Show More
@@ -1,3731 +1,3740 b'' | |||||
1 | # revlog.py - storage back-end for mercurial |
|
1 | # revlog.py - storage back-end for mercurial | |
2 | # coding: utf8 |
|
2 | # coding: utf8 | |
3 | # |
|
3 | # | |
4 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> |
|
4 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> | |
5 | # |
|
5 | # | |
6 | # This software may be used and distributed according to the terms of the |
|
6 | # This software may be used and distributed according to the terms of the | |
7 | # GNU General Public License version 2 or any later version. |
|
7 | # GNU General Public License version 2 or any later version. | |
8 |
|
8 | |||
9 | """Storage back-end for Mercurial. |
|
9 | """Storage back-end for Mercurial. | |
10 |
|
10 | |||
11 | This provides efficient delta storage with O(1) retrieve and append |
|
11 | This provides efficient delta storage with O(1) retrieve and append | |
12 | and O(changes) merge between branches. |
|
12 | and O(changes) merge between branches. | |
13 | """ |
|
13 | """ | |
14 |
|
14 | |||
15 |
|
15 | |||
16 | import binascii |
|
16 | import binascii | |
17 | import collections |
|
17 | import collections | |
18 | import contextlib |
|
18 | import contextlib | |
19 | import io |
|
19 | import io | |
20 | import os |
|
20 | import os | |
21 | import struct |
|
21 | import struct | |
22 | import weakref |
|
22 | import weakref | |
23 | import zlib |
|
23 | import zlib | |
24 |
|
24 | |||
25 | # import stuff from node for others to import from revlog |
|
25 | # import stuff from node for others to import from revlog | |
26 | from .node import ( |
|
26 | from .node import ( | |
27 | bin, |
|
27 | bin, | |
28 | hex, |
|
28 | hex, | |
29 | nullrev, |
|
29 | nullrev, | |
30 | sha1nodeconstants, |
|
30 | sha1nodeconstants, | |
31 | short, |
|
31 | short, | |
32 | wdirrev, |
|
32 | wdirrev, | |
33 | ) |
|
33 | ) | |
34 | from .i18n import _ |
|
34 | from .i18n import _ | |
35 | from .revlogutils.constants import ( |
|
35 | from .revlogutils.constants import ( | |
36 | ALL_KINDS, |
|
36 | ALL_KINDS, | |
37 | CHANGELOGV2, |
|
37 | CHANGELOGV2, | |
38 | COMP_MODE_DEFAULT, |
|
38 | COMP_MODE_DEFAULT, | |
39 | COMP_MODE_INLINE, |
|
39 | COMP_MODE_INLINE, | |
40 | COMP_MODE_PLAIN, |
|
40 | COMP_MODE_PLAIN, | |
41 | DELTA_BASE_REUSE_NO, |
|
41 | DELTA_BASE_REUSE_NO, | |
42 | DELTA_BASE_REUSE_TRY, |
|
42 | DELTA_BASE_REUSE_TRY, | |
43 | ENTRY_RANK, |
|
43 | ENTRY_RANK, | |
44 | FEATURES_BY_VERSION, |
|
44 | FEATURES_BY_VERSION, | |
45 | FLAG_GENERALDELTA, |
|
45 | FLAG_GENERALDELTA, | |
46 | FLAG_INLINE_DATA, |
|
46 | FLAG_INLINE_DATA, | |
47 | INDEX_HEADER, |
|
47 | INDEX_HEADER, | |
48 | KIND_CHANGELOG, |
|
48 | KIND_CHANGELOG, | |
49 | KIND_FILELOG, |
|
49 | KIND_FILELOG, | |
50 | RANK_UNKNOWN, |
|
50 | RANK_UNKNOWN, | |
51 | REVLOGV0, |
|
51 | REVLOGV0, | |
52 | REVLOGV1, |
|
52 | REVLOGV1, | |
53 | REVLOGV1_FLAGS, |
|
53 | REVLOGV1_FLAGS, | |
54 | REVLOGV2, |
|
54 | REVLOGV2, | |
55 | REVLOGV2_FLAGS, |
|
55 | REVLOGV2_FLAGS, | |
56 | REVLOG_DEFAULT_FLAGS, |
|
56 | REVLOG_DEFAULT_FLAGS, | |
57 | REVLOG_DEFAULT_FORMAT, |
|
57 | REVLOG_DEFAULT_FORMAT, | |
58 | REVLOG_DEFAULT_VERSION, |
|
58 | REVLOG_DEFAULT_VERSION, | |
59 | SUPPORTED_FLAGS, |
|
59 | SUPPORTED_FLAGS, | |
60 | ) |
|
60 | ) | |
61 | from .revlogutils.flagutil import ( |
|
61 | from .revlogutils.flagutil import ( | |
62 | REVIDX_DEFAULT_FLAGS, |
|
62 | REVIDX_DEFAULT_FLAGS, | |
63 | REVIDX_ELLIPSIS, |
|
63 | REVIDX_ELLIPSIS, | |
64 | REVIDX_EXTSTORED, |
|
64 | REVIDX_EXTSTORED, | |
65 | REVIDX_FLAGS_ORDER, |
|
65 | REVIDX_FLAGS_ORDER, | |
66 | REVIDX_HASCOPIESINFO, |
|
66 | REVIDX_HASCOPIESINFO, | |
67 | REVIDX_ISCENSORED, |
|
67 | REVIDX_ISCENSORED, | |
68 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
68 | REVIDX_RAWTEXT_CHANGING_FLAGS, | |
69 | ) |
|
69 | ) | |
70 | from .thirdparty import attr |
|
70 | from .thirdparty import attr | |
71 | from . import ( |
|
71 | from . import ( | |
72 | ancestor, |
|
72 | ancestor, | |
73 | dagop, |
|
73 | dagop, | |
74 | error, |
|
74 | error, | |
75 | mdiff, |
|
75 | mdiff, | |
76 | policy, |
|
76 | policy, | |
77 | pycompat, |
|
77 | pycompat, | |
78 | revlogutils, |
|
78 | revlogutils, | |
79 | templatefilters, |
|
79 | templatefilters, | |
80 | util, |
|
80 | util, | |
81 | ) |
|
81 | ) | |
82 | from .interfaces import ( |
|
82 | from .interfaces import ( | |
83 | repository, |
|
83 | repository, | |
84 | util as interfaceutil, |
|
84 | util as interfaceutil, | |
85 | ) |
|
85 | ) | |
86 | from .revlogutils import ( |
|
86 | from .revlogutils import ( | |
87 | deltas as deltautil, |
|
87 | deltas as deltautil, | |
88 | docket as docketutil, |
|
88 | docket as docketutil, | |
89 | flagutil, |
|
89 | flagutil, | |
90 | nodemap as nodemaputil, |
|
90 | nodemap as nodemaputil, | |
91 | randomaccessfile, |
|
91 | randomaccessfile, | |
92 | revlogv0, |
|
92 | revlogv0, | |
93 | rewrite, |
|
93 | rewrite, | |
94 | sidedata as sidedatautil, |
|
94 | sidedata as sidedatautil, | |
95 | ) |
|
95 | ) | |
96 | from .utils import ( |
|
96 | from .utils import ( | |
97 | storageutil, |
|
97 | storageutil, | |
98 | stringutil, |
|
98 | stringutil, | |
99 | ) |
|
99 | ) | |
100 |
|
100 | |||
101 | # blanked usage of all the name to prevent pyflakes constraints |
|
101 | # blanked usage of all the name to prevent pyflakes constraints | |
102 | # We need these name available in the module for extensions. |
|
102 | # We need these name available in the module for extensions. | |
103 |
|
103 | |||
104 | REVLOGV0 |
|
104 | REVLOGV0 | |
105 | REVLOGV1 |
|
105 | REVLOGV1 | |
106 | REVLOGV2 |
|
106 | REVLOGV2 | |
107 | CHANGELOGV2 |
|
107 | CHANGELOGV2 | |
108 | FLAG_INLINE_DATA |
|
108 | FLAG_INLINE_DATA | |
109 | FLAG_GENERALDELTA |
|
109 | FLAG_GENERALDELTA | |
110 | REVLOG_DEFAULT_FLAGS |
|
110 | REVLOG_DEFAULT_FLAGS | |
111 | REVLOG_DEFAULT_FORMAT |
|
111 | REVLOG_DEFAULT_FORMAT | |
112 | REVLOG_DEFAULT_VERSION |
|
112 | REVLOG_DEFAULT_VERSION | |
113 | REVLOGV1_FLAGS |
|
113 | REVLOGV1_FLAGS | |
114 | REVLOGV2_FLAGS |
|
114 | REVLOGV2_FLAGS | |
115 | REVIDX_ISCENSORED |
|
115 | REVIDX_ISCENSORED | |
116 | REVIDX_ELLIPSIS |
|
116 | REVIDX_ELLIPSIS | |
117 | REVIDX_HASCOPIESINFO |
|
117 | REVIDX_HASCOPIESINFO | |
118 | REVIDX_EXTSTORED |
|
118 | REVIDX_EXTSTORED | |
119 | REVIDX_DEFAULT_FLAGS |
|
119 | REVIDX_DEFAULT_FLAGS | |
120 | REVIDX_FLAGS_ORDER |
|
120 | REVIDX_FLAGS_ORDER | |
121 | REVIDX_RAWTEXT_CHANGING_FLAGS |
|
121 | REVIDX_RAWTEXT_CHANGING_FLAGS | |
122 |
|
122 | |||
123 | parsers = policy.importmod('parsers') |
|
123 | parsers = policy.importmod('parsers') | |
124 | rustancestor = policy.importrust('ancestor') |
|
124 | rustancestor = policy.importrust('ancestor') | |
125 | rustdagop = policy.importrust('dagop') |
|
125 | rustdagop = policy.importrust('dagop') | |
126 | rustrevlog = policy.importrust('revlog') |
|
126 | rustrevlog = policy.importrust('revlog') | |
127 |
|
127 | |||
128 | # Aliased for performance. |
|
128 | # Aliased for performance. | |
129 | _zlibdecompress = zlib.decompress |
|
129 | _zlibdecompress = zlib.decompress | |
130 |
|
130 | |||
131 | # max size of inline data embedded into a revlog |
|
131 | # max size of inline data embedded into a revlog | |
132 | _maxinline = 131072 |
|
132 | _maxinline = 131072 | |
133 |
|
133 | |||
134 | # Flag processors for REVIDX_ELLIPSIS. |
|
134 | # Flag processors for REVIDX_ELLIPSIS. | |
135 | def ellipsisreadprocessor(rl, text): |
|
135 | def ellipsisreadprocessor(rl, text): | |
136 | return text, False |
|
136 | return text, False | |
137 |
|
137 | |||
138 |
|
138 | |||
139 | def ellipsiswriteprocessor(rl, text): |
|
139 | def ellipsiswriteprocessor(rl, text): | |
140 | return text, False |
|
140 | return text, False | |
141 |
|
141 | |||
142 |
|
142 | |||
143 | def ellipsisrawprocessor(rl, text): |
|
143 | def ellipsisrawprocessor(rl, text): | |
144 | return False |
|
144 | return False | |
145 |
|
145 | |||
146 |
|
146 | |||
147 | ellipsisprocessor = ( |
|
147 | ellipsisprocessor = ( | |
148 | ellipsisreadprocessor, |
|
148 | ellipsisreadprocessor, | |
149 | ellipsiswriteprocessor, |
|
149 | ellipsiswriteprocessor, | |
150 | ellipsisrawprocessor, |
|
150 | ellipsisrawprocessor, | |
151 | ) |
|
151 | ) | |
152 |
|
152 | |||
153 |
|
153 | |||
154 | def _verify_revision(rl, skipflags, state, node): |
|
154 | def _verify_revision(rl, skipflags, state, node): | |
155 | """Verify the integrity of the given revlog ``node`` while providing a hook |
|
155 | """Verify the integrity of the given revlog ``node`` while providing a hook | |
156 | point for extensions to influence the operation.""" |
|
156 | point for extensions to influence the operation.""" | |
157 | if skipflags: |
|
157 | if skipflags: | |
158 | state[b'skipread'].add(node) |
|
158 | state[b'skipread'].add(node) | |
159 | else: |
|
159 | else: | |
160 | # Side-effect: read content and verify hash. |
|
160 | # Side-effect: read content and verify hash. | |
161 | rl.revision(node) |
|
161 | rl.revision(node) | |
162 |
|
162 | |||
163 |
|
163 | |||
164 | # True if a fast implementation for persistent-nodemap is available |
|
164 | # True if a fast implementation for persistent-nodemap is available | |
165 | # |
|
165 | # | |
166 | # We also consider we have a "fast" implementation in "pure" python because |
|
166 | # We also consider we have a "fast" implementation in "pure" python because | |
167 | # people using pure don't really have performance consideration (and a |
|
167 | # people using pure don't really have performance consideration (and a | |
168 | # wheelbarrow of other slowness source) |
|
168 | # wheelbarrow of other slowness source) | |
169 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or hasattr( |
|
169 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or hasattr( | |
170 | parsers, 'BaseIndexObject' |
|
170 | parsers, 'BaseIndexObject' | |
171 | ) |
|
171 | ) | |
172 |
|
172 | |||
173 |
|
173 | |||
174 | @interfaceutil.implementer(repository.irevisiondelta) |
|
174 | @interfaceutil.implementer(repository.irevisiondelta) | |
175 | @attr.s(slots=True) |
|
175 | @attr.s(slots=True) | |
176 | class revlogrevisiondelta: |
|
176 | class revlogrevisiondelta: | |
177 | node = attr.ib() |
|
177 | node = attr.ib() | |
178 | p1node = attr.ib() |
|
178 | p1node = attr.ib() | |
179 | p2node = attr.ib() |
|
179 | p2node = attr.ib() | |
180 | basenode = attr.ib() |
|
180 | basenode = attr.ib() | |
181 | flags = attr.ib() |
|
181 | flags = attr.ib() | |
182 | baserevisionsize = attr.ib() |
|
182 | baserevisionsize = attr.ib() | |
183 | revision = attr.ib() |
|
183 | revision = attr.ib() | |
184 | delta = attr.ib() |
|
184 | delta = attr.ib() | |
185 | sidedata = attr.ib() |
|
185 | sidedata = attr.ib() | |
186 | protocol_flags = attr.ib() |
|
186 | protocol_flags = attr.ib() | |
187 | linknode = attr.ib(default=None) |
|
187 | linknode = attr.ib(default=None) | |
188 |
|
188 | |||
189 |
|
189 | |||
190 | @interfaceutil.implementer(repository.iverifyproblem) |
|
190 | @interfaceutil.implementer(repository.iverifyproblem) | |
191 | @attr.s(frozen=True) |
|
191 | @attr.s(frozen=True) | |
192 | class revlogproblem: |
|
192 | class revlogproblem: | |
193 | warning = attr.ib(default=None) |
|
193 | warning = attr.ib(default=None) | |
194 | error = attr.ib(default=None) |
|
194 | error = attr.ib(default=None) | |
195 | node = attr.ib(default=None) |
|
195 | node = attr.ib(default=None) | |
196 |
|
196 | |||
197 |
|
197 | |||
198 | def parse_index_v1(data, inline): |
|
198 | def parse_index_v1(data, inline): | |
199 | # call the C implementation to parse the index data |
|
199 | # call the C implementation to parse the index data | |
200 | index, cache = parsers.parse_index2(data, inline) |
|
200 | index, cache = parsers.parse_index2(data, inline) | |
201 | return index, cache |
|
201 | return index, cache | |
202 |
|
202 | |||
203 |
|
203 | |||
204 | def parse_index_v2(data, inline): |
|
204 | def parse_index_v2(data, inline): | |
205 | # call the C implementation to parse the index data |
|
205 | # call the C implementation to parse the index data | |
206 | index, cache = parsers.parse_index2(data, inline, format=REVLOGV2) |
|
206 | index, cache = parsers.parse_index2(data, inline, format=REVLOGV2) | |
207 | return index, cache |
|
207 | return index, cache | |
208 |
|
208 | |||
209 |
|
209 | |||
210 | def parse_index_cl_v2(data, inline): |
|
210 | def parse_index_cl_v2(data, inline): | |
211 | # call the C implementation to parse the index data |
|
211 | # call the C implementation to parse the index data | |
212 | index, cache = parsers.parse_index2(data, inline, format=CHANGELOGV2) |
|
212 | index, cache = parsers.parse_index2(data, inline, format=CHANGELOGV2) | |
213 | return index, cache |
|
213 | return index, cache | |
214 |
|
214 | |||
215 |
|
215 | |||
216 | if hasattr(parsers, 'parse_index_devel_nodemap'): |
|
216 | if hasattr(parsers, 'parse_index_devel_nodemap'): | |
217 |
|
217 | |||
218 | def parse_index_v1_nodemap(data, inline): |
|
218 | def parse_index_v1_nodemap(data, inline): | |
219 | index, cache = parsers.parse_index_devel_nodemap(data, inline) |
|
219 | index, cache = parsers.parse_index_devel_nodemap(data, inline) | |
220 | return index, cache |
|
220 | return index, cache | |
221 |
|
221 | |||
222 |
|
222 | |||
223 | else: |
|
223 | else: | |
224 | parse_index_v1_nodemap = None |
