | |
671 | self.createmode = _calcmode(vfs) |
|
728 | self.createmode = _calcmode(vfs) | |
672 | vfs.createmode = self.createmode |
|
729 | vfs.createmode = self.createmode | |
673 | self.rawvfs = vfs |
|
730 | self.rawvfs = vfs | |
674 | self.vfs = vfsmod.filtervfs(vfs, encodedir) |
|
731 | self.vfs = vfsmod.filtervfs(vfs, encodedir) | |
675 | self.opener = self.vfs |
|
732 | self.opener = self.vfs | |
676 |
|
733 | |||
677 | def join(self, f): |
|
734 | def join(self, f): | |
678 | return self.path + b'/' + encodedir(f) |
|
735 | return self.path + b'/' + encodedir(f) | |
679 |
|
736 | |||
680 | def _walk(self, relpath, recurse, undecodable=None): |
|
737 | def _walk(self, relpath, recurse, undecodable=None): | |
681 | '''yields (revlog_type, unencoded, size)''' |
|
738 | '''yields (revlog_type, unencoded, size)''' | |
682 | path = self.path |
|
739 | path = self.path | |
683 | if relpath: |
|
740 | if relpath: | |
684 | path += b'/' + relpath |
|
741 | path += b'/' + relpath | |
685 | striplen = len(self.path) + 1 |
|
742 | striplen = len(self.path) + 1 | |
686 | l = [] |
|
743 | l = [] | |
687 | if self.rawvfs.isdir(path): |
|
744 | if self.rawvfs.isdir(path): | |
688 | visit = [path] |
|
745 | visit = [path] | |
689 | readdir = self.rawvfs.readdir |
|
746 | readdir = self.rawvfs.readdir | |
690 | while visit: |
|
747 | while visit: | |
691 | p = visit.pop() |
|
748 | p = visit.pop() | |
692 | for f, kind, st in readdir(p, stat=True): |
|
749 | for f, kind, st in readdir(p, stat=True): | |
693 | fp = p + b'/' + f |
|
750 | fp = p + b'/' + f | |
694 | rl_type = is_revlog(f, kind, st) |
|
751 | rl_type = is_revlog(f, kind, st) | |
695 | if rl_type is not None: |
|
752 | if rl_type is not None: | |
696 | n = util.pconvert(fp[striplen:]) |
|
753 | n = util.pconvert(fp[striplen:]) | |
697 | l.append((decodedir(n), (rl_type, st.st_size))) |
|
754 | l.append((decodedir(n), (rl_type, st.st_size))) | |
698 | elif kind == stat.S_IFDIR and recurse: |
|
755 | elif kind == stat.S_IFDIR and recurse: | |
699 | visit.append(fp) |
|
756 | visit.append(fp) | |
700 |
|
757 | |||
701 | l.sort() |
|
758 | l.sort() | |
702 | return l |
|
759 | return l | |
703 |
|
760 | |||
704 | def changelog(self, trypending, concurrencychecker=None): |
|
761 | def changelog(self, trypending, concurrencychecker=None): | |
705 | return changelog.changelog( |
|
762 | return changelog.changelog( | |
706 | self.vfs, |
|
763 | self.vfs, | |
707 | trypending=trypending, |
|
764 | trypending=trypending, | |
708 | concurrencychecker=concurrencychecker, |
|
765 | concurrencychecker=concurrencychecker, | |
709 | ) |
|
766 | ) | |
710 |
|
767 | |||
711 | def manifestlog(self, repo, storenarrowmatch): |
|
768 | def manifestlog(self, repo, storenarrowmatch): | |
712 | rootstore = manifest.manifestrevlog(repo.nodeconstants, self.vfs) |
|
769 | rootstore = manifest.manifestrevlog(repo.nodeconstants, self.vfs) | |
713 | return manifest.manifestlog(self.vfs, repo, rootstore, storenarrowmatch) |
|
770 | return manifest.manifestlog(self.vfs, repo, rootstore, storenarrowmatch) | |
714 |
|
771 | |||
715 | def data_entries( |
|
772 | def data_entries( | |
716 | self, matcher=None, undecodable=None |
|
773 | self, matcher=None, undecodable=None | |
717 | ) -> Generator[BaseStoreEntry, None, None]: |
|
774 | ) -> Generator[BaseStoreEntry, None, None]: | |
718 | """Like walk, but excluding the changelog and root manifest. |
|
775 | """Like walk, but excluding the changelog and root manifest. | |
719 |
|
776 | |||
720 | When [undecodable] is None, revlogs names that can't be |
|
777 | When [undecodable] is None, revlogs names that can't be | |
721 | decoded cause an exception. When it is provided, it should |
|
778 | decoded cause an exception. When it is provided, it should | |
722 | be a list and the filenames that can't be decoded are added |
|
779 | be a list and the filenames that can't be decoded are added | |
723 | to it instead. This is very rarely needed.""" |
|
780 | to it instead. This is very rarely needed.""" | |
724 | dirs = [ |
|
781 | dirs = [ | |
725 | (b'data', FILEFLAGS_FILELOG, False), |
|
782 | (b'data', FILEFLAGS_FILELOG, False), | |
726 | (b'meta', FILEFLAGS_MANIFESTLOG, True), |
|
783 | (b'meta', FILEFLAGS_MANIFESTLOG, True), | |
727 | ] |
|
784 | ] | |
728 | for base_dir, rl_type, strip_filename in dirs: |
|
785 | for base_dir, rl_type, strip_filename in dirs: | |
729 | files = self._walk(base_dir, True, undecodable=undecodable) |
|
786 | files = self._walk(base_dir, True, undecodable=undecodable) | |
730 | files = (f for f in files if f[1][0] is not None) |
|
787 | files = (f for f in files if f[1][0] is not None) | |
731 | for revlog, details in _gather_revlog(files): |
|
788 | for revlog, details in _gather_revlog(files): | |
732 | file_details = {} |
|
789 | file_details = {} | |
733 | revlog_target_id = revlog.split(b'/', 1)[1] |
|
790 | revlog_target_id = revlog.split(b'/', 1)[1] | |
734 | if strip_filename and b'/' in revlog: |
|
791 | if strip_filename and b'/' in revlog: | |
735 | revlog_target_id = revlog_target_id.rsplit(b'/', 1)[0] |
|
792 | revlog_target_id = revlog_target_id.rsplit(b'/', 1)[0] | |
736 | revlog_target_id += b'/' |
|
793 | revlog_target_id += b'/' | |
737 | for ext, (t, s) in sorted(details.items()): |
|
794 | for ext, (t, s) in sorted(details.items()): | |
738 | file_details[ext] = { |
|
795 | file_details[ext] = { | |
739 | 'is_volatile': bool(t & FILEFLAGS_VOLATILE), |
|
796 | 'is_volatile': bool(t & FILEFLAGS_VOLATILE), | |
740 | 'file_size': s, |
|
797 | 'file_size': s, | |
741 | } |
|
798 | } | |
742 | yield RevlogStoreEntry( |
|
799 | yield RevlogStoreEntry( | |
743 | path_prefix=revlog, |
|
800 | path_prefix=revlog, | |
744 | revlog_type=rl_type, |
|
801 | revlog_type=rl_type, | |
745 | target_id=revlog_target_id, |
|
802 | target_id=revlog_target_id, | |
746 | details=file_details, |
|
803 | details=file_details, | |
747 | ) |
|
804 | ) | |
748 |
|
805 | |||
749 | def top_entries( |
|
806 | def top_entries( | |
750 | self, phase=False, obsolescence=False |
|
807 | self, phase=False, obsolescence=False | |
751 | ) -> Generator[BaseStoreEntry, None, None]: |
|
808 | ) -> Generator[BaseStoreEntry, None, None]: | |
752 | if phase and self.vfs.exists(b'phaseroots'): |
|
809 | if phase and self.vfs.exists(b'phaseroots'): | |
753 | yield SimpleStoreEntry( |
|
810 | yield SimpleStoreEntry( | |
754 | entry_path=b'phaseroots', |
|
811 | entry_path=b'phaseroots', | |
755 | is_volatile=True, |
|
812 | is_volatile=True, | |
756 | ) |
|
813 | ) | |
757 |
|
814 | |||
758 | if obsolescence and self.vfs.exists(b'obsstore'): |
|
815 | if obsolescence and self.vfs.exists(b'obsstore'): | |
759 | # XXX if we had the file size it could be non-volatile |
|
816 | # XXX if we had the file size it could be non-volatile | |
760 | yield SimpleStoreEntry( |
|
817 | yield SimpleStoreEntry( | |
761 | entry_path=b'obsstore', |
|
818 | entry_path=b'obsstore', | |
762 | is_volatile=True, |
|
819 | is_volatile=True, | |
763 | ) |
|
820 | ) | |
764 |
|
821 | |||
765 | files = reversed(self._walk(b'', False)) |
|
822 | files = reversed(self._walk(b'', False)) | |
766 |
|
823 | |||
767 | changelogs = collections.defaultdict(dict) |
|
824 | changelogs = collections.defaultdict(dict) | |
768 | manifestlogs = collections.defaultdict(dict) |
|
825 | manifestlogs = collections.defaultdict(dict) | |
769 |
|
826 | |||
770 | for u, (t, s) in files: |
|
827 | for u, (t, s) in files: | |
771 | if u.startswith(b'00changelog'): |
|
828 | if u.startswith(b'00changelog'): | |
772 | name, ext = _split_revlog_ext(u) |
|
829 | name, ext = _split_revlog_ext(u) | |
773 | changelogs[name][ext] = (t, s) |
|
830 | changelogs[name][ext] = (t, s) | |
774 | elif u.startswith(b'00manifest'): |
|
831 | elif u.startswith(b'00manifest'): | |
775 | name, ext = _split_revlog_ext(u) |
|
832 | name, ext = _split_revlog_ext(u) | |
776 | manifestlogs[name][ext] = (t, s) |
|
833 | manifestlogs[name][ext] = (t, s) | |
777 | else: |
|
834 | else: | |
778 | yield SimpleStoreEntry( |
|
835 | yield SimpleStoreEntry( | |
779 | entry_path=u, |
|
836 | entry_path=u, | |
780 | is_volatile=bool(t & FILEFLAGS_VOLATILE), |
|
837 | is_volatile=bool(t & FILEFLAGS_VOLATILE), | |
781 | file_size=s, |
|
838 | file_size=s, | |
782 | ) |
|
839 | ) | |
783 | # yield manifest before changelog |
|
840 | # yield manifest before changelog | |
784 | top_rl = [ |
|
841 | top_rl = [ | |
785 | (manifestlogs, FILEFLAGS_MANIFESTLOG), |
|
842 | (manifestlogs, FILEFLAGS_MANIFESTLOG), | |
786 | (changelogs, FILEFLAGS_CHANGELOG), |
|
843 | (changelogs, FILEFLAGS_CHANGELOG), | |
787 | ] |
|
844 | ] | |
788 | assert len(manifestlogs) <= 1 |
|
845 | assert len(manifestlogs) <= 1 | |
789 | assert len(changelogs) <= 1 |
|
846 | assert len(changelogs) <= 1 | |
790 | for data, revlog_type in top_rl: |
|
847 | for data, revlog_type in top_rl: | |
791 | for revlog, details in sorted(data.items()): |
|
848 | for revlog, details in sorted(data.items()): | |
792 | file_details = {} |
|
849 | file_details = {} | |
793 | for ext, (t, s) in details.items(): |
|
850 | for ext, (t, s) in details.items(): | |
794 | file_details[ext] = { |
|
851 | file_details[ext] = { | |
795 | 'is_volatile': bool(t & FILEFLAGS_VOLATILE), |
|
852 | 'is_volatile': bool(t & FILEFLAGS_VOLATILE), | |
796 | 'file_size': s, |
|
853 | 'file_size': s, | |
797 | } |
|
854 | } | |
798 | yield RevlogStoreEntry( |
|
855 | yield RevlogStoreEntry( | |
799 | path_prefix=revlog, |
|
856 | path_prefix=revlog, | |
800 | revlog_type=revlog_type, |
|
857 | revlog_type=revlog_type, | |
801 | target_id=b'', |
|
858 | target_id=b'', | |
802 | details=file_details, |
|
859 | details=file_details, | |
803 | ) |
|
860 | ) | |
804 |
|
861 | |||
805 | def walk( |
|
862 | def walk( | |
806 | self, matcher=None, phase=False, obsolescence=False |
|
863 | self, matcher=None, phase=False, obsolescence=False | |
807 | ) -> Generator[BaseStoreEntry, None, None]: |
|
864 | ) -> Generator[BaseStoreEntry, None, None]: | |
808 | """return files related to data storage (ie: revlogs) |
|
865 | """return files related to data storage (ie: revlogs) | |
809 |
|
866 | |||
810 | yields instance from BaseStoreEntry subclasses |
|
867 | yields instance from BaseStoreEntry subclasses | |
811 |
|
868 | |||
812 | if a matcher is passed, storage files of only those tracked paths |
|
869 | if a matcher is passed, storage files of only those tracked paths | |
813 | are passed with matches the matcher |
|
870 | are passed with matches the matcher | |
814 | """ |
|
871 | """ | |
815 | # yield data files first |
|
872 | # yield data files first | |
816 | for x in self.data_entries(matcher): |
|
873 | for x in self.data_entries(matcher): | |
817 | yield x |
|
874 | yield x | |
818 | for x in self.top_entries(phase=phase, obsolescence=obsolescence): |
|
875 | for x in self.top_entries(phase=phase, obsolescence=obsolescence): | |
819 | yield x |
|
876 | yield x | |
820 |
|
877 | |||
821 | def copylist(self): |
|
878 | def copylist(self): | |
822 | return _data |
|
879 | return _data | |
823 |
|
880 | |||
824 | def write(self, tr): |
|
881 | def write(self, tr): | |
825 | pass |
|
882 | pass | |
826 |
|
883 | |||
827 | def invalidatecaches(self): |
|
884 | def invalidatecaches(self): | |
828 | pass |
|
885 | pass | |
829 |
|
886 | |||
830 | def markremoved(self, fn): |
|
887 | def markremoved(self, fn): | |
831 | pass |
|
888 | pass | |
832 |
|
889 | |||
833 | def __contains__(self, path): |
|
890 | def __contains__(self, path): | |
834 | '''Checks if the store contains path''' |
|
891 | '''Checks if the store contains path''' | |
835 | path = b"/".join((b"data", path)) |
|
892 | path = b"/".join((b"data", path)) | |
836 | # file? |
|
893 | # file? | |
837 | if self.vfs.exists(path + b".i"): |
|
894 | if self.vfs.exists(path + b".i"): | |
838 | return True |
|
895 | return True | |
839 | # dir? |
|
896 | # dir? | |
840 | if not path.endswith(b"/"): |
|
897 | if not path.endswith(b"/"): | |
841 | path = path + b"/" |
|
898 | path = path + b"/" | |
842 | return self.vfs.exists(path) |
|
899 | return self.vfs.exists(path) | |
843 |
|
900 | |||
844 |
|
901 | |||
845 | class encodedstore(basicstore): |
|
902 | class encodedstore(basicstore): | |
846 | def __init__(self, path, vfstype): |
|
903 | def __init__(self, path, vfstype): | |
847 | vfs = vfstype(path + b'/store') |
|
904 | vfs = vfstype(path + b'/store') | |
848 | self.path = vfs.base |
|
905 | self.path = vfs.base | |
849 | self.createmode = _calcmode(vfs) |
|
906 | self.createmode = _calcmode(vfs) | |
850 | vfs.createmode = self.createmode |
|
907 | vfs.createmode = self.createmode | |
851 | self.rawvfs = vfs |
|
908 | self.rawvfs = vfs | |
852 | self.vfs = vfsmod.filtervfs(vfs, encodefilename) |
|
909 | self.vfs = vfsmod.filtervfs(vfs, encodefilename) | |
853 | self.opener = self.vfs |
|
910 | self.opener = self.vfs | |
854 |
|
911 | |||
855 | def _walk(self, relpath, recurse, undecodable=None): |
|
912 | def _walk(self, relpath, recurse, undecodable=None): | |
856 | old = super()._walk(relpath, recurse) |
|
913 | old = super()._walk(relpath, recurse) | |
857 | new = [] |
|
914 | new = [] | |
858 | for f1, value in old: |
|
915 | for f1, value in old: | |
859 | try: |
|
916 | try: | |
860 | f2 = decodefilename(f1) |
|
917 | f2 = decodefilename(f1) | |
861 | except KeyError: |
|
918 | except KeyError: | |
862 | if undecodable is None: |
|
919 | if undecodable is None: | |
863 | msg = _(b'undecodable revlog name %s') % f1 |
|
920 | msg = _(b'undecodable revlog name %s') % f1 | |
864 | raise error.StorageError(msg) |
|
921 | raise error.StorageError(msg) | |
865 | else: |
|
922 | else: | |
866 | undecodable.append(f1) |
|
923 | undecodable.append(f1) | |
867 | continue |
|
924 | continue | |
868 | new.append((f2, value)) |
|
925 | new.append((f2, value)) | |
869 | return new |
|
926 | return new | |
870 |
|
927 | |||
871 | def data_entries( |
|
928 | def data_entries( | |
872 | self, matcher=None, undecodable=None |
|
929 | self, matcher=None, undecodable=None | |
873 | ) -> Generator[BaseStoreEntry, None, None]: |
|
930 | ) -> Generator[BaseStoreEntry, None, None]: | |
874 | entries = super(encodedstore, self).data_entries( |
|
931 | entries = super(encodedstore, self).data_entries( | |
875 | undecodable=undecodable |
|
932 | undecodable=undecodable | |
876 | ) |
|
933 | ) | |
877 | for entry in entries: |
|
934 | for entry in entries: | |
878 | if _match_tracked_entry(entry, matcher): |
|
