Show More
@@ -1,3401 +1,3388 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 | from __future__ import absolute_import |
|
15 | from __future__ import absolute_import | |
16 |
|
16 | |||
17 | import binascii |
|
17 | import binascii | |
18 | import collections |
|
18 | import collections | |
19 | import contextlib |
|
19 | import contextlib | |
20 | import errno |
|
20 | import errno | |
21 | import io |
|
21 | import io | |
22 | import os |
|
22 | import os | |
23 | import struct |
|
23 | import struct | |
24 | import zlib |
|
24 | import zlib | |
25 |
|
25 | |||
26 | # import stuff from node for others to import from revlog |
|
26 | # import stuff from node for others to import from revlog | |
27 | from .node import ( |
|
27 | from .node import ( | |
28 | bin, |
|
28 | bin, | |
29 | hex, |
|
29 | hex, | |
30 | nullrev, |
|
30 | nullrev, | |
31 | sha1nodeconstants, |
|
31 | sha1nodeconstants, | |
32 | short, |
|
32 | short, | |
33 | wdirrev, |
|
33 | wdirrev, | |
34 | ) |
|
34 | ) | |
35 | from .i18n import _ |
|
35 | from .i18n import _ | |
36 | from .pycompat import getattr |
|
36 | from .pycompat import getattr | |
37 | from .revlogutils.constants import ( |
|
37 | from .revlogutils.constants import ( | |
38 | ALL_KINDS, |
|
38 | ALL_KINDS, | |
39 | CHANGELOGV2, |
|
39 | CHANGELOGV2, | |
40 | COMP_MODE_DEFAULT, |
|
40 | COMP_MODE_DEFAULT, | |
41 | COMP_MODE_INLINE, |
|
41 | COMP_MODE_INLINE, | |
42 | COMP_MODE_PLAIN, |
|
42 | COMP_MODE_PLAIN, | |
43 | FEATURES_BY_VERSION, |
|
43 | FEATURES_BY_VERSION, | |
44 | FLAG_GENERALDELTA, |
|
44 | FLAG_GENERALDELTA, | |
45 | FLAG_INLINE_DATA, |
|
45 | FLAG_INLINE_DATA, | |
46 | INDEX_HEADER, |
|
46 | INDEX_HEADER, | |
47 | KIND_CHANGELOG, |
|
47 | KIND_CHANGELOG, | |
48 | REVLOGV0, |
|
48 | REVLOGV0, | |
49 | REVLOGV1, |
|
49 | REVLOGV1, | |
50 | REVLOGV1_FLAGS, |
|
50 | REVLOGV1_FLAGS, | |
51 | REVLOGV2, |
|
51 | REVLOGV2, | |
52 | REVLOGV2_FLAGS, |
|
52 | REVLOGV2_FLAGS, | |
53 | REVLOG_DEFAULT_FLAGS, |
|
53 | REVLOG_DEFAULT_FLAGS, | |
54 | REVLOG_DEFAULT_FORMAT, |
|
54 | REVLOG_DEFAULT_FORMAT, | |
55 | REVLOG_DEFAULT_VERSION, |
|
55 | REVLOG_DEFAULT_VERSION, | |
56 | SUPPORTED_FLAGS, |
|
56 | SUPPORTED_FLAGS, | |
57 | ) |
|
57 | ) | |
58 | from .revlogutils.flagutil import ( |
|
58 | from .revlogutils.flagutil import ( | |
59 | REVIDX_DEFAULT_FLAGS, |
|
59 | REVIDX_DEFAULT_FLAGS, | |
60 | REVIDX_ELLIPSIS, |
|
60 | REVIDX_ELLIPSIS, | |
61 | REVIDX_EXTSTORED, |
|
61 | REVIDX_EXTSTORED, | |
62 | REVIDX_FLAGS_ORDER, |
|
62 | REVIDX_FLAGS_ORDER, | |
63 | REVIDX_HASCOPIESINFO, |
|
63 | REVIDX_HASCOPIESINFO, | |
64 | REVIDX_ISCENSORED, |
|
64 | REVIDX_ISCENSORED, | |
65 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
65 | REVIDX_RAWTEXT_CHANGING_FLAGS, | |
66 | ) |
|
66 | ) | |
67 | from .thirdparty import attr |
|
67 | from .thirdparty import attr | |
68 | from . import ( |
|
68 | from . import ( | |
69 | ancestor, |
|
69 | ancestor, | |
70 | dagop, |
|
70 | dagop, | |
71 | error, |
|
71 | error, | |
72 | mdiff, |
|
72 | mdiff, | |
73 | policy, |
|
73 | policy, | |
74 | pycompat, |
|
74 | pycompat, | |
75 | revlogutils, |
|
75 | revlogutils, | |
76 | templatefilters, |
|
76 | templatefilters, | |
77 | util, |
|
77 | util, | |
78 | ) |
|
78 | ) | |
79 | from .interfaces import ( |
|
79 | from .interfaces import ( | |
80 | repository, |
|
80 | repository, | |
81 | util as interfaceutil, |
|
81 | util as interfaceutil, | |
82 | ) |
|
82 | ) | |
83 | from .revlogutils import ( |
|
83 | from .revlogutils import ( | |
84 | censor, |
|
84 | censor, | |
85 | deltas as deltautil, |
|
85 | deltas as deltautil, | |
86 | docket as docketutil, |
|
86 | docket as docketutil, | |
87 | flagutil, |
|
87 | flagutil, | |
88 | nodemap as nodemaputil, |
|
88 | nodemap as nodemaputil, | |
89 | revlogv0, |
|
89 | revlogv0, | |
90 | sidedata as sidedatautil, |
|
90 | sidedata as sidedatautil, | |
91 | ) |
|
91 | ) | |
92 | from .utils import ( |
|
92 | from .utils import ( | |
93 | storageutil, |
|
93 | storageutil, | |
94 | stringutil, |
|
94 | stringutil, | |
95 | ) |
|
95 | ) | |
96 |
|
96 | |||
97 | # blanked usage of all the name to prevent pyflakes constraints |
|
97 | # blanked usage of all the name to prevent pyflakes constraints | |
98 | # We need these name available in the module for extensions. |
|
98 | # We need these name available in the module for extensions. | |
99 |
|
99 | |||
100 | REVLOGV0 |
|
100 | REVLOGV0 | |
101 | REVLOGV1 |
|
101 | REVLOGV1 | |
102 | REVLOGV2 |
|
102 | REVLOGV2 | |
103 | FLAG_INLINE_DATA |
|
103 | FLAG_INLINE_DATA | |
104 | FLAG_GENERALDELTA |
|
104 | FLAG_GENERALDELTA | |
105 | REVLOG_DEFAULT_FLAGS |
|
105 | REVLOG_DEFAULT_FLAGS | |
106 | REVLOG_DEFAULT_FORMAT |
|
106 | REVLOG_DEFAULT_FORMAT | |
107 | REVLOG_DEFAULT_VERSION |
|
107 | REVLOG_DEFAULT_VERSION | |
108 | REVLOGV1_FLAGS |
|
108 | REVLOGV1_FLAGS | |
109 | REVLOGV2_FLAGS |
|
109 | REVLOGV2_FLAGS | |
110 | REVIDX_ISCENSORED |
|
110 | REVIDX_ISCENSORED | |
111 | REVIDX_ELLIPSIS |
|
111 | REVIDX_ELLIPSIS | |
112 | REVIDX_HASCOPIESINFO |
|
112 | REVIDX_HASCOPIESINFO | |
113 | REVIDX_EXTSTORED |
|
113 | REVIDX_EXTSTORED | |
114 | REVIDX_DEFAULT_FLAGS |
|
114 | REVIDX_DEFAULT_FLAGS | |
115 | REVIDX_FLAGS_ORDER |
|
115 | REVIDX_FLAGS_ORDER | |
116 | REVIDX_RAWTEXT_CHANGING_FLAGS |
|
116 | REVIDX_RAWTEXT_CHANGING_FLAGS | |
117 |
|
117 | |||
118 | parsers = policy.importmod('parsers') |
|
118 | parsers = policy.importmod('parsers') | |
119 | rustancestor = policy.importrust('ancestor') |
|
119 | rustancestor = policy.importrust('ancestor') | |
120 | rustdagop = policy.importrust('dagop') |
|
120 | rustdagop = policy.importrust('dagop') | |
121 | rustrevlog = policy.importrust('revlog') |
|
121 | rustrevlog = policy.importrust('revlog') | |
122 |
|
122 | |||
123 | # Aliased for performance. |
|
123 | # Aliased for performance. | |
124 | _zlibdecompress = zlib.decompress |
|
124 | _zlibdecompress = zlib.decompress | |
125 |
|
125 | |||
126 | # max size of revlog with inline data |
|
126 | # max size of revlog with inline data | |
127 | _maxinline = 131072 |
|
127 | _maxinline = 131072 | |
128 | _chunksize = 1048576 |
|
128 | _chunksize = 1048576 | |
129 |
|
129 | |||
130 | # Flag processors for REVIDX_ELLIPSIS. |
|
130 | # Flag processors for REVIDX_ELLIPSIS. | |
131 | def ellipsisreadprocessor(rl, text): |
|
131 | def ellipsisreadprocessor(rl, text): | |
132 | return text, False |
|
132 | return text, False | |
133 |
|
133 | |||
134 |
|
134 | |||
135 | def ellipsiswriteprocessor(rl, text): |
|
135 | def ellipsiswriteprocessor(rl, text): | |
136 | return text, False |
|
136 | return text, False | |
137 |
|
137 | |||
138 |
|
138 | |||
139 | def ellipsisrawprocessor(rl, text): |
|
139 | def ellipsisrawprocessor(rl, text): | |
140 | return False |
|
140 | return False | |
141 |
|
141 | |||
142 |
|
142 | |||
143 | ellipsisprocessor = ( |
|
143 | ellipsisprocessor = ( | |
144 | ellipsisreadprocessor, |
|
144 | ellipsisreadprocessor, | |
145 | ellipsiswriteprocessor, |
|
145 | ellipsiswriteprocessor, | |
146 | ellipsisrawprocessor, |
|
146 | ellipsisrawprocessor, | |
147 | ) |
|
147 | ) | |
148 |
|
148 | |||
149 |
|
149 | |||
150 | def _verify_revision(rl, skipflags, state, node): |
|
150 | def _verify_revision(rl, skipflags, state, node): | |
151 | """Verify the integrity of the given revlog ``node`` while providing a hook |
|
151 | """Verify the integrity of the given revlog ``node`` while providing a hook | |
152 | point for extensions to influence the operation.""" |
|
152 | point for extensions to influence the operation.""" | |
153 | if skipflags: |
|
153 | if skipflags: | |
154 | state[b'skipread'].add(node) |
|
154 | state[b'skipread'].add(node) | |
155 | else: |
|
155 | else: | |
156 | # Side-effect: read content and verify hash. |
|
156 | # Side-effect: read content and verify hash. | |
157 | rl.revision(node) |
|
157 | rl.revision(node) | |
158 |
|
158 | |||
159 |
|
159 | |||
160 | # True if a fast implementation for persistent-nodemap is available |
|
160 | # True if a fast implementation for persistent-nodemap is available | |
161 | # |
|
161 | # | |
162 | # We also consider we have a "fast" implementation in "pure" python because |
|
162 | # We also consider we have a "fast" implementation in "pure" python because | |
163 | # people using pure don't really have performance consideration (and a |
|
163 | # people using pure don't really have performance consideration (and a | |
164 | # wheelbarrow of other slowness source) |
|
164 | # wheelbarrow of other slowness source) | |
165 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or util.safehasattr( |
|
165 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or util.safehasattr( | |
166 | parsers, 'BaseIndexObject' |
|
166 | parsers, 'BaseIndexObject' | |
167 | ) |
|
167 | ) | |
168 |
|
168 | |||
169 |
|
169 | |||
170 | @attr.s(slots=True, frozen=True) |
|
|||
171 | class _revisioninfo(object): |
|
|||
172 | """Information about a revision that allows building its fulltext |
|
|||
173 | node: expected hash of the revision |
|
|||
174 | p1, p2: parent revs of the revision |
|
|||
175 | btext: built text cache consisting of a one-element list |
|
|||
176 | cachedelta: (baserev, uncompressed_delta) or None |
|
|||
177 | flags: flags associated to the revision storage |
|
|||
178 |
|
||||
179 | One of btext[0] or cachedelta must be set. |
|
|||
180 | """ |
|
|||
181 |
|
||||
182 | node = attr.ib() |
|
|||
183 | p1 = attr.ib() |
|
|||
184 | p2 = attr.ib() |
|
|||
185 | btext = attr.ib() |
|
|||
186 | textlen = attr.ib() |
|
|||
187 | cachedelta = attr.ib() |
|
|||
188 | flags = attr.ib() |
|
|||
189 |
|
||||
190 |
|
||||
191 | @interfaceutil.implementer(repository.irevisiondelta) |
|
170 | @interfaceutil.implementer(repository.irevisiondelta) | |
192 | @attr.s(slots=True) |
|
171 | @attr.s(slots=True) | |
193 | class revlogrevisiondelta(object): |
|
172 | class revlogrevisiondelta(object): | |
194 | node = attr.ib() |
|
173 | node = attr.ib() | |
195 | p1node = attr.ib() |
|
174 | p1node = attr.ib() | |
196 | p2node = attr.ib() |
|
175 | p2node = attr.ib() | |
197 | basenode = attr.ib() |
|
176 | basenode = attr.ib() | |
198 | flags = attr.ib() |
|
177 | flags = attr.ib() | |
199 | baserevisionsize = attr.ib() |
|
178 | baserevisionsize = attr.ib() | |
200 | revision = attr.ib() |
|
179 | revision = attr.ib() | |
201 | delta = attr.ib() |
|
180 | delta = attr.ib() | |
202 | sidedata = attr.ib() |
|
181 | sidedata = attr.ib() | |
203 | protocol_flags = attr.ib() |
|
182 | protocol_flags = attr.ib() | |
204 | linknode = attr.ib(default=None) |
|
183 | linknode = attr.ib(default=None) | |
205 |
|
184 | |||
206 |
|
185 | |||
207 | @interfaceutil.implementer(repository.iverifyproblem) |
|
186 | @interfaceutil.implementer(repository.iverifyproblem) | |
208 | @attr.s(frozen=True) |
|
187 | @attr.s(frozen=True) | |
209 | class revlogproblem(object): |
|
188 | class revlogproblem(object): | |
210 | warning = attr.ib(default=None) |
|
189 | warning = attr.ib(default=None) | |
211 | error = attr.ib(default=None) |
|
190 | error = attr.ib(default=None) | |
212 | node = attr.ib(default=None) |
|
191 | node = attr.ib(default=None) | |
213 |
|
192 | |||
214 |
|
193 | |||
215 | def parse_index_v1(data, inline): |
|
194 | def parse_index_v1(data, inline): | |
216 | # call the C implementation to parse the index data |
|
195 | # call the C implementation to parse the index data | |
217 | index, cache = parsers.parse_index2(data, inline) |
|
196 | index, cache = parsers.parse_index2(data, inline) | |
218 | return index, cache |
|
197 | return index, cache | |
219 |
|
198 | |||
220 |
|
199 | |||
221 | def parse_index_v2(data, inline): |
|
200 | def parse_index_v2(data, inline): | |
222 | # call the C implementation to parse the index data |
|
201 | # call the C implementation to parse the index data | |
223 | index, cache = parsers.parse_index2(data, inline, revlogv2=True) |
|
202 | index, cache = parsers.parse_index2(data, inline, revlogv2=True) | |
224 | return index, cache |
|
203 | return index, cache | |
225 |
|
204 | |||
226 |
|
205 | |||
227 | def parse_index_cl_v2(data, inline): |
|
206 | def parse_index_cl_v2(data, inline): | |
228 | # call the C implementation to parse the index data |
|
207 | # call the C implementation to parse the index data | |
229 | assert not inline |
|
208 | assert not inline | |
230 | from .pure.parsers import parse_index_cl_v2 |
|
209 | from .pure.parsers import parse_index_cl_v2 | |
231 |
|
210 | |||
232 | index, cache = parse_index_cl_v2(data) |
|
211 | index, cache = parse_index_cl_v2(data) | |
233 | return index, cache |
|
212 | return index, cache | |
234 |
|
213 | |||
235 |
|
214 | |||
236 | if util.safehasattr(parsers, 'parse_index_devel_nodemap'): |
|
215 | if util.safehasattr(parsers, 'parse_index_devel_nodemap'): | |
237 |
|
216 | |||
238 | def parse_index_v1_nodemap(data, inline): |
|
217 | def parse_index_v1_nodemap(data, inline): | |
239 | index, cache = parsers.parse_index_devel_nodemap(data, inline) |
|
218 | index, cache = parsers.parse_index_devel_nodemap(data, inline) | |
240 | return index, cache |
|
219 | return index, cache | |
241 |
|
220 | |||
242 |
|
221 | |||
243 | else: |
|
222 | else: | |
244 | parse_index_v1_nodemap = None |
|
223 | parse_index_v1_nodemap = None | |
245 |
|
224 | |||
246 |
|
225 | |||
247 | def parse_index_v1_mixed(data, inline): |
|
226 | def parse_index_v1_mixed(data, inline): | |
248 | index, cache = parse_index_v1(data, inline) |
|
227 | index, cache = parse_index_v1(data, inline) | |
249 | return rustrevlog.MixedIndex(index), cache |
|
228 | return rustrevlog.MixedIndex(index), cache | |
250 |
|
229 | |||
251 |
|
230 | |||
252 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte |
|
231 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte | |
253 | # signed integer) |
|
232 | # signed integer) | |
254 | _maxentrysize = 0x7FFFFFFF |
|
233 | _maxentrysize = 0x7FFFFFFF | |
255 |
|
234 | |||
256 | PARTIAL_READ_MSG = _( |
|
235 | PARTIAL_READ_MSG = _( | |
257 | b'partial read of revlog %s; expected %d bytes from offset %d, got %d' |
|
236 | b'partial read of revlog %s; expected %d bytes from offset %d, got %d' | |
258 | ) |
|
237 | ) | |
259 |
|
238 | |||
260 | FILE_TOO_SHORT_MSG = _( |
|
239 | FILE_TOO_SHORT_MSG = _( | |
261 | b'cannot read from revlog %s;' |
|
240 | b'cannot read from revlog %s;' | |
262 | b' expected %d bytes from offset %d, data size is %d' |
|
241 | b' expected %d bytes from offset %d, data size is %d' | |
263 | ) |
|
242 | ) | |
264 |
|
243 | |||
265 |
|
244 | |||
266 | class revlog(object): |
|
245 | class revlog(object): | |
267 | """ |
|
246 | """ | |
268 | the underlying revision storage object |
|
247 | the underlying revision storage object | |
269 |
|
248 | |||
270 | 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. | |
271 |
|
250 | |||
272 | The index is a file with a fixed record size containing |
|
251 | The index is a file with a fixed record size containing | |
273 | information on each revision, including its nodeid (hash), the |
|
252 | information on each revision, including its nodeid (hash), the | |
274 | 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 | |
275 | 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 | |
276 | 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 | |
277 | data. |
|
256 | data. | |
278 |
|
257 | |||
279 | The revision data itself is a linear collection of data chunks. |
|
258 | The revision data itself is a linear collection of data chunks. | |
280 | Each chunk represents a revision and is usually represented as a |
|
259 | Each chunk represents a revision and is usually represented as a | |
281 | delta against the previous chunk. To bound lookup time, runs of |
|
260 | delta against the previous chunk. To bound lookup time, runs of | |
282 | 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 | |
283 | version data. This makes retrieval of a version proportional to |
|
262 | version data. This makes retrieval of a version proportional to | |
284 | its size, or O(1) relative to the number of revisions. |
|
263 | its size, or O(1) relative to the number of revisions. | |
285 |
|
264 | |||
286 | 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 | |
287 | 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 | |
288 | 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 | |
289 | for locking while reading. |
|
268 | for locking while reading. | |
290 |
|
269 | |||
291 | If checkambig, indexfile is opened with checkambig=True at |
|
270 | If checkambig, indexfile is opened with checkambig=True at | |
292 | writing, to avoid file stat ambiguity. |
|
271 | writing, to avoid file stat ambiguity. | |
293 |
|
272 | |||
294 | If mmaplargeindex is True, and an mmapindexthreshold is set, the |
|
273 | If mmaplargeindex is True, and an mmapindexthreshold is set, the | |
295 | 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 | |
296 | configured threshold. |
|
275 | configured threshold. | |
297 |
|
276 | |||
298 | If censorable is True, the revlog can have censored revisions. |
|
277 | If censorable is True, the revlog can have censored revisions. | |
299 |
|
278 | |||
300 | 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 | |
301 | compression for the data content. |
|
280 | compression for the data content. | |
302 |
|
281 | |||
303 | `concurrencychecker` is an optional function that receives 3 arguments: a |
|
282 | `concurrencychecker` is an optional function that receives 3 arguments: a | |
304 | 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 | |
305 | 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 | |
306 | raising). |
|
285 | raising). | |
307 |
|
286 | |||
308 | 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 | |
309 | index entry. |
|
288 | index entry. | |
310 | """ |
|
289 | """ | |
311 |
|
290 | |||
312 | _flagserrorclass = error.RevlogError |
|
291 | _flagserrorclass = error.RevlogError | |
313 |
|
292 | |||
314 | def __init__( |
|
293 | def __init__( | |
315 | self, |
|
294 | self, | |
316 | opener, |
|
295 | opener, | |
317 | target, |
|
296 | target, | |
318 | radix, |
|
297 | radix, | |
319 | postfix=None, # only exist for `tmpcensored` now |
|
298 | postfix=None, # only exist for `tmpcensored` now | |
320 | checkambig=False, |
|
299 | checkambig=False, | |
321 | mmaplargeindex=False, |
|
300 | mmaplargeindex=False, | |
322 | censorable=False, |
|
301 | censorable=False, | |
323 | upperboundcomp=None, |
|
302 | upperboundcomp=None, | |
324 | persistentnodemap=False, |
|
303 | persistentnodemap=False, | |
325 | concurrencychecker=None, |
|
304 | concurrencychecker=None, | |
326 | trypending=False, |
|
305 | trypending=False, | |
327 | ): |
|
306 | ): | |
328 | """ |
|
307 | """ | |
329 | create a revlog object |
|
308 | create a revlog object | |
330 |
|
309 | |||
331 | opener is a function that abstracts the file opening operation |
|
310 | opener is a function that abstracts the file opening operation | |
332 | and can be used to implement COW semantics or the like. |
|
311 | and can be used to implement COW semantics or the like. | |
333 |
|
312 | |||
334 | `target`: a (KIND, ID) tuple that identify the content stored in |
|
313 | `target`: a (KIND, ID) tuple that identify the content stored in | |
335 | this revlog. It help the rest of the code to understand what the revlog |
|
314 | this revlog. It help the rest of the code to understand what the revlog | |
336 | is about without having to resort to heuristic and index filename |
|
315 | is about without having to resort to heuristic and index filename | |
337 | analysis. Note: that this must be reliably be set by normal code, but |
|
316 | analysis. Note: that this must be reliably be set by normal code, but | |
338 | that test, debug, or performance measurement code might not set this to |
|
317 | that test, debug, or performance measurement code might not set this to | |
339 | accurate value. |
|
318 | accurate value. | |
340 | """ |
|
319 | """ | |
341 | self.upperboundcomp = upperboundcomp |
|
320 | self.upperboundcomp = upperboundcomp | |
342 |
|
321 | |||
343 | self.radix = radix |
|
322 | self.radix = radix | |
344 |
|
323 | |||
345 | self._docket_file = None |
|
324 | self._docket_file = None | |
346 | self._indexfile = None |
|
325 | self._indexfile = None | |
347 | self._datafile = None |
|
326 | self._datafile = None | |
348 | self._sidedatafile = None |
|
327 | self._sidedatafile = None | |
349 | self._nodemap_file = None |
|
328 | self._nodemap_file = None | |
350 | self.postfix = postfix |
|
329 | self.postfix = postfix | |
351 | self._trypending = trypending |
|
330 | self._trypending = trypending | |
352 | self.opener = opener |
|
331 | self.opener = opener | |
353 | if persistentnodemap: |
|
332 | if persistentnodemap: | |
354 | self._nodemap_file = nodemaputil.get_nodemap_file(self) |
|
333 | self._nodemap_file = nodemaputil.get_nodemap_file(self) | |
355 |
|
334 | |||
356 | assert target[0] in ALL_KINDS |
|
335 | assert target[0] in ALL_KINDS | |
357 | assert len(target) == 2 |
|
336 | assert len(target) == 2 | |
358 | self.target = target |
|
337 | self.target = target | |
359 | # When True, indexfile is opened with checkambig=True at writing, to |
|
338 | # When True, indexfile is opened with checkambig=True at writing, to | |
360 | # avoid file stat ambiguity. |
|
339 | # avoid file stat ambiguity. | |
361 | self._checkambig = checkambig |
|
340 | self._checkambig = checkambig | |
362 | self._mmaplargeindex = mmaplargeindex |
|
341 | self._mmaplargeindex = mmaplargeindex | |
363 | self._censorable = censorable |
|
342 | self._censorable = censorable | |
364 | # 3-tuple of (node, rev, text) for a raw revision. |
|
343 | # 3-tuple of (node, rev, text) for a raw revision. | |
365 | self._revisioncache = None |
|
344 | self._revisioncache = None | |
366 | # Maps rev to chain base rev. |
|
345 | # Maps rev to chain base rev. | |
367 | self._chainbasecache = util.lrucachedict(100) |
|
346 | self._chainbasecache = util.lrucachedict(100) | |
368 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. |
|
347 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. | |
369 | self._chunkcache = (0, b'') |
|
348 | self._chunkcache = (0, b'') | |
370 | # How much data to read and cache into the raw revlog data cache. |
|
349 | # How much data to read and cache into the raw revlog data cache. | |
371 | self._chunkcachesize = 65536 |
|
350 | self._chunkcachesize = 65536 | |
372 | self._maxchainlen = None |
|
351 | self._maxchainlen = None | |
373 | self._deltabothparents = True |
|
352 | self._deltabothparents = True | |
374 | self.index = None |
|
353 | self.index = None | |
375 | self._docket = None |
|
354 | self._docket = None | |
376 | self._nodemap_docket = None |
|
355 | self._nodemap_docket = None | |
377 | # Mapping of partial identifiers to full nodes. |
|
356 | # Mapping of partial identifiers to full nodes. | |
378 | self._pcache = {} |
|
357 | self._pcache = {} | |
379 | # Mapping of revision integer to full node. |
|
358 | # Mapping of revision integer to full node. | |
380 | self._compengine = b'zlib' |
|
359 | self._compengine = b'zlib' | |
381 | self._compengineopts = {} |
|
360 | self._compengineopts = {} | |
382 | self._maxdeltachainspan = -1 |
|
361 | self._maxdeltachainspan = -1 | |
383 | self._withsparseread = False |
|
362 | self._withsparseread = False | |
384 | self._sparserevlog = False |
|
363 | self._sparserevlog = False | |
385 | self.hassidedata = False |
|
364 | self.hassidedata = False | |
386 | self._srdensitythreshold = 0.50 |
|
365 | self._srdensitythreshold = 0.50 | |
387 | self._srmingapsize = 262144 |
|
366 | self._srmingapsize = 262144 | |
388 |
|
367 | |||
389 | # Make copy of flag processors so each revlog instance can support |
|
368 | # Make copy of flag processors so each revlog instance can support | |
390 | # custom flags. |
|
369 | # custom flags. | |
391 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
370 | self._flagprocessors = dict(flagutil.flagprocessors) | |
392 |
|
371 | |||
393 | # 3-tuple of file handles being used for active writing. |
|
372 | # 3-tuple of file handles being used for active writing. | |
394 | self._writinghandles = None |
|
373 | self._writinghandles = None | |
395 | # prevent nesting of addgroup |
|
374 | # prevent nesting of addgroup | |
396 | self._adding_group = None |
|
375 | self._adding_group = None | |
397 |
|
376 | |||
398 | self._loadindex() |
|
377 | self._loadindex() | |
399 |
|
378 | |||
400 | self._concurrencychecker = concurrencychecker |
|
379 | self._concurrencychecker = concurrencychecker | |
401 |
|
380 | |||
402 | def _init_opts(self): |
|
381 | def _init_opts(self): | |
403 | """process options (from above/config) to setup associated default revlog mode |
|
382 | """process options (from above/config) to setup associated default revlog mode | |
404 |
|
383 | |||
405 | These values might be affected when actually reading on disk information. |
|
384 | These values might be affected when actually reading on disk information. | |
406 |
|
385 | |||
407 | The relevant values are returned for use in _loadindex(). |
|
386 | The relevant values are returned for use in _loadindex(). | |
408 |
|
387 | |||
409 | * newversionflags: |
|
388 | * newversionflags: | |
410 | version header to use if we need to create a new revlog |
|
389 | version header to use if we need to create a new revlog | |
411 |
|
390 | |||
412 | * mmapindexthreshold: |
|
391 | * mmapindexthreshold: | |
413 | minimal index size for start to use mmap |
|
392 | minimal index size for start to use mmap | |
414 |
|
393 | |||
415 | * force_nodemap: |
|
394 | * force_nodemap: | |
416 | force the usage of a "development" version of the nodemap code |
|
395 | force the usage of a "development" version of the nodemap code | |
417 | """ |
|
396 | """ | |
418 | mmapindexthreshold = None |
|
397 | mmapindexthreshold = None | |
419 | opts = self.opener.options |
|
398 | opts = self.opener.options | |
420 |
|
399 | |||
421 | if b'changelogv2' in opts and self.revlog_kind == KIND_CHANGELOG: |
|
400 | if b'changelogv2' in opts and self.revlog_kind == KIND_CHANGELOG: | |
422 | new_header = CHANGELOGV2 |
|
401 | new_header = CHANGELOGV2 | |
423 | elif b'revlogv2' in opts: |
|
402 | elif b'revlogv2' in opts: | |
424 | new_header = REVLOGV2 |
|
403 | new_header = REVLOGV2 | |
425 | elif b'revlogv1' in opts: |
|
404 | elif b'revlogv1' in opts: | |
426 | new_header = REVLOGV1 | FLAG_INLINE_DATA |
|
405 | new_header = REVLOGV1 | FLAG_INLINE_DATA | |
427 | if b'generaldelta' in opts: |
|
406 | if b'generaldelta' in opts: | |
428 | new_header |= FLAG_GENERALDELTA |
|
407 | new_header |= FLAG_GENERALDELTA | |
429 | elif b'revlogv0' in self.opener.options: |
|
408 | elif b'revlogv0' in self.opener.options: | |
430 | new_header = REVLOGV0 |
|
409 | new_header = REVLOGV0 | |
431 | else: |
|
410 | else: | |
432 | new_header = REVLOG_DEFAULT_VERSION |
|
411 | new_header = REVLOG_DEFAULT_VERSION | |
433 |
|
412 | |||
434 | if b'chunkcachesize' in opts: |
|
413 | if b'chunkcachesize' in opts: | |
435 | self._chunkcachesize = opts[b'chunkcachesize'] |
|
414 | self._chunkcachesize = opts[b'chunkcachesize'] | |
436 | if b'maxchainlen' in opts: |
|
415 | if b'maxchainlen' in opts: | |
437 | self._maxchainlen = opts[b'maxchainlen'] |
|
416 | self._maxchainlen = opts[b'maxchainlen'] | |
438 | if b'deltabothparents' in opts: |
|
417 | if b'deltabothparents' in opts: | |
439 | self._deltabothparents = opts[b'deltabothparents'] |
|
418 | self._deltabothparents = opts[b'deltabothparents'] | |
440 | self._lazydelta = bool(opts.get(b'lazydelta', True)) |
|
419 | self._lazydelta = bool(opts.get(b'lazydelta', True)) | |
441 | self._lazydeltabase = False |
|
420 | self._lazydeltabase = False | |
442 | if self._lazydelta: |
|
421 | if self._lazydelta: | |
443 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) |
|
422 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) | |
444 | if b'compengine' in opts: |
|
423 | if b'compengine' in opts: | |
445 | self._compengine = opts[b'compengine'] |
|
424 | self._compengine = opts[b'compengine'] | |
446 | if b'zlib.level' in opts: |
|
425 | if b'zlib.level' in opts: | |
447 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] |
|
426 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] | |
448 | if b'zstd.level' in opts: |
|
427 | if b'zstd.level' in opts: | |
449 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] |
|
428 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] | |
450 | if b'maxdeltachainspan' in opts: |
|
429 | if b'maxdeltachainspan' in opts: | |
451 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] |
|
430 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] | |
452 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: |
|
431 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: | |
453 | mmapindexthreshold = opts[b'mmapindexthreshold'] |
|
432 | mmapindexthreshold = opts[b'mmapindexthreshold'] | |
454 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) |
|
433 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) | |
455 | withsparseread = bool(opts.get(b'with-sparse-read', False)) |
|
434 | withsparseread = bool(opts.get(b'with-sparse-read', False)) | |
456 | # sparse-revlog forces sparse-read |
|
435 | # sparse-revlog forces sparse-read | |
457 | self._withsparseread = self._sparserevlog or withsparseread |
|
436 | self._withsparseread = self._sparserevlog or withsparseread | |
458 | if b'sparse-read-density-threshold' in opts: |
|
437 | if b'sparse-read-density-threshold' in opts: | |
459 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] |
|
438 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] | |
460 | if b'sparse-read-min-gap-size' in opts: |
|
439 | if b'sparse-read-min-gap-size' in opts: | |
461 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] |
|
440 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] | |
462 | if opts.get(b'enableellipsis'): |
|
441 | if opts.get(b'enableellipsis'): | |
463 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor |
|
442 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor | |
464 |
|
443 | |||
465 | # revlog v0 doesn't have flag processors |
|
444 | # revlog v0 doesn't have flag processors | |
466 | for flag, processor in pycompat.iteritems( |
|
445 | for flag, processor in pycompat.iteritems( | |
467 | opts.get(b'flagprocessors', {}) |
|
446 | opts.get(b'flagprocessors', {}) | |
468 | ): |
|
447 | ): | |
469 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) |
|
448 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) | |
470 |
|
449 | |||
471 | if self._chunkcachesize <= 0: |
|
450 | if self._chunkcachesize <= 0: | |
472 | raise error.RevlogError( |
|
451 | raise error.RevlogError( | |
473 | _(b'revlog chunk cache size %r is not greater than 0') |
|
452 | _(b'revlog chunk cache size %r is not greater than 0') | |
474 | % self._chunkcachesize |
|
453 | % self._chunkcachesize | |
475 | ) |
|
454 | ) | |
476 | elif self._chunkcachesize & (self._chunkcachesize - 1): |
|
455 | elif self._chunkcachesize & (self._chunkcachesize - 1): | |
477 | raise error.RevlogError( |
|
456 | raise error.RevlogError( | |
478 | _(b'revlog chunk cache size %r is not a power of 2') |
|
457 | _(b'revlog chunk cache size %r is not a power of 2') | |
479 | % self._chunkcachesize |
|
458 | % self._chunkcachesize | |
480 | ) |
|
459 | ) | |
481 | force_nodemap = opts.get(b'devel-force-nodemap', False) |
|
460 | force_nodemap = opts.get(b'devel-force-nodemap', False) | |
482 | return new_header, mmapindexthreshold, force_nodemap |
|
461 | return new_header, mmapindexthreshold, force_nodemap | |
483 |
|
462 | |||
484 | def _get_data(self, filepath, mmap_threshold, size=None): |
|
463 | def _get_data(self, filepath, mmap_threshold, size=None): | |
485 | """return a file content with or without mmap |
|
464 | """return a file content with or without mmap | |
486 |
|
465 | |||
487 | If the file is missing return the empty string""" |
|
466 | If the file is missing return the empty string""" | |
488 | try: |
|
467 | try: | |
489 | with self.opener(filepath) as fp: |
|
468 | with self.opener(filepath) as fp: | |
490 | if mmap_threshold is not None: |
|
469 | if mmap_threshold is not None: | |
491 | file_size = self.opener.fstat(fp).st_size |
|
470 | file_size = self.opener.fstat(fp).st_size | |
492 | if file_size >= mmap_threshold: |
|
471 | if file_size >= mmap_threshold: | |
493 | if size is not None: |
|
472 | if size is not None: | |
494 | # avoid potentiel mmap crash |
|
473 | # avoid potentiel mmap crash | |
495 | size = min(file_size, size) |
|
474 | size = min(file_size, size) | |
496 | # TODO: should .close() to release resources without |
|
475 | # TODO: should .close() to release resources without | |
497 | # relying on Python GC |
|
476 | # relying on Python GC | |
498 | if size is None: |
|
477 | if size is None: | |
499 | return util.buffer(util.mmapread(fp)) |
|
478 | return util.buffer(util.mmapread(fp)) | |
500 | else: |
|
479 | else: | |
501 | return util.buffer(util.mmapread(fp, size)) |
|
480 | return util.buffer(util.mmapread(fp, size)) | |
502 | if size is None: |
|
481 | if size is None: | |
503 | return fp.read() |
|
482 | return fp.read() | |
504 | else: |
|
483 | else: | |
505 | return fp.read(size) |
|
484 | return fp.read(size) | |
506 | except IOError as inst: |
|
485 | except IOError as inst: | |
507 | if inst.errno != errno.ENOENT: |
|
486 | if inst.errno != errno.ENOENT: | |
508 | raise |
|
487 | raise | |
509 | return b'' |
|
488 | return b'' | |
510 |
|
489 | |||
511 | def _loadindex(self): |
|
490 | def _loadindex(self): | |
512 |
|
491 | |||
513 | new_header, mmapindexthreshold, force_nodemap = self._init_opts() |
|
492 | new_header, mmapindexthreshold, force_nodemap = self._init_opts() | |
514 |
|
493 | |||
515 | if self.postfix is not None: |
|
494 | if self.postfix is not None: | |
516 | entry_point = b'%s.i.%s' % (self.radix, self.postfix) |
|
495 | entry_point = b'%s.i.%s' % (self.radix, self.postfix) | |
517 | elif self._trypending and self.opener.exists(b'%s.i.a' % self.radix): |
|
496 | elif self._trypending and self.opener.exists(b'%s.i.a' % self.radix): | |
518 | entry_point = b'%s.i.a' % self.radix |
|
497 | entry_point = b'%s.i.a' % self.radix | |
519 | else: |
|
498 | else: | |
520 | entry_point = b'%s.i' % self.radix |
|
499 | entry_point = b'%s.i' % self.radix | |
521 |
|
500 | |||
522 | entry_data = b'' |
|
501 | entry_data = b'' | |
523 | self._initempty = True |
|
502 | self._initempty = True | |
524 | entry_data = self._get_data(entry_point, mmapindexthreshold) |
|
503 | entry_data = self._get_data(entry_point, mmapindexthreshold) | |
525 | if len(entry_data) > 0: |
|
504 | if len(entry_data) > 0: | |
526 | header = INDEX_HEADER.unpack(entry_data[:4])[0] |
|
505 | header = INDEX_HEADER.unpack(entry_data[:4])[0] | |
527 | self._initempty = False |
|
506 | self._initempty = False | |
528 | else: |
|
507 | else: | |
529 | header = new_header |
|
508 | header = new_header | |
530 |
|
509 | |||
531 | self._format_flags = header & ~0xFFFF |
|
510 | self._format_flags = header & ~0xFFFF | |
532 | self._format_version = header & 0xFFFF |
|
511 | self._format_version = header & 0xFFFF | |
533 |
|
512 | |||
534 | supported_flags = SUPPORTED_FLAGS.get(self._format_version) |
|
513 | supported_flags = SUPPORTED_FLAGS.get(self._format_version) | |
535 | if supported_flags is None: |
|
514 | if supported_flags is None: | |
536 | msg = _(b'unknown version (%d) in revlog %s') |
|
515 | msg = _(b'unknown version (%d) in revlog %s') | |
537 | msg %= (self._format_version, self.display_id) |
|
516 | msg %= (self._format_version, self.display_id) | |
538 | raise error.RevlogError(msg) |
|
517 | raise error.RevlogError(msg) | |
539 | elif self._format_flags & ~supported_flags: |
|
518 | elif self._format_flags & ~supported_flags: | |
540 | msg = _(b'unknown flags (%#04x) in version %d revlog %s') |
|
519 | msg = _(b'unknown flags (%#04x) in version %d revlog %s') | |
541 | display_flag = self._format_flags >> 16 |
|
520 | display_flag = self._format_flags >> 16 | |
542 | msg %= (display_flag, self._format_version, self.display_id) |
|
521 | msg %= (display_flag, self._format_version, self.display_id) | |
543 | raise error.RevlogError(msg) |
|
522 | raise error.RevlogError(msg) | |
544 |
|
523 | |||
545 | features = FEATURES_BY_VERSION[self._format_version] |
|
524 | features = FEATURES_BY_VERSION[self._format_version] | |
546 | self._inline = features[b'inline'](self._format_flags) |
|
525 | self._inline = features[b'inline'](self._format_flags) | |
547 | self._generaldelta = features[b'generaldelta'](self._format_flags) |
|
526 | self._generaldelta = features[b'generaldelta'](self._format_flags) | |
548 | self.hassidedata = features[b'sidedata'] |
|
527 | self.hassidedata = features[b'sidedata'] | |
549 |
|
528 | |||
550 | if not features[b'docket']: |
|
529 | if not features[b'docket']: | |
551 | self._indexfile = entry_point |
|
530 | self._indexfile = entry_point | |
552 | index_data = entry_data |
|
531 | index_data = entry_data | |
553 | else: |
|
532 | else: | |
554 | self._docket_file = entry_point |
|
533 | self._docket_file = entry_point | |
555 | if self._initempty: |
|
534 | if self._initempty: | |
556 | self._docket = docketutil.default_docket(self, header) |
|
535 | self._docket = docketutil.default_docket(self, header) | |
557 | else: |
|
536 | else: | |
558 | self._docket = docketutil.parse_docket( |
|
537 | self._docket = docketutil.parse_docket( | |
559 | self, entry_data, use_pending=self._trypending |
|
538 | self, entry_data, use_pending=self._trypending | |
560 | ) |
|
539 | ) | |
561 | self._indexfile = self._docket.index_filepath() |
|
540 | self._indexfile = self._docket.index_filepath() | |
562 | index_data = b'' |
|
541 | index_data = b'' | |
563 | index_size = self._docket.index_end |
|
542 | index_size = self._docket.index_end | |
564 | if index_size > 0: |
|
543 | if index_size > 0: | |
565 | index_data = self._get_data( |
|
544 | index_data = self._get_data( | |
566 | self._indexfile, mmapindexthreshold, size=index_size |
|
545 | self._indexfile, mmapindexthreshold, size=index_size | |
567 | ) |
|
546 | ) | |
568 | if len(index_data) < index_size: |
|
547 | if len(index_data) < index_size: | |
569 | msg = _(b'too few index data for %s: got %d, expected %d') |
|
548 | msg = _(b'too few index data for %s: got %d, expected %d') | |
570 | msg %= (self.display_id, len(index_data), index_size) |
|
549 | msg %= (self.display_id, len(index_data), index_size) | |
571 | raise error.RevlogError(msg) |
|
550 | raise error.RevlogError(msg) | |
572 |
|
551 | |||
573 | self._inline = False |
|
552 | self._inline = False | |
574 | # generaldelta implied by version 2 revlogs. |
|
553 | # generaldelta implied by version 2 revlogs. | |
575 | self._generaldelta = True |
|
554 | self._generaldelta = True | |
576 | # the logic for persistent nodemap will be dealt with within the |
|
555 | # the logic for persistent nodemap will be dealt with within the | |
577 | # main docket, so disable it for now. |
|
556 | # main docket, so disable it for now. | |
578 | self._nodemap_file = None |
|
557 | self._nodemap_file = None | |
579 |
|
558 | |||
580 | if self._docket is not None: |
|
559 | if self._docket is not None: | |
581 | self._datafile = self._docket.data_filepath() |
|
560 | self._datafile = self._docket.data_filepath() | |
582 | self._sidedatafile = self._docket.sidedata_filepath() |
|
561 | self._sidedatafile = self._docket.sidedata_filepath() | |
583 | elif self.postfix is None: |
|
562 | elif self.postfix is None: | |
584 | self._datafile = b'%s.d' % self.radix |
|
563 | self._datafile = b'%s.d' % self.radix | |
585 | else: |
|
564 | else: | |
586 | self._datafile = b'%s.d.%s' % (self.radix, self.postfix) |
|
565 | self._datafile = b'%s.d.%s' % (self.radix, self.postfix) | |
587 |
|
566 | |||
588 | self.nodeconstants = sha1nodeconstants |
|
567 | self.nodeconstants = sha1nodeconstants | |
589 | self.nullid = self.nodeconstants.nullid |
|
568 | self.nullid = self.nodeconstants.nullid | |
590 |
|
569 | |||
591 | # sparse-revlog can't be on without general-delta (issue6056) |
|
570 | # sparse-revlog can't be on without general-delta (issue6056) | |
592 | if not self._generaldelta: |
|
571 | if not self._generaldelta: | |
593 | self._sparserevlog = False |
|
572 | self._sparserevlog = False | |
594 |
|
573 | |||
595 | self._storedeltachains = True |
|
574 | self._storedeltachains = True | |
596 |
|
575 | |||
597 | devel_nodemap = ( |
|
576 | devel_nodemap = ( | |
598 | self._nodemap_file |
|
577 | self._nodemap_file | |
599 | and force_nodemap |
|
578 | and force_nodemap | |
600 | and parse_index_v1_nodemap is not None |
|
579 | and parse_index_v1_nodemap is not None | |
601 | ) |
|
580 | ) | |
602 |
|
581 | |||
603 | use_rust_index = False |
|
582 | use_rust_index = False | |
604 | if rustrevlog is not None: |
|
583 | if rustrevlog is not None: | |
605 | if self._nodemap_file is not None: |
|
584 | if self._nodemap_file is not None: | |
606 | use_rust_index = True |
|
585 | use_rust_index = True | |
607 | else: |
|
586 | else: | |
608 | use_rust_index = self.opener.options.get(b'rust.index') |
|
587 | use_rust_index = self.opener.options.get(b'rust.index') | |
609 |
|
588 | |||
610 | self._parse_index = parse_index_v1 |
|
589 | self._parse_index = parse_index_v1 | |
611 | if self._format_version == REVLOGV0: |
|
590 | if self._format_version == REVLOGV0: | |
612 | self._parse_index = revlogv0.parse_index_v0 |
|
591 | self._parse_index = revlogv0.parse_index_v0 | |
613 | elif self._format_version == REVLOGV2: |
|
592 | elif self._format_version == REVLOGV2: | |
614 | self._parse_index = parse_index_v2 |
|
593 | self._parse_index = parse_index_v2 | |
615 | elif self._format_version == CHANGELOGV2: |
|
594 | elif self._format_version == CHANGELOGV2: | |