|
224 | parse_index_v1_nodemap = None | |
225 |
|
225 | |||
226 |
|
226 | |||
227 | def parse_index_v1_mixed(data, inline): |
|
227 | def parse_index_v1_mixed(data, inline): | |
228 | index, cache = parse_index_v1(data, inline) |
|
228 | index, cache = parse_index_v1(data, inline) | |
229 | return rustrevlog.MixedIndex(index), cache |
|
229 | return rustrevlog.MixedIndex(index), cache | |
230 |
|
230 | |||
231 |
|
231 | |||
232 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte |
|
232 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte | |
233 | # signed integer) |
|
233 | # signed integer) | |
234 | _maxentrysize = 0x7FFFFFFF |
|
234 | _maxentrysize = 0x7FFFFFFF | |
235 |
|
235 | |||
236 | FILE_TOO_SHORT_MSG = _( |
|
236 | FILE_TOO_SHORT_MSG = _( | |
237 | b'cannot read from revlog %s;' |
|
237 | b'cannot read from revlog %s;' | |
238 | b' expected %d bytes from offset %d, data size is %d' |
|
238 | b' expected %d bytes from offset %d, data size is %d' | |
239 | ) |
|
239 | ) | |
240 |
|
240 | |||
241 | hexdigits = b'0123456789abcdefABCDEF' |
|
241 | hexdigits = b'0123456789abcdefABCDEF' | |
242 |
|
242 | |||
243 |
|
243 | |||
|
244 | class _Config: | |||
|
245 | def copy(self): | |||
|
246 | return self.__class__(**self.__dict__) | |||
|
247 | ||||
|
248 | ||||
244 | @attr.s() |
|
249 | @attr.s() | |
245 | class FeatureConfig: |
|
250 | class FeatureConfig(_Config): | |
246 | """Hold configuration values about the available revlog features""" |
|
251 | """Hold configuration values about the available revlog features""" | |
247 |
|
252 | |||
248 | # the default compression engine |
|
253 | # the default compression engine | |
249 | compression_engine = attr.ib(default=b'zlib') |
|
254 | compression_engine = attr.ib(default=b'zlib') | |
250 | # compression engines options |
|
255 | # compression engines options | |
251 | compression_engine_options = attr.ib(default=attr.Factory(dict)) |
|
256 | compression_engine_options = attr.ib(default=attr.Factory(dict)) | |
252 |
|
257 | |||
253 | # can we use censor on this revlog |
|
258 | # can we use censor on this revlog | |
254 | censorable = attr.ib(default=False) |
|
259 | censorable = attr.ib(default=False) | |
255 | # does this revlog use the "side data" feature |
|
260 | # does this revlog use the "side data" feature | |
256 | has_side_data = attr.ib(default=False) |
|
261 | has_side_data = attr.ib(default=False) | |
257 | # might remove rank configuration once the computation has no impact |
|
262 | # might remove rank configuration once the computation has no impact | |
258 | compute_rank = attr.ib(default=False) |
|
263 | compute_rank = attr.ib(default=False) | |
259 | # parent order is supposed to be semantically irrelevant, so we |
|
264 | # parent order is supposed to be semantically irrelevant, so we | |
260 | # normally resort parents to ensure that the first parent is non-null, |
|
265 | # normally resort parents to ensure that the first parent is non-null, | |
261 | # if there is a non-null parent at all. |
|
266 | # if there is a non-null parent at all. | |
262 | # filelog abuses the parent order as flag to mark some instances of |
|
267 | # filelog abuses the parent order as flag to mark some instances of | |
263 | # meta-encoded files, so allow it to disable this behavior. |
|
268 | # meta-encoded files, so allow it to disable this behavior. | |
264 | canonical_parent_order = attr.ib(default=False) |
|
269 | canonical_parent_order = attr.ib(default=False) | |
265 | # can ellipsis commit be used |
|
270 | # can ellipsis commit be used | |
266 | enable_ellipsis = attr.ib(default=False) |
|
271 | enable_ellipsis = attr.ib(default=False) | |
267 |
|
272 | |||
|
273 | def copy(self): | |||
|
274 | new = super().copy() | |||
|
275 | new.compression_engine_options = self.compression_engine_options.copy() | |||
|
276 | return new | |||
|
277 | ||||
268 |
|
278 | |||
269 | @attr.s() |
|
279 | @attr.s() | |
270 | class DataConfig: |
|
280 | class DataConfig(_Config): | |
271 | """Hold configuration value about how the revlog data are read""" |
|
281 | """Hold configuration value about how the revlog data are read""" | |
272 |
|
282 | |||
273 | # should we try to open the "pending" version of the revlog |
|
283 | # should we try to open the "pending" version of the revlog | |
274 | try_pending = attr.ib(default=False) |
|
284 | try_pending = attr.ib(default=False) | |
275 | # should we try to open the "splitted" version of the revlog |
|
285 | # should we try to open the "splitted" version of the revlog | |
276 | try_split = attr.ib(default=False) |
|
286 | try_split = attr.ib(default=False) | |
277 | # When True, indexfile should be opened with checkambig=True at writing, |
|
287 | # When True, indexfile should be opened with checkambig=True at writing, | |
278 | # to avoid file stat ambiguity. |
|
288 | # to avoid file stat ambiguity. | |
279 | check_ambig = attr.ib(default=False) |
|
289 | check_ambig = attr.ib(default=False) | |
280 |
|
290 | |||
281 | # If true, use mmap instead of reading to deal with large index |
|
291 | # If true, use mmap instead of reading to deal with large index | |
282 | mmap_large_index = attr.ib(default=False) |
|
292 | mmap_large_index = attr.ib(default=False) | |
283 | # how much data is large |
|
293 | # how much data is large | |
284 | mmap_index_threshold = attr.ib(default=None) |
|
294 | mmap_index_threshold = attr.ib(default=None) | |
285 | # How much data to read and cache into the raw revlog data cache. |
|
295 | # How much data to read and cache into the raw revlog data cache. | |
286 | chunk_cache_size = attr.ib(default=65536) |
|
296 | chunk_cache_size = attr.ib(default=65536) | |
287 |
|
297 | |||
288 | # Allow sparse reading of the revlog data |
|
298 | # Allow sparse reading of the revlog data | |
289 | with_sparse_read = attr.ib(default=False) |
|
299 | with_sparse_read = attr.ib(default=False) | |
290 | # minimal density of a sparse read chunk |
|
300 | # minimal density of a sparse read chunk | |
291 | sr_density_threshold = attr.ib(default=0.50) |
|
301 | sr_density_threshold = attr.ib(default=0.50) | |
292 | # minimal size of data we skip when performing sparse read |
|
302 | # minimal size of data we skip when performing sparse read | |
293 | sr_min_gap_size = attr.ib(default=262144) |
|
303 | sr_min_gap_size = attr.ib(default=262144) | |
294 |
|
304 | |||
295 | # are delta encoded against arbitrary bases. |
|
305 | # are delta encoded against arbitrary bases. | |
296 | generaldelta = attr.ib(default=False) |
|
306 | generaldelta = attr.ib(default=False) | |
297 |
|
307 | |||
298 |
|
308 | |||
299 | @attr.s() |
|
309 | @attr.s() | |
300 | class DeltaConfig: |
|
310 | class DeltaConfig(_Config): | |
301 | """Hold configuration value about how new delta are computed |
|
311 | """Hold configuration value about how new delta are computed | |
302 |
|
312 | |||
303 | Some attributes are duplicated from DataConfig to help havign each object |
|
313 | Some attributes are duplicated from DataConfig to help havign each object | |
304 | self contained. |
|
314 | self contained. | |
305 | """ |
|
315 | """ | |
306 |
|
316 | |||
307 | # can delta be encoded against arbitrary bases. |
|
317 | # can delta be encoded against arbitrary bases. | |
308 | general_delta = attr.ib(default=False) |
|
318 | general_delta = attr.ib(default=False) | |
309 | # Allow sparse writing of the revlog data |
|
319 | # Allow sparse writing of the revlog data | |
310 | sparse_revlog = attr.ib(default=False) |
|
320 | sparse_revlog = attr.ib(default=False) | |
311 | # maximum length of a delta chain |
|
321 | # maximum length of a delta chain | |
312 | max_chain_len = attr.ib(default=None) |
|
322 | max_chain_len = attr.ib(default=None) | |
313 | # Maximum distance between delta chain base start and end |
|
323 | # Maximum distance between delta chain base start and end | |
314 | max_deltachain_span = attr.ib(default=-1) |
|
324 | max_deltachain_span = attr.ib(default=-1) | |
315 | # If `upper_bound_comp` is not None, this is the expected maximal gain from |
|
325 | # If `upper_bound_comp` is not None, this is the expected maximal gain from | |
316 | # compression for the data content. |
|
326 | # compression for the data content. | |
317 | upper_bound_comp = attr.ib(default=None) |
|
327 | upper_bound_comp = attr.ib(default=None) | |
318 | # Should we try a delta against both parent |
|
328 | # Should we try a delta against both parent | |
319 | delta_both_parents = attr.ib(default=True) |
|
329 | delta_both_parents = attr.ib(default=True) | |
320 | # Test delta base candidate group by chunk of this maximal size. |
|
330 | # Test delta base candidate group by chunk of this maximal size. | |
321 | candidate_group_chunk_size = attr.ib(default=0) |
|
331 | candidate_group_chunk_size = attr.ib(default=0) | |
322 | # Should we display debug information about delta computation |
|
332 | # Should we display debug information about delta computation | |
323 | debug_delta = attr.ib(default=False) |
|
333 | debug_delta = attr.ib(default=False) | |
324 | # trust incoming delta by default |
|
334 | # trust incoming delta by default | |
325 | lazy_delta = attr.ib(default=True) |
|
335 | lazy_delta = attr.ib(default=True) | |
326 | # trust the base of incoming delta by default |
|
336 | # trust the base of incoming delta by default | |
327 | lazy_delta_base = attr.ib(default=False) |
|
337 | lazy_delta_base = attr.ib(default=False) | |
328 |
|
338 | |||
329 |
|
339 | |||
330 | class revlog: |
|
340 | class revlog: | |
331 | """ |
|
341 | """ | |
332 | the underlying revision storage object |
|
342 | the underlying revision storage object | |
333 |
|
343 | |||
334 | A revlog consists of two parts, an index and the revision data. |
|
344 | A revlog consists of two parts, an index and the revision data. | |
335 |
|
345 | |||
336 | The index is a file with a fixed record size containing |
|
346 | The index is a file with a fixed record size containing | |
337 | information on each revision, including its nodeid (hash), the |
|
347 | information on each revision, including its nodeid (hash), the | |
338 | nodeids of its parents, the position and offset of its data within |
|
348 | nodeids of its parents, the position and offset of its data within | |
339 | the data file, and the revision it's based on. Finally, each entry |
|
349 | the data file, and the revision it's based on. Finally, each entry | |
340 | contains a linkrev entry that can serve as a pointer to external |
|
350 | contains a linkrev entry that can serve as a pointer to external | |
341 | data. |
|
351 | data. | |
342 |
|
352 | |||
343 | The revision data itself is a linear collection of data chunks. |
|
353 | The revision data itself is a linear collection of data chunks. | |
344 | Each chunk represents a revision and is usually represented as a |
|
354 | Each chunk represents a revision and is usually represented as a | |
345 | delta against the previous chunk. To bound lookup time, runs of |
|
355 | delta against the previous chunk. To bound lookup time, runs of | |
346 | deltas are limited to about 2 times the length of the original |
|
356 | deltas are limited to about 2 times the length of the original | |
347 | version data. This makes retrieval of a version proportional to |
|
357 | version data. This makes retrieval of a version proportional to | |
348 | its size, or O(1) relative to the number of revisions. |
|
358 | its size, or O(1) relative to the number of revisions. | |
349 |
|
359 | |||
350 | Both pieces of the revlog are written to in an append-only |
|
360 | Both pieces of the revlog are written to in an append-only | |
351 | fashion, which means we never need to rewrite a file to insert or |
|
361 | fashion, which means we never need to rewrite a file to insert or | |
352 | remove data, and can use some simple techniques to avoid the need |
|
362 | remove data, and can use some simple techniques to avoid the need | |
353 | for locking while reading. |
|
363 | for locking while reading. | |
354 |
|
364 | |||
355 | If checkambig, indexfile is opened with checkambig=True at |
|
365 | If checkambig, indexfile is opened with checkambig=True at | |
356 | writing, to avoid file stat ambiguity. |
|
366 | writing, to avoid file stat ambiguity. | |
357 |
|
367 | |||
358 | If mmaplargeindex is True, and an mmapindexthreshold is set, the |
|
368 | If mmaplargeindex is True, and an mmapindexthreshold is set, the | |
359 | index will be mmapped rather than read if it is larger than the |
|
369 | index will be mmapped rather than read if it is larger than the | |
360 | configured threshold. |
|
370 | configured threshold. | |
361 |
|
371 | |||
362 | If censorable is True, the revlog can have censored revisions. |
|
372 | If censorable is True, the revlog can have censored revisions. | |
363 |
|
373 | |||
364 | If `upperboundcomp` is not None, this is the expected maximal gain from |
|
374 | If `upperboundcomp` is not None, this is the expected maximal gain from | |
365 | compression for the data content. |
|
375 | compression for the data content. | |
366 |
|
376 | |||
367 | `concurrencychecker` is an optional function that receives 3 arguments: a |
|
377 | `concurrencychecker` is an optional function that receives 3 arguments: a | |
368 | file handle, a filename, and an expected position. It should check whether |
|