935 | if _match_tracked_entry(entry, matcher): | |
879 | yield entry |
|
936 | yield entry | |
880 |
|
937 | |||
881 | def join(self, f): |
|
938 | def join(self, f): | |
882 | return self.path + b'/' + encodefilename(f) |
|
939 | return self.path + b'/' + encodefilename(f) | |
883 |
|
940 | |||
884 | def copylist(self): |
|
941 | def copylist(self): | |
885 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in _data] |
|
942 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in _data] | |
886 |
|
943 | |||
887 |
|
944 | |||
888 | class fncache: |
|
945 | class fncache: | |
889 | # the filename used to be partially encoded |
|
946 | # the filename used to be partially encoded | |
890 | # hence the encodedir/decodedir dance |
|
947 | # hence the encodedir/decodedir dance | |
891 | def __init__(self, vfs): |
|
948 | def __init__(self, vfs): | |
892 | self.vfs = vfs |
|
949 | self.vfs = vfs | |
893 | self._ignores = set() |
|
950 | self._ignores = set() | |
894 | self.entries = None |
|
951 | self.entries = None | |
895 | self._dirty = False |
|
952 | self._dirty = False | |
896 | # set of new additions to fncache |
|
953 | # set of new additions to fncache | |
897 | self.addls = set() |
|
954 | self.addls = set() | |
898 |
|
955 | |||
899 | def ensureloaded(self, warn=None): |
|
956 | def ensureloaded(self, warn=None): | |
900 | """read the fncache file if not already read. |
|
957 | """read the fncache file if not already read. | |
901 |
|
958 | |||
902 | If the file on disk is corrupted, raise. If warn is provided, |
|
959 | If the file on disk is corrupted, raise. If warn is provided, | |
903 | warn and keep going instead.""" |
|
960 | warn and keep going instead.""" | |
904 | if self.entries is None: |
|
961 | if self.entries is None: | |
905 | self._load(warn) |
|
962 | self._load(warn) | |
906 |
|
963 | |||
907 | def _load(self, warn=None): |
|
964 | def _load(self, warn=None): | |
908 | '''fill the entries from the fncache file''' |
|
965 | '''fill the entries from the fncache file''' | |
909 | self._dirty = False |
|
966 | self._dirty = False | |
910 | try: |
|
967 | try: | |
911 | fp = self.vfs(b'fncache', mode=b'rb') |
|
968 | fp = self.vfs(b'fncache', mode=b'rb') | |
912 | except IOError: |
|
969 | except IOError: | |
913 | # skip nonexistent file |
|
970 | # skip nonexistent file | |
914 | self.entries = set() |
|
971 | self.entries = set() | |
915 | return |
|
972 | return | |
916 |
|
973 | |||
917 | self.entries = set() |
|
974 | self.entries = set() | |
918 | chunk = b'' |
|
975 | chunk = b'' | |
919 | for c in iter(functools.partial(fp.read, fncache_chunksize), b''): |
|
976 | for c in iter(functools.partial(fp.read, fncache_chunksize), b''): | |
920 | chunk += c |
|
977 | chunk += c | |
921 | try: |
|
978 | try: | |
922 | p = chunk.rindex(b'\n') |
|
979 | p = chunk.rindex(b'\n') | |
923 | self.entries.update(decodedir(chunk[: p + 1]).splitlines()) |
|
980 | self.entries.update(decodedir(chunk[: p + 1]).splitlines()) | |
924 | chunk = chunk[p + 1 :] |
|
981 | chunk = chunk[p + 1 :] | |
925 | except ValueError: |
|
982 | except ValueError: | |
926 | # substring '\n' not found, maybe the entry is bigger than the |
|
983 | # substring '\n' not found, maybe the entry is bigger than the | |
927 | # chunksize, so let's keep iterating |
|
984 | # chunksize, so let's keep iterating | |
928 | pass |
|
985 | pass | |
929 |
|
986 | |||
930 | if chunk: |
|
987 | if chunk: | |
931 | msg = _(b"fncache does not ends with a newline") |
|
988 | msg = _(b"fncache does not ends with a newline") | |
932 | if warn: |
|
989 | if warn: | |
933 | warn(msg + b'\n') |
|
990 | warn(msg + b'\n') | |
934 | else: |
|
991 | else: | |
935 | raise error.Abort( |
|
992 | raise error.Abort( | |
936 | msg, |
|
993 | msg, | |
937 | hint=_( |
|
994 | hint=_( | |
938 | b"use 'hg debugrebuildfncache' to " |
|
995 | b"use 'hg debugrebuildfncache' to " | |
939 | b"rebuild the fncache" |
|
996 | b"rebuild the fncache" | |
940 | ), |
|
997 | ), | |
941 | ) |
|
998 | ) | |
942 | self._checkentries(fp, warn) |
|
999 | self._checkentries(fp, warn) | |
943 | fp.close() |
|
1000 | fp.close() | |
944 |
|
1001 | |||
945 | def _checkentries(self, fp, warn): |
|
1002 | def _checkentries(self, fp, warn): | |
946 | """make sure there is no empty string in entries""" |
|
1003 | """make sure there is no empty string in entries""" | |
947 | if b'' in self.entries: |
|
1004 | if b'' in self.entries: | |
948 | fp.seek(0) |
|
1005 | fp.seek(0) | |
949 | for n, line in enumerate(fp): |
|
1006 | for n, line in enumerate(fp): | |
950 | if not line.rstrip(b'\n'): |
|
1007 | if not line.rstrip(b'\n'): | |
951 | t = _(b'invalid entry in fncache, line %d') % (n + 1) |
|
1008 | t = _(b'invalid entry in fncache, line %d') % (n + 1) | |
952 | if warn: |
|
1009 | if warn: | |
953 | warn(t + b'\n') |
|
1010 | warn(t + b'\n') | |
954 | else: |
|
1011 | else: | |
955 | raise error.Abort(t) |
|
1012 | raise error.Abort(t) | |
956 |
|
1013 | |||
957 | def write(self, tr): |
|
1014 | def write(self, tr): | |
958 | if self._dirty: |
|
1015 | if self._dirty: | |
959 | assert self.entries is not None |
|
1016 | assert self.entries is not None | |
960 | self.entries = self.entries | self.addls |
|
1017 | self.entries = self.entries | self.addls | |
961 | self.addls = set() |
|
1018 | self.addls = set() | |
962 | tr.addbackup(b'fncache') |
|
1019 | tr.addbackup(b'fncache') | |
963 | fp = self.vfs(b'fncache', mode=b'wb', atomictemp=True) |
|
1020 | fp = self.vfs(b'fncache', mode=b'wb', atomictemp=True) | |
964 | if self.entries: |
|
1021 | if self.entries: | |
965 | fp.write(encodedir(b'\n'.join(self.entries) + b'\n')) |
|
1022 | fp.write(encodedir(b'\n'.join(self.entries) + b'\n')) | |
966 | fp.close() |
|
1023 | fp.close() | |
967 | self._dirty = False |
|
1024 | self._dirty = False | |
968 | if self.addls: |
|
1025 | if self.addls: | |
969 | # if we have just new entries, let's append them to the fncache |
|
1026 | # if we have just new entries, let's append them to the fncache | |
970 | tr.addbackup(b'fncache') |
|
1027 | tr.addbackup(b'fncache') | |
971 | fp = self.vfs(b'fncache', mode=b'ab', atomictemp=True) |
|
1028 | fp = self.vfs(b'fncache', mode=b'ab', atomictemp=True) | |
972 | if self.addls: |
|
1029 | if self.addls: | |
973 | fp.write(encodedir(b'\n'.join(self.addls) + b'\n')) |
|
1030 | fp.write(encodedir(b'\n'.join(self.addls) + b'\n')) | |
974 | fp.close() |
|
1031 | fp.close() | |
975 | self.entries = None |
|
1032 | self.entries = None | |
976 | self.addls = set() |
|
1033 | self.addls = set() | |
977 |
|
1034 | |||
978 | def addignore(self, fn): |
|
1035 | def addignore(self, fn): | |
979 | self._ignores.add(fn) |
|
1036 | self._ignores.add(fn) | |
980 |
|
1037 | |||
981 | def add(self, fn): |
|
1038 | def add(self, fn): | |
982 | if fn in self._ignores: |
|
1039 | if fn in self._ignores: | |
983 | return |
|
1040 | return | |
984 | if self.entries is None: |
|
1041 | if self.entries is None: | |
985 | self._load() |
|
1042 | self._load() | |
986 | if fn not in self.entries: |
|
1043 | if fn not in self.entries: | |
987 | self.addls.add(fn) |
|
1044 | self.addls.add(fn) | |
988 |
|
1045 | |||
989 | def remove(self, fn): |
|
1046 | def remove(self, fn): | |
990 | if self.entries is None: |
|
1047 | if self.entries is None: | |
991 | self._load() |
|
1048 | self._load() | |
992 | if fn in self.addls: |
|
1049 | if fn in self.addls: | |
993 | self.addls.remove(fn) |
|
1050 | self.addls.remove(fn) | |
994 | return |
|
1051 | return | |
995 | try: |
|
1052 | try: | |
996 | self.entries.remove(fn) |
|
1053 | self.entries.remove(fn) | |
997 | self._dirty = True |
|
1054 | self._dirty = True | |
998 | except KeyError: |
|
1055 | except KeyError: | |
999 | pass |
|
1056 | pass | |
1000 |
|
1057 | |||
1001 | def __contains__(self, fn): |
|
1058 | def __contains__(self, fn): | |
1002 | if fn in self.addls: |
|
1059 | if fn in self.addls: | |
1003 | return True |
|
1060 | return True | |
1004 | if self.entries is None: |
|
1061 | if self.entries is None: | |
1005 | self._load() |
|
1062 | self._load() | |
1006 | return fn in self.entries |
|
1063 | return fn in self.entries | |
1007 |
|
1064 | |||
1008 | def __iter__(self): |
|
1065 | def __iter__(self): | |
1009 | if self.entries is None: |
|
1066 | if self.entries is None: | |
1010 | self._load() |
|
1067 | self._load() | |
1011 | return iter(self.entries | self.addls) |
|
1068 | return iter(self.entries | self.addls) | |
1012 |
|
1069 | |||
1013 |
|
1070 | |||
1014 | class _fncachevfs(vfsmod.proxyvfs): |
|
1071 | class _fncachevfs(vfsmod.proxyvfs): | |
1015 | def __init__(self, vfs, fnc, encode): |
|
1072 | def __init__(self, vfs, fnc, encode): | |
1016 | vfsmod.proxyvfs.__init__(self, vfs) |
|
1073 | vfsmod.proxyvfs.__init__(self, vfs) | |
1017 | self.fncache = fnc |
|
1074 | self.fncache = fnc | |
1018 | self.encode = encode |
|
1075 | self.encode = encode | |
1019 |
|
1076 | |||
1020 | def __call__(self, path, mode=b'r', *args, **kw): |
|
1077 | def __call__(self, path, mode=b'r', *args, **kw): | |
1021 | encoded = self.encode(path) |
|
1078 | encoded = self.encode(path) | |
1022 | if ( |
|
1079 | if ( | |
1023 | mode not in (b'r', b'rb') |
|
1080 | mode not in (b'r', b'rb') | |
1024 | and (path.startswith(b'data/') or path.startswith(b'meta/')) |
|
1081 | and (path.startswith(b'data/') or path.startswith(b'meta/')) | |
1025 | and revlog_type(path) is not None |
|
1082 | and revlog_type(path) is not None | |
1026 | ): |
|
1083 | ): | |
1027 | # do not trigger a fncache load when adding a file that already is |
|
1084 | # do not trigger a fncache load when adding a file that already is | |
1028 | # known to exist. |
|
1085 | # known to exist. | |
1029 | notload = self.fncache.entries is None and self.vfs.exists(encoded) |
|
1086 | notload = self.fncache.entries is None and self.vfs.exists(encoded) | |
1030 | if notload and b'r+' in mode and not self.vfs.stat(encoded).st_size: |
|
1087 | if notload and b'r+' in mode and not self.vfs.stat(encoded).st_size: | |
1031 | # when appending to an existing file, if the file has size zero, |
|
1088 | # when appending to an existing file, if the file has size zero, | |
1032 | # it should be considered as missing. Such zero-size files are |
|
1089 | # it should be considered as missing. Such zero-size files are | |
1033 | # the result of truncation when a transaction is aborted. |
|
1090 | # the result of truncation when a transaction is aborted. | |
1034 | notload = False |
|
1091 | notload = False | |
1035 | if not notload: |
|
1092 | if not notload: | |
1036 | self.fncache.add(path) |
|
1093 | self.fncache.add(path) | |
1037 | return self.vfs(encoded, mode, *args, **kw) |
|
1094 | return self.vfs(encoded, mode, *args, **kw) | |
1038 |
|
1095 | |||
1039 | def join(self, path): |
|
1096 | def join(self, path): | |
1040 | if path: |
|
1097 | if path: | |
1041 | return self.vfs.join(self.encode(path)) |
|
1098 | return self.vfs.join(self.encode(path)) | |
1042 | else: |
|
1099 | else: | |
1043 | return self.vfs.join(path) |
|
1100 | return self.vfs.join(path) | |
1044 |
|
1101 | |||
1045 | def register_file(self, path): |
|
1102 | def register_file(self, path): | |
1046 | """generic hook point to lets fncache steer its stew""" |
|
1103 | """generic hook point to lets fncache steer its stew""" | |
1047 | if path.startswith(b'data/') or path.startswith(b'meta/'): |
|
1104 | if path.startswith(b'data/') or path.startswith(b'meta/'): | |
1048 | self.fncache.add(path) |
|
1105 | self.fncache.add(path) | |
1049 |
|
1106 | |||
1050 |
|
1107 | |||
1051 | class fncachestore(basicstore): |
|
1108 | class fncachestore(basicstore): | |
1052 | def __init__(self, path, vfstype, dotencode): |
|
1109 | def __init__(self, path, vfstype, dotencode): | |
1053 | if dotencode: |
|
1110 | if dotencode: | |
1054 | encode = _pathencode |
|
1111 | encode = _pathencode | |
1055 | else: |
|
1112 | else: | |
1056 | encode = _plainhybridencode |
|
1113 | encode = _plainhybridencode | |
1057 | self.encode = encode |
|
1114 | self.encode = encode | |
1058 | vfs = vfstype(path + b'/store') |
|
1115 | vfs = vfstype(path + b'/store') | |
1059 | self.path = vfs.base |
|
1116 | self.path = vfs.base | |
1060 | self.pathsep = self.path + b'/' |
|
1117 | self.pathsep = self.path + b'/' | |
1061 | self.createmode = _calcmode(vfs) |
|
1118 | self.createmode = _calcmode(vfs) | |
1062 | vfs.createmode = self.createmode |
|
1119 | vfs.createmode = self.createmode | |
1063 | self.rawvfs = vfs |
|
1120 | self.rawvfs = vfs | |
1064 | fnc = fncache(vfs) |
|
1121 | fnc = fncache(vfs) | |
1065 | self.fncache = fnc |
|
1122 | self.fncache = fnc | |
1066 | self.vfs = _fncachevfs(vfs, fnc, encode) |
|
1123 | self.vfs = _fncachevfs(vfs, fnc, encode) | |
1067 | self.opener = self.vfs |
|
1124 | self.opener = self.vfs | |
1068 |
|
1125 | |||
1069 | def join(self, f): |
|
1126 | def join(self, f): | |
1070 | return self.pathsep + self.encode(f) |
|
1127 | return self.pathsep + self.encode(f) | |
1071 |
|
1128 | |||
1072 | def getsize(self, path): |
|
1129 | def getsize(self, path): | |
1073 | return self.rawvfs.stat(path).st_size |
|
1130 | return self.rawvfs.stat(path).st_size | |
1074 |
|
1131 | |||
1075 | def data_entries( |
|
1132 | def data_entries( | |
1076 | self, matcher=None, undecodable=None |
|
1133 | self, matcher=None, undecodable=None | |
1077 | ) -> Generator[BaseStoreEntry, None, None]: |
|
1134 | ) -> Generator[BaseStoreEntry, None, None]: | |
1078 | files = ((f, revlog_type(f)) for f in self.fncache) |
|
1135 | files = ((f, revlog_type(f)) for f in self.fncache) | |
1079 | # Note: all files in fncache should be revlog related, However the |
|
1136 | # Note: all files in fncache should be revlog related, However the | |
1080 | # fncache might contains such file added by previous version of |
|
1137 | # fncache might contains such file added by previous version of | |
1081 | # Mercurial. |
|
1138 | # Mercurial. | |
1082 | files = (f for f in files if f[1] is not None) |
|
1139 | files = (f for f in files if f[1] is not None) | |
1083 | by_revlog = _gather_revlog(files) |
|
1140 | by_revlog = _gather_revlog(files) | |
1084 | for revlog, details in by_revlog: |
|
1141 | for revlog, details in by_revlog: | |
1085 | file_details = {} |
|
1142 | file_details = {} | |
1086 | if revlog.startswith(b'data/'): |
|
1143 | if revlog.startswith(b'data/'): | |
1087 | rl_type = FILEFLAGS_FILELOG |
|
1144 | rl_type = FILEFLAGS_FILELOG | |
1088 | revlog_target_id = revlog.split(b'/', 1)[1] |
|