616 | self._parse_index = parse_index_cl_v2 |
|
595 | self._parse_index = parse_index_cl_v2 | |
617 | elif devel_nodemap: |
|
596 | elif devel_nodemap: | |
618 | self._parse_index = parse_index_v1_nodemap |
|
597 | self._parse_index = parse_index_v1_nodemap | |
619 | elif use_rust_index: |
|
598 | elif use_rust_index: | |
620 | self._parse_index = parse_index_v1_mixed |
|
599 | self._parse_index = parse_index_v1_mixed | |
621 | try: |
|
600 | try: | |
622 | d = self._parse_index(index_data, self._inline) |
|
601 | d = self._parse_index(index_data, self._inline) | |
623 | index, _chunkcache = d |
|
602 | index, _chunkcache = d | |
624 | use_nodemap = ( |
|
603 | use_nodemap = ( | |
625 | not self._inline |
|
604 | not self._inline | |
626 | and self._nodemap_file is not None |
|
605 | and self._nodemap_file is not None | |
627 | and util.safehasattr(index, 'update_nodemap_data') |
|
606 | and util.safehasattr(index, 'update_nodemap_data') | |
628 | ) |
|
607 | ) | |
629 | if use_nodemap: |
|
608 | if use_nodemap: | |
630 | nodemap_data = nodemaputil.persisted_data(self) |
|
609 | nodemap_data = nodemaputil.persisted_data(self) | |
631 | if nodemap_data is not None: |
|
610 | if nodemap_data is not None: | |
632 | docket = nodemap_data[0] |
|
611 | docket = nodemap_data[0] | |
633 | if ( |
|
612 | if ( | |
634 | len(d[0]) > docket.tip_rev |
|
613 | len(d[0]) > docket.tip_rev | |
635 | and d[0][docket.tip_rev][7] == docket.tip_node |
|
614 | and d[0][docket.tip_rev][7] == docket.tip_node | |
636 | ): |
|
615 | ): | |
637 | # no changelog tampering |
|
616 | # no changelog tampering | |
638 | self._nodemap_docket = docket |
|
617 | self._nodemap_docket = docket | |
639 | index.update_nodemap_data(*nodemap_data) |
|
618 | index.update_nodemap_data(*nodemap_data) | |
640 | except (ValueError, IndexError): |
|
619 | except (ValueError, IndexError): | |
641 | raise error.RevlogError( |
|
620 | raise error.RevlogError( | |
642 | _(b"index %s is corrupted") % self.display_id |
|
621 | _(b"index %s is corrupted") % self.display_id | |
643 | ) |
|
622 | ) | |
644 | self.index, self._chunkcache = d |
|
623 | self.index, self._chunkcache = d | |
645 | if not self._chunkcache: |
|
624 | if not self._chunkcache: | |
646 | self._chunkclear() |
|
625 | self._chunkclear() | |
647 | # revnum -> (chain-length, sum-delta-length) |
|
626 | # revnum -> (chain-length, sum-delta-length) | |
648 | self._chaininfocache = util.lrucachedict(500) |
|
627 | self._chaininfocache = util.lrucachedict(500) | |
649 | # revlog header -> revlog compressor |
|
628 | # revlog header -> revlog compressor | |
650 | self._decompressors = {} |
|
629 | self._decompressors = {} | |
651 |
|
630 | |||
652 | @util.propertycache |
|
631 | @util.propertycache | |
653 | def revlog_kind(self): |
|
632 | def revlog_kind(self): | |
654 | return self.target[0] |
|
633 | return self.target[0] | |
655 |
|
634 | |||
656 | @util.propertycache |
|
635 | @util.propertycache | |
657 | def display_id(self): |
|
636 | def display_id(self): | |
658 | """The public facing "ID" of the revlog that we use in message""" |
|
637 | """The public facing "ID" of the revlog that we use in message""" | |
659 | # Maybe we should build a user facing representation of |
|
638 | # Maybe we should build a user facing representation of | |
660 | # revlog.target instead of using `self.radix` |
|
639 | # revlog.target instead of using `self.radix` | |
661 | return self.radix |
|
640 | return self.radix | |
662 |
|
641 | |||
663 | def _get_decompressor(self, t): |
|
642 | def _get_decompressor(self, t): | |
664 | try: |
|
643 | try: | |
665 | compressor = self._decompressors[t] |
|
644 | compressor = self._decompressors[t] | |
666 | except KeyError: |
|
645 | except KeyError: | |
667 | try: |
|
646 | try: | |
668 | engine = util.compengines.forrevlogheader(t) |
|
647 | engine = util.compengines.forrevlogheader(t) | |
669 | compressor = engine.revlogcompressor(self._compengineopts) |
|
648 | compressor = engine.revlogcompressor(self._compengineopts) | |
670 | self._decompressors[t] = compressor |
|
649 | self._decompressors[t] = compressor | |
671 | except KeyError: |
|
650 | except KeyError: | |
672 | raise error.RevlogError( |
|
651 | raise error.RevlogError( | |
673 | _(b'unknown compression type %s') % binascii.hexlify(t) |
|
652 | _(b'unknown compression type %s') % binascii.hexlify(t) | |
674 | ) |
|
653 | ) | |
675 | return compressor |
|
654 | return compressor | |
676 |
|
655 | |||
677 | @util.propertycache |
|
656 | @util.propertycache | |
678 | def _compressor(self): |
|
657 | def _compressor(self): | |
679 | engine = util.compengines[self._compengine] |
|
658 | engine = util.compengines[self._compengine] | |
680 | return engine.revlogcompressor(self._compengineopts) |
|
659 | return engine.revlogcompressor(self._compengineopts) | |
681 |
|
660 | |||
682 | @util.propertycache |
|
661 | @util.propertycache | |
683 | def _decompressor(self): |
|
662 | def _decompressor(self): | |
684 | """the default decompressor""" |
|
663 | """the default decompressor""" | |
685 | if self._docket is None: |
|
664 | if self._docket is None: | |
686 | return None |
|
665 | return None | |
687 | t = self._docket.default_compression_header |
|
666 | t = self._docket.default_compression_header | |
688 | c = self._get_decompressor(t) |
|
667 | c = self._get_decompressor(t) | |
689 | return c.decompress |
|
668 | return c.decompress | |
690 |
|
669 | |||
691 | def _indexfp(self): |
|
670 | def _indexfp(self): | |
692 | """file object for the revlog's index file""" |
|
671 | """file object for the revlog's index file""" | |
693 | return self.opener(self._indexfile, mode=b"r") |
|
672 | return self.opener(self._indexfile, mode=b"r") | |
694 |
|
673 | |||
695 | def __index_write_fp(self): |
|
674 | def __index_write_fp(self): | |
696 | # You should not use this directly and use `_writing` instead |
|
675 | # You should not use this directly and use `_writing` instead | |
697 | try: |
|
676 | try: | |
698 | f = self.opener( |
|
677 | f = self.opener( | |
699 | self._indexfile, mode=b"r+", checkambig=self._checkambig |
|
678 | self._indexfile, mode=b"r+", checkambig=self._checkambig | |
700 | ) |
|
679 | ) | |
701 | if self._docket is None: |
|
680 | if self._docket is None: | |
702 | f.seek(0, os.SEEK_END) |
|
681 | f.seek(0, os.SEEK_END) | |
703 | else: |
|
682 | else: | |
704 | f.seek(self._docket.index_end, os.SEEK_SET) |
|
683 | f.seek(self._docket.index_end, os.SEEK_SET) | |
705 | return f |
|
684 | return f | |
706 | except IOError as inst: |
|
685 | except IOError as inst: | |
707 | if inst.errno != errno.ENOENT: |
|
686 | if inst.errno != errno.ENOENT: | |
708 | raise |
|
687 | raise | |
709 | return self.opener( |
|
688 | return self.opener( | |
710 | self._indexfile, mode=b"w+", checkambig=self._checkambig |
|
689 | self._indexfile, mode=b"w+", checkambig=self._checkambig | |
711 | ) |
|
690 | ) | |
712 |
|
691 | |||
713 | def __index_new_fp(self): |
|
692 | def __index_new_fp(self): | |
714 | # You should not use this unless you are upgrading from inline revlog |
|
693 | # You should not use this unless you are upgrading from inline revlog | |
715 | return self.opener( |
|
694 | return self.opener( | |
716 | self._indexfile, |
|
695 | self._indexfile, | |
717 | mode=b"w", |
|
696 | mode=b"w", | |
718 | checkambig=self._checkambig, |
|
697 | checkambig=self._checkambig, | |
719 | atomictemp=True, |
|
698 | atomictemp=True, | |
720 | ) |
|
699 | ) | |
721 |
|
700 | |||
722 | def _datafp(self, mode=b'r'): |
|
701 | def _datafp(self, mode=b'r'): | |
723 | """file object for the revlog's data file""" |
|
702 | """file object for the revlog's data file""" | |
724 | return self.opener(self._datafile, mode=mode) |
|
703 | return self.opener(self._datafile, mode=mode) | |
725 |
|
704 | |||
726 | @contextlib.contextmanager |
|
705 | @contextlib.contextmanager | |
727 | def _datareadfp(self, existingfp=None): |
|
706 | def _datareadfp(self, existingfp=None): | |
728 | """file object suitable to read data""" |
|
707 | """file object suitable to read data""" | |
729 | # Use explicit file handle, if given. |
|
708 | # Use explicit file handle, if given. | |
730 | if existingfp is not None: |
|
709 | if existingfp is not None: | |
731 | yield existingfp |
|
710 | yield existingfp | |
732 |
|
711 | |||
733 | # Use a file handle being actively used for writes, if available. |
|
712 | # Use a file handle being actively used for writes, if available. | |
734 | # There is some danger to doing this because reads will seek the |
|
713 | # There is some danger to doing this because reads will seek the | |
735 | # file. However, _writeentry() performs a SEEK_END before all writes, |
|
714 | # file. However, _writeentry() performs a SEEK_END before all writes, | |
736 | # so we should be safe. |
|
715 | # so we should be safe. | |
737 | elif self._writinghandles: |
|
716 | elif self._writinghandles: | |
738 | if self._inline: |
|
717 | if self._inline: | |
739 | yield self._writinghandles[0] |
|
718 | yield self._writinghandles[0] | |
740 | else: |
|
719 | else: | |
741 | yield self._writinghandles[1] |
|
720 | yield self._writinghandles[1] | |
742 |
|
721 | |||
743 | # Otherwise open a new file handle. |
|
722 | # Otherwise open a new file handle. | |
744 | else: |
|
723 | else: | |
745 | if self._inline: |
|
724 | if self._inline: | |
746 | func = self._indexfp |
|
725 | func = self._indexfp | |
747 | else: |
|
726 | else: | |
748 | func = self._datafp |
|
727 | func = self._datafp | |
749 | with func() as fp: |
|
728 | with func() as fp: | |
750 | yield fp |
|
729 | yield fp | |
751 |
|
730 | |||
752 | @contextlib.contextmanager |
|
731 | @contextlib.contextmanager | |
753 | def _sidedatareadfp(self): |
|
732 | def _sidedatareadfp(self): | |
754 | """file object suitable to read sidedata""" |
|
733 | """file object suitable to read sidedata""" | |
755 | if self._writinghandles: |
|
734 | if self._writinghandles: | |
756 | yield self._writinghandles[2] |
|
735 | yield self._writinghandles[2] | |
757 | else: |
|
736 | else: | |
758 | with self.opener(self._sidedatafile) as fp: |
|
737 | with self.opener(self._sidedatafile) as fp: | |
759 | yield fp |
|
738 | yield fp | |
760 |
|
739 | |||
761 | def tiprev(self): |
|
740 | def tiprev(self): | |
762 | return len(self.index) - 1 |
|
741 | return len(self.index) - 1 | |
763 |
|
742 | |||
764 | def tip(self): |
|
743 | def tip(self): | |
765 | return self.node(self.tiprev()) |
|
744 | return self.node(self.tiprev()) | |
766 |
|
745 | |||
767 | def __contains__(self, rev): |
|
746 | def __contains__(self, rev): | |
768 | return 0 <= rev < len(self) |
|
747 | return 0 <= rev < len(self) | |
769 |
|
748 | |||
770 | def __len__(self): |
|
749 | def __len__(self): | |
771 | return len(self.index) |
|
750 | return len(self.index) | |
772 |
|
751 | |||
773 | def __iter__(self): |
|
752 | def __iter__(self): | |
774 | return iter(pycompat.xrange(len(self))) |
|
753 | return iter(pycompat.xrange(len(self))) | |
775 |
|
754 | |||
776 | def revs(self, start=0, stop=None): |
|
755 | def revs(self, start=0, stop=None): | |
777 | """iterate over all rev in this revlog (from start to stop)""" |
|
756 | """iterate over all rev in this revlog (from start to stop)""" | |
778 | return storageutil.iterrevs(len(self), start=start, stop=stop) |
|
757 | return storageutil.iterrevs(len(self), start=start, stop=stop) | |
779 |
|
758 | |||
780 | @property |
|
759 | @property | |
781 | def nodemap(self): |
|
760 | def nodemap(self): | |
782 | msg = ( |
|
761 | msg = ( | |
783 | b"revlog.nodemap is deprecated, " |
|
762 | b"revlog.nodemap is deprecated, " | |
784 | b"use revlog.index.[has_node|rev|get_rev]" |
|
763 | b"use revlog.index.[has_node|rev|get_rev]" | |
785 | ) |
|
764 | ) | |
786 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
765 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) | |
787 | return self.index.nodemap |
|
766 | return self.index.nodemap | |
788 |
|
767 | |||
789 | @property |
|
768 | @property | |
790 | def _nodecache(self): |
|
769 | def _nodecache(self): | |
791 | msg = b"revlog._nodecache is deprecated, use revlog.index.nodemap" |
|
770 | msg = b"revlog._nodecache is deprecated, use revlog.index.nodemap" | |
792 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
771 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) | |
793 | return self.index.nodemap |
|
772 | return self.index.nodemap | |
794 |
|
773 | |||
795 | def hasnode(self, node): |
|
774 | def hasnode(self, node): | |
796 | try: |
|
775 | try: | |
797 | self.rev(node) |
|
776 | self.rev(node) | |
798 | return True |
|
777 | return True | |
799 | except KeyError: |
|
778 | except KeyError: | |
800 | return False |
|
779 | return False | |
801 |
|
780 | |||
802 | def candelta(self, baserev, rev): |
|
781 | def candelta(self, baserev, rev): | |
803 | """whether two revisions (baserev, rev) can be delta-ed or not""" |
|
782 | """whether two revisions (baserev, rev) can be delta-ed or not""" | |
804 | # Disable delta if either rev requires a content-changing flag |
|
783 | # Disable delta if either rev requires a content-changing flag | |
805 | # processor (ex. LFS). This is because such flag processor can alter |
|
784 | # processor (ex. LFS). This is because such flag processor can alter | |
806 | # the rawtext content that the delta will be based on, and two clients |
|
785 | # the rawtext content that the delta will be based on, and two clients | |
807 | # could have a same revlog node with different flags (i.e. different |
|
786 | # could have a same revlog node with different flags (i.e. different | |
808 | # rawtext contents) and the delta could be incompatible. |
|
787 | # rawtext contents) and the delta could be incompatible. | |
809 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
788 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( | |
810 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS |
|
789 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS | |
811 | ): |
|
790 | ): | |
812 | return False |
|
791 | return False | |
813 | return True |
|
792 | return True | |
814 |
|
793 | |||
815 | def update_caches(self, transaction): |
|
794 | def update_caches(self, transaction): | |
816 | if self._nodemap_file is not None: |
|
795 | if self._nodemap_file is not None: | |
817 | if transaction is None: |
|
796 | if transaction is None: | |
818 | nodemaputil.update_persistent_nodemap(self) |
|
797 | nodemaputil.update_persistent_nodemap(self) | |
819 | else: |
|
798 | else: | |
820 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
799 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
821 |
|
800 | |||
822 | def clearcaches(self): |
|
801 | def clearcaches(self): | |
823 | self._revisioncache = None |
|
802 | self._revisioncache = None | |
824 | self._chainbasecache.clear() |
|
803 | self._chainbasecache.clear() | |
825 | self._chunkcache = (0, b'') |
|
804 | self._chunkcache = (0, b'') | |
826 | self._pcache = {} |
|
805 | self._pcache = {} | |
827 | self._nodemap_docket = None |
|
806 | self._nodemap_docket = None | |
828 | self.index.clearcaches() |
|
807 | self.index.clearcaches() | |
829 | # The python code is the one responsible for validating the docket, we |
|
808 | # The python code is the one responsible for validating the docket, we | |
830 | # end up having to refresh it here. |
|
809 | # end up having to refresh it here. | |
831 | use_nodemap = ( |
|
810 | use_nodemap = ( | |
832 | not self._inline |
|
811 | not self._inline | |
833 | and self._nodemap_file is not None |
|
812 | and self._nodemap_file is not None | |
834 | and util.safehasattr(self.index, 'update_nodemap_data') |
|
813 | and util.safehasattr(self.index, 'update_nodemap_data') | |
835 | ) |
|
814 | ) | |
836 | if use_nodemap: |
|
815 | if use_nodemap: | |
837 | nodemap_data = nodemaputil.persisted_data(self) |
|
816 | nodemap_data = nodemaputil.persisted_data(self) | |
838 | if nodemap_data is not None: |
|
817 | if nodemap_data is not None: | |
839 | self._nodemap_docket = nodemap_data[0] |
|
818 | self._nodemap_docket = nodemap_data[0] | |
840 | self.index.update_nodemap_data(*nodemap_data) |
|
819 | self.index.update_nodemap_data(*nodemap_data) | |
841 |
|
820 | |||
842 | def rev(self, node): |
|
821 | def rev(self, node): | |
843 | try: |
|
822 | try: | |
844 | return self.index.rev(node) |
|
823 | return self.index.rev(node) | |
845 | except TypeError: |
|
824 | except TypeError: | |
846 | raise |
|
825 | raise | |
847 | except error.RevlogError: |
|
826 | except error.RevlogError: | |
848 | # parsers.c radix tree lookup failed |
|
827 | # parsers.c radix tree lookup failed | |
849 | if ( |
|
828 | if ( | |
850 | node == self.nodeconstants.wdirid |
|
829 | node == self.nodeconstants.wdirid | |
851 | or node in self.nodeconstants.wdirfilenodeids |
|
830 | or node in self.nodeconstants.wdirfilenodeids | |
852 | ): |
|
831 | ): | |
853 | raise error.WdirUnsupported |
|
832 | raise error.WdirUnsupported | |
854 | raise error.LookupError(node, self.display_id, _(b'no node')) |
|
833 | raise error.LookupError(node, self.display_id, _(b'no node')) | |
855 |
|
834 | |||
856 | # Accessors for index entries. |
|
835 | # Accessors for index entries. | |
857 |
|
836 | |||
858 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes |
|
837 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes | |
859 | # are flags. |
|
838 | # are flags. | |
860 | def start(self, rev): |
|
839 | def start(self, rev): | |
861 | return int(self.index[rev][0] >> 16) |
|
840 | return int(self.index[rev][0] >> 16) | |
862 |
|
841 | |||
863 | def sidedata_cut_off(self, rev): |
|
842 | def sidedata_cut_off(self, rev): | |
864 | sd_cut_off = self.index[rev][8] |
|
843 | sd_cut_off = self.index[rev][8] | |
865 | if sd_cut_off != 0: |
|
844 | if sd_cut_off != 0: | |
866 | return sd_cut_off |
|
845 | return sd_cut_off | |
867 | # This is some annoying dance, because entries without sidedata |
|
846 | # This is some annoying dance, because entries without sidedata | |
868 | # currently use 0 as their ofsset. (instead of previous-offset + |
|
847 | # currently use 0 as their ofsset. (instead of previous-offset + | |
869 | # previous-size) |
|
848 | # previous-size) | |
870 | # |
|
849 | # | |
871 | # We should reconsider this sidedata β 0 sidata_offset policy. |
|
850 | # We should reconsider this sidedata β 0 sidata_offset policy. | |
872 | # In the meantime, we need this. |
|
851 | # In the meantime, we need this. | |
873 | while 0 <= rev: |
|
852 | while 0 <= rev: | |
874 | e = self.index[rev] |
|
853 | e = self.index[rev] | |
875 | if e[9] != 0: |
|
854 | if e[9] != 0: | |
876 | return e[8] + e[9] |
|
855 | return e[8] + e[9] | |
877 | rev -= 1 |
|
856 | rev -= 1 | |
878 | return 0 |
|
857 | return 0 | |
879 |
|
858 | |||
880 | def flags(self, rev): |
|
859 | def flags(self, rev): | |
881 | return self.index[rev][0] & 0xFFFF |
|
860 | return self.index[rev][0] & 0xFFFF | |
882 |
|
861 | |||
883 | def length(self, rev): |
|
862 | def length(self, rev): | |
884 | return self.index[rev][1] |
|
863 | return self.index[rev][1] | |
885 |
|
864 | |||
886 | def sidedata_length(self, rev): |
|
865 | def sidedata_length(self, rev): | |
887 | if not self.hassidedata: |
|
866 | if not self.hassidedata: | |
888 | return 0 |
|
867 | return 0 | |
889 | return self.index[rev][9] |
|
868 | return self.index[rev][9] | |
890 |
|
869 | |||
891 | def rawsize(self, rev): |
|
870 | def rawsize(self, rev): | |
892 | """return the length of the uncompressed text for a given revision""" |
|
871 | """return the length of the uncompressed text for a given revision""" | |
893 | l = self.index[rev][2] |
|
872 | l = self.index[rev][2] | |
894 | if l >= 0: |
|
873 | if l >= 0: | |
895 | return l |
|
874 | return l | |
896 |
|
875 | |||
897 | t = self.rawdata(rev) |
|
876 | t = self.rawdata(rev) | |
898 | return len(t) |
|
877 | return len(t) | |
899 |
|
878 | |||
900 | def size(self, rev): |
|
879 | def size(self, rev): | |
901 | """length of non-raw text (processed by a "read" flag processor)""" |
|
880 | """length of non-raw text (processed by a "read" flag processor)""" | |
902 | # fast path: if no "read" flag processor could change the content, |
|
881 | # fast path: if no "read" flag processor could change the content, | |
903 | # size is rawsize. note: ELLIPSIS is known to not change the content. |
|
882 | # size is rawsize. note: ELLIPSIS is known to not change the content. | |
904 | flags = self.flags(rev) |
|
883 | flags = self.flags(rev) | |
905 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: |
|
884 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: | |
906 | return self.rawsize(rev) |
|
885 | return self.rawsize(rev) | |
907 |
|
886 | |||
908 | return len(self.revision(rev, raw=False)) |
|
887 | return len(self.revision(rev, raw=False)) | |
909 |
|
888 | |||
910 | def chainbase(self, rev): |
|
889 | def chainbase(self, rev): | |
911 | base = self._chainbasecache.get(rev) |
|
890 | base = self._chainbasecache.get(rev) | |
912 | if base is not None: |
|
891 | if base is not None: | |
913 | return base |
|
892 | return base | |
914 |
|
893 | |||
915 | index = self.index |
|
894 | index = self.index | |
916 | iterrev = rev |
|
895 | iterrev = rev | |
917 | base = index[iterrev][3] |
|
896 | base = index[iterrev][3] | |
918 | while base != iterrev: |
|
897 | while base != iterrev: | |
919 | iterrev = base |
|
898 | iterrev = base | |
920 | base = index[iterrev][3] |
|
899 | base = index[iterrev][3] | |
921 |
|
900 | |||
922 | self._chainbasecache[rev] = base |
|
901 | self._chainbasecache[rev] = base | |
923 | return base |
|
902 | return base | |
924 |
|
903 | |||
925 | def linkrev(self, rev): |
|
904 | def linkrev(self, rev): | |
926 | return self.index[rev][4] |
|
905 | return self.index[rev][4] | |
927 |
|
906 | |||
928 | def parentrevs(self, rev): |
|
907 | def parentrevs(self, rev): | |
929 | try: |
|
908 | try: | |
930 | entry = self.index[rev] |
|
909 | entry = self.index[rev] | |
931 | except IndexError: |
|
910 | except IndexError: | |
932 | if rev == wdirrev: |
|
911 | if rev == wdirrev: | |
933 | raise error.WdirUnsupported |
|
912 | raise error.WdirUnsupported | |
934 | raise |
|
913 | raise | |
935 | if entry[5] == nullrev: |
|
914 | if entry[5] == nullrev: | |
936 | return entry[6], entry[5] |
|
915 | return entry[6], entry[5] | |
937 | else: |
|
916 | else: | |
938 | return entry[5], entry[6] |
|
917 | return entry[5], entry[6] | |
939 |
|
918 | |||
940 | # fast parentrevs(rev) where rev isn't filtered |
|
919 | # fast parentrevs(rev) where rev isn't filtered | |
941 | _uncheckedparentrevs = parentrevs |
|
920 | _uncheckedparentrevs = parentrevs | |
942 |
|
921 | |||
943 | def node(self, rev): |
|
922 | def node(self, rev): | |
944 | try: |
|
923 | try: | |
945 | return self.index[rev][7] |
|
924 | return self.index[rev][7] | |
946 | except IndexError: |
|
925 | except IndexError: | |
947 | if rev == wdirrev: |
|
926 | if rev == wdirrev: | |
948 | raise error.WdirUnsupported |
|
927 | raise error.WdirUnsupported | |
949 | raise |
|
928 | raise | |
950 |
|
929 | |||
951 | # Derived from index values. |
|
930 | # Derived from index values. | |
952 |
|
931 | |||
953 | def end(self, rev): |
|
932 | def end(self, rev): | |
954 | return self.start(rev) + self.length(rev) |
|
933 | return self.start(rev) + self.length(rev) | |
955 |
|
934 | |||
956 | def parents(self, node): |
|
935 | def parents(self, node): | |
957 | i = self.index |
|
936 | i = self.index | |
958 | d = i[self.rev(node)] |
|
937 | d = i[self.rev(node)] | |
959 | # inline node() to avoid function call overhead |
|
938 | # inline node() to avoid function call overhead | |
960 | if d[5] == self.nullid: |
|
939 | if d[5] == self.nullid: | |
961 | return i[d[6]][7], i[d[5]][7] |
|
940 | return i[d[6]][7], i[d[5]][7] | |
962 | else: |
|
941 | else: | |
963 | return i[d[5]][7], i[d[6]][7] |
|
942 | return i[d[5]][7], i[d[6]][7] | |
964 |
|
943 | |||
965 | def chainlen(self, rev): |
|
944 | def chainlen(self, rev): | |
966 | return self._chaininfo(rev)[0] |
|
945 | return self._chaininfo(rev)[0] | |
967 |
|
946 | |||
968 | def _chaininfo(self, rev): |
|
947 | def _chaininfo(self, rev): | |
969 | chaininfocache = self._chaininfocache |
|
948 | chaininfocache = self._chaininfocache | |
970 | if rev in chaininfocache: |
|
949 | if rev in chaininfocache: | |
971 | return chaininfocache[rev] |
|
950 | return chaininfocache[rev] | |
972 | index = self.index |
|
951 | index = self.index | |
973 | generaldelta = self._generaldelta |
|
952 | generaldelta = self._generaldelta | |
974 | iterrev = rev |
|
953 | iterrev = rev | |
975 | e = index[iterrev] |
|
954 | e = index[iterrev] | |
976 | clen = 0 |
|
955 | clen = 0 | |
977 | compresseddeltalen = 0 |
|
956 | compresseddeltalen = 0 | |
978 | while iterrev != e[3]: |
|
957 | while iterrev != e[3]: | |
979 | clen += 1 |
|
958 | clen += 1 | |
980 | compresseddeltalen += e[1] |
|
959 | compresseddeltalen += e[1] | |
981 | if generaldelta: |
|
960 | if generaldelta: | |
982 | iterrev = e[3] |
|
961 | iterrev = e[3] | |
983 | else: |
|
962 | else: | |
984 | iterrev -= 1 |
|
963 | iterrev -= 1 | |
985 | if iterrev in chaininfocache: |
|
964 | if iterrev in chaininfocache: | |
986 | t = chaininfocache[iterrev] |
|
965 | t = chaininfocache[iterrev] | |
987 | clen += t[0] |
|
966 | clen += t[0] | |
988 | compresseddeltalen += t[1] |
|
967 | compresseddeltalen += t[1] | |
989 | break |
|
968 | break | |
990 | e = index[iterrev] |
|
969 | e = index[iterrev] | |
991 | else: |
|
970 | else: | |
992 | # Add text length of base since decompressing that also takes |
|
971 | # Add text length of base since decompressing that also takes | |
993 | # work. For cache hits the length is already included. |
|
972 | # work. For cache hits the length is already included. | |
994 | compresseddeltalen += e[1] |
|
973 | compresseddeltalen += e[1] | |
995 | r = (clen, compresseddeltalen) |
|
974 | r = (clen, compresseddeltalen) | |
996 | chaininfocache[rev] = r |
|
975 | chaininfocache[rev] = r | |
997 | return r |
|
976 | return r | |
998 |
|
977 | |||
999 | def _deltachain(self, rev, stoprev=None): |
|
978 | def _deltachain(self, rev, stoprev=None): | |
1000 | """Obtain the delta chain for a revision. |
|
979 | """Obtain the delta chain for a revision. | |
1001 |
|
980 | |||
1002 | ``stoprev`` specifies a revision to stop at. If not specified, we |
|
981 | ``stoprev`` specifies a revision to stop at. If not specified, we | |
1003 | stop at the base of the chain. |
|
982 | stop at the base of the chain. | |
1004 |
|
983 | |||
1005 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of |
|
984 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of | |
1006 | revs in ascending order and ``stopped`` is a bool indicating whether |
|
985 | revs in ascending order and ``stopped`` is a bool indicating whether | |
1007 | ``stoprev`` was hit. |
|
986 | ``stoprev`` was hit. | |
1008 | """ |
|
987 | """ | |
1009 | # Try C implementation. |
|
988 | # Try C implementation. | |
1010 | try: |
|
989 | try: | |
1011 | return self.index.deltachain(rev, stoprev, self._generaldelta) |
|
990 | return self.index.deltachain(rev, stoprev, self._generaldelta) | |
1012 | except AttributeError: |
|
991 | except AttributeError: | |
1013 | pass |
|
992 | pass | |
1014 |
|
993 | |||
1015 | chain = [] |
|
994 | chain = [] | |
1016 |
|
995 | |||
1017 | # Alias to prevent attribute lookup in tight loop. |
|
996 | # Alias to prevent attribute lookup in tight loop. | |
1018 | index = self.index |
|
997 | index = self.index | |
1019 | generaldelta = self._generaldelta |
|
998 | generaldelta = self._generaldelta | |
1020 |
|
999 | |||
1021 | iterrev = rev |
|
1000 | iterrev = rev | |
1022 | e = index[iterrev] |
|
1001 | e = index[iterrev] | |
1023 | while iterrev != e[3] and iterrev != stoprev: |
|
1002 | while iterrev != e[3] and iterrev != stoprev: | |
1024 | chain.append(iterrev) |
|
1003 | chain.append(iterrev) | |
1025 | if generaldelta: |
|
1004 | if generaldelta: | |
1026 | iterrev = e[3] |
|
1005 | iterrev = e[3] | |
1027 | else: |
|
1006 | else: | |
1028 | iterrev -= 1 |
|
1007 | iterrev -= 1 | |
1029 | e = index[iterrev] |
|
1008 | e = index[iterrev] | |
1030 |
|
1009 | |||
1031 | if iterrev == stoprev: |
|
1010 | if iterrev == stoprev: | |
1032 | stopped = True |
|
1011 | stopped = True | |
1033 | else: |
|
1012 | else: | |
1034 | chain.append(iterrev) |
|
1013 | chain.append(iterrev) | |
1035 | stopped = False |
|
1014 | stopped = False | |
1036 |
|
1015 | |||
1037 | chain.reverse() |
|
1016 | chain.reverse() | |
1038 | return chain, stopped |
|
1017 | return chain, stopped | |
1039 |
|
1018 | |||
1040 | def ancestors(self, revs, stoprev=0, inclusive=False): |
|
1019 | def ancestors(self, revs, stoprev=0, inclusive=False): | |
1041 | """Generate the ancestors of 'revs' in reverse revision order. |
|
1020 | """Generate the ancestors of 'revs' in reverse revision order. | |
1042 | Does not generate revs lower than stoprev. |
|
1021 | Does not generate revs lower than stoprev. | |
1043 |
|
1022 | |||
1044 | See the documentation for ancestor.lazyancestors for more details.""" |
|
1023 | See the documentation for ancestor.lazyancestors for more details.""" | |
1045 |
|
1024 | |||
1046 | # first, make sure start revisions aren't filtered |
|
1025 | # first, make sure start revisions aren't filtered | |
1047 | revs = list(revs) |
|
1026 | revs = list(revs) | |
1048 | checkrev = self.node |
|
1027 | checkrev = self.node | |
1049 | for r in revs: |
|
1028 | for r in revs: | |
1050 | checkrev(r) |
|
1029 | checkrev(r) | |
1051 | # and we're sure ancestors aren't filtered as well |
|
1030 | # and we're sure ancestors aren't filtered as well | |
1052 |
|
1031 | |||
1053 | if rustancestor is not None and self.index.rust_ext_compat: |
|
1032 | if rustancestor is not None and self.index.rust_ext_compat: | |
1054 | lazyancestors = rustancestor.LazyAncestors |
|
1033 | lazyancestors = rustancestor.LazyAncestors | |
1055 | arg = self.index |
|
1034 | arg = self.index | |
1056 | else: |
|
1035 | else: | |
1057 | lazyancestors = ancestor.lazyancestors |
|
1036 | lazyancestors = ancestor.lazyancestors | |
1058 | arg = self._uncheckedparentrevs |
|
1037 | arg = self._uncheckedparentrevs | |
1059 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) |
|
1038 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) | |
1060 |
|
1039 | |||
1061 | def descendants(self, revs): |
|
1040 | def descendants(self, revs): | |
1062 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
1041 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) | |
1063 |
|
1042 | |||
1064 | def findcommonmissing(self, common=None, heads=None): |
|
1043 | def findcommonmissing(self, common=None, heads=None): | |
1065 | """Return a tuple of the ancestors of common and the ancestors of heads |
|
1044 | """Return a tuple of the ancestors of common and the ancestors of heads | |
1066 | that are not ancestors of common. In revset terminology, we return the |
|
1045 | that are not ancestors of common. In revset terminology, we return the | |
1067 | tuple: |
|
1046 | tuple: | |
1068 |
|
1047 | |||
1069 | ::common, (::heads) - (::common) |
|
1048 | ::common, (::heads) - (::common) | |
1070 |
|
1049 | |||
1071 | The list is sorted by revision number, meaning it is |
|
1050 | The list is sorted by revision number, meaning it is | |
1072 | topologically sorted. |
|
1051 | topologically sorted. | |
1073 |
|
1052 | |||
1074 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1053 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1075 | not supplied, uses all of the revlog's heads. If common is not |
|
1054 | not supplied, uses all of the revlog's heads. If common is not | |
1076 | supplied, uses nullid.""" |
|
1055 | supplied, uses nullid.""" | |
1077 | if common is None: |
|
1056 | if common is None: | |
1078 | common = [self.nullid] |
|
1057 | common = [self.nullid] | |
1079 | if heads is None: |
|
1058 | if heads is None: | |
1080 | heads = self.heads() |
|
1059 | heads = self.heads() | |
1081 |
|
1060 | |||
1082 | common = [self.rev(n) for n in common] |
|
1061 | common = [self.rev(n) for n in common] | |
1083 | heads = [self.rev(n) for n in heads] |
|
1062 | heads = [self.rev(n) for n in heads] | |
1084 |
|
1063 | |||
1085 | # we want the ancestors, but inclusive |
|
1064 | # we want the ancestors, but inclusive | |
1086 | class lazyset(object): |
|
1065 | class lazyset(object): | |
1087 | def __init__(self, lazyvalues): |
|
1066 | def __init__(self, lazyvalues): | |
1088 | self.addedvalues = set() |
|
1067 | self.addedvalues = set() | |
1089 | self.lazyvalues = lazyvalues |
|
1068 | self.lazyvalues = lazyvalues | |
1090 |
|
1069 | |||
1091 | def __contains__(self, value): |
|
1070 | def __contains__(self, value): | |
1092 | return value in self.addedvalues or value in self.lazyvalues |
|
1071 | return value in self.addedvalues or value in self.lazyvalues | |
1093 |
|
1072 | |||
1094 | def __iter__(self): |
|
1073 | def __iter__(self): | |
1095 | added = self.addedvalues |
|
1074 | added = self.addedvalues | |
1096 | for r in added: |
|
1075 | for r in added: | |
1097 | yield r |
|
1076 | yield r | |
1098 | for r in self.lazyvalues: |
|
1077 | for r in self.lazyvalues: | |
1099 | if not r in added: |
|
1078 | if not r in added: | |
1100 | yield r |
|
1079 | yield r | |
1101 |
|
1080 | |||
1102 | def add(self, value): |
|
1081 | def add(self, value): | |
1103 | self.addedvalues.add(value) |
|
1082 | self.addedvalues.add(value) | |
1104 |
|
1083 | |||
1105 | def update(self, values): |
|
1084 | def update(self, values): | |
1106 | self.addedvalues.update(values) |
|
1085 | self.addedvalues.update(values) | |
1107 |
|
1086 | |||
1108 | has = lazyset(self.ancestors(common)) |
|
1087 | has = lazyset(self.ancestors(common)) | |
1109 | has.add(nullrev) |
|
1088 | has.add(nullrev) | |
1110 | has.update(common) |
|
1089 | has.update(common) | |
1111 |
|
1090 | |||
1112 | # take all ancestors from heads that aren't in has |
|
1091 | # take all ancestors from heads that aren't in has | |
1113 | missing = set() |
|
1092 | missing = set() | |
1114 | visit = collections.deque(r for r in heads if r not in has) |
|
1093 | visit = collections.deque(r for r in heads if r not in has) | |
1115 | while visit: |
|
1094 | while visit: | |
1116 | r = visit.popleft() |
|
1095 | r = visit.popleft() | |
1117 | if r in missing: |
|
1096 | if r in missing: | |
1118 | continue |
|
1097 | continue | |
1119 | else: |
|
1098 | else: | |
1120 | missing.add(r) |
|
1099 | missing.add(r) | |
1121 | for p in self.parentrevs(r): |
|
1100 | for p in self.parentrevs(r): | |
1122 | if p not in has: |
|
1101 | if p not in has: | |
1123 | visit.append(p) |
|
1102 | visit.append(p) | |
1124 | missing = list(missing) |
|
1103 | missing = list(missing) | |
1125 | missing.sort() |
|
1104 | missing.sort() | |
1126 | return has, [self.node(miss) for miss in missing] |
|
1105 | return has, [self.node(miss) for miss in missing] | |
1127 |
|
1106 | |||
1128 | def incrementalmissingrevs(self, common=None): |
|
1107 | def incrementalmissingrevs(self, common=None): | |
1129 | """Return an object that can be used to incrementally compute the |
|
1108 | """Return an object that can be used to incrementally compute the | |
1130 | revision numbers of the ancestors of arbitrary sets that are not |
|
1109 | revision numbers of the ancestors of arbitrary sets that are not | |
1131 | ancestors of common. This is an ancestor.incrementalmissingancestors |
|
1110 | ancestors of common. This is an ancestor.incrementalmissingancestors | |
1132 | object. |
|
1111 | object. | |
1133 |
|
1112 | |||
1134 | 'common' is a list of revision numbers. If common is not supplied, uses |
|
1113 | 'common' is a list of revision numbers. If common is not supplied, uses | |
1135 | nullrev. |
|
1114 | nullrev. | |
1136 | """ |
|
1115 | """ | |
1137 | if common is None: |
|
1116 | if common is None: | |
1138 | common = [nullrev] |
|
1117 | common = [nullrev] | |
1139 |
|
1118 | |||
1140 | if rustancestor is not None and self.index.rust_ext_compat: |
|
1119 | if rustancestor is not None and self.index.rust_ext_compat: | |
1141 | return rustancestor.MissingAncestors(self.index, common) |
|
1120 | return rustancestor.MissingAncestors(self.index, common) | |
1142 | return ancestor.incrementalmissingancestors(self.parentrevs, common) |
|
1121 | return ancestor.incrementalmissingancestors(self.parentrevs, common) | |
1143 |
|
1122 | |||
1144 | def findmissingrevs(self, common=None, heads=None): |
|
1123 | def findmissingrevs(self, common=None, heads=None): | |
1145 | """Return the revision numbers of the ancestors of heads that |
|
1124 | """Return the revision numbers of the ancestors of heads that | |
1146 | are not ancestors of common. |
|
1125 | are not ancestors of common. | |
1147 |
|
1126 | |||
1148 | More specifically, return a list of revision numbers corresponding to |
|
1127 | More specifically, return a list of revision numbers corresponding to | |
1149 | nodes N such that every N satisfies the following constraints: |
|
1128 | nodes N such that every N satisfies the following constraints: | |
1150 |
|
1129 | |||
1151 | 1. N is an ancestor of some node in 'heads' |
|
1130 | 1. N is an ancestor of some node in 'heads' | |
1152 | 2. N is not an ancestor of any node in 'common' |
|
1131 | 2. N is not an ancestor of any node in 'common' | |
1153 |
|
1132 | |||
1154 | The list is sorted by revision number, meaning it is |
|
1133 | The list is sorted by revision number, meaning it is | |
1155 | topologically sorted. |
|
1134 | topologically sorted. | |
1156 |
|
1135 | |||
1157 | 'heads' and 'common' are both lists of revision numbers. If heads is |
|
1136 | 'heads' and 'common' are both lists of revision numbers. If heads is | |
1158 | not supplied, uses all of the revlog's heads. If common is not |
|
1137 | not supplied, uses all of the revlog's heads. If common is not | |
1159 | supplied, uses nullid.""" |
|
1138 | supplied, uses nullid.""" | |
1160 | if common is None: |
|
1139 | if common is None: | |
1161 | common = [nullrev] |
|
1140 | common = [nullrev] | |
1162 | if heads is None: |
|
1141 | if heads is None: | |
1163 | heads = self.headrevs() |
|
1142 | heads = self.headrevs() | |
1164 |
|
1143 | |||
1165 | inc = self.incrementalmissingrevs(common=common) |
|
1144 | inc = self.incrementalmissingrevs(common=common) | |
1166 | return inc.missingancestors(heads) |
|
1145 | return inc.missingancestors(heads) | |
1167 |
|
1146 | |||
1168 | def findmissing(self, common=None, heads=None): |
|
1147 | def findmissing(self, common=None, heads=None): | |
1169 | """Return the ancestors of heads that are not ancestors of common. |
|
1148 | """Return the ancestors of heads that are not ancestors of common. | |
1170 |
|
1149 | |||
1171 | More specifically, return a list of nodes N such that every N |
|
1150 | More specifically, return a list of nodes N such that every N | |
1172 | satisfies the following constraints: |
|
1151 | satisfies the following constraints: | |
1173 |
|
1152 | |||
1174 | 1. N is an ancestor of some node in 'heads' |
|
1153 | 1. N is an ancestor of some node in 'heads' | |
1175 | 2. N is not an ancestor of any node in 'common' |
|
1154 | 2. N is not an ancestor of any node in 'common' | |
1176 |
|
1155 | |||
1177 | The list is sorted by revision number, meaning it is |
|
1156 | The list is sorted by revision number, meaning it is | |
1178 | topologically sorted. |
|
1157 | topologically sorted. | |
1179 |
|
1158 | |||
1180 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1159 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1181 | not supplied, uses all of the revlog's heads. If common is not |
|
1160 | not supplied, uses all of the revlog's heads. If common is not | |
1182 | supplied, uses nullid.""" |
|
1161 | supplied, uses nullid.""" | |
1183 | if common is None: |
|
1162 | if common is None: | |
1184 | common = [self.nullid] |
|
1163 | common = [self.nullid] | |
1185 | if heads is None: |
|
1164 | if heads is None: | |
1186 | heads = self.heads() |
|
1165 | heads = self.heads() | |
1187 |
|
1166 | |||
1188 | common = [self.rev(n) for n in common] |
|
1167 | common = [self.rev(n) for n in common] | |
1189 | heads = [self.rev(n) for n in heads] |
|
1168 | heads = [self.rev(n) for n in heads] | |
1190 |
|
1169 | |||
1191 | inc = self.incrementalmissingrevs(common=common) |
|
1170 | inc = self.incrementalmissingrevs(common=common) | |
1192 | return [self.node(r) for r in inc.missingancestors(heads)] |
|
1171 | return [self.node(r) for r in inc.missingancestors(heads)] | |
1193 |
|
1172 | |||
1194 | def nodesbetween(self, roots=None, heads=None): |
|
1173 | def nodesbetween(self, roots=None, heads=None): | |
1195 | """Return a topological path from 'roots' to 'heads'. |
|
1174 | """Return a topological path from 'roots' to 'heads'. | |
1196 |
|
1175 | |||
1197 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a |
|
1176 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a | |
1198 | topologically sorted list of all nodes N that satisfy both of |
|
1177 | topologically sorted list of all nodes N that satisfy both of | |
1199 | these constraints: |
|
1178 | these constraints: | |
1200 |
|
1179 | |||
1201 | 1. N is a descendant of some node in 'roots' |
|
1180 | 1. N is a descendant of some node in 'roots' | |
1202 | 2. N is an ancestor of some node in 'heads' |
|
1181 | 2. N is an ancestor of some node in 'heads' | |
1203 |
|
1182 | |||
1204 | Every node is considered to be both a descendant and an ancestor |
|
1183 | Every node is considered to be both a descendant and an ancestor | |
1205 | of itself, so every reachable node in 'roots' and 'heads' will be |
|
1184 | of itself, so every reachable node in 'roots' and 'heads' will be | |
1206 | included in 'nodes'. |
|
1185 | included in 'nodes'. | |
1207 |
|
1186 | |||