378 | file handle, a filename, and an expected position. It should check whether | |
369 | the current position in the file handle is valid, and log/warn/fail (by |
|
379 | the current position in the file handle is valid, and log/warn/fail (by | |
370 | raising). |
|
380 | raising). | |
371 |
|
381 | |||
372 | See mercurial/revlogutils/contants.py for details about the content of an |
|
382 | See mercurial/revlogutils/contants.py for details about the content of an | |
373 | index entry. |
|
383 | index entry. | |
374 | """ |
|
384 | """ | |
375 |
|
385 | |||
376 | _flagserrorclass = error.RevlogError |
|
386 | _flagserrorclass = error.RevlogError | |
377 |
|
387 | |||
378 | @staticmethod |
|
388 | @staticmethod | |
379 | def is_inline_index(header_bytes): |
|
389 | def is_inline_index(header_bytes): | |
380 | """Determine if a revlog is inline from the initial bytes of the index""" |
|
390 | """Determine if a revlog is inline from the initial bytes of the index""" | |
381 | header = INDEX_HEADER.unpack(header_bytes)[0] |
|
391 | header = INDEX_HEADER.unpack(header_bytes)[0] | |
382 |
|
392 | |||
383 | _format_flags = header & ~0xFFFF |
|
393 | _format_flags = header & ~0xFFFF | |
384 | _format_version = header & 0xFFFF |
|
394 | _format_version = header & 0xFFFF | |
385 |
|
395 | |||
386 | features = FEATURES_BY_VERSION[_format_version] |
|
396 | features = FEATURES_BY_VERSION[_format_version] | |
387 | return features[b'inline'](_format_flags) |
|
397 | return features[b'inline'](_format_flags) | |
388 |
|
398 | |||
389 | def __init__( |
|
399 | def __init__( | |
390 | self, |
|
400 | self, | |
391 | opener, |
|
401 | opener, | |
392 | target, |
|
402 | target, | |
393 | radix, |
|
403 | radix, | |
394 | postfix=None, # only exist for `tmpcensored` now |
|
404 | postfix=None, # only exist for `tmpcensored` now | |
395 | checkambig=False, |
|
405 | checkambig=False, | |
396 | mmaplargeindex=False, |
|
406 | mmaplargeindex=False, | |
397 | censorable=False, |
|
407 | censorable=False, | |
398 | upperboundcomp=None, |
|
408 | upperboundcomp=None, | |
399 | persistentnodemap=False, |
|
409 | persistentnodemap=False, | |
400 | concurrencychecker=None, |
|
410 | concurrencychecker=None, | |
401 | trypending=False, |
|
411 | trypending=False, | |
402 | try_split=False, |
|
412 | try_split=False, | |
403 | canonical_parent_order=True, |
|
413 | canonical_parent_order=True, | |
404 | ): |
|
414 | ): | |
405 | """ |
|
415 | """ | |
406 | create a revlog object |
|
416 | create a revlog object | |
407 |
|
417 | |||
408 | opener is a function that abstracts the file opening operation |
|
418 | opener is a function that abstracts the file opening operation | |
409 | and can be used to implement COW semantics or the like. |
|
419 | and can be used to implement COW semantics or the like. | |
410 |
|
420 | |||
411 | `target`: a (KIND, ID) tuple that identify the content stored in |
|
421 | `target`: a (KIND, ID) tuple that identify the content stored in | |
412 | this revlog. It help the rest of the code to understand what the revlog |
|
422 | this revlog. It help the rest of the code to understand what the revlog | |
413 | is about without having to resort to heuristic and index filename |
|
423 | is about without having to resort to heuristic and index filename | |
414 | analysis. Note: that this must be reliably be set by normal code, but |
|
424 | analysis. Note: that this must be reliably be set by normal code, but | |
415 | that test, debug, or performance measurement code might not set this to |
|
425 | that test, debug, or performance measurement code might not set this to | |
416 | accurate value. |
|
426 | accurate value. | |
417 | """ |
|
427 | """ | |
418 | self.upperboundcomp = upperboundcomp |
|
428 | self.upperboundcomp = upperboundcomp | |
419 |
|
429 | |||
420 | self.radix = radix |
|
430 | self.radix = radix | |
421 |
|
431 | |||
422 | self._docket_file = None |
|
432 | self._docket_file = None | |
423 | self._indexfile = None |
|
433 | self._indexfile = None | |
424 | self._datafile = None |
|
434 | self._datafile = None | |
425 | self._sidedatafile = None |
|
435 | self._sidedatafile = None | |
426 | self._nodemap_file = None |
|
436 | self._nodemap_file = None | |
427 | self.postfix = postfix |
|
437 | self.postfix = postfix | |
428 | self._trypending = trypending |
|
438 | self._trypending = trypending | |
429 | self._try_split = try_split |
|
439 | self._try_split = try_split | |
430 | self.opener = opener |
|
440 | self.opener = opener | |
431 | if persistentnodemap: |
|
441 | if persistentnodemap: | |
432 | self._nodemap_file = nodemaputil.get_nodemap_file(self) |
|
442 | self._nodemap_file = nodemaputil.get_nodemap_file(self) | |
433 |
|
443 | |||
434 | assert target[0] in ALL_KINDS |
|
444 | assert target[0] in ALL_KINDS | |
435 | assert len(target) == 2 |
|
445 | assert len(target) == 2 | |
436 | self.target = target |
|
446 | self.target = target | |
437 | self.feature_config = FeatureConfig( |
|
447 | self.feature_config = FeatureConfig( | |
438 | censorable=censorable, |
|
448 | censorable=censorable, | |
439 | canonical_parent_order=canonical_parent_order, |
|
449 | canonical_parent_order=canonical_parent_order, | |
440 | ) |
|
450 | ) | |
441 | self.data_config = DataConfig( |
|
451 | self.data_config = DataConfig( | |
442 | check_ambig=checkambig, |
|
452 | check_ambig=checkambig, | |
443 | mmap_large_index=mmaplargeindex, |
|
453 | mmap_large_index=mmaplargeindex, | |
444 | ) |
|
454 | ) | |
445 | self.delta_config = DeltaConfig() |
|
455 | self.delta_config = DeltaConfig() | |
446 |
|
456 | |||
447 | # 3-tuple of (node, rev, text) for a raw revision. |
|
457 | # 3-tuple of (node, rev, text) for a raw revision. | |
448 | self._revisioncache = None |
|
458 | self._revisioncache = None | |
449 | # Maps rev to chain base rev. |
|
459 | # Maps rev to chain base rev. | |
450 | self._chainbasecache = util.lrucachedict(100) |
|
460 | self._chainbasecache = util.lrucachedict(100) | |
451 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. |
|
461 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. | |
452 | self._chunkcache = (0, b'') |
|
462 | self._chunkcache = (0, b'') | |
453 |
|
463 | |||
454 | self.index = None |
|
464 | self.index = None | |
455 | self._docket = None |
|
465 | self._docket = None | |
456 | self._nodemap_docket = None |
|
466 | self._nodemap_docket = None | |
457 | # Mapping of partial identifiers to full nodes. |
|
467 | # Mapping of partial identifiers to full nodes. | |
458 | self._pcache = {} |
|
468 | self._pcache = {} | |
459 |
|
469 | |||
460 | # other optionnals features |
|
470 | # other optionnals features | |
461 |
|
471 | |||
462 | # Make copy of flag processors so each revlog instance can support |
|
472 | # Make copy of flag processors so each revlog instance can support | |
463 | # custom flags. |
|
473 | # custom flags. | |
464 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
474 | self._flagprocessors = dict(flagutil.flagprocessors) | |
465 |
|
475 | |||
466 | # 3-tuple of file handles being used for active writing. |
|
476 | # 3-tuple of file handles being used for active writing. | |
467 | self._writinghandles = None |
|
477 | self._writinghandles = None | |
468 | # prevent nesting of addgroup |
|
478 | # prevent nesting of addgroup | |
469 | self._adding_group = None |
|
479 | self._adding_group = None | |
470 |
|
480 | |||
471 | self._loadindex() |
|
481 | self._loadindex() | |
472 |
|
482 | |||
473 | self._concurrencychecker = concurrencychecker |
|
483 | self._concurrencychecker = concurrencychecker | |
474 |
|
484 | |||
475 | @property |
|
485 | @property | |
476 | def _generaldelta(self): |
|
486 | def _generaldelta(self): | |
477 | """temporary compatibility proxy""" |
|
487 | """temporary compatibility proxy""" | |
478 | return self.delta_config.general_delta |
|
488 | return self.delta_config.general_delta | |
479 |
|
489 | |||
480 | @property |
|
490 | @property | |
481 | def _checkambig(self): |
|
491 | def _checkambig(self): | |
482 | """temporary compatibility proxy""" |
|
492 | """temporary compatibility proxy""" | |
483 | return self.data_config.check_ambig |
|
493 | return self.data_config.check_ambig | |
484 |
|
494 | |||
485 | @property |
|
495 | @property | |
486 | def _mmaplargeindex(self): |
|
496 | def _mmaplargeindex(self): | |
487 | """temporary compatibility proxy""" |
|
497 | """temporary compatibility proxy""" | |
488 | return self.data_config.mmap_large_index |
|
498 | return self.data_config.mmap_large_index | |
489 |
|
499 | |||
490 | @property |
|
500 | @property | |
491 | def _censorable(self): |
|
501 | def _censorable(self): | |
492 | """temporary compatibility proxy""" |
|
502 | """temporary compatibility proxy""" | |
493 | return self.feature_config.censorable |
|
503 | return self.feature_config.censorable | |
494 |
|
504 | |||
495 | @property |
|
505 | @property | |
496 | def _chunkcachesize(self): |
|
506 | def _chunkcachesize(self): | |
497 | """temporary compatibility proxy""" |
|
507 | """temporary compatibility proxy""" | |
498 | return self.data_config.chunk_cache_size |
|
508 | return self.data_config.chunk_cache_size | |
499 |
|
509 | |||
500 | @property |
|
510 | @property | |
501 | def _maxchainlen(self): |
|
511 | def _maxchainlen(self): | |
502 | """temporary compatibility proxy""" |
|
512 | """temporary compatibility proxy""" | |
503 | return self.delta_config.max_chain_len |
|
513 | return self.delta_config.max_chain_len | |
504 |
|
514 | |||
505 | @property |
|
515 | @property | |
506 | def _deltabothparents(self): |
|
516 | def _deltabothparents(self): | |
507 | """temporary compatibility proxy""" |
|
517 | """temporary compatibility proxy""" | |
508 | return self.delta_config.delta_both_parents |
|
518 | return self.delta_config.delta_both_parents | |
509 |
|
519 | |||
510 | @property |
|
520 | @property | |
511 | def _candidate_group_chunk_size(self): |
|
521 | def _candidate_group_chunk_size(self): | |
512 | """temporary compatibility proxy""" |
|
522 | """temporary compatibility proxy""" | |
513 | return self.delta_config.candidate_group_chunk_size |
|
523 | return self.delta_config.candidate_group_chunk_size | |
514 |
|
524 | |||
515 | @property |
|
525 | @property | |
516 | def _debug_delta(self): |
|
526 | def _debug_delta(self): | |
517 | """temporary compatibility proxy""" |
|
527 | """temporary compatibility proxy""" | |
518 | return self.delta_config.debug_delta |
|
528 | return self.delta_config.debug_delta | |
519 |
|
529 | |||
520 | @property |
|
530 | @property | |
521 | def _compengine(self): |
|
531 | def _compengine(self): | |
522 | """temporary compatibility proxy""" |
|
532 | """temporary compatibility proxy""" | |
523 | return self.feature_config.compression_engine |
|
533 | return self.feature_config.compression_engine | |
524 |
|
534 | |||
525 | @property |
|
535 | @property | |
526 | def _compengineopts(self): |
|
536 | def _compengineopts(self): | |
527 | """temporary compatibility proxy""" |
|
537 | """temporary compatibility proxy""" | |
528 | return self.feature_config.compression_engine_options |
|
538 | return self.feature_config.compression_engine_options | |
529 |
|
539 | |||
530 | @property |
|
540 | @property | |
531 | def _maxdeltachainspan(self): |
|
541 | def _maxdeltachainspan(self): | |
532 | """temporary compatibility proxy""" |
|
542 | """temporary compatibility proxy""" | |
533 | return self.delta_config.max_deltachain_span |
|
543 | return self.delta_config.max_deltachain_span | |
534 |
|
544 | |||
535 | @property |
|
545 | @property | |
536 | def _withsparseread(self): |
|
546 | def _withsparseread(self): | |
537 | """temporary compatibility proxy""" |
|
547 | """temporary compatibility proxy""" | |
538 | return self.data_config.with_sparse_read |
|
548 | return self.data_config.with_sparse_read | |
539 |
|
549 | |||
540 | @property |
|
550 | @property | |
541 | def _sparserevlog(self): |
|
551 | def _sparserevlog(self): | |
542 | """temporary compatibility proxy""" |
|
552 | """temporary compatibility proxy""" | |
543 | return self.delta_config.sparse_revlog |
|
553 | return self.delta_config.sparse_revlog | |
544 |
|
554 | |||
545 | @property |
|
555 | @property | |
546 | def hassidedata(self): |
|
556 | def hassidedata(self): | |
547 | """temporary compatibility proxy""" |
|
557 | """temporary compatibility proxy""" | |
548 | return self.feature_config.has_side_data |
|
558 | return self.feature_config.has_side_data | |
549 |
|
559 | |||
550 | @property |
|
560 | @property | |
551 | def _srdensitythreshold(self): |
|
561 | def _srdensitythreshold(self): | |
552 | """temporary compatibility proxy""" |
|
562 | """temporary compatibility proxy""" | |
553 | return self.data_config.sr_density_threshold |
|
563 | return self.data_config.sr_density_threshold | |
554 |
|
564 | |||
555 | @property |
|
565 | @property | |
556 | def _srmingapsize(self): |
|
566 | def _srmingapsize(self): | |
557 | """temporary compatibility proxy""" |
|
567 | """temporary compatibility proxy""" | |
558 | return self.data_config.sr_min_gap_size |
|
568 | return self.data_config.sr_min_gap_size | |
559 |
|
569 | |||
560 | @property |
|
570 | @property | |
561 | def _compute_rank(self): |
|
571 | def _compute_rank(self): | |
562 | """temporary compatibility proxy""" |
|
572 | """temporary compatibility proxy""" | |
563 | return self.feature_config.compute_rank |
|