1145 | revlog_target_id = revlog.split(b'/', 1)[1] | |
1089 | elif revlog.startswith(b'meta/'): |
|
1146 | elif revlog.startswith(b'meta/'): | |
1090 | rl_type = FILEFLAGS_MANIFESTLOG |
|
1147 | rl_type = FILEFLAGS_MANIFESTLOG | |
1091 | # drop the initial directory and the `00manifest` file part |
|
1148 | # drop the initial directory and the `00manifest` file part | |
1092 | tmp = revlog.split(b'/', 1)[1] |
|
1149 | tmp = revlog.split(b'/', 1)[1] | |
1093 | revlog_target_id = tmp.rsplit(b'/', 1)[0] + b'/' |
|
1150 | revlog_target_id = tmp.rsplit(b'/', 1)[0] + b'/' | |
1094 | else: |
|
1151 | else: | |
1095 | # unreachable |
|
1152 | # unreachable | |
1096 | assert False, revlog |
|
1153 | assert False, revlog | |
1097 | for ext, t in details.items(): |
|
1154 | for ext, t in details.items(): | |
1098 | file_details[ext] = { |
|
1155 | file_details[ext] = { | |
1099 | 'is_volatile': bool(t & FILEFLAGS_VOLATILE), |
|
1156 | 'is_volatile': bool(t & FILEFLAGS_VOLATILE), | |
1100 | } |
|
1157 | } | |
1101 | entry = RevlogStoreEntry( |
|
1158 | entry = RevlogStoreEntry( | |
1102 | path_prefix=revlog, |
|
1159 | path_prefix=revlog, | |
1103 | revlog_type=rl_type, |
|
1160 | revlog_type=rl_type, | |
1104 | target_id=revlog_target_id, |
|
1161 | target_id=revlog_target_id, | |
1105 | details=file_details, |
|
1162 | details=file_details, | |
1106 | ) |
|
1163 | ) | |
1107 | if _match_tracked_entry(entry, matcher): |
|
1164 | if _match_tracked_entry(entry, matcher): | |
1108 | yield entry |
|
1165 | yield entry | |
1109 |
|
1166 | |||
1110 | def copylist(self): |
|
1167 | def copylist(self): | |
1111 | d = ( |
|
1168 | d = ( | |
1112 | b'bookmarks', |
|
1169 | b'bookmarks', | |
1113 | b'narrowspec', |
|
1170 | b'narrowspec', | |
1114 | b'data', |
|
1171 | b'data', | |
1115 | b'meta', |
|
1172 | b'meta', | |
1116 | b'dh', |
|
1173 | b'dh', | |
1117 | b'fncache', |
|
1174 | b'fncache', | |
1118 | b'phaseroots', |
|
1175 | b'phaseroots', | |
1119 | b'obsstore', |
|
1176 | b'obsstore', | |
1120 | b'00manifest.d', |
|
1177 | b'00manifest.d', | |
1121 | b'00manifest.i', |
|
1178 | b'00manifest.i', | |
1122 | b'00changelog.d', |
|
1179 | b'00changelog.d', | |
1123 | b'00changelog.i', |
|
1180 | b'00changelog.i', | |
1124 | b'requires', |
|
1181 | b'requires', | |
1125 | ) |
|
1182 | ) | |
1126 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in d] |
|
1183 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in d] | |
1127 |
|
1184 | |||
1128 | def write(self, tr): |
|
1185 | def write(self, tr): | |
1129 | self.fncache.write(tr) |
|
1186 | self.fncache.write(tr) | |
1130 |
|
1187 | |||
1131 | def invalidatecaches(self): |
|
1188 | def invalidatecaches(self): | |
1132 | self.fncache.entries = None |
|
1189 | self.fncache.entries = None | |
1133 | self.fncache.addls = set() |
|
1190 | self.fncache.addls = set() | |
1134 |
|
1191 | |||
1135 | def markremoved(self, fn): |
|
1192 | def markremoved(self, fn): | |
1136 | self.fncache.remove(fn) |
|
1193 | self.fncache.remove(fn) | |
1137 |
|
1194 | |||
1138 | def _exists(self, f): |
|
1195 | def _exists(self, f): | |
1139 | ef = self.encode(f) |
|
1196 | ef = self.encode(f) | |
1140 | try: |
|
1197 | try: | |
1141 | self.getsize(ef) |
|
1198 | self.getsize(ef) | |
1142 | return True |
|
1199 | return True | |
1143 | except FileNotFoundError: |
|
1200 | except FileNotFoundError: | |
1144 | return False |
|
1201 | return False | |
1145 |
|
1202 | |||
1146 | def __contains__(self, path): |
|
1203 | def __contains__(self, path): | |
1147 | '''Checks if the store contains path''' |
|
1204 | '''Checks if the store contains path''' | |
1148 | path = b"/".join((b"data", path)) |
|
1205 | path = b"/".join((b"data", path)) | |
1149 | # check for files (exact match) |
|
1206 | # check for files (exact match) | |
1150 | e = path + b'.i' |
|
1207 | e = path + b'.i' | |
1151 | if e in self.fncache and self._exists(e): |
|
1208 | if e in self.fncache and self._exists(e): | |
1152 | return True |
|
1209 | return True | |
1153 | # now check for directories (prefix match) |
|
1210 | # now check for directories (prefix match) | |
1154 | if not path.endswith(b'/'): |
|
1211 | if not path.endswith(b'/'): | |
1155 | path += b'/' |
|
1212 | path += b'/' | |
1156 | for e in self.fncache: |
|
1213 | for e in self.fncache: | |
1157 | if e.startswith(path) and self._exists(e): |
|
1214 | if e.startswith(path) and self._exists(e): | |
1158 | return True |
|
1215 | return True | |
1159 | return False |
|
1216 | return False |
@@ -1,966 +1,970 b'' | |||||
1 | # streamclone.py - producing and consuming streaming repository data |
|
1 | # streamclone.py - producing and consuming streaming repository data | |
2 | # |
|
2 | # | |
3 | # Copyright 2015 Gregory Szorc <gregory.szorc@gmail.com> |
|
3 | # Copyright 2015 Gregory Szorc <gregory.szorc@gmail.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 |
|
8 | |||
9 | import contextlib |
|
9 | import contextlib | |
10 | import os |
|
10 | import os | |
11 | import struct |
|
11 | import struct | |
12 |
|
12 | |||
13 | from .i18n import _ |
|
13 | from .i18n import _ | |
14 | from .interfaces import repository |
|
14 | from .interfaces import repository | |
15 | from . import ( |
|
15 | from . import ( | |
16 | bookmarks, |
|
16 | bookmarks, | |
17 | bundle2 as bundle2mod, |
|
17 | bundle2 as bundle2mod, | |
18 | cacheutil, |
|
18 | cacheutil, | |
19 | error, |
|
19 | error, | |
20 | narrowspec, |
|
20 | narrowspec, | |
21 | phases, |
|
21 | phases, | |
22 | pycompat, |
|
22 | pycompat, | |
23 | requirements as requirementsmod, |
|
23 | requirements as requirementsmod, | |
24 | scmutil, |
|
24 | scmutil, | |
25 | store, |
|
25 | store, | |
26 | transaction, |
|
26 | transaction, | |
27 | util, |
|
27 | util, | |
28 | ) |
|
28 | ) | |
29 | from .revlogutils import ( |
|
29 | from .revlogutils import ( | |
30 | nodemap, |
|
30 | nodemap, | |
31 | ) |
|
31 | ) | |
32 |
|
32 | |||
33 |
|
33 | |||
34 | def new_stream_clone_requirements(default_requirements, streamed_requirements): |
|
34 | def new_stream_clone_requirements(default_requirements, streamed_requirements): | |
35 | """determine the final set of requirement for a new stream clone |
|
35 | """determine the final set of requirement for a new stream clone | |
36 |
|
36 | |||
37 | this method combine the "default" requirements that a new repository would |
|
37 | this method combine the "default" requirements that a new repository would | |
38 | use with the constaint we get from the stream clone content. We keep local |
|
38 | use with the constaint we get from the stream clone content. We keep local | |
39 | configuration choice when possible. |
|
39 | configuration choice when possible. | |
40 | """ |
|
40 | """ | |
41 | requirements = set(default_requirements) |
|
41 | requirements = set(default_requirements) | |
42 | requirements -= requirementsmod.STREAM_FIXED_REQUIREMENTS |
|
42 | requirements -= requirementsmod.STREAM_FIXED_REQUIREMENTS | |
43 | requirements.update(streamed_requirements) |
|
43 | requirements.update(streamed_requirements) | |
44 | return requirements |
|
44 | return requirements | |
45 |
|
45 | |||
46 |
|
46 | |||
47 | def streamed_requirements(repo): |
|
47 | def streamed_requirements(repo): | |
48 | """the set of requirement the new clone will have to support |
|
48 | """the set of requirement the new clone will have to support | |
49 |
|
49 | |||
50 | This is used for advertising the stream options and to generate the actual |
|
50 | This is used for advertising the stream options and to generate the actual | |
51 | stream content.""" |
|
51 | stream content.""" | |
52 | requiredformats = ( |
|
52 | requiredformats = ( | |
53 | repo.requirements & requirementsmod.STREAM_FIXED_REQUIREMENTS |
|
53 | repo.requirements & requirementsmod.STREAM_FIXED_REQUIREMENTS | |
54 | ) |
|
54 | ) | |
55 | return requiredformats |
|
55 | return requiredformats | |
56 |
|
56 | |||
57 |
|
57 | |||
58 | def canperformstreamclone(pullop, bundle2=False): |
|
58 | def canperformstreamclone(pullop, bundle2=False): | |
59 | """Whether it is possible to perform a streaming clone as part of pull. |
|
59 | """Whether it is possible to perform a streaming clone as part of pull. | |
60 |
|
60 | |||
61 | ``bundle2`` will cause the function to consider stream clone through |
|
61 | ``bundle2`` will cause the function to consider stream clone through | |
62 | bundle2 and only through bundle2. |
|
62 | bundle2 and only through bundle2. | |
63 |
|
63 | |||
64 | Returns a tuple of (supported, requirements). ``supported`` is True if |
|
64 | Returns a tuple of (supported, requirements). ``supported`` is True if | |
65 | streaming clone is supported and False otherwise. ``requirements`` is |
|
65 | streaming clone is supported and False otherwise. ``requirements`` is | |
66 | a set of repo requirements from the remote, or ``None`` if stream clone |
|
66 | a set of repo requirements from the remote, or ``None`` if stream clone | |
67 | isn't supported. |
|
67 | isn't supported. | |
68 | """ |
|
68 | """ | |
69 | repo = pullop.repo |
|
69 | repo = pullop.repo | |
70 | remote = pullop.remote |
|
70 | remote = pullop.remote | |
71 |
|
71 | |||
72 | # should we consider streaming clone at all ? |
|
72 | # should we consider streaming clone at all ? | |
73 | streamrequested = pullop.streamclonerequested |
|
73 | streamrequested = pullop.streamclonerequested | |
74 | # If we don't have a preference, let the server decide for us. This |
|
74 | # If we don't have a preference, let the server decide for us. This | |
75 | # likely only comes into play in LANs. |
|
75 | # likely only comes into play in LANs. | |
76 | if streamrequested is None: |
|
76 | if streamrequested is None: | |
77 | # The server can advertise whether to prefer streaming clone. |
|
77 | # The server can advertise whether to prefer streaming clone. | |
78 | streamrequested = remote.capable(b'stream-preferred') |
|
78 | streamrequested = remote.capable(b'stream-preferred') | |
79 | if not streamrequested: |
|
79 | if not streamrequested: | |
80 | return False, None |
|
80 | return False, None | |
81 |
|
81 | |||
82 | # Streaming clone only works on an empty destination repository |
|
82 | # Streaming clone only works on an empty destination repository | |
83 | if len(repo): |
|
83 | if len(repo): | |
84 | return False, None |
|
84 | return False, None | |
85 |
|
85 | |||
86 | # Streaming clone only works if all data is being requested. |
|
86 | # Streaming clone only works if all data is being requested. | |
87 | if pullop.heads: |
|
87 | if pullop.heads: | |
88 | return False, None |
|
88 | return False, None | |
89 |
|
89 | |||
90 | bundle2supported = False |
|
90 | bundle2supported = False | |
91 | if pullop.canusebundle2: |
|
91 | if pullop.canusebundle2: | |
92 | local_caps = bundle2mod.getrepocaps(repo, role=b'client') |
|
92 | local_caps = bundle2mod.getrepocaps(repo, role=b'client') | |
93 | local_supported = set(local_caps.get(b'stream', [])) |
|
93 | local_supported = set(local_caps.get(b'stream', [])) | |
94 | remote_supported = set(pullop.remotebundle2caps.get(b'stream', [])) |
|
94 | remote_supported = set(pullop.remotebundle2caps.get(b'stream', [])) | |
95 | bundle2supported = bool(local_supported & remote_supported) |
|
95 | bundle2supported = bool(local_supported & remote_supported) | |
96 | # else |
|
96 | # else | |
97 | # Server doesn't support bundle2 stream clone or doesn't support |
|
97 | # Server doesn't support bundle2 stream clone or doesn't support | |
98 | # the versions we support. Fall back and possibly allow legacy. |
|
98 | # the versions we support. Fall back and possibly allow legacy. | |
99 |
|
99 | |||
100 | # Ensures legacy code path uses available bundle2. |
|
100 | # Ensures legacy code path uses available bundle2. | |
101 | if bundle2supported and not bundle2: |
|
101 | if bundle2supported and not bundle2: | |
102 | return False, None |
|
102 | return False, None | |
103 | # Ensures bundle2 doesn't try to do a stream clone if it isn't supported. |
|
103 | # Ensures bundle2 doesn't try to do a stream clone if it isn't supported. | |
104 | elif bundle2 and not bundle2supported: |
|
104 | elif bundle2 and not bundle2supported: | |
105 | return False, None |
|
105 | return False, None | |
106 |
|
106 | |||
107 | # In order for stream clone to work, the client has to support all the |
|
107 | # In order for stream clone to work, the client has to support all the | |
108 | # requirements advertised by the server. |
|
108 | # requirements advertised by the server. | |
109 | # |
|
109 | # | |
110 | # The server advertises its requirements via the "stream" and "streamreqs" |
|
110 | # The server advertises its requirements via the "stream" and "streamreqs" | |
111 | # capability. "stream" (a value-less capability) is advertised if and only |
|
111 | # capability. "stream" (a value-less capability) is advertised if and only | |
112 | # if the only requirement is "revlogv1." Else, the "streamreqs" capability |
|
112 | # if the only requirement is "revlogv1." Else, the "streamreqs" capability | |
113 | # is advertised and contains a comma-delimited list of requirements. |
|
113 | # is advertised and contains a comma-delimited list of requirements. | |
114 | requirements = set() |
|
114 | requirements = set() | |
115 | if remote.capable(b'stream'): |
|
115 | if remote.capable(b'stream'): | |
116 | requirements.add(requirementsmod.REVLOGV1_REQUIREMENT) |
|
116 | requirements.add(requirementsmod.REVLOGV1_REQUIREMENT) | |
117 | else: |
|
117 | else: | |
118 | streamreqs = remote.capable(b'streamreqs') |
|
118 | streamreqs = remote.capable(b'streamreqs') | |
119 | # This is weird and shouldn't happen with modern servers. |
|
119 | # This is weird and shouldn't happen with modern servers. | |
120 | if not streamreqs: |
|
120 | if not streamreqs: | |
121 | pullop.repo.ui.warn( |
|
121 | pullop.repo.ui.warn( | |
122 | _( |
|
122 | _( | |
123 | b'warning: stream clone requested but server has them ' |
|
123 | b'warning: stream clone requested but server has them ' | |
124 | b'disabled\n' |
|
124 | b'disabled\n' | |
125 | ) |
|
125 | ) | |
126 | ) |
|
126 | ) | |
127 | return False, None |
|
127 | return False, None | |
128 |
|
128 | |||
129 | streamreqs = set(streamreqs.split(b',')) |
|
129 | streamreqs = set(streamreqs.split(b',')) | |
130 | # Server requires something we don't support. Bail. |
|
130 | # Server requires something we don't support. Bail. | |
131 | missingreqs = streamreqs - repo.supported |
|