1208 | 'outroots' is the list of reachable nodes in 'roots', i.e., the |
|
1187 | 'outroots' is the list of reachable nodes in 'roots', i.e., the | |
1209 | subset of 'roots' that is returned in 'nodes'. Likewise, |
|
1188 | subset of 'roots' that is returned in 'nodes'. Likewise, | |
1210 | 'outheads' is the subset of 'heads' that is also in 'nodes'. |
|
1189 | 'outheads' is the subset of 'heads' that is also in 'nodes'. | |
1211 |
|
1190 | |||
1212 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is |
|
1191 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is | |
1213 | unspecified, uses nullid as the only root. If 'heads' is |
|
1192 | unspecified, uses nullid as the only root. If 'heads' is | |
1214 | unspecified, uses list of all of the revlog's heads.""" |
|
1193 | unspecified, uses list of all of the revlog's heads.""" | |
1215 | nonodes = ([], [], []) |
|
1194 | nonodes = ([], [], []) | |
1216 | if roots is not None: |
|
1195 | if roots is not None: | |
1217 | roots = list(roots) |
|
1196 | roots = list(roots) | |
1218 | if not roots: |
|
1197 | if not roots: | |
1219 | return nonodes |
|
1198 | return nonodes | |
1220 | lowestrev = min([self.rev(n) for n in roots]) |
|
1199 | lowestrev = min([self.rev(n) for n in roots]) | |
1221 | else: |
|
1200 | else: | |
1222 | roots = [self.nullid] # Everybody's a descendant of nullid |
|
1201 | roots = [self.nullid] # Everybody's a descendant of nullid | |
1223 | lowestrev = nullrev |
|
1202 | lowestrev = nullrev | |
1224 | if (lowestrev == nullrev) and (heads is None): |
|
1203 | if (lowestrev == nullrev) and (heads is None): | |
1225 | # We want _all_ the nodes! |
|
1204 | # We want _all_ the nodes! | |
1226 | return ( |
|
1205 | return ( | |
1227 | [self.node(r) for r in self], |
|
1206 | [self.node(r) for r in self], | |
1228 | [self.nullid], |
|
1207 | [self.nullid], | |
1229 | list(self.heads()), |
|
1208 | list(self.heads()), | |
1230 | ) |
|
1209 | ) | |
1231 | if heads is None: |
|
1210 | if heads is None: | |
1232 | # All nodes are ancestors, so the latest ancestor is the last |
|
1211 | # All nodes are ancestors, so the latest ancestor is the last | |
1233 | # node. |
|
1212 | # node. | |
1234 | highestrev = len(self) - 1 |
|
1213 | highestrev = len(self) - 1 | |
1235 | # Set ancestors to None to signal that every node is an ancestor. |
|
1214 | # Set ancestors to None to signal that every node is an ancestor. | |
1236 | ancestors = None |
|
1215 | ancestors = None | |
1237 | # Set heads to an empty dictionary for later discovery of heads |
|
1216 | # Set heads to an empty dictionary for later discovery of heads | |
1238 | heads = {} |
|
1217 | heads = {} | |
1239 | else: |
|
1218 | else: | |
1240 | heads = list(heads) |
|
1219 | heads = list(heads) | |
1241 | if not heads: |
|
1220 | if not heads: | |
1242 | return nonodes |
|
1221 | return nonodes | |
1243 | ancestors = set() |
|
1222 | ancestors = set() | |
1244 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
1223 | # Turn heads into a dictionary so we can remove 'fake' heads. | |
1245 | # Also, later we will be using it to filter out the heads we can't |
|
1224 | # Also, later we will be using it to filter out the heads we can't | |
1246 | # find from roots. |
|
1225 | # find from roots. | |
1247 | heads = dict.fromkeys(heads, False) |
|
1226 | heads = dict.fromkeys(heads, False) | |
1248 | # Start at the top and keep marking parents until we're done. |
|
1227 | # Start at the top and keep marking parents until we're done. | |
1249 | nodestotag = set(heads) |
|
1228 | nodestotag = set(heads) | |
1250 | # Remember where the top was so we can use it as a limit later. |
|
1229 | # Remember where the top was so we can use it as a limit later. | |
1251 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
1230 | highestrev = max([self.rev(n) for n in nodestotag]) | |
1252 | while nodestotag: |
|
1231 | while nodestotag: | |
1253 | # grab a node to tag |
|
1232 | # grab a node to tag | |
1254 | n = nodestotag.pop() |
|
1233 | n = nodestotag.pop() | |
1255 | # Never tag nullid |
|
1234 | # Never tag nullid | |
1256 | if n == self.nullid: |
|
1235 | if n == self.nullid: | |
1257 | continue |
|
1236 | continue | |
1258 | # A node's revision number represents its place in a |
|
1237 | # A node's revision number represents its place in a | |
1259 | # topologically sorted list of nodes. |
|
1238 | # topologically sorted list of nodes. | |
1260 | r = self.rev(n) |
|
1239 | r = self.rev(n) | |
1261 | if r >= lowestrev: |
|
1240 | if r >= lowestrev: | |
1262 | if n not in ancestors: |
|
1241 | if n not in ancestors: | |
1263 | # If we are possibly a descendant of one of the roots |
|
1242 | # If we are possibly a descendant of one of the roots | |
1264 | # and we haven't already been marked as an ancestor |
|
1243 | # and we haven't already been marked as an ancestor | |
1265 | ancestors.add(n) # Mark as ancestor |
|
1244 | ancestors.add(n) # Mark as ancestor | |
1266 | # Add non-nullid parents to list of nodes to tag. |
|
1245 | # Add non-nullid parents to list of nodes to tag. | |
1267 | nodestotag.update( |
|
1246 | nodestotag.update( | |
1268 | [p for p in self.parents(n) if p != self.nullid] |
|
1247 | [p for p in self.parents(n) if p != self.nullid] | |
1269 | ) |
|
1248 | ) | |
1270 | elif n in heads: # We've seen it before, is it a fake head? |
|
1249 | elif n in heads: # We've seen it before, is it a fake head? | |
1271 | # So it is, real heads should not be the ancestors of |
|
1250 | # So it is, real heads should not be the ancestors of | |
1272 | # any other heads. |
|
1251 | # any other heads. | |
1273 | heads.pop(n) |
|
1252 | heads.pop(n) | |
1274 | if not ancestors: |
|
1253 | if not ancestors: | |
1275 | return nonodes |
|
1254 | return nonodes | |
1276 | # Now that we have our set of ancestors, we want to remove any |
|
1255 | # Now that we have our set of ancestors, we want to remove any | |
1277 | # roots that are not ancestors. |
|
1256 | # roots that are not ancestors. | |
1278 |
|
1257 | |||
1279 | # If one of the roots was nullid, everything is included anyway. |
|
1258 | # If one of the roots was nullid, everything is included anyway. | |
1280 | if lowestrev > nullrev: |
|
1259 | if lowestrev > nullrev: | |
1281 | # But, since we weren't, let's recompute the lowest rev to not |
|
1260 | # But, since we weren't, let's recompute the lowest rev to not | |
1282 | # include roots that aren't ancestors. |
|
1261 | # include roots that aren't ancestors. | |
1283 |
|
1262 | |||
1284 | # Filter out roots that aren't ancestors of heads |
|
1263 | # Filter out roots that aren't ancestors of heads | |
1285 | roots = [root for root in roots if root in ancestors] |
|
1264 | roots = [root for root in roots if root in ancestors] | |
1286 | # Recompute the lowest revision |
|
1265 | # Recompute the lowest revision | |
1287 | if roots: |
|
1266 | if roots: | |
1288 | lowestrev = min([self.rev(root) for root in roots]) |
|
1267 | lowestrev = min([self.rev(root) for root in roots]) | |
1289 | else: |
|
1268 | else: | |
1290 | # No more roots? Return empty list |
|
1269 | # No more roots? Return empty list | |
1291 | return nonodes |
|
1270 | return nonodes | |
1292 | else: |
|
1271 | else: | |
1293 | # We are descending from nullid, and don't need to care about |
|
1272 | # We are descending from nullid, and don't need to care about | |
1294 | # any other roots. |
|
1273 | # any other roots. | |
1295 | lowestrev = nullrev |
|
1274 | lowestrev = nullrev | |
1296 | roots = [self.nullid] |
|
1275 | roots = [self.nullid] | |
1297 | # Transform our roots list into a set. |
|
1276 | # Transform our roots list into a set. | |
1298 | descendants = set(roots) |
|
1277 | descendants = set(roots) | |
1299 | # Also, keep the original roots so we can filter out roots that aren't |
|
1278 | # Also, keep the original roots so we can filter out roots that aren't | |
1300 | # 'real' roots (i.e. are descended from other roots). |
|
1279 | # 'real' roots (i.e. are descended from other roots). | |
1301 | roots = descendants.copy() |
|
1280 | roots = descendants.copy() | |
1302 | # Our topologically sorted list of output nodes. |
|
1281 | # Our topologically sorted list of output nodes. | |
1303 | orderedout = [] |
|
1282 | orderedout = [] | |
1304 | # Don't start at nullid since we don't want nullid in our output list, |
|
1283 | # Don't start at nullid since we don't want nullid in our output list, | |
1305 | # and if nullid shows up in descendants, empty parents will look like |
|
1284 | # and if nullid shows up in descendants, empty parents will look like | |
1306 | # they're descendants. |
|
1285 | # they're descendants. | |
1307 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): |
|
1286 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): | |
1308 | n = self.node(r) |
|
1287 | n = self.node(r) | |
1309 | isdescendant = False |
|
1288 | isdescendant = False | |
1310 | if lowestrev == nullrev: # Everybody is a descendant of nullid |
|
1289 | if lowestrev == nullrev: # Everybody is a descendant of nullid | |
1311 | isdescendant = True |
|
1290 | isdescendant = True | |
1312 | elif n in descendants: |
|
1291 | elif n in descendants: | |
1313 | # n is already a descendant |
|
1292 | # n is already a descendant | |
1314 | isdescendant = True |
|
1293 | isdescendant = True | |
1315 | # This check only needs to be done here because all the roots |
|
1294 | # This check only needs to be done here because all the roots | |
1316 | # will start being marked is descendants before the loop. |
|
1295 | # will start being marked is descendants before the loop. | |
1317 | if n in roots: |
|
1296 | if n in roots: | |
1318 | # If n was a root, check if it's a 'real' root. |
|
1297 | # If n was a root, check if it's a 'real' root. | |
1319 | p = tuple(self.parents(n)) |
|
1298 | p = tuple(self.parents(n)) | |
1320 | # If any of its parents are descendants, it's not a root. |
|
1299 | # If any of its parents are descendants, it's not a root. | |
1321 | if (p[0] in descendants) or (p[1] in descendants): |
|
1300 | if (p[0] in descendants) or (p[1] in descendants): | |
1322 | roots.remove(n) |
|
1301 | roots.remove(n) | |
1323 | else: |
|
1302 | else: | |
1324 | p = tuple(self.parents(n)) |
|
1303 | p = tuple(self.parents(n)) | |
1325 | # A node is a descendant if either of its parents are |
|
1304 | # A node is a descendant if either of its parents are | |
1326 | # descendants. (We seeded the dependents list with the roots |
|
1305 | # descendants. (We seeded the dependents list with the roots | |
1327 | # up there, remember?) |
|
1306 | # up there, remember?) | |
1328 | if (p[0] in descendants) or (p[1] in descendants): |
|
1307 | if (p[0] in descendants) or (p[1] in descendants): | |
1329 | descendants.add(n) |
|
1308 | descendants.add(n) | |
1330 | isdescendant = True |
|
1309 | isdescendant = True | |
1331 | if isdescendant and ((ancestors is None) or (n in ancestors)): |
|
1310 | if isdescendant and ((ancestors is None) or (n in ancestors)): | |
1332 | # Only include nodes that are both descendants and ancestors. |
|
1311 | # Only include nodes that are both descendants and ancestors. | |
1333 | orderedout.append(n) |
|
1312 | orderedout.append(n) | |
1334 | if (ancestors is not None) and (n in heads): |
|
1313 | if (ancestors is not None) and (n in heads): | |
1335 | # We're trying to figure out which heads are reachable |
|
1314 | # We're trying to figure out which heads are reachable | |
1336 | # from roots. |
|
1315 | # from roots. | |
1337 | # Mark this head as having been reached |
|
1316 | # Mark this head as having been reached | |
1338 | heads[n] = True |
|
1317 | heads[n] = True | |
1339 | elif ancestors is None: |
|
1318 | elif ancestors is None: | |
1340 | # Otherwise, we're trying to discover the heads. |
|
1319 | # Otherwise, we're trying to discover the heads. | |
1341 | # Assume this is a head because if it isn't, the next step |
|
1320 | # Assume this is a head because if it isn't, the next step | |
1342 | # will eventually remove it. |
|
1321 | # will eventually remove it. | |
1343 | heads[n] = True |
|
1322 | heads[n] = True | |
1344 | # But, obviously its parents aren't. |
|
1323 | # But, obviously its parents aren't. | |
1345 | for p in self.parents(n): |
|
1324 | for p in self.parents(n): | |
1346 | heads.pop(p, None) |
|
1325 | heads.pop(p, None) | |
1347 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] |
|
1326 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] | |
1348 | roots = list(roots) |
|
1327 | roots = list(roots) | |
1349 | assert orderedout |
|
1328 | assert orderedout | |
1350 | assert roots |
|
1329 | assert roots | |
1351 | assert heads |
|
1330 | assert heads | |
1352 | return (orderedout, roots, heads) |
|
1331 | return (orderedout, roots, heads) | |
1353 |
|
1332 | |||
1354 | def headrevs(self, revs=None): |
|
1333 | def headrevs(self, revs=None): | |
1355 | if revs is None: |
|
1334 | if revs is None: | |
1356 | try: |
|
1335 | try: | |
1357 | return self.index.headrevs() |
|
1336 | return self.index.headrevs() | |
1358 | except AttributeError: |
|
1337 | except AttributeError: | |
1359 | return self._headrevs() |
|
1338 | return self._headrevs() | |
1360 | if rustdagop is not None and self.index.rust_ext_compat: |
|
1339 | if rustdagop is not None and self.index.rust_ext_compat: | |
1361 | return rustdagop.headrevs(self.index, revs) |
|
1340 | return rustdagop.headrevs(self.index, revs) | |
1362 | return dagop.headrevs(revs, self._uncheckedparentrevs) |
|
1341 | return dagop.headrevs(revs, self._uncheckedparentrevs) | |
1363 |
|
1342 | |||
1364 | def computephases(self, roots): |
|
1343 | def computephases(self, roots): | |
1365 | return self.index.computephasesmapsets(roots) |
|
1344 | return self.index.computephasesmapsets(roots) | |
1366 |
|
1345 | |||
1367 | def _headrevs(self): |
|
1346 | def _headrevs(self): | |
1368 | count = len(self) |
|
1347 | count = len(self) | |
1369 | if not count: |
|
1348 | if not count: | |
1370 | return [nullrev] |
|
1349 | return [nullrev] | |
1371 | # we won't iter over filtered rev so nobody is a head at start |
|
1350 | # we won't iter over filtered rev so nobody is a head at start | |
1372 | ishead = [0] * (count + 1) |
|
1351 | ishead = [0] * (count + 1) | |
1373 | index = self.index |
|
1352 | index = self.index | |
1374 | for r in self: |
|
1353 | for r in self: | |
1375 | ishead[r] = 1 # I may be an head |
|
1354 | ishead[r] = 1 # I may be an head | |
1376 | e = index[r] |
|
1355 | e = index[r] | |
1377 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not |
|
1356 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not | |
1378 | return [r for r, val in enumerate(ishead) if val] |
|
1357 | return [r for r, val in enumerate(ishead) if val] | |
1379 |
|
1358 | |||
1380 | def heads(self, start=None, stop=None): |
|
1359 | def heads(self, start=None, stop=None): | |
1381 | """return the list of all nodes that have no children |
|
1360 | """return the list of all nodes that have no children | |
1382 |
|
1361 | |||
1383 | if start is specified, only heads that are descendants of |
|
1362 | if start is specified, only heads that are descendants of | |
1384 | start will be returned |
|
1363 | start will be returned | |
1385 | if stop is specified, it will consider all the revs from stop |
|
1364 | if stop is specified, it will consider all the revs from stop | |
1386 | as if they had no children |
|
1365 | as if they had no children | |
1387 | """ |
|
1366 | """ | |
1388 | if start is None and stop is None: |
|
1367 | if start is None and stop is None: | |
1389 | if not len(self): |
|
1368 | if not len(self): | |
1390 | return [self.nullid] |
|
1369 | return [self.nullid] | |
1391 | return [self.node(r) for r in self.headrevs()] |
|
1370 | return [self.node(r) for r in self.headrevs()] | |
1392 |
|
1371 | |||
1393 | if start is None: |
|
1372 | if start is None: | |
1394 | start = nullrev |
|
1373 | start = nullrev | |
1395 | else: |
|
1374 | else: | |
1396 | start = self.rev(start) |
|
1375 | start = self.rev(start) | |
1397 |
|
1376 | |||
1398 | stoprevs = {self.rev(n) for n in stop or []} |
|
1377 | stoprevs = {self.rev(n) for n in stop or []} | |
1399 |
|
1378 | |||
1400 | revs = dagop.headrevssubset( |
|
1379 | revs = dagop.headrevssubset( | |
1401 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs |
|
1380 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs | |
1402 | ) |
|
1381 | ) | |
1403 |
|
1382 | |||
1404 | return [self.node(rev) for rev in revs] |
|
1383 | return [self.node(rev) for rev in revs] | |
1405 |
|
1384 | |||
1406 | def children(self, node): |
|
1385 | def children(self, node): | |
1407 | """find the children of a given node""" |
|
1386 | """find the children of a given node""" | |
1408 | c = [] |
|
1387 | c = [] | |
1409 | p = self.rev(node) |
|
1388 | p = self.rev(node) | |
1410 | for r in self.revs(start=p + 1): |
|
1389 | for r in self.revs(start=p + 1): | |
1411 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
1390 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
1412 | if prevs: |
|
1391 | if prevs: | |
1413 | for pr in prevs: |
|
1392 | for pr in prevs: | |
1414 | if pr == p: |
|
1393 | if pr == p: | |
1415 | c.append(self.node(r)) |
|
1394 | c.append(self.node(r)) | |
1416 | elif p == nullrev: |
|
1395 | elif p == nullrev: | |
1417 | c.append(self.node(r)) |
|
1396 | c.append(self.node(r)) | |
1418 | return c |
|
1397 | return c | |
1419 |
|
1398 | |||
1420 | def commonancestorsheads(self, a, b): |
|
1399 | def commonancestorsheads(self, a, b): | |
1421 | """calculate all the heads of the common ancestors of nodes a and b""" |
|
1400 | """calculate all the heads of the common ancestors of nodes a and b""" | |
1422 | a, b = self.rev(a), self.rev(b) |
|
1401 | a, b = self.rev(a), self.rev(b) | |
1423 | ancs = self._commonancestorsheads(a, b) |
|
1402 | ancs = self._commonancestorsheads(a, b) | |
1424 | return pycompat.maplist(self.node, ancs) |
|
1403 | return pycompat.maplist(self.node, ancs) | |
1425 |
|
1404 | |||
1426 | def _commonancestorsheads(self, *revs): |
|
1405 | def _commonancestorsheads(self, *revs): | |
1427 | """calculate all the heads of the common ancestors of revs""" |
|
1406 | """calculate all the heads of the common ancestors of revs""" | |
1428 | try: |
|
1407 | try: | |
1429 | ancs = self.index.commonancestorsheads(*revs) |
|
1408 | ancs = self.index.commonancestorsheads(*revs) | |
1430 | except (AttributeError, OverflowError): # C implementation failed |
|
1409 | except (AttributeError, OverflowError): # C implementation failed | |
1431 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) |
|
1410 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) | |
1432 | return ancs |
|
1411 | return ancs | |
1433 |
|
1412 | |||
1434 | def isancestor(self, a, b): |
|
1413 | def isancestor(self, a, b): | |
1435 | """return True if node a is an ancestor of node b |
|
1414 | """return True if node a is an ancestor of node b | |
1436 |
|
1415 | |||
1437 | A revision is considered an ancestor of itself.""" |
|
1416 | A revision is considered an ancestor of itself.""" | |
1438 | a, b = self.rev(a), self.rev(b) |
|
1417 | a, b = self.rev(a), self.rev(b) | |
1439 | return self.isancestorrev(a, b) |
|
1418 | return self.isancestorrev(a, b) | |
1440 |
|
1419 | |||
1441 | def isancestorrev(self, a, b): |
|
1420 | def isancestorrev(self, a, b): | |
1442 | """return True if revision a is an ancestor of revision b |
|
1421 | """return True if revision a is an ancestor of revision b | |
1443 |
|
1422 | |||
1444 | A revision is considered an ancestor of itself. |
|
1423 | A revision is considered an ancestor of itself. | |
1445 |
|
1424 | |||
1446 | The implementation of this is trivial but the use of |
|
1425 | The implementation of this is trivial but the use of | |
1447 | reachableroots is not.""" |
|
1426 | reachableroots is not.""" | |
1448 | if a == nullrev: |
|
1427 | if a == nullrev: | |
1449 | return True |
|
1428 | return True | |
1450 | elif a == b: |
|
1429 | elif a == b: | |
1451 | return True |
|
1430 | return True | |
1452 | elif a > b: |
|
1431 | elif a > b: | |
1453 | return False |
|
1432 | return False | |
1454 | return bool(self.reachableroots(a, [b], [a], includepath=False)) |
|
1433 | return bool(self.reachableroots(a, [b], [a], includepath=False)) | |
1455 |
|
1434 | |||
1456 | def reachableroots(self, minroot, heads, roots, includepath=False): |
|
1435 | def reachableroots(self, minroot, heads, roots, includepath=False): | |
1457 | """return (heads(::(<roots> and <roots>::<heads>))) |
|
1436 | """return (heads(::(<roots> and <roots>::<heads>))) | |
1458 |
|
1437 | |||
1459 | If includepath is True, return (<roots>::<heads>).""" |
|
1438 | If includepath is True, return (<roots>::<heads>).""" | |
1460 | try: |
|
1439 | try: | |
1461 | return self.index.reachableroots2( |
|
1440 | return self.index.reachableroots2( | |
1462 | minroot, heads, roots, includepath |
|
1441 | minroot, heads, roots, includepath | |
1463 | ) |
|
1442 | ) | |
1464 | except AttributeError: |
|
1443 | except AttributeError: | |
1465 | return dagop._reachablerootspure( |
|
1444 | return dagop._reachablerootspure( | |
1466 | self.parentrevs, minroot, roots, heads, includepath |
|
1445 | self.parentrevs, minroot, roots, heads, includepath | |
1467 | ) |
|
1446 | ) | |
1468 |
|
1447 | |||
1469 | def ancestor(self, a, b): |
|
1448 | def ancestor(self, a, b): | |
1470 | """calculate the "best" common ancestor of nodes a and b""" |
|
1449 | """calculate the "best" common ancestor of nodes a and b""" | |
1471 |
|
1450 | |||
1472 | a, b = self.rev(a), self.rev(b) |
|
1451 | a, b = self.rev(a), self.rev(b) | |
1473 | try: |
|
1452 | try: | |
1474 | ancs = self.index.ancestors(a, b) |
|
1453 | ancs = self.index.ancestors(a, b) | |
1475 | except (AttributeError, OverflowError): |
|
1454 | except (AttributeError, OverflowError): | |
1476 | ancs = ancestor.ancestors(self.parentrevs, a, b) |
|
1455 | ancs = ancestor.ancestors(self.parentrevs, a, b) | |
1477 | if ancs: |
|
1456 | if ancs: | |
1478 | # choose a consistent winner when there's a tie |
|
1457 | # choose a consistent winner when there's a tie | |
1479 | return min(map(self.node, ancs)) |
|
1458 | return min(map(self.node, ancs)) | |
1480 | return self.nullid |
|
1459 | return self.nullid | |
1481 |
|
1460 | |||
1482 | def _match(self, id): |
|
1461 | def _match(self, id): | |
1483 | if isinstance(id, int): |
|
1462 | if isinstance(id, int): | |
1484 | # rev |
|
1463 | # rev | |
1485 | return self.node(id) |
|
1464 | return self.node(id) | |
1486 | if len(id) == self.nodeconstants.nodelen: |
|
1465 | if len(id) == self.nodeconstants.nodelen: | |
1487 | # possibly a binary node |
|
1466 | # possibly a binary node | |
1488 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
1467 | # odds of a binary node being all hex in ASCII are 1 in 10**25 | |
1489 | try: |
|
1468 | try: | |
1490 | node = id |
|
1469 | node = id | |
1491 | self.rev(node) # quick search the index |
|
1470 | self.rev(node) # quick search the index | |
1492 | return node |
|
1471 | return node | |
1493 | except error.LookupError: |
|
1472 | except error.LookupError: | |
1494 | pass # may be partial hex id |
|
1473 | pass # may be partial hex id | |
1495 | try: |
|
1474 | try: | |
1496 | # str(rev) |
|
1475 | # str(rev) | |
1497 | rev = int(id) |
|
1476 | rev = int(id) | |
1498 | if b"%d" % rev != id: |
|
1477 | if b"%d" % rev != id: | |
1499 | raise ValueError |
|
1478 | raise ValueError | |
1500 | if rev < 0: |
|
1479 | if rev < 0: | |
1501 | rev = len(self) + rev |
|
1480 | rev = len(self) + rev | |
1502 | if rev < 0 or rev >= len(self): |
|
1481 | if rev < 0 or rev >= len(self): | |
1503 | raise ValueError |
|
1482 | raise ValueError | |
1504 | return self.node(rev) |
|
1483 | return self.node(rev) | |
1505 | except (ValueError, OverflowError): |
|
1484 | except (ValueError, OverflowError): | |
1506 | pass |
|
1485 | pass | |
1507 | if len(id) == 2 * self.nodeconstants.nodelen: |
|
1486 | if len(id) == 2 * self.nodeconstants.nodelen: | |
1508 | try: |
|
1487 | try: | |
1509 | # a full hex nodeid? |
|
1488 | # a full hex nodeid? | |
1510 | node = bin(id) |
|
1489 | node = bin(id) | |
1511 | self.rev(node) |
|
1490 | self.rev(node) | |
1512 | return node |
|
1491 | return node | |
1513 | except (TypeError, error.LookupError): |
|
1492 | except (TypeError, error.LookupError): | |
1514 | pass |
|
1493 | pass | |
1515 |
|
1494 | |||
1516 | def _partialmatch(self, id): |
|
1495 | def _partialmatch(self, id): | |
1517 | # we don't care wdirfilenodeids as they should be always full hash |
|
1496 | # we don't care wdirfilenodeids as they should be always full hash | |
1518 | maybewdir = self.nodeconstants.wdirhex.startswith(id) |
|
1497 | maybewdir = self.nodeconstants.wdirhex.startswith(id) | |
1519 | ambiguous = False |
|
1498 | ambiguous = False | |
1520 | try: |
|
1499 | try: | |
1521 | partial = self.index.partialmatch(id) |
|
1500 | partial = self.index.partialmatch(id) | |
1522 | if partial and self.hasnode(partial): |
|
1501 | if partial and self.hasnode(partial): | |
1523 | if maybewdir: |
|
1502 | if maybewdir: | |
1524 | # single 'ff...' match in radix tree, ambiguous with wdir |
|
1503 | # single 'ff...' match in radix tree, ambiguous with wdir | |
1525 | ambiguous = True |
|
1504 | ambiguous = True | |
1526 | else: |
|
1505 | else: | |
1527 | return partial |
|
1506 | return partial | |
1528 | elif maybewdir: |
|
1507 | elif maybewdir: | |
1529 | # no 'ff...' match in radix tree, wdir identified |
|
1508 | # no 'ff...' match in radix tree, wdir identified | |
1530 | raise error.WdirUnsupported |
|
1509 | raise error.WdirUnsupported | |
1531 | else: |
|
1510 | else: | |
1532 | return None |
|
1511 | return None | |
1533 | except error.RevlogError: |
|
1512 | except error.RevlogError: | |
1534 | # parsers.c radix tree lookup gave multiple matches |
|
1513 | # parsers.c radix tree lookup gave multiple matches | |
1535 | # fast path: for unfiltered changelog, radix tree is accurate |
|
1514 | # fast path: for unfiltered changelog, radix tree is accurate | |
1536 | if not getattr(self, 'filteredrevs', None): |
|
1515 | if not getattr(self, 'filteredrevs', None): | |
1537 | ambiguous = True |
|
1516 | ambiguous = True | |
1538 | # fall through to slow path that filters hidden revisions |
|
1517 | # fall through to slow path that filters hidden revisions | |
1539 | except (AttributeError, ValueError): |
|
1518 | except (AttributeError, ValueError): | |
1540 | # we are pure python, or key was too short to search radix tree |
|
1519 | # we are pure python, or key was too short to search radix tree | |
1541 | pass |
|
1520 | pass | |
1542 | if ambiguous: |
|
1521 | if ambiguous: | |
1543 | raise error.AmbiguousPrefixLookupError( |
|
1522 | raise error.AmbiguousPrefixLookupError( | |
1544 | id, self.display_id, _(b'ambiguous identifier') |
|
1523 | id, self.display_id, _(b'ambiguous identifier') | |
1545 | ) |
|
1524 | ) | |
1546 |
|
1525 | |||
1547 | if id in self._pcache: |
|
1526 | if id in self._pcache: | |
1548 | return self._pcache[id] |
|
1527 | return self._pcache[id] | |
1549 |
|
1528 | |||
1550 | if len(id) <= 40: |
|
1529 | if len(id) <= 40: | |
1551 | try: |
|
1530 | try: | |
1552 | # hex(node)[:...] |
|
1531 | # hex(node)[:...] | |
1553 | l = len(id) // 2 # grab an even number of digits |
|
1532 | l = len(id) // 2 # grab an even number of digits | |
1554 | prefix = bin(id[: l * 2]) |
|
1533 | prefix = bin(id[: l * 2]) | |
1555 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] |
|
1534 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] | |
1556 | nl = [ |
|
1535 | nl = [ | |
1557 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) |
|
1536 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) | |
1558 | ] |
|
1537 | ] | |
1559 | if self.nodeconstants.nullhex.startswith(id): |
|
1538 | if self.nodeconstants.nullhex.startswith(id): | |
1560 | nl.append(self.nullid) |
|
1539 | nl.append(self.nullid) | |
1561 | if len(nl) > 0: |
|
1540 | if len(nl) > 0: | |
1562 | if len(nl) == 1 and not maybewdir: |
|
1541 | if len(nl) == 1 and not maybewdir: | |
1563 | self._pcache[id] = nl[0] |
|
1542 | self._pcache[id] = nl[0] | |
1564 | return nl[0] |
|
1543 | return nl[0] | |
1565 | raise error.AmbiguousPrefixLookupError( |
|
1544 | raise error.AmbiguousPrefixLookupError( | |
1566 | id, self.display_id, _(b'ambiguous identifier') |
|
1545 | id, self.display_id, _(b'ambiguous identifier') | |
1567 | ) |
|
1546 | ) | |
1568 | if maybewdir: |
|
1547 | if maybewdir: | |
1569 | raise error.WdirUnsupported |
|
1548 | raise error.WdirUnsupported | |
1570 | return None |
|
1549 | return None | |
1571 | except TypeError: |
|
1550 | except TypeError: | |
1572 | pass |
|
1551 | pass | |
1573 |
|
1552 | |||
1574 | def lookup(self, id): |
|
1553 | def lookup(self, id): | |
1575 | """locate a node based on: |
|
1554 | """locate a node based on: | |
1576 | - revision number or str(revision number) |
|
1555 | - revision number or str(revision number) | |
1577 | - nodeid or subset of hex nodeid |
|
1556 | - nodeid or subset of hex nodeid | |
1578 | """ |
|
1557 | """ | |
1579 | n = self._match(id) |
|
1558 | n = self._match(id) | |
1580 | if n is not None: |
|
1559 | if n is not None: | |
1581 | return n |
|
1560 | return n | |
1582 | n = self._partialmatch(id) |
|
1561 | n = self._partialmatch(id) | |
1583 | if n: |
|
1562 | if n: | |
1584 | return n |
|
1563 | return n | |
1585 |
|
1564 | |||
1586 | raise error.LookupError(id, self.display_id, _(b'no match found')) |
|
1565 | raise error.LookupError(id, self.display_id, _(b'no match found')) | |
1587 |
|
1566 | |||
1588 | def shortest(self, node, minlength=1): |
|
1567 | def shortest(self, node, minlength=1): | |
1589 | """Find the shortest unambiguous prefix that matches node.""" |
|
1568 | """Find the shortest unambiguous prefix that matches node.""" | |
1590 |
|
1569 | |||
1591 | def isvalid(prefix): |
|
1570 | def isvalid(prefix): | |
1592 | try: |
|
1571 | try: | |
1593 | matchednode = self._partialmatch(prefix) |
|
1572 | matchednode = self._partialmatch(prefix) | |
1594 | except error.AmbiguousPrefixLookupError: |
|
1573 | except error.AmbiguousPrefixLookupError: | |
1595 | return False |
|
1574 | return False | |
1596 | except error.WdirUnsupported: |
|
1575 | except error.WdirUnsupported: | |
1597 | # single 'ff...' match |
|
1576 | # single 'ff...' match | |
1598 | return True |
|
1577 | return True | |
1599 | if matchednode is None: |
|
1578 | if matchednode is None: | |
1600 | raise error.LookupError(node, self.display_id, _(b'no node')) |
|
1579 | raise error.LookupError(node, self.display_id, _(b'no node')) | |
1601 | return True |
|
1580 | return True | |
1602 |
|
1581 | |||
1603 | def maybewdir(prefix): |
|
1582 | def maybewdir(prefix): | |
1604 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) |
|
1583 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) | |
1605 |
|
1584 | |||
1606 | hexnode = hex(node) |
|
1585 | hexnode = hex(node) | |
1607 |
|
1586 | |||
1608 | def disambiguate(hexnode, minlength): |
|
1587 | def disambiguate(hexnode, minlength): | |
1609 | """Disambiguate against wdirid.""" |
|
1588 | """Disambiguate against wdirid.""" | |
1610 | for length in range(minlength, len(hexnode) + 1): |
|
1589 | for length in range(minlength, len(hexnode) + 1): | |
1611 | prefix = hexnode[:length] |
|
1590 | prefix = hexnode[:length] | |
1612 | if not maybewdir(prefix): |
|
1591 | if not maybewdir(prefix): | |
1613 | return prefix |
|
1592 | return prefix | |
1614 |
|
1593 | |||
1615 | if not getattr(self, 'filteredrevs', None): |
|
1594 | if not getattr(self, 'filteredrevs', None): | |
1616 | try: |
|
1595 | try: | |
1617 | length = max(self.index.shortest(node), minlength) |
|
1596 | length = max(self.index.shortest(node), minlength) | |
1618 | return disambiguate(hexnode, length) |
|
1597 | return disambiguate(hexnode, length) | |
1619 | except error.RevlogError: |
|
1598 | except error.RevlogError: | |
1620 | if node != self.nodeconstants.wdirid: |
|
1599 | if node != self.nodeconstants.wdirid: | |
1621 | raise error.LookupError( |
|
1600 | raise error.LookupError( | |
1622 | node, self.display_id, _(b'no node') |
|
1601 | node, self.display_id, _(b'no node') | |
1623 | ) |
|
1602 | ) | |
1624 | except AttributeError: |
|
1603 | except AttributeError: | |
1625 | # Fall through to pure code |
|
1604 | # Fall through to pure code | |
1626 | pass |
|
1605 | pass | |
1627 |
|
1606 | |||
1628 | if node == self.nodeconstants.wdirid: |
|
1607 | if node == self.nodeconstants.wdirid: | |
1629 | for length in range(minlength, len(hexnode) + 1): |
|
1608 | for length in range(minlength, len(hexnode) + 1): | |
1630 | prefix = hexnode[:length] |
|
1609 | prefix = hexnode[:length] | |
1631 | if isvalid(prefix): |
|
1610 | if isvalid(prefix): | |
1632 | return prefix |
|
1611 | return prefix | |
1633 |
|
1612 | |||
1634 | for length in range(minlength, len(hexnode) + 1): |
|
1613 | for length in range(minlength, len(hexnode) + 1): | |
1635 | prefix = hexnode[:length] |
|
1614 | prefix = hexnode[:length] | |
1636 | if isvalid(prefix): |
|
1615 | if isvalid(prefix): | |
1637 | return disambiguate(hexnode, length) |
|
1616 | return disambiguate(hexnode, length) | |
1638 |
|
1617 | |||
1639 | def cmp(self, node, text): |
|
1618 | def cmp(self, node, text): | |
1640 | """compare text with a given file revision |
|
1619 | """compare text with a given file revision | |
1641 |
|
1620 | |||
1642 | returns True if text is different than what is stored. |
|
1621 | returns True if text is different than what is stored. | |
1643 | """ |
|
1622 | """ | |
1644 | p1, p2 = self.parents(node) |
|
1623 | p1, p2 = self.parents(node) | |
1645 | return storageutil.hashrevisionsha1(text, p1, p2) != node |
|
1624 | return storageutil.hashrevisionsha1(text, p1, p2) != node | |
1646 |
|
1625 | |||
1647 | def _cachesegment(self, offset, data): |
|
1626 | def _cachesegment(self, offset, data): | |
1648 | """Add a segment to the revlog cache. |
|
1627 | """Add a segment to the revlog cache. | |
1649 |
|
1628 | |||
1650 | Accepts an absolute offset and the data that is at that location. |
|
1629 | Accepts an absolute offset and the data that is at that location. | |
1651 | """ |
|
1630 | """ | |
1652 | o, d = self._chunkcache |
|
1631 | o, d = self._chunkcache | |
1653 | # try to add to existing cache |
|
1632 | # try to add to existing cache | |
1654 | if o + len(d) == offset and len(d) + len(data) < _chunksize: |
|
1633 | if o + len(d) == offset and len(d) + len(data) < _chunksize: | |
1655 | self._chunkcache = o, d + data |
|
1634 | self._chunkcache = o, d + data | |
1656 | else: |
|
1635 | else: | |
1657 | self._chunkcache = offset, data |
|
1636 | self._chunkcache = offset, data | |
1658 |
|
1637 | |||
1659 | def _readsegment(self, offset, length, df=None): |
|
1638 | def _readsegment(self, offset, length, df=None): | |
1660 | """Load a segment of raw data from the revlog. |
|
1639 | """Load a segment of raw data from the revlog. | |
1661 |
|
1640 | |||
1662 | Accepts an absolute offset, length to read, and an optional existing |
|
1641 | Accepts an absolute offset, length to read, and an optional existing | |
1663 | file handle to read from. |
|
1642 | file handle to read from. | |
1664 |
|
1643 | |||
1665 | If an existing file handle is passed, it will be seeked and the |
|
1644 | If an existing file handle is passed, it will be seeked and the | |
1666 | original seek position will NOT be restored. |
|
1645 | original seek position will NOT be restored. | |
1667 |
|
1646 | |||
1668 | Returns a str or buffer of raw byte data. |
|
1647 | Returns a str or buffer of raw byte data. | |
1669 |
|
1648 | |||
1670 | Raises if the requested number of bytes could not be read. |
|
1649 | Raises if the requested number of bytes could not be read. | |
1671 | """ |
|
1650 | """ | |
1672 | # Cache data both forward and backward around the requested |
|
1651 | # Cache data both forward and backward around the requested | |
1673 | # data, in a fixed size window. This helps speed up operations |
|
1652 | # data, in a fixed size window. This helps speed up operations | |
1674 | # involving reading the revlog backwards. |
|
1653 | # involving reading the revlog backwards. | |
1675 | cachesize = self._chunkcachesize |
|
1654 | cachesize = self._chunkcachesize | |
1676 | realoffset = offset & ~(cachesize - 1) |
|
1655 | realoffset = offset & ~(cachesize - 1) | |
1677 | reallength = ( |
|
1656 | reallength = ( | |
1678 | (offset + length + cachesize) & ~(cachesize - 1) |
|
1657 | (offset + length + cachesize) & ~(cachesize - 1) | |
1679 | ) - realoffset |
|
1658 | ) - realoffset | |
1680 | with self._datareadfp(df) as df: |
|
1659 | with self._datareadfp(df) as df: | |
1681 | df.seek(realoffset) |
|
1660 | df.seek(realoffset) | |
1682 | d = df.read(reallength) |
|
1661 | d = df.read(reallength) | |
1683 |
|
1662 | |||
1684 | self._cachesegment(realoffset, d) |
|
1663 | self._cachesegment(realoffset, d) | |
1685 | if offset != realoffset or reallength != length: |
|
1664 | if offset != realoffset or reallength != length: | |
1686 | startoffset = offset - realoffset |
|
1665 | startoffset = offset - realoffset | |
1687 | if len(d) - startoffset < length: |
|
1666 | if len(d) - startoffset < length: | |
1688 | filename = self._indexfile if self._inline else self._datafile |
|
1667 | filename = self._indexfile if self._inline else self._datafile | |
1689 | got = len(d) - startoffset |
|
1668 | got = len(d) - startoffset | |
1690 | m = PARTIAL_READ_MSG % (filename, length, offset, got) |
|
1669 | m = PARTIAL_READ_MSG % (filename, length, offset, got) | |
1691 | raise error.RevlogError(m) |
|
1670 | raise error.RevlogError(m) | |
1692 | return util.buffer(d, startoffset, length) |
|
1671 | return util.buffer(d, startoffset, length) | |
1693 |
|
1672 | |||
1694 | if len(d) < length: |
|
1673 | if len(d) < length: | |
1695 | filename = self._indexfile if self._inline else self._datafile |
|
1674 | filename = self._indexfile if self._inline else self._datafile | |
1696 | got = len(d) - startoffset |
|
1675 | got = len(d) - startoffset | |
1697 | m = PARTIAL_READ_MSG % (filename, length, offset, got) |
|
1676 | m = PARTIAL_READ_MSG % (filename, length, offset, got) | |
1698 | raise error.RevlogError(m) |
|
1677 | raise error.RevlogError(m) | |
1699 |
|
1678 | |||
1700 | return d |
|
1679 | return d | |
1701 |
|
1680 | |||
1702 | def _getsegment(self, offset, length, df=None): |
|
1681 | def _getsegment(self, offset, length, df=None): | |
1703 | """Obtain a segment of raw data from the revlog. |
|
1682 | """Obtain a segment of raw data from the revlog. | |
1704 |
|
1683 | |||
1705 | Accepts an absolute offset, length of bytes to obtain, and an |
|
1684 | Accepts an absolute offset, length of bytes to obtain, and an | |
1706 | optional file handle to the already-opened revlog. If the file |
|
1685 | optional file handle to the already-opened revlog. If the file | |
1707 | handle is used, it's original seek position will not be preserved. |
|
1686 | handle is used, it's original seek position will not be preserved. | |
1708 |
|
1687 | |||
1709 | Requests for data may be returned from a cache. |
|
1688 | Requests for data may be returned from a cache. | |
1710 |
|
1689 | |||
1711 | Returns a str or a buffer instance of raw byte data. |
|
1690 | Returns a str or a buffer instance of raw byte data. | |
1712 | """ |
|
1691 | """ | |
1713 | o, d = self._chunkcache |
|
1692 | o, d = self._chunkcache | |
1714 | l = len(d) |
|
1693 | l = len(d) | |
1715 |
|
1694 | |||
1716 | # is it in the cache? |
|
1695 | # is it in the cache? | |
1717 | cachestart = offset - o |
|
1696 | cachestart = offset - o | |
1718 | cacheend = cachestart + length |
|
1697 | cacheend = cachestart + length | |
1719 | if cachestart >= 0 and cacheend <= l: |
|
1698 | if cachestart >= 0 and cacheend <= l: | |
1720 | if cachestart == 0 and cacheend == l: |
|
1699 | if cachestart == 0 and cacheend == l: | |
1721 | return d # avoid a copy |
|
1700 | return d # avoid a copy | |
1722 | return util.buffer(d, cachestart, cacheend - cachestart) |
|
1701 | return util.buffer(d, cachestart, cacheend - cachestart) | |
1723 |
|
1702 | |||
1724 | return self._readsegment(offset, length, df=df) |
|
1703 | return self._readsegment(offset, length, df=df) | |
1725 |
|
1704 | |||
1726 | def _getsegmentforrevs(self, startrev, endrev, df=None): |
|
1705 | def _getsegmentforrevs(self, startrev, endrev, df=None): | |
1727 | """Obtain a segment of raw data corresponding to a range of revisions. |
|
1706 | """Obtain a segment of raw data corresponding to a range of revisions. | |
1728 |
|
1707 | |||
1729 | Accepts the start and end revisions and an optional already-open |
|
1708 | Accepts the start and end revisions and an optional already-open | |
1730 | file handle to be used for reading. If the file handle is read, its |
|
1709 | file handle to be used for reading. If the file handle is read, its | |
1731 | seek position will not be preserved. |
|
1710 | seek position will not be preserved. | |
1732 |
|
1711 | |||
1733 | Requests for data may be satisfied by a cache. |
|
1712 | Requests for data may be satisfied by a cache. | |
1734 |
|
1713 | |||
1735 | Returns a 2-tuple of (offset, data) for the requested range of |
|
1714 | Returns a 2-tuple of (offset, data) for the requested range of | |
1736 | revisions. Offset is the integer offset from the beginning of the |
|
1715 | revisions. Offset is the integer offset from the beginning of the | |
1737 | revlog and data is a str or buffer of the raw byte data. |
|
1716 | revlog and data is a str or buffer of the raw byte data. | |
1738 |
|
1717 | |||
1739 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` |
|
1718 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` | |
1740 | to determine where each revision's data begins and ends. |
|