573 | return self.feature_config.compute_rank | |
564 |
|
574 | |||
565 | @property |
|
575 | @property | |
566 | def canonical_parent_order(self): |
|
576 | def canonical_parent_order(self): | |
567 | """temporary compatibility proxy""" |
|
577 | """temporary compatibility proxy""" | |
568 | return self.feature_config.canonical_parent_order |
|
578 | return self.feature_config.canonical_parent_order | |
569 |
|
579 | |||
570 | @property |
|
580 | @property | |
571 | def _lazydelta(self): |
|
581 | def _lazydelta(self): | |
572 | """temporary compatibility proxy""" |
|
582 | """temporary compatibility proxy""" | |
573 | return self.delta_config.lazy_delta |
|
583 | return self.delta_config.lazy_delta | |
574 |
|
584 | |||
575 | @property |
|
585 | @property | |
576 | def _lazydeltabase(self): |
|
586 | def _lazydeltabase(self): | |
577 | """temporary compatibility proxy""" |
|
587 | """temporary compatibility proxy""" | |
578 | return self.delta_config.lazy_delta_base |
|
588 | return self.delta_config.lazy_delta_base | |
579 |
|
589 | |||
580 | def _init_opts(self): |
|
590 | def _init_opts(self): | |
581 | """process options (from above/config) to setup associated default revlog mode |
|
591 | """process options (from above/config) to setup associated default revlog mode | |
582 |
|
592 | |||
583 | These values might be affected when actually reading on disk information. |
|
593 | These values might be affected when actually reading on disk information. | |
584 |
|
594 | |||
585 | The relevant values are returned for use in _loadindex(). |
|
595 | The relevant values are returned for use in _loadindex(). | |
586 |
|
596 | |||
587 | * newversionflags: |
|
597 | * newversionflags: | |
588 | version header to use if we need to create a new revlog |
|
598 | version header to use if we need to create a new revlog | |
589 |
|
599 | |||
590 | * mmapindexthreshold: |
|
600 | * mmapindexthreshold: | |
591 | minimal index size for start to use mmap |
|
601 | minimal index size for start to use mmap | |
592 |
|
602 | |||
593 | * force_nodemap: |
|
603 | * force_nodemap: | |
594 | force the usage of a "development" version of the nodemap code |
|
604 | force the usage of a "development" version of the nodemap code | |
595 | """ |
|
605 | """ | |
596 | mmapindexthreshold = None |
|
606 | mmapindexthreshold = None | |
597 | opts = self.opener.options |
|
607 | opts = self.opener.options | |
598 |
|
608 | |||
599 | if b'changelogv2' in opts and self.revlog_kind == KIND_CHANGELOG: |
|
609 | if b'changelogv2' in opts and self.revlog_kind == KIND_CHANGELOG: | |
600 | new_header = CHANGELOGV2 |
|
610 | new_header = CHANGELOGV2 | |
601 | compute_rank = opts.get(b'changelogv2.compute-rank', True) |
|
611 | compute_rank = opts.get(b'changelogv2.compute-rank', True) | |
602 | self.feature_config.compute_rank = compute_rank |
|
612 | self.feature_config.compute_rank = compute_rank | |
603 | elif b'revlogv2' in opts: |
|
613 | elif b'revlogv2' in opts: | |
604 | new_header = REVLOGV2 |
|
614 | new_header = REVLOGV2 | |
605 | elif b'revlogv1' in opts: |
|
615 | elif b'revlogv1' in opts: | |
606 | new_header = REVLOGV1 | FLAG_INLINE_DATA |
|
616 | new_header = REVLOGV1 | FLAG_INLINE_DATA | |
607 | if b'generaldelta' in opts: |
|
617 | if b'generaldelta' in opts: | |
608 | new_header |= FLAG_GENERALDELTA |
|
618 | new_header |= FLAG_GENERALDELTA | |
609 | elif b'revlogv0' in self.opener.options: |
|
619 | elif b'revlogv0' in self.opener.options: | |
610 | new_header = REVLOGV0 |
|
620 | new_header = REVLOGV0 | |
611 | else: |
|
621 | else: | |
612 | new_header = REVLOG_DEFAULT_VERSION |
|
622 | new_header = REVLOG_DEFAULT_VERSION | |
613 |
|
623 | |||
614 | if b'chunkcachesize' in opts: |
|
624 | if b'chunkcachesize' in opts: | |
615 | self.data_config.chunk_cache_size = opts[b'chunkcachesize'] |
|
625 | self.data_config.chunk_cache_size = opts[b'chunkcachesize'] | |
616 | if b'maxchainlen' in opts: |
|
626 | if b'maxchainlen' in opts: | |
617 | self.delta_config.max_chain_len = opts[b'maxchainlen'] |
|
627 | self.delta_config.max_chain_len = opts[b'maxchainlen'] | |
618 | if b'deltabothparents' in opts: |
|
628 | if b'deltabothparents' in opts: | |
619 | self.delta_config.delta_both_parents = opts[b'deltabothparents'] |
|
629 | self.delta_config.delta_both_parents = opts[b'deltabothparents'] | |
620 | dps_cgds = opts.get(b'delta-parent-search.candidate-group-chunk-size') |
|
630 | dps_cgds = opts.get(b'delta-parent-search.candidate-group-chunk-size') | |
621 | if dps_cgds: |
|
631 | if dps_cgds: | |
622 | self.delta_config.candidate_group_chunk_size = dps_cgds |
|
632 | self.delta_config.candidate_group_chunk_size = dps_cgds | |
623 | if b'lazydelta' in opts: |
|
633 | if b'lazydelta' in opts: | |
624 | self.delta_config.lazy_delta = bool(opts[b'lazydelta']) |
|
634 | self.delta_config.lazy_delta = bool(opts[b'lazydelta']) | |
625 | if self._lazydelta and b'lazydeltabase' in opts: |
|
635 | if self._lazydelta and b'lazydeltabase' in opts: | |
626 | self.delta_config.lazy_delta_base = opts[b'lazydeltabase'] |
|
636 | self.delta_config.lazy_delta_base = opts[b'lazydeltabase'] | |
627 | if b'debug-delta' in opts: |
|
637 | if b'debug-delta' in opts: | |
628 | self.delta_config.debug_delta = opts[b'debug-delta'] |
|
638 | self.delta_config.debug_delta = opts[b'debug-delta'] | |
629 | if b'compengine' in opts: |
|
639 | if b'compengine' in opts: | |
630 | self.feature_config.compression_engine = opts[b'compengine'] |
|
640 | self.feature_config.compression_engine = opts[b'compengine'] | |
631 | comp_engine_opts = self.feature_config.compression_engine_options |
|
641 | comp_engine_opts = self.feature_config.compression_engine_options | |
632 | if b'zlib.level' in opts: |
|
642 | if b'zlib.level' in opts: | |
633 | comp_engine_opts[b'zlib.level'] = opts[b'zlib.level'] |
|
643 | comp_engine_opts[b'zlib.level'] = opts[b'zlib.level'] | |
634 | if b'zstd.level' in opts: |
|
644 | if b'zstd.level' in opts: | |
635 | comp_engine_opts[b'zstd.level'] = opts[b'zstd.level'] |
|
645 | comp_engine_opts[b'zstd.level'] = opts[b'zstd.level'] | |
636 | if b'maxdeltachainspan' in opts: |
|
646 | if b'maxdeltachainspan' in opts: | |
637 | self.delta_config.max_deltachain_span = opts[b'maxdeltachainspan'] |
|
647 | self.delta_config.max_deltachain_span = opts[b'maxdeltachainspan'] | |
638 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: |
|
648 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: | |
639 | mmapindexthreshold = opts[b'mmapindexthreshold'] |
|
649 | mmapindexthreshold = opts[b'mmapindexthreshold'] | |
640 | self.data_config.mmap_index_threshold = mmapindexthreshold |
|
650 | self.data_config.mmap_index_threshold = mmapindexthreshold | |
641 | if b'sparse-revlog' in opts: |
|
651 | if b'sparse-revlog' in opts: | |
642 | self.delta_config.sparse_revlog = bool(opts[b'sparse-revlog']) |
|
652 | self.delta_config.sparse_revlog = bool(opts[b'sparse-revlog']) | |
643 | if self.delta_config.sparse_revlog: |
|
653 | if self.delta_config.sparse_revlog: | |
644 | # sparse-revlog forces sparse-read |
|
654 | # sparse-revlog forces sparse-read | |
645 | self.data_config.with_sparse_read = True |
|
655 | self.data_config.with_sparse_read = True | |
646 | elif b'with-sparse-read' in opts: |
|
656 | elif b'with-sparse-read' in opts: | |
647 | self.data_config.with_sparse_read = bool(opts[b'with-sparse-read']) |
|
657 | self.data_config.with_sparse_read = bool(opts[b'with-sparse-read']) | |
648 | if b'sparse-read-density-threshold' in opts: |
|
658 | if b'sparse-read-density-threshold' in opts: | |
649 | self.data_config.sr_density_threshold = opts[ |
|
659 | self.data_config.sr_density_threshold = opts[ | |
650 | b'sparse-read-density-threshold' |
|
660 | b'sparse-read-density-threshold' | |
651 | ] |
|
661 | ] | |
652 | if b'sparse-read-min-gap-size' in opts: |
|
662 | if b'sparse-read-min-gap-size' in opts: | |
653 | self.data_config.sr_min_gap_size = opts[b'sparse-read-min-gap-size'] |
|
663 | self.data_config.sr_min_gap_size = opts[b'sparse-read-min-gap-size'] | |
654 | if opts.get(b'enableellipsis'): |
|
664 | if opts.get(b'enableellipsis'): | |
655 | self.feature_config.enable_ellipsis = True |
|
665 | self.feature_config.enable_ellipsis = True | |
656 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor |
|
666 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor | |
657 |
|
667 | |||
658 | # revlog v0 doesn't have flag processors |
|
668 | # revlog v0 doesn't have flag processors | |
659 | for flag, processor in opts.get(b'flagprocessors', {}).items(): |
|
669 | for flag, processor in opts.get(b'flagprocessors', {}).items(): | |
660 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) |
|
670 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) | |
661 |
|
671 | |||
662 | chunk_cache_size = self.data_config.chunk_cache_size |
|
672 | chunk_cache_size = self.data_config.chunk_cache_size | |
663 | if chunk_cache_size <= 0: |
|
673 | if chunk_cache_size <= 0: | |
664 | raise error.RevlogError( |
|
674 | raise error.RevlogError( | |
665 | _(b'revlog chunk cache size %r is not greater than 0') |
|
675 | _(b'revlog chunk cache size %r is not greater than 0') | |
666 | % chunk_cache_size |
|
676 | % chunk_cache_size | |
667 | ) |
|
677 | ) | |
668 | elif chunk_cache_size & (chunk_cache_size - 1): |
|
678 | elif chunk_cache_size & (chunk_cache_size - 1): | |
669 | raise error.RevlogError( |
|
679 | raise error.RevlogError( | |
670 | _(b'revlog chunk cache size %r is not a power of 2') |
|
680 | _(b'revlog chunk cache size %r is not a power of 2') | |
671 | % chunk_cache_size |
|
681 | % chunk_cache_size | |
672 | ) |
|
682 | ) | |
673 | force_nodemap = opts.get(b'devel-force-nodemap', False) |
|
683 | force_nodemap = opts.get(b'devel-force-nodemap', False) | |
674 | return new_header, mmapindexthreshold, force_nodemap |
|
684 | return new_header, mmapindexthreshold, force_nodemap | |
675 |
|
685 | |||
676 | def _get_data(self, filepath, mmap_threshold, size=None): |
|
686 | def _get_data(self, filepath, mmap_threshold, size=None): | |
677 | """return a file content with or without mmap |
|
687 | """return a file content with or without mmap | |
678 |
|
688 | |||
679 | If the file is missing return the empty string""" |
|
689 | If the file is missing return the empty string""" | |
680 | try: |
|
690 | try: | |
681 | with self.opener(filepath) as fp: |
|
691 | with self.opener(filepath) as fp: | |
682 | if mmap_threshold is not None: |
|
692 | if mmap_threshold is not None: | |
683 | file_size = self.opener.fstat(fp).st_size |
|
693 | file_size = self.opener.fstat(fp).st_size | |
684 | if file_size >= mmap_threshold: |
|
694 | if file_size >= mmap_threshold: | |
685 | if size is not None: |
|
695 | if size is not None: | |
686 | # avoid potentiel mmap crash |
|
696 | # avoid potentiel mmap crash | |
687 | size = min(file_size, size) |
|
697 | size = min(file_size, size) | |
688 | # TODO: should .close() to release resources without |
|
698 | # TODO: should .close() to release resources without | |
689 | # relying on Python GC |
|
699 | # relying on Python GC | |
690 | if size is None: |
|
700 | if size is None: | |
691 | return util.buffer(util.mmapread(fp)) |
|
701 | return util.buffer(util.mmapread(fp)) | |
692 | else: |
|
702 | else: | |
693 | return util.buffer(util.mmapread(fp, size)) |
|
703 | return util.buffer(util.mmapread(fp, size)) | |
694 | if size is None: |
|
704 | if size is None: | |
695 | return fp.read() |
|
705 | return fp.read() | |
696 | else: |
|
706 | else: | |
697 | return fp.read(size) |
|
707 | return fp.read(size) | |
698 | except FileNotFoundError: |
|
708 | except FileNotFoundError: | |
699 | return b'' |
|
709 | return b'' | |
700 |
|
710 | |||
701 | def get_streams(self, max_linkrev, force_inline=False): |
|
711 | def get_streams(self, max_linkrev, force_inline=False): | |
702 | """return a list of streams that represent this revlog |
|
712 | """return a list of streams that represent this revlog | |
703 |
|
713 | |||
704 | This is used by stream-clone to do bytes to bytes copies of a repository. |
|
714 | This is used by stream-clone to do bytes to bytes copies of a repository. | |
705 |
|
715 | |||
706 | This streams data for all revisions that refer to a changelog revision up |
|
716 | This streams data for all revisions that refer to a changelog revision up | |
707 | to `max_linkrev`. |
|
717 | to `max_linkrev`. | |
708 |
|
718 | |||
709 | If `force_inline` is set, it enforces that the stream will represent an inline revlog. |
|
719 | If `force_inline` is set, it enforces that the stream will represent an inline revlog. | |
710 |
|
720 | |||
711 | It returns is a list of three-tuple: |
|
721 | It returns is a list of three-tuple: | |
712 |
|
722 | |||
713 | [ |
|
723 | [ | |
714 | (filename, bytes_stream, stream_size), |
|
724 | (filename, bytes_stream, stream_size), | |
715 | β¦ |
|
725 | β¦ | |
716 | ] |
|
726 | ] | |
717 | """ |
|
727 | """ | |
718 | n = len(self) |
|
728 | n = len(self) | |
719 | index = self.index |
|
729 | index = self.index | |
720 | while n > 0: |
|
730 | while n > 0: | |
721 | linkrev = index[n - 1][4] |
|
731 | linkrev = index[n - 1][4] | |
722 | if linkrev < max_linkrev: |
|
732 | if linkrev < max_linkrev: | |
723 | break |
|
733 | break | |
724 | # note: this loop will rarely go through multiple iterations, since |
|