131 | missingreqs = streamreqs - repo.supported | |
132 | if missingreqs: |
|
132 | if missingreqs: | |
133 | pullop.repo.ui.warn( |
|
133 | pullop.repo.ui.warn( | |
134 | _( |
|
134 | _( | |
135 | b'warning: stream clone requested but client is missing ' |
|
135 | b'warning: stream clone requested but client is missing ' | |
136 | b'requirements: %s\n' |
|
136 | b'requirements: %s\n' | |
137 | ) |
|
137 | ) | |
138 | % b', '.join(sorted(missingreqs)) |
|
138 | % b', '.join(sorted(missingreqs)) | |
139 | ) |
|
139 | ) | |
140 | pullop.repo.ui.warn( |
|
140 | pullop.repo.ui.warn( | |
141 | _( |
|
141 | _( | |
142 | b'(see https://www.mercurial-scm.org/wiki/MissingRequirement ' |
|
142 | b'(see https://www.mercurial-scm.org/wiki/MissingRequirement ' | |
143 | b'for more information)\n' |
|
143 | b'for more information)\n' | |
144 | ) |
|
144 | ) | |
145 | ) |
|
145 | ) | |
146 | return False, None |
|
146 | return False, None | |
147 | requirements = streamreqs |
|
147 | requirements = streamreqs | |
148 |
|
148 | |||
149 | return True, requirements |
|
149 | return True, requirements | |
150 |
|
150 | |||
151 |
|
151 | |||
152 | def maybeperformlegacystreamclone(pullop): |
|
152 | def maybeperformlegacystreamclone(pullop): | |
153 | """Possibly perform a legacy stream clone operation. |
|
153 | """Possibly perform a legacy stream clone operation. | |
154 |
|
154 | |||
155 | Legacy stream clones are performed as part of pull but before all other |
|
155 | Legacy stream clones are performed as part of pull but before all other | |
156 | operations. |
|
156 | operations. | |
157 |
|
157 | |||
158 | A legacy stream clone will not be performed if a bundle2 stream clone is |
|
158 | A legacy stream clone will not be performed if a bundle2 stream clone is | |
159 | supported. |
|
159 | supported. | |
160 | """ |
|
160 | """ | |
161 | from . import localrepo |
|
161 | from . import localrepo | |
162 |
|
162 | |||
163 | supported, requirements = canperformstreamclone(pullop) |
|
163 | supported, requirements = canperformstreamclone(pullop) | |
164 |
|
164 | |||
165 | if not supported: |
|
165 | if not supported: | |
166 | return |
|
166 | return | |
167 |
|
167 | |||
168 | repo = pullop.repo |
|
168 | repo = pullop.repo | |
169 | remote = pullop.remote |
|
169 | remote = pullop.remote | |
170 |
|
170 | |||
171 | # Save remote branchmap. We will use it later to speed up branchcache |
|
171 | # Save remote branchmap. We will use it later to speed up branchcache | |
172 | # creation. |
|
172 | # creation. | |
173 | rbranchmap = None |
|
173 | rbranchmap = None | |
174 | if remote.capable(b'branchmap'): |
|
174 | if remote.capable(b'branchmap'): | |
175 | with remote.commandexecutor() as e: |
|
175 | with remote.commandexecutor() as e: | |
176 | rbranchmap = e.callcommand(b'branchmap', {}).result() |
|
176 | rbranchmap = e.callcommand(b'branchmap', {}).result() | |
177 |
|
177 | |||
178 | repo.ui.status(_(b'streaming all changes\n')) |
|
178 | repo.ui.status(_(b'streaming all changes\n')) | |
179 |
|
179 | |||
180 | with remote.commandexecutor() as e: |
|
180 | with remote.commandexecutor() as e: | |
181 | fp = e.callcommand(b'stream_out', {}).result() |
|
181 | fp = e.callcommand(b'stream_out', {}).result() | |
182 |
|
182 | |||
183 | # TODO strictly speaking, this code should all be inside the context |
|
183 | # TODO strictly speaking, this code should all be inside the context | |
184 | # manager because the context manager is supposed to ensure all wire state |
|
184 | # manager because the context manager is supposed to ensure all wire state | |
185 | # is flushed when exiting. But the legacy peers don't do this, so it |
|
185 | # is flushed when exiting. But the legacy peers don't do this, so it | |
186 | # doesn't matter. |
|
186 | # doesn't matter. | |
187 | l = fp.readline() |
|
187 | l = fp.readline() | |
188 | try: |
|
188 | try: | |
189 | resp = int(l) |
|
189 | resp = int(l) | |
190 | except ValueError: |
|
190 | except ValueError: | |
191 | raise error.ResponseError( |
|
191 | raise error.ResponseError( | |
192 | _(b'unexpected response from remote server:'), l |
|
192 | _(b'unexpected response from remote server:'), l | |
193 | ) |
|
193 | ) | |
194 | if resp == 1: |
|
194 | if resp == 1: | |
195 | raise error.Abort(_(b'operation forbidden by server')) |
|
195 | raise error.Abort(_(b'operation forbidden by server')) | |
196 | elif resp == 2: |
|
196 | elif resp == 2: | |
197 | raise error.Abort(_(b'locking the remote repository failed')) |
|
197 | raise error.Abort(_(b'locking the remote repository failed')) | |
198 | elif resp != 0: |
|
198 | elif resp != 0: | |
199 | raise error.Abort(_(b'the server sent an unknown error code')) |
|
199 | raise error.Abort(_(b'the server sent an unknown error code')) | |
200 |
|
200 | |||
201 | l = fp.readline() |
|
201 | l = fp.readline() | |
202 | try: |
|
202 | try: | |
203 | filecount, bytecount = map(int, l.split(b' ', 1)) |
|
203 | filecount, bytecount = map(int, l.split(b' ', 1)) | |
204 | except (ValueError, TypeError): |
|
204 | except (ValueError, TypeError): | |
205 | raise error.ResponseError( |
|
205 | raise error.ResponseError( | |
206 | _(b'unexpected response from remote server:'), l |
|
206 | _(b'unexpected response from remote server:'), l | |
207 | ) |
|
207 | ) | |
208 |
|
208 | |||
209 | with repo.lock(): |
|
209 | with repo.lock(): | |
210 | consumev1(repo, fp, filecount, bytecount) |
|
210 | consumev1(repo, fp, filecount, bytecount) | |
211 | repo.requirements = new_stream_clone_requirements( |
|
211 | repo.requirements = new_stream_clone_requirements( | |
212 | repo.requirements, |
|
212 | repo.requirements, | |
213 | requirements, |
|
213 | requirements, | |
214 | ) |
|
214 | ) | |
215 | repo.svfs.options = localrepo.resolvestorevfsoptions( |
|
215 | repo.svfs.options = localrepo.resolvestorevfsoptions( | |
216 | repo.ui, repo.requirements, repo.features |
|
216 | repo.ui, repo.requirements, repo.features | |
217 | ) |
|
217 | ) | |
218 | scmutil.writereporequirements(repo) |
|
218 | scmutil.writereporequirements(repo) | |
219 | nodemap.post_stream_cleanup(repo) |
|
219 | nodemap.post_stream_cleanup(repo) | |
220 |
|
220 | |||
221 | if rbranchmap: |
|
221 | if rbranchmap: | |
222 | repo._branchcaches.replace(repo, rbranchmap) |
|
222 | repo._branchcaches.replace(repo, rbranchmap) | |
223 |
|
223 | |||
224 | repo.invalidate() |
|
224 | repo.invalidate() | |
225 |
|
225 | |||
226 |
|
226 | |||
227 | def allowservergeneration(repo): |
|
227 | def allowservergeneration(repo): | |
228 | """Whether streaming clones are allowed from the server.""" |
|
228 | """Whether streaming clones are allowed from the server.""" | |
229 | if repository.REPO_FEATURE_STREAM_CLONE not in repo.features: |
|
229 | if repository.REPO_FEATURE_STREAM_CLONE not in repo.features: | |
230 | return False |
|
230 | return False | |
231 |
|
231 | |||
232 | if not repo.ui.configbool(b'server', b'uncompressed', untrusted=True): |
|
232 | if not repo.ui.configbool(b'server', b'uncompressed', untrusted=True): | |
233 | return False |
|
233 | return False | |
234 |
|
234 | |||
235 | # The way stream clone works makes it impossible to hide secret changesets. |
|
235 | # The way stream clone works makes it impossible to hide secret changesets. | |
236 | # So don't allow this by default. |
|
236 | # So don't allow this by default. | |
237 | secret = phases.hassecret(repo) |
|
237 | secret = phases.hassecret(repo) | |
238 | if secret: |
|
238 | if secret: | |
239 | return repo.ui.configbool(b'server', b'uncompressedallowsecret') |
|
239 | return repo.ui.configbool(b'server', b'uncompressedallowsecret') | |
240 |
|
240 | |||
241 | return True |
|
241 | return True | |
242 |
|
242 | |||
243 |
|
243 | |||
244 | # This is it's own function so extensions can override it. |
|
244 | # This is it's own function so extensions can override it. | |
245 | def _walkstreamfiles(repo, matcher=None, phase=False, obsolescence=False): |
|
245 | def _walkstreamfiles(repo, matcher=None, phase=False, obsolescence=False): | |
246 | return repo.store.walk(matcher, phase=phase, obsolescence=obsolescence) |
|
246 | return repo.store.walk(matcher, phase=phase, obsolescence=obsolescence) | |
247 |
|
247 | |||
248 |
|
248 | |||
249 | def generatev1(repo): |
|
249 | def generatev1(repo): | |
250 | """Emit content for version 1 of a streaming clone. |
|
250 | """Emit content for version 1 of a streaming clone. | |
251 |
|
251 | |||
252 | This returns a 3-tuple of (file count, byte size, data iterator). |
|
252 | This returns a 3-tuple of (file count, byte size, data iterator). | |
253 |
|
253 | |||
254 | The data iterator consists of N entries for each file being transferred. |
|
254 | The data iterator consists of N entries for each file being transferred. | |
255 | Each file entry starts as a line with the file name and integer size |
|
255 | Each file entry starts as a line with the file name and integer size | |
256 | delimited by a null byte. |
|
256 | delimited by a null byte. | |
257 |
|
257 | |||
258 | The raw file data follows. Following the raw file data is the next file |
|
258 | The raw file data follows. Following the raw file data is the next file | |
259 | entry, or EOF. |
|
259 | entry, or EOF. | |
260 |
|
260 | |||
261 | When used on the wire protocol, an additional line indicating protocol |
|
261 | When used on the wire protocol, an additional line indicating protocol | |
262 | success will be prepended to the stream. This function is not responsible |
|
262 | success will be prepended to the stream. This function is not responsible | |
263 | for adding it. |
|
263 | for adding it. | |
264 |
|
264 | |||
265 | This function will obtain a repository lock to ensure a consistent view of |
|
265 | This function will obtain a repository lock to ensure a consistent view of | |
266 | the store is captured. It therefore may raise LockError. |
|
266 | the store is captured. It therefore may raise LockError. | |
267 | """ |
|
267 | """ | |
268 | entries = [] |
|
268 | entries = [] | |
269 | total_bytes = 0 |
|
269 | total_bytes = 0 | |
270 | # Get consistent snapshot of repo, lock during scan. |
|
270 | # Get consistent snapshot of repo, lock during scan. | |
271 | with repo.lock(): |
|
271 | with repo.lock(): | |
272 | repo.ui.debug(b'scanning\n') |
|
272 | repo.ui.debug(b'scanning\n') | |
273 | for entry in _walkstreamfiles(repo): |
|
273 | for entry in _walkstreamfiles(repo): | |
274 | for f in entry.files(): |
|
274 | for f in entry.files(): | |
275 | file_size = f.file_size(repo.store.vfs) |
|
275 | file_size = f.file_size(repo.store.vfs) | |
276 | if file_size: |
|
276 | if file_size: | |
277 | entries.append((f.unencoded_path, file_size)) |
|
277 | entries.append((f.unencoded_path, file_size)) | |
278 | total_bytes += file_size |
|
278 | total_bytes += file_size | |
279 | _test_sync_point_walk_1(repo) |
|
279 | _test_sync_point_walk_1(repo) | |
280 | _test_sync_point_walk_2(repo) |
|
280 | _test_sync_point_walk_2(repo) | |
281 |
|
281 | |||
282 | repo.ui.debug( |
|
282 | repo.ui.debug( | |
283 | b'%d files, %d bytes to transfer\n' % (len(entries), total_bytes) |
|
283 | b'%d files, %d bytes to transfer\n' % (len(entries), total_bytes) | |
284 | ) |
|
284 | ) | |
285 |
|
285 | |||
286 | svfs = repo.svfs |
|
286 | svfs = repo.svfs | |
287 | debugflag = repo.ui.debugflag |
|
287 | debugflag = repo.ui.debugflag | |
288 |
|
288 | |||
289 | def emitrevlogdata(): |
|
289 | def emitrevlogdata(): | |
290 | for name, size in entries: |
|
290 | for name, size in entries: | |
291 | if debugflag: |
|
291 | if debugflag: | |
292 | repo.ui.debug(b'sending %s (%d bytes)\n' % (name, size)) |
|
292 | repo.ui.debug(b'sending %s (%d bytes)\n' % (name, size)) | |
293 | # partially encode name over the wire for backwards compat |
|
293 | # partially encode name over the wire for backwards compat | |
294 | yield b'%s\0%d\n' % (store.encodedir(name), size) |
|
294 | yield b'%s\0%d\n' % (store.encodedir(name), size) | |
295 | # auditing at this stage is both pointless (paths are already |
|
295 | # auditing at this stage is both pointless (paths are already | |
296 | # trusted by the local repo) and expensive |
|
296 | # trusted by the local repo) and expensive | |
297 | with svfs(name, b'rb', auditpath=False) as fp: |
|
297 | with svfs(name, b'rb', auditpath=False) as fp: | |
298 | if size <= 65536: |
|
298 | if size <= 65536: | |
299 | yield fp.read(size) |
|
299 | yield fp.read(size) | |
300 | else: |
|
300 | else: | |
301 | for chunk in util.filechunkiter(fp, limit=size): |
|
301 | for chunk in util.filechunkiter(fp, limit=size): | |
302 | yield chunk |
|
302 | yield chunk | |
303 |
|
303 | |||
304 | return len(entries), total_bytes, emitrevlogdata() |
|
304 | return len(entries), total_bytes, emitrevlogdata() | |
305 |
|
305 | |||
306 |
|
306 | |||
307 | def generatev1wireproto(repo): |
|
307 | def generatev1wireproto(repo): | |
308 | """Emit content for version 1 of streaming clone suitable for the wire. |
|
308 | """Emit content for version 1 of streaming clone suitable for the wire. | |
309 |
|
309 | |||
310 | This is the data output from ``generatev1()`` with 2 header lines. The |
|
310 | This is the data output from ``generatev1()`` with 2 header lines. The | |
311 | first line indicates overall success. The 2nd contains the file count and |
|
311 | first line indicates overall success. The 2nd contains the file count and | |
312 | byte size of payload. |
|
312 | byte size of payload. | |
313 |
|
313 | |||
314 | The success line contains "0" for success, "1" for stream generation not |
|
314 | The success line contains "0" for success, "1" for stream generation not | |
315 | allowed, and "2" for error locking the repository (possibly indicating |
|
315 | allowed, and "2" for error locking the repository (possibly indicating | |
316 | a permissions error for the server process). |
|
316 | a permissions error for the server process). | |
317 | """ |
|
317 | """ | |
318 | if not allowservergeneration(repo): |
|
318 | if not allowservergeneration(repo): | |
319 | yield b'1\n' |
|
319 | yield b'1\n' | |
320 | return |
|
320 | return | |
321 |
|
321 | |||
322 | try: |
|
322 | try: | |
323 | filecount, bytecount, it = generatev1(repo) |
|
323 | filecount, bytecount, it = generatev1(repo) | |
324 | except error.LockError: |
|
324 | except error.LockError: | |
325 | yield b'2\n' |
|
325 | yield b'2\n' | |
326 | return |
|
326 | return | |
327 |
|
327 | |||
328 | # Indicates successful response. |