1719 | to determine where each revision's data begins and ends. | |
1741 | """ |
|
1720 | """ | |
1742 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons |
|
1721 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons | |
1743 | # (functions are expensive). |
|
1722 | # (functions are expensive). | |
1744 | index = self.index |
|
1723 | index = self.index | |
1745 | istart = index[startrev] |
|
1724 | istart = index[startrev] | |
1746 | start = int(istart[0] >> 16) |
|
1725 | start = int(istart[0] >> 16) | |
1747 | if startrev == endrev: |
|
1726 | if startrev == endrev: | |
1748 | end = start + istart[1] |
|
1727 | end = start + istart[1] | |
1749 | else: |
|
1728 | else: | |
1750 | iend = index[endrev] |
|
1729 | iend = index[endrev] | |
1751 | end = int(iend[0] >> 16) + iend[1] |
|
1730 | end = int(iend[0] >> 16) + iend[1] | |
1752 |
|
1731 | |||
1753 | if self._inline: |
|
1732 | if self._inline: | |
1754 | start += (startrev + 1) * self.index.entry_size |
|
1733 | start += (startrev + 1) * self.index.entry_size | |
1755 | end += (endrev + 1) * self.index.entry_size |
|
1734 | end += (endrev + 1) * self.index.entry_size | |
1756 | length = end - start |
|
1735 | length = end - start | |
1757 |
|
1736 | |||
1758 | return start, self._getsegment(start, length, df=df) |
|
1737 | return start, self._getsegment(start, length, df=df) | |
1759 |
|
1738 | |||
1760 | def _chunk(self, rev, df=None): |
|
1739 | def _chunk(self, rev, df=None): | |
1761 | """Obtain a single decompressed chunk for a revision. |
|
1740 | """Obtain a single decompressed chunk for a revision. | |
1762 |
|
1741 | |||
1763 | Accepts an integer revision and an optional already-open file handle |
|
1742 | Accepts an integer revision and an optional already-open file handle | |
1764 | to be used for reading. If used, the seek position of the file will not |
|
1743 | to be used for reading. If used, the seek position of the file will not | |
1765 | be preserved. |
|
1744 | be preserved. | |
1766 |
|
1745 | |||
1767 | Returns a str holding uncompressed data for the requested revision. |
|
1746 | Returns a str holding uncompressed data for the requested revision. | |
1768 | """ |
|
1747 | """ | |
1769 | compression_mode = self.index[rev][10] |
|
1748 | compression_mode = self.index[rev][10] | |
1770 | data = self._getsegmentforrevs(rev, rev, df=df)[1] |
|
1749 | data = self._getsegmentforrevs(rev, rev, df=df)[1] | |
1771 | if compression_mode == COMP_MODE_PLAIN: |
|
1750 | if compression_mode == COMP_MODE_PLAIN: | |
1772 | return data |
|
1751 | return data | |
1773 | elif compression_mode == COMP_MODE_DEFAULT: |
|
1752 | elif compression_mode == COMP_MODE_DEFAULT: | |
1774 | return self._decompressor(data) |
|
1753 | return self._decompressor(data) | |
1775 | elif compression_mode == COMP_MODE_INLINE: |
|
1754 | elif compression_mode == COMP_MODE_INLINE: | |
1776 | return self.decompress(data) |
|
1755 | return self.decompress(data) | |
1777 | else: |
|
1756 | else: | |
1778 | msg = 'unknown compression mode %d' |
|
1757 | msg = 'unknown compression mode %d' | |
1779 | msg %= compression_mode |
|
1758 | msg %= compression_mode | |
1780 | raise error.RevlogError(msg) |
|
1759 | raise error.RevlogError(msg) | |
1781 |
|
1760 | |||
1782 | def _chunks(self, revs, df=None, targetsize=None): |
|
1761 | def _chunks(self, revs, df=None, targetsize=None): | |
1783 | """Obtain decompressed chunks for the specified revisions. |
|
1762 | """Obtain decompressed chunks for the specified revisions. | |
1784 |
|
1763 | |||
1785 | Accepts an iterable of numeric revisions that are assumed to be in |
|
1764 | Accepts an iterable of numeric revisions that are assumed to be in | |
1786 | ascending order. Also accepts an optional already-open file handle |
|
1765 | ascending order. Also accepts an optional already-open file handle | |
1787 | to be used for reading. If used, the seek position of the file will |
|
1766 | to be used for reading. If used, the seek position of the file will | |
1788 | not be preserved. |
|
1767 | not be preserved. | |
1789 |
|
1768 | |||
1790 | This function is similar to calling ``self._chunk()`` multiple times, |
|
1769 | This function is similar to calling ``self._chunk()`` multiple times, | |
1791 | but is faster. |
|
1770 | but is faster. | |
1792 |
|
1771 | |||
1793 | Returns a list with decompressed data for each requested revision. |
|
1772 | Returns a list with decompressed data for each requested revision. | |
1794 | """ |
|
1773 | """ | |
1795 | if not revs: |
|
1774 | if not revs: | |
1796 | return [] |
|
1775 | return [] | |
1797 | start = self.start |
|
1776 | start = self.start | |
1798 | length = self.length |
|
1777 | length = self.length | |
1799 | inline = self._inline |
|
1778 | inline = self._inline | |
1800 | iosize = self.index.entry_size |
|
1779 | iosize = self.index.entry_size | |
1801 | buffer = util.buffer |
|
1780 | buffer = util.buffer | |
1802 |
|
1781 | |||
1803 | l = [] |
|
1782 | l = [] | |
1804 | ladd = l.append |
|
1783 | ladd = l.append | |
1805 |
|
1784 | |||
1806 | if not self._withsparseread: |
|
1785 | if not self._withsparseread: | |
1807 | slicedchunks = (revs,) |
|
1786 | slicedchunks = (revs,) | |
1808 | else: |
|
1787 | else: | |
1809 | slicedchunks = deltautil.slicechunk( |
|
1788 | slicedchunks = deltautil.slicechunk( | |
1810 | self, revs, targetsize=targetsize |
|
1789 | self, revs, targetsize=targetsize | |
1811 | ) |
|
1790 | ) | |
1812 |
|
1791 | |||
1813 | for revschunk in slicedchunks: |
|
1792 | for revschunk in slicedchunks: | |
1814 | firstrev = revschunk[0] |
|
1793 | firstrev = revschunk[0] | |
1815 | # Skip trailing revisions with empty diff |
|
1794 | # Skip trailing revisions with empty diff | |
1816 | for lastrev in revschunk[::-1]: |
|
1795 | for lastrev in revschunk[::-1]: | |
1817 | if length(lastrev) != 0: |
|
1796 | if length(lastrev) != 0: | |
1818 | break |
|
1797 | break | |
1819 |
|
1798 | |||
1820 | try: |
|
1799 | try: | |
1821 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) |
|
1800 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) | |
1822 | except OverflowError: |
|
1801 | except OverflowError: | |
1823 | # issue4215 - we can't cache a run of chunks greater than |
|
1802 | # issue4215 - we can't cache a run of chunks greater than | |
1824 | # 2G on Windows |
|
1803 | # 2G on Windows | |
1825 | return [self._chunk(rev, df=df) for rev in revschunk] |
|
1804 | return [self._chunk(rev, df=df) for rev in revschunk] | |
1826 |
|
1805 | |||
1827 | decomp = self.decompress |
|
1806 | decomp = self.decompress | |
1828 | # self._decompressor might be None, but will not be used in that case |
|
1807 | # self._decompressor might be None, but will not be used in that case | |
1829 | def_decomp = self._decompressor |
|
1808 | def_decomp = self._decompressor | |
1830 | for rev in revschunk: |
|
1809 | for rev in revschunk: | |
1831 | chunkstart = start(rev) |
|
1810 | chunkstart = start(rev) | |
1832 | if inline: |
|
1811 | if inline: | |
1833 | chunkstart += (rev + 1) * iosize |
|
1812 | chunkstart += (rev + 1) * iosize | |
1834 | chunklength = length(rev) |
|
1813 | chunklength = length(rev) | |
1835 | comp_mode = self.index[rev][10] |
|
1814 | comp_mode = self.index[rev][10] | |
1836 | c = buffer(data, chunkstart - offset, chunklength) |
|
1815 | c = buffer(data, chunkstart - offset, chunklength) | |
1837 | if comp_mode == COMP_MODE_PLAIN: |
|
1816 | if comp_mode == COMP_MODE_PLAIN: | |
1838 | ladd(c) |
|
1817 | ladd(c) | |
1839 | elif comp_mode == COMP_MODE_INLINE: |
|
1818 | elif comp_mode == COMP_MODE_INLINE: | |
1840 | ladd(decomp(c)) |
|
1819 | ladd(decomp(c)) | |
1841 | elif comp_mode == COMP_MODE_DEFAULT: |
|
1820 | elif comp_mode == COMP_MODE_DEFAULT: | |
1842 | ladd(def_decomp(c)) |
|
1821 | ladd(def_decomp(c)) | |
1843 | else: |
|
1822 | else: | |
1844 | msg = 'unknown compression mode %d' |
|
1823 | msg = 'unknown compression mode %d' | |
1845 | msg %= comp_mode |
|
1824 | msg %= comp_mode | |
1846 | raise error.RevlogError(msg) |
|
1825 | raise error.RevlogError(msg) | |
1847 |
|
1826 | |||
1848 | return l |
|
1827 | return l | |
1849 |
|
1828 | |||
1850 | def _chunkclear(self): |
|
1829 | def _chunkclear(self): | |
1851 | """Clear the raw chunk cache.""" |
|
1830 | """Clear the raw chunk cache.""" | |
1852 | self._chunkcache = (0, b'') |
|
1831 | self._chunkcache = (0, b'') | |
1853 |
|
1832 | |||
1854 | def deltaparent(self, rev): |
|
1833 | def deltaparent(self, rev): | |
1855 | """return deltaparent of the given revision""" |
|
1834 | """return deltaparent of the given revision""" | |
1856 | base = self.index[rev][3] |
|
1835 | base = self.index[rev][3] | |
1857 | if base == rev: |
|
1836 | if base == rev: | |
1858 | return nullrev |
|
1837 | return nullrev | |
1859 | elif self._generaldelta: |
|
1838 | elif self._generaldelta: | |
1860 | return base |
|
1839 | return base | |
1861 | else: |
|
1840 | else: | |
1862 | return rev - 1 |
|
1841 | return rev - 1 | |
1863 |
|
1842 | |||
1864 | def issnapshot(self, rev): |
|
1843 | def issnapshot(self, rev): | |
1865 | """tells whether rev is a snapshot""" |
|
1844 | """tells whether rev is a snapshot""" | |
1866 | if not self._sparserevlog: |
|
1845 | if not self._sparserevlog: | |
1867 | return self.deltaparent(rev) == nullrev |
|
1846 | return self.deltaparent(rev) == nullrev | |
1868 | elif util.safehasattr(self.index, b'issnapshot'): |
|
1847 | elif util.safehasattr(self.index, b'issnapshot'): | |
1869 | # directly assign the method to cache the testing and access |
|
1848 | # directly assign the method to cache the testing and access | |
1870 | self.issnapshot = self.index.issnapshot |
|
1849 | self.issnapshot = self.index.issnapshot | |
1871 | return self.issnapshot(rev) |
|
1850 | return self.issnapshot(rev) | |
1872 | if rev == nullrev: |
|
1851 | if rev == nullrev: | |
1873 | return True |
|
1852 | return True | |
1874 | entry = self.index[rev] |
|
1853 | entry = self.index[rev] | |
1875 | base = entry[3] |
|
1854 | base = entry[3] | |
1876 | if base == rev: |
|
1855 | if base == rev: | |
1877 | return True |
|
1856 | return True | |
1878 | if base == nullrev: |
|
1857 | if base == nullrev: | |
1879 | return True |
|
1858 | return True | |
1880 | p1 = entry[5] |
|
1859 | p1 = entry[5] | |
1881 | p2 = entry[6] |
|
1860 | p2 = entry[6] | |
1882 | if base == p1 or base == p2: |
|
1861 | if base == p1 or base == p2: | |
1883 | return False |
|
1862 | return False | |
1884 | return self.issnapshot(base) |
|
1863 | return self.issnapshot(base) | |
1885 |
|
1864 | |||
1886 | def snapshotdepth(self, rev): |
|
1865 | def snapshotdepth(self, rev): | |
1887 | """number of snapshot in the chain before this one""" |
|
1866 | """number of snapshot in the chain before this one""" | |
1888 | if not self.issnapshot(rev): |
|
1867 | if not self.issnapshot(rev): | |
1889 | raise error.ProgrammingError(b'revision %d not a snapshot') |
|
1868 | raise error.ProgrammingError(b'revision %d not a snapshot') | |
1890 | return len(self._deltachain(rev)[0]) - 1 |
|
1869 | return len(self._deltachain(rev)[0]) - 1 | |
1891 |
|
1870 | |||
1892 | def revdiff(self, rev1, rev2): |
|
1871 | def revdiff(self, rev1, rev2): | |
1893 | """return or calculate a delta between two revisions |
|
1872 | """return or calculate a delta between two revisions | |
1894 |
|
1873 | |||
1895 | The delta calculated is in binary form and is intended to be written to |
|
1874 | The delta calculated is in binary form and is intended to be written to | |
1896 | revlog data directly. So this function needs raw revision data. |
|
1875 | revlog data directly. So this function needs raw revision data. | |
1897 | """ |
|
1876 | """ | |
1898 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: |
|
1877 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: | |
1899 | return bytes(self._chunk(rev2)) |
|
1878 | return bytes(self._chunk(rev2)) | |
1900 |
|
1879 | |||
1901 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) |
|
1880 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) | |
1902 |
|
1881 | |||
1903 | def _processflags(self, text, flags, operation, raw=False): |
|
1882 | def _processflags(self, text, flags, operation, raw=False): | |
1904 | """deprecated entry point to access flag processors""" |
|
1883 | """deprecated entry point to access flag processors""" | |
1905 | msg = b'_processflag(...) use the specialized variant' |
|
1884 | msg = b'_processflag(...) use the specialized variant' | |
1906 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1885 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) | |
1907 | if raw: |
|
1886 | if raw: | |
1908 | return text, flagutil.processflagsraw(self, text, flags) |
|
1887 | return text, flagutil.processflagsraw(self, text, flags) | |
1909 | elif operation == b'read': |
|
1888 | elif operation == b'read': | |
1910 | return flagutil.processflagsread(self, text, flags) |
|
1889 | return flagutil.processflagsread(self, text, flags) | |
1911 | else: # write operation |
|
1890 | else: # write operation | |
1912 | return flagutil.processflagswrite(self, text, flags) |
|
1891 | return flagutil.processflagswrite(self, text, flags) | |
1913 |
|
1892 | |||
1914 | def revision(self, nodeorrev, _df=None, raw=False): |
|
1893 | def revision(self, nodeorrev, _df=None, raw=False): | |
1915 | """return an uncompressed revision of a given node or revision |
|
1894 | """return an uncompressed revision of a given node or revision | |
1916 | number. |
|
1895 | number. | |
1917 |
|
1896 | |||
1918 | _df - an existing file handle to read from. (internal-only) |
|
1897 | _df - an existing file handle to read from. (internal-only) | |
1919 | raw - an optional argument specifying if the revision data is to be |
|
1898 | raw - an optional argument specifying if the revision data is to be | |
1920 | treated as raw data when applying flag transforms. 'raw' should be set |
|
1899 | treated as raw data when applying flag transforms. 'raw' should be set | |
1921 | to True when generating changegroups or in debug commands. |
|
1900 | to True when generating changegroups or in debug commands. | |
1922 | """ |
|
1901 | """ | |
1923 | if raw: |
|
1902 | if raw: | |
1924 | msg = ( |
|
1903 | msg = ( | |
1925 | b'revlog.revision(..., raw=True) is deprecated, ' |
|
1904 | b'revlog.revision(..., raw=True) is deprecated, ' | |
1926 | b'use revlog.rawdata(...)' |
|
1905 | b'use revlog.rawdata(...)' | |
1927 | ) |
|
1906 | ) | |
1928 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1907 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) | |
1929 | return self._revisiondata(nodeorrev, _df, raw=raw) |
|
1908 | return self._revisiondata(nodeorrev, _df, raw=raw) | |
1930 |
|
1909 | |||
1931 | def sidedata(self, nodeorrev, _df=None): |
|
1910 | def sidedata(self, nodeorrev, _df=None): | |
1932 | """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 | |
1933 |
|
1912 | |||
1934 | This function currently return a dictionary. However, more advanced |
|
1913 | This function currently return a dictionary. However, more advanced | |
1935 | 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 | |
1936 | efficient/lazy code. |
|
1915 | efficient/lazy code. | |
1937 | """ |
|
1916 | """ | |
1938 | # deal with <nodeorrev> argument type |
|
1917 | # deal with <nodeorrev> argument type | |
1939 | if isinstance(nodeorrev, int): |
|
1918 | if isinstance(nodeorrev, int): | |
1940 | rev = nodeorrev |
|
1919 | rev = nodeorrev | |
1941 | else: |
|
1920 | else: | |
1942 | rev = self.rev(nodeorrev) |
|
1921 | rev = self.rev(nodeorrev) | |
1943 | return self._sidedata(rev) |
|
1922 | return self._sidedata(rev) | |
1944 |
|
1923 | |||
1945 | def _revisiondata(self, nodeorrev, _df=None, raw=False): |
|
1924 | def _revisiondata(self, nodeorrev, _df=None, raw=False): | |
1946 | # deal with <nodeorrev> argument type |
|
1925 | # deal with <nodeorrev> argument type | |
1947 | if isinstance(nodeorrev, int): |
|
1926 | if isinstance(nodeorrev, int): | |
1948 | rev = nodeorrev |
|
1927 | rev = nodeorrev | |
1949 | node = self.node(rev) |
|
1928 | node = self.node(rev) | |
1950 | else: |
|
1929 | else: | |
1951 | node = nodeorrev |
|
1930 | node = nodeorrev | |
1952 | rev = None |
|
1931 | rev = None | |
1953 |
|
1932 | |||
1954 | # fast path the special `nullid` rev |
|
1933 | # fast path the special `nullid` rev | |
1955 | if node == self.nullid: |
|
1934 | if node == self.nullid: | |
1956 | return b"" |
|
1935 | return b"" | |
1957 |
|
1936 | |||
1958 | # ``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 | |
1959 | # revision or might need to be processed to retrieve the revision. |
|
1938 | # revision or might need to be processed to retrieve the revision. | |
1960 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) |
|
1939 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) | |
1961 |
|
1940 | |||
1962 | if raw and validated: |
|
1941 | if raw and validated: | |
1963 | # 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 | |
1964 | # text is cached, we can exit early. |
|
1943 | # text is cached, we can exit early. | |
1965 | return rawtext |
|
1944 | return rawtext | |
1966 | if rev is None: |
|
1945 | if rev is None: | |
1967 | rev = self.rev(node) |
|
1946 | rev = self.rev(node) | |
1968 | # the revlog's flag for this revision |
|
1947 | # the revlog's flag for this revision | |
1969 | # (usually alter its state or content) |
|
1948 | # (usually alter its state or content) | |
1970 | flags = self.flags(rev) |
|
1949 | flags = self.flags(rev) | |
1971 |
|
1950 | |||
1972 | if validated and flags == REVIDX_DEFAULT_FLAGS: |
|
1951 | if validated and flags == REVIDX_DEFAULT_FLAGS: | |
1973 | # no extra flags set, no flag processor runs, text = rawtext |
|
1952 | # no extra flags set, no flag processor runs, text = rawtext | |
1974 | return rawtext |
|
1953 | return rawtext | |
1975 |
|
1954 | |||
1976 | if raw: |
|
1955 | if raw: | |
1977 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
1956 | validatehash = flagutil.processflagsraw(self, rawtext, flags) | |
1978 | text = rawtext |
|
1957 | text = rawtext | |
1979 | else: |
|
1958 | else: | |
1980 | r = flagutil.processflagsread(self, rawtext, flags) |
|
1959 | r = flagutil.processflagsread(self, rawtext, flags) | |
1981 | text, validatehash = r |
|
1960 | text, validatehash = r | |
1982 | if validatehash: |
|
1961 | if validatehash: | |
1983 | self.checkhash(text, node, rev=rev) |
|
1962 | self.checkhash(text, node, rev=rev) | |
1984 | if not validated: |
|
1963 | if not validated: | |
1985 | self._revisioncache = (node, rev, rawtext) |
|
1964 | self._revisioncache = (node, rev, rawtext) | |
1986 |
|
1965 | |||
1987 | return text |
|
1966 | return text | |
1988 |
|
1967 | |||
1989 | def _rawtext(self, node, rev, _df=None): |
|
1968 | def _rawtext(self, node, rev, _df=None): | |
1990 | """return the possibly unvalidated rawtext for a revision |
|
1969 | """return the possibly unvalidated rawtext for a revision | |
1991 |
|
1970 | |||
1992 | returns (rev, rawtext, validated) |
|
1971 | returns (rev, rawtext, validated) | |
1993 | """ |
|
1972 | """ | |
1994 |
|
1973 | |||
1995 | # revision in the cache (could be useful to apply delta) |
|
1974 | # revision in the cache (could be useful to apply delta) | |
1996 | cachedrev = None |
|
1975 | cachedrev = None | |
1997 | # An intermediate text to apply deltas to |
|
1976 | # An intermediate text to apply deltas to | |
1998 | basetext = None |
|
1977 | basetext = None | |
1999 |
|
1978 | |||
2000 | # Check if we have the entry in cache |
|
1979 | # Check if we have the entry in cache | |
2001 | # The cache entry looks like (node, rev, rawtext) |
|
1980 | # The cache entry looks like (node, rev, rawtext) | |
2002 | if self._revisioncache: |
|
1981 | if self._revisioncache: | |
2003 | if self._revisioncache[0] == node: |
|
1982 | if self._revisioncache[0] == node: | |
2004 | return (rev, self._revisioncache[2], True) |
|
1983 | return (rev, self._revisioncache[2], True) | |
2005 | cachedrev = self._revisioncache[1] |
|
1984 | cachedrev = self._revisioncache[1] | |
2006 |
|
1985 | |||
2007 | if rev is None: |
|
1986 | if rev is None: | |
2008 | rev = self.rev(node) |
|
1987 | rev = self.rev(node) | |
2009 |
|
1988 | |||
2010 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) |
|
1989 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) | |
2011 | if stopped: |
|
1990 | if stopped: | |
2012 | basetext = self._revisioncache[2] |
|
1991 | basetext = self._revisioncache[2] | |
2013 |
|
1992 | |||
2014 | # drop cache to save memory, the caller is expected to |
|
1993 | # drop cache to save memory, the caller is expected to | |
2015 | # update self._revisioncache after validating the text |
|
1994 | # update self._revisioncache after validating the text | |
2016 | self._revisioncache = None |
|
1995 | self._revisioncache = None | |
2017 |
|
1996 | |||
2018 | targetsize = None |
|
1997 | targetsize = None | |
2019 | rawsize = self.index[rev][2] |
|
1998 | rawsize = self.index[rev][2] | |
2020 | if 0 <= rawsize: |
|
1999 | if 0 <= rawsize: | |
2021 | targetsize = 4 * rawsize |
|
2000 | targetsize = 4 * rawsize | |
2022 |
|
2001 | |||
2023 | bins = self._chunks(chain, df=_df, targetsize=targetsize) |
|
2002 | bins = self._chunks(chain, df=_df, targetsize=targetsize) | |
2024 | if basetext is None: |
|
2003 | if basetext is None: | |
2025 | basetext = bytes(bins[0]) |
|
2004 | basetext = bytes(bins[0]) | |
2026 | bins = bins[1:] |
|
2005 | bins = bins[1:] | |
2027 |
|
2006 | |||
2028 | rawtext = mdiff.patches(basetext, bins) |
|
2007 | rawtext = mdiff.patches(basetext, bins) | |
2029 | del basetext # let us have a chance to free memory early |
|
2008 | del basetext # let us have a chance to free memory early | |
2030 | return (rev, rawtext, False) |
|
2009 | return (rev, rawtext, False) | |
2031 |
|
2010 | |||
2032 | def _sidedata(self, rev): |
|
2011 | def _sidedata(self, rev): | |
2033 | """Return the sidedata for a given revision number.""" |
|
2012 | """Return the sidedata for a given revision number.""" | |
2034 | index_entry = self.index[rev] |
|
2013 | index_entry = self.index[rev] | |
2035 | sidedata_offset = index_entry[8] |
|
2014 | sidedata_offset = index_entry[8] | |
2036 | sidedata_size = index_entry[9] |
|
2015 | sidedata_size = index_entry[9] | |
2037 |
|
2016 | |||
2038 | if self._inline: |
|
2017 | if self._inline: | |
2039 | sidedata_offset += self.index.entry_size * (1 + rev) |
|
2018 | sidedata_offset += self.index.entry_size * (1 + rev) | |
2040 | if sidedata_size == 0: |
|
2019 | if sidedata_size == 0: | |
2041 | return {} |
|
2020 | return {} | |
2042 |
|
2021 | |||
2043 | # XXX this need caching, as we do for data |
|
2022 | # XXX this need caching, as we do for data | |
2044 | with self._sidedatareadfp() as sdf: |
|
2023 | with self._sidedatareadfp() as sdf: | |
2045 | if self._docket.sidedata_end < sidedata_offset + sidedata_size: |
|
2024 | if self._docket.sidedata_end < sidedata_offset + sidedata_size: | |
2046 | filename = self._sidedatafile |
|
2025 | filename = self._sidedatafile | |
2047 | end = self._docket.sidedata_end |
|
2026 | end = self._docket.sidedata_end | |
2048 | offset = sidedata_offset |
|
2027 | offset = sidedata_offset | |
2049 | length = sidedata_size |
|
2028 | length = sidedata_size | |
2050 | m = FILE_TOO_SHORT_MSG % (filename, length, offset, end) |
|
2029 | m = FILE_TOO_SHORT_MSG % (filename, length, offset, end) | |
2051 | raise error.RevlogError(m) |
|
2030 | raise error.RevlogError(m) | |
2052 |
|
2031 | |||
2053 | sdf.seek(sidedata_offset, os.SEEK_SET) |
|
2032 | sdf.seek(sidedata_offset, os.SEEK_SET) | |
2054 | comp_segment = sdf.read(sidedata_size) |
|
2033 | comp_segment = sdf.read(sidedata_size) | |
2055 |
|
2034 | |||
2056 | if len(comp_segment) < sidedata_size: |
|
2035 | if len(comp_segment) < sidedata_size: | |
2057 | filename = self._sidedatafile |
|
2036 | filename = self._sidedatafile | |
2058 | length = sidedata_size |
|
2037 | length = sidedata_size | |
2059 | offset = sidedata_offset |
|
2038 | offset = sidedata_offset | |
2060 | got = len(comp_segment) |
|
2039 | got = len(comp_segment) | |
2061 | m = PARTIAL_READ_MSG % (filename, length, offset, got) |
|
2040 | m = PARTIAL_READ_MSG % (filename, length, offset, got) | |
2062 | raise error.RevlogError(m) |
|
2041 | raise error.RevlogError(m) | |
2063 |
|
2042 | |||
2064 | comp = self.index[rev][11] |
|
2043 | comp = self.index[rev][11] | |
2065 | if comp == COMP_MODE_PLAIN: |
|
2044 | if comp == COMP_MODE_PLAIN: | |
2066 | segment = comp_segment |
|
2045 | segment = comp_segment | |
2067 | elif comp == COMP_MODE_DEFAULT: |
|
2046 | elif comp == COMP_MODE_DEFAULT: | |
2068 | segment = self._decompressor(comp_segment) |
|
2047 | segment = self._decompressor(comp_segment) | |
2069 | elif comp == COMP_MODE_INLINE: |
|
2048 | elif comp == COMP_MODE_INLINE: | |
2070 | segment = self.decompress(comp_segment) |
|
2049 | segment = self.decompress(comp_segment) | |
2071 | else: |
|
2050 | else: | |
2072 | msg = 'unknown compression mode %d' |
|
2051 | msg = 'unknown compression mode %d' | |
2073 | msg %= comp |
|
2052 | msg %= comp | |
2074 | raise error.RevlogError(msg) |
|
2053 | raise error.RevlogError(msg) | |
2075 |
|
2054 | |||
2076 | sidedata = sidedatautil.deserialize_sidedata(segment) |
|
2055 | sidedata = sidedatautil.deserialize_sidedata(segment) | |
2077 | return sidedata |
|
2056 | return sidedata | |
2078 |
|
2057 | |||
2079 | def rawdata(self, nodeorrev, _df=None): |
|
2058 | def rawdata(self, nodeorrev, _df=None): | |
2080 | """return an uncompressed raw data of a given node or revision number. |
|
2059 | """return an uncompressed raw data of a given node or revision number. | |
2081 |
|
2060 | |||
2082 | _df - an existing file handle to read from. (internal-only) |
|
2061 | _df - an existing file handle to read from. (internal-only) | |
2083 | """ |
|
2062 | """ | |
2084 | return self._revisiondata(nodeorrev, _df, raw=True) |
|
2063 | return self._revisiondata(nodeorrev, _df, raw=True) | |
2085 |
|
2064 | |||
2086 | def hash(self, text, p1, p2): |
|
2065 | def hash(self, text, p1, p2): | |
2087 | """Compute a node hash. |
|
2066 | """Compute a node hash. | |
2088 |
|
2067 | |||
2089 | Available as a function so that subclasses can replace the hash |
|
2068 | Available as a function so that subclasses can replace the hash | |
2090 | as needed. |
|
2069 | as needed. | |
2091 | """ |
|
2070 | """ | |
2092 | return storageutil.hashrevisionsha1(text, p1, p2) |
|
2071 | return storageutil.hashrevisionsha1(text, p1, p2) | |
2093 |
|
2072 | |||
2094 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
2073 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
2095 | """Check node hash integrity. |
|
2074 | """Check node hash integrity. | |
2096 |
|
2075 | |||
2097 | Available as a function so that subclasses can extend hash mismatch |
|
2076 | Available as a function so that subclasses can extend hash mismatch | |
2098 | behaviors as needed. |
|
2077 | behaviors as needed. | |
2099 | """ |
|
2078 | """ | |
2100 | try: |
|
2079 | try: | |
2101 | if p1 is None and p2 is None: |
|
2080 | if p1 is None and p2 is None: | |
2102 | p1, p2 = self.parents(node) |
|
2081 | p1, p2 = self.parents(node) | |
2103 | if node != self.hash(text, p1, p2): |
|
2082 | if node != self.hash(text, p1, p2): | |
2104 | # Clear the revision cache on hash failure. The revision cache |
|
2083 | # Clear the revision cache on hash failure. The revision cache | |
2105 | # only stores the raw revision and clearing the cache does have |
|
2084 | # only stores the raw revision and clearing the cache does have | |
2106 | # the side-effect that we won't have a cache hit when the raw |
|
2085 | # the side-effect that we won't have a cache hit when the raw | |
2107 | # revision data is accessed. But this case should be rare and |
|
2086 | # revision data is accessed. But this case should be rare and | |
2108 | # it is extra work to teach the cache about the hash |
|
2087 | # it is extra work to teach the cache about the hash | |
2109 | # verification state. |
|
2088 | # verification state. | |
2110 | if self._revisioncache and self._revisioncache[0] == node: |
|
2089 | if self._revisioncache and self._revisioncache[0] == node: | |
2111 | self._revisioncache = None |
|
2090 | self._revisioncache = None | |
2112 |
|
2091 | |||
2113 | revornode = rev |
|
2092 | revornode = rev | |
2114 | if revornode is None: |
|
2093 | if revornode is None: | |
2115 | revornode = templatefilters.short(hex(node)) |
|
2094 | revornode = templatefilters.short(hex(node)) | |
2116 | raise error.RevlogError( |
|
2095 | raise error.RevlogError( | |
2117 | _(b"integrity check failed on %s:%s") |
|
2096 | _(b"integrity check failed on %s:%s") | |
2118 | % (self.display_id, pycompat.bytestr(revornode)) |
|
2097 | % (self.display_id, pycompat.bytestr(revornode)) | |
2119 | ) |
|
2098 | ) | |
2120 | except error.RevlogError: |
|
2099 | except error.RevlogError: | |
2121 | if self._censorable and storageutil.iscensoredtext(text): |
|
2100 | if self._censorable and storageutil.iscensoredtext(text): | |
2122 | raise error.CensoredNodeError(self.display_id, node, text) |
|
2101 | raise error.CensoredNodeError(self.display_id, node, text) | |
2123 | raise |
|
2102 | raise | |
2124 |
|
2103 | |||
2125 | def _enforceinlinesize(self, tr): |
|
2104 | def _enforceinlinesize(self, tr): | |
2126 | """Check if the revlog is too big for inline and convert if so. |
|
2105 | """Check if the revlog is too big for inline and convert if so. | |
2127 |
|
2106 | |||
2128 | This should be called after revisions are added to the revlog. If the |
|
2107 | This should be called after revisions are added to the revlog. If the | |
2129 | revlog has grown too large to be an inline revlog, it will convert it |
|
2108 | revlog has grown too large to be an inline revlog, it will convert it | |
2130 | to use multiple index and data files. |
|
2109 | to use multiple index and data files. | |
2131 | """ |
|
2110 | """ | |
2132 | tiprev = len(self) - 1 |
|
2111 | tiprev = len(self) - 1 | |
2133 | total_size = self.start(tiprev) + self.length(tiprev) |
|
2112 | total_size = self.start(tiprev) + self.length(tiprev) | |
2134 | if not self._inline or total_size < _maxinline: |
|
2113 | if not self._inline or total_size < _maxinline: | |
2135 | return |
|
2114 | return | |
2136 |
|
2115 | |||
2137 | troffset = tr.findoffset(self._indexfile) |
|
2116 | troffset = tr.findoffset(self._indexfile) | |
2138 | if troffset is None: |
|
2117 | if troffset is None: | |
2139 | raise error.RevlogError( |
|
2118 | raise error.RevlogError( | |
2140 | _(b"%s not found in the transaction") % self._indexfile |
|
2119 | _(b"%s not found in the transaction") % self._indexfile | |
2141 | ) |
|
2120 | ) | |
2142 | trindex = 0 |
|
2121 | trindex = 0 | |
2143 | tr.add(self._datafile, 0) |
|
2122 | tr.add(self._datafile, 0) | |
2144 |
|
2123 | |||
2145 | existing_handles = False |
|
2124 | existing_handles = False | |
2146 | if self._writinghandles is not None: |
|
2125 | if self._writinghandles is not None: | |
2147 | existing_handles = True |
|
2126 | existing_handles = True | |
2148 | fp = self._writinghandles[0] |
|
2127 | fp = self._writinghandles[0] | |
2149 | fp.flush() |
|
2128 | fp.flush() | |
2150 | fp.close() |
|
2129 | fp.close() | |
2151 | # We can't use the cached file handle after close(). So prevent |
|
2130 | # We can't use the cached file handle after close(). So prevent | |
2152 | # its usage. |
|
2131 | # its usage. | |
2153 | self._writinghandles = None |
|
2132 | self._writinghandles = None | |
2154 |
|
2133 | |||
2155 | new_dfh = self._datafp(b'w+') |
|
2134 | new_dfh = self._datafp(b'w+') | |
2156 | new_dfh.truncate(0) # drop any potentially existing data |
|
2135 | new_dfh.truncate(0) # drop any potentially existing data | |
2157 | try: |
|
2136 | try: | |
2158 | with self._indexfp() as read_ifh: |
|
2137 | with self._indexfp() as read_ifh: | |
2159 | for r in self: |
|
2138 | for r in self: | |
2160 | new_dfh.write(self._getsegmentforrevs(r, r, df=read_ifh)[1]) |
|
2139 | new_dfh.write(self._getsegmentforrevs(r, r, df=read_ifh)[1]) | |
2161 | if troffset <= self.start(r) + r * self.index.entry_size: |
|
2140 | if troffset <= self.start(r) + r * self.index.entry_size: | |
2162 | trindex = r |
|
2141 | trindex = r | |
2163 | new_dfh.flush() |
|
2142 | new_dfh.flush() | |
2164 |
|
2143 | |||
2165 | with self.__index_new_fp() as fp: |
|
2144 | with self.__index_new_fp() as fp: | |
2166 | self._format_flags &= ~FLAG_INLINE_DATA |
|
2145 | self._format_flags &= ~FLAG_INLINE_DATA | |
2167 | self._inline = False |
|
2146 | self._inline = False | |
2168 | for i in self: |
|
2147 | for i in self: | |
2169 | e = self.index.entry_binary(i) |
|
2148 | e = self.index.entry_binary(i) | |
2170 | if i == 0 and self._docket is None: |
|
2149 | if i == 0 and self._docket is None: | |
2171 | header = self._format_flags | self._format_version |
|
2150 | header = self._format_flags | self._format_version | |
2172 | header = self.index.pack_header(header) |
|
2151 | header = self.index.pack_header(header) | |
2173 | e = header + e |
|
2152 | e = header + e | |
2174 | fp.write(e) |
|
2153 | fp.write(e) | |
2175 | if self._docket is not None: |
|
2154 | if self._docket is not None: | |
2176 | self._docket.index_end = fp.tell() |
|
2155 | self._docket.index_end = fp.tell() | |
2177 |
|
2156 | |||
2178 | # There is a small transactional race here. If the rename of |
|
2157 | # There is a small transactional race here. If the rename of | |
2179 | # the index fails, we should remove the datafile. It is more |
|
2158 | # the index fails, we should remove the datafile. It is more | |
2180 | # important to ensure that the data file is not truncated |
|
2159 | # important to ensure that the data file is not truncated | |
2181 | # when the index is replaced as otherwise data is lost. |
|
2160 | # when the index is replaced as otherwise data is lost. | |
2182 | tr.replace(self._datafile, self.start(trindex)) |
|
2161 | tr.replace(self._datafile, self.start(trindex)) | |
2183 |
|
2162 | |||
2184 | # the temp file replace the real index when we exit the context |
|
2163 | # the temp file replace the real index when we exit the context | |
2185 | # manager |
|
2164 | # manager | |
2186 |
|
2165 | |||
2187 | tr.replace(self._indexfile, trindex * self.index.entry_size) |
|
2166 | tr.replace(self._indexfile, trindex * self.index.entry_size) | |
2188 | nodemaputil.setup_persistent_nodemap(tr, self) |
|
2167 | nodemaputil.setup_persistent_nodemap(tr, self) | |
2189 | self._chunkclear() |
|
2168 | self._chunkclear() | |
2190 |
|
2169 | |||
2191 | if existing_handles: |
|
2170 | if existing_handles: | |
2192 | # switched from inline to conventional reopen the index |
|
2171 | # switched from inline to conventional reopen the index | |
2193 | ifh = self.__index_write_fp() |
|
2172 | ifh = self.__index_write_fp() | |
2194 | self._writinghandles = (ifh, new_dfh, None) |
|
2173 | self._writinghandles = (ifh, new_dfh, None) | |
2195 | new_dfh = None |
|
2174 | new_dfh = None | |
2196 | finally: |
|
2175 | finally: | |
2197 | if new_dfh is not None: |
|
2176 | if new_dfh is not None: | |
2198 | new_dfh.close() |
|
2177 | new_dfh.close() | |
2199 |
|
2178 | |||
2200 | def _nodeduplicatecallback(self, transaction, node): |
|
2179 | def _nodeduplicatecallback(self, transaction, node): | |
2201 | """called when trying to add a node already stored.""" |
|
2180 | """called when trying to add a node already stored.""" | |
2202 |
|
2181 | |||
2203 | @contextlib.contextmanager |
|
2182 | @contextlib.contextmanager | |
2204 | def _writing(self, transaction): |
|
2183 | def _writing(self, transaction): | |
2205 | if self._trypending: |
|
2184 | if self._trypending: | |
2206 | msg = b'try to write in a `trypending` revlog: %s' |
|
2185 | msg = b'try to write in a `trypending` revlog: %s' | |
2207 | msg %= self.display_id |
|
2186 | msg %= self.display_id | |
2208 | raise error.ProgrammingError(msg) |
|
2187 | raise error.ProgrammingError(msg) | |
2209 | if self._writinghandles is not None: |
|
2188 | if self._writinghandles is not None: | |
2210 | yield |
|
2189 | yield | |
2211 | else: |
|
2190 | else: | |
2212 | ifh = dfh = sdfh = None |
|
2191 | ifh = dfh = sdfh = None | |
2213 | try: |
|
2192 | try: | |
2214 | r = len(self) |
|
2193 | r = len(self) | |
2215 | # opening the data file. |
|
2194 | # opening the data file. | |
2216 | dsize = 0 |
|
2195 | dsize = 0 | |
2217 | if r: |
|
2196 | if r: | |
2218 | dsize = self.end(r - 1) |
|
2197 | dsize = self.end(r - 1) | |
2219 | dfh = None |
|
2198 | dfh = None | |
2220 | if not self._inline: |
|
2199 | if not self._inline: | |
2221 | try: |
|
2200 | try: | |
2222 | dfh = self._datafp(b"r+") |
|
2201 | dfh = self._datafp(b"r+") | |
2223 | if self._docket is None: |
|
2202 | if self._docket is None: | |
2224 | dfh.seek(0, os.SEEK_END) |
|
2203 | dfh.seek(0, os.SEEK_END) | |
2225 | else: |
|
2204 | else: | |
2226 | dfh.seek(self._docket.data_end, os.SEEK_SET) |
|
2205 | dfh.seek(self._docket.data_end, os.SEEK_SET) | |
2227 | except IOError as inst: |
|
2206 | except IOError as inst: | |
2228 | if inst.errno != errno.ENOENT: |
|
2207 | if inst.errno != errno.ENOENT: | |
2229 | raise |
|
2208 | raise | |
2230 | dfh = self._datafp(b"w+") |
|
2209 | dfh = self._datafp(b"w+") | |
2231 | transaction.add(self._datafile, dsize) |
|
2210 | transaction.add(self._datafile, dsize) | |
2232 | if self._sidedatafile is not None: |
|
2211 | if self._sidedatafile is not None: | |
2233 | try: |
|
2212 | try: | |
2234 | sdfh = self.opener(self._sidedatafile, mode=b"r+") |
|
2213 | sdfh = self.opener(self._sidedatafile, mode=b"r+") | |
2235 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) |
|
2214 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) | |
2236 | except IOError as inst: |
|
2215 | except IOError as inst: | |
2237 | if inst.errno != errno.ENOENT: |
|
2216 | if inst.errno != errno.ENOENT: | |
2238 | raise |
|
2217 | raise | |
2239 | sdfh = self.opener(self._sidedatafile, mode=b"w+") |
|
2218 | sdfh = self.opener(self._sidedatafile, mode=b"w+") | |
2240 | transaction.add( |
|
2219 | transaction.add( | |
2241 | self._sidedatafile, self._docket.sidedata_end |
|
2220 | self._sidedatafile, self._docket.sidedata_end | |
2242 | ) |
|
2221 | ) | |
2243 |
|
2222 | |||
2244 | # opening the index file. |
|
2223 | # opening the index file. | |
2245 | isize = r * self.index.entry_size |
|
2224 | isize = r * self.index.entry_size | |
2246 | ifh = self.__index_write_fp() |
|
2225 | ifh = self.__index_write_fp() | |
2247 | if self._inline: |
|
2226 | if self._inline: | |
2248 | transaction.add(self._indexfile, dsize + isize) |
|
2227 | transaction.add(self._indexfile, dsize + isize) | |
2249 | else: |
|
2228 | else: | |
2250 | transaction.add(self._indexfile, isize) |
|
2229 | transaction.add(self._indexfile, isize) | |
2251 | # exposing all file handle for writing. |
|
2230 | # exposing all file handle for writing. | |
2252 | self._writinghandles = (ifh, dfh, sdfh) |
|
2231 | self._writinghandles = (ifh, dfh, sdfh) | |
2253 | yield |
|
2232 | yield | |
2254 | if self._docket is not None: |
|
2233 | if self._docket is not None: | |
2255 | self._write_docket(transaction) |
|
2234 | self._write_docket(transaction) | |
2256 | finally: |
|
2235 | finally: | |
2257 | self._writinghandles = None |
|
2236 | self._writinghandles = None | |
2258 | if dfh is not None: |
|
2237 | if dfh is not None: | |
2259 | dfh.close() |
|
2238 | dfh.close() | |
2260 | if sdfh is not None: |
|
2239 | if sdfh is not None: | |
2261 | dfh.close() |
|
2240 | dfh.close() | |
2262 | # closing the index file last to avoid exposing referent to |
|
2241 | # closing the index file last to avoid exposing referent to | |
2263 | # potential unflushed data content. |
|
2242 | # potential unflushed data content. | |
2264 | if ifh is not None: |
|
2243 | if ifh is not None: | |
2265 | ifh.close() |
|
2244 | ifh.close() | |
2266 |
|
2245 | |||
2267 | def _write_docket(self, transaction): |
|
2246 | def _write_docket(self, transaction): | |
2268 | """write the current docket on disk |
|
2247 | """write the current docket on disk | |
2269 |
|
2248 | |||
2270 | Exist as a method to help changelog to implement transaction logic |
|
2249 | Exist as a method to help changelog to implement transaction logic | |
2271 |
|
2250 | |||
2272 | We could also imagine using the same transaction logic for all revlog |
|
2251 | We could also imagine using the same transaction logic for all revlog | |
2273 | since docket are cheap.""" |
|
2252 | since docket are cheap.""" | |
2274 | self._docket.write(transaction) |
|
2253 | self._docket.write(transaction) | |
2275 |
|
2254 | |||
2276 | def addrevision( |
|
2255 | def addrevision( | |
2277 | self, |
|
2256 | self, | |
2278 | text, |
|
2257 | text, | |
2279 | transaction, |
|
2258 | transaction, | |
2280 | link, |
|
2259 | link, | |
2281 | p1, |
|
2260 | p1, | |
2282 | p2, |
|
2261 | p2, | |
2283 | cachedelta=None, |
|
2262 | cachedelta=None, | |
2284 | node=None, |
|
2263 | node=None, | |
2285 | flags=REVIDX_DEFAULT_FLAGS, |
|
2264 | flags=REVIDX_DEFAULT_FLAGS, | |
2286 | deltacomputer=None, |
|
2265 | deltacomputer=None, | |
2287 | sidedata=None, |
|
2266 | sidedata=None, | |
2288 | ): |
|
2267 | ): | |
2289 | """add a revision to the log |
|
2268 | """add a revision to the log | |
2290 |
|
2269 | |||
2291 | text - the revision data to add |
|
2270 | text - the revision data to add | |
2292 | transaction - the transaction object used for rollback |
|
2271 | transaction - the transaction object used for rollback | |
2293 | link - the linkrev data to add |
|
2272 | link - the linkrev data to add | |
2294 | p1, p2 - the parent nodeids of the revision |
|
2273 | p1, p2 - the parent nodeids of the revision | |
2295 | cachedelta - an optional precomputed delta |
|
2274 | cachedelta - an optional precomputed delta | |