734 | # note: this loop will rarely go through multiple iterations, since | |
725 | # it only traverses commits created during the current streaming |
|
735 | # it only traverses commits created during the current streaming | |
726 | # pull operation. |
|
736 | # pull operation. | |
727 | # |
|
737 | # | |
728 | # If this become a problem, using a binary search should cap the |
|
738 | # If this become a problem, using a binary search should cap the | |
729 | # runtime of this. |
|
739 | # runtime of this. | |
730 | n = n - 1 |
|
740 | n = n - 1 | |
731 | if n == 0: |
|
741 | if n == 0: | |
732 | # no data to send |
|
742 | # no data to send | |
733 | return [] |
|
743 | return [] | |
734 | index_size = n * index.entry_size |
|
744 | index_size = n * index.entry_size | |
735 | data_size = self.end(n - 1) |
|
745 | data_size = self.end(n - 1) | |
736 |
|
746 | |||
737 | # XXX we might have been split (or stripped) since the object |
|
747 | # XXX we might have been split (or stripped) since the object | |
738 | # initialization, We need to close this race too, but having a way to |
|
748 | # initialization, We need to close this race too, but having a way to | |
739 | # pre-open the file we feed to the revlog and never closing them before |
|
749 | # pre-open the file we feed to the revlog and never closing them before | |
740 | # we are done streaming. |
|
750 | # we are done streaming. | |
741 |
|
751 | |||
742 | if self._inline: |
|
752 | if self._inline: | |
743 |
|
753 | |||
744 | def get_stream(): |
|
754 | def get_stream(): | |
745 | with self._indexfp() as fp: |
|
755 | with self._indexfp() as fp: | |
746 | yield None |
|
756 | yield None | |
747 | size = index_size + data_size |
|
757 | size = index_size + data_size | |
748 | if size <= 65536: |
|
758 | if size <= 65536: | |
749 | yield fp.read(size) |
|
759 | yield fp.read(size) | |
750 | else: |
|
760 | else: | |
751 | yield from util.filechunkiter(fp, limit=size) |
|
761 | yield from util.filechunkiter(fp, limit=size) | |
752 |
|
762 | |||
753 | inline_stream = get_stream() |
|
763 | inline_stream = get_stream() | |
754 | next(inline_stream) |
|
764 | next(inline_stream) | |
755 | return [ |
|
765 | return [ | |
756 | (self._indexfile, inline_stream, index_size + data_size), |
|
766 | (self._indexfile, inline_stream, index_size + data_size), | |
757 | ] |
|
767 | ] | |
758 | elif force_inline: |
|
768 | elif force_inline: | |
759 |
|
769 | |||
760 | def get_stream(): |
|
770 | def get_stream(): | |
761 | with self.reading(): |
|
771 | with self.reading(): | |
762 | yield None |
|
772 | yield None | |
763 |
|
773 | |||
764 | for rev in range(n): |
|
774 | for rev in range(n): | |
765 | idx = self.index.entry_binary(rev) |
|
775 | idx = self.index.entry_binary(rev) | |
766 | if rev == 0 and self._docket is None: |
|
776 | if rev == 0 and self._docket is None: | |
767 | # re-inject the inline flag |
|
777 | # re-inject the inline flag | |
768 | header = self._format_flags |
|
778 | header = self._format_flags | |
769 | header |= self._format_version |
|
779 | header |= self._format_version | |
770 | header |= FLAG_INLINE_DATA |
|
780 | header |= FLAG_INLINE_DATA | |
771 | header = self.index.pack_header(header) |
|
781 | header = self.index.pack_header(header) | |
772 | idx = header + idx |
|
782 | idx = header + idx | |
773 | yield idx |
|
783 | yield idx | |
774 | yield self._getsegmentforrevs(rev, rev)[1] |
|
784 | yield self._getsegmentforrevs(rev, rev)[1] | |
775 |
|
785 | |||
776 | inline_stream = get_stream() |
|
786 | inline_stream = get_stream() | |
777 | next(inline_stream) |
|
787 | next(inline_stream) | |
778 | return [ |
|
788 | return [ | |
779 | (self._indexfile, inline_stream, index_size + data_size), |
|
789 | (self._indexfile, inline_stream, index_size + data_size), | |
780 | ] |
|
790 | ] | |
781 | else: |
|
791 | else: | |
782 |
|
792 | |||
783 | def get_index_stream(): |
|
793 | def get_index_stream(): | |
784 | with self._indexfp() as fp: |
|
794 | with self._indexfp() as fp: | |
785 | yield None |
|
795 | yield None | |
786 | if index_size <= 65536: |
|
796 | if index_size <= 65536: | |
787 | yield fp.read(index_size) |
|
797 | yield fp.read(index_size) | |
788 | else: |
|
798 | else: | |
789 | yield from util.filechunkiter(fp, limit=index_size) |
|
799 | yield from util.filechunkiter(fp, limit=index_size) | |
790 |
|
800 | |||
791 | def get_data_stream(): |
|
801 | def get_data_stream(): | |
792 | with self._datafp() as fp: |
|
802 | with self._datafp() as fp: | |
793 | yield None |
|
803 | yield None | |
794 | if data_size <= 65536: |
|
804 | if data_size <= 65536: | |
795 | yield fp.read(data_size) |
|
805 | yield fp.read(data_size) | |
796 | else: |
|
806 | else: | |
797 | yield from util.filechunkiter(fp, limit=data_size) |
|
807 | yield from util.filechunkiter(fp, limit=data_size) | |
798 |
|
808 | |||
799 | index_stream = get_index_stream() |
|
809 | index_stream = get_index_stream() | |
800 | next(index_stream) |
|
810 | next(index_stream) | |
801 | data_stream = get_data_stream() |
|
811 | data_stream = get_data_stream() | |
802 | next(data_stream) |
|
812 | next(data_stream) | |
803 | return [ |
|
813 | return [ | |
804 | (self._datafile, data_stream, data_size), |
|
814 | (self._datafile, data_stream, data_size), | |
805 | (self._indexfile, index_stream, index_size), |
|
815 | (self._indexfile, index_stream, index_size), | |
806 | ] |
|
816 | ] | |
807 |
|
817 | |||
808 | def _loadindex(self, docket=None): |
|
818 | def _loadindex(self, docket=None): | |
809 |
|
819 | |||
810 | new_header, mmapindexthreshold, force_nodemap = self._init_opts() |
|
820 | new_header, mmapindexthreshold, force_nodemap = self._init_opts() | |
811 |
|
821 | |||
812 | if self.postfix is not None: |
|
822 | if self.postfix is not None: | |
813 | entry_point = b'%s.i.%s' % (self.radix, self.postfix) |
|
823 | entry_point = b'%s.i.%s' % (self.radix, self.postfix) | |
814 | elif self._trypending and self.opener.exists(b'%s.i.a' % self.radix): |
|
824 | elif self._trypending and self.opener.exists(b'%s.i.a' % self.radix): | |
815 | entry_point = b'%s.i.a' % self.radix |
|
825 | entry_point = b'%s.i.a' % self.radix | |
816 | elif self._try_split and self.opener.exists(self._split_index_file): |
|
826 | elif self._try_split and self.opener.exists(self._split_index_file): | |
817 | entry_point = self._split_index_file |
|
827 | entry_point = self._split_index_file | |
818 | else: |
|
828 | else: | |
819 | entry_point = b'%s.i' % self.radix |
|
829 | entry_point = b'%s.i' % self.radix | |
820 |
|
830 | |||
821 | if docket is not None: |
|
831 | if docket is not None: | |
822 | self._docket = docket |
|
832 | self._docket = docket | |
823 | self._docket_file = entry_point |
|
833 | self._docket_file = entry_point | |
824 | else: |
|
834 | else: | |
825 | self._initempty = True |
|
835 | self._initempty = True | |
826 | entry_data = self._get_data(entry_point, mmapindexthreshold) |
|
836 | entry_data = self._get_data(entry_point, mmapindexthreshold) | |
827 | if len(entry_data) > 0: |
|
837 | if len(entry_data) > 0: | |
828 | header = INDEX_HEADER.unpack(entry_data[:4])[0] |
|
838 | header = INDEX_HEADER.unpack(entry_data[:4])[0] | |
829 | self._initempty = False |
|
839 | self._initempty = False | |
830 | else: |
|
840 | else: | |
831 | header = new_header |
|
841 | header = new_header | |
832 |
|
842 | |||
833 | self._format_flags = header & ~0xFFFF |
|
843 | self._format_flags = header & ~0xFFFF | |
834 | self._format_version = header & 0xFFFF |
|
844 | self._format_version = header & 0xFFFF | |
835 |
|
845 | |||
836 | supported_flags = SUPPORTED_FLAGS.get(self._format_version) |
|
846 | supported_flags = SUPPORTED_FLAGS.get(self._format_version) | |
837 | if supported_flags is None: |
|
847 | if supported_flags is None: | |
838 | msg = _(b'unknown version (%d) in revlog %s') |
|
848 | msg = _(b'unknown version (%d) in revlog %s') | |
839 | msg %= (self._format_version, self.display_id) |
|
849 | msg %= (self._format_version, self.display_id) | |
840 | raise error.RevlogError(msg) |
|
850 | raise error.RevlogError(msg) | |
841 | elif self._format_flags & ~supported_flags: |
|
851 | elif self._format_flags & ~supported_flags: | |
842 | msg = _(b'unknown flags (%#04x) in version %d revlog %s') |
|
852 | msg = _(b'unknown flags (%#04x) in version %d revlog %s') | |
843 | display_flag = self._format_flags >> 16 |
|
853 | display_flag = self._format_flags >> 16 | |
844 | msg %= (display_flag, self._format_version, self.display_id) |
|
854 | msg %= (display_flag, self._format_version, self.display_id) | |
845 | raise error.RevlogError(msg) |
|
855 | raise error.RevlogError(msg) | |
846 |
|
856 | |||
847 | features = FEATURES_BY_VERSION[self._format_version] |
|
857 | features = FEATURES_BY_VERSION[self._format_version] | |
848 | self._inline = features[b'inline'](self._format_flags) |
|
858 | self._inline = features[b'inline'](self._format_flags) | |
849 | self.delta_config.general_delta = features[b'generaldelta']( |
|
859 | self.delta_config.general_delta = features[b'generaldelta']( | |
850 | self._format_flags |
|
860 | self._format_flags | |
851 | ) |
|
861 | ) | |
852 | self.feature_config.has_side_data = features[b'sidedata'] |
|
862 | self.feature_config.has_side_data = features[b'sidedata'] | |
853 |
|
863 | |||
854 | if not features[b'docket']: |
|
864 | if not features[b'docket']: | |
855 | self._indexfile = entry_point |
|
865 | self._indexfile = entry_point | |
856 | index_data = entry_data |
|
866 | index_data = entry_data | |
857 | else: |
|
867 | else: | |
858 | self._docket_file = entry_point |
|
868 | self._docket_file = entry_point | |
859 | if self._initempty: |
|
869 | if self._initempty: | |
860 | self._docket = docketutil.default_docket(self, header) |
|
870 | self._docket = docketutil.default_docket(self, header) | |
861 | else: |
|
871 | else: | |
862 | self._docket = docketutil.parse_docket( |
|
872 | self._docket = docketutil.parse_docket( | |
863 | self, entry_data, use_pending=self._trypending |
|
873 | self, entry_data, use_pending=self._trypending | |
864 | ) |
|
874 | ) | |
865 |
|
875 | |||
866 | if self._docket is not None: |
|
876 | if self._docket is not None: | |
867 | self._indexfile = self._docket.index_filepath() |
|
877 | self._indexfile = self._docket.index_filepath() | |
868 | index_data = b'' |
|
878 | index_data = b'' | |
869 | index_size = self._docket.index_end |
|
879 | index_size = self._docket.index_end | |
870 | if index_size > 0: |
|
880 | if index_size > 0: | |
871 | index_data = self._get_data( |
|
881 | index_data = self._get_data( | |
872 | self._indexfile, mmapindexthreshold, size=index_size |
|
882 | self._indexfile, mmapindexthreshold, size=index_size | |
873 | ) |
|
883 | ) | |
874 | if len(index_data) < index_size: |
|
884 | if len(index_data) < index_size: | |
875 | msg = _(b'too few index data for %s: got %d, expected %d') |
|
885 | msg = _(b'too few index data for %s: got %d, expected %d') | |
876 | msg %= (self.display_id, len(index_data), index_size) |
|
886 | msg %= (self.display_id, len(index_data), index_size) | |
877 | raise error.RevlogError(msg) |
|
887 | raise error.RevlogError(msg) | |
878 |
|
888 | |||
879 | self._inline = False |
|
889 | self._inline = False | |
880 | # generaldelta implied by version 2 revlogs. |
|
890 | # generaldelta implied by version 2 revlogs. | |
881 | self.delta_config.general_delta = True |
|
891 | self.delta_config.general_delta = True | |
882 | # the logic for persistent nodemap will be dealt with within the |
|
892 | # the logic for persistent nodemap will be dealt with within the | |
883 | # main docket, so disable it for now. |
|
893 | # main docket, so disable it for now. | |
884 | self._nodemap_file = None |
|
894 | self._nodemap_file = None | |
885 |
|
895 | |||
886 | if self._docket is not None: |
|
896 | if self._docket is not None: | |
887 | self._datafile = self._docket.data_filepath() |
|
897 | self._datafile = self._docket.data_filepath() | |
888 | self._sidedatafile = self._docket.sidedata_filepath() |
|
898 | self._sidedatafile = self._docket.sidedata_filepath() | |
889 | elif self.postfix is None: |
|
899 | elif self.postfix is None: | |
890 | self._datafile = b'%s.d' % self.radix |
|
900 | self._datafile = b'%s.d' % self.radix | |
891 | else: |
|
901 | else: | |
892 | self._datafile = b'%s.d.%s' % (self.radix, self.postfix) |
|
902 | self._datafile = b'%s.d.%s' % (self.radix, self.postfix) | |
893 |
|
903 | |||
894 | self.nodeconstants = sha1nodeconstants |
|
904 | self.nodeconstants = sha1nodeconstants | |
895 | self.nullid = self.nodeconstants.nullid |
|
905 | self.nullid = self.nodeconstants.nullid | |
896 |
|
906 | |||
897 | # sparse-revlog can't be on without general-delta (issue6056) |
|
907 | # sparse-revlog can't be on without general-delta (issue6056) | |
898 | if not self._generaldelta: |
|
908 | if not self._generaldelta: | |
899 | self.delta_config.sparse_revlog = False |
|
909 | self.delta_config.sparse_revlog = False | |
900 |
|
910 | |||
901 | self._storedeltachains = True |
|
911 | self._storedeltachains = True | |
902 |
|
912 | |||
903 | devel_nodemap = ( |
|
913 | devel_nodemap = ( | |
904 | self._nodemap_file |
|
914 | self._nodemap_file | |
905 | and force_nodemap |
|
915 | and force_nodemap | |
906 | and parse_index_v1_nodemap is not None |
|
916 | and parse_index_v1_nodemap is not None | |
907 | ) |
|
917 | ) | |
908 |