|
328 | # Indicates successful response. | |
329 | yield b'0\n' |
|
329 | yield b'0\n' | |
330 | yield b'%d %d\n' % (filecount, bytecount) |
|
330 | yield b'%d %d\n' % (filecount, bytecount) | |
331 | for chunk in it: |
|
331 | for chunk in it: | |
332 | yield chunk |
|
332 | yield chunk | |
333 |
|
333 | |||
334 |
|
334 | |||
335 | def generatebundlev1(repo, compression=b'UN'): |
|
335 | def generatebundlev1(repo, compression=b'UN'): | |
336 | """Emit content for version 1 of a stream clone bundle. |
|
336 | """Emit content for version 1 of a stream clone bundle. | |
337 |
|
337 | |||
338 | The first 4 bytes of the output ("HGS1") denote this as stream clone |
|
338 | The first 4 bytes of the output ("HGS1") denote this as stream clone | |
339 | bundle version 1. |
|
339 | bundle version 1. | |
340 |
|
340 | |||
341 | The next 2 bytes indicate the compression type. Only "UN" is currently |
|
341 | The next 2 bytes indicate the compression type. Only "UN" is currently | |
342 | supported. |
|
342 | supported. | |
343 |
|
343 | |||
344 | The next 16 bytes are two 64-bit big endian unsigned integers indicating |
|
344 | The next 16 bytes are two 64-bit big endian unsigned integers indicating | |
345 | file count and byte count, respectively. |
|
345 | file count and byte count, respectively. | |
346 |
|
346 | |||
347 | The next 2 bytes is a 16-bit big endian unsigned short declaring the length |
|
347 | The next 2 bytes is a 16-bit big endian unsigned short declaring the length | |
348 | of the requirements string, including a trailing \0. The following N bytes |
|
348 | of the requirements string, including a trailing \0. The following N bytes | |
349 | are the requirements string, which is ASCII containing a comma-delimited |
|
349 | are the requirements string, which is ASCII containing a comma-delimited | |
350 | list of repo requirements that are needed to support the data. |
|
350 | list of repo requirements that are needed to support the data. | |
351 |
|
351 | |||
352 | The remaining content is the output of ``generatev1()`` (which may be |
|
352 | The remaining content is the output of ``generatev1()`` (which may be | |
353 | compressed in the future). |
|
353 | compressed in the future). | |
354 |
|
354 | |||
355 | Returns a tuple of (requirements, data generator). |
|
355 | Returns a tuple of (requirements, data generator). | |
356 | """ |
|
356 | """ | |
357 | if compression != b'UN': |
|
357 | if compression != b'UN': | |
358 | raise ValueError(b'we do not support the compression argument yet') |
|
358 | raise ValueError(b'we do not support the compression argument yet') | |
359 |
|
359 | |||
360 | requirements = streamed_requirements(repo) |
|
360 | requirements = streamed_requirements(repo) | |
361 | requires = b','.join(sorted(requirements)) |
|
361 | requires = b','.join(sorted(requirements)) | |
362 |
|
362 | |||
363 | def gen(): |
|
363 | def gen(): | |
364 | yield b'HGS1' |
|
364 | yield b'HGS1' | |
365 | yield compression |
|
365 | yield compression | |
366 |
|
366 | |||
367 | filecount, bytecount, it = generatev1(repo) |
|
367 | filecount, bytecount, it = generatev1(repo) | |
368 | repo.ui.status( |
|
368 | repo.ui.status( | |
369 | _(b'writing %d bytes for %d files\n') % (bytecount, filecount) |
|
369 | _(b'writing %d bytes for %d files\n') % (bytecount, filecount) | |
370 | ) |
|
370 | ) | |
371 |
|
371 | |||
372 | yield struct.pack(b'>QQ', filecount, bytecount) |
|
372 | yield struct.pack(b'>QQ', filecount, bytecount) | |
373 | yield struct.pack(b'>H', len(requires) + 1) |
|
373 | yield struct.pack(b'>H', len(requires) + 1) | |
374 | yield requires + b'\0' |
|
374 | yield requires + b'\0' | |
375 |
|
375 | |||
376 | # This is where we'll add compression in the future. |
|
376 | # This is where we'll add compression in the future. | |
377 | assert compression == b'UN' |
|
377 | assert compression == b'UN' | |
378 |
|
378 | |||
379 | progress = repo.ui.makeprogress( |
|
379 | progress = repo.ui.makeprogress( | |
380 | _(b'bundle'), total=bytecount, unit=_(b'bytes') |
|
380 | _(b'bundle'), total=bytecount, unit=_(b'bytes') | |
381 | ) |
|
381 | ) | |
382 | progress.update(0) |
|
382 | progress.update(0) | |
383 |
|
383 | |||
384 | for chunk in it: |
|
384 | for chunk in it: | |
385 | progress.increment(step=len(chunk)) |
|
385 | progress.increment(step=len(chunk)) | |
386 | yield chunk |
|
386 | yield chunk | |
387 |
|
387 | |||
388 | progress.complete() |
|
388 | progress.complete() | |
389 |
|
389 | |||
390 | return requirements, gen() |
|
390 | return requirements, gen() | |
391 |
|
391 | |||
392 |
|
392 | |||
393 | def consumev1(repo, fp, filecount, bytecount): |
|
393 | def consumev1(repo, fp, filecount, bytecount): | |
394 | """Apply the contents from version 1 of a streaming clone file handle. |
|
394 | """Apply the contents from version 1 of a streaming clone file handle. | |
395 |
|
395 | |||
396 | This takes the output from "stream_out" and applies it to the specified |
|
396 | This takes the output from "stream_out" and applies it to the specified | |
397 | repository. |
|
397 | repository. | |
398 |
|
398 | |||
399 | Like "stream_out," the status line added by the wire protocol is not |
|
399 | Like "stream_out," the status line added by the wire protocol is not | |
400 | handled by this function. |
|
400 | handled by this function. | |
401 | """ |
|
401 | """ | |
402 | with repo.lock(): |
|
402 | with repo.lock(): | |
403 | repo.ui.status( |
|
403 | repo.ui.status( | |
404 | _(b'%d files to transfer, %s of data\n') |
|
404 | _(b'%d files to transfer, %s of data\n') | |
405 | % (filecount, util.bytecount(bytecount)) |
|
405 | % (filecount, util.bytecount(bytecount)) | |
406 | ) |
|
406 | ) | |
407 | progress = repo.ui.makeprogress( |
|
407 | progress = repo.ui.makeprogress( | |
408 | _(b'clone'), total=bytecount, unit=_(b'bytes') |
|
408 | _(b'clone'), total=bytecount, unit=_(b'bytes') | |
409 | ) |
|
409 | ) | |
410 | progress.update(0) |
|
410 | progress.update(0) | |
411 | start = util.timer() |
|
411 | start = util.timer() | |
412 |
|
412 | |||
413 | # TODO: get rid of (potential) inconsistency |
|
413 | # TODO: get rid of (potential) inconsistency | |
414 | # |
|
414 | # | |
415 | # If transaction is started and any @filecache property is |
|
415 | # If transaction is started and any @filecache property is | |
416 | # changed at this point, it causes inconsistency between |
|
416 | # changed at this point, it causes inconsistency between | |
417 | # in-memory cached property and streamclone-ed file on the |
|
417 | # in-memory cached property and streamclone-ed file on the | |
418 | # disk. Nested transaction prevents transaction scope "clone" |
|
418 | # disk. Nested transaction prevents transaction scope "clone" | |
419 | # below from writing in-memory changes out at the end of it, |
|
419 | # below from writing in-memory changes out at the end of it, | |
420 | # even though in-memory changes are discarded at the end of it |
|
420 | # even though in-memory changes are discarded at the end of it | |
421 | # regardless of transaction nesting. |
|
421 | # regardless of transaction nesting. | |
422 | # |
|
422 | # | |
423 | # But transaction nesting can't be simply prohibited, because |
|
423 | # But transaction nesting can't be simply prohibited, because | |
424 | # nesting occurs also in ordinary case (e.g. enabling |
|
424 | # nesting occurs also in ordinary case (e.g. enabling | |
425 | # clonebundles). |
|
425 | # clonebundles). | |
426 |
|
426 | |||
427 | with repo.transaction(b'clone'): |
|
427 | with repo.transaction(b'clone'): | |
428 | with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount): |
|
428 | with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount): | |
429 | for i in range(filecount): |
|
429 | for i in range(filecount): | |
430 | # XXX doesn't support '\n' or '\r' in filenames |
|
430 | # XXX doesn't support '\n' or '\r' in filenames | |
431 | l = fp.readline() |
|
431 | l = fp.readline() | |
432 | try: |
|
432 | try: | |
433 | name, size = l.split(b'\0', 1) |
|
433 | name, size = l.split(b'\0', 1) | |
434 | size = int(size) |
|
434 | size = int(size) | |
435 | except (ValueError, TypeError): |
|
435 | except (ValueError, TypeError): | |
436 | raise error.ResponseError( |
|
436 | raise error.ResponseError( | |
437 | _(b'unexpected response from remote server:'), l |
|
437 | _(b'unexpected response from remote server:'), l | |
438 | ) |
|
438 | ) | |
439 | if repo.ui.debugflag: |
|
439 | if repo.ui.debugflag: | |
440 | repo.ui.debug( |
|
440 | repo.ui.debug( | |
441 | b'adding %s (%s)\n' % (name, util.bytecount(size)) |
|
441 | b'adding %s (%s)\n' % (name, util.bytecount(size)) | |
442 | ) |
|
442 | ) | |
443 | # for backwards compat, name was partially encoded |
|
443 | # for backwards compat, name was partially encoded | |
444 | path = store.decodedir(name) |
|
444 | path = store.decodedir(name) | |
445 | with repo.svfs(path, b'w', backgroundclose=True) as ofp: |
|
445 | with repo.svfs(path, b'w', backgroundclose=True) as ofp: | |
446 | for chunk in util.filechunkiter(fp, limit=size): |
|
446 | for chunk in util.filechunkiter(fp, limit=size): | |
447 | progress.increment(step=len(chunk)) |
|
447 | progress.increment(step=len(chunk)) | |
448 | ofp.write(chunk) |
|
448 | ofp.write(chunk) | |
449 |
|
449 | |||
450 | # force @filecache properties to be reloaded from |
|
450 | # force @filecache properties to be reloaded from | |
451 | # streamclone-ed file at next access |
|
451 | # streamclone-ed file at next access | |
452 | repo.invalidate(clearfilecache=True) |
|
452 | repo.invalidate(clearfilecache=True) | |
453 |
|
453 | |||
454 | elapsed = util.timer() - start |
|
454 | elapsed = util.timer() - start | |
455 | if elapsed <= 0: |
|
455 | if elapsed <= 0: | |
456 | elapsed = 0.001 |
|
456 | elapsed = 0.001 | |
457 | progress.complete() |
|
457 | progress.complete() | |
458 | repo.ui.status( |
|
458 | repo.ui.status( | |
459 | _(b'transferred %s in %.1f seconds (%s/sec)\n') |
|
459 | _(b'transferred %s in %.1f seconds (%s/sec)\n') | |
460 | % ( |
|
460 | % ( | |
461 | util.bytecount(bytecount), |
|
461 | util.bytecount(bytecount), | |
462 | elapsed, |
|
462 | elapsed, | |
463 | util.bytecount(bytecount / elapsed), |
|
463 | util.bytecount(bytecount / elapsed), | |
464 | ) |
|
464 | ) | |
465 | ) |
|
465 | ) | |
466 |
|
466 | |||
467 |
|
467 | |||
468 | def readbundle1header(fp): |
|
468 | def readbundle1header(fp): | |
469 | compression = fp.read(2) |
|
469 | compression = fp.read(2) | |
470 | if compression != b'UN': |
|
470 | if compression != b'UN': | |
471 | raise error.Abort( |
|
471 | raise error.Abort( | |
472 | _( |
|
472 | _( | |
473 | b'only uncompressed stream clone bundles are ' |
|
473 | b'only uncompressed stream clone bundles are ' | |
474 | b'supported; got %s' |
|
474 | b'supported; got %s' | |
475 | ) |
|
475 | ) | |
476 | % compression |
|
476 | % compression | |
477 | ) |
|
477 | ) | |
478 |
|
478 | |||
479 | filecount, bytecount = struct.unpack(b'>QQ', fp.read(16)) |
|
479 | filecount, bytecount = struct.unpack(b'>QQ', fp.read(16)) | |
480 | requireslen = struct.unpack(b'>H', fp.read(2))[0] |
|
480 | requireslen = struct.unpack(b'>H', fp.read(2))[0] | |
481 | requires = fp.read(requireslen) |
|
481 | requires = fp.read(requireslen) | |
482 |
|
482 | |||
483 | if not requires.endswith(b'\0'): |
|
483 | if not requires.endswith(b'\0'): | |
484 | raise error.Abort( |
|
484 | raise error.Abort( | |
485 | _( |
|
485 | _( | |
486 | b'malformed stream clone bundle: ' |
|
486 | b'malformed stream clone bundle: ' | |
487 | b'requirements not properly encoded' |
|
487 | b'requirements not properly encoded' | |
488 | ) |
|
488 | ) | |
489 | ) |
|
489 | ) | |
490 |
|
490 | |||
491 | requirements = set(requires.rstrip(b'\0').split(b',')) |
|
491 | requirements = set(requires.rstrip(b'\0').split(b',')) | |
492 |
|
492 | |||
493 | return filecount, bytecount, requirements |
|
493 | return filecount, bytecount, requirements | |
494 |
|
494 | |||
495 |
|
495 | |||
496 | def applybundlev1(repo, fp): |
|
496 | def applybundlev1(repo, fp): | |
497 | """Apply the content from a stream clone bundle version 1. |
|
497 | """Apply the content from a stream clone bundle version 1. | |
498 |
|
498 | |||
499 | We assume the 4 byte header has been read and validated and the file handle |
|
499 | We assume the 4 byte header has been read and validated and the file handle | |
500 | is at the 2 byte compression identifier. |
|
500 | is at the 2 byte compression identifier. | |
501 | """ |
|
501 | """ | |
502 | if len(repo): |
|
502 | if len(repo): | |
503 | raise error.Abort( |
|
503 | raise error.Abort( | |
504 | _(b'cannot apply stream clone bundle on non-empty repo') |
|
504 | _(b'cannot apply stream clone bundle on non-empty repo') | |
505 | ) |
|
505 | ) | |
506 |
|
506 | |||
507 | filecount, bytecount, requirements = readbundle1header(fp) |
|
507 | filecount, bytecount, requirements = readbundle1header(fp) | |
508 | missingreqs = requirements - repo.supported |
|
508 | missingreqs = requirements - repo.supported | |
509 | if missingreqs: |
|
509 | if missingreqs: | |
510 | raise error.Abort( |
|
510 | raise error.Abort( | |
511 | _(b'unable to apply stream clone: unsupported format: %s') |
|
511 | _(b'unable to apply stream clone: unsupported format: %s') | |
512 | % b', '.join(sorted(missingreqs)) |
|
512 | % b', '.join(sorted(missingreqs)) | |
513 | ) |
|
513 | ) | |
514 |
|
514 | |||
515 | consumev1(repo, fp, filecount, bytecount) |
|
515 | consumev1(repo, fp, filecount, bytecount) | |
516 | nodemap.post_stream_cleanup(repo) |
|
516 | nodemap.post_stream_cleanup(repo) | |
517 |
|
517 | |||
518 |
|
518 | |||
519 | class streamcloneapplier: |
|
519 | class streamcloneapplier: | |
520 | """Class to manage applying streaming clone bundles. |
|
520 | """Class to manage applying streaming clone bundles. | |
521 |
|
521 | |||
522 | We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle |
|
522 | We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle | |
523 | readers to perform bundle type-specific functionality. |
|
523 | readers to perform bundle type-specific functionality. | |
524 | """ |
|
524 | """ | |
525 |
|
525 | |||
526 | def __init__(self, fh): |
|
526 | def __init__(self, fh): | |
527 | self._fh = fh |
|
527 | self._fh = fh | |
528 |
|
528 | |||
529 | def apply(self, repo): |
|
529 | def apply(self, repo): | |
530 | return applybundlev1(repo, self._fh) |
|