2296 | node - nodeid of revision; typically node is not specified, and it is |
|
2275 | node - nodeid of revision; typically node is not specified, and it is | |
2297 | computed by default as hash(text, p1, p2), however subclasses might |
|
2276 | computed by default as hash(text, p1, p2), however subclasses might | |
2298 | use different hashing method (and override checkhash() in such case) |
|
2277 | use different hashing method (and override checkhash() in such case) | |
2299 | flags - the known flags to set on the revision |
|
2278 | flags - the known flags to set on the revision | |
2300 | deltacomputer - an optional deltacomputer instance shared between |
|
2279 | deltacomputer - an optional deltacomputer instance shared between | |
2301 | multiple calls |
|
2280 | multiple calls | |
2302 | """ |
|
2281 | """ | |
2303 | if link == nullrev: |
|
2282 | if link == nullrev: | |
2304 | raise error.RevlogError( |
|
2283 | raise error.RevlogError( | |
2305 | _(b"attempted to add linkrev -1 to %s") % self.display_id |
|
2284 | _(b"attempted to add linkrev -1 to %s") % self.display_id | |
2306 | ) |
|
2285 | ) | |
2307 |
|
2286 | |||
2308 | if sidedata is None: |
|
2287 | if sidedata is None: | |
2309 | sidedata = {} |
|
2288 | sidedata = {} | |
2310 | elif sidedata and not self.hassidedata: |
|
2289 | elif sidedata and not self.hassidedata: | |
2311 | raise error.ProgrammingError( |
|
2290 | raise error.ProgrammingError( | |
2312 | _(b"trying to add sidedata to a revlog who don't support them") |
|
2291 | _(b"trying to add sidedata to a revlog who don't support them") | |
2313 | ) |
|
2292 | ) | |
2314 |
|
2293 | |||
2315 | if flags: |
|
2294 | if flags: | |
2316 | node = node or self.hash(text, p1, p2) |
|
2295 | node = node or self.hash(text, p1, p2) | |
2317 |
|
2296 | |||
2318 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) |
|
2297 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) | |
2319 |
|
2298 | |||
2320 | # If the flag processor modifies the revision data, ignore any provided |
|
2299 | # If the flag processor modifies the revision data, ignore any provided | |
2321 | # cachedelta. |
|
2300 | # cachedelta. | |
2322 | if rawtext != text: |
|
2301 | if rawtext != text: | |
2323 | cachedelta = None |
|
2302 | cachedelta = None | |
2324 |
|
2303 | |||
2325 | if len(rawtext) > _maxentrysize: |
|
2304 | if len(rawtext) > _maxentrysize: | |
2326 | raise error.RevlogError( |
|
2305 | raise error.RevlogError( | |
2327 | _( |
|
2306 | _( | |
2328 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" |
|
2307 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" | |
2329 | ) |
|
2308 | ) | |
2330 | % (self.display_id, len(rawtext)) |
|
2309 | % (self.display_id, len(rawtext)) | |
2331 | ) |
|
2310 | ) | |
2332 |
|
2311 | |||
2333 | node = node or self.hash(rawtext, p1, p2) |
|
2312 | node = node or self.hash(rawtext, p1, p2) | |
2334 | rev = self.index.get_rev(node) |
|
2313 | rev = self.index.get_rev(node) | |
2335 | if rev is not None: |
|
2314 | if rev is not None: | |
2336 | return rev |
|
2315 | return rev | |
2337 |
|
2316 | |||
2338 | if validatehash: |
|
2317 | if validatehash: | |
2339 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
2318 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
2340 |
|
2319 | |||
2341 | return self.addrawrevision( |
|
2320 | return self.addrawrevision( | |
2342 | rawtext, |
|
2321 | rawtext, | |
2343 | transaction, |
|
2322 | transaction, | |
2344 | link, |
|
2323 | link, | |
2345 | p1, |
|
2324 | p1, | |
2346 | p2, |
|
2325 | p2, | |
2347 | node, |
|
2326 | node, | |
2348 | flags, |
|
2327 | flags, | |
2349 | cachedelta=cachedelta, |
|
2328 | cachedelta=cachedelta, | |
2350 | deltacomputer=deltacomputer, |
|
2329 | deltacomputer=deltacomputer, | |
2351 | sidedata=sidedata, |
|
2330 | sidedata=sidedata, | |
2352 | ) |
|
2331 | ) | |
2353 |
|
2332 | |||
2354 | def addrawrevision( |
|
2333 | def addrawrevision( | |
2355 | self, |
|
2334 | self, | |
2356 | rawtext, |
|
2335 | rawtext, | |
2357 | transaction, |
|
2336 | transaction, | |
2358 | link, |
|
2337 | link, | |
2359 | p1, |
|
2338 | p1, | |
2360 | p2, |
|
2339 | p2, | |
2361 | node, |
|
2340 | node, | |
2362 | flags, |
|
2341 | flags, | |
2363 | cachedelta=None, |
|
2342 | cachedelta=None, | |
2364 | deltacomputer=None, |
|
2343 | deltacomputer=None, | |
2365 | sidedata=None, |
|
2344 | sidedata=None, | |
2366 | ): |
|
2345 | ): | |
2367 | """add a raw revision with known flags, node and parents |
|
2346 | """add a raw revision with known flags, node and parents | |
2368 | useful when reusing a revision not stored in this revlog (ex: received |
|
2347 | useful when reusing a revision not stored in this revlog (ex: received | |
2369 | over wire, or read from an external bundle). |
|
2348 | over wire, or read from an external bundle). | |
2370 | """ |
|
2349 | """ | |
2371 | with self._writing(transaction): |
|
2350 | with self._writing(transaction): | |
2372 | return self._addrevision( |
|
2351 | return self._addrevision( | |
2373 | node, |
|
2352 | node, | |
2374 | rawtext, |
|
2353 | rawtext, | |
2375 | transaction, |
|
2354 | transaction, | |
2376 | link, |
|
2355 | link, | |
2377 | p1, |
|
2356 | p1, | |
2378 | p2, |
|
2357 | p2, | |
2379 | flags, |
|
2358 | flags, | |
2380 | cachedelta, |
|
2359 | cachedelta, | |
2381 | deltacomputer=deltacomputer, |
|
2360 | deltacomputer=deltacomputer, | |
2382 | sidedata=sidedata, |
|
2361 | sidedata=sidedata, | |
2383 | ) |
|
2362 | ) | |
2384 |
|
2363 | |||
2385 | def compress(self, data): |
|
2364 | def compress(self, data): | |
2386 | """Generate a possibly-compressed representation of data.""" |
|
2365 | """Generate a possibly-compressed representation of data.""" | |
2387 | if not data: |
|
2366 | if not data: | |
2388 | return b'', data |
|
2367 | return b'', data | |
2389 |
|
2368 | |||
2390 | compressed = self._compressor.compress(data) |
|
2369 | compressed = self._compressor.compress(data) | |
2391 |
|
2370 | |||
2392 | if compressed: |
|
2371 | if compressed: | |
2393 | # The revlog compressor added the header in the returned data. |
|
2372 | # The revlog compressor added the header in the returned data. | |
2394 | return b'', compressed |
|
2373 | return b'', compressed | |
2395 |
|
2374 | |||
2396 | if data[0:1] == b'\0': |
|
2375 | if data[0:1] == b'\0': | |
2397 | return b'', data |
|
2376 | return b'', data | |
2398 | return b'u', data |
|
2377 | return b'u', data | |
2399 |
|
2378 | |||
2400 | def decompress(self, data): |
|
2379 | def decompress(self, data): | |
2401 | """Decompress a revlog chunk. |
|
2380 | """Decompress a revlog chunk. | |
2402 |
|
2381 | |||
2403 | The chunk is expected to begin with a header identifying the |
|
2382 | The chunk is expected to begin with a header identifying the | |
2404 | format type so it can be routed to an appropriate decompressor. |
|
2383 | format type so it can be routed to an appropriate decompressor. | |
2405 | """ |
|
2384 | """ | |
2406 | if not data: |
|
2385 | if not data: | |
2407 | return data |
|
2386 | return data | |
2408 |
|
2387 | |||
2409 | # Revlogs are read much more frequently than they are written and many |
|
2388 | # Revlogs are read much more frequently than they are written and many | |
2410 | # chunks only take microseconds to decompress, so performance is |
|
2389 | # chunks only take microseconds to decompress, so performance is | |
2411 | # important here. |
|
2390 | # important here. | |
2412 | # |
|
2391 | # | |
2413 | # We can make a few assumptions about revlogs: |
|
2392 | # We can make a few assumptions about revlogs: | |
2414 | # |
|
2393 | # | |
2415 | # 1) the majority of chunks will be compressed (as opposed to inline |
|
2394 | # 1) the majority of chunks will be compressed (as opposed to inline | |
2416 | # raw data). |
|
2395 | # raw data). | |
2417 | # 2) decompressing *any* data will likely by at least 10x slower than |
|
2396 | # 2) decompressing *any* data will likely by at least 10x slower than | |
2418 | # returning raw inline data. |
|
2397 | # returning raw inline data. | |
2419 | # 3) we want to prioritize common and officially supported compression |
|
2398 | # 3) we want to prioritize common and officially supported compression | |
2420 | # engines |
|
2399 | # engines | |
2421 | # |
|
2400 | # | |
2422 | # It follows that we want to optimize for "decompress compressed data |
|
2401 | # It follows that we want to optimize for "decompress compressed data | |
2423 | # when encoded with common and officially supported compression engines" |
|
2402 | # when encoded with common and officially supported compression engines" | |
2424 | # case over "raw data" and "data encoded by less common or non-official |
|
2403 | # case over "raw data" and "data encoded by less common or non-official | |
2425 | # compression engines." That is why we have the inline lookup first |
|
2404 | # compression engines." That is why we have the inline lookup first | |
2426 | # followed by the compengines lookup. |
|
2405 | # followed by the compengines lookup. | |
2427 | # |
|
2406 | # | |
2428 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib |
|
2407 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib | |
2429 | # compressed chunks. And this matters for changelog and manifest reads. |
|
2408 | # compressed chunks. And this matters for changelog and manifest reads. | |
2430 | t = data[0:1] |
|
2409 | t = data[0:1] | |
2431 |
|
2410 | |||
2432 | if t == b'x': |
|
2411 | if t == b'x': | |
2433 | try: |
|
2412 | try: | |
2434 | return _zlibdecompress(data) |
|
2413 | return _zlibdecompress(data) | |
2435 | except zlib.error as e: |
|
2414 | except zlib.error as e: | |
2436 | raise error.RevlogError( |
|
2415 | raise error.RevlogError( | |
2437 | _(b'revlog decompress error: %s') |
|
2416 | _(b'revlog decompress error: %s') | |
2438 | % stringutil.forcebytestr(e) |
|
2417 | % stringutil.forcebytestr(e) | |
2439 | ) |
|
2418 | ) | |
2440 | # '\0' is more common than 'u' so it goes first. |
|
2419 | # '\0' is more common than 'u' so it goes first. | |
2441 | elif t == b'\0': |
|
2420 | elif t == b'\0': | |
2442 | return data |
|
2421 | return data | |
2443 | elif t == b'u': |
|
2422 | elif t == b'u': | |
2444 | return util.buffer(data, 1) |
|
2423 | return util.buffer(data, 1) | |
2445 |
|
2424 | |||
2446 | compressor = self._get_decompressor(t) |
|
2425 | compressor = self._get_decompressor(t) | |
2447 |
|
2426 | |||
2448 | return compressor.decompress(data) |
|
2427 | return compressor.decompress(data) | |
2449 |
|
2428 | |||
2450 | def _addrevision( |
|
2429 | def _addrevision( | |
2451 | self, |
|
2430 | self, | |
2452 | node, |
|
2431 | node, | |
2453 | rawtext, |
|
2432 | rawtext, | |
2454 | transaction, |
|
2433 | transaction, | |
2455 | link, |
|
2434 | link, | |
2456 | p1, |
|
2435 | p1, | |
2457 | p2, |
|
2436 | p2, | |
2458 | flags, |
|
2437 | flags, | |
2459 | cachedelta, |
|
2438 | cachedelta, | |
2460 | alwayscache=False, |
|
2439 | alwayscache=False, | |
2461 | deltacomputer=None, |
|
2440 | deltacomputer=None, | |
2462 | sidedata=None, |
|
2441 | sidedata=None, | |
2463 | ): |
|
2442 | ): | |
2464 | """internal function to add revisions to the log |
|
2443 | """internal function to add revisions to the log | |
2465 |
|
2444 | |||
2466 | see addrevision for argument descriptions. |
|
2445 | see addrevision for argument descriptions. | |
2467 |
|
2446 | |||
2468 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. |
|
2447 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. | |
2469 |
|
2448 | |||
2470 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will |
|
2449 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will | |
2471 | be used. |
|
2450 | be used. | |
2472 |
|
2451 | |||
2473 | invariants: |
|
2452 | invariants: | |
2474 | - rawtext is optional (can be None); if not set, cachedelta must be set. |
|
2453 | - rawtext is optional (can be None); if not set, cachedelta must be set. | |
2475 | if both are set, they must correspond to each other. |
|
2454 | if both are set, they must correspond to each other. | |
2476 | """ |
|
2455 | """ | |
2477 | if node == self.nullid: |
|
2456 | if node == self.nullid: | |
2478 | raise error.RevlogError( |
|
2457 | raise error.RevlogError( | |
2479 | _(b"%s: attempt to add null revision") % self.display_id |
|
2458 | _(b"%s: attempt to add null revision") % self.display_id | |
2480 | ) |
|
2459 | ) | |
2481 | if ( |
|
2460 | if ( | |
2482 | node == self.nodeconstants.wdirid |
|
2461 | node == self.nodeconstants.wdirid | |
2483 | or node in self.nodeconstants.wdirfilenodeids |
|
2462 | or node in self.nodeconstants.wdirfilenodeids | |
2484 | ): |
|
2463 | ): | |
2485 | raise error.RevlogError( |
|
2464 | raise error.RevlogError( | |
2486 | _(b"%s: attempt to add wdir revision") % self.display_id |
|
2465 | _(b"%s: attempt to add wdir revision") % self.display_id | |
2487 | ) |
|
2466 | ) | |
2488 | if self._writinghandles is None: |
|
2467 | if self._writinghandles is None: | |
2489 | msg = b'adding revision outside `revlog._writing` context' |
|
2468 | msg = b'adding revision outside `revlog._writing` context' | |
2490 | raise error.ProgrammingError(msg) |
|
2469 | raise error.ProgrammingError(msg) | |
2491 |
|
2470 | |||
2492 | if self._inline: |
|
2471 | if self._inline: | |
2493 | fh = self._writinghandles[0] |
|
2472 | fh = self._writinghandles[0] | |
2494 | else: |
|
2473 | else: | |
2495 | fh = self._writinghandles[1] |
|
2474 | fh = self._writinghandles[1] | |
2496 |
|
2475 | |||
2497 | btext = [rawtext] |
|
2476 | btext = [rawtext] | |
2498 |
|
2477 | |||
2499 | curr = len(self) |
|
2478 | curr = len(self) | |
2500 | prev = curr - 1 |
|
2479 | prev = curr - 1 | |
2501 |
|
2480 | |||
2502 | offset = self._get_data_offset(prev) |
|
2481 | offset = self._get_data_offset(prev) | |
2503 |
|
2482 | |||
2504 | if self._concurrencychecker: |
|
2483 | if self._concurrencychecker: | |
2505 | ifh, dfh, sdfh = self._writinghandles |
|
2484 | ifh, dfh, sdfh = self._writinghandles | |
2506 | # XXX no checking for the sidedata file |
|
2485 | # XXX no checking for the sidedata file | |
2507 | if self._inline: |
|
2486 | if self._inline: | |
2508 | # offset is "as if" it were in the .d file, so we need to add on |
|
2487 | # offset is "as if" it were in the .d file, so we need to add on | |
2509 | # the size of the entry metadata. |
|
2488 | # the size of the entry metadata. | |
2510 | self._concurrencychecker( |
|
2489 | self._concurrencychecker( | |
2511 | ifh, self._indexfile, offset + curr * self.index.entry_size |
|
2490 | ifh, self._indexfile, offset + curr * self.index.entry_size | |
2512 | ) |
|
2491 | ) | |
2513 | else: |
|
2492 | else: | |
2514 | # Entries in the .i are a consistent size. |
|
2493 | # Entries in the .i are a consistent size. | |
2515 | self._concurrencychecker( |
|
2494 | self._concurrencychecker( | |
2516 | ifh, self._indexfile, curr * self.index.entry_size |
|
2495 | ifh, self._indexfile, curr * self.index.entry_size | |
2517 | ) |
|
2496 | ) | |
2518 | self._concurrencychecker(dfh, self._datafile, offset) |
|
2497 | self._concurrencychecker(dfh, self._datafile, offset) | |
2519 |
|
2498 | |||
2520 | p1r, p2r = self.rev(p1), self.rev(p2) |
|
2499 | p1r, p2r = self.rev(p1), self.rev(p2) | |
2521 |
|
2500 | |||
2522 | # full versions are inserted when the needed deltas |
|
2501 | # full versions are inserted when the needed deltas | |
2523 | # become comparable to the uncompressed text |
|
2502 | # become comparable to the uncompressed text | |
2524 | if rawtext is None: |
|
2503 | if rawtext is None: | |
2525 | # need rawtext size, before changed by flag processors, which is |
|
2504 | # need rawtext size, before changed by flag processors, which is | |
2526 | # the non-raw size. use revlog explicitly to avoid filelog's extra |
|
2505 | # the non-raw size. use revlog explicitly to avoid filelog's extra | |
2527 | # logic that might remove metadata size. |
|
2506 | # logic that might remove metadata size. | |
2528 | textlen = mdiff.patchedsize( |
|
2507 | textlen = mdiff.patchedsize( | |
2529 | revlog.size(self, cachedelta[0]), cachedelta[1] |
|
2508 | revlog.size(self, cachedelta[0]), cachedelta[1] | |
2530 | ) |
|
2509 | ) | |
2531 | else: |
|
2510 | else: | |
2532 | textlen = len(rawtext) |
|
2511 | textlen = len(rawtext) | |
2533 |
|
2512 | |||
2534 | if deltacomputer is None: |
|
2513 | if deltacomputer is None: | |
2535 | deltacomputer = deltautil.deltacomputer(self) |
|
2514 | deltacomputer = deltautil.deltacomputer(self) | |
2536 |
|
2515 | |||
2537 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) |
|
2516 | revinfo = revlogutils.revisioninfo( | |
|
2517 | node, | |||
|
2518 | p1, | |||
|
2519 | p2, | |||
|
2520 | btext, | |||
|
2521 | textlen, | |||
|
2522 | cachedelta, | |||
|
2523 | flags, | |||
|
2524 | ) | |||
2538 |
|
2525 | |||
2539 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) |
|
2526 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) | |
2540 |
|
2527 | |||
2541 | compression_mode = COMP_MODE_INLINE |
|
2528 | compression_mode = COMP_MODE_INLINE | |
2542 | if self._docket is not None: |
|
2529 | if self._docket is not None: | |
2543 | h, d = deltainfo.data |
|
2530 | h, d = deltainfo.data | |
2544 | if not h and not d: |
|
2531 | if not h and not d: | |
2545 | # not data to store at all... declare them uncompressed |
|
2532 | # not data to store at all... declare them uncompressed | |
2546 | compression_mode = COMP_MODE_PLAIN |
|
2533 | compression_mode = COMP_MODE_PLAIN | |
2547 | elif not h: |
|
2534 | elif not h: | |
2548 | t = d[0:1] |
|
2535 | t = d[0:1] | |
2549 | if t == b'\0': |
|
2536 | if t == b'\0': | |
2550 | compression_mode = COMP_MODE_PLAIN |
|
2537 | compression_mode = COMP_MODE_PLAIN | |
2551 | elif t == self._docket.default_compression_header: |
|
2538 | elif t == self._docket.default_compression_header: | |
2552 | compression_mode = COMP_MODE_DEFAULT |
|
2539 | compression_mode = COMP_MODE_DEFAULT | |
2553 | elif h == b'u': |
|
2540 | elif h == b'u': | |
2554 | # we have a more efficient way to declare uncompressed |
|
2541 | # we have a more efficient way to declare uncompressed | |
2555 | h = b'' |
|
2542 | h = b'' | |
2556 | compression_mode = COMP_MODE_PLAIN |
|
2543 | compression_mode = COMP_MODE_PLAIN | |
2557 | deltainfo = deltautil.drop_u_compression(deltainfo) |
|
2544 | deltainfo = deltautil.drop_u_compression(deltainfo) | |
2558 |
|
2545 | |||
2559 | sidedata_compression_mode = COMP_MODE_INLINE |
|
2546 | sidedata_compression_mode = COMP_MODE_INLINE | |
2560 | if sidedata and self.hassidedata: |
|
2547 | if sidedata and self.hassidedata: | |
2561 | sidedata_compression_mode = COMP_MODE_PLAIN |
|
2548 | sidedata_compression_mode = COMP_MODE_PLAIN | |
2562 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) |
|
2549 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) | |
2563 | sidedata_offset = self._docket.sidedata_end |
|
2550 | sidedata_offset = self._docket.sidedata_end | |
2564 | h, comp_sidedata = self.compress(serialized_sidedata) |
|
2551 | h, comp_sidedata = self.compress(serialized_sidedata) | |
2565 | if ( |
|
2552 | if ( | |
2566 | h != b'u' |
|
2553 | h != b'u' | |
2567 | and comp_sidedata[0:1] != b'\0' |
|
2554 | and comp_sidedata[0:1] != b'\0' | |
2568 | and len(comp_sidedata) < len(serialized_sidedata) |
|
2555 | and len(comp_sidedata) < len(serialized_sidedata) | |
2569 | ): |
|
2556 | ): | |
2570 | assert not h |
|
2557 | assert not h | |
2571 | if ( |
|
2558 | if ( | |
2572 | comp_sidedata[0:1] |
|
2559 | comp_sidedata[0:1] | |
2573 | == self._docket.default_compression_header |
|
2560 | == self._docket.default_compression_header | |
2574 | ): |
|
2561 | ): | |
2575 | sidedata_compression_mode = COMP_MODE_DEFAULT |
|
2562 | sidedata_compression_mode = COMP_MODE_DEFAULT | |
2576 | serialized_sidedata = comp_sidedata |
|
2563 | serialized_sidedata = comp_sidedata | |
2577 | else: |
|
2564 | else: | |
2578 | sidedata_compression_mode = COMP_MODE_INLINE |
|
2565 | sidedata_compression_mode = COMP_MODE_INLINE | |
2579 | serialized_sidedata = comp_sidedata |
|
2566 | serialized_sidedata = comp_sidedata | |
2580 | else: |
|
2567 | else: | |
2581 | serialized_sidedata = b"" |
|
2568 | serialized_sidedata = b"" | |
2582 | # Don't store the offset if the sidedata is empty, that way |
|
2569 | # Don't store the offset if the sidedata is empty, that way | |
2583 | # we can easily detect empty sidedata and they will be no different |
|
2570 | # we can easily detect empty sidedata and they will be no different | |
2584 | # than ones we manually add. |
|
2571 | # than ones we manually add. | |
2585 | sidedata_offset = 0 |
|
2572 | sidedata_offset = 0 | |
2586 |
|
2573 | |||
2587 | e = revlogutils.entry( |
|
2574 | e = revlogutils.entry( | |
2588 | flags=flags, |
|
2575 | flags=flags, | |
2589 | data_offset=offset, |
|
2576 | data_offset=offset, | |
2590 | data_compressed_length=deltainfo.deltalen, |
|
2577 | data_compressed_length=deltainfo.deltalen, | |
2591 | data_uncompressed_length=textlen, |
|
2578 | data_uncompressed_length=textlen, | |
2592 | data_compression_mode=compression_mode, |
|
2579 | data_compression_mode=compression_mode, | |
2593 | data_delta_base=deltainfo.base, |
|
2580 | data_delta_base=deltainfo.base, | |
2594 | link_rev=link, |
|
2581 | link_rev=link, | |
2595 | parent_rev_1=p1r, |
|
2582 | parent_rev_1=p1r, | |
2596 | parent_rev_2=p2r, |
|
2583 | parent_rev_2=p2r, | |
2597 | node_id=node, |
|
2584 | node_id=node, | |
2598 | sidedata_offset=sidedata_offset, |
|
2585 | sidedata_offset=sidedata_offset, | |
2599 | sidedata_compressed_length=len(serialized_sidedata), |
|
2586 | sidedata_compressed_length=len(serialized_sidedata), | |
2600 | sidedata_compression_mode=sidedata_compression_mode, |
|
2587 | sidedata_compression_mode=sidedata_compression_mode, | |
2601 | ) |
|
2588 | ) | |
2602 |
|
2589 | |||
2603 | self.index.append(e) |
|
2590 | self.index.append(e) | |
2604 | entry = self.index.entry_binary(curr) |
|
2591 | entry = self.index.entry_binary(curr) | |
2605 | if curr == 0 and self._docket is None: |
|
2592 | if curr == 0 and self._docket is None: | |
2606 | header = self._format_flags | self._format_version |
|
2593 | header = self._format_flags | self._format_version | |
2607 | header = self.index.pack_header(header) |
|
2594 | header = self.index.pack_header(header) | |
2608 | entry = header + entry |
|
2595 | entry = header + entry | |
2609 | self._writeentry( |
|
2596 | self._writeentry( | |
2610 | transaction, |
|
2597 | transaction, | |
2611 | entry, |
|
2598 | entry, | |
2612 | deltainfo.data, |
|
2599 | deltainfo.data, | |
2613 | link, |
|
2600 | link, | |
2614 | offset, |
|
2601 | offset, | |
2615 | serialized_sidedata, |
|
2602 | serialized_sidedata, | |
2616 | sidedata_offset, |
|
2603 | sidedata_offset, | |
2617 | ) |
|
2604 | ) | |
2618 |
|
2605 | |||
2619 | rawtext = btext[0] |
|
2606 | rawtext = btext[0] | |
2620 |
|
2607 | |||
2621 | if alwayscache and rawtext is None: |
|
2608 | if alwayscache and rawtext is None: | |
2622 | rawtext = deltacomputer.buildtext(revinfo, fh) |
|
2609 | rawtext = deltacomputer.buildtext(revinfo, fh) | |
2623 |
|
2610 | |||
2624 | if type(rawtext) == bytes: # only accept immutable objects |
|
2611 | if type(rawtext) == bytes: # only accept immutable objects | |
2625 | self._revisioncache = (node, curr, rawtext) |
|
2612 | self._revisioncache = (node, curr, rawtext) | |
2626 | self._chainbasecache[curr] = deltainfo.chainbase |
|
2613 | self._chainbasecache[curr] = deltainfo.chainbase | |
2627 | return curr |
|
2614 | return curr | |
2628 |
|
2615 | |||
2629 | def _get_data_offset(self, prev): |
|
2616 | def _get_data_offset(self, prev): | |
2630 | """Returns the current offset in the (in-transaction) data file. |
|
2617 | """Returns the current offset in the (in-transaction) data file. | |
2631 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket |
|
2618 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket | |
2632 | file to store that information: since sidedata can be rewritten to the |
|
2619 | file to store that information: since sidedata can be rewritten to the | |
2633 | end of the data file within a transaction, you can have cases where, for |
|
2620 | end of the data file within a transaction, you can have cases where, for | |
2634 | example, rev `n` does not have sidedata while rev `n - 1` does, leading |
|
2621 | example, rev `n` does not have sidedata while rev `n - 1` does, leading | |
2635 | to `n - 1`'s sidedata being written after `n`'s data. |
|
2622 | to `n - 1`'s sidedata being written after `n`'s data. | |
2636 |
|
2623 | |||
2637 | TODO cache this in a docket file before getting out of experimental.""" |
|
2624 | TODO cache this in a docket file before getting out of experimental.""" | |
2638 | if self._docket is None: |
|
2625 | if self._docket is None: | |
2639 | return self.end(prev) |
|
2626 | return self.end(prev) | |
2640 | else: |
|
2627 | else: | |
2641 | return self._docket.data_end |
|
2628 | return self._docket.data_end | |
2642 |
|
2629 | |||
2643 | def _writeentry( |
|
2630 | def _writeentry( | |
2644 | self, transaction, entry, data, link, offset, sidedata, sidedata_offset |
|
2631 | self, transaction, entry, data, link, offset, sidedata, sidedata_offset | |
2645 | ): |
|
2632 | ): | |
2646 | # Files opened in a+ mode have inconsistent behavior on various |
|
2633 | # Files opened in a+ mode have inconsistent behavior on various | |
2647 | # platforms. Windows requires that a file positioning call be made |
|
2634 | # platforms. Windows requires that a file positioning call be made | |
2648 | # when the file handle transitions between reads and writes. See |
|
2635 | # when the file handle transitions between reads and writes. See | |
2649 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
2636 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other | |
2650 | # platforms, Python or the platform itself can be buggy. Some versions |
|
2637 | # platforms, Python or the platform itself can be buggy. Some versions | |
2651 | # of Solaris have been observed to not append at the end of the file |
|
2638 | # of Solaris have been observed to not append at the end of the file | |
2652 | # if the file was seeked to before the end. See issue4943 for more. |
|
2639 | # if the file was seeked to before the end. See issue4943 for more. | |
2653 | # |
|
2640 | # | |
2654 | # We work around this issue by inserting a seek() before writing. |
|
2641 | # We work around this issue by inserting a seek() before writing. | |
2655 | # Note: This is likely not necessary on Python 3. However, because |
|
2642 | # Note: This is likely not necessary on Python 3. However, because | |
2656 | # the file handle is reused for reads and may be seeked there, we need |
|
2643 | # the file handle is reused for reads and may be seeked there, we need | |
2657 | # to be careful before changing this. |
|
2644 | # to be careful before changing this. | |
2658 | if self._writinghandles is None: |
|
2645 | if self._writinghandles is None: | |
2659 | msg = b'adding revision outside `revlog._writing` context' |
|
2646 | msg = b'adding revision outside `revlog._writing` context' | |
2660 | raise error.ProgrammingError(msg) |
|
2647 | raise error.ProgrammingError(msg) | |
2661 | ifh, dfh, sdfh = self._writinghandles |
|
2648 | ifh, dfh, sdfh = self._writinghandles | |
2662 | if self._docket is None: |
|
2649 | if self._docket is None: | |
2663 | ifh.seek(0, os.SEEK_END) |
|
2650 | ifh.seek(0, os.SEEK_END) | |
2664 | else: |
|
2651 | else: | |
2665 | ifh.seek(self._docket.index_end, os.SEEK_SET) |
|
2652 | ifh.seek(self._docket.index_end, os.SEEK_SET) | |
2666 | if dfh: |
|
2653 | if dfh: | |
2667 | if self._docket is None: |
|
2654 | if self._docket is None: | |
2668 | dfh.seek(0, os.SEEK_END) |
|
2655 | dfh.seek(0, os.SEEK_END) | |
2669 | else: |
|
2656 | else: | |
2670 | dfh.seek(self._docket.data_end, os.SEEK_SET) |
|
2657 | dfh.seek(self._docket.data_end, os.SEEK_SET) | |
2671 | if sdfh: |
|
2658 | if sdfh: | |
2672 | sdfh.seek(self._docket.sidedata_end, os.SEEK_SET) |
|
2659 | sdfh.seek(self._docket.sidedata_end, os.SEEK_SET) | |
2673 |
|
2660 | |||
2674 | curr = len(self) - 1 |
|
2661 | curr = len(self) - 1 | |
2675 | if not self._inline: |
|
2662 | if not self._inline: | |
2676 | transaction.add(self._datafile, offset) |
|
2663 | transaction.add(self._datafile, offset) | |
2677 | if self._sidedatafile: |
|
2664 | if self._sidedatafile: | |
2678 | transaction.add(self._sidedatafile, sidedata_offset) |
|
2665 | transaction.add(self._sidedatafile, sidedata_offset) | |
2679 | transaction.add(self._indexfile, curr * len(entry)) |
|
2666 | transaction.add(self._indexfile, curr * len(entry)) | |
2680 | if data[0]: |
|
2667 | if data[0]: | |
2681 | dfh.write(data[0]) |
|
2668 | dfh.write(data[0]) | |
2682 | dfh.write(data[1]) |
|
2669 | dfh.write(data[1]) | |
2683 | if sidedata: |
|
2670 | if sidedata: | |
2684 | sdfh.write(sidedata) |
|
2671 | sdfh.write(sidedata) | |
2685 | ifh.write(entry) |
|
2672 | ifh.write(entry) | |
2686 | else: |
|
2673 | else: | |
2687 | offset += curr * self.index.entry_size |
|
2674 | offset += curr * self.index.entry_size | |
2688 | transaction.add(self._indexfile, offset) |
|
2675 | transaction.add(self._indexfile, offset) | |
2689 | ifh.write(entry) |
|
2676 | ifh.write(entry) | |
2690 | ifh.write(data[0]) |
|
2677 | ifh.write(data[0]) | |
2691 | ifh.write(data[1]) |
|
2678 | ifh.write(data[1]) | |
2692 | assert not sidedata |
|
2679 | assert not sidedata | |
2693 | self._enforceinlinesize(transaction) |
|
2680 | self._enforceinlinesize(transaction) | |
2694 | if self._docket is not None: |
|
2681 | if self._docket is not None: | |
2695 | self._docket.index_end = self._writinghandles[0].tell() |
|
2682 | self._docket.index_end = self._writinghandles[0].tell() | |
2696 | self._docket.data_end = self._writinghandles[1].tell() |
|
2683 | self._docket.data_end = self._writinghandles[1].tell() | |
2697 | self._docket.sidedata_end = self._writinghandles[2].tell() |
|
2684 | self._docket.sidedata_end = self._writinghandles[2].tell() | |
2698 |
|
2685 | |||
2699 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
2686 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
2700 |
|
2687 | |||
2701 | def addgroup( |
|
2688 | def addgroup( | |
2702 | self, |
|
2689 | self, | |
2703 | deltas, |
|
2690 | deltas, | |
2704 | linkmapper, |
|
2691 | linkmapper, | |
2705 | transaction, |
|
2692 | transaction, | |
2706 | alwayscache=False, |
|
2693 | alwayscache=False, | |
2707 | addrevisioncb=None, |
|
2694 | addrevisioncb=None, | |
2708 | duplicaterevisioncb=None, |
|
2695 | duplicaterevisioncb=None, | |
2709 | ): |
|
2696 | ): | |
2710 | """ |
|
2697 | """ | |
2711 | add a delta group |
|
2698 | add a delta group | |
2712 |
|
2699 | |||
2713 | given a set of deltas, add them to the revision log. the |
|
2700 | given a set of deltas, add them to the revision log. the | |
2714 | first delta is against its parent, which should be in our |
|
2701 | first delta is against its parent, which should be in our | |
2715 | log, the rest are against the previous delta. |
|
2702 | log, the rest are against the previous delta. | |
2716 |
|
2703 | |||
2717 | If ``addrevisioncb`` is defined, it will be called with arguments of |
|
2704 | If ``addrevisioncb`` is defined, it will be called with arguments of | |
2718 | this revlog and the node that was added. |
|
2705 | this revlog and the node that was added. | |
2719 | """ |
|
2706 | """ | |
2720 |
|
2707 | |||
2721 | if self._adding_group: |
|
2708 | if self._adding_group: | |
2722 | raise error.ProgrammingError(b'cannot nest addgroup() calls') |
|
2709 | raise error.ProgrammingError(b'cannot nest addgroup() calls') | |
2723 |
|
2710 | |||
2724 | self._adding_group = True |
|
2711 | self._adding_group = True | |
2725 | empty = True |
|
2712 | empty = True | |
2726 | try: |
|
2713 | try: | |
2727 | with self._writing(transaction): |
|
2714 | with self._writing(transaction): | |
2728 | deltacomputer = deltautil.deltacomputer(self) |
|
2715 | deltacomputer = deltautil.deltacomputer(self) | |
2729 | # loop through our set of deltas |
|
2716 | # loop through our set of deltas | |
2730 | for data in deltas: |
|
2717 | for data in deltas: | |
2731 | ( |
|
2718 | ( | |
2732 | node, |
|
2719 | node, | |
2733 | p1, |
|
2720 | p1, | |
2734 | p2, |
|
2721 | p2, | |
2735 | linknode, |
|
2722 | linknode, | |
2736 | deltabase, |
|
2723 | deltabase, | |
2737 | delta, |
|
2724 | delta, | |
2738 | flags, |
|
2725 | flags, | |
2739 | sidedata, |
|
2726 | sidedata, | |
2740 | ) = data |
|
2727 | ) = data | |
2741 | link = linkmapper(linknode) |
|
2728 | link = linkmapper(linknode) | |
2742 | flags = flags or REVIDX_DEFAULT_FLAGS |
|
2729 | flags = flags or REVIDX_DEFAULT_FLAGS | |
2743 |
|
2730 | |||
2744 | rev = self.index.get_rev(node) |
|
2731 | rev = self.index.get_rev(node) | |
2745 | if rev is not None: |
|
2732 | if rev is not None: | |
2746 | # this can happen if two branches make the same change |
|
2733 | # this can happen if two branches make the same change | |
2747 | self._nodeduplicatecallback(transaction, rev) |
|
2734 | self._nodeduplicatecallback(transaction, rev) | |
2748 | if duplicaterevisioncb: |
|
2735 | if duplicaterevisioncb: | |
2749 | duplicaterevisioncb(self, rev) |
|
2736 | duplicaterevisioncb(self, rev) | |
2750 | empty = False |
|
2737 | empty = False | |
2751 | continue |
|
2738 | continue | |
2752 |
|
2739 | |||
2753 | for p in (p1, p2): |
|
2740 | for p in (p1, p2): | |
2754 | if not self.index.has_node(p): |
|
2741 | if not self.index.has_node(p): | |
2755 | raise error.LookupError( |
|
2742 | raise error.LookupError( | |
2756 | p, self.radix, _(b'unknown parent') |
|
2743 | p, self.radix, _(b'unknown parent') | |
2757 | ) |
|
2744 | ) | |
2758 |
|
2745 | |||
2759 | if not self.index.has_node(deltabase): |
|
2746 | if not self.index.has_node(deltabase): | |
2760 | raise error.LookupError( |
|
2747 | raise error.LookupError( | |
2761 | deltabase, self.display_id, _(b'unknown delta base') |
|
2748 | deltabase, self.display_id, _(b'unknown delta base') | |
2762 | ) |
|
2749 | ) | |
2763 |
|
2750 | |||
2764 | baserev = self.rev(deltabase) |
|
2751 | baserev = self.rev(deltabase) | |
2765 |
|
2752 | |||
2766 | if baserev != nullrev and self.iscensored(baserev): |
|
2753 | if baserev != nullrev and self.iscensored(baserev): | |
2767 | # if base is censored, delta must be full replacement in a |
|
2754 | # if base is censored, delta must be full replacement in a | |
2768 | # single patch operation |
|
2755 | # single patch operation | |
2769 | hlen = struct.calcsize(b">lll") |
|
2756 | hlen = struct.calcsize(b">lll") | |
2770 | oldlen = self.rawsize(baserev) |
|
2757 | oldlen = self.rawsize(baserev) | |
2771 | newlen = len(delta) - hlen |
|
2758 | newlen = len(delta) - hlen | |
2772 | if delta[:hlen] != mdiff.replacediffheader( |
|
2759 | if delta[:hlen] != mdiff.replacediffheader( | |
2773 | oldlen, newlen |
|
2760 | oldlen, newlen | |
2774 | ): |
|
2761 | ): | |
2775 | raise error.CensoredBaseError( |
|
2762 | raise error.CensoredBaseError( | |
2776 | self.display_id, self.node(baserev) |
|
2763 | self.display_id, self.node(baserev) | |
2777 | ) |
|
2764 | ) | |
2778 |
|
2765 | |||
2779 | if not flags and self._peek_iscensored(baserev, delta): |
|
2766 | if not flags and self._peek_iscensored(baserev, delta): | |
2780 | flags |= REVIDX_ISCENSORED |
|
2767 | flags |= REVIDX_ISCENSORED | |
2781 |
|
2768 | |||
2782 | # We assume consumers of addrevisioncb will want to retrieve |
|
2769 | # We assume consumers of addrevisioncb will want to retrieve | |
2783 | # the added revision, which will require a call to |
|
2770 | # the added revision, which will require a call to | |
2784 | # revision(). revision() will fast path if there is a cache |
|
2771 | # revision(). revision() will fast path if there is a cache | |
2785 | # hit. So, we tell _addrevision() to always cache in this case. |
|
2772 | # hit. So, we tell _addrevision() to always cache in this case. | |
2786 | # We're only using addgroup() in the context of changegroup |
|
2773 | # We're only using addgroup() in the context of changegroup | |
2787 | # generation so the revision data can always be handled as raw |
|
2774 | # generation so the revision data can always be handled as raw | |
2788 | # by the flagprocessor. |
|
2775 | # by the flagprocessor. | |
2789 | rev = self._addrevision( |
|
2776 | rev = self._addrevision( | |
2790 | node, |
|
2777 | node, | |
2791 | None, |
|
2778 | None, | |
2792 | transaction, |
|
2779 | transaction, | |
2793 | link, |
|
2780 | link, | |
2794 | p1, |
|
2781 | p1, | |
2795 | p2, |
|
2782 | p2, | |
2796 | flags, |
|
2783 | flags, | |
2797 | (baserev, delta), |
|
2784 | (baserev, delta), | |
2798 | alwayscache=alwayscache, |
|
2785 | alwayscache=alwayscache, | |
2799 | deltacomputer=deltacomputer, |
|
2786 | deltacomputer=deltacomputer, | |
2800 | sidedata=sidedata, |
|
2787 | sidedata=sidedata, | |
2801 | ) |
|
2788 | ) | |
2802 |
|
2789 | |||
2803 | if addrevisioncb: |
|
2790 | if addrevisioncb: | |
2804 | addrevisioncb(self, rev) |
|
2791 | addrevisioncb(self, rev) | |
2805 | empty = False |
|
2792 | empty = False | |
2806 | finally: |
|
2793 | finally: | |
2807 | self._adding_group = False |
|
2794 | self._adding_group = False | |
2808 | return not empty |
|
2795 | return not empty | |
2809 |
|
2796 | |||
2810 | def iscensored(self, rev): |
|
2797 | def iscensored(self, rev): | |
2811 | """Check if a file revision is censored.""" |
|
2798 | """Check if a file revision is censored.""" | |
2812 | if not self._censorable: |
|
2799 | if not self._censorable: | |
2813 | return False |
|
2800 | return False | |
2814 |
|
2801 | |||
2815 | return self.flags(rev) & REVIDX_ISCENSORED |
|
2802 | return self.flags(rev) & REVIDX_ISCENSORED | |
2816 |
|
2803 | |||
2817 | def _peek_iscensored(self, baserev, delta): |
|
2804 | def _peek_iscensored(self, baserev, delta): | |
2818 | """Quickly check if a delta produces a censored revision.""" |
|
2805 | """Quickly check if a delta produces a censored revision.""" | |
2819 | if not self._censorable: |
|
2806 | if not self._censorable: | |
2820 | return False |
|
2807 | return False | |
2821 |
|
2808 | |||
2822 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) |
|
2809 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) | |
2823 |
|
2810 | |||
2824 | def getstrippoint(self, minlink): |
|
2811 | def getstrippoint(self, minlink): | |
2825 | """find the minimum rev that must be stripped to strip the linkrev |
|
2812 | """find the minimum rev that must be stripped to strip the linkrev | |
2826 |
|
2813 | |||
2827 | Returns a tuple containing the minimum rev and a set of all revs that |
|
2814 | Returns a tuple containing the minimum rev and a set of all revs that | |
2828 | have linkrevs that will be broken by this strip. |
|
2815 | have linkrevs that will be broken by this strip. | |
2829 | """ |
|
2816 | """ | |
2830 | return storageutil.resolvestripinfo( |
|
2817 | return storageutil.resolvestripinfo( | |
2831 | minlink, |
|
2818 | minlink, | |
2832 | len(self) - 1, |
|
2819 | len(self) - 1, | |
2833 | self.headrevs(), |
|
2820 | self.headrevs(), | |
2834 | self.linkrev, |
|
2821 | self.linkrev, | |
2835 | self.parentrevs, |
|
2822 | self.parentrevs, | |
2836 | ) |
|
2823 | ) | |
2837 |
|
2824 | |||
2838 | def strip(self, minlink, transaction): |
|
2825 | def strip(self, minlink, transaction): | |
2839 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
2826 | """truncate the revlog on the first revision with a linkrev >= minlink | |
2840 |
|
2827 | |||
2841 | This function is called when we're stripping revision minlink and |
|
2828 | This function is called when we're stripping revision minlink and | |
2842 | its descendants from the repository. |
|
2829 | its descendants from the repository. | |
2843 |
|
2830 | |||
2844 | We have to remove all revisions with linkrev >= minlink, because |
|
2831 | We have to remove all revisions with linkrev >= minlink, because | |
2845 | the equivalent changelog revisions will be renumbered after the |
|
2832 | the equivalent changelog revisions will be renumbered after the | |
2846 | strip. |
|
2833 | strip. | |
2847 |
|
2834 | |||
2848 | So we truncate the revlog on the first of these revisions, and |
|
2835 | So we truncate the revlog on the first of these revisions, and | |
2849 | trust that the caller has saved the revisions that shouldn't be |
|
2836 | trust that the caller has saved the revisions that shouldn't be | |
2850 | removed and that it'll re-add them after this truncation. |
|
2837 | removed and that it'll re-add them after this truncation. | |
2851 | """ |
|
2838 | """ | |
2852 | if len(self) == 0: |
|
2839 | if len(self) == 0: | |
2853 | return |
|
2840 | return | |
2854 |
|
2841 | |||
2855 | rev, _ = self.getstrippoint(minlink) |
|
2842 | rev, _ = self.getstrippoint(minlink) | |
2856 | if rev == len(self): |
|
2843 | if rev == len(self): | |
2857 | return |
|
2844 | return | |
2858 |
|
2845 | |||
2859 | # first truncate the files on disk |
|
2846 | # first truncate the files on disk | |
2860 | data_end = self.start(rev) |
|
2847 | data_end = self.start(rev) | |
2861 | if not self._inline: |
|
2848 | if not self._inline: | |
2862 | transaction.add(self._datafile, data_end) |
|
2849 | transaction.add(self._datafile, data_end) | |
2863 | end = rev * self.index.entry_size |
|
2850 | end = rev * self.index.entry_size | |
2864 | else: |
|
2851 | else: | |
2865 | end = data_end + (rev * self.index.entry_size) |
|
2852 | end = data_end + (rev * self.index.entry_size) | |
2866 |
|
2853 | |||
2867 | if self._sidedatafile: |
|
2854 | if self._sidedatafile: | |
2868 | sidedata_end = self.sidedata_cut_off(rev) |
|
2855 | sidedata_end = self.sidedata_cut_off(rev) | |
2869 | transaction.add(self._sidedatafile, sidedata_end) |
|
2856 | transaction.add(self._sidedatafile, sidedata_end) | |