|
918 | |||
909 | use_rust_index = False |
|
919 | use_rust_index = False | |
910 | if rustrevlog is not None: |
|
920 | if rustrevlog is not None: | |
911 | if self._nodemap_file is not None: |
|
921 | if self._nodemap_file is not None: | |
912 | use_rust_index = True |
|
922 | use_rust_index = True | |
913 | else: |
|
923 | else: | |
914 | use_rust_index = self.opener.options.get(b'rust.index') |
|
924 | use_rust_index = self.opener.options.get(b'rust.index') | |
915 |
|
925 | |||
916 | self._parse_index = parse_index_v1 |
|
926 | self._parse_index = parse_index_v1 | |
917 | if self._format_version == REVLOGV0: |
|
927 | if self._format_version == REVLOGV0: | |
918 | self._parse_index = revlogv0.parse_index_v0 |
|
928 | self._parse_index = revlogv0.parse_index_v0 | |
919 | elif self._format_version == REVLOGV2: |
|
929 | elif self._format_version == REVLOGV2: | |
920 | self._parse_index = parse_index_v2 |
|
930 | self._parse_index = parse_index_v2 | |
921 | elif self._format_version == CHANGELOGV2: |
|
931 | elif self._format_version == CHANGELOGV2: | |
922 | self._parse_index = parse_index_cl_v2 |
|
932 | self._parse_index = parse_index_cl_v2 | |
923 | elif devel_nodemap: |
|
933 | elif devel_nodemap: | |
924 | self._parse_index = parse_index_v1_nodemap |
|
934 | self._parse_index = parse_index_v1_nodemap | |
925 | elif use_rust_index: |
|
935 | elif use_rust_index: | |
926 | self._parse_index = parse_index_v1_mixed |
|
936 | self._parse_index = parse_index_v1_mixed | |
927 | try: |
|
937 | try: | |
928 | d = self._parse_index(index_data, self._inline) |
|
938 | d = self._parse_index(index_data, self._inline) | |
929 | index, chunkcache = d |
|
939 | index, chunkcache = d | |
930 | use_nodemap = ( |
|
940 | use_nodemap = ( | |
931 | not self._inline |
|
941 | not self._inline | |
932 | and self._nodemap_file is not None |
|
942 | and self._nodemap_file is not None | |
933 | and hasattr(index, 'update_nodemap_data') |
|
943 | and hasattr(index, 'update_nodemap_data') | |
934 | ) |
|
944 | ) | |
935 | if use_nodemap: |
|
945 | if use_nodemap: | |
936 | nodemap_data = nodemaputil.persisted_data(self) |
|
946 | nodemap_data = nodemaputil.persisted_data(self) | |
937 | if nodemap_data is not None: |
|
947 | if nodemap_data is not None: | |
938 | docket = nodemap_data[0] |
|
948 | docket = nodemap_data[0] | |
939 | if ( |
|
949 | if ( | |
940 | len(d[0]) > docket.tip_rev |
|
950 | len(d[0]) > docket.tip_rev | |
941 | and d[0][docket.tip_rev][7] == docket.tip_node |
|
951 | and d[0][docket.tip_rev][7] == docket.tip_node | |
942 | ): |
|
952 | ): | |
943 | # no changelog tampering |
|
953 | # no changelog tampering | |
944 | self._nodemap_docket = docket |
|
954 | self._nodemap_docket = docket | |
945 | index.update_nodemap_data(*nodemap_data) |
|
955 | index.update_nodemap_data(*nodemap_data) | |
946 | except (ValueError, IndexError): |
|
956 | except (ValueError, IndexError): | |
947 | raise error.RevlogError( |
|
957 | raise error.RevlogError( | |
948 | _(b"index %s is corrupted") % self.display_id |
|
958 | _(b"index %s is corrupted") % self.display_id | |
949 | ) |
|
959 | ) | |
950 | self.index = index |
|
960 | self.index = index | |
951 | self._segmentfile = randomaccessfile.randomaccessfile( |
|
961 | self._segmentfile = randomaccessfile.randomaccessfile( | |
952 | self.opener, |
|
962 | self.opener, | |
953 | (self._indexfile if self._inline else self._datafile), |
|
963 | (self._indexfile if self._inline else self._datafile), | |
954 | self._chunkcachesize, |
|
964 | self._chunkcachesize, | |
955 | chunkcache, |
|
965 | chunkcache, | |
956 | ) |
|
966 | ) | |
957 | self._segmentfile_sidedata = randomaccessfile.randomaccessfile( |
|
967 | self._segmentfile_sidedata = randomaccessfile.randomaccessfile( | |
958 | self.opener, |
|
968 | self.opener, | |
959 | self._sidedatafile, |
|
969 | self._sidedatafile, | |
960 | self._chunkcachesize, |
|
970 | self._chunkcachesize, | |
961 | ) |
|
971 | ) | |
962 | # revnum -> (chain-length, sum-delta-length) |
|
972 | # revnum -> (chain-length, sum-delta-length) | |
963 | self._chaininfocache = util.lrucachedict(500) |
|
973 | self._chaininfocache = util.lrucachedict(500) | |
964 | # revlog header -> revlog compressor |
|
974 | # revlog header -> revlog compressor | |
965 | self._decompressors = {} |
|
975 | self._decompressors = {} | |
966 |
|
976 | |||
967 | def get_revlog(self): |
|
977 | def get_revlog(self): | |
968 | """simple function to mirror API of other not-really-revlog API""" |
|
978 | """simple function to mirror API of other not-really-revlog API""" | |
969 | return self |
|
979 | return self | |
970 |
|
980 | |||
971 | @util.propertycache |
|
981 | @util.propertycache | |
972 | def revlog_kind(self): |
|
982 | def revlog_kind(self): | |
973 | return self.target[0] |
|
983 | return self.target[0] | |
974 |
|
984 | |||
975 | @util.propertycache |
|
985 | @util.propertycache | |
976 | def display_id(self): |
|
986 | def display_id(self): | |
977 | """The public facing "ID" of the revlog that we use in message""" |
|
987 | """The public facing "ID" of the revlog that we use in message""" | |
978 | if self.revlog_kind == KIND_FILELOG: |
|
988 | if self.revlog_kind == KIND_FILELOG: | |
979 | # Reference the file without the "data/" prefix, so it is familiar |
|
989 | # Reference the file without the "data/" prefix, so it is familiar | |
980 | # to the user. |
|
990 | # to the user. | |
981 | return self.target[1] |
|
991 | return self.target[1] | |
982 | else: |
|
992 | else: | |
983 | return self.radix |
|
993 | return self.radix | |
984 |
|
994 | |||
985 | def _get_decompressor(self, t): |
|
995 | def _get_decompressor(self, t): | |
986 | try: |
|
996 | try: | |
987 | compressor = self._decompressors[t] |
|
997 | compressor = self._decompressors[t] | |
988 | except KeyError: |
|
998 | except KeyError: | |
989 | try: |
|
999 | try: | |
990 | engine = util.compengines.forrevlogheader(t) |
|
1000 | engine = util.compengines.forrevlogheader(t) | |
991 | compressor = engine.revlogcompressor(self._compengineopts) |
|
1001 | compressor = engine.revlogcompressor(self._compengineopts) | |
992 | self._decompressors[t] = compressor |
|
1002 | self._decompressors[t] = compressor | |
993 | except KeyError: |
|
1003 | except KeyError: | |
994 | raise error.RevlogError( |
|
1004 | raise error.RevlogError( | |
995 | _(b'unknown compression type %s') % binascii.hexlify(t) |
|
1005 | _(b'unknown compression type %s') % binascii.hexlify(t) | |
996 | ) |
|
1006 | ) | |
997 | return compressor |
|
1007 | return compressor | |
998 |
|
1008 | |||
999 | @util.propertycache |
|
1009 | @util.propertycache | |
1000 | def _compressor(self): |
|
1010 | def _compressor(self): | |
1001 | engine = util.compengines[self._compengine] |
|
1011 | engine = util.compengines[self._compengine] | |
1002 | return engine.revlogcompressor(self._compengineopts) |
|
1012 | return engine.revlogcompressor(self._compengineopts) | |
1003 |
|
1013 | |||
1004 | @util.propertycache |
|
1014 | @util.propertycache | |
1005 | def _decompressor(self): |
|
1015 | def _decompressor(self): | |
1006 | """the default decompressor""" |
|
1016 | """the default decompressor""" | |
1007 | if self._docket is None: |
|
1017 | if self._docket is None: | |
1008 | return None |
|
1018 | return None | |
1009 | t = self._docket.default_compression_header |
|
1019 | t = self._docket.default_compression_header | |
1010 | c = self._get_decompressor(t) |
|
1020 | c = self._get_decompressor(t) | |
1011 | return c.decompress |
|
1021 | return c.decompress | |
1012 |
|
1022 | |||
1013 | def _indexfp(self): |
|
1023 | def _indexfp(self): | |
1014 | """file object for the revlog's index file""" |
|
1024 | """file object for the revlog's index file""" | |
1015 | return self.opener(self._indexfile, mode=b"r") |
|
1025 | return self.opener(self._indexfile, mode=b"r") | |
1016 |
|
1026 | |||
1017 | def __index_write_fp(self): |
|
1027 | def __index_write_fp(self): | |
1018 | # You should not use this directly and use `_writing` instead |
|
1028 | # You should not use this directly and use `_writing` instead | |
1019 | try: |
|
1029 | try: | |
1020 | f = self.opener( |
|
1030 | f = self.opener( | |
1021 | self._indexfile, mode=b"r+", checkambig=self._checkambig |
|
1031 | self._indexfile, mode=b"r+", checkambig=self._checkambig | |
1022 | ) |
|
1032 | ) | |
1023 | if self._docket is None: |
|
1033 | if self._docket is None: | |
1024 | f.seek(0, os.SEEK_END) |
|
1034 | f.seek(0, os.SEEK_END) | |
1025 | else: |
|
1035 | else: | |
1026 | f.seek(self._docket.index_end, os.SEEK_SET) |
|
1036 | f.seek(self._docket.index_end, os.SEEK_SET) | |
1027 | return f |
|
1037 | return f | |
1028 | except FileNotFoundError: |
|
1038 | except FileNotFoundError: | |
1029 | return self.opener( |
|
1039 | return self.opener( | |
1030 | self._indexfile, mode=b"w+", checkambig=self._checkambig |
|
1040 | self._indexfile, mode=b"w+", checkambig=self._checkambig | |
1031 | ) |
|
1041 | ) | |
1032 |
|
1042 | |||
1033 | def __index_new_fp(self): |
|
1043 | def __index_new_fp(self): | |
1034 | # You should not use this unless you are upgrading from inline revlog |
|
1044 | # You should not use this unless you are upgrading from inline revlog | |
1035 | return self.opener( |
|
1045 | return self.opener( | |
1036 | self._indexfile, |
|
1046 | self._indexfile, | |
1037 | mode=b"w", |
|
1047 | mode=b"w", | |
1038 | checkambig=self._checkambig, |
|
1048 | checkambig=self._checkambig, | |
1039 | atomictemp=True, |
|
1049 | atomictemp=True, | |
1040 | ) |
|
1050 | ) | |
1041 |
|
1051 | |||
1042 | def _datafp(self, mode=b'r'): |
|
1052 | def _datafp(self, mode=b'r'): | |
1043 | """file object for the revlog's data file""" |
|
1053 | """file object for the revlog's data file""" | |
1044 | return self.opener(self._datafile, mode=mode) |
|
1054 | return self.opener(self._datafile, mode=mode) | |
1045 |
|
1055 | |||
1046 | @contextlib.contextmanager |
|
1056 | @contextlib.contextmanager | |
1047 | def _sidedatareadfp(self): |
|
1057 | def _sidedatareadfp(self): | |
1048 | """file object suitable to read sidedata""" |
|
1058 | """file object suitable to read sidedata""" | |
1049 | if self._writinghandles: |
|
1059 | if self._writinghandles: | |
1050 | yield self._writinghandles[2] |
|
1060 | yield self._writinghandles[2] | |
1051 | else: |
|
1061 | else: | |
1052 | with self.opener(self._sidedatafile) as fp: |
|
1062 | with self.opener(self._sidedatafile) as fp: | |
1053 | yield fp |
|
1063 | yield fp | |
1054 |
|
1064 | |||
1055 | def tiprev(self): |
|
1065 | def tiprev(self): | |
1056 | return len(self.index) - 1 |
|
1066 | return len(self.index) - 1 | |
1057 |
|
1067 | |||
1058 | def tip(self): |
|
1068 | def tip(self): | |
1059 | return self.node(self.tiprev()) |
|
1069 | return self.node(self.tiprev()) | |
1060 |
|
1070 | |||
1061 | def __contains__(self, rev): |
|
1071 | def __contains__(self, rev): | |
1062 | return 0 <= rev < len(self) |
|
1072 | return 0 <= rev < len(self) | |
1063 |
|
1073 | |||
1064 | def __len__(self): |
|
1074 | def __len__(self): | |
1065 | return len(self.index) |
|
1075 | return len(self.index) | |
1066 |
|
1076 | |||
1067 | def __iter__(self): |
|
1077 | def __iter__(self): | |
1068 | return iter(range(len(self))) |
|
1078 | return iter(range(len(self))) | |
1069 |
|
1079 | |||
1070 | def revs(self, start=0, stop=None): |
|
1080 | def revs(self, start=0, stop=None): | |
1071 | """iterate over all rev in this revlog (from start to stop)""" |
|
1081 | """iterate over all rev in this revlog (from start to stop)""" | |
1072 | return storageutil.iterrevs(len(self), start=start, stop=stop) |
|
1082 | return storageutil.iterrevs(len(self), start=start, stop=stop) | |
1073 |
|
1083 | |||
1074 | def hasnode(self, node): |
|
1084 | def hasnode(self, node): | |
1075 | try: |
|
1085 | try: | |
1076 | self.rev(node) |
|
1086 | self.rev(node) | |
1077 | return True |
|
1087 | return True | |
1078 | except KeyError: |
|
1088 | except KeyError: | |
1079 | return False |
|
1089 | return False | |
1080 |
|
1090 | |||
1081 | def _candelta(self, baserev, rev): |
|
1091 | def _candelta(self, baserev, rev): | |
1082 | """whether two revisions (baserev, rev) can be delta-ed or not""" |
|
1092 | """whether two revisions (baserev, rev) can be delta-ed or not""" | |
1083 | # Disable delta if either rev requires a content-changing flag |
|
1093 | # Disable delta if either rev requires a content-changing flag | |
1084 | # processor (ex. LFS). This is because such flag processor can alter |
|
1094 | # processor (ex. LFS). This is because such flag processor can alter | |
1085 | # the rawtext content that the delta will be based on, and two clients |
|
1095 | # the rawtext content that the delta will be based on, and two clients | |
1086 | # could have a same revlog node with different flags (i.e. different |
|
1096 | # could have a same revlog node with different flags (i.e. different | |
1087 | # rawtext contents) and the delta could be incompatible. |
|
1097 | # rawtext contents) and the delta could be incompatible. | |
1088 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
1098 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( | |
1089 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS |
|
1099 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS | |
1090 | ): |
|
1100 | ): | |
1091 | return False |
|
1101 | return False | |