530 | return applybundlev1(repo, self._fh) | |
531 |
|
531 | |||
532 |
|
532 | |||
533 | # type of file to stream |
|
533 | # type of file to stream | |
534 | _fileappend = 0 # append only file |
|
534 | _fileappend = 0 # append only file | |
535 | _filefull = 1 # full snapshot file |
|
535 | _filefull = 1 # full snapshot file | |
536 |
|
536 | |||
537 | # Source of the file |
|
537 | # Source of the file | |
538 | _srcstore = b's' # store (svfs) |
|
538 | _srcstore = b's' # store (svfs) | |
539 | _srccache = b'c' # cache (cache) |
|
539 | _srccache = b'c' # cache (cache) | |
540 |
|
540 | |||
541 | # This is it's own function so extensions can override it. |
|
541 | # This is it's own function so extensions can override it. | |
542 | def _walkstreamfullstorefiles(repo): |
|
542 | def _walkstreamfullstorefiles(repo): | |
543 | """list snapshot file from the store""" |
|
543 | """list snapshot file from the store""" | |
544 | fnames = [] |
|
544 | fnames = [] | |
545 | if not repo.publishing(): |
|
545 | if not repo.publishing(): | |
546 | fnames.append(b'phaseroots') |
|
546 | fnames.append(b'phaseroots') | |
547 | return fnames |
|
547 | return fnames | |
548 |
|
548 | |||
549 |
|
549 | |||
550 | def _filterfull(entry, copy, vfsmap): |
|
550 | def _filterfull(entry, copy, vfsmap): | |
551 | """actually copy the snapshot files""" |
|
551 | """actually copy the snapshot files""" | |
552 | src, name, ftype, data = entry |
|
552 | src, name, ftype, data = entry | |
553 | if ftype != _filefull: |
|
553 | if ftype != _filefull: | |
554 | return entry |
|
554 | return entry | |
555 | return (src, name, ftype, copy(vfsmap[src].join(name))) |
|
555 | return (src, name, ftype, copy(vfsmap[src].join(name))) | |
556 |
|
556 | |||
557 |
|
557 | |||
558 | class TempCopyManager: |
|
558 | class TempCopyManager: | |
559 | """Manage temporary backup of volatile file during stream clone |
|
559 | """Manage temporary backup of volatile file during stream clone | |
560 |
|
560 | |||
561 | This should be used as a Python context, the copies will be discarded when |
|
561 | This should be used as a Python context, the copies will be discarded when | |
562 | exiting the context. |
|
562 | exiting the context. | |
563 |
|
563 | |||
564 | A copy can be done by calling the object on the real path (encoded full |
|
564 | A copy can be done by calling the object on the real path (encoded full | |
565 | path) |
|
565 | path) | |
566 |
|
566 | |||
567 | The backup path can be retrieved using the __getitem__ protocol, obj[path]. |
|
567 | The backup path can be retrieved using the __getitem__ protocol, obj[path]. | |
568 | On file without backup, it will return the unmodified path. (equivalent to |
|
568 | On file without backup, it will return the unmodified path. (equivalent to | |
569 | `dict.get(x, x)`) |
|
569 | `dict.get(x, x)`) | |
570 | """ |
|
570 | """ | |
571 |
|
571 | |||
572 | def __init__(self): |
|
572 | def __init__(self): | |
573 | self._copies = None |
|
573 | self._copies = None | |
574 | self._dst_dir = None |
|
574 | self._dst_dir = None | |
575 |
|
575 | |||
576 | def __enter__(self): |
|
576 | def __enter__(self): | |
577 | if self._copies is not None: |
|
577 | if self._copies is not None: | |
578 | msg = "Copies context already open" |
|
578 | msg = "Copies context already open" | |
579 | raise error.ProgrammingError(msg) |
|
579 | raise error.ProgrammingError(msg) | |
580 | self._copies = {} |
|
580 | self._copies = {} | |
581 | self._dst_dir = pycompat.mkdtemp(prefix=b'hg-clone-') |
|
581 | self._dst_dir = pycompat.mkdtemp(prefix=b'hg-clone-') | |
582 | return self |
|
582 | return self | |
583 |
|
583 | |||
584 | def __call__(self, src): |
|
584 | def __call__(self, src): | |
585 | """create a backup of the file at src""" |
|
585 | """create a backup of the file at src""" | |
586 | prefix = os.path.basename(src) |
|
586 | prefix = os.path.basename(src) | |
587 | fd, dst = pycompat.mkstemp(prefix=prefix, dir=self._dst_dir) |
|
587 | fd, dst = pycompat.mkstemp(prefix=prefix, dir=self._dst_dir) | |
588 | os.close(fd) |
|
588 | os.close(fd) | |
589 | self._copies[src] = dst |
|
589 | self._copies[src] = dst | |
590 | util.copyfiles(src, dst, hardlink=True) |
|
590 | util.copyfiles(src, dst, hardlink=True) | |
591 | return dst |
|
591 | return dst | |
592 |
|
592 | |||
593 | def __getitem__(self, src): |
|
593 | def __getitem__(self, src): | |
594 | """return the path to a valid version of `src` |
|
594 | """return the path to a valid version of `src` | |
595 |
|
595 | |||
596 | If the file has no backup, the path of the file is returned |
|
596 | If the file has no backup, the path of the file is returned | |
597 | unmodified.""" |
|
597 | unmodified.""" | |
598 | return self._copies.get(src, src) |
|
598 | return self._copies.get(src, src) | |
599 |
|
599 | |||
600 | def __exit__(self, *args, **kwars): |
|
600 | def __exit__(self, *args, **kwars): | |
601 | """discard all backups""" |
|
601 | """discard all backups""" | |
602 | for tmp in self._copies.values(): |
|
602 | for tmp in self._copies.values(): | |
603 | util.tryunlink(tmp) |
|
603 | util.tryunlink(tmp) | |
604 | util.tryrmdir(self._dst_dir) |
|
604 | util.tryrmdir(self._dst_dir) | |
605 | self._copies = None |
|
605 | self._copies = None | |
606 | self._dst_dir = None |
|
606 | self._dst_dir = None | |
607 |
|
607 | |||
608 |
|
608 | |||
609 | def _makemap(repo): |
|
609 | def _makemap(repo): | |
610 | """make a (src -> vfs) map for the repo""" |
|
610 | """make a (src -> vfs) map for the repo""" | |
611 | vfsmap = { |
|
611 | vfsmap = { | |
612 | _srcstore: repo.svfs, |
|
612 | _srcstore: repo.svfs, | |
613 | _srccache: repo.cachevfs, |
|
613 | _srccache: repo.cachevfs, | |
614 | } |
|
614 | } | |
615 | # we keep repo.vfs out of the on purpose, ther are too many danger there |
|
615 | # we keep repo.vfs out of the on purpose, ther are too many danger there | |
616 | # (eg: .hg/hgrc) |
|
616 | # (eg: .hg/hgrc) | |
617 | assert repo.vfs not in vfsmap.values() |
|
617 | assert repo.vfs not in vfsmap.values() | |
618 |
|
618 | |||
619 | return vfsmap |
|
619 | return vfsmap | |
620 |
|
620 | |||
621 |
|
621 | |||
622 | def _emit2(repo, entries): |
|
622 | def _emit2(repo, entries): | |
623 | """actually emit the stream bundle""" |
|
623 | """actually emit the stream bundle""" | |
624 | vfsmap = _makemap(repo) |
|
624 | vfsmap = _makemap(repo) | |
625 | # we keep repo.vfs out of the on purpose, ther are too many danger there |
|
625 | # we keep repo.vfs out of the on purpose, ther are too many danger there | |
626 | # (eg: .hg/hgrc), |
|
626 | # (eg: .hg/hgrc), | |
627 | # |
|
627 | # | |
628 | # this assert is duplicated (from _makemap) as author might think this is |
|
628 | # this assert is duplicated (from _makemap) as author might think this is | |
629 | # fine, while this is really not fine. |
|
629 | # fine, while this is really not fine. | |
630 | if repo.vfs in vfsmap.values(): |
|
630 | if repo.vfs in vfsmap.values(): | |
631 | raise error.ProgrammingError( |
|
631 | raise error.ProgrammingError( | |
632 | b'repo.vfs must not be added to vfsmap for security reasons' |
|
632 | b'repo.vfs must not be added to vfsmap for security reasons' | |
633 | ) |
|
633 | ) | |
634 |
|
634 | |||
635 | # translate the vfs one |
|
635 | # translate the vfs one | |
636 | entries = [(vfs_key, vfsmap[vfs_key], e) for (vfs_key, e) in entries] |
|
636 | entries = [(vfs_key, vfsmap[vfs_key], e) for (vfs_key, e) in entries] | |
637 |
|
637 | |||
|
638 | max_linkrev = len(repo) | |||
638 | file_count = totalfilesize = 0 |
|
639 | file_count = totalfilesize = 0 | |
639 | # record the expected size of every file |
|
640 | # record the expected size of every file | |
640 | for k, vfs, e in entries: |
|
641 | for k, vfs, e in entries: | |
641 | for f in e.files(): |
|
642 | for f in e.files(): | |
642 | file_count += 1 |
|
643 | file_count += 1 | |
643 | totalfilesize += f.file_size(vfs) |
|
644 | totalfilesize += f.file_size(vfs) | |
644 |
|
645 | |||
645 | progress = repo.ui.makeprogress( |
|
646 | progress = repo.ui.makeprogress( | |
646 | _(b'bundle'), total=totalfilesize, unit=_(b'bytes') |
|
647 | _(b'bundle'), total=totalfilesize, unit=_(b'bytes') | |
647 | ) |
|
648 | ) | |
648 | progress.update(0) |
|
649 | progress.update(0) | |
649 | with TempCopyManager() as copy, progress: |
|
650 | with TempCopyManager() as copy, progress: | |
650 | # create a copy of volatile files |
|
651 | # create a copy of volatile files | |
651 | for k, vfs, e in entries: |
|
652 | for k, vfs, e in entries: | |
652 | for f in e.files(): |
|
653 | for f in e.files(): | |
653 | if f.is_volatile: |
|
654 | if f.is_volatile: | |
654 | copy(vfs.join(f.unencoded_path)) |
|
655 | copy(vfs.join(f.unencoded_path)) | |
655 | # the first yield release the lock on the repository |
|
656 | # the first yield release the lock on the repository | |
656 | yield file_count, totalfilesize |
|
657 | yield file_count, totalfilesize | |
657 | totalbytecount = 0 |
|
658 | totalbytecount = 0 | |
658 |
|
659 | |||
659 | for src, vfs, e in entries: |
|
660 | for src, vfs, e in entries: | |
660 |
|
|
661 | entry_streams = e.get_streams( | |
|
662 | repo=repo, vfs=vfs, copies=copy, max_changeset=max_linkrev | |||
|
663 | ) | |||
|
664 | for name, stream, size in entry_streams: | |||
661 | yield src |
|
665 | yield src | |
662 | yield util.uvarintencode(len(name)) |
|
666 | yield util.uvarintencode(len(name)) | |
663 | yield util.uvarintencode(size) |
|
667 | yield util.uvarintencode(size) | |
664 | yield name |
|
668 | yield name | |
665 | bytecount = 0 |
|
669 | bytecount = 0 | |
666 | for chunk in stream: |
|
670 | for chunk in stream: | |
667 | bytecount += len(chunk) |
|
671 | bytecount += len(chunk) | |
668 | totalbytecount += len(chunk) |
|
672 | totalbytecount += len(chunk) | |
669 | progress.update(totalbytecount) |
|
673 | progress.update(totalbytecount) | |
670 | yield chunk |
|
674 | yield chunk | |
671 | if bytecount != size: |
|
675 | if bytecount != size: | |
672 | # Would most likely be caused by a race due to `hg |
|
676 | # Would most likely be caused by a race due to `hg | |
673 | # strip` or a revlog split |
|
677 | # strip` or a revlog split | |
674 | msg = _( |
|
678 | msg = _( | |
675 | b'clone could only read %d bytes from %s, but ' |
|
679 | b'clone could only read %d bytes from %s, but ' | |
676 | b'expected %d bytes' |
|
680 | b'expected %d bytes' | |
677 | ) |
|
681 | ) | |
678 | raise error.Abort(msg % (bytecount, name, size)) |
|
682 | raise error.Abort(msg % (bytecount, name, size)) | |
679 |
|
683 | |||
680 |
|
684 | |||
681 | def _test_sync_point_walk_1(repo): |
|
685 | def _test_sync_point_walk_1(repo): | |
682 | """a function for synchronisation during tests""" |
|
686 | """a function for synchronisation during tests""" | |
683 |
|
687 | |||
684 |
|
688 | |||
685 | def _test_sync_point_walk_2(repo): |
|
689 | def _test_sync_point_walk_2(repo): | |
686 | """a function for synchronisation during tests""" |
|
690 | """a function for synchronisation during tests""" | |
687 |
|
691 | |||
688 |
|
692 | |||
689 | def _entries_walk(repo, includes, excludes, includeobsmarkers): |
|
693 | def _entries_walk(repo, includes, excludes, includeobsmarkers): | |
690 | """emit a seris of files information useful to clone a repo |
|
694 | """emit a seris of files information useful to clone a repo | |
691 |
|
695 | |||
692 | return (vfs-key, entry) iterator |
|
696 | return (vfs-key, entry) iterator | |
693 |
|
697 | |||
694 | Where `entry` is StoreEntry. (used even for cache entries) |
|
698 | Where `entry` is StoreEntry. (used even for cache entries) | |
695 | """ |
|
699 | """ | |
696 | assert repo._currentlock(repo._lockref) is not None |
|
700 | assert repo._currentlock(repo._lockref) is not None | |
697 |
|
701 | |||
698 | matcher = None |
|
702 | matcher = None | |
699 | if includes or excludes: |
|
703 | if includes or excludes: | |
700 | matcher = narrowspec.match(repo.root, includes, excludes) |
|
704 | matcher = narrowspec.match(repo.root, includes, excludes) | |
701 |
|
705 | |||
702 | phase = not repo.publishing() |
|
706 | phase = not repo.publishing() | |
703 | entries = _walkstreamfiles( |
|
707 | entries = _walkstreamfiles( | |
704 | repo, |
|
708 | repo, | |
705 | matcher, |
|
709 | matcher, | |
706 | phase=phase, |
|
710 | phase=phase, | |
707 | obsolescence=includeobsmarkers, |
|
711 | obsolescence=includeobsmarkers, | |
708 | ) |
|
712 | ) | |
709 | for entry in entries: |
|
713 | for entry in entries: | |
710 | yield (_srcstore, entry) |
|
714 | yield (_srcstore, entry) | |
711 |
|
715 | |||
712 | for name in cacheutil.cachetocopy(repo): |
|
716 | for name in cacheutil.cachetocopy(repo): | |
713 | if repo.cachevfs.exists(name): |
|
717 | if repo.cachevfs.exists(name): | |
714 | # not really a StoreEntry, but close enough |
|
718 | # not really a StoreEntry, but close enough | |
715 | entry = store.SimpleStoreEntry( |
|
719 | entry = store.SimpleStoreEntry( | |
716 | entry_path=name, |
|
720 | entry_path=name, | |
717 | is_volatile=True, |
|
721 | is_volatile=True, | |
718 | ) |
|
722 | ) | |
719 | yield (_srccache, entry) |
|
723 | yield (_srccache, entry) | |
720 |
|
724 | |||
721 |
|
725 | |||
722 | def generatev2(repo, includes, excludes, includeobsmarkers): |
|
726 | def generatev2(repo, includes, excludes, includeobsmarkers): | |
723 | """Emit content for version 2 of a streaming clone. |
|
727 | """Emit content for version 2 of a streaming clone. | |
724 |
|
728 | |||
725 | the data stream consists the following entries: |
|
729 | the data stream consists the following entries: | |
726 | 1) A char representing the file destination (eg: store or cache) |
|
730 | 1) A char representing the file destination (eg: store or cache) | |
727 | 2) A varint containing the length of the filename |
|
731 | 2) A varint containing the length of the filename | |
728 | 3) A varint containing the length of file data |
|
732 | 3) A varint containing the length of file data | |
729 | 4) N bytes containing the filename (the internal, store-agnostic form) |
|
733 | 4) N bytes containing the filename (the internal, store-agnostic form) | |
730 | 5) N bytes containing the file data |
|
734 | 5) N bytes containing the file data | |
731 |
|
735 | |||
732 | Returns a 3-tuple of (file count, file size, data iterator). |
|
736 | Returns a 3-tuple of (file count, file size, data iterator). | |
733 | """ |
|
737 | """ | |
734 |
|
738 | |||
735 | with repo.lock(): |
|