2870 |
|
2857 | |||
2871 | transaction.add(self._indexfile, end) |
|
2858 | transaction.add(self._indexfile, end) | |
2872 | if self._docket is not None: |
|
2859 | if self._docket is not None: | |
2873 | # XXX we could, leverage the docket while stripping. However it is |
|
2860 | # XXX we could, leverage the docket while stripping. However it is | |
2874 | # not powerfull enough at the time of this comment |
|
2861 | # not powerfull enough at the time of this comment | |
2875 | self._docket.index_end = end |
|
2862 | self._docket.index_end = end | |
2876 | self._docket.data_end = data_end |
|
2863 | self._docket.data_end = data_end | |
2877 | self._docket.sidedata_end = sidedata_end |
|
2864 | self._docket.sidedata_end = sidedata_end | |
2878 | self._docket.write(transaction, stripping=True) |
|
2865 | self._docket.write(transaction, stripping=True) | |
2879 |
|
2866 | |||
2880 | # then reset internal state in memory to forget those revisions |
|
2867 | # then reset internal state in memory to forget those revisions | |
2881 | self._revisioncache = None |
|
2868 | self._revisioncache = None | |
2882 | self._chaininfocache = util.lrucachedict(500) |
|
2869 | self._chaininfocache = util.lrucachedict(500) | |
2883 | self._chunkclear() |
|
2870 | self._chunkclear() | |
2884 |
|
2871 | |||
2885 | del self.index[rev:-1] |
|
2872 | del self.index[rev:-1] | |
2886 |
|
2873 | |||
2887 | def checksize(self): |
|
2874 | def checksize(self): | |
2888 | """Check size of index and data files |
|
2875 | """Check size of index and data files | |
2889 |
|
2876 | |||
2890 | return a (dd, di) tuple. |
|
2877 | return a (dd, di) tuple. | |
2891 | - dd: extra bytes for the "data" file |
|
2878 | - dd: extra bytes for the "data" file | |
2892 | - di: extra bytes for the "index" file |
|
2879 | - di: extra bytes for the "index" file | |
2893 |
|
2880 | |||
2894 | A healthy revlog will return (0, 0). |
|
2881 | A healthy revlog will return (0, 0). | |
2895 | """ |
|
2882 | """ | |
2896 | expected = 0 |
|
2883 | expected = 0 | |
2897 | if len(self): |
|
2884 | if len(self): | |
2898 | expected = max(0, self.end(len(self) - 1)) |
|
2885 | expected = max(0, self.end(len(self) - 1)) | |
2899 |
|
2886 | |||
2900 | try: |
|
2887 | try: | |
2901 | with self._datafp() as f: |
|
2888 | with self._datafp() as f: | |
2902 | f.seek(0, io.SEEK_END) |
|
2889 | f.seek(0, io.SEEK_END) | |
2903 | actual = f.tell() |
|
2890 | actual = f.tell() | |
2904 | dd = actual - expected |
|
2891 | dd = actual - expected | |
2905 | except IOError as inst: |
|
2892 | except IOError as inst: | |
2906 | if inst.errno != errno.ENOENT: |
|
2893 | if inst.errno != errno.ENOENT: | |
2907 | raise |
|
2894 | raise | |
2908 | dd = 0 |
|
2895 | dd = 0 | |
2909 |
|
2896 | |||
2910 | try: |
|
2897 | try: | |
2911 | f = self.opener(self._indexfile) |
|
2898 | f = self.opener(self._indexfile) | |
2912 | f.seek(0, io.SEEK_END) |
|
2899 | f.seek(0, io.SEEK_END) | |
2913 | actual = f.tell() |
|
2900 | actual = f.tell() | |
2914 | f.close() |
|
2901 | f.close() | |
2915 | s = self.index.entry_size |
|
2902 | s = self.index.entry_size | |
2916 | i = max(0, actual // s) |
|
2903 | i = max(0, actual // s) | |
2917 | di = actual - (i * s) |
|
2904 | di = actual - (i * s) | |
2918 | if self._inline: |
|
2905 | if self._inline: | |
2919 | databytes = 0 |
|
2906 | databytes = 0 | |
2920 | for r in self: |
|
2907 | for r in self: | |
2921 | databytes += max(0, self.length(r)) |
|
2908 | databytes += max(0, self.length(r)) | |
2922 | dd = 0 |
|
2909 | dd = 0 | |
2923 | di = actual - len(self) * s - databytes |
|
2910 | di = actual - len(self) * s - databytes | |
2924 | except IOError as inst: |
|
2911 | except IOError as inst: | |
2925 | if inst.errno != errno.ENOENT: |
|
2912 | if inst.errno != errno.ENOENT: | |
2926 | raise |
|
2913 | raise | |
2927 | di = 0 |
|
2914 | di = 0 | |
2928 |
|
2915 | |||
2929 | return (dd, di) |
|
2916 | return (dd, di) | |
2930 |
|
2917 | |||
2931 | def files(self): |
|
2918 | def files(self): | |
2932 | res = [self._indexfile] |
|
2919 | res = [self._indexfile] | |
2933 | if not self._inline: |
|
2920 | if not self._inline: | |
2934 | res.append(self._datafile) |
|
2921 | res.append(self._datafile) | |
2935 | return res |
|
2922 | return res | |
2936 |
|
2923 | |||
2937 | def emitrevisions( |
|
2924 | def emitrevisions( | |
2938 | self, |
|
2925 | self, | |
2939 | nodes, |
|
2926 | nodes, | |
2940 | nodesorder=None, |
|
2927 | nodesorder=None, | |
2941 | revisiondata=False, |
|
2928 | revisiondata=False, | |
2942 | assumehaveparentrevisions=False, |
|
2929 | assumehaveparentrevisions=False, | |
2943 | deltamode=repository.CG_DELTAMODE_STD, |
|
2930 | deltamode=repository.CG_DELTAMODE_STD, | |
2944 | sidedata_helpers=None, |
|
2931 | sidedata_helpers=None, | |
2945 | ): |
|
2932 | ): | |
2946 | if nodesorder not in (b'nodes', b'storage', b'linear', None): |
|
2933 | if nodesorder not in (b'nodes', b'storage', b'linear', None): | |
2947 | raise error.ProgrammingError( |
|
2934 | raise error.ProgrammingError( | |
2948 | b'unhandled value for nodesorder: %s' % nodesorder |
|
2935 | b'unhandled value for nodesorder: %s' % nodesorder | |
2949 | ) |
|
2936 | ) | |
2950 |
|
2937 | |||
2951 | if nodesorder is None and not self._generaldelta: |
|
2938 | if nodesorder is None and not self._generaldelta: | |
2952 | nodesorder = b'storage' |
|
2939 | nodesorder = b'storage' | |
2953 |
|
2940 | |||
2954 | if ( |
|
2941 | if ( | |
2955 | not self._storedeltachains |
|
2942 | not self._storedeltachains | |
2956 | and deltamode != repository.CG_DELTAMODE_PREV |
|
2943 | and deltamode != repository.CG_DELTAMODE_PREV | |
2957 | ): |
|
2944 | ): | |
2958 | deltamode = repository.CG_DELTAMODE_FULL |
|
2945 | deltamode = repository.CG_DELTAMODE_FULL | |
2959 |
|
2946 | |||
2960 | return storageutil.emitrevisions( |
|
2947 | return storageutil.emitrevisions( | |
2961 | self, |
|
2948 | self, | |
2962 | nodes, |
|
2949 | nodes, | |
2963 | nodesorder, |
|
2950 | nodesorder, | |
2964 | revlogrevisiondelta, |
|
2951 | revlogrevisiondelta, | |
2965 | deltaparentfn=self.deltaparent, |
|
2952 | deltaparentfn=self.deltaparent, | |
2966 | candeltafn=self.candelta, |
|
2953 | candeltafn=self.candelta, | |
2967 | rawsizefn=self.rawsize, |
|
2954 | rawsizefn=self.rawsize, | |
2968 | revdifffn=self.revdiff, |
|
2955 | revdifffn=self.revdiff, | |
2969 | flagsfn=self.flags, |
|
2956 | flagsfn=self.flags, | |
2970 | deltamode=deltamode, |
|
2957 | deltamode=deltamode, | |
2971 | revisiondata=revisiondata, |
|
2958 | revisiondata=revisiondata, | |
2972 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
2959 | assumehaveparentrevisions=assumehaveparentrevisions, | |
2973 | sidedata_helpers=sidedata_helpers, |
|
2960 | sidedata_helpers=sidedata_helpers, | |
2974 | ) |
|
2961 | ) | |
2975 |
|
2962 | |||
2976 | DELTAREUSEALWAYS = b'always' |
|
2963 | DELTAREUSEALWAYS = b'always' | |
2977 | DELTAREUSESAMEREVS = b'samerevs' |
|
2964 | DELTAREUSESAMEREVS = b'samerevs' | |
2978 | DELTAREUSENEVER = b'never' |
|
2965 | DELTAREUSENEVER = b'never' | |
2979 |
|
2966 | |||
2980 | DELTAREUSEFULLADD = b'fulladd' |
|
2967 | DELTAREUSEFULLADD = b'fulladd' | |
2981 |
|
2968 | |||
2982 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} |
|
2969 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} | |
2983 |
|
2970 | |||
2984 | def clone( |
|
2971 | def clone( | |
2985 | self, |
|
2972 | self, | |
2986 | tr, |
|
2973 | tr, | |
2987 | destrevlog, |
|
2974 | destrevlog, | |
2988 | addrevisioncb=None, |
|
2975 | addrevisioncb=None, | |
2989 | deltareuse=DELTAREUSESAMEREVS, |
|
2976 | deltareuse=DELTAREUSESAMEREVS, | |
2990 | forcedeltabothparents=None, |
|
2977 | forcedeltabothparents=None, | |
2991 | sidedata_helpers=None, |
|
2978 | sidedata_helpers=None, | |
2992 | ): |
|
2979 | ): | |
2993 | """Copy this revlog to another, possibly with format changes. |
|
2980 | """Copy this revlog to another, possibly with format changes. | |
2994 |
|
2981 | |||
2995 | The destination revlog will contain the same revisions and nodes. |
|
2982 | The destination revlog will contain the same revisions and nodes. | |
2996 | However, it may not be bit-for-bit identical due to e.g. delta encoding |
|
2983 | However, it may not be bit-for-bit identical due to e.g. delta encoding | |
2997 | differences. |
|
2984 | differences. | |
2998 |
|
2985 | |||
2999 | The ``deltareuse`` argument control how deltas from the existing revlog |
|
2986 | The ``deltareuse`` argument control how deltas from the existing revlog | |
3000 | are preserved in the destination revlog. The argument can have the |
|
2987 | are preserved in the destination revlog. The argument can have the | |
3001 | following values: |
|
2988 | following values: | |
3002 |
|
2989 | |||
3003 | DELTAREUSEALWAYS |
|
2990 | DELTAREUSEALWAYS | |
3004 | Deltas will always be reused (if possible), even if the destination |
|
2991 | Deltas will always be reused (if possible), even if the destination | |
3005 | revlog would not select the same revisions for the delta. This is the |
|
2992 | revlog would not select the same revisions for the delta. This is the | |
3006 | fastest mode of operation. |
|
2993 | fastest mode of operation. | |
3007 | DELTAREUSESAMEREVS |
|
2994 | DELTAREUSESAMEREVS | |
3008 | Deltas will be reused if the destination revlog would pick the same |
|
2995 | Deltas will be reused if the destination revlog would pick the same | |
3009 | revisions for the delta. This mode strikes a balance between speed |
|
2996 | revisions for the delta. This mode strikes a balance between speed | |
3010 | and optimization. |
|
2997 | and optimization. | |
3011 | DELTAREUSENEVER |
|
2998 | DELTAREUSENEVER | |
3012 | Deltas will never be reused. This is the slowest mode of execution. |
|
2999 | Deltas will never be reused. This is the slowest mode of execution. | |
3013 | This mode can be used to recompute deltas (e.g. if the diff/delta |
|
3000 | This mode can be used to recompute deltas (e.g. if the diff/delta | |
3014 | algorithm changes). |
|
3001 | algorithm changes). | |
3015 | DELTAREUSEFULLADD |
|
3002 | DELTAREUSEFULLADD | |
3016 | Revision will be re-added as if their were new content. This is |
|
3003 | Revision will be re-added as if their were new content. This is | |
3017 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. |
|
3004 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. | |
3018 | eg: large file detection and handling. |
|
3005 | eg: large file detection and handling. | |
3019 |
|
3006 | |||
3020 | Delta computation can be slow, so the choice of delta reuse policy can |
|
3007 | Delta computation can be slow, so the choice of delta reuse policy can | |
3021 | significantly affect run time. |
|
3008 | significantly affect run time. | |
3022 |
|
3009 | |||
3023 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between |
|
3010 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between | |
3024 | two extremes. Deltas will be reused if they are appropriate. But if the |
|
3011 | two extremes. Deltas will be reused if they are appropriate. But if the | |
3025 | delta could choose a better revision, it will do so. This means if you |
|
3012 | delta could choose a better revision, it will do so. This means if you | |
3026 | are converting a non-generaldelta revlog to a generaldelta revlog, |
|
3013 | are converting a non-generaldelta revlog to a generaldelta revlog, | |
3027 | deltas will be recomputed if the delta's parent isn't a parent of the |
|
3014 | deltas will be recomputed if the delta's parent isn't a parent of the | |
3028 | revision. |
|
3015 | revision. | |
3029 |
|
3016 | |||
3030 | In addition to the delta policy, the ``forcedeltabothparents`` |
|
3017 | In addition to the delta policy, the ``forcedeltabothparents`` | |
3031 | argument controls whether to force compute deltas against both parents |
|
3018 | argument controls whether to force compute deltas against both parents | |
3032 | for merges. By default, the current default is used. |
|
3019 | for merges. By default, the current default is used. | |
3033 |
|
3020 | |||
3034 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on |
|
3021 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on | |
3035 | `sidedata_helpers`. |
|
3022 | `sidedata_helpers`. | |
3036 | """ |
|
3023 | """ | |
3037 | if deltareuse not in self.DELTAREUSEALL: |
|
3024 | if deltareuse not in self.DELTAREUSEALL: | |
3038 | raise ValueError( |
|
3025 | raise ValueError( | |
3039 | _(b'value for deltareuse invalid: %s') % deltareuse |
|
3026 | _(b'value for deltareuse invalid: %s') % deltareuse | |
3040 | ) |
|
3027 | ) | |
3041 |
|
3028 | |||
3042 | if len(destrevlog): |
|
3029 | if len(destrevlog): | |
3043 | raise ValueError(_(b'destination revlog is not empty')) |
|
3030 | raise ValueError(_(b'destination revlog is not empty')) | |
3044 |
|
3031 | |||
3045 | if getattr(self, 'filteredrevs', None): |
|
3032 | if getattr(self, 'filteredrevs', None): | |
3046 | raise ValueError(_(b'source revlog has filtered revisions')) |
|
3033 | raise ValueError(_(b'source revlog has filtered revisions')) | |
3047 | if getattr(destrevlog, 'filteredrevs', None): |
|
3034 | if getattr(destrevlog, 'filteredrevs', None): | |
3048 | raise ValueError(_(b'destination revlog has filtered revisions')) |
|
3035 | raise ValueError(_(b'destination revlog has filtered revisions')) | |
3049 |
|
3036 | |||
3050 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, |
|
3037 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, | |
3051 | # if possible. |
|
3038 | # if possible. | |
3052 | oldlazydelta = destrevlog._lazydelta |
|
3039 | oldlazydelta = destrevlog._lazydelta | |
3053 | oldlazydeltabase = destrevlog._lazydeltabase |
|
3040 | oldlazydeltabase = destrevlog._lazydeltabase | |
3054 | oldamd = destrevlog._deltabothparents |
|
3041 | oldamd = destrevlog._deltabothparents | |
3055 |
|
3042 | |||
3056 | try: |
|
3043 | try: | |
3057 | if deltareuse == self.DELTAREUSEALWAYS: |
|
3044 | if deltareuse == self.DELTAREUSEALWAYS: | |
3058 | destrevlog._lazydeltabase = True |
|
3045 | destrevlog._lazydeltabase = True | |
3059 | destrevlog._lazydelta = True |
|
3046 | destrevlog._lazydelta = True | |
3060 | elif deltareuse == self.DELTAREUSESAMEREVS: |
|
3047 | elif deltareuse == self.DELTAREUSESAMEREVS: | |
3061 | destrevlog._lazydeltabase = False |
|
3048 | destrevlog._lazydeltabase = False | |
3062 | destrevlog._lazydelta = True |
|
3049 | destrevlog._lazydelta = True | |
3063 | elif deltareuse == self.DELTAREUSENEVER: |
|
3050 | elif deltareuse == self.DELTAREUSENEVER: | |
3064 | destrevlog._lazydeltabase = False |
|
3051 | destrevlog._lazydeltabase = False | |
3065 | destrevlog._lazydelta = False |
|
3052 | destrevlog._lazydelta = False | |
3066 |
|
3053 | |||
3067 | destrevlog._deltabothparents = forcedeltabothparents or oldamd |
|
3054 | destrevlog._deltabothparents = forcedeltabothparents or oldamd | |
3068 |
|
3055 | |||
3069 | self._clone( |
|
3056 | self._clone( | |
3070 | tr, |
|
3057 | tr, | |
3071 | destrevlog, |
|
3058 | destrevlog, | |
3072 | addrevisioncb, |
|
3059 | addrevisioncb, | |
3073 | deltareuse, |
|
3060 | deltareuse, | |
3074 | forcedeltabothparents, |
|
3061 | forcedeltabothparents, | |
3075 | sidedata_helpers, |
|
3062 | sidedata_helpers, | |
3076 | ) |
|
3063 | ) | |
3077 |
|
3064 | |||
3078 | finally: |
|
3065 | finally: | |
3079 | destrevlog._lazydelta = oldlazydelta |
|
3066 | destrevlog._lazydelta = oldlazydelta | |
3080 | destrevlog._lazydeltabase = oldlazydeltabase |
|
3067 | destrevlog._lazydeltabase = oldlazydeltabase | |
3081 | destrevlog._deltabothparents = oldamd |
|
3068 | destrevlog._deltabothparents = oldamd | |
3082 |
|
3069 | |||
3083 | def _clone( |
|
3070 | def _clone( | |
3084 | self, |
|
3071 | self, | |
3085 | tr, |
|
3072 | tr, | |
3086 | destrevlog, |
|
3073 | destrevlog, | |
3087 | addrevisioncb, |
|
3074 | addrevisioncb, | |
3088 | deltareuse, |
|
3075 | deltareuse, | |
3089 | forcedeltabothparents, |
|
3076 | forcedeltabothparents, | |
3090 | sidedata_helpers, |
|
3077 | sidedata_helpers, | |
3091 | ): |
|
3078 | ): | |
3092 | """perform the core duty of `revlog.clone` after parameter processing""" |
|
3079 | """perform the core duty of `revlog.clone` after parameter processing""" | |
3093 | deltacomputer = deltautil.deltacomputer(destrevlog) |
|
3080 | deltacomputer = deltautil.deltacomputer(destrevlog) | |
3094 | index = self.index |
|
3081 | index = self.index | |
3095 | for rev in self: |
|
3082 | for rev in self: | |
3096 | entry = index[rev] |
|
3083 | entry = index[rev] | |
3097 |
|
3084 | |||
3098 | # Some classes override linkrev to take filtered revs into |
|
3085 | # Some classes override linkrev to take filtered revs into | |
3099 | # account. Use raw entry from index. |
|
3086 | # account. Use raw entry from index. | |
3100 | flags = entry[0] & 0xFFFF |
|
3087 | flags = entry[0] & 0xFFFF | |
3101 | linkrev = entry[4] |
|
3088 | linkrev = entry[4] | |
3102 | p1 = index[entry[5]][7] |
|
3089 | p1 = index[entry[5]][7] | |
3103 | p2 = index[entry[6]][7] |
|
3090 | p2 = index[entry[6]][7] | |
3104 | node = entry[7] |
|
3091 | node = entry[7] | |
3105 |
|
3092 | |||
3106 | # (Possibly) reuse the delta from the revlog if allowed and |
|
3093 | # (Possibly) reuse the delta from the revlog if allowed and | |
3107 | # the revlog chunk is a delta. |
|
3094 | # the revlog chunk is a delta. | |
3108 | cachedelta = None |
|
3095 | cachedelta = None | |
3109 | rawtext = None |
|
3096 | rawtext = None | |
3110 | if deltareuse == self.DELTAREUSEFULLADD: |
|
3097 | if deltareuse == self.DELTAREUSEFULLADD: | |
3111 | text = self._revisiondata(rev) |
|
3098 | text = self._revisiondata(rev) | |
3112 | sidedata = self.sidedata(rev) |
|
3099 | sidedata = self.sidedata(rev) | |
3113 |
|
3100 | |||
3114 | if sidedata_helpers is not None: |
|
3101 | if sidedata_helpers is not None: | |
3115 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
3102 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
3116 | self, sidedata_helpers, sidedata, rev |
|
3103 | self, sidedata_helpers, sidedata, rev | |
3117 | ) |
|
3104 | ) | |
3118 | flags = flags | new_flags[0] & ~new_flags[1] |
|
3105 | flags = flags | new_flags[0] & ~new_flags[1] | |
3119 |
|
3106 | |||
3120 | destrevlog.addrevision( |
|
3107 | destrevlog.addrevision( | |
3121 | text, |
|
3108 | text, | |
3122 | tr, |
|
3109 | tr, | |
3123 | linkrev, |
|
3110 | linkrev, | |
3124 | p1, |
|
3111 | p1, | |
3125 | p2, |
|
3112 | p2, | |
3126 | cachedelta=cachedelta, |
|
3113 | cachedelta=cachedelta, | |
3127 | node=node, |
|
3114 | node=node, | |
3128 | flags=flags, |
|
3115 | flags=flags, | |
3129 | deltacomputer=deltacomputer, |
|
3116 | deltacomputer=deltacomputer, | |
3130 | sidedata=sidedata, |
|
3117 | sidedata=sidedata, | |
3131 | ) |
|
3118 | ) | |
3132 | else: |
|
3119 | else: | |
3133 | if destrevlog._lazydelta: |
|
3120 | if destrevlog._lazydelta: | |
3134 | dp = self.deltaparent(rev) |
|
3121 | dp = self.deltaparent(rev) | |
3135 | if dp != nullrev: |
|
3122 | if dp != nullrev: | |
3136 | cachedelta = (dp, bytes(self._chunk(rev))) |
|
3123 | cachedelta = (dp, bytes(self._chunk(rev))) | |
3137 |
|
3124 | |||
3138 | sidedata = None |
|
3125 | sidedata = None | |
3139 | if not cachedelta: |
|
3126 | if not cachedelta: | |
3140 | rawtext = self._revisiondata(rev) |
|
3127 | rawtext = self._revisiondata(rev) | |
3141 | sidedata = self.sidedata(rev) |
|
3128 | sidedata = self.sidedata(rev) | |
3142 | if sidedata is None: |
|
3129 | if sidedata is None: | |
3143 | sidedata = self.sidedata(rev) |
|
3130 | sidedata = self.sidedata(rev) | |
3144 |
|
3131 | |||
3145 | if sidedata_helpers is not None: |
|
3132 | if sidedata_helpers is not None: | |
3146 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
3133 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
3147 | self, sidedata_helpers, sidedata, rev |
|
3134 | self, sidedata_helpers, sidedata, rev | |
3148 | ) |
|
3135 | ) | |
3149 | flags = flags | new_flags[0] & ~new_flags[1] |
|
3136 | flags = flags | new_flags[0] & ~new_flags[1] | |
3150 |
|
3137 | |||
3151 | with destrevlog._writing(tr): |
|
3138 | with destrevlog._writing(tr): | |
3152 | destrevlog._addrevision( |
|
3139 | destrevlog._addrevision( | |
3153 | node, |
|
3140 | node, | |
3154 | rawtext, |
|
3141 | rawtext, | |
3155 | tr, |
|
3142 | tr, | |
3156 | linkrev, |
|
3143 | linkrev, | |
3157 | p1, |
|
3144 | p1, | |
3158 | p2, |
|
3145 | p2, | |
3159 | flags, |
|
3146 | flags, | |
3160 | cachedelta, |
|
3147 | cachedelta, | |
3161 | deltacomputer=deltacomputer, |
|
3148 | deltacomputer=deltacomputer, | |
3162 | sidedata=sidedata, |
|
3149 | sidedata=sidedata, | |
3163 | ) |
|
3150 | ) | |
3164 |
|
3151 | |||
3165 | if addrevisioncb: |
|
3152 | if addrevisioncb: | |
3166 | addrevisioncb(self, rev, node) |
|
3153 | addrevisioncb(self, rev, node) | |
3167 |
|
3154 | |||
3168 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
3155 | def censorrevision(self, tr, censornode, tombstone=b''): | |
3169 | if self._format_version == REVLOGV0: |
|
3156 | if self._format_version == REVLOGV0: | |
3170 | raise error.RevlogError( |
|
3157 | raise error.RevlogError( | |
3171 | _(b'cannot censor with version %d revlogs') |
|
3158 | _(b'cannot censor with version %d revlogs') | |
3172 | % self._format_version |
|
3159 | % self._format_version | |
3173 | ) |
|
3160 | ) | |
3174 | elif self._format_version == REVLOGV1: |
|
3161 | elif self._format_version == REVLOGV1: | |
3175 | censor.v1_censor(self, tr, censornode, tombstone) |
|
3162 | censor.v1_censor(self, tr, censornode, tombstone) | |
3176 | else: |
|
3163 | else: | |
3177 | # revlog v2 |
|
3164 | # revlog v2 | |
3178 | raise error.RevlogError( |
|
3165 | raise error.RevlogError( | |
3179 | _(b'cannot censor with version %d revlogs') |
|
3166 | _(b'cannot censor with version %d revlogs') | |
3180 | % self._format_version |
|
3167 | % self._format_version | |
3181 | ) |
|
3168 | ) | |
3182 |
|
3169 | |||
3183 | def verifyintegrity(self, state): |
|
3170 | def verifyintegrity(self, state): | |
3184 | """Verifies the integrity of the revlog. |
|
3171 | """Verifies the integrity of the revlog. | |
3185 |
|
3172 | |||
3186 | Yields ``revlogproblem`` instances describing problems that are |
|
3173 | Yields ``revlogproblem`` instances describing problems that are | |
3187 | found. |
|
3174 | found. | |
3188 | """ |
|
3175 | """ | |
3189 | dd, di = self.checksize() |
|
3176 | dd, di = self.checksize() | |
3190 | if dd: |
|
3177 | if dd: | |
3191 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) |
|
3178 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) | |
3192 | if di: |
|
3179 | if di: | |
3193 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) |
|
3180 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) | |
3194 |
|
3181 | |||
3195 | version = self._format_version |
|
3182 | version = self._format_version | |
3196 |
|
3183 | |||
3197 | # The verifier tells us what version revlog we should be. |
|
3184 | # The verifier tells us what version revlog we should be. | |
3198 | if version != state[b'expectedversion']: |
|
3185 | if version != state[b'expectedversion']: | |
3199 | yield revlogproblem( |
|
3186 | yield revlogproblem( | |
3200 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") |
|
3187 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") | |
3201 | % (self.display_id, version, state[b'expectedversion']) |
|
3188 | % (self.display_id, version, state[b'expectedversion']) | |
3202 | ) |
|
3189 | ) | |
3203 |
|
3190 | |||
3204 | state[b'skipread'] = set() |
|
3191 | state[b'skipread'] = set() | |
3205 | state[b'safe_renamed'] = set() |
|
3192 | state[b'safe_renamed'] = set() | |
3206 |
|
3193 | |||
3207 | for rev in self: |
|
3194 | for rev in self: | |
3208 | node = self.node(rev) |
|
3195 | node = self.node(rev) | |
3209 |
|
3196 | |||
3210 | # Verify contents. 4 cases to care about: |
|
3197 | # Verify contents. 4 cases to care about: | |
3211 | # |
|
3198 | # | |
3212 | # common: the most common case |
|
3199 | # common: the most common case | |
3213 | # rename: with a rename |
|
3200 | # rename: with a rename | |
3214 | # meta: file content starts with b'\1\n', the metadata |
|
3201 | # meta: file content starts with b'\1\n', the metadata | |
3215 | # header defined in filelog.py, but without a rename |
|
3202 | # header defined in filelog.py, but without a rename | |
3216 | # ext: content stored externally |
|
3203 | # ext: content stored externally | |
3217 | # |
|
3204 | # | |
3218 | # More formally, their differences are shown below: |
|
3205 | # More formally, their differences are shown below: | |
3219 | # |
|
3206 | # | |
3220 | # | common | rename | meta | ext |
|
3207 | # | common | rename | meta | ext | |
3221 | # ------------------------------------------------------- |
|
3208 | # ------------------------------------------------------- | |
3222 | # flags() | 0 | 0 | 0 | not 0 |
|
3209 | # flags() | 0 | 0 | 0 | not 0 | |
3223 | # renamed() | False | True | False | ? |
|
3210 | # renamed() | False | True | False | ? | |
3224 | # rawtext[0:2]=='\1\n'| False | True | True | ? |
|
3211 | # rawtext[0:2]=='\1\n'| False | True | True | ? | |
3225 | # |
|
3212 | # | |
3226 | # "rawtext" means the raw text stored in revlog data, which |
|
3213 | # "rawtext" means the raw text stored in revlog data, which | |
3227 | # could be retrieved by "rawdata(rev)". "text" |
|
3214 | # could be retrieved by "rawdata(rev)". "text" | |
3228 | # mentioned below is "revision(rev)". |
|
3215 | # mentioned below is "revision(rev)". | |
3229 | # |
|
3216 | # | |
3230 | # There are 3 different lengths stored physically: |
|
3217 | # There are 3 different lengths stored physically: | |
3231 | # 1. L1: rawsize, stored in revlog index |
|
3218 | # 1. L1: rawsize, stored in revlog index | |
3232 | # 2. L2: len(rawtext), stored in revlog data |
|
3219 | # 2. L2: len(rawtext), stored in revlog data | |
3233 | # 3. L3: len(text), stored in revlog data if flags==0, or |
|
3220 | # 3. L3: len(text), stored in revlog data if flags==0, or | |
3234 | # possibly somewhere else if flags!=0 |
|
3221 | # possibly somewhere else if flags!=0 | |
3235 | # |
|
3222 | # | |
3236 | # L1 should be equal to L2. L3 could be different from them. |
|
3223 | # L1 should be equal to L2. L3 could be different from them. | |
3237 | # "text" may or may not affect commit hash depending on flag |
|
3224 | # "text" may or may not affect commit hash depending on flag | |
3238 | # processors (see flagutil.addflagprocessor). |
|
3225 | # processors (see flagutil.addflagprocessor). | |
3239 | # |
|
3226 | # | |
3240 | # | common | rename | meta | ext |
|
3227 | # | common | rename | meta | ext | |
3241 | # ------------------------------------------------- |
|
3228 | # ------------------------------------------------- | |
3242 | # rawsize() | L1 | L1 | L1 | L1 |
|
3229 | # rawsize() | L1 | L1 | L1 | L1 | |
3243 | # size() | L1 | L2-LM | L1(*) | L1 (?) |
|
3230 | # size() | L1 | L2-LM | L1(*) | L1 (?) | |
3244 | # len(rawtext) | L2 | L2 | L2 | L2 |
|
3231 | # len(rawtext) | L2 | L2 | L2 | L2 | |
3245 | # len(text) | L2 | L2 | L2 | L3 |
|
3232 | # len(text) | L2 | L2 | L2 | L3 | |
3246 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) |
|
3233 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) | |
3247 | # |
|
3234 | # | |
3248 | # LM: length of metadata, depending on rawtext |
|
3235 | # LM: length of metadata, depending on rawtext | |
3249 | # (*): not ideal, see comment in filelog.size |
|
3236 | # (*): not ideal, see comment in filelog.size | |
3250 | # (?): could be "- len(meta)" if the resolved content has |
|
3237 | # (?): could be "- len(meta)" if the resolved content has | |
3251 | # rename metadata |
|
3238 | # rename metadata | |
3252 | # |
|
3239 | # | |
3253 | # Checks needed to be done: |
|
3240 | # Checks needed to be done: | |
3254 | # 1. length check: L1 == L2, in all cases. |
|
3241 | # 1. length check: L1 == L2, in all cases. | |
3255 | # 2. hash check: depending on flag processor, we may need to |
|
3242 | # 2. hash check: depending on flag processor, we may need to | |
3256 | # use either "text" (external), or "rawtext" (in revlog). |
|
3243 | # use either "text" (external), or "rawtext" (in revlog). | |
3257 |
|
3244 | |||
3258 | try: |
|
3245 | try: | |
3259 | skipflags = state.get(b'skipflags', 0) |
|
3246 | skipflags = state.get(b'skipflags', 0) | |
3260 | if skipflags: |
|
3247 | if skipflags: | |
3261 | skipflags &= self.flags(rev) |
|
3248 | skipflags &= self.flags(rev) | |
3262 |
|
3249 | |||
3263 | _verify_revision(self, skipflags, state, node) |
|
3250 | _verify_revision(self, skipflags, state, node) | |
3264 |
|
3251 | |||
3265 | l1 = self.rawsize(rev) |
|
3252 | l1 = self.rawsize(rev) | |
3266 | l2 = len(self.rawdata(node)) |
|
3253 | l2 = len(self.rawdata(node)) | |
3267 |
|
3254 | |||
3268 | if l1 != l2: |
|
3255 | if l1 != l2: | |
3269 | yield revlogproblem( |
|
3256 | yield revlogproblem( | |
3270 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), |
|
3257 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), | |
3271 | node=node, |
|
3258 | node=node, | |
3272 | ) |
|
3259 | ) | |
3273 |
|
3260 | |||
3274 | except error.CensoredNodeError: |
|
3261 | except error.CensoredNodeError: | |
3275 | if state[b'erroroncensored']: |
|
3262 | if state[b'erroroncensored']: | |
3276 | yield revlogproblem( |
|
3263 | yield revlogproblem( | |
3277 | error=_(b'censored file data'), node=node |
|
3264 | error=_(b'censored file data'), node=node | |
3278 | ) |
|
3265 | ) | |
3279 | state[b'skipread'].add(node) |
|
3266 | state[b'skipread'].add(node) | |
3280 | except Exception as e: |
|
3267 | except Exception as e: | |
3281 | yield revlogproblem( |
|
3268 | yield revlogproblem( | |
3282 | error=_(b'unpacking %s: %s') |
|
3269 | error=_(b'unpacking %s: %s') | |
3283 | % (short(node), stringutil.forcebytestr(e)), |
|
3270 | % (short(node), stringutil.forcebytestr(e)), | |
3284 | node=node, |
|
3271 | node=node, | |
3285 | ) |
|
3272 | ) | |
3286 | state[b'skipread'].add(node) |
|
3273 | state[b'skipread'].add(node) | |
3287 |
|
3274 | |||
3288 | def storageinfo( |
|
3275 | def storageinfo( | |
3289 | self, |
|
3276 | self, | |
3290 | exclusivefiles=False, |
|
3277 | exclusivefiles=False, | |
3291 | sharedfiles=False, |
|
3278 | sharedfiles=False, | |
3292 | revisionscount=False, |
|
3279 | revisionscount=False, | |
3293 | trackedsize=False, |
|
3280 | trackedsize=False, | |
3294 | storedsize=False, |
|
3281 | storedsize=False, | |
3295 | ): |
|
3282 | ): | |
3296 | d = {} |
|
3283 | d = {} | |
3297 |
|
3284 | |||
3298 | if exclusivefiles: |
|
3285 | if exclusivefiles: | |
3299 | d[b'exclusivefiles'] = [(self.opener, self._indexfile)] |
|
3286 | d[b'exclusivefiles'] = [(self.opener, self._indexfile)] | |
3300 | if not self._inline: |
|
3287 | if not self._inline: | |
3301 | d[b'exclusivefiles'].append((self.opener, self._datafile)) |
|
3288 | d[b'exclusivefiles'].append((self.opener, self._datafile)) | |
3302 |
|
3289 | |||
3303 | if sharedfiles: |
|
3290 | if sharedfiles: | |
3304 | d[b'sharedfiles'] = [] |
|
3291 | d[b'sharedfiles'] = [] | |
3305 |
|
3292 | |||
3306 | if revisionscount: |
|
3293 | if revisionscount: | |
3307 | d[b'revisionscount'] = len(self) |
|
3294 | d[b'revisionscount'] = len(self) | |
3308 |
|
3295 | |||
3309 | if trackedsize: |
|
3296 | if trackedsize: | |
3310 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) |
|
3297 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) | |
3311 |
|
3298 | |||
3312 | if storedsize: |
|
3299 | if storedsize: | |
3313 | d[b'storedsize'] = sum( |
|
3300 | d[b'storedsize'] = sum( | |
3314 | self.opener.stat(path).st_size for path in self.files() |
|
3301 | self.opener.stat(path).st_size for path in self.files() | |
3315 | ) |
|
3302 | ) | |
3316 |
|
3303 | |||
3317 | return d |
|
3304 | return d | |
3318 |
|
3305 | |||
3319 | def rewrite_sidedata(self, transaction, helpers, startrev, endrev): |
|
3306 | def rewrite_sidedata(self, transaction, helpers, startrev, endrev): | |
3320 | if not self.hassidedata: |
|
3307 | if not self.hassidedata: | |
3321 | return |
|
3308 | return | |
3322 | # revlog formats with sidedata support does not support inline |
|
3309 | # revlog formats with sidedata support does not support inline | |
3323 | assert not self._inline |
|
3310 | assert not self._inline | |
3324 | if not helpers[1] and not helpers[2]: |
|
3311 | if not helpers[1] and not helpers[2]: | |
3325 | # Nothing to generate or remove |
|
3312 | # Nothing to generate or remove | |
3326 | return |
|
3313 | return | |
3327 |
|
3314 | |||
3328 | new_entries = [] |
|
3315 | new_entries = [] | |
3329 | # append the new sidedata |
|
3316 | # append the new sidedata | |
3330 | with self._writing(transaction): |
|
3317 | with self._writing(transaction): | |
3331 | ifh, dfh, sdfh = self._writinghandles |
|
3318 | ifh, dfh, sdfh = self._writinghandles | |
3332 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) |
|
3319 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) | |
3333 |
|
3320 | |||
3334 | current_offset = sdfh.tell() |
|
3321 | current_offset = sdfh.tell() | |
3335 | for rev in range(startrev, endrev + 1): |
|
3322 | for rev in range(startrev, endrev + 1): | |
3336 | entry = self.index[rev] |
|
3323 | entry = self.index[rev] | |
3337 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( |
|
3324 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( | |
3338 | store=self, |
|
3325 | store=self, | |
3339 | sidedata_helpers=helpers, |
|
3326 | sidedata_helpers=helpers, | |
3340 | sidedata={}, |
|
3327 | sidedata={}, | |
3341 | rev=rev, |
|
3328 | rev=rev, | |
3342 | ) |
|
3329 | ) | |
3343 |
|
3330 | |||
3344 | serialized_sidedata = sidedatautil.serialize_sidedata( |
|
3331 | serialized_sidedata = sidedatautil.serialize_sidedata( | |
3345 | new_sidedata |
|
3332 | new_sidedata | |
3346 | ) |
|
3333 | ) | |
3347 |
|
3334 | |||
3348 | sidedata_compression_mode = COMP_MODE_INLINE |
|
3335 | sidedata_compression_mode = COMP_MODE_INLINE | |
3349 | if serialized_sidedata and self.hassidedata: |
|
3336 | if serialized_sidedata and self.hassidedata: | |
3350 | sidedata_compression_mode = COMP_MODE_PLAIN |
|
3337 | sidedata_compression_mode = COMP_MODE_PLAIN | |
3351 | h, comp_sidedata = self.compress(serialized_sidedata) |
|
3338 | h, comp_sidedata = self.compress(serialized_sidedata) | |
3352 | if ( |
|
3339 | if ( | |
3353 | h != b'u' |
|
3340 | h != b'u' | |
3354 | and comp_sidedata[0] != b'\0' |
|
3341 | and comp_sidedata[0] != b'\0' | |
3355 | and len(comp_sidedata) < len(serialized_sidedata) |
|
3342 | and len(comp_sidedata) < len(serialized_sidedata) | |
3356 | ): |
|
3343 | ): | |
3357 | assert not h |
|
3344 | assert not h | |
3358 | if ( |
|
3345 | if ( | |
3359 | comp_sidedata[0] |
|
3346 | comp_sidedata[0] | |
3360 | == self._docket.default_compression_header |
|
3347 | == self._docket.default_compression_header | |
3361 | ): |
|
3348 | ): | |
3362 | sidedata_compression_mode = COMP_MODE_DEFAULT |
|
3349 | sidedata_compression_mode = COMP_MODE_DEFAULT | |
3363 | serialized_sidedata = comp_sidedata |
|
3350 | serialized_sidedata = comp_sidedata | |
3364 | else: |
|
3351 | else: | |
3365 | sidedata_compression_mode = COMP_MODE_INLINE |
|
3352 | sidedata_compression_mode = COMP_MODE_INLINE | |
3366 | serialized_sidedata = comp_sidedata |
|
3353 | serialized_sidedata = comp_sidedata | |
3367 | if entry[8] != 0 or entry[9] != 0: |
|
3354 | if entry[8] != 0 or entry[9] != 0: | |
3368 | # rewriting entries that already have sidedata is not |
|
3355 | # rewriting entries that already have sidedata is not | |
3369 | # supported yet, because it introduces garbage data in the |
|
3356 | # supported yet, because it introduces garbage data in the | |
3370 | # revlog. |
|
3357 | # revlog. | |
3371 | msg = b"rewriting existing sidedata is not supported yet" |
|
3358 | msg = b"rewriting existing sidedata is not supported yet" | |
3372 | raise error.Abort(msg) |
|
3359 | raise error.Abort(msg) | |
3373 |
|
3360 | |||
3374 | # Apply (potential) flags to add and to remove after running |
|
3361 | # Apply (potential) flags to add and to remove after running | |
3375 | # the sidedata helpers |
|
3362 | # the sidedata helpers | |
3376 | new_offset_flags = entry[0] | flags[0] & ~flags[1] |
|
3363 | new_offset_flags = entry[0] | flags[0] & ~flags[1] | |
3377 | entry_update = ( |
|
3364 | entry_update = ( | |
3378 | current_offset, |
|
3365 | current_offset, | |
3379 | len(serialized_sidedata), |
|
3366 | len(serialized_sidedata), | |
3380 | new_offset_flags, |
|
3367 | new_offset_flags, | |
3381 | sidedata_compression_mode, |
|
3368 | sidedata_compression_mode, | |
3382 | ) |
|
3369 | ) | |
3383 |
|
3370 | |||
3384 | # the sidedata computation might have move the file cursors around |
|
3371 | # the sidedata computation might have move the file cursors around | |
3385 | sdfh.seek(current_offset, os.SEEK_SET) |
|
3372 | sdfh.seek(current_offset, os.SEEK_SET) | |
3386 | sdfh.write(serialized_sidedata) |
|
3373 | sdfh.write(serialized_sidedata) | |
3387 | new_entries.append(entry_update) |
|
3374 | new_entries.append(entry_update) | |
3388 | current_offset += len(serialized_sidedata) |
|
3375 | current_offset += len(serialized_sidedata) | |
3389 | self._docket.sidedata_end = sdfh.tell() |
|
3376 | self._docket.sidedata_end = sdfh.tell() | |
3390 |
|
3377 | |||
3391 | # rewrite the new index entries |
|
3378 | # rewrite the new index entries | |
3392 | ifh.seek(startrev * self.index.entry_size) |
|
3379 | ifh.seek(startrev * self.index.entry_size) | |
3393 | for i, e in enumerate(new_entries): |
|
3380 | for i, e in enumerate(new_entries): | |
3394 | rev = startrev + i |
|
3381 | rev = startrev + i | |
3395 | self.index.replace_sidedata_info(rev, *e) |
|
3382 | self.index.replace_sidedata_info(rev, *e) | |
3396 | packed = self.index.entry_binary(rev) |
|
3383 | packed = self.index.entry_binary(rev) | |
3397 | if rev == 0 and self._docket is None: |
|
3384 | if rev == 0 and self._docket is None: | |
3398 | header = self._format_flags | self._format_version |
|
3385 | header = self._format_flags | self._format_version | |
3399 | header = self.index.pack_header(header) |
|
3386 | header = self.index.pack_header(header) | |
3400 | packed = header + packed |
|
3387 | packed = header + packed | |
3401 | ifh.write(packed) |
|
3388 | ifh.write(packed) |
@@ -1,58 +1,80 b'' | |||||
1 | # mercurial.revlogutils -- basic utilities for revlog |
|
1 | # mercurial.revlogutils -- basic utilities for revlog | |
2 | # |
|
2 | # | |
3 | # Copyright 2019 Pierre-Yves David <pierre-yves.david@octobus.net> |
|
3 | # Copyright 2019 Pierre-Yves David <pierre-yves.david@octobus.net> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
|
10 | from ..thirdparty import attr | |||
10 | from ..interfaces import repository |
|
11 | from ..interfaces import repository | |
11 |
|
12 | |||
12 | # See mercurial.revlogutils.constants for doc |
|
13 | # See mercurial.revlogutils.constants for doc | |
13 | COMP_MODE_INLINE = 2 |
|
14 | COMP_MODE_INLINE = 2 | |
14 |
|
15 | |||
15 |
|
16 | |||
16 | def offset_type(offset, type): |
|
17 | def offset_type(offset, type): | |
17 | if (type & ~repository.REVISION_FLAGS_KNOWN) != 0: |
|
18 | if (type & ~repository.REVISION_FLAGS_KNOWN) != 0: | |
18 | raise ValueError(b'unknown revlog index flags: %d' % type) |
|
19 | raise ValueError(b'unknown revlog index flags: %d' % type) | |
19 | return int(int(offset) << 16 | type) |
|
20 | return int(int(offset) << 16 | type) | |
20 |
|
21 | |||
21 |
|
22 | |||
22 | def entry( |
|
23 | def entry( | |
23 | data_offset, |
|
24 | data_offset, | |
24 | data_compressed_length, |
|
25 | data_compressed_length, | |
25 | data_delta_base, |
|
26 | data_delta_base, | |
26 | link_rev, |
|
27 | link_rev, | |
27 | parent_rev_1, |
|
28 | parent_rev_1, | |
28 | parent_rev_2, |
|
29 | parent_rev_2, | |
29 | node_id, |
|
30 | node_id, | |
30 | flags=0, |
|
31 | flags=0, | |
31 | data_uncompressed_length=-1, |
|
32 | data_uncompressed_length=-1, | |
32 | data_compression_mode=COMP_MODE_INLINE, |
|
33 | data_compression_mode=COMP_MODE_INLINE, | |
33 | sidedata_offset=0, |
|
34 | sidedata_offset=0, | |
34 | sidedata_compressed_length=0, |
|
35 | sidedata_compressed_length=0, | |
35 | sidedata_compression_mode=COMP_MODE_INLINE, |
|
36 | sidedata_compression_mode=COMP_MODE_INLINE, | |
36 | ): |
|
37 | ): | |
37 | """Build one entry from symbolic name |
|
38 | """Build one entry from symbolic name | |
38 |
|
39 | |||
39 | This is useful to abstract the actual detail of how we build the entry |
|
40 | This is useful to abstract the actual detail of how we build the entry | |
40 | tuple for caller who don't care about it. |
|
41 | tuple for caller who don't care about it. | |
41 |
|
42 | |||
42 | This should always be called using keyword arguments. Some arguments have |
|
43 | This should always be called using keyword arguments. Some arguments have | |
43 | default value, this match the value used by index version that does not store such data. |
|
44 | default value, this match the value used by index version that does not store such data. | |
44 | """ |
|
45 | """ | |
45 | return ( |
|
46 | return ( | |
46 | offset_type(data_offset, flags), |
|
47 | offset_type(data_offset, flags), | |
47 | data_compressed_length, |
|
48 | data_compressed_length, | |
48 | data_uncompressed_length, |
|
49 | data_uncompressed_length, | |
49 | data_delta_base, |
|
50 | data_delta_base, | |
50 | link_rev, |
|
51 | link_rev, | |
51 | parent_rev_1, |
|
52 | parent_rev_1, | |
52 | parent_rev_2, |
|
53 | parent_rev_2, | |
53 | node_id, |
|
54 | node_id, | |
54 | sidedata_offset, |
|
55 | sidedata_offset, | |
55 | sidedata_compressed_length, |
|
56 | sidedata_compressed_length, | |
56 | data_compression_mode, |
|
57 | data_compression_mode, | |
57 | sidedata_compression_mode, |
|