1092 | return True |
|
1102 | return True | |
1093 |
|
1103 | |||
1094 | def update_caches(self, transaction): |
|
1104 | def update_caches(self, transaction): | |
1095 | """update on disk cache |
|
1105 | """update on disk cache | |
1096 |
|
1106 | |||
1097 | If a transaction is passed, the update may be delayed to transaction |
|
1107 | If a transaction is passed, the update may be delayed to transaction | |
1098 | commit.""" |
|
1108 | commit.""" | |
1099 | if self._nodemap_file is not None: |
|
1109 | if self._nodemap_file is not None: | |
1100 | if transaction is None: |
|
1110 | if transaction is None: | |
1101 | nodemaputil.update_persistent_nodemap(self) |
|
1111 | nodemaputil.update_persistent_nodemap(self) | |
1102 | else: |
|
1112 | else: | |
1103 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
1113 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
1104 |
|
1114 | |||
1105 | def clearcaches(self): |
|
1115 | def clearcaches(self): | |
1106 | """Clear in-memory caches""" |
|
1116 | """Clear in-memory caches""" | |
1107 | self._revisioncache = None |
|
1117 | self._revisioncache = None | |
1108 | self._chainbasecache.clear() |
|
1118 | self._chainbasecache.clear() | |
1109 | self._segmentfile.clear_cache() |
|
1119 | self._segmentfile.clear_cache() | |
1110 | self._segmentfile_sidedata.clear_cache() |
|
1120 | self._segmentfile_sidedata.clear_cache() | |
1111 | self._pcache = {} |
|
1121 | self._pcache = {} | |
1112 | self._nodemap_docket = None |
|
1122 | self._nodemap_docket = None | |
1113 | self.index.clearcaches() |
|
1123 | self.index.clearcaches() | |
1114 | # The python code is the one responsible for validating the docket, we |
|
1124 | # The python code is the one responsible for validating the docket, we | |
1115 | # end up having to refresh it here. |
|
1125 | # end up having to refresh it here. | |
1116 | use_nodemap = ( |
|
1126 | use_nodemap = ( | |
1117 | not self._inline |
|
1127 | not self._inline | |
1118 | and self._nodemap_file is not None |
|
1128 | and self._nodemap_file is not None | |
1119 | and hasattr(self.index, 'update_nodemap_data') |
|
1129 | and hasattr(self.index, 'update_nodemap_data') | |
1120 | ) |
|
1130 | ) | |
1121 | if use_nodemap: |
|
1131 | if use_nodemap: | |
1122 | nodemap_data = nodemaputil.persisted_data(self) |
|
1132 | nodemap_data = nodemaputil.persisted_data(self) | |
1123 | if nodemap_data is not None: |
|
1133 | if nodemap_data is not None: | |
1124 | self._nodemap_docket = nodemap_data[0] |
|
1134 | self._nodemap_docket = nodemap_data[0] | |
1125 | self.index.update_nodemap_data(*nodemap_data) |
|
1135 | self.index.update_nodemap_data(*nodemap_data) | |
1126 |
|
1136 | |||
1127 | def rev(self, node): |
|
1137 | def rev(self, node): | |
1128 | """return the revision number associated with a <nodeid>""" |
|
1138 | """return the revision number associated with a <nodeid>""" | |
1129 | try: |
|
1139 | try: | |
1130 | return self.index.rev(node) |
|
1140 | return self.index.rev(node) | |
1131 | except TypeError: |
|
1141 | except TypeError: | |
1132 | raise |
|
1142 | raise | |
1133 | except error.RevlogError: |
|
1143 | except error.RevlogError: | |
1134 | # parsers.c radix tree lookup failed |
|
1144 | # parsers.c radix tree lookup failed | |
1135 | if ( |
|
1145 | if ( | |
1136 | node == self.nodeconstants.wdirid |
|
1146 | node == self.nodeconstants.wdirid | |
1137 | or node in self.nodeconstants.wdirfilenodeids |
|
1147 | or node in self.nodeconstants.wdirfilenodeids | |
1138 | ): |
|
1148 | ): | |
1139 | raise error.WdirUnsupported |
|
1149 | raise error.WdirUnsupported | |
1140 | raise error.LookupError(node, self.display_id, _(b'no node')) |
|
1150 | raise error.LookupError(node, self.display_id, _(b'no node')) | |
1141 |
|
1151 | |||
1142 | # Accessors for index entries. |
|
1152 | # Accessors for index entries. | |
1143 |
|
1153 | |||
1144 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes |
|
1154 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes | |
1145 | # are flags. |
|
1155 | # are flags. | |
1146 | def start(self, rev): |
|
1156 | def start(self, rev): | |
1147 | return int(self.index[rev][0] >> 16) |
|
1157 | return int(self.index[rev][0] >> 16) | |
1148 |
|
1158 | |||
1149 | def sidedata_cut_off(self, rev): |
|
1159 | def sidedata_cut_off(self, rev): | |
1150 | sd_cut_off = self.index[rev][8] |
|
1160 | sd_cut_off = self.index[rev][8] | |
1151 | if sd_cut_off != 0: |
|
1161 | if sd_cut_off != 0: | |
1152 | return sd_cut_off |
|
1162 | return sd_cut_off | |
1153 | # This is some annoying dance, because entries without sidedata |
|
1163 | # This is some annoying dance, because entries without sidedata | |
1154 | # currently use 0 as their ofsset. (instead of previous-offset + |
|
1164 | # currently use 0 as their ofsset. (instead of previous-offset + | |
1155 | # previous-size) |
|
1165 | # previous-size) | |
1156 | # |
|
1166 | # | |
1157 | # We should reconsider this sidedata β 0 sidata_offset policy. |
|
1167 | # We should reconsider this sidedata β 0 sidata_offset policy. | |
1158 | # In the meantime, we need this. |
|
1168 | # In the meantime, we need this. | |
1159 | while 0 <= rev: |
|
1169 | while 0 <= rev: | |
1160 | e = self.index[rev] |
|
1170 | e = self.index[rev] | |
1161 | if e[9] != 0: |
|
1171 | if e[9] != 0: | |
1162 | return e[8] + e[9] |
|
1172 | return e[8] + e[9] | |
1163 | rev -= 1 |
|
1173 | rev -= 1 | |
1164 | return 0 |
|
1174 | return 0 | |
1165 |
|
1175 | |||
1166 | def flags(self, rev): |
|
1176 | def flags(self, rev): | |
1167 | return self.index[rev][0] & 0xFFFF |
|
1177 | return self.index[rev][0] & 0xFFFF | |
1168 |
|
1178 | |||
1169 | def length(self, rev): |
|
1179 | def length(self, rev): | |
1170 | return self.index[rev][1] |
|
1180 | return self.index[rev][1] | |
1171 |
|
1181 | |||
1172 | def sidedata_length(self, rev): |
|
1182 | def sidedata_length(self, rev): | |
1173 | if not self.hassidedata: |
|
1183 | if not self.hassidedata: | |
1174 | return 0 |
|
1184 | return 0 | |
1175 | return self.index[rev][9] |
|
1185 | return self.index[rev][9] | |
1176 |
|
1186 | |||
1177 | def rawsize(self, rev): |
|
1187 | def rawsize(self, rev): | |
1178 | """return the length of the uncompressed text for a given revision""" |
|
1188 | """return the length of the uncompressed text for a given revision""" | |
1179 | l = self.index[rev][2] |
|
1189 | l = self.index[rev][2] | |
1180 | if l >= 0: |
|
1190 | if l >= 0: | |
1181 | return l |
|
1191 | return l | |
1182 |
|
1192 | |||
1183 | t = self.rawdata(rev) |
|
1193 | t = self.rawdata(rev) | |
1184 | return len(t) |
|
1194 | return len(t) | |
1185 |
|
1195 | |||
1186 | def size(self, rev): |
|
1196 | def size(self, rev): | |
1187 | """length of non-raw text (processed by a "read" flag processor)""" |
|
1197 | """length of non-raw text (processed by a "read" flag processor)""" | |
1188 | # fast path: if no "read" flag processor could change the content, |
|
1198 | # fast path: if no "read" flag processor could change the content, | |
1189 | # size is rawsize. note: ELLIPSIS is known to not change the content. |
|
1199 | # size is rawsize. note: ELLIPSIS is known to not change the content. | |
1190 | flags = self.flags(rev) |
|
1200 | flags = self.flags(rev) | |
1191 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: |
|
1201 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: | |
1192 | return self.rawsize(rev) |
|
1202 | return self.rawsize(rev) | |
1193 |
|
1203 | |||
1194 | return len(self.revision(rev)) |
|
1204 | return len(self.revision(rev)) | |
1195 |
|
1205 | |||
1196 | def fast_rank(self, rev): |
|
1206 | def fast_rank(self, rev): | |
1197 | """Return the rank of a revision if already known, or None otherwise. |
|
1207 | """Return the rank of a revision if already known, or None otherwise. | |
1198 |
|
1208 | |||
1199 | The rank of a revision is the size of the sub-graph it defines as a |
|
1209 | The rank of a revision is the size of the sub-graph it defines as a | |
1200 | head. Equivalently, the rank of a revision `r` is the size of the set |
|
1210 | head. Equivalently, the rank of a revision `r` is the size of the set | |
1201 | `ancestors(r)`, `r` included. |
|
1211 | `ancestors(r)`, `r` included. | |
1202 |
|
1212 | |||
1203 | This method returns the rank retrieved from the revlog in constant |
|
1213 | This method returns the rank retrieved from the revlog in constant | |
1204 | time. It makes no attempt at computing unknown values for versions of |
|
1214 | time. It makes no attempt at computing unknown values for versions of | |
1205 | the revlog which do not persist the rank. |
|
1215 | the revlog which do not persist the rank. | |
1206 | """ |
|
1216 | """ | |
1207 | rank = self.index[rev][ENTRY_RANK] |
|
1217 | rank = self.index[rev][ENTRY_RANK] | |
1208 | if self._format_version != CHANGELOGV2 or rank == RANK_UNKNOWN: |
|
1218 | if self._format_version != CHANGELOGV2 or rank == RANK_UNKNOWN: | |
1209 | return None |
|
1219 | return None | |
1210 | if rev == nullrev: |
|
1220 | if rev == nullrev: | |
1211 | return 0 # convention |
|
1221 | return 0 # convention | |
1212 | return rank |
|
1222 | return rank | |
1213 |
|
1223 | |||
1214 | def chainbase(self, rev): |
|
1224 | def chainbase(self, rev): | |
1215 | base = self._chainbasecache.get(rev) |
|
1225 | base = self._chainbasecache.get(rev) | |
1216 | if base is not None: |
|
1226 | if base is not None: | |
1217 | return base |
|
1227 | return base | |
1218 |
|
1228 | |||
1219 | index = self.index |
|
1229 | index = self.index | |
1220 | iterrev = rev |
|
1230 | iterrev = rev | |
1221 | base = index[iterrev][3] |
|
1231 | base = index[iterrev][3] | |
1222 | while base != iterrev: |
|
1232 | while base != iterrev: | |
1223 | iterrev = base |
|
1233 | iterrev = base | |
1224 | base = index[iterrev][3] |
|
1234 | base = index[iterrev][3] | |
1225 |
|
1235 | |||
1226 | self._chainbasecache[rev] = base |
|
1236 | self._chainbasecache[rev] = base | |
1227 | return base |
|
1237 | return base | |
1228 |
|
1238 | |||
1229 | def linkrev(self, rev): |
|
1239 | def linkrev(self, rev): | |
1230 | return self.index[rev][4] |
|
1240 | return self.index[rev][4] | |
1231 |
|
1241 | |||
1232 | def parentrevs(self, rev): |
|
1242 | def parentrevs(self, rev): | |
1233 | try: |
|
1243 | try: | |
1234 | entry = self.index[rev] |
|
1244 | entry = self.index[rev] | |
1235 | except IndexError: |
|
1245 | except IndexError: | |
1236 | if rev == wdirrev: |
|
1246 | if rev == wdirrev: | |
1237 | raise error.WdirUnsupported |
|
1247 | raise error.WdirUnsupported | |
1238 | raise |
|
1248 | raise | |
1239 |
|
1249 | |||
1240 | if self.canonical_parent_order and entry[5] == nullrev: |
|
1250 | if self.canonical_parent_order and entry[5] == nullrev: | |
1241 | return entry[6], entry[5] |
|
1251 | return entry[6], entry[5] | |
1242 | else: |
|
1252 | else: | |
1243 | return entry[5], entry[6] |
|
1253 | return entry[5], entry[6] | |
1244 |
|
1254 | |||
1245 | # fast parentrevs(rev) where rev isn't filtered |
|
1255 | # fast parentrevs(rev) where rev isn't filtered | |
1246 | _uncheckedparentrevs = parentrevs |
|
1256 | _uncheckedparentrevs = parentrevs | |
1247 |
|
1257 | |||
1248 | def node(self, rev): |
|
1258 | def node(self, rev): | |
1249 | try: |
|
1259 | try: | |
1250 | return self.index[rev][7] |
|
1260 | return self.index[rev][7] | |
1251 | except IndexError: |
|
1261 | except IndexError: | |
1252 | if rev == wdirrev: |
|
1262 | if rev == wdirrev: | |
1253 | raise error.WdirUnsupported |
|
1263 | raise error.WdirUnsupported | |
1254 | raise |
|
1264 | raise | |
1255 |
|
1265 | |||
1256 | # Derived from index values. |
|
1266 | # Derived from index values. | |
1257 |
|
1267 | |||
1258 | def end(self, rev): |
|
1268 | def end(self, rev): | |
1259 | return self.start(rev) + self.length(rev) |
|
1269 | return self.start(rev) + self.length(rev) | |
1260 |
|
1270 | |||
1261 | def parents(self, node): |
|
1271 | def parents(self, node): | |
1262 | i = self.index |
|
1272 | i = self.index | |
1263 | d = i[self.rev(node)] |
|
1273 | d = i[self.rev(node)] | |
1264 | # inline node() to avoid function call overhead |
|
1274 | # inline node() to avoid function call overhead | |
1265 | if self.canonical_parent_order and d[5] == self.nullid: |
|
1275 | if self.canonical_parent_order and d[5] == self.nullid: | |
1266 | return i[d[6]][7], i[d[5]][7] |
|
1276 | return i[d[6]][7], i[d[5]][7] | |
1267 | else: |
|
1277 | else: | |
1268 | return i[d[5]][7], i[d[6]][7] |
|
1278 | return i[d[5]][7], i[d[6]][7] | |
1269 |
|
1279 | |||
1270 | def chainlen(self, rev): |
|
1280 | def chainlen(self, rev): | |
1271 | return self._chaininfo(rev)[0] |
|
1281 | return self._chaininfo(rev)[0] | |
1272 |
|
1282 | |||
1273 | def _chaininfo(self, rev): |
|
1283 | def _chaininfo(self, rev): | |
1274 | chaininfocache = self._chaininfocache |
|
1284 | chaininfocache = self._chaininfocache | |
1275 | if rev in chaininfocache: |
|
1285 | if rev in chaininfocache: | |
1276 | return chaininfocache[rev] |
|
1286 | return chaininfocache[rev] | |
1277 | index = self.index |
|
1287 | index = self.index | |
1278 | generaldelta = self._generaldelta |
|
1288 | generaldelta = self._generaldelta | |
1279 | iterrev = rev |
|
1289 | iterrev = rev | |
1280 | e = index[iterrev] |
|
1290 | e = index[iterrev] | |
1281 | clen = 0 |
|
1291 | clen = 0 | |
1282 | compresseddeltalen = 0 |
|
1292 | compresseddeltalen = 0 | |