739 | with repo.lock(): | |
736 |
|
740 | |||
737 | repo.ui.debug(b'scanning\n') |
|
741 | repo.ui.debug(b'scanning\n') | |
738 |
|
742 | |||
739 | entries = _entries_walk( |
|
743 | entries = _entries_walk( | |
740 | repo, |
|
744 | repo, | |
741 | includes=includes, |
|
745 | includes=includes, | |
742 | excludes=excludes, |
|
746 | excludes=excludes, | |
743 | includeobsmarkers=includeobsmarkers, |
|
747 | includeobsmarkers=includeobsmarkers, | |
744 | ) |
|
748 | ) | |
745 |
|
749 | |||
746 | chunks = _emit2(repo, entries) |
|
750 | chunks = _emit2(repo, entries) | |
747 | first = next(chunks) |
|
751 | first = next(chunks) | |
748 | file_count, total_file_size = first |
|
752 | file_count, total_file_size = first | |
749 | _test_sync_point_walk_1(repo) |
|
753 | _test_sync_point_walk_1(repo) | |
750 | _test_sync_point_walk_2(repo) |
|
754 | _test_sync_point_walk_2(repo) | |
751 |
|
755 | |||
752 | return file_count, total_file_size, chunks |
|
756 | return file_count, total_file_size, chunks | |
753 |
|
757 | |||
754 |
|
758 | |||
755 | def generatev3(repo, includes, excludes, includeobsmarkers): |
|
759 | def generatev3(repo, includes, excludes, includeobsmarkers): | |
756 | return generatev2(repo, includes, excludes, includeobsmarkers) |
|
760 | return generatev2(repo, includes, excludes, includeobsmarkers) | |
757 |
|
761 | |||
758 |
|
762 | |||
759 | @contextlib.contextmanager |
|
763 | @contextlib.contextmanager | |
760 | def nested(*ctxs): |
|
764 | def nested(*ctxs): | |
761 | this = ctxs[0] |
|
765 | this = ctxs[0] | |
762 | rest = ctxs[1:] |
|
766 | rest = ctxs[1:] | |
763 | with this: |
|
767 | with this: | |
764 | if rest: |
|
768 | if rest: | |
765 | with nested(*rest): |
|
769 | with nested(*rest): | |
766 | yield |
|
770 | yield | |
767 | else: |
|
771 | else: | |
768 | yield |
|
772 | yield | |
769 |
|
773 | |||
770 |
|
774 | |||
771 | def consumev2(repo, fp, filecount, filesize): |
|
775 | def consumev2(repo, fp, filecount, filesize): | |
772 | """Apply the contents from a version 2 streaming clone. |
|
776 | """Apply the contents from a version 2 streaming clone. | |
773 |
|
777 | |||
774 | Data is read from an object that only needs to provide a ``read(size)`` |
|
778 | Data is read from an object that only needs to provide a ``read(size)`` | |
775 | method. |
|
779 | method. | |
776 | """ |
|
780 | """ | |
777 | with repo.lock(): |
|
781 | with repo.lock(): | |
778 | repo.ui.status( |
|
782 | repo.ui.status( | |
779 | _(b'%d files to transfer, %s of data\n') |
|
783 | _(b'%d files to transfer, %s of data\n') | |
780 | % (filecount, util.bytecount(filesize)) |
|
784 | % (filecount, util.bytecount(filesize)) | |
781 | ) |
|
785 | ) | |
782 |
|
786 | |||
783 | start = util.timer() |
|
787 | start = util.timer() | |
784 | progress = repo.ui.makeprogress( |
|
788 | progress = repo.ui.makeprogress( | |
785 | _(b'clone'), total=filesize, unit=_(b'bytes') |
|
789 | _(b'clone'), total=filesize, unit=_(b'bytes') | |
786 | ) |
|
790 | ) | |
787 | progress.update(0) |
|
791 | progress.update(0) | |
788 |
|
792 | |||
789 | vfsmap = _makemap(repo) |
|
793 | vfsmap = _makemap(repo) | |
790 | # we keep repo.vfs out of the on purpose, ther are too many danger |
|
794 | # we keep repo.vfs out of the on purpose, ther are too many danger | |
791 | # there (eg: .hg/hgrc), |
|
795 | # there (eg: .hg/hgrc), | |
792 | # |
|
796 | # | |
793 | # this assert is duplicated (from _makemap) as author might think this |
|
797 | # this assert is duplicated (from _makemap) as author might think this | |
794 | # is fine, while this is really not fine. |
|
798 | # is fine, while this is really not fine. | |
795 | if repo.vfs in vfsmap.values(): |
|
799 | if repo.vfs in vfsmap.values(): | |
796 | raise error.ProgrammingError( |
|
800 | raise error.ProgrammingError( | |
797 | b'repo.vfs must not be added to vfsmap for security reasons' |
|
801 | b'repo.vfs must not be added to vfsmap for security reasons' | |
798 | ) |
|
802 | ) | |
799 |
|
803 | |||
800 | with repo.transaction(b'clone'): |
|
804 | with repo.transaction(b'clone'): | |
801 | ctxs = (vfs.backgroundclosing(repo.ui) for vfs in vfsmap.values()) |
|
805 | ctxs = (vfs.backgroundclosing(repo.ui) for vfs in vfsmap.values()) | |
802 | with nested(*ctxs): |
|
806 | with nested(*ctxs): | |
803 | for i in range(filecount): |
|
807 | for i in range(filecount): | |
804 | src = util.readexactly(fp, 1) |
|
808 | src = util.readexactly(fp, 1) | |
805 | vfs = vfsmap[src] |
|
809 | vfs = vfsmap[src] | |
806 | namelen = util.uvarintdecodestream(fp) |
|
810 | namelen = util.uvarintdecodestream(fp) | |
807 | datalen = util.uvarintdecodestream(fp) |
|
811 | datalen = util.uvarintdecodestream(fp) | |
808 |
|
812 | |||
809 | name = util.readexactly(fp, namelen) |
|
813 | name = util.readexactly(fp, namelen) | |
810 |
|
814 | |||
811 | if repo.ui.debugflag: |
|
815 | if repo.ui.debugflag: | |
812 | repo.ui.debug( |
|
816 | repo.ui.debug( | |
813 | b'adding [%s] %s (%s)\n' |
|
817 | b'adding [%s] %s (%s)\n' | |
814 | % (src, name, util.bytecount(datalen)) |
|
818 | % (src, name, util.bytecount(datalen)) | |
815 | ) |
|
819 | ) | |
816 |
|
820 | |||
817 | with vfs(name, b'w') as ofp: |
|
821 | with vfs(name, b'w') as ofp: | |
818 | for chunk in util.filechunkiter(fp, limit=datalen): |
|
822 | for chunk in util.filechunkiter(fp, limit=datalen): | |
819 | progress.increment(step=len(chunk)) |
|
823 | progress.increment(step=len(chunk)) | |
820 | ofp.write(chunk) |
|
824 | ofp.write(chunk) | |
821 |
|
825 | |||
822 | # force @filecache properties to be reloaded from |
|
826 | # force @filecache properties to be reloaded from | |
823 | # streamclone-ed file at next access |
|
827 | # streamclone-ed file at next access | |
824 | repo.invalidate(clearfilecache=True) |
|
828 | repo.invalidate(clearfilecache=True) | |
825 |
|
829 | |||
826 | elapsed = util.timer() - start |
|
830 | elapsed = util.timer() - start | |
827 | if elapsed <= 0: |
|
831 | if elapsed <= 0: | |
828 | elapsed = 0.001 |
|
832 | elapsed = 0.001 | |
829 | repo.ui.status( |
|
833 | repo.ui.status( | |
830 | _(b'transferred %s in %.1f seconds (%s/sec)\n') |
|
834 | _(b'transferred %s in %.1f seconds (%s/sec)\n') | |
831 | % ( |
|
835 | % ( | |
832 | util.bytecount(progress.pos), |
|
836 | util.bytecount(progress.pos), | |
833 | elapsed, |
|
837 | elapsed, | |
834 | util.bytecount(progress.pos / elapsed), |
|
838 | util.bytecount(progress.pos / elapsed), | |
835 | ) |
|
839 | ) | |
836 | ) |
|
840 | ) | |
837 | progress.complete() |
|
841 | progress.complete() | |
838 |
|
842 | |||
839 |
|
843 | |||
840 | def applybundlev2(repo, fp, filecount, filesize, requirements): |
|
844 | def applybundlev2(repo, fp, filecount, filesize, requirements): | |
841 | from . import localrepo |
|
845 | from . import localrepo | |
842 |
|
846 | |||
843 | missingreqs = [r for r in requirements if r not in repo.supported] |
|
847 | missingreqs = [r for r in requirements if r not in repo.supported] | |
844 | if missingreqs: |
|
848 | if missingreqs: | |
845 | raise error.Abort( |
|
849 | raise error.Abort( | |
846 | _(b'unable to apply stream clone: unsupported format: %s') |
|
850 | _(b'unable to apply stream clone: unsupported format: %s') | |
847 | % b', '.join(sorted(missingreqs)) |
|
851 | % b', '.join(sorted(missingreqs)) | |
848 | ) |
|
852 | ) | |
849 |
|
853 | |||
850 | consumev2(repo, fp, filecount, filesize) |
|
854 | consumev2(repo, fp, filecount, filesize) | |
851 |
|
855 | |||
852 | repo.requirements = new_stream_clone_requirements( |
|
856 | repo.requirements = new_stream_clone_requirements( | |
853 | repo.requirements, |
|
857 | repo.requirements, | |
854 | requirements, |
|
858 | requirements, | |
855 | ) |
|
859 | ) | |
856 | repo.svfs.options = localrepo.resolvestorevfsoptions( |
|
860 | repo.svfs.options = localrepo.resolvestorevfsoptions( | |
857 | repo.ui, repo.requirements, repo.features |
|
861 | repo.ui, repo.requirements, repo.features | |
858 | ) |
|
862 | ) | |
859 | scmutil.writereporequirements(repo) |
|
863 | scmutil.writereporequirements(repo) | |
860 | nodemap.post_stream_cleanup(repo) |
|
864 | nodemap.post_stream_cleanup(repo) | |
861 |
|
865 | |||
862 |
|
866 | |||
863 | def _copy_files(src_vfs_map, dst_vfs_map, entries, progress): |
|
867 | def _copy_files(src_vfs_map, dst_vfs_map, entries, progress): | |
864 | hardlink = [True] |
|
868 | hardlink = [True] | |
865 |
|
869 | |||
866 | def copy_used(): |
|
870 | def copy_used(): | |
867 | hardlink[0] = False |
|
871 | hardlink[0] = False | |
868 | progress.topic = _(b'copying') |
|
872 | progress.topic = _(b'copying') | |
869 |
|
873 | |||
870 | for k, path in entries: |
|
874 | for k, path in entries: | |
871 | src_vfs = src_vfs_map[k] |
|
875 | src_vfs = src_vfs_map[k] | |
872 | dst_vfs = dst_vfs_map[k] |
|
876 | dst_vfs = dst_vfs_map[k] | |
873 | src_path = src_vfs.join(path) |
|
877 | src_path = src_vfs.join(path) | |
874 | dst_path = dst_vfs.join(path) |
|
878 | dst_path = dst_vfs.join(path) | |
875 | # We cannot use dirname and makedirs of dst_vfs here because the store |
|
879 | # We cannot use dirname and makedirs of dst_vfs here because the store | |
876 | # encoding confuses them. See issue 6581 for details. |
|
880 | # encoding confuses them. See issue 6581 for details. | |
877 | dirname = os.path.dirname(dst_path) |
|
881 | dirname = os.path.dirname(dst_path) | |
878 | if not os.path.exists(dirname): |
|
882 | if not os.path.exists(dirname): | |
879 | util.makedirs(dirname) |
|
883 | util.makedirs(dirname) | |
880 | dst_vfs.register_file(path) |
|
884 | dst_vfs.register_file(path) | |
881 | # XXX we could use the #nb_bytes argument. |
|
885 | # XXX we could use the #nb_bytes argument. | |
882 | util.copyfile( |
|
886 | util.copyfile( | |
883 | src_path, |
|
887 | src_path, | |
884 | dst_path, |
|
888 | dst_path, | |
885 | hardlink=hardlink[0], |
|
889 | hardlink=hardlink[0], | |
886 | no_hardlink_cb=copy_used, |
|
890 | no_hardlink_cb=copy_used, | |
887 | check_fs_hardlink=False, |
|
891 | check_fs_hardlink=False, | |
888 | ) |
|
892 | ) | |
889 | progress.increment() |
|
893 | progress.increment() | |
890 | return hardlink[0] |
|
894 | return hardlink[0] | |
891 |
|
895 | |||
892 |
|
896 | |||
893 | def local_copy(src_repo, dest_repo): |
|
897 | def local_copy(src_repo, dest_repo): | |
894 | """copy all content from one local repository to another |
|
898 | """copy all content from one local repository to another | |
895 |
|
899 | |||
896 | This is useful for local clone""" |
|
900 | This is useful for local clone""" | |
897 | src_store_requirements = { |
|
901 | src_store_requirements = { | |
898 | r |
|
902 | r | |
899 | for r in src_repo.requirements |
|
903 | for r in src_repo.requirements | |
900 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS |
|
904 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS | |
901 | } |
|
905 | } | |
902 | dest_store_requirements = { |
|
906 | dest_store_requirements = { | |
903 | r |
|
907 | r | |
904 | for r in dest_repo.requirements |
|
908 | for r in dest_repo.requirements | |
905 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS |
|
909 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS | |
906 | } |
|
910 | } | |
907 | assert src_store_requirements == dest_store_requirements |
|
911 | assert src_store_requirements == dest_store_requirements | |
908 |
|
912 | |||
909 | with dest_repo.lock(): |
|
913 | with dest_repo.lock(): | |
910 | with src_repo.lock(): |
|
914 | with src_repo.lock(): | |
911 |
|
915 | |||
912 | # bookmark is not integrated to the streaming as it might use the |
|
916 | # bookmark is not integrated to the streaming as it might use the | |
913 | # `repo.vfs` and they are too many sentitive data accessible |
|
917 | # `repo.vfs` and they are too many sentitive data accessible | |
914 | # through `repo.vfs` to expose it to streaming clone. |
|
918 | # through `repo.vfs` to expose it to streaming clone. | |
915 | src_book_vfs = bookmarks.bookmarksvfs(src_repo) |
|
919 | src_book_vfs = bookmarks.bookmarksvfs(src_repo) | |
916 | srcbookmarks = src_book_vfs.join(b'bookmarks') |
|
920 | srcbookmarks = src_book_vfs.join(b'bookmarks') | |
917 | bm_count = 0 |
|
921 | bm_count = 0 | |
918 | if os.path.exists(srcbookmarks): |
|
922 | if os.path.exists(srcbookmarks): | |
919 | bm_count = 1 |
|
923 | bm_count = 1 | |
920 |
|
924 | |||
921 | entries = _entries_walk( |
|
925 | entries = _entries_walk( | |
922 | src_repo, |
|
926 | src_repo, | |
923 | includes=None, |
|
927 | includes=None, | |
924 | excludes=None, |
|
928 | excludes=None, | |
925 | includeobsmarkers=True, |
|
929 | includeobsmarkers=True, | |
926 | ) |
|
930 | ) | |
927 | entries = list(entries) |
|
931 | entries = list(entries) | |
928 | src_vfs_map = _makemap(src_repo) |
|
932 | src_vfs_map = _makemap(src_repo) | |
929 | dest_vfs_map = _makemap(dest_repo) |
|
933 | dest_vfs_map = _makemap(dest_repo) | |
930 | total_files = sum(len(e[1].files()) for e in entries) + bm_count |
|
934 | total_files = sum(len(e[1].files()) for e in entries) + bm_count | |
931 | progress = src_repo.ui.makeprogress( |
|
935 | progress = src_repo.ui.makeprogress( | |
932 | topic=_(b'linking'), |
|
936 | topic=_(b'linking'), | |
933 | total=total_files, |
|
937 | total=total_files, | |
934 | unit=_(b'files'), |
|
938 | unit=_(b'files'), | |
935 | ) |
|
939 | ) | |
936 | # copy files |
|
940 | # copy files | |
937 | # |
|
941 | # | |
938 | # We could copy the full file while the source repository is locked |
|
942 | # We could copy the full file while the source repository is locked | |
939 | # and the other one without the lock. However, in the linking case, |
|
943 | # and the other one without the lock. However, in the linking case, | |
940 | # this would also requires checks that nobody is appending any data |
|
944 | # this would also requires checks that nobody is appending any data | |
941 | # to the files while we do the clone, so this is not done yet. We |
|
945 | # to the files while we do the clone, so this is not done yet. We | |
942 | # could do this blindly when copying files. |
|
946 | # could do this blindly when copying files. | |
943 | files = [ |
|
947 | files = [ | |
944 | (vfs_key, f.unencoded_path) |
|
948 | (vfs_key, f.unencoded_path) | |
945 | for vfs_key, e in entries |
|
949 | for vfs_key, e in entries | |
946 | for f in e.files() |
|
950 | for f in e.files() | |
947 | ] |
|
951 | ] | |
948 | hardlink = _copy_files(src_vfs_map, dest_vfs_map, files, progress) |
|