58 | sidedata_compression_mode, | |
58 | ) |
|
59 | ) | |
|
60 | ||||
|
61 | ||||
|
62 | @attr.s(slots=True, frozen=True) | |||
|
63 | class revisioninfo(object): | |||
|
64 | """Information about a revision that allows building its fulltext | |||
|
65 | node: expected hash of the revision | |||
|
66 | p1, p2: parent revs of the revision | |||
|
67 | btext: built text cache consisting of a one-element list | |||
|
68 | cachedelta: (baserev, uncompressed_delta) or None | |||
|
69 | flags: flags associated to the revision storage | |||
|
70 | ||||
|
71 | One of btext[0] or cachedelta must be set. | |||
|
72 | """ | |||
|
73 | ||||
|
74 | node = attr.ib() | |||
|
75 | p1 = attr.ib() | |||
|
76 | p2 = attr.ib() | |||
|
77 | btext = attr.ib() | |||
|
78 | textlen = attr.ib() | |||
|
79 | cachedelta = attr.ib() | |||
|
80 | flags = attr.ib() |
@@ -1,1106 +1,1106 b'' | |||||
1 | # revlogdeltas.py - Logic around delta computation for revlog |
|
1 | # revlogdeltas.py - Logic around delta computation for revlog | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> |
|
3 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> | |
4 | # Copyright 2018 Octobus <contact@octobus.net> |
|
4 | # Copyright 2018 Octobus <contact@octobus.net> | |
5 | # |
|
5 | # | |
6 | # This software may be used and distributed according to the terms of the |
|
6 | # This software may be used and distributed according to the terms of the | |
7 | # GNU General Public License version 2 or any later version. |
|
7 | # GNU General Public License version 2 or any later version. | |
8 | """Helper class to compute deltas stored inside revlogs""" |
|
8 | """Helper class to compute deltas stored inside revlogs""" | |
9 |
|
9 | |||
10 | from __future__ import absolute_import |
|
10 | from __future__ import absolute_import | |
11 |
|
11 | |||
12 | import collections |
|
12 | import collections | |
13 | import struct |
|
13 | import struct | |
14 |
|
14 | |||
15 | # import stuff from node for others to import from revlog |
|
15 | # import stuff from node for others to import from revlog | |
16 | from ..node import nullrev |
|
16 | from ..node import nullrev | |
17 | from ..i18n import _ |
|
17 | from ..i18n import _ | |
18 | from ..pycompat import getattr |
|
18 | from ..pycompat import getattr | |
19 |
|
19 | |||
20 | from .constants import ( |
|
20 | from .constants import ( | |
21 | REVIDX_ISCENSORED, |
|
21 | REVIDX_ISCENSORED, | |
22 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
22 | REVIDX_RAWTEXT_CHANGING_FLAGS, | |
23 | ) |
|
23 | ) | |
24 |
|
24 | |||
25 | from ..thirdparty import attr |
|
25 | from ..thirdparty import attr | |
26 |
|
26 | |||
27 | from .. import ( |
|
27 | from .. import ( | |
28 | error, |
|
28 | error, | |
29 | mdiff, |
|
29 | mdiff, | |
30 | util, |
|
30 | util, | |
31 | ) |
|
31 | ) | |
32 |
|
32 | |||
33 | from . import flagutil |
|
33 | from . import flagutil | |
34 |
|
34 | |||
35 | # maximum <delta-chain-data>/<revision-text-length> ratio |
|
35 | # maximum <delta-chain-data>/<revision-text-length> ratio | |
36 | LIMIT_DELTA2TEXT = 2 |
|
36 | LIMIT_DELTA2TEXT = 2 | |
37 |
|
37 | |||
38 |
|
38 | |||
39 | class _testrevlog(object): |
|
39 | class _testrevlog(object): | |
40 | """minimalist fake revlog to use in doctests""" |
|
40 | """minimalist fake revlog to use in doctests""" | |
41 |
|
41 | |||
42 | def __init__(self, data, density=0.5, mingap=0, snapshot=()): |
|
42 | def __init__(self, data, density=0.5, mingap=0, snapshot=()): | |
43 | """data is an list of revision payload boundaries""" |
|
43 | """data is an list of revision payload boundaries""" | |
44 | self._data = data |
|
44 | self._data = data | |
45 | self._srdensitythreshold = density |
|
45 | self._srdensitythreshold = density | |
46 | self._srmingapsize = mingap |
|
46 | self._srmingapsize = mingap | |
47 | self._snapshot = set(snapshot) |
|
47 | self._snapshot = set(snapshot) | |
48 | self.index = None |
|
48 | self.index = None | |
49 |
|
49 | |||
50 | def start(self, rev): |
|
50 | def start(self, rev): | |
51 | if rev == nullrev: |
|
51 | if rev == nullrev: | |
52 | return 0 |
|
52 | return 0 | |
53 | if rev == 0: |
|
53 | if rev == 0: | |
54 | return 0 |
|
54 | return 0 | |
55 | return self._data[rev - 1] |
|
55 | return self._data[rev - 1] | |
56 |
|
56 | |||
57 | def end(self, rev): |
|
57 | def end(self, rev): | |
58 | if rev == nullrev: |
|
58 | if rev == nullrev: | |
59 | return 0 |
|
59 | return 0 | |
60 | return self._data[rev] |
|
60 | return self._data[rev] | |
61 |
|
61 | |||
62 | def length(self, rev): |
|
62 | def length(self, rev): | |
63 | return self.end(rev) - self.start(rev) |
|
63 | return self.end(rev) - self.start(rev) | |
64 |
|
64 | |||
65 | def __len__(self): |
|
65 | def __len__(self): | |
66 | return len(self._data) |
|
66 | return len(self._data) | |
67 |
|
67 | |||
68 | def issnapshot(self, rev): |
|
68 | def issnapshot(self, rev): | |
69 | if rev == nullrev: |
|
69 | if rev == nullrev: | |
70 | return True |
|
70 | return True | |
71 | return rev in self._snapshot |
|
71 | return rev in self._snapshot | |
72 |
|
72 | |||
73 |
|
73 | |||
74 | def slicechunk(revlog, revs, targetsize=None): |
|
74 | def slicechunk(revlog, revs, targetsize=None): | |
75 | """slice revs to reduce the amount of unrelated data to be read from disk. |
|
75 | """slice revs to reduce the amount of unrelated data to be read from disk. | |
76 |
|
76 | |||
77 | ``revs`` is sliced into groups that should be read in one time. |
|
77 | ``revs`` is sliced into groups that should be read in one time. | |
78 | Assume that revs are sorted. |
|
78 | Assume that revs are sorted. | |
79 |
|
79 | |||
80 | The initial chunk is sliced until the overall density (payload/chunks-span |
|
80 | The initial chunk is sliced until the overall density (payload/chunks-span | |
81 | ratio) is above `revlog._srdensitythreshold`. No gap smaller than |
|
81 | ratio) is above `revlog._srdensitythreshold`. No gap smaller than | |
82 | `revlog._srmingapsize` is skipped. |
|
82 | `revlog._srmingapsize` is skipped. | |
83 |
|
83 | |||
84 | If `targetsize` is set, no chunk larger than `targetsize` will be yield. |
|
84 | If `targetsize` is set, no chunk larger than `targetsize` will be yield. | |
85 | For consistency with other slicing choice, this limit won't go lower than |
|
85 | For consistency with other slicing choice, this limit won't go lower than | |
86 | `revlog._srmingapsize`. |
|
86 | `revlog._srmingapsize`. | |
87 |
|
87 | |||
88 | If individual revisions chunk are larger than this limit, they will still |
|
88 | If individual revisions chunk are larger than this limit, they will still | |
89 | be raised individually. |
|
89 | be raised individually. | |
90 |
|
90 | |||
91 | >>> data = [ |
|
91 | >>> data = [ | |
92 | ... 5, #00 (5) |
|
92 | ... 5, #00 (5) | |
93 | ... 10, #01 (5) |
|
93 | ... 10, #01 (5) | |
94 | ... 12, #02 (2) |
|
94 | ... 12, #02 (2) | |
95 | ... 12, #03 (empty) |
|
95 | ... 12, #03 (empty) | |
96 | ... 27, #04 (15) |
|
96 | ... 27, #04 (15) | |
97 | ... 31, #05 (4) |
|
97 | ... 31, #05 (4) | |
98 | ... 31, #06 (empty) |
|
98 | ... 31, #06 (empty) | |
99 | ... 42, #07 (11) |
|
99 | ... 42, #07 (11) | |
100 | ... 47, #08 (5) |
|
100 | ... 47, #08 (5) | |
101 | ... 47, #09 (empty) |
|
101 | ... 47, #09 (empty) | |
102 | ... 48, #10 (1) |
|
102 | ... 48, #10 (1) | |
103 | ... 51, #11 (3) |
|
103 | ... 51, #11 (3) | |
104 | ... 74, #12 (23) |
|
104 | ... 74, #12 (23) | |
105 | ... 85, #13 (11) |
|
105 | ... 85, #13 (11) | |
106 | ... 86, #14 (1) |
|
106 | ... 86, #14 (1) | |
107 | ... 91, #15 (5) |
|
107 | ... 91, #15 (5) | |
108 | ... ] |
|
108 | ... ] | |
109 | >>> revlog = _testrevlog(data, snapshot=range(16)) |
|
109 | >>> revlog = _testrevlog(data, snapshot=range(16)) | |
110 |
|
110 | |||
111 | >>> list(slicechunk(revlog, list(range(16)))) |
|
111 | >>> list(slicechunk(revlog, list(range(16)))) | |
112 | [[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]] |
|
112 | [[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]] | |
113 | >>> list(slicechunk(revlog, [0, 15])) |
|
113 | >>> list(slicechunk(revlog, [0, 15])) | |
114 | [[0], [15]] |
|
114 | [[0], [15]] | |
115 | >>> list(slicechunk(revlog, [0, 11, 15])) |
|
115 | >>> list(slicechunk(revlog, [0, 11, 15])) | |
116 | [[0], [11], [15]] |
|
116 | [[0], [11], [15]] | |
117 | >>> list(slicechunk(revlog, [0, 11, 13, 15])) |
|
117 | >>> list(slicechunk(revlog, [0, 11, 13, 15])) | |
118 | [[0], [11, 13, 15]] |
|
118 | [[0], [11, 13, 15]] | |
119 | >>> list(slicechunk(revlog, [1, 2, 3, 5, 8, 10, 11, 14])) |
|
119 | >>> list(slicechunk(revlog, [1, 2, 3, 5, 8, 10, 11, 14])) | |
120 | [[1, 2], [5, 8, 10, 11], [14]] |
|
120 | [[1, 2], [5, 8, 10, 11], [14]] | |
121 |
|
121 | |||
122 | Slicing with a maximum chunk size |
|
122 | Slicing with a maximum chunk size | |
123 | >>> list(slicechunk(revlog, [0, 11, 13, 15], targetsize=15)) |
|
123 | >>> list(slicechunk(revlog, [0, 11, 13, 15], targetsize=15)) | |
124 | [[0], [11], [13], [15]] |
|
124 | [[0], [11], [13], [15]] | |
125 | >>> list(slicechunk(revlog, [0, 11, 13, 15], targetsize=20)) |
|
125 | >>> list(slicechunk(revlog, [0, 11, 13, 15], targetsize=20)) | |
126 | [[0], [11], [13, 15]] |
|
126 | [[0], [11], [13, 15]] | |
127 |
|
127 | |||
128 | Slicing involving nullrev |
|
128 | Slicing involving nullrev | |
129 | >>> list(slicechunk(revlog, [-1, 0, 11, 13, 15], targetsize=20)) |
|
129 | >>> list(slicechunk(revlog, [-1, 0, 11, 13, 15], targetsize=20)) | |
130 | [[-1, 0], [11], [13, 15]] |
|
130 | [[-1, 0], [11], [13, 15]] | |
131 | >>> list(slicechunk(revlog, [-1, 13, 15], targetsize=5)) |
|
131 | >>> list(slicechunk(revlog, [-1, 13, 15], targetsize=5)) | |
132 | [[-1], [13], [15]] |
|
132 | [[-1], [13], [15]] | |
133 | """ |
|
133 | """ | |
134 | if targetsize is not None: |
|
134 | if targetsize is not None: | |
135 | targetsize = max(targetsize, revlog._srmingapsize) |
|
135 | targetsize = max(targetsize, revlog._srmingapsize) | |
136 | # targetsize should not be specified when evaluating delta candidates: |
|
136 | # targetsize should not be specified when evaluating delta candidates: | |
137 | # * targetsize is used to ensure we stay within specification when reading, |
|
137 | # * targetsize is used to ensure we stay within specification when reading, | |
138 | densityslicing = getattr(revlog.index, 'slicechunktodensity', None) |
|
138 | densityslicing = getattr(revlog.index, 'slicechunktodensity', None) | |
139 | if densityslicing is None: |
|
139 | if densityslicing is None: | |
140 | densityslicing = lambda x, y, z: _slicechunktodensity(revlog, x, y, z) |
|
140 | densityslicing = lambda x, y, z: _slicechunktodensity(revlog, x, y, z) | |
141 | for chunk in densityslicing( |
|
141 | for chunk in densityslicing( | |
142 | revs, revlog._srdensitythreshold, revlog._srmingapsize |
|
142 | revs, revlog._srdensitythreshold, revlog._srmingapsize | |
143 | ): |
|
143 | ): | |
144 | for subchunk in _slicechunktosize(revlog, chunk, targetsize): |
|
144 | for subchunk in _slicechunktosize(revlog, chunk, targetsize): | |
145 | yield subchunk |
|
145 | yield subchunk | |
146 |
|
146 | |||
147 |
|
147 | |||
148 | def _slicechunktosize(revlog, revs, targetsize=None): |
|
148 | def _slicechunktosize(revlog, revs, targetsize=None): | |
149 | """slice revs to match the target size |
|
149 | """slice revs to match the target size | |
150 |
|
150 | |||
151 | This is intended to be used on chunk that density slicing selected by that |
|
151 | This is intended to be used on chunk that density slicing selected by that | |
152 | are still too large compared to the read garantee of revlog. This might |
|
152 | are still too large compared to the read garantee of revlog. This might | |
153 | happens when "minimal gap size" interrupted the slicing or when chain are |
|
153 | happens when "minimal gap size" interrupted the slicing or when chain are | |
154 | built in a way that create large blocks next to each other. |
|
154 | built in a way that create large blocks next to each other. | |
155 |
|
155 | |||
156 | >>> data = [ |
|
156 | >>> data = [ | |
157 | ... 3, #0 (3) |
|
157 | ... 3, #0 (3) | |
158 | ... 5, #1 (2) |
|
158 | ... 5, #1 (2) | |
159 | ... 6, #2 (1) |
|
159 | ... 6, #2 (1) | |
160 | ... 8, #3 (2) |
|
160 | ... 8, #3 (2) | |
161 | ... 8, #4 (empty) |
|
161 | ... 8, #4 (empty) | |
162 | ... 11, #5 (3) |
|
162 | ... 11, #5 (3) | |
163 | ... 12, #6 (1) |
|
163 | ... 12, #6 (1) | |
164 | ... 13, #7 (1) |
|
164 | ... 13, #7 (1) | |
165 | ... 14, #8 (1) |
|
165 | ... 14, #8 (1) | |
166 | ... ] |
|
166 | ... ] | |
167 |
|
167 | |||
168 | == All snapshots cases == |
|
168 | == All snapshots cases == | |
169 | >>> revlog = _testrevlog(data, snapshot=range(9)) |
|
169 | >>> revlog = _testrevlog(data, snapshot=range(9)) | |
170 |
|
170 | |||
171 | Cases where chunk is already small enough |
|
171 | Cases where chunk is already small enough | |
172 | >>> list(_slicechunktosize(revlog, [0], 3)) |
|
172 | >>> list(_slicechunktosize(revlog, [0], 3)) | |
173 | [[0]] |
|
173 | [[0]] | |
174 | >>> list(_slicechunktosize(revlog, [6, 7], 3)) |
|
174 | >>> list(_slicechunktosize(revlog, [6, 7], 3)) | |
175 | [[6, 7]] |
|
175 | [[6, 7]] | |
176 | >>> list(_slicechunktosize(revlog, [0], None)) |
|
176 | >>> list(_slicechunktosize(revlog, [0], None)) | |
177 | [[0]] |
|
177 | [[0]] | |
178 | >>> list(_slicechunktosize(revlog, [6, 7], None)) |
|
178 | >>> list(_slicechunktosize(revlog, [6, 7], None)) | |
179 | [[6, 7]] |
|
179 | [[6, 7]] | |
180 |
|
180 | |||
181 | cases where we need actual slicing |
|
181 | cases where we need actual slicing | |
182 | >>> list(_slicechunktosize(revlog, [0, 1], 3)) |
|
182 | >>> list(_slicechunktosize(revlog, [0, 1], 3)) | |
183 | [[0], [1]] |
|
183 | [[0], [1]] | |
184 | >>> list(_slicechunktosize(revlog, [1, 3], 3)) |
|
184 | >>> list(_slicechunktosize(revlog, [1, 3], 3)) | |
185 | [[1], [3]] |
|
185 | [[1], [3]] | |
186 | >>> list(_slicechunktosize(revlog, [1, 2, 3], 3)) |
|
186 | >>> list(_slicechunktosize(revlog, [1, 2, 3], 3)) | |
187 | [[1, 2], [3]] |
|
187 | [[1, 2], [3]] | |
188 | >>> list(_slicechunktosize(revlog, [3, 5], 3)) |
|
188 | >>> list(_slicechunktosize(revlog, [3, 5], 3)) | |
189 | [[3], [5]] |
|
189 | [[3], [5]] | |
190 | >>> list(_slicechunktosize(revlog, [3, 4, 5], 3)) |
|
190 | >>> list(_slicechunktosize(revlog, [3, 4, 5], 3)) | |
191 | [[3], [5]] |
|
191 | [[3], [5]] | |
192 | >>> list(_slicechunktosize(revlog, [5, 6, 7, 8], 3)) |
|
192 | >>> list(_slicechunktosize(revlog, [5, 6, 7, 8], 3)) | |
193 | [[5], [6, 7, 8]] |
|
193 | [[5], [6, 7, 8]] | |
194 | >>> list(_slicechunktosize(revlog, [0, 1, 2, 3, 4, 5, 6, 7, 8], 3)) |
|
194 | >>> list(_slicechunktosize(revlog, [0, 1, 2, 3, 4, 5, 6, 7, 8], 3)) | |
195 | [[0], [1, 2], [3], [5], [6, 7, 8]] |
|
195 | [[0], [1, 2], [3], [5], [6, 7, 8]] | |
196 |
|
196 | |||
197 | Case with too large individual chunk (must return valid chunk) |
|
197 | Case with too large individual chunk (must return valid chunk) | |
198 | >>> list(_slicechunktosize(revlog, [0, 1], 2)) |
|
198 | >>> list(_slicechunktosize(revlog, [0, 1], 2)) | |
199 | [[0], [1]] |
|
199 | [[0], [1]] | |
200 | >>> list(_slicechunktosize(revlog, [1, 3], 1)) |
|
200 | >>> list(_slicechunktosize(revlog, [1, 3], 1)) | |
201 | [[1], [3]] |
|
201 | [[1], [3]] | |
202 | >>> list(_slicechunktosize(revlog, [3, 4, 5], 2)) |
|
202 | >>> list(_slicechunktosize(revlog, [3, 4, 5], 2)) | |
203 | [[3], [5]] |
|
203 | [[3], [5]] | |
204 |
|
204 | |||
205 | == No Snapshot cases == |
|
205 | == No Snapshot cases == | |
206 | >>> revlog = _testrevlog(data) |
|
206 | >>> revlog = _testrevlog(data) | |
207 |
|
207 | |||
208 | Cases where chunk is already small enough |
|
208 | Cases where chunk is already small enough | |
209 | >>> list(_slicechunktosize(revlog, [0], 3)) |
|
209 | >>> list(_slicechunktosize(revlog, [0], 3)) | |
210 | [[0]] |
|
210 | [[0]] | |
211 | >>> list(_slicechunktosize(revlog, [6, 7], 3)) |
|
211 | >>> list(_slicechunktosize(revlog, [6, 7], 3)) | |
212 | [[6, 7]] |
|
212 | [[6, 7]] | |
213 | >>> list(_slicechunktosize(revlog, [0], None)) |
|
213 | >>> list(_slicechunktosize(revlog, [0], None)) | |
214 | [[0]] |
|
214 | [[0]] | |
215 | >>> list(_slicechunktosize(revlog, [6, 7], None)) |
|
215 | >>> list(_slicechunktosize(revlog, [6, 7], None)) | |
216 | [[6, 7]] |
|
216 | [[6, 7]] | |
217 |
|
217 | |||
218 | cases where we need actual slicing |
|
218 | cases where we need actual slicing | |
219 | >>> list(_slicechunktosize(revlog, [0, 1], 3)) |
|
219 | >>> list(_slicechunktosize(revlog, [0, 1], 3)) | |
220 | [[0], [1]] |
|
220 | [[0], [1]] | |
221 | >>> list(_slicechunktosize(revlog, [1, 3], 3)) |
|
221 | >>> list(_slicechunktosize(revlog, [1, 3], 3)) | |
222 | [[1], [3]] |
|
222 | [[1], [3]] | |
223 | >>> list(_slicechunktosize(revlog, [1, 2, 3], 3)) |
|
223 | >>> list(_slicechunktosize(revlog, [1, 2, 3], 3)) | |
224 | [[1], [2, 3]] |
|
224 | [[1], [2, 3]] | |
225 | >>> list(_slicechunktosize(revlog, [3, 5], 3)) |
|
225 | >>> list(_slicechunktosize(revlog, [3, 5], 3)) | |
226 | [[3], [5]] |
|
226 | [[3], [5]] | |
227 | >>> list(_slicechunktosize(revlog, [3, 4, 5], 3)) |
|
227 | >>> list(_slicechunktosize(revlog, [3, 4, 5], 3)) | |
228 | [[3], [4, 5]] |
|
228 | [[3], [4, 5]] | |
229 | >>> list(_slicechunktosize(revlog, [5, 6, 7, 8], 3)) |
|
229 | >>> list(_slicechunktosize(revlog, [5, 6, 7, 8], 3)) | |
230 | [[5], [6, 7, 8]] |
|
230 | [[5], [6, 7, 8]] | |
231 | >>> list(_slicechunktosize(revlog, [0, 1, 2, 3, 4, 5, 6, 7, 8], 3)) |
|
231 | >>> list(_slicechunktosize(revlog, [0, 1, 2, 3, 4, 5, 6, 7, 8], 3)) | |
232 | [[0], [1, 2], [3], [5], [6, 7, 8]] |
|
232 | [[0], [1, 2], [3], [5], [6, 7, 8]] | |
233 |
|
233 | |||
234 | Case with too large individual chunk (must return valid chunk) |
|
234 | Case with too large individual chunk (must return valid chunk) | |
235 | >>> list(_slicechunktosize(revlog, [0, 1], 2)) |
|
235 | >>> list(_slicechunktosize(revlog, [0, 1], 2)) | |
236 | [[0], [1]] |
|
236 | [[0], [1]] | |
237 | >>> list(_slicechunktosize(revlog, [1, 3], 1)) |
|
237 | >>> list(_slicechunktosize(revlog, [1, 3], 1)) | |
238 | [[1], [3]] |
|
238 | [[1], [3]] | |
239 | >>> list(_slicechunktosize(revlog, [3, 4, 5], 2)) |
|
239 | >>> list(_slicechunktosize(revlog, [3, 4, 5], 2)) | |
240 | [[3], [5]] |
|
240 | [[3], [5]] | |
241 |
|
241 | |||
242 | == mixed case == |
|
242 | == mixed case == | |
243 | >>> revlog = _testrevlog(data, snapshot=[0, 1, 2]) |
|
243 | >>> revlog = _testrevlog(data, snapshot=[0, 1, 2]) | |
244 | >>> list(_slicechunktosize(revlog, list(range(9)), 5)) |
|
244 | >>> list(_slicechunktosize(revlog, list(range(9)), 5)) | |
245 | [[0, 1], [2], [3, 4, 5], [6, 7, 8]] |
|
245 | [[0, 1], [2], [3, 4, 5], [6, 7, 8]] | |
246 | """ |
|
246 | """ | |
247 | assert targetsize is None or 0 <= targetsize |
|
247 | assert targetsize is None or 0 <= targetsize | |
248 | startdata = revlog.start(revs[0]) |
|
248 | startdata = revlog.start(revs[0]) | |
249 | enddata = revlog.end(revs[-1]) |
|
249 | enddata = revlog.end(revs[-1]) | |
250 | fullspan = enddata - startdata |
|
250 | fullspan = enddata - startdata | |
251 | if targetsize is None or fullspan <= targetsize: |
|
251 | if targetsize is None or fullspan <= targetsize: | |
252 | yield revs |
|
252 | yield revs | |
253 | return |
|
253 | return | |
254 |
|
254 | |||
255 | startrevidx = 0 |
|
255 | startrevidx = 0 | |
256 | endrevidx = 1 |
|
256 | endrevidx = 1 | |
257 | iterrevs = enumerate(revs) |
|
257 | iterrevs = enumerate(revs) | |
258 | next(iterrevs) # skip first rev. |
|
258 | next(iterrevs) # skip first rev. | |
259 | # first step: get snapshots out of the way |
|
259 | # first step: get snapshots out of the way | |
260 | for idx, r in iterrevs: |
|
260 | for idx, r in iterrevs: | |
261 | span = revlog.end(r) - startdata |
|
261 | span = revlog.end(r) - startdata | |
262 | snapshot = revlog.issnapshot(r) |
|
262 | snapshot = revlog.issnapshot(r) | |
263 | if span <= targetsize and snapshot: |
|
263 | if span <= targetsize and snapshot: | |
264 | endrevidx = idx + 1 |
|
264 | endrevidx = idx + 1 | |
265 | else: |
|
265 | else: | |
266 | chunk = _trimchunk(revlog, revs, startrevidx, endrevidx) |
|
266 | chunk = _trimchunk(revlog, revs, startrevidx, endrevidx) | |
267 | if chunk: |
|
267 | if chunk: | |
268 | yield chunk |
|
268 | yield chunk | |
269 | startrevidx = idx |
|
269 | startrevidx = idx | |
270 | startdata = revlog.start(r) |
|
270 | startdata = revlog.start(r) | |
271 | endrevidx = idx + 1 |
|
271 | endrevidx = idx + 1 | |
272 | if not snapshot: |
|
272 | if not snapshot: | |
273 | break |
|
273 | break | |
274 |
|
274 | |||
275 | # for the others, we use binary slicing to quickly converge toward valid |
|
275 | # for the others, we use binary slicing to quickly converge toward valid | |
276 | # chunks (otherwise, we might end up looking for start/end of many |
|
276 | # chunks (otherwise, we might end up looking for start/end of many | |
277 | # revisions). This logic is not looking for the perfect slicing point, it |
|
277 | # revisions). This logic is not looking for the perfect slicing point, it | |
278 | # focuses on quickly converging toward valid chunks. |
|
278 | # focuses on quickly converging toward valid chunks. | |
279 | nbitem = len(revs) |
|
279 | nbitem = len(revs) | |
280 | while (enddata - startdata) > targetsize: |
|
280 | while (enddata - startdata) > targetsize: | |
281 | endrevidx = nbitem |
|
281 | endrevidx = nbitem | |
282 | if nbitem - startrevidx <= 1: |
|
282 | if nbitem - startrevidx <= 1: | |
283 | break # protect against individual chunk larger than limit |
|
283 | break # protect against individual chunk larger than limit | |
284 | localenddata = revlog.end(revs[endrevidx - 1]) |
|
284 | localenddata = revlog.end(revs[endrevidx - 1]) | |
285 | span = localenddata - startdata |
|
285 | span = localenddata - startdata | |
286 | while span > targetsize: |
|
286 | while span > targetsize: | |
287 | if endrevidx - startrevidx <= 1: |
|
287 | if endrevidx - startrevidx <= 1: | |
288 | break # protect against individual chunk larger than limit |
|
288 | break # protect against individual chunk larger than limit | |
289 | endrevidx -= (endrevidx - startrevidx) // 2 |
|
289 | endrevidx -= (endrevidx - startrevidx) // 2 | |
290 | localenddata = revlog.end(revs[endrevidx - 1]) |
|
290 | localenddata = revlog.end(revs[endrevidx - 1]) | |
291 | span = localenddata - startdata |
|
291 | span = localenddata - startdata | |
292 | chunk = _trimchunk(revlog, revs, startrevidx, endrevidx) |
|
292 | chunk = _trimchunk(revlog, revs, startrevidx, endrevidx) | |
293 | if chunk: |
|
293 | if chunk: | |
294 | yield chunk |
|
294 | yield chunk | |
295 | startrevidx = endrevidx |
|
295 | startrevidx = endrevidx | |
296 | startdata = revlog.start(revs[startrevidx]) |
|
296 | startdata = revlog.start(revs[startrevidx]) | |
297 |
|
297 | |||
298 | chunk = _trimchunk(revlog, revs, startrevidx) |
|
298 | chunk = _trimchunk(revlog, revs, startrevidx) | |
299 | if chunk: |
|
299 | if chunk: | |
300 | yield chunk |
|
300 | yield chunk | |
301 |
|
301 | |||
302 |
|
302 | |||
303 | def _slicechunktodensity(revlog, revs, targetdensity=0.5, mingapsize=0): |
|
303 | def _slicechunktodensity(revlog, revs, targetdensity=0.5, mingapsize=0): | |
304 | """slice revs to reduce the amount of unrelated data to be read from disk. |
|
304 | """slice revs to reduce the amount of unrelated data to be read from disk. | |
305 |
|
305 | |||
306 | ``revs`` is sliced into groups that should be read in one time. |
|
306 | ``revs`` is sliced into groups that should be read in one time. | |
307 | Assume that revs are sorted. |
|
307 | Assume that revs are sorted. | |
308 |
|
308 | |||
309 | The initial chunk is sliced until the overall density (payload/chunks-span |
|
309 | The initial chunk is sliced until the overall density (payload/chunks-span | |
310 | ratio) is above `targetdensity`. No gap smaller than `mingapsize` is |
|
310 | ratio) is above `targetdensity`. No gap smaller than `mingapsize` is | |
311 | skipped. |
|
311 | skipped. | |
312 |
|
312 | |||
313 | >>> revlog = _testrevlog([ |
|
313 | >>> revlog = _testrevlog([ | |
314 | ... 5, #00 (5) |
|
314 | ... 5, #00 (5) | |
315 | ... 10, #01 (5) |
|
315 | ... 10, #01 (5) | |
316 | ... 12, #02 (2) |
|
316 | ... 12, #02 (2) | |
317 | ... 12, #03 (empty) |
|
317 | ... 12, #03 (empty) | |
318 | ... 27, #04 (15) |
|
318 | ... 27, #04 (15) | |
319 | ... 31, #05 (4) |
|
319 | ... 31, #05 (4) | |
320 | ... 31, #06 (empty) |
|
320 | ... 31, #06 (empty) | |
321 | ... 42, #07 (11) |
|
321 | ... 42, #07 (11) | |
322 | ... 47, #08 (5) |
|
322 | ... 47, #08 (5) | |
323 | ... 47, #09 (empty) |
|
323 | ... 47, #09 (empty) | |
324 | ... 48, #10 (1) |
|
324 | ... 48, #10 (1) | |
325 | ... 51, #11 (3) |
|
325 | ... 51, #11 (3) | |
326 | ... 74, #12 (23) |
|
326 | ... 74, #12 (23) | |
327 | ... 85, #13 (11) |
|
327 | ... 85, #13 (11) | |
328 | ... 86, #14 (1) |
|
328 | ... 86, #14 (1) | |
329 | ... 91, #15 (5) |
|
329 | ... 91, #15 (5) | |
330 | ... ]) |
|
330 | ... ]) | |
331 |
|
331 | |||
332 | >>> list(_slicechunktodensity(revlog, list(range(16)))) |
|
332 | >>> list(_slicechunktodensity(revlog, list(range(16)))) | |
333 | [[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]] |
|
333 | [[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]] | |
334 | >>> list(_slicechunktodensity(revlog, [0, 15])) |
|
334 | >>> list(_slicechunktodensity(revlog, [0, 15])) | |
335 | [[0], [15]] |
|
335 | [[0], [15]] | |
336 | >>> list(_slicechunktodensity(revlog, [0, 11, 15])) |
|
336 | >>> list(_slicechunktodensity(revlog, [0, 11, 15])) | |
337 | [[0], [11], [15]] |
|
337 | [[0], [11], [15]] | |
338 | >>> list(_slicechunktodensity(revlog, [0, 11, 13, 15])) |
|
338 | >>> list(_slicechunktodensity(revlog, [0, 11, 13, 15])) | |
339 | [[0], [11, 13, 15]] |
|
339 | [[0], [11, 13, 15]] | |
340 | >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14])) |
|
340 | >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14])) | |
341 | [[1, 2], [5, 8, 10, 11], [14]] |
|
341 | [[1, 2], [5, 8, 10, 11], [14]] | |
342 | >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14], |
|
342 | >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14], | |
343 | ... mingapsize=20)) |
|
343 | ... mingapsize=20)) | |
344 | [[1, 2, 3, 5, 8, 10, 11], [14]] |
|
344 | [[1, 2, 3, 5, 8, 10, 11], [14]] | |
345 | >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14], |
|
345 | >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14], | |
346 | ... targetdensity=0.95)) |
|
346 | ... targetdensity=0.95)) | |
347 | [[1, 2], [5], [8, 10, 11], [14]] |
|
347 | [[1, 2], [5], [8, 10, 11], [14]] | |
348 | >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14], |
|
348 | >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14], | |
349 | ... targetdensity=0.95, mingapsize=12)) |
|
349 | ... targetdensity=0.95, mingapsize=12)) | |
350 | [[1, 2], [5, 8, 10, 11], [14]] |
|
350 | [[1, 2], [5, 8, 10, 11], [14]] | |
351 | """ |
|
351 | """ | |
352 | start = revlog.start |
|
352 | start = revlog.start | |
353 | length = revlog.length |
|
353 | length = revlog.length | |
354 |
|
354 | |||
355 | if len(revs) <= 1: |
|
355 | if len(revs) <= 1: | |
356 | yield revs |
|
356 | yield revs | |
357 | return |
|
357 | return | |
358 |
|
358 | |||
359 | deltachainspan = segmentspan(revlog, revs) |
|
359 | deltachainspan = segmentspan(revlog, revs) | |
360 |
|
360 | |||
361 | if deltachainspan < mingapsize: |
|
361 | if deltachainspan < mingapsize: | |
362 | yield revs |
|
362 | yield revs | |
363 | return |
|
363 | return | |
364 |
|
364 | |||
365 | readdata = deltachainspan |
|
365 | readdata = deltachainspan | |
366 | chainpayload = sum(length(r) for r in revs) |
|
366 | chainpayload = sum(length(r) for r in revs) | |
367 |
|
367 | |||
368 | if deltachainspan: |
|
368 | if deltachainspan: | |
369 | density = chainpayload / float(deltachainspan) |
|
369 | density = chainpayload / float(deltachainspan) | |
370 | else: |
|
370 | else: | |
371 | density = 1.0 |
|
371 | density = 1.0 | |
372 |
|
372 | |||
373 | if density >= targetdensity: |
|
373 | if density >= targetdensity: | |
374 | yield revs |
|
374 | yield revs | |
375 | return |
|
375 | return | |
376 |
|
376 | |||
377 | # Store the gaps in a heap to have them sorted by decreasing size |
|
377 | # Store the gaps in a heap to have them sorted by decreasing size | |
378 | gaps = [] |
|
378 | gaps = [] | |
379 | prevend = None |
|
379 | prevend = None | |
380 | for i, rev in enumerate(revs): |
|
380 | for i, rev in enumerate(revs): | |
381 | revstart = start(rev) |
|
381 | revstart = start(rev) | |
382 | revlen = length(rev) |
|
382 | revlen = length(rev) | |
383 |
|
383 | |||
384 | # Skip empty revisions to form larger holes |
|
384 | # Skip empty revisions to form larger holes | |
385 | if revlen == 0: |
|
385 | if revlen == 0: | |
386 | continue |
|
386 | continue | |
387 |
|
387 | |||
388 | if prevend is not None: |
|
388 | if prevend is not None: | |
389 | gapsize = revstart - prevend |
|
389 | gapsize = revstart - prevend | |
390 | # only consider holes that are large enough |
|
390 | # only consider holes that are large enough | |
391 | if gapsize > mingapsize: |
|
391 | if gapsize > mingapsize: | |
392 | gaps.append((gapsize, i)) |
|
392 | gaps.append((gapsize, i)) | |
393 |
|
393 | |||
394 | prevend = revstart + revlen |
|
394 | prevend = revstart + revlen | |
395 | # sort the gaps to pop them from largest to small |
|
395 | # sort the gaps to pop them from largest to small | |
396 | gaps.sort() |
|
396 | gaps.sort() | |
397 |
|
397 | |||
398 | # Collect the indices of the largest holes until the density is acceptable |
|
398 | # Collect the indices of the largest holes until the density is acceptable | |
399 | selected = [] |
|
399 | selected = [] | |
400 | while gaps and density < targetdensity: |
|
400 | while gaps and density < targetdensity: | |
401 | gapsize, gapidx = gaps.pop() |
|
401 | gapsize, gapidx = gaps.pop() | |
402 |
|
402 | |||
403 | selected.append(gapidx) |
|
403 | selected.append(gapidx) | |
404 |
|
404 | |||
405 | # the gap sizes are stored as negatives to be sorted decreasingly |
|
405 | # the gap sizes are stored as negatives to be sorted decreasingly | |
406 | # by the heap |
|
406 | # by the heap | |
407 | readdata -= gapsize |
|
407 | readdata -= gapsize | |
408 | if readdata > 0: |
|
408 | if readdata > 0: | |
409 | density = chainpayload / float(readdata) |
|
409 | density = chainpayload / float(readdata) | |
410 | else: |
|
410 | else: | |
411 | density = 1.0 |
|
411 | density = 1.0 | |
412 | selected.sort() |
|
412 | selected.sort() | |
413 |
|
413 | |||
414 | # Cut the revs at collected indices |
|
414 | # Cut the revs at collected indices | |
415 | previdx = 0 |
|
415 | previdx = 0 | |
416 | for idx in selected: |
|
416 | for idx in selected: | |
417 |
|
417 | |||
418 | chunk = _trimchunk(revlog, revs, previdx, idx) |
|
418 | chunk = _trimchunk(revlog, revs, previdx, idx) | |
419 | if chunk: |
|
419 | if chunk: | |
420 | yield chunk |
|
420 | yield chunk | |
421 |
|
421 | |||
422 | previdx = idx |
|
422 | previdx = idx | |
423 |
|
423 | |||
424 | chunk = _trimchunk(revlog, revs, previdx) |
|
424 | chunk = _trimchunk(revlog, revs, previdx) | |
425 | if chunk: |
|
425 | if chunk: | |
426 | yield chunk |
|
426 | yield chunk | |
427 |
|
427 | |||
428 |
|
428 | |||
429 | def _trimchunk(revlog, revs, startidx, endidx=None): |
|
429 | def _trimchunk(revlog, revs, startidx, endidx=None): | |
430 | """returns revs[startidx:endidx] without empty trailing revs |
|
430 | """returns revs[startidx:endidx] without empty trailing revs | |
431 |
|
431 | |||
432 | Doctest Setup |
|
432 | Doctest Setup | |
433 | >>> revlog = _testrevlog([ |
|
433 | >>> revlog = _testrevlog([ | |
434 | ... 5, #0 |
|
434 | ... 5, #0 | |
435 | ... 10, #1 |
|
435 | ... 10, #1 | |
436 | ... 12, #2 |
|
436 | ... 12, #2 | |
437 | ... 12, #3 (empty) |
|
437 | ... 12, #3 (empty) | |
438 | ... 17, #4 |
|
438 | ... 17, #4 | |
439 | ... 21, #5 |
|
439 | ... 21, #5 | |
440 | ... 21, #6 (empty) |
|
440 | ... 21, #6 (empty) | |
441 | ... ]) |
|
441 | ... ]) | |
442 |
|
442 | |||
443 | Contiguous cases: |
|
443 | Contiguous cases: | |
444 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 0) |
|
444 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 0) | |
445 | [0, 1, 2, 3, 4, 5] |
|
445 | [0, 1, 2, 3, 4, 5] | |
446 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 0, 5) |
|
446 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 0, 5) | |
447 | [0, 1, 2, 3, 4] |
|
447 | [0, 1, 2, 3, 4] | |
448 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 0, 4) |
|
448 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 0, 4) | |
449 | [0, 1, 2] |
|
449 | [0, 1, 2] | |
450 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 2, 4) |
|
450 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 2, 4) | |
451 | [2] |
|
451 | [2] | |
452 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 3) |
|
452 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 3) | |
453 | [3, 4, 5] |
|
453 | [3, 4, 5] | |
454 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 3, 5) |
|
454 | >>> _trimchunk(revlog, [0, 1, 2, 3, 4, 5, 6], 3, 5) | |
455 | [3, 4] |
|
455 | [3, 4] | |
456 |
|
456 | |||
457 | Discontiguous cases: |
|
457 | Discontiguous cases: | |
458 | >>> _trimchunk(revlog, [1, 3, 5, 6], 0) |
|
458 | >>> _trimchunk(revlog, [1, 3, 5, 6], 0) | |
459 | [1, 3, 5] |
|
459 | [1, 3, 5] | |
460 | >>> _trimchunk(revlog, [1, 3, 5, 6], 0, 2) |
|
460 | >>> _trimchunk(revlog, [1, 3, 5, 6], 0, 2) | |
461 | [1] |
|
461 | [1] | |
462 | >>> _trimchunk(revlog, [1, 3, 5, 6], 1, 3) |
|
462 | >>> _trimchunk(revlog, [1, 3, 5, 6], 1, 3) | |
463 | [3, 5] |
|
463 | [3, 5] | |
464 | >>> _trimchunk(revlog, [1, 3, 5, 6], 1) |
|
464 | >>> _trimchunk(revlog, [1, 3, 5, 6], 1) | |
465 | [3, 5] |
|
465 | [3, 5] | |
466 | """ |
|
466 | """ | |
467 | length = revlog.length |
|
467 | length = revlog.length | |
468 |
|
468 | |||
469 | if endidx is None: |
|
469 | if endidx is None: | |
470 | endidx = len(revs) |
|
470 | endidx = len(revs) | |
471 |
|
471 | |||
472 | # If we have a non-emtpy delta candidate, there are nothing to trim |
|
472 | # If we have a non-emtpy delta candidate, there are nothing to trim | |
473 | if revs[endidx - 1] < len(revlog): |
|
473 | if revs[endidx - 1] < len(revlog): | |
474 | # Trim empty revs at the end, except the very first revision of a chain |
|
474 | # Trim empty revs at the end, except the very first revision of a chain | |
475 | while ( |
|
475 | while ( | |
476 | endidx > 1 and endidx > startidx and length(revs[endidx - 1]) == 0 |
|
476 | endidx > 1 and endidx > startidx and length(revs[endidx - 1]) == 0 | |
477 | ): |
|
477 | ): | |
478 | endidx -= 1 |
|
478 | endidx -= 1 | |
479 |
|
479 | |||
480 | return revs[startidx:endidx] |
|
480 | return revs[startidx:endidx] | |
481 |
|
481 | |||
482 |
|
482 | |||
483 | def segmentspan(revlog, revs): |
|
483 | def segmentspan(revlog, revs): | |
484 | """Get the byte span of a segment of revisions |
|
484 | """Get the byte span of a segment of revisions | |
485 |
|
485 | |||
486 | revs is a sorted array of revision numbers |
|
486 | revs is a sorted array of revision numbers | |
487 |
|
487 | |||
488 | >>> revlog = _testrevlog([ |
|
488 | >>> revlog = _testrevlog([ | |
489 | ... 5, #0 |
|
489 | ... 5, #0 | |
490 | ... 10, #1 |
|
490 | ... 10, #1 | |
491 | ... 12, #2 |
|
491 | ... 12, #2 | |
492 | ... 12, #3 (empty) |
|
492 | ... 12, #3 (empty) | |
493 | ... 17, #4 |
|
493 | ... 17, #4 | |
494 | ... ]) |
|
494 | ... ]) | |
495 |
|
495 | |||
496 | >>> segmentspan(revlog, [0, 1, 2, 3, 4]) |
|
496 | >>> segmentspan(revlog, [0, 1, 2, 3, 4]) | |
497 | 17 |
|
497 | 17 | |
498 | >>> segmentspan(revlog, [0, 4]) |
|
498 | >>> segmentspan(revlog, [0, 4]) | |
499 | 17 |
|
499 | 17 | |
500 | >>> segmentspan(revlog, [3, 4]) |
|
500 | >>> segmentspan(revlog, [3, 4]) | |
501 | 5 |
|
501 | 5 | |
502 | >>> segmentspan(revlog, [1, 2, 3,]) |
|
502 | >>> segmentspan(revlog, [1, 2, 3,]) | |
503 | 7 |
|
503 | 7 | |
504 | >>> segmentspan(revlog, [1, 3]) |
|
504 | >>> segmentspan(revlog, [1, 3]) | |
505 | 7 |
|
505 | 7 | |
506 | """ |
|
506 | """ | |
507 | if not revs: |
|
507 | if not revs: | |
508 | return 0 |
|
508 | return 0 | |
509 | end = revlog.end(revs[-1]) |
|
509 | end = revlog.end(revs[-1]) | |
510 | return end - revlog.start(revs[0]) |
|
510 | return end - revlog.start(revs[0]) | |
511 |
|
511 | |||
512 |
|
512 | |||
513 | def _textfromdelta(fh, revlog, baserev, delta, p1, p2, flags, expectednode): |
|
513 | def _textfromdelta(fh, revlog, baserev, delta, p1, p2, flags, expectednode): | |
514 | """build full text from a (base, delta) pair and other metadata""" |
|
514 | """build full text from a (base, delta) pair and other metadata""" | |
515 | # special case deltas which replace entire base; no need to decode |
|
515 | # special case deltas which replace entire base; no need to decode | |
516 | # base revision. this neatly avoids censored bases, which throw when |
|
516 | # base revision. this neatly avoids censored bases, which throw when | |
517 | # they're decoded. |
|
517 | # they're decoded. | |
518 | hlen = struct.calcsize(b">lll") |
|
518 | hlen = struct.calcsize(b">lll") | |
519 | if delta[:hlen] == mdiff.replacediffheader( |
|
519 | if delta[:hlen] == mdiff.replacediffheader( | |
520 | revlog.rawsize(baserev), len(delta) - hlen |
|
520 | revlog.rawsize(baserev), len(delta) - hlen | |
521 | ): |
|
521 | ): | |
522 | fulltext = delta[hlen:] |
|
522 | fulltext = delta[hlen:] | |
523 | else: |
|
523 | else: | |
524 | # deltabase is rawtext before changed by flag processors, which is |
|
524 | # deltabase is rawtext before changed by flag processors, which is | |
525 | # equivalent to non-raw text |
|
525 | # equivalent to non-raw text | |
526 | basetext = revlog.revision(baserev, _df=fh, raw=False) |
|
526 | basetext = revlog.revision(baserev, _df=fh, raw=False) | |
527 | fulltext = mdiff.patch(basetext, delta) |
|
527 | fulltext = mdiff.patch(basetext, delta) | |
528 |
|
528 | |||
529 | try: |
|
529 | try: | |
530 | validatehash = flagutil.processflagsraw(revlog, fulltext, flags) |
|
530 | validatehash = flagutil.processflagsraw(revlog, fulltext, flags) | |
531 | if validatehash: |
|
531 | if validatehash: | |
532 | revlog.checkhash(fulltext, expectednode, p1=p1, p2=p2) |
|
532 | revlog.checkhash(fulltext, expectednode, p1=p1, p2=p2) | |
533 | if flags & REVIDX_ISCENSORED: |
|
533 | if flags & REVIDX_ISCENSORED: | |
534 | raise error.StorageError( |
|
534 | raise error.StorageError( | |
535 | _(b'node %s is not censored') % expectednode |
|
535 | _(b'node %s is not censored') % expectednode | |
536 | ) |
|
536 | ) | |
537 | except error.CensoredNodeError: |
|
537 | except error.CensoredNodeError: | |
538 | # must pass the censored index flag to add censored revisions |
|
538 | # must pass the censored index flag to add censored revisions | |
539 | if not flags & REVIDX_ISCENSORED: |
|
539 | if not flags & REVIDX_ISCENSORED: | |
540 | raise |
|
540 | raise | |
541 | return fulltext |
|
541 | return fulltext | |
542 |
|
542 | |||
543 |
|
543 | |||
544 | @attr.s(slots=True, frozen=True) |
|
544 | @attr.s(slots=True, frozen=True) | |
545 | class _deltainfo(object): |
|
545 | class _deltainfo(object): | |
546 | distance = attr.ib() |
|
546 | distance = attr.ib() | |
547 | deltalen = attr.ib() |
|
547 | deltalen = attr.ib() | |
548 | data = attr.ib() |
|
548 | data = attr.ib() | |
549 | base = attr.ib() |
|
549 | base = attr.ib() | |
550 | chainbase = attr.ib() |
|
550 | chainbase = attr.ib() | |
551 | chainlen = attr.ib() |
|
551 | chainlen = attr.ib() | |
552 | compresseddeltalen = attr.ib() |
|
552 | compresseddeltalen = attr.ib() | |
553 | snapshotdepth = attr.ib() |
|
553 | snapshotdepth = attr.ib() | |
554 |
|
554 | |||
555 |
|
555 | |||
556 | def drop_u_compression(delta): |
|
556 | def drop_u_compression(delta): | |
557 | """turn into a "u" (no-compression) into no-compression without header |
|
557 | """turn into a "u" (no-compression) into no-compression without header | |
558 |
|
558 | |||
559 | This is useful for revlog format that has better compression method. |
|
559 | This is useful for revlog format that has better compression method. | |
560 | """ |
|
560 | """ | |
561 | assert delta.data[0] == b'u', delta.data[0] |
|
561 | assert delta.data[0] == b'u', delta.data[0] | |
562 | return _deltainfo( |
|
562 | return _deltainfo( | |
563 | delta.distance, |
|
563 | delta.distance, | |
564 | delta.deltalen - 1, |
|
564 | delta.deltalen - 1, | |
565 | (b'', delta.data[1]), |
|
565 | (b'', delta.data[1]), | |
566 | delta.base, |
|
566 | delta.base, | |
567 | delta.chainbase, |
|
567 | delta.chainbase, | |
568 | delta.chainlen, |
|
568 | delta.chainlen, | |
569 | delta.compresseddeltalen, |
|
569 | delta.compresseddeltalen, | |
570 | delta.snapshotdepth, |
|
570 | delta.snapshotdepth, | |
571 | ) |
|
571 | ) | |
572 |
|
572 | |||
573 |
|
573 | |||
574 | def isgooddeltainfo(revlog, deltainfo, revinfo): |
|
574 | def isgooddeltainfo(revlog, deltainfo, revinfo): | |
575 | """Returns True if the given delta is good. Good means that it is within |
|
575 | """Returns True if the given delta is good. Good means that it is within | |
576 | the disk span, disk size, and chain length bounds that we know to be |
|
576 | the disk span, disk size, and chain length bounds that we know to be | |
577 | performant.""" |
|
577 | performant.""" | |
578 | if deltainfo is None: |
|
578 | if deltainfo is None: | |
579 | return False |
|
579 | return False | |
580 |
|
580 | |||
581 | # - 'deltainfo.distance' is the distance from the base revision -- |
|
581 | # - 'deltainfo.distance' is the distance from the base revision -- | |
582 | # bounding it limits the amount of I/O we need to do. |
|
582 | # bounding it limits the amount of I/O we need to do. | |