952 | hardlink = _copy_files(src_vfs_map, dest_vfs_map, files, progress) | |
949 |
|
953 | |||
950 | # copy bookmarks over |
|
954 | # copy bookmarks over | |
951 | if bm_count: |
|
955 | if bm_count: | |
952 | dst_book_vfs = bookmarks.bookmarksvfs(dest_repo) |
|
956 | dst_book_vfs = bookmarks.bookmarksvfs(dest_repo) | |
953 | dstbookmarks = dst_book_vfs.join(b'bookmarks') |
|
957 | dstbookmarks = dst_book_vfs.join(b'bookmarks') | |
954 | util.copyfile(srcbookmarks, dstbookmarks) |
|
958 | util.copyfile(srcbookmarks, dstbookmarks) | |
955 | progress.complete() |
|
959 | progress.complete() | |
956 | if hardlink: |
|
960 | if hardlink: | |
957 | msg = b'linked %d files\n' |
|
961 | msg = b'linked %d files\n' | |
958 | else: |
|
962 | else: | |
959 | msg = b'copied %d files\n' |
|
963 | msg = b'copied %d files\n' | |
960 | src_repo.ui.debug(msg % total_files) |
|
964 | src_repo.ui.debug(msg % total_files) | |
961 |
|
965 | |||
962 | with dest_repo.transaction(b"localclone") as tr: |
|
966 | with dest_repo.transaction(b"localclone") as tr: | |
963 | dest_repo.store.write(tr) |
|
967 | dest_repo.store.write(tr) | |
964 |
|
968 | |||
965 | # clean up transaction file as they do not make sense |
|
969 | # clean up transaction file as they do not make sense | |
966 | transaction.cleanup_undo_files(dest_repo.ui.warn, dest_repo.vfs_map) |
|
970 | transaction.cleanup_undo_files(dest_repo.ui.warn, dest_repo.vfs_map) |
@@ -1,150 +1,150 b'' | |||||
1 | Test stream cloning while a revlog split happens |
|
1 | Test stream cloning while a revlog split happens | |
2 | ------------------------------------------------ |
|
2 | ------------------------------------------------ | |
3 |
|
3 | |||
4 | #testcases stream-bundle2-v2 stream-bundle2-v3 |
|
4 | #testcases stream-bundle2-v2 stream-bundle2-v3 | |
5 |
|
5 | |||
6 | #if stream-bundle2-v3 |
|
6 | #if stream-bundle2-v3 | |
7 | $ cat << EOF >> $HGRCPATH |
|
7 | $ cat << EOF >> $HGRCPATH | |
8 | > [experimental] |
|
8 | > [experimental] | |
9 | > stream-v3 = yes |
|
9 | > stream-v3 = yes | |
10 | > EOF |
|
10 | > EOF | |
11 | #endif |
|
11 | #endif | |
12 |
|
12 | |||
13 | setup a repository for tests |
|
13 | setup a repository for tests | |
14 | ---------------------------- |
|
14 | ---------------------------- | |
15 |
|
15 | |||
16 | $ cat >> $HGRCPATH << EOF |
|
16 | $ cat >> $HGRCPATH << EOF | |
17 | > [format] |
|
17 | > [format] | |
18 | > # skip compression to make it easy to trigger a split |
|
18 | > # skip compression to make it easy to trigger a split | |
19 | > revlog-compression=none |
|
19 | > revlog-compression=none | |
20 | > EOF |
|
20 | > EOF | |
21 |
|
21 | |||
22 | $ hg init server |
|
22 | $ hg init server | |
23 | $ cd server |
|
23 | $ cd server | |
24 | $ file="some-file" |
|
24 | $ file="some-file" | |
25 | $ printf '%20d' '1' > $file |
|
25 | $ printf '%20d' '1' > $file | |
26 | $ hg commit -Aqma |
|
26 | $ hg commit -Aqma | |
27 | $ printf '%1024d' '1' > $file |
|
27 | $ printf '%1024d' '1' > $file | |
28 | $ hg commit -Aqmb |
|
28 | $ hg commit -Aqmb | |
29 | $ printf '%20d' '1' > $file |
|
29 | $ printf '%20d' '1' > $file | |
30 | $ hg commit -Aqmc |
|
30 | $ hg commit -Aqmc | |
31 |
|
31 | |||
32 | check the revlog is inline |
|
32 | check the revlog is inline | |
33 |
|
33 | |||
34 | $ f -s .hg/store/data/some-file* |
|
34 | $ f -s .hg/store/data/some-file* | |
35 | .hg/store/data/some-file.i: size=1259 |
|
35 | .hg/store/data/some-file.i: size=1259 | |
36 | $ hg debug-revlog-index some-file |
|
36 | $ hg debug-revlog-index some-file | |
37 | rev linkrev nodeid p1-nodeid p2-nodeid |
|
37 | rev linkrev nodeid p1-nodeid p2-nodeid | |
38 | 0 0 ed70cecbc103 000000000000 000000000000 |
|
38 | 0 0 ed70cecbc103 000000000000 000000000000 | |
39 | 1 1 7241018db64c ed70cecbc103 000000000000 |
|
39 | 1 1 7241018db64c ed70cecbc103 000000000000 | |
40 | 2 2 fa1120531cc1 7241018db64c 000000000000 |
|
40 | 2 2 fa1120531cc1 7241018db64c 000000000000 | |
41 | $ cd .. |
|
41 | $ cd .. | |
42 |
|
42 | |||
43 | setup synchronisation file |
|
43 | setup synchronisation file | |
44 |
|
44 | |||
45 | $ HG_TEST_STREAM_WALKED_FILE_1="$TESTTMP/sync_file_walked_1" |
|
45 | $ HG_TEST_STREAM_WALKED_FILE_1="$TESTTMP/sync_file_walked_1" | |
46 | $ export HG_TEST_STREAM_WALKED_FILE_1 |
|
46 | $ export HG_TEST_STREAM_WALKED_FILE_1 | |
47 | $ HG_TEST_STREAM_WALKED_FILE_2="$TESTTMP/sync_file_walked_2" |
|
47 | $ HG_TEST_STREAM_WALKED_FILE_2="$TESTTMP/sync_file_walked_2" | |
48 | $ export HG_TEST_STREAM_WALKED_FILE_2 |
|
48 | $ export HG_TEST_STREAM_WALKED_FILE_2 | |
49 | $ HG_TEST_STREAM_WALKED_FILE_3="$TESTTMP/sync_file_walked_3" |
|
49 | $ HG_TEST_STREAM_WALKED_FILE_3="$TESTTMP/sync_file_walked_3" | |
50 | $ export HG_TEST_STREAM_WALKED_FILE_3 |
|
50 | $ export HG_TEST_STREAM_WALKED_FILE_3 | |
51 |
|
51 | |||
52 |
|
52 | |||
53 | Test stream-clone raced by a revlog-split |
|
53 | Test stream-clone raced by a revlog-split | |
54 | ========================================= |
|
54 | ========================================= | |
55 |
|
55 | |||
56 | Test stream-clone where the file is split right after the lock section is done |
|
56 | Test stream-clone where the file is split right after the lock section is done | |
57 |
|
57 | |||
58 | Start the server |
|
58 | Start the server | |
59 |
|
59 | |||
60 | $ hg serve -R server \ |
|
60 | $ hg serve -R server \ | |
61 | > -p $HGPORT1 -d --error errors.log --pid-file=hg.pid \ |
|
61 | > -p $HGPORT1 -d --error errors.log --pid-file=hg.pid \ | |
62 | > --config extensions.stream_steps="$RUNTESTDIR/testlib/ext-stream-clone-steps.py" |
|
62 | > --config extensions.stream_steps="$RUNTESTDIR/testlib/ext-stream-clone-steps.py" | |
63 | $ cat hg.pid >> $DAEMON_PIDS |
|
63 | $ cat hg.pid >> $DAEMON_PIDS | |
64 |
|
64 | |||
65 | Start a client doing a streaming clone |
|
65 | Start a client doing a streaming clone | |
66 |
|
66 | |||
67 | $ (hg clone -q --stream -U http://localhost:$HGPORT1 clone-while-split > client.log 2>&1; touch "$HG_TEST_STREAM_WALKED_FILE_3") & |
|
67 | $ (hg clone -q --stream -U http://localhost:$HGPORT1 clone-while-split > client.log 2>&1; touch "$HG_TEST_STREAM_WALKED_FILE_3") & | |
68 |
|
68 | |||
69 | Wait for the server to be done collecting data |
|
69 | Wait for the server to be done collecting data | |
70 |
|
70 | |||
71 | $ $RUNTESTDIR/testlib/wait-on-file 10 $HG_TEST_STREAM_WALKED_FILE_1 |
|
71 | $ $RUNTESTDIR/testlib/wait-on-file 10 $HG_TEST_STREAM_WALKED_FILE_1 | |
72 |
|
72 | |||
73 | trigger a split |
|
73 | trigger a split | |
74 |
|
74 | |||
75 | $ dd if=/dev/zero of=server/$file bs=1k count=128 > /dev/null 2>&1 |
|
75 | $ dd if=/dev/zero of=server/$file bs=1k count=128 > /dev/null 2>&1 | |
76 | $ hg -R server ci -m "triggering a split" --config ui.timeout.warn=-1 |
|
76 | $ hg -R server ci -m "triggering a split" --config ui.timeout.warn=-1 | |
77 |
|
77 | |||
78 | unlock the stream generation |
|
78 | unlock the stream generation | |
79 |
|
79 | |||
80 | $ touch $HG_TEST_STREAM_WALKED_FILE_2 |
|
80 | $ touch $HG_TEST_STREAM_WALKED_FILE_2 | |
81 |
|
81 | |||
82 | wait for the client to be done cloning. |
|
82 | wait for the client to be done cloning. | |
83 |
|
83 | |||
84 | $ $RUNTESTDIR/testlib/wait-on-file 10 $HG_TEST_STREAM_WALKED_FILE_3 |
|
84 | $ $RUNTESTDIR/testlib/wait-on-file 10 $HG_TEST_STREAM_WALKED_FILE_3 | |
85 |
|
85 | |||
86 | Check everything is fine |
|
86 | Check everything is fine | |
87 |
|
87 | |||
88 | $ cat client.log |
|
88 | $ cat client.log | |
89 |
remote: abort: unexpected error: |
|
89 | remote: abort: unexpected error: expected 0 bytes but 1067 provided for data/some-file.d (known-bad-output !) | |
90 | abort: pull failed on remote (known-bad-output !) |
|
90 | abort: pull failed on remote (known-bad-output !) | |
91 | $ tail -2 errors.log |
|
91 | $ tail -2 errors.log | |
92 |
mercurial.error.Abort: |
|
92 | mercurial.error.Abort: expected 0 bytes but 1067 provided for data/some-file.d (known-bad-output !) | |
93 | (known-bad-output !) |
|
93 | (known-bad-output !) | |
94 |
$ |
|
94 | $ hg -R clone-while-split verify | |
95 | checking changesets (missing-correct-output !) |
|
95 | checking changesets (missing-correct-output !) | |
96 | checking manifests (missing-correct-output !) |
|
96 | checking manifests (missing-correct-output !) | |
97 | crosschecking files in changesets and manifests (missing-correct-output !) |
|
97 | crosschecking files in changesets and manifests (missing-correct-output !) | |
98 | checking files (missing-correct-output !) |
|
98 | checking files (missing-correct-output !) | |
99 | checking dirstate (missing-correct-output !) |
|
99 | checking dirstate (missing-correct-output !) | |
100 | checked 3 changesets with 3 changes to 1 files (missing-correct-output !) |
|
100 | checked 3 changesets with 3 changes to 1 files (missing-correct-output !) | |
101 | abort: repository clone-while-split not found (known-bad-output !) |
|
101 | abort: repository clone-while-split not found (known-bad-output !) | |
102 | [255] |
|
102 | [255] | |
103 | $ hg -R clone-while-split tip |
|
103 | $ hg -R clone-while-split tip | |
104 | changeset: 2:dbd9854c38a6 (missing-correct-output !) |
|
104 | changeset: 2:dbd9854c38a6 (missing-correct-output !) | |
105 | tag: tip (missing-correct-output !) |
|
105 | tag: tip (missing-correct-output !) | |
106 | user: test (missing-correct-output !) |
|
106 | user: test (missing-correct-output !) | |
107 | date: Thu Jan 01 00:00:00 1970 +0000 (missing-correct-output !) |
|
107 | date: Thu Jan 01 00:00:00 1970 +0000 (missing-correct-output !) | |
108 | summary: c (missing-correct-output !) |
|
108 | summary: c (missing-correct-output !) | |
109 | (missing-correct-output !) |
|
109 | (missing-correct-output !) | |
110 | abort: repository clone-while-split not found (known-bad-output !) |
|
110 | abort: repository clone-while-split not found (known-bad-output !) | |
111 | [255] |
|
111 | [255] | |
112 | $ hg -R clone-while-split debug-revlog-index some-file |
|
112 | $ hg -R clone-while-split debug-revlog-index some-file | |
113 | rev linkrev nodeid p1-nodeid p2-nodeid (missing-correct-output !) |
|
113 | rev linkrev nodeid p1-nodeid p2-nodeid (missing-correct-output !) | |
114 | 0 0 ed70cecbc103 000000000000 000000000000 (missing-correct-output !) |
|
114 | 0 0 ed70cecbc103 000000000000 000000000000 (missing-correct-output !) | |
115 | 1 1 7241018db64c ed70cecbc103 000000000000 (missing-correct-output !) |
|
115 | 1 1 7241018db64c ed70cecbc103 000000000000 (missing-correct-output !) | |
116 | 2 2 fa1120531cc1 7241018db64c 000000000000 (missing-correct-output !) |
|
116 | 2 2 fa1120531cc1 7241018db64c 000000000000 (missing-correct-output !) | |
117 | abort: repository clone-while-split not found (known-bad-output !) |
|
117 | abort: repository clone-while-split not found (known-bad-output !) | |
118 | [255] |
|
118 | [255] | |
119 |
|
119 | |||
120 | subsequent pull work |
|
120 | subsequent pull work | |
121 |
|
121 | |||
122 | $ hg -R clone-while-split pull |
|
122 | $ hg -R clone-while-split pull | |
123 | pulling from http://localhost:$HGPORT1/ (missing-correct-output !) |
|
123 | pulling from http://localhost:$HGPORT1/ (missing-correct-output !) | |
124 | searching for changes (missing-correct-output !) |
|
124 | searching for changes (missing-correct-output !) | |
125 | adding changesets (missing-correct-output !) |
|
125 | adding changesets (missing-correct-output !) | |
126 | adding manifests (missing-correct-output !) |
|
126 | adding manifests (missing-correct-output !) | |
127 | adding file changes (missing-correct-output !) |
|
127 | adding file changes (missing-correct-output !) | |
128 | added 1 changesets with 1 changes to 1 files (missing-correct-output !) |
|
128 | added 1 changesets with 1 changes to 1 files (missing-correct-output !) | |
129 | new changesets df05c6cb1406 (missing-correct-output !) |
|
129 | new changesets df05c6cb1406 (missing-correct-output !) | |
130 | (run 'hg update' to get a working copy) (missing-correct-output !) |
|
130 | (run 'hg update' to get a working copy) (missing-correct-output !) | |
131 | abort: repository clone-while-split not found (known-bad-output !) |
|
131 | abort: repository clone-while-split not found (known-bad-output !) | |
132 | [255] |
|
132 | [255] | |
133 |
|
133 | |||
134 | $ hg -R clone-while-split debug-revlog-index some-file |
|
134 | $ hg -R clone-while-split debug-revlog-index some-file | |
135 | rev linkrev nodeid p1-nodeid p2-nodeid (missing-correct-output !) |
|
135 | rev linkrev nodeid p1-nodeid p2-nodeid (missing-correct-output !) | |
136 | 0 0 ed70cecbc103 000000000000 000000000000 (missing-correct-output !) |
|
136 | 0 0 ed70cecbc103 000000000000 000000000000 (missing-correct-output !) | |
137 | 1 1 7241018db64c ed70cecbc103 000000000000 (missing-correct-output !) |
|
137 | 1 1 7241018db64c ed70cecbc103 000000000000 (missing-correct-output !) | |
138 | 2 2 fa1120531cc1 7241018db64c 000000000000 (missing-correct-output !) |
|
138 | 2 2 fa1120531cc1 7241018db64c 000000000000 (missing-correct-output !) | |
139 | 3 3 a631378adaa3 fa1120531cc1 000000000000 (missing-correct-output !) |
|
139 | 3 3 a631378adaa3 fa1120531cc1 000000000000 (missing-correct-output !) | |
140 | abort: repository clone-while-split not found (known-bad-output !) |
|
140 | abort: repository clone-while-split not found (known-bad-output !) | |
141 | [255] |
|
141 | [255] | |
142 | $ hg -R clone-while-split verify |
|
142 | $ hg -R clone-while-split verify | |
143 | checking changesets (missing-correct-output !) |
|
143 | checking changesets (missing-correct-output !) | |
144 | checking manifests (missing-correct-output !) |
|
144 | checking manifests (missing-correct-output !) | |
145 | crosschecking files in changesets and manifests (missing-correct-output !) |
|
145 | crosschecking files in changesets and manifests (missing-correct-output !) | |
146 | checking files (missing-correct-output !) |
|
146 | checking files (missing-correct-output !) | |
147 | checking dirstate (missing-correct-output !) |
|
147 | checking dirstate (missing-correct-output !) | |
148 | checked 4 changesets with 4 changes to 1 files (missing-correct-output !) |
|
148 | checked 4 changesets with 4 changes to 1 files (missing-correct-output !) | |
149 | abort: repository clone-while-split not found (known-bad-output !) |
|
149 | abort: repository clone-while-split not found (known-bad-output !) | |
150 | [255] |
|
150 | [255] |
General Comments 0
You need to be logged in to leave comments.
Login now