583 | # - 'deltainfo.compresseddeltalen' is the sum of the total size of |
|
583 | # - 'deltainfo.compresseddeltalen' is the sum of the total size of | |
584 | # deltas we need to apply -- bounding it limits the amount of CPU |
|
584 | # deltas we need to apply -- bounding it limits the amount of CPU | |
585 | # we consume. |
|
585 | # we consume. | |
586 |
|
586 | |||
587 | textlen = revinfo.textlen |
|
587 | textlen = revinfo.textlen | |
588 | defaultmax = textlen * 4 |
|
588 | defaultmax = textlen * 4 | |
589 | maxdist = revlog._maxdeltachainspan |
|
589 | maxdist = revlog._maxdeltachainspan | |
590 | if not maxdist: |
|
590 | if not maxdist: | |
591 | maxdist = deltainfo.distance # ensure the conditional pass |
|
591 | maxdist = deltainfo.distance # ensure the conditional pass | |
592 | maxdist = max(maxdist, defaultmax) |
|
592 | maxdist = max(maxdist, defaultmax) | |
593 |
|
593 | |||
594 | # Bad delta from read span: |
|
594 | # Bad delta from read span: | |
595 | # |
|
595 | # | |
596 | # If the span of data read is larger than the maximum allowed. |
|
596 | # If the span of data read is larger than the maximum allowed. | |
597 | # |
|
597 | # | |
598 | # In the sparse-revlog case, we rely on the associated "sparse reading" |
|
598 | # In the sparse-revlog case, we rely on the associated "sparse reading" | |
599 | # to avoid issue related to the span of data. In theory, it would be |
|
599 | # to avoid issue related to the span of data. In theory, it would be | |
600 | # possible to build pathological revlog where delta pattern would lead |
|
600 | # possible to build pathological revlog where delta pattern would lead | |
601 | # to too many reads. However, they do not happen in practice at all. So |
|
601 | # to too many reads. However, they do not happen in practice at all. So | |
602 | # we skip the span check entirely. |
|
602 | # we skip the span check entirely. | |
603 | if not revlog._sparserevlog and maxdist < deltainfo.distance: |
|
603 | if not revlog._sparserevlog and maxdist < deltainfo.distance: | |
604 | return False |
|
604 | return False | |
605 |
|
605 | |||
606 | # Bad delta from new delta size: |
|
606 | # Bad delta from new delta size: | |
607 | # |
|
607 | # | |
608 | # If the delta size is larger than the target text, storing the |
|
608 | # If the delta size is larger than the target text, storing the | |
609 | # delta will be inefficient. |
|
609 | # delta will be inefficient. | |
610 | if textlen < deltainfo.deltalen: |
|
610 | if textlen < deltainfo.deltalen: | |
611 | return False |
|
611 | return False | |
612 |
|
612 | |||
613 | # Bad delta from cumulated payload size: |
|
613 | # Bad delta from cumulated payload size: | |
614 | # |
|
614 | # | |
615 | # If the sum of delta get larger than K * target text length. |
|
615 | # If the sum of delta get larger than K * target text length. | |
616 | if textlen * LIMIT_DELTA2TEXT < deltainfo.compresseddeltalen: |
|
616 | if textlen * LIMIT_DELTA2TEXT < deltainfo.compresseddeltalen: | |
617 | return False |
|
617 | return False | |
618 |
|
618 | |||
619 | # Bad delta from chain length: |
|
619 | # Bad delta from chain length: | |
620 | # |
|
620 | # | |
621 | # If the number of delta in the chain gets too high. |
|
621 | # If the number of delta in the chain gets too high. | |
622 | if revlog._maxchainlen and revlog._maxchainlen < deltainfo.chainlen: |
|
622 | if revlog._maxchainlen and revlog._maxchainlen < deltainfo.chainlen: | |
623 | return False |
|
623 | return False | |
624 |
|
624 | |||
625 | # bad delta from intermediate snapshot size limit |
|
625 | # bad delta from intermediate snapshot size limit | |
626 | # |
|
626 | # | |
627 | # If an intermediate snapshot size is higher than the limit. The |
|
627 | # If an intermediate snapshot size is higher than the limit. The | |
628 | # limit exist to prevent endless chain of intermediate delta to be |
|
628 | # limit exist to prevent endless chain of intermediate delta to be | |
629 | # created. |
|
629 | # created. | |
630 | if ( |
|
630 | if ( | |
631 | deltainfo.snapshotdepth is not None |
|
631 | deltainfo.snapshotdepth is not None | |
632 | and (textlen >> deltainfo.snapshotdepth) < deltainfo.deltalen |
|
632 | and (textlen >> deltainfo.snapshotdepth) < deltainfo.deltalen | |
633 | ): |
|
633 | ): | |
634 | return False |
|
634 | return False | |
635 |
|
635 | |||
636 | # bad delta if new intermediate snapshot is larger than the previous |
|
636 | # bad delta if new intermediate snapshot is larger than the previous | |
637 | # snapshot |
|
637 | # snapshot | |
638 | if ( |
|
638 | if ( | |
639 | deltainfo.snapshotdepth |
|
639 | deltainfo.snapshotdepth | |
640 | and revlog.length(deltainfo.base) < deltainfo.deltalen |
|
640 | and revlog.length(deltainfo.base) < deltainfo.deltalen | |
641 | ): |
|
641 | ): | |
642 | return False |
|
642 | return False | |
643 |
|
643 | |||
644 | return True |
|
644 | return True | |
645 |
|
645 | |||
646 |
|
646 | |||
647 | # If a revision's full text is that much bigger than a base candidate full |
|
647 | # If a revision's full text is that much bigger than a base candidate full | |
648 | # text's, it is very unlikely that it will produce a valid delta. We no longer |
|
648 | # text's, it is very unlikely that it will produce a valid delta. We no longer | |
649 | # consider these candidates. |
|
649 | # consider these candidates. | |
650 | LIMIT_BASE2TEXT = 500 |
|
650 | LIMIT_BASE2TEXT = 500 | |
651 |
|
651 | |||
652 |
|
652 | |||
653 | def _candidategroups(revlog, textlen, p1, p2, cachedelta): |
|
653 | def _candidategroups(revlog, textlen, p1, p2, cachedelta): | |
654 | """Provides group of revision to be tested as delta base |
|
654 | """Provides group of revision to be tested as delta base | |
655 |
|
655 | |||
656 | This top level function focus on emitting groups with unique and worthwhile |
|
656 | This top level function focus on emitting groups with unique and worthwhile | |
657 | content. See _raw_candidate_groups for details about the group order. |
|
657 | content. See _raw_candidate_groups for details about the group order. | |
658 | """ |
|
658 | """ | |
659 | # should we try to build a delta? |
|
659 | # should we try to build a delta? | |
660 | if not (len(revlog) and revlog._storedeltachains): |
|
660 | if not (len(revlog) and revlog._storedeltachains): | |
661 | yield None |
|
661 | yield None | |
662 | return |
|
662 | return | |
663 |
|
663 | |||
664 | deltalength = revlog.length |
|
664 | deltalength = revlog.length | |
665 | deltaparent = revlog.deltaparent |
|
665 | deltaparent = revlog.deltaparent | |
666 | sparse = revlog._sparserevlog |
|
666 | sparse = revlog._sparserevlog | |
667 | good = None |
|
667 | good = None | |
668 |
|
668 | |||
669 | deltas_limit = textlen * LIMIT_DELTA2TEXT |
|
669 | deltas_limit = textlen * LIMIT_DELTA2TEXT | |
670 |
|
670 | |||
671 | tested = {nullrev} |
|
671 | tested = {nullrev} | |
672 | candidates = _refinedgroups(revlog, p1, p2, cachedelta) |
|
672 | candidates = _refinedgroups(revlog, p1, p2, cachedelta) | |
673 | while True: |
|
673 | while True: | |
674 | temptative = candidates.send(good) |
|
674 | temptative = candidates.send(good) | |
675 | if temptative is None: |
|
675 | if temptative is None: | |
676 | break |
|
676 | break | |
677 | group = [] |
|
677 | group = [] | |
678 | for rev in temptative: |
|
678 | for rev in temptative: | |
679 | # skip over empty delta (no need to include them in a chain) |
|
679 | # skip over empty delta (no need to include them in a chain) | |
680 | while revlog._generaldelta and not ( |
|
680 | while revlog._generaldelta and not ( | |
681 | rev == nullrev or rev in tested or deltalength(rev) |
|
681 | rev == nullrev or rev in tested or deltalength(rev) | |
682 | ): |
|
682 | ): | |
683 | tested.add(rev) |
|
683 | tested.add(rev) | |
684 | rev = deltaparent(rev) |
|
684 | rev = deltaparent(rev) | |
685 | # no need to try a delta against nullrev, this will be done as a |
|
685 | # no need to try a delta against nullrev, this will be done as a | |
686 | # last resort. |
|
686 | # last resort. | |
687 | if rev == nullrev: |
|
687 | if rev == nullrev: | |
688 | continue |
|
688 | continue | |
689 | # filter out revision we tested already |
|
689 | # filter out revision we tested already | |
690 | if rev in tested: |
|
690 | if rev in tested: | |
691 | continue |
|
691 | continue | |
692 | tested.add(rev) |
|
692 | tested.add(rev) | |
693 | # filter out delta base that will never produce good delta |
|
693 | # filter out delta base that will never produce good delta | |
694 | if deltas_limit < revlog.length(rev): |
|
694 | if deltas_limit < revlog.length(rev): | |
695 | continue |
|
695 | continue | |
696 | if sparse and revlog.rawsize(rev) < (textlen // LIMIT_BASE2TEXT): |
|
696 | if sparse and revlog.rawsize(rev) < (textlen // LIMIT_BASE2TEXT): | |
697 | continue |
|
697 | continue | |
698 | # no delta for rawtext-changing revs (see "candelta" for why) |
|
698 | # no delta for rawtext-changing revs (see "candelta" for why) | |
699 | if revlog.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS: |
|
699 | if revlog.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS: | |
700 | continue |
|
700 | continue | |
701 | # If we reach here, we are about to build and test a delta. |
|
701 | # If we reach here, we are about to build and test a delta. | |
702 | # The delta building process will compute the chaininfo in all |
|
702 | # The delta building process will compute the chaininfo in all | |
703 | # case, since that computation is cached, it is fine to access it |
|
703 | # case, since that computation is cached, it is fine to access it | |
704 | # here too. |
|
704 | # here too. | |
705 | chainlen, chainsize = revlog._chaininfo(rev) |
|
705 | chainlen, chainsize = revlog._chaininfo(rev) | |
706 | # if chain will be too long, skip base |
|
706 | # if chain will be too long, skip base | |
707 | if revlog._maxchainlen and chainlen >= revlog._maxchainlen: |
|
707 | if revlog._maxchainlen and chainlen >= revlog._maxchainlen: | |
708 | continue |
|
708 | continue | |
709 | # if chain already have too much data, skip base |
|
709 | # if chain already have too much data, skip base | |
710 | if deltas_limit < chainsize: |
|
710 | if deltas_limit < chainsize: | |
711 | continue |
|
711 | continue | |
712 | if sparse and revlog.upperboundcomp is not None: |
|
712 | if sparse and revlog.upperboundcomp is not None: | |
713 | maxcomp = revlog.upperboundcomp |
|
713 | maxcomp = revlog.upperboundcomp | |
714 | basenotsnap = (p1, p2, nullrev) |
|
714 | basenotsnap = (p1, p2, nullrev) | |
715 | if rev not in basenotsnap and revlog.issnapshot(rev): |
|
715 | if rev not in basenotsnap and revlog.issnapshot(rev): | |
716 | snapshotdepth = revlog.snapshotdepth(rev) |
|
716 | snapshotdepth = revlog.snapshotdepth(rev) | |
717 | # If text is significantly larger than the base, we can |
|
717 | # If text is significantly larger than the base, we can | |
718 | # expect the resulting delta to be proportional to the size |
|
718 | # expect the resulting delta to be proportional to the size | |
719 | # difference |
|
719 | # difference | |
720 | revsize = revlog.rawsize(rev) |
|
720 | revsize = revlog.rawsize(rev) | |
721 | rawsizedistance = max(textlen - revsize, 0) |
|
721 | rawsizedistance = max(textlen - revsize, 0) | |
722 | # use an estimate of the compression upper bound. |
|
722 | # use an estimate of the compression upper bound. | |
723 | lowestrealisticdeltalen = rawsizedistance // maxcomp |
|
723 | lowestrealisticdeltalen = rawsizedistance // maxcomp | |
724 |
|
724 | |||
725 | # check the absolute constraint on the delta size |
|
725 | # check the absolute constraint on the delta size | |
726 | snapshotlimit = textlen >> snapshotdepth |
|
726 | snapshotlimit = textlen >> snapshotdepth | |
727 | if snapshotlimit < lowestrealisticdeltalen: |
|
727 | if snapshotlimit < lowestrealisticdeltalen: | |
728 | # delta lower bound is larger than accepted upper bound |
|
728 | # delta lower bound is larger than accepted upper bound | |
729 | continue |
|
729 | continue | |
730 |
|
730 | |||
731 | # check the relative constraint on the delta size |
|
731 | # check the relative constraint on the delta size | |
732 | revlength = revlog.length(rev) |
|
732 | revlength = revlog.length(rev) | |
733 | if revlength < lowestrealisticdeltalen: |
|
733 | if revlength < lowestrealisticdeltalen: | |
734 | # delta probable lower bound is larger than target base |
|
734 | # delta probable lower bound is larger than target base | |
735 | continue |
|
735 | continue | |
736 |
|
736 | |||
737 | group.append(rev) |
|
737 | group.append(rev) | |
738 | if group: |
|
738 | if group: | |
739 | # XXX: in the sparse revlog case, group can become large, |
|
739 | # XXX: in the sparse revlog case, group can become large, | |
740 | # impacting performances. Some bounding or slicing mecanism |
|
740 | # impacting performances. Some bounding or slicing mecanism | |
741 | # would help to reduce this impact. |
|
741 | # would help to reduce this impact. | |
742 | good = yield tuple(group) |
|
742 | good = yield tuple(group) | |
743 | yield None |
|
743 | yield None | |
744 |
|
744 | |||
745 |
|
745 | |||
746 | def _findsnapshots(revlog, cache, start_rev): |
|
746 | def _findsnapshots(revlog, cache, start_rev): | |
747 | """find snapshot from start_rev to tip""" |
|
747 | """find snapshot from start_rev to tip""" | |
748 | if util.safehasattr(revlog.index, b'findsnapshots'): |
|
748 | if util.safehasattr(revlog.index, b'findsnapshots'): | |
749 | revlog.index.findsnapshots(cache, start_rev) |
|
749 | revlog.index.findsnapshots(cache, start_rev) | |
750 | else: |
|
750 | else: | |
751 | deltaparent = revlog.deltaparent |
|
751 | deltaparent = revlog.deltaparent | |
752 | issnapshot = revlog.issnapshot |
|
752 | issnapshot = revlog.issnapshot | |
753 | for rev in revlog.revs(start_rev): |
|
753 | for rev in revlog.revs(start_rev): | |
754 | if issnapshot(rev): |
|
754 | if issnapshot(rev): | |
755 | cache[deltaparent(rev)].append(rev) |
|
755 | cache[deltaparent(rev)].append(rev) | |
756 |
|
756 | |||
757 |
|
757 | |||
758 | def _refinedgroups(revlog, p1, p2, cachedelta): |
|
758 | def _refinedgroups(revlog, p1, p2, cachedelta): | |
759 | good = None |
|
759 | good = None | |
760 | # First we try to reuse a the delta contained in the bundle. |
|
760 | # First we try to reuse a the delta contained in the bundle. | |
761 | # (or from the source revlog) |
|
761 | # (or from the source revlog) | |
762 | # |
|
762 | # | |
763 | # This logic only applies to general delta repositories and can be disabled |
|
763 | # This logic only applies to general delta repositories and can be disabled | |
764 | # through configuration. Disabling reuse source delta is useful when |
|
764 | # through configuration. Disabling reuse source delta is useful when | |
765 | # we want to make sure we recomputed "optimal" deltas. |
|
765 | # we want to make sure we recomputed "optimal" deltas. | |
766 | if cachedelta and revlog._generaldelta and revlog._lazydeltabase: |
|
766 | if cachedelta and revlog._generaldelta and revlog._lazydeltabase: | |
767 | # Assume what we received from the server is a good choice |
|
767 | # Assume what we received from the server is a good choice | |
768 | # build delta will reuse the cache |
|
768 | # build delta will reuse the cache | |
769 | good = yield (cachedelta[0],) |
|
769 | good = yield (cachedelta[0],) | |
770 | if good is not None: |
|
770 | if good is not None: | |
771 | yield None |
|
771 | yield None | |
772 | return |
|
772 | return | |
773 | snapshots = collections.defaultdict(list) |
|
773 | snapshots = collections.defaultdict(list) | |
774 | for candidates in _rawgroups(revlog, p1, p2, cachedelta, snapshots): |
|
774 | for candidates in _rawgroups(revlog, p1, p2, cachedelta, snapshots): | |
775 | good = yield candidates |
|
775 | good = yield candidates | |
776 | if good is not None: |
|
776 | if good is not None: | |
777 | break |
|
777 | break | |
778 |
|
778 | |||
779 | # If sparse revlog is enabled, we can try to refine the available deltas |
|
779 | # If sparse revlog is enabled, we can try to refine the available deltas | |
780 | if not revlog._sparserevlog: |
|
780 | if not revlog._sparserevlog: | |
781 | yield None |
|
781 | yield None | |
782 | return |
|
782 | return | |
783 |
|
783 | |||
784 | # if we have a refinable value, try to refine it |
|
784 | # if we have a refinable value, try to refine it | |
785 | if good is not None and good not in (p1, p2) and revlog.issnapshot(good): |
|
785 | if good is not None and good not in (p1, p2) and revlog.issnapshot(good): | |
786 | # refine snapshot down |
|
786 | # refine snapshot down | |
787 | previous = None |
|
787 | previous = None | |
788 | while previous != good: |
|
788 | while previous != good: | |
789 | previous = good |
|
789 | previous = good | |
790 | base = revlog.deltaparent(good) |
|
790 | base = revlog.deltaparent(good) | |
791 | if base == nullrev: |
|
791 | if base == nullrev: | |
792 | break |
|
792 | break | |
793 | good = yield (base,) |
|
793 | good = yield (base,) | |
794 | # refine snapshot up |
|
794 | # refine snapshot up | |
795 | if not snapshots: |
|
795 | if not snapshots: | |
796 | _findsnapshots(revlog, snapshots, good + 1) |
|
796 | _findsnapshots(revlog, snapshots, good + 1) | |
797 | previous = None |
|
797 | previous = None | |
798 | while good != previous: |
|
798 | while good != previous: | |
799 | previous = good |
|
799 | previous = good | |
800 | children = tuple(sorted(c for c in snapshots[good])) |
|
800 | children = tuple(sorted(c for c in snapshots[good])) | |
801 | good = yield children |
|
801 | good = yield children | |
802 |
|
802 | |||
803 | # we have found nothing |
|
803 | # we have found nothing | |
804 | yield None |
|
804 | yield None | |
805 |
|
805 | |||
806 |
|
806 | |||
807 | def _rawgroups(revlog, p1, p2, cachedelta, snapshots=None): |
|
807 | def _rawgroups(revlog, p1, p2, cachedelta, snapshots=None): | |
808 | """Provides group of revision to be tested as delta base |
|
808 | """Provides group of revision to be tested as delta base | |
809 |
|
809 | |||
810 | This lower level function focus on emitting delta theorically interresting |
|
810 | This lower level function focus on emitting delta theorically interresting | |
811 | without looking it any practical details. |
|
811 | without looking it any practical details. | |
812 |
|
812 | |||
813 | The group order aims at providing fast or small candidates first. |
|
813 | The group order aims at providing fast or small candidates first. | |
814 | """ |
|
814 | """ | |
815 | gdelta = revlog._generaldelta |
|
815 | gdelta = revlog._generaldelta | |
816 | # gate sparse behind general-delta because of issue6056 |
|
816 | # gate sparse behind general-delta because of issue6056 | |
817 | sparse = gdelta and revlog._sparserevlog |
|
817 | sparse = gdelta and revlog._sparserevlog | |
818 | curr = len(revlog) |
|
818 | curr = len(revlog) | |
819 | prev = curr - 1 |
|
819 | prev = curr - 1 | |
820 | deltachain = lambda rev: revlog._deltachain(rev)[0] |
|
820 | deltachain = lambda rev: revlog._deltachain(rev)[0] | |
821 |
|
821 | |||
822 | if gdelta: |
|
822 | if gdelta: | |
823 | # exclude already lazy tested base if any |
|
823 | # exclude already lazy tested base if any | |
824 | parents = [p for p in (p1, p2) if p != nullrev] |
|
824 | parents = [p for p in (p1, p2) if p != nullrev] | |
825 |
|
825 | |||
826 | if not revlog._deltabothparents and len(parents) == 2: |
|
826 | if not revlog._deltabothparents and len(parents) == 2: | |
827 | parents.sort() |
|
827 | parents.sort() | |
828 | # To minimize the chance of having to build a fulltext, |
|
828 | # To minimize the chance of having to build a fulltext, | |
829 | # pick first whichever parent is closest to us (max rev) |
|
829 | # pick first whichever parent is closest to us (max rev) | |
830 | yield (parents[1],) |
|
830 | yield (parents[1],) | |
831 | # then the other one (min rev) if the first did not fit |
|
831 | # then the other one (min rev) if the first did not fit | |
832 | yield (parents[0],) |
|
832 | yield (parents[0],) | |
833 | elif len(parents) > 0: |
|
833 | elif len(parents) > 0: | |
834 | # Test all parents (1 or 2), and keep the best candidate |
|
834 | # Test all parents (1 or 2), and keep the best candidate | |
835 | yield parents |
|
835 | yield parents | |
836 |
|
836 | |||
837 | if sparse and parents: |
|
837 | if sparse and parents: | |
838 | if snapshots is None: |
|
838 | if snapshots is None: | |
839 | # map: base-rev: snapshot-rev |
|
839 | # map: base-rev: snapshot-rev | |
840 | snapshots = collections.defaultdict(list) |
|
840 | snapshots = collections.defaultdict(list) | |
841 | # See if we can use an existing snapshot in the parent chains to use as |
|
841 | # See if we can use an existing snapshot in the parent chains to use as | |
842 | # a base for a new intermediate-snapshot |
|
842 | # a base for a new intermediate-snapshot | |
843 | # |
|
843 | # | |
844 | # search for snapshot in parents delta chain |
|
844 | # search for snapshot in parents delta chain | |
845 | # map: snapshot-level: snapshot-rev |
|
845 | # map: snapshot-level: snapshot-rev | |
846 | parents_snaps = collections.defaultdict(set) |
|
846 | parents_snaps = collections.defaultdict(set) | |
847 | candidate_chains = [deltachain(p) for p in parents] |
|
847 | candidate_chains = [deltachain(p) for p in parents] | |
848 | for chain in candidate_chains: |
|
848 | for chain in candidate_chains: | |
849 | for idx, s in enumerate(chain): |
|
849 | for idx, s in enumerate(chain): | |
850 | if not revlog.issnapshot(s): |
|
850 | if not revlog.issnapshot(s): | |
851 | break |
|
851 | break | |
852 | parents_snaps[idx].add(s) |
|
852 | parents_snaps[idx].add(s) | |
853 | snapfloor = min(parents_snaps[0]) + 1 |
|
853 | snapfloor = min(parents_snaps[0]) + 1 | |
854 | _findsnapshots(revlog, snapshots, snapfloor) |
|
854 | _findsnapshots(revlog, snapshots, snapfloor) | |
855 | # search for the highest "unrelated" revision |
|
855 | # search for the highest "unrelated" revision | |
856 | # |
|
856 | # | |
857 | # Adding snapshots used by "unrelated" revision increase the odd we |
|
857 | # Adding snapshots used by "unrelated" revision increase the odd we | |
858 | # reuse an independant, yet better snapshot chain. |
|
858 | # reuse an independant, yet better snapshot chain. | |
859 | # |
|
859 | # | |
860 | # XXX instead of building a set of revisions, we could lazily enumerate |
|
860 | # XXX instead of building a set of revisions, we could lazily enumerate | |
861 | # over the chains. That would be more efficient, however we stick to |
|
861 | # over the chains. That would be more efficient, however we stick to | |
862 | # simple code for now. |
|
862 | # simple code for now. | |
863 | all_revs = set() |
|
863 | all_revs = set() | |
864 | for chain in candidate_chains: |
|
864 | for chain in candidate_chains: | |
865 | all_revs.update(chain) |
|
865 | all_revs.update(chain) | |
866 | other = None |
|
866 | other = None | |
867 | for r in revlog.revs(prev, snapfloor): |
|
867 | for r in revlog.revs(prev, snapfloor): | |
868 | if r not in all_revs: |
|
868 | if r not in all_revs: | |
869 | other = r |
|
869 | other = r | |
870 | break |
|
870 | break | |
871 | if other is not None: |
|
871 | if other is not None: | |
872 | # To avoid unfair competition, we won't use unrelated intermediate |
|
872 | # To avoid unfair competition, we won't use unrelated intermediate | |
873 | # snapshot that are deeper than the ones from the parent delta |
|
873 | # snapshot that are deeper than the ones from the parent delta | |
874 | # chain. |
|
874 | # chain. | |
875 | max_depth = max(parents_snaps.keys()) |
|
875 | max_depth = max(parents_snaps.keys()) | |
876 | chain = deltachain(other) |
|
876 | chain = deltachain(other) | |
877 | for idx, s in enumerate(chain): |
|
877 | for idx, s in enumerate(chain): | |
878 | if s < snapfloor: |
|
878 | if s < snapfloor: | |
879 | continue |
|
879 | continue | |
880 | if max_depth < idx: |
|
880 | if max_depth < idx: | |
881 | break |
|
881 | break | |
882 | if not revlog.issnapshot(s): |
|
882 | if not revlog.issnapshot(s): | |
883 | break |
|
883 | break | |
884 | parents_snaps[idx].add(s) |
|
884 | parents_snaps[idx].add(s) | |
885 | # Test them as possible intermediate snapshot base |
|
885 | # Test them as possible intermediate snapshot base | |
886 | # We test them from highest to lowest level. High level one are more |
|
886 | # We test them from highest to lowest level. High level one are more | |
887 | # likely to result in small delta |
|
887 | # likely to result in small delta | |
888 | floor = None |
|
888 | floor = None | |
889 | for idx, snaps in sorted(parents_snaps.items(), reverse=True): |
|
889 | for idx, snaps in sorted(parents_snaps.items(), reverse=True): | |
890 | siblings = set() |
|
890 | siblings = set() | |
891 | for s in snaps: |
|
891 | for s in snaps: | |
892 | siblings.update(snapshots[s]) |
|
892 | siblings.update(snapshots[s]) | |
893 | # Before considering making a new intermediate snapshot, we check |
|
893 | # Before considering making a new intermediate snapshot, we check | |
894 | # if an existing snapshot, children of base we consider, would be |
|
894 | # if an existing snapshot, children of base we consider, would be | |
895 | # suitable. |
|
895 | # suitable. | |
896 | # |
|
896 | # | |
897 | # It give a change to reuse a delta chain "unrelated" to the |
|
897 | # It give a change to reuse a delta chain "unrelated" to the | |
898 | # current revision instead of starting our own. Without such |
|
898 | # current revision instead of starting our own. Without such | |
899 | # re-use, topological branches would keep reopening new chains. |
|
899 | # re-use, topological branches would keep reopening new chains. | |
900 | # Creating more and more snapshot as the repository grow. |
|
900 | # Creating more and more snapshot as the repository grow. | |
901 |
|
901 | |||
902 | if floor is not None: |
|
902 | if floor is not None: | |
903 | # We only do this for siblings created after the one in our |
|
903 | # We only do this for siblings created after the one in our | |
904 | # parent's delta chain. Those created before has less chances |
|
904 | # parent's delta chain. Those created before has less chances | |
905 | # to be valid base since our ancestors had to create a new |
|
905 | # to be valid base since our ancestors had to create a new | |
906 | # snapshot. |
|
906 | # snapshot. | |
907 | siblings = [r for r in siblings if floor < r] |
|
907 | siblings = [r for r in siblings if floor < r] | |
908 | yield tuple(sorted(siblings)) |
|
908 | yield tuple(sorted(siblings)) | |
909 | # then test the base from our parent's delta chain. |
|
909 | # then test the base from our parent's delta chain. | |
910 | yield tuple(sorted(snaps)) |
|
910 | yield tuple(sorted(snaps)) | |
911 | floor = min(snaps) |
|
911 | floor = min(snaps) | |
912 | # No suitable base found in the parent chain, search if any full |
|
912 | # No suitable base found in the parent chain, search if any full | |
913 | # snapshots emitted since parent's base would be a suitable base for an |
|
913 | # snapshots emitted since parent's base would be a suitable base for an | |
914 | # intermediate snapshot. |
|
914 | # intermediate snapshot. | |
915 | # |
|
915 | # | |
916 | # It give a chance to reuse a delta chain unrelated to the current |
|
916 | # It give a chance to reuse a delta chain unrelated to the current | |
917 | # revisions instead of starting our own. Without such re-use, |
|
917 | # revisions instead of starting our own. Without such re-use, | |
918 | # topological branches would keep reopening new full chains. Creating |
|
918 | # topological branches would keep reopening new full chains. Creating | |
919 | # more and more snapshot as the repository grow. |
|
919 | # more and more snapshot as the repository grow. | |
920 | yield tuple(snapshots[nullrev]) |
|
920 | yield tuple(snapshots[nullrev]) | |
921 |
|
921 | |||
922 | if not sparse: |
|
922 | if not sparse: | |
923 | # other approach failed try against prev to hopefully save us a |
|
923 | # other approach failed try against prev to hopefully save us a | |
924 | # fulltext. |
|
924 | # fulltext. | |
925 | yield (prev,) |
|
925 | yield (prev,) | |
926 |
|
926 | |||
927 |
|
927 | |||
928 | class deltacomputer(object): |
|
928 | class deltacomputer(object): | |
929 | def __init__(self, revlog): |
|
929 | def __init__(self, revlog): | |
930 | self.revlog = revlog |
|
930 | self.revlog = revlog | |
931 |
|
931 | |||
932 | def buildtext(self, revinfo, fh): |
|
932 | def buildtext(self, revinfo, fh): | |
933 | """Builds a fulltext version of a revision |
|
933 | """Builds a fulltext version of a revision | |
934 |
|
934 | |||
935 |
revinfo: |
|
935 | revinfo: revisioninfo instance that contains all needed info | |
936 | fh: file handle to either the .i or the .d revlog file, |
|
936 | fh: file handle to either the .i or the .d revlog file, | |
937 | depending on whether it is inlined or not |
|
937 | depending on whether it is inlined or not | |
938 | """ |
|
938 | """ | |
939 | btext = revinfo.btext |
|
939 | btext = revinfo.btext | |
940 | if btext[0] is not None: |
|
940 | if btext[0] is not None: | |
941 | return btext[0] |
|
941 | return btext[0] | |
942 |
|
942 | |||
943 | revlog = self.revlog |
|
943 | revlog = self.revlog | |
944 | cachedelta = revinfo.cachedelta |
|
944 | cachedelta = revinfo.cachedelta | |
945 | baserev = cachedelta[0] |
|
945 | baserev = cachedelta[0] | |
946 | delta = cachedelta[1] |
|
946 | delta = cachedelta[1] | |
947 |
|
947 | |||
948 | fulltext = btext[0] = _textfromdelta( |
|
948 | fulltext = btext[0] = _textfromdelta( | |
949 | fh, |
|
949 | fh, | |
950 | revlog, |
|
950 | revlog, | |
951 | baserev, |
|
951 | baserev, | |
952 | delta, |
|
952 | delta, | |
953 | revinfo.p1, |
|
953 | revinfo.p1, | |
954 | revinfo.p2, |
|
954 | revinfo.p2, | |
955 | revinfo.flags, |
|
955 | revinfo.flags, | |
956 | revinfo.node, |
|
956 | revinfo.node, | |
957 | ) |
|
957 | ) | |
958 | return fulltext |
|
958 | return fulltext | |
959 |
|
959 | |||
960 | def _builddeltadiff(self, base, revinfo, fh): |
|
960 | def _builddeltadiff(self, base, revinfo, fh): | |
961 | revlog = self.revlog |
|
961 | revlog = self.revlog | |
962 | t = self.buildtext(revinfo, fh) |
|
962 | t = self.buildtext(revinfo, fh) | |
963 | if revlog.iscensored(base): |
|
963 | if revlog.iscensored(base): | |
964 | # deltas based on a censored revision must replace the |
|
964 | # deltas based on a censored revision must replace the | |
965 | # full content in one patch, so delta works everywhere |
|
965 | # full content in one patch, so delta works everywhere | |
966 | header = mdiff.replacediffheader(revlog.rawsize(base), len(t)) |
|
966 | header = mdiff.replacediffheader(revlog.rawsize(base), len(t)) | |
967 | delta = header + t |
|
967 | delta = header + t | |
968 | else: |
|
968 | else: | |
969 | ptext = revlog.rawdata(base, _df=fh) |
|
969 | ptext = revlog.rawdata(base, _df=fh) | |
970 | delta = mdiff.textdiff(ptext, t) |
|
970 | delta = mdiff.textdiff(ptext, t) | |
971 |
|
971 | |||
972 | return delta |
|
972 | return delta | |
973 |
|
973 | |||
974 | def _builddeltainfo(self, revinfo, base, fh): |
|
974 | def _builddeltainfo(self, revinfo, base, fh): | |
975 | # can we use the cached delta? |
|
975 | # can we use the cached delta? | |
976 | revlog = self.revlog |
|
976 | revlog = self.revlog | |
977 | chainbase = revlog.chainbase(base) |
|
977 | chainbase = revlog.chainbase(base) | |
978 | if revlog._generaldelta: |
|
978 | if revlog._generaldelta: | |
979 | deltabase = base |
|
979 | deltabase = base | |
980 | else: |
|
980 | else: | |
981 | deltabase = chainbase |
|
981 | deltabase = chainbase | |
982 | snapshotdepth = None |
|
982 | snapshotdepth = None | |
983 | if revlog._sparserevlog and deltabase == nullrev: |
|
983 | if revlog._sparserevlog and deltabase == nullrev: | |
984 | snapshotdepth = 0 |
|
984 | snapshotdepth = 0 | |
985 | elif revlog._sparserevlog and revlog.issnapshot(deltabase): |
|
985 | elif revlog._sparserevlog and revlog.issnapshot(deltabase): | |
986 | # A delta chain should always be one full snapshot, |
|
986 | # A delta chain should always be one full snapshot, | |
987 | # zero or more semi-snapshots, and zero or more deltas |
|
987 | # zero or more semi-snapshots, and zero or more deltas | |
988 | p1, p2 = revlog.rev(revinfo.p1), revlog.rev(revinfo.p2) |
|
988 | p1, p2 = revlog.rev(revinfo.p1), revlog.rev(revinfo.p2) | |
989 | if deltabase not in (p1, p2) and revlog.issnapshot(deltabase): |
|
989 | if deltabase not in (p1, p2) and revlog.issnapshot(deltabase): | |
990 | snapshotdepth = len(revlog._deltachain(deltabase)[0]) |
|
990 | snapshotdepth = len(revlog._deltachain(deltabase)[0]) | |
991 | delta = None |
|
991 | delta = None | |
992 | if revinfo.cachedelta: |
|
992 | if revinfo.cachedelta: | |
993 | cachebase, cachediff = revinfo.cachedelta |
|
993 | cachebase, cachediff = revinfo.cachedelta | |
994 | # check if the diff still apply |
|
994 | # check if the diff still apply | |
995 | currentbase = cachebase |
|
995 | currentbase = cachebase | |
996 | while ( |
|
996 | while ( | |
997 | currentbase != nullrev |
|
997 | currentbase != nullrev | |
998 | and currentbase != base |
|
998 | and currentbase != base | |
999 | and self.revlog.length(currentbase) == 0 |
|
999 | and self.revlog.length(currentbase) == 0 | |
1000 | ): |
|
1000 | ): | |
1001 | currentbase = self.revlog.deltaparent(currentbase) |
|
1001 | currentbase = self.revlog.deltaparent(currentbase) | |
1002 | if self.revlog._lazydelta and currentbase == base: |
|
1002 | if self.revlog._lazydelta and currentbase == base: | |
1003 | delta = revinfo.cachedelta[1] |
|
1003 | delta = revinfo.cachedelta[1] | |
1004 | if delta is None: |
|
1004 | if delta is None: | |
1005 | delta = self._builddeltadiff(base, revinfo, fh) |
|
1005 | delta = self._builddeltadiff(base, revinfo, fh) | |
1006 | # snapshotdept need to be neither None nor 0 level snapshot |
|
1006 | # snapshotdept need to be neither None nor 0 level snapshot | |
1007 | if revlog.upperboundcomp is not None and snapshotdepth: |
|
1007 | if revlog.upperboundcomp is not None and snapshotdepth: | |
1008 | lowestrealisticdeltalen = len(delta) // revlog.upperboundcomp |
|
1008 | lowestrealisticdeltalen = len(delta) // revlog.upperboundcomp | |
1009 | snapshotlimit = revinfo.textlen >> snapshotdepth |
|
1009 | snapshotlimit = revinfo.textlen >> snapshotdepth | |
1010 | if snapshotlimit < lowestrealisticdeltalen: |
|
1010 | if snapshotlimit < lowestrealisticdeltalen: | |
1011 | return None |
|
1011 | return None | |
1012 | if revlog.length(base) < lowestrealisticdeltalen: |
|
1012 | if revlog.length(base) < lowestrealisticdeltalen: | |
1013 | return None |
|
1013 | return None | |
1014 | header, data = revlog.compress(delta) |
|
1014 | header, data = revlog.compress(delta) | |
1015 | deltalen = len(header) + len(data) |
|
1015 | deltalen = len(header) + len(data) | |
1016 | offset = revlog.end(len(revlog) - 1) |
|
1016 | offset = revlog.end(len(revlog) - 1) | |
1017 | dist = deltalen + offset - revlog.start(chainbase) |
|
1017 | dist = deltalen + offset - revlog.start(chainbase) | |
1018 | chainlen, compresseddeltalen = revlog._chaininfo(base) |
|
1018 | chainlen, compresseddeltalen = revlog._chaininfo(base) | |
1019 | chainlen += 1 |
|
1019 | chainlen += 1 | |
1020 | compresseddeltalen += deltalen |
|
1020 | compresseddeltalen += deltalen | |
1021 |
|
1021 | |||
1022 | return _deltainfo( |
|
1022 | return _deltainfo( | |
1023 | dist, |
|
1023 | dist, | |
1024 | deltalen, |
|
1024 | deltalen, | |
1025 | (header, data), |
|
1025 | (header, data), | |
1026 | deltabase, |
|
1026 | deltabase, | |
1027 | chainbase, |
|
1027 | chainbase, | |
1028 | chainlen, |
|
1028 | chainlen, | |
1029 | compresseddeltalen, |
|
1029 | compresseddeltalen, | |
1030 | snapshotdepth, |
|
1030 | snapshotdepth, | |
1031 | ) |
|
1031 | ) | |
1032 |
|
1032 | |||
1033 | def _fullsnapshotinfo(self, fh, revinfo): |
|
1033 | def _fullsnapshotinfo(self, fh, revinfo): | |
1034 | curr = len(self.revlog) |
|
1034 | curr = len(self.revlog) | |
1035 | rawtext = self.buildtext(revinfo, fh) |
|
1035 | rawtext = self.buildtext(revinfo, fh) | |
1036 | data = self.revlog.compress(rawtext) |
|
1036 | data = self.revlog.compress(rawtext) | |
1037 | compresseddeltalen = deltalen = dist = len(data[1]) + len(data[0]) |
|
1037 | compresseddeltalen = deltalen = dist = len(data[1]) + len(data[0]) | |
1038 | deltabase = chainbase = curr |
|
1038 | deltabase = chainbase = curr | |
1039 | snapshotdepth = 0 |
|
1039 | snapshotdepth = 0 | |
1040 | chainlen = 1 |
|
1040 | chainlen = 1 | |
1041 |
|
1041 | |||
1042 | return _deltainfo( |
|
1042 | return _deltainfo( | |
1043 | dist, |
|
1043 | dist, | |
1044 | deltalen, |
|
1044 | deltalen, | |
1045 | data, |
|
1045 | data, | |
1046 | deltabase, |
|
1046 | deltabase, | |
1047 | chainbase, |
|
1047 | chainbase, | |
1048 | chainlen, |
|
1048 | chainlen, | |
1049 | compresseddeltalen, |
|
1049 | compresseddeltalen, | |
1050 | snapshotdepth, |
|
1050 | snapshotdepth, | |
1051 | ) |
|
1051 | ) | |
1052 |
|
1052 | |||
1053 | def finddeltainfo(self, revinfo, fh): |
|
1053 | def finddeltainfo(self, revinfo, fh): | |
1054 | """Find an acceptable delta against a candidate revision |
|
1054 | """Find an acceptable delta against a candidate revision | |
1055 |
|
1055 | |||
1056 | revinfo: information about the revision (instance of _revisioninfo) |
|
1056 | revinfo: information about the revision (instance of _revisioninfo) | |
1057 | fh: file handle to either the .i or the .d revlog file, |
|
1057 | fh: file handle to either the .i or the .d revlog file, | |
1058 | depending on whether it is inlined or not |
|
1058 | depending on whether it is inlined or not | |
1059 |
|
1059 | |||
1060 | Returns the first acceptable candidate revision, as ordered by |
|
1060 | Returns the first acceptable candidate revision, as ordered by | |
1061 | _candidategroups |
|
1061 | _candidategroups | |
1062 |
|
1062 | |||
1063 | If no suitable deltabase is found, we return delta info for a full |
|
1063 | If no suitable deltabase is found, we return delta info for a full | |
1064 | snapshot. |
|
1064 | snapshot. | |
1065 | """ |
|
1065 | """ | |
1066 | if not revinfo.textlen: |
|
1066 | if not revinfo.textlen: | |
1067 | return self._fullsnapshotinfo(fh, revinfo) |
|
1067 | return self._fullsnapshotinfo(fh, revinfo) | |
1068 |
|
1068 | |||
1069 | # no delta for flag processor revision (see "candelta" for why) |
|
1069 | # no delta for flag processor revision (see "candelta" for why) | |
1070 | # not calling candelta since only one revision needs test, also to |
|
1070 | # not calling candelta since only one revision needs test, also to | |
1071 | # avoid overhead fetching flags again. |
|
1071 | # avoid overhead fetching flags again. | |
1072 | if revinfo.flags & REVIDX_RAWTEXT_CHANGING_FLAGS: |
|
1072 | if revinfo.flags & REVIDX_RAWTEXT_CHANGING_FLAGS: | |
1073 | return self._fullsnapshotinfo(fh, revinfo) |
|
1073 | return self._fullsnapshotinfo(fh, revinfo) | |
1074 |
|
1074 | |||
1075 | cachedelta = revinfo.cachedelta |
|
1075 | cachedelta = revinfo.cachedelta | |
1076 | p1 = revinfo.p1 |
|
1076 | p1 = revinfo.p1 | |
1077 | p2 = revinfo.p2 |
|
1077 | p2 = revinfo.p2 | |
1078 | revlog = self.revlog |
|
1078 | revlog = self.revlog | |
1079 |
|
1079 | |||
1080 | deltainfo = None |
|
1080 | deltainfo = None | |
1081 | p1r, p2r = revlog.rev(p1), revlog.rev(p2) |
|
1081 | p1r, p2r = revlog.rev(p1), revlog.rev(p2) | |
1082 | groups = _candidategroups( |
|
1082 | groups = _candidategroups( | |
1083 | self.revlog, revinfo.textlen, p1r, p2r, cachedelta |
|
1083 | self.revlog, revinfo.textlen, p1r, p2r, cachedelta | |
1084 | ) |
|
1084 | ) | |
1085 | candidaterevs = next(groups) |
|
1085 | candidaterevs = next(groups) | |
1086 | while candidaterevs is not None: |
|
1086 | while candidaterevs is not None: | |
1087 | nominateddeltas = [] |
|
1087 | nominateddeltas = [] | |
1088 | if deltainfo is not None: |
|
1088 | if deltainfo is not None: | |
1089 | # if we already found a good delta, |
|
1089 | # if we already found a good delta, | |
1090 | # challenge it against refined candidates |
|
1090 | # challenge it against refined candidates | |
1091 | nominateddeltas.append(deltainfo) |
|
1091 | nominateddeltas.append(deltainfo) | |
1092 | for candidaterev in candidaterevs: |
|
1092 | for candidaterev in candidaterevs: | |
1093 | candidatedelta = self._builddeltainfo(revinfo, candidaterev, fh) |
|
1093 | candidatedelta = self._builddeltainfo(revinfo, candidaterev, fh) | |
1094 | if candidatedelta is not None: |
|
1094 | if candidatedelta is not None: | |
1095 | if isgooddeltainfo(self.revlog, candidatedelta, revinfo): |
|
1095 | if isgooddeltainfo(self.revlog, candidatedelta, revinfo): | |
1096 | nominateddeltas.append(candidatedelta) |
|
1096 | nominateddeltas.append(candidatedelta) | |
1097 | if nominateddeltas: |
|
1097 | if nominateddeltas: | |
1098 | deltainfo = min(nominateddeltas, key=lambda x: x.deltalen) |
|
1098 | deltainfo = min(nominateddeltas, key=lambda x: x.deltalen) | |
1099 | if deltainfo is not None: |
|
1099 | if deltainfo is not None: | |
1100 | candidaterevs = groups.send(deltainfo.base) |
|
1100 | candidaterevs = groups.send(deltainfo.base) | |
1101 | else: |
|
1101 | else: | |
1102 | candidaterevs = next(groups) |
|
1102 | candidaterevs = next(groups) | |
1103 |
|
1103 | |||
1104 | if deltainfo is None: |
|
1104 | if deltainfo is None: | |
1105 | deltainfo = self._fullsnapshotinfo(fh, revinfo) |
|
1105 | deltainfo = self._fullsnapshotinfo(fh, revinfo) | |
1106 | return deltainfo |
|
1106 | return deltainfo |
General Comments 0
You need to be logged in to leave comments.
Login now