1283 | while iterrev != e[3]: |
|
1293 | while iterrev != e[3]: | |
1284 | clen += 1 |
|
1294 | clen += 1 | |
1285 | compresseddeltalen += e[1] |
|
1295 | compresseddeltalen += e[1] | |
1286 | if generaldelta: |
|
1296 | if generaldelta: | |
1287 | iterrev = e[3] |
|
1297 | iterrev = e[3] | |
1288 | else: |
|
1298 | else: | |
1289 | iterrev -= 1 |
|
1299 | iterrev -= 1 | |
1290 | if iterrev in chaininfocache: |
|
1300 | if iterrev in chaininfocache: | |
1291 | t = chaininfocache[iterrev] |
|
1301 | t = chaininfocache[iterrev] | |
1292 | clen += t[0] |
|
1302 | clen += t[0] | |
1293 | compresseddeltalen += t[1] |
|
1303 | compresseddeltalen += t[1] | |
1294 | break |
|
1304 | break | |
1295 | e = index[iterrev] |
|
1305 | e = index[iterrev] | |
1296 | else: |
|
1306 | else: | |
1297 | # Add text length of base since decompressing that also takes |
|
1307 | # Add text length of base since decompressing that also takes | |
1298 | # work. For cache hits the length is already included. |
|
1308 | # work. For cache hits the length is already included. | |
1299 | compresseddeltalen += e[1] |
|
1309 | compresseddeltalen += e[1] | |
1300 | r = (clen, compresseddeltalen) |
|
1310 | r = (clen, compresseddeltalen) | |
1301 | chaininfocache[rev] = r |
|
1311 | chaininfocache[rev] = r | |
1302 | return r |
|
1312 | return r | |
1303 |
|
1313 | |||
1304 | def _deltachain(self, rev, stoprev=None): |
|
1314 | def _deltachain(self, rev, stoprev=None): | |
1305 | """Obtain the delta chain for a revision. |
|
1315 | """Obtain the delta chain for a revision. | |
1306 |
|
1316 | |||
1307 | ``stoprev`` specifies a revision to stop at. If not specified, we |
|
1317 | ``stoprev`` specifies a revision to stop at. If not specified, we | |
1308 | stop at the base of the chain. |
|
1318 | stop at the base of the chain. | |
1309 |
|
1319 | |||
1310 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of |
|
1320 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of | |
1311 | revs in ascending order and ``stopped`` is a bool indicating whether |
|
1321 | revs in ascending order and ``stopped`` is a bool indicating whether | |
1312 | ``stoprev`` was hit. |
|
1322 | ``stoprev`` was hit. | |
1313 | """ |
|
1323 | """ | |
1314 | # Try C implementation. |
|
1324 | # Try C implementation. | |
1315 | try: |
|
1325 | try: | |
1316 | return self.index.deltachain(rev, stoprev, self._generaldelta) |
|
1326 | return self.index.deltachain(rev, stoprev, self._generaldelta) | |
1317 | except AttributeError: |
|
1327 | except AttributeError: | |
1318 | pass |
|
1328 | pass | |
1319 |
|
1329 | |||
1320 | chain = [] |
|
1330 | chain = [] | |
1321 |
|
1331 | |||
1322 | # Alias to prevent attribute lookup in tight loop. |
|
1332 | # Alias to prevent attribute lookup in tight loop. | |
1323 | index = self.index |
|
1333 | index = self.index | |
1324 | generaldelta = self._generaldelta |
|
1334 | generaldelta = self._generaldelta | |
1325 |
|
1335 | |||
1326 | iterrev = rev |
|
1336 | iterrev = rev | |
1327 | e = index[iterrev] |
|
1337 | e = index[iterrev] | |
1328 | while iterrev != e[3] and iterrev != stoprev: |
|
1338 | while iterrev != e[3] and iterrev != stoprev: | |
1329 | chain.append(iterrev) |
|
1339 | chain.append(iterrev) | |
1330 | if generaldelta: |
|
1340 | if generaldelta: | |
1331 | iterrev = e[3] |
|
1341 | iterrev = e[3] | |
1332 | else: |
|
1342 | else: | |
1333 | iterrev -= 1 |
|
1343 | iterrev -= 1 | |
1334 | e = index[iterrev] |
|
1344 | e = index[iterrev] | |
1335 |
|
1345 | |||
1336 | if iterrev == stoprev: |
|
1346 | if iterrev == stoprev: | |
1337 | stopped = True |
|
1347 | stopped = True | |
1338 | else: |
|
1348 | else: | |
1339 | chain.append(iterrev) |
|
1349 | chain.append(iterrev) | |
1340 | stopped = False |
|
1350 | stopped = False | |
1341 |
|
1351 | |||
1342 | chain.reverse() |
|
1352 | chain.reverse() | |
1343 | return chain, stopped |
|
1353 | return chain, stopped | |
1344 |
|
1354 | |||
1345 | def ancestors(self, revs, stoprev=0, inclusive=False): |
|
1355 | def ancestors(self, revs, stoprev=0, inclusive=False): | |
1346 | """Generate the ancestors of 'revs' in reverse revision order. |
|
1356 | """Generate the ancestors of 'revs' in reverse revision order. | |
1347 | Does not generate revs lower than stoprev. |
|
1357 | Does not generate revs lower than stoprev. | |
1348 |
|
1358 | |||
1349 | See the documentation for ancestor.lazyancestors for more details.""" |
|
1359 | See the documentation for ancestor.lazyancestors for more details.""" | |
1350 |
|
1360 | |||
1351 | # first, make sure start revisions aren't filtered |
|
1361 | # first, make sure start revisions aren't filtered | |
1352 | revs = list(revs) |
|
1362 | revs = list(revs) | |
1353 | checkrev = self.node |
|
1363 | checkrev = self.node | |
1354 | for r in revs: |
|
1364 | for r in revs: | |
1355 | checkrev(r) |
|
1365 | checkrev(r) | |
1356 | # and we're sure ancestors aren't filtered as well |
|
1366 | # and we're sure ancestors aren't filtered as well | |
1357 |
|
1367 | |||
1358 | if rustancestor is not None and self.index.rust_ext_compat: |
|
1368 | if rustancestor is not None and self.index.rust_ext_compat: | |
1359 | lazyancestors = rustancestor.LazyAncestors |
|
1369 | lazyancestors = rustancestor.LazyAncestors | |
1360 | arg = self.index |
|
1370 | arg = self.index | |
1361 | else: |
|
1371 | else: | |
1362 | lazyancestors = ancestor.lazyancestors |
|
1372 | lazyancestors = ancestor.lazyancestors | |
1363 | arg = self._uncheckedparentrevs |
|
1373 | arg = self._uncheckedparentrevs | |
1364 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) |
|
1374 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) | |
1365 |
|
1375 | |||
1366 | def descendants(self, revs): |
|
1376 | def descendants(self, revs): | |
1367 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
1377 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) | |
1368 |
|
1378 | |||
1369 | def findcommonmissing(self, common=None, heads=None): |
|
1379 | def findcommonmissing(self, common=None, heads=None): | |
1370 | """Return a tuple of the ancestors of common and the ancestors of heads |
|
1380 | """Return a tuple of the ancestors of common and the ancestors of heads | |
1371 | that are not ancestors of common. In revset terminology, we return the |
|
1381 | that are not ancestors of common. In revset terminology, we return the | |
1372 | tuple: |
|
1382 | tuple: | |
1373 |
|
1383 | |||
1374 | ::common, (::heads) - (::common) |
|
1384 | ::common, (::heads) - (::common) | |
1375 |
|
1385 | |||
1376 | The list is sorted by revision number, meaning it is |
|
1386 | The list is sorted by revision number, meaning it is | |
1377 | topologically sorted. |
|
1387 | topologically sorted. | |
1378 |
|
1388 | |||
1379 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1389 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1380 | not supplied, uses all of the revlog's heads. If common is not |
|
1390 | not supplied, uses all of the revlog's heads. If common is not | |
1381 | supplied, uses nullid.""" |
|
1391 | supplied, uses nullid.""" | |
1382 | if common is None: |
|
1392 | if common is None: | |
1383 | common = [self.nullid] |
|
1393 | common = [self.nullid] | |
1384 | if heads is None: |
|
1394 | if heads is None: | |
1385 | heads = self.heads() |
|
1395 | heads = self.heads() | |
1386 |
|
1396 | |||
1387 | common = [self.rev(n) for n in common] |
|
1397 | common = [self.rev(n) for n in common] | |
1388 | heads = [self.rev(n) for n in heads] |
|
1398 | heads = [self.rev(n) for n in heads] | |
1389 |
|
1399 | |||
1390 | # we want the ancestors, but inclusive |
|
1400 | # we want the ancestors, but inclusive | |
1391 | class lazyset: |
|
1401 | class lazyset: | |
1392 | def __init__(self, lazyvalues): |
|
1402 | def __init__(self, lazyvalues): | |
1393 | self.addedvalues = set() |
|
1403 | self.addedvalues = set() | |
1394 | self.lazyvalues = lazyvalues |
|
1404 | self.lazyvalues = lazyvalues | |
1395 |
|
1405 | |||
1396 | def __contains__(self, value): |
|
1406 | def __contains__(self, value): | |
1397 | return value in self.addedvalues or value in self.lazyvalues |
|
1407 | return value in self.addedvalues or value in self.lazyvalues | |
1398 |
|
1408 | |||
1399 | def __iter__(self): |
|
1409 | def __iter__(self): | |
1400 | added = self.addedvalues |
|
1410 | added = self.addedvalues | |
1401 | for r in added: |
|
1411 | for r in added: | |
1402 | yield r |
|
1412 | yield r | |
1403 | for r in self.lazyvalues: |
|
1413 | for r in self.lazyvalues: | |
1404 | if not r in added: |
|
1414 | if not r in added: | |
1405 | yield r |
|
1415 | yield r | |
1406 |
|
1416 | |||
1407 | def add(self, value): |
|
1417 | def add(self, value): | |
1408 | self.addedvalues.add(value) |
|
1418 | self.addedvalues.add(value) | |
1409 |
|
1419 | |||
1410 | def update(self, values): |
|
1420 | def update(self, values): | |
1411 | self.addedvalues.update(values) |
|
1421 | self.addedvalues.update(values) | |
1412 |
|
1422 | |||
1413 | has = lazyset(self.ancestors(common)) |
|
1423 | has = lazyset(self.ancestors(common)) | |
1414 | has.add(nullrev) |
|
1424 | has.add(nullrev) | |
1415 | has.update(common) |
|
1425 | has.update(common) | |
1416 |
|
1426 | |||
1417 | # take all ancestors from heads that aren't in has |
|
1427 | # take all ancestors from heads that aren't in has | |
1418 | missing = set() |
|
1428 | missing = set() | |
1419 | visit = collections.deque(r for r in heads if r not in has) |
|
1429 | visit = collections.deque(r for r in heads if r not in has) | |
1420 | while visit: |
|
1430 | while visit: | |
1421 | r = visit.popleft() |
|
1431 | r = visit.popleft() | |
1422 | if r in missing: |
|
1432 | if r in missing: | |
1423 | continue |
|
1433 | continue | |
1424 | else: |
|
1434 | else: | |
1425 | missing.add(r) |
|
1435 | missing.add(r) | |
1426 | for p in self.parentrevs(r): |
|
1436 | for p in self.parentrevs(r): | |
1427 | if p not in has: |
|
1437 | if p not in has: | |
1428 | visit.append(p) |
|
1438 | visit.append(p) | |
1429 | missing = list(missing) |
|
1439 | missing = list(missing) | |
1430 | missing.sort() |
|
1440 | missing.sort() | |
1431 | return has, [self.node(miss) for miss in missing] |
|
1441 | return has, [self.node(miss) for miss in missing] | |
1432 |
|
1442 | |||
1433 | def incrementalmissingrevs(self, common=None): |
|
1443 | def incrementalmissingrevs(self, common=None): | |
1434 | """Return an object that can be used to incrementally compute the |
|
1444 | """Return an object that can be used to incrementally compute the | |
1435 | revision numbers of the ancestors of arbitrary sets that are not |
|
1445 | revision numbers of the ancestors of arbitrary sets that are not | |
1436 | ancestors of common. This is an ancestor.incrementalmissingancestors |
|
1446 | ancestors of common. This is an ancestor.incrementalmissingancestors | |
1437 | object. |
|
1447 | object. | |
1438 |
|
1448 | |||
1439 | 'common' is a list of revision numbers. If common is not supplied, uses |
|
1449 | 'common' is a list of revision numbers. If common is not supplied, uses | |
1440 | nullrev. |
|
1450 | nullrev. | |
1441 | """ |
|
1451 | """ | |
1442 | if common is None: |
|
1452 | if common is None: | |
1443 | common = [nullrev] |
|
1453 | common = [nullrev] | |
1444 |
|
1454 | |||
1445 | if rustancestor is not None and self.index.rust_ext_compat: |
|
1455 | if rustancestor is not None and self.index.rust_ext_compat: | |
1446 | return rustancestor.MissingAncestors(self.index, common) |
|
1456 | return rustancestor.MissingAncestors(self.index, common) | |
1447 | return ancestor.incrementalmissingancestors(self.parentrevs, common) |
|
1457 | return ancestor.incrementalmissingancestors(self.parentrevs, common) | |
1448 |
|
1458 | |||
1449 | def findmissingrevs(self, common=None, heads=None): |
|
1459 | def findmissingrevs(self, common=None, heads=None): | |
1450 | """Return the revision numbers of the ancestors of heads that |
|
1460 | """Return the revision numbers of the ancestors of heads that | |
1451 | are not ancestors of common. |
|
1461 | are not ancestors of common. | |
1452 |
|
1462 | |||
1453 | More specifically, return a list of revision numbers corresponding to |
|
1463 | More specifically, return a list of revision numbers corresponding to | |
1454 | nodes N such that every N satisfies the following constraints: |
|
1464 | nodes N such that every N satisfies the following constraints: | |
1455 |
|
1465 | |||
1456 | 1. N is an ancestor of some node in 'heads' |
|
1466 | 1. N is an ancestor of some node in 'heads' | |
1457 | 2. N is not an ancestor of any node in 'common' |
|
1467 | 2. N is not an ancestor of any node in 'common' | |
1458 |
|
1468 | |||
1459 | The list is sorted by revision number, meaning it is |
|
1469 | The list is sorted by revision number, meaning it is | |
1460 | topologically sorted. |
|
1470 | topologically sorted. | |
1461 |
|
1471 | |||
1462 | 'heads' and 'common' are both lists of revision numbers. If heads is |
|
1472 | 'heads' and 'common' are both lists of revision numbers. If heads is | |
1463 | not supplied, uses all of the revlog's heads. If common is not |
|
1473 | not supplied, uses all of the revlog's heads. If common is not | |
1464 | supplied, uses nullid.""" |
|
1474 | supplied, uses nullid.""" | |
1465 | if common is None: |
|
1475 | if common is None: | |
1466 | common = [nullrev] |
|
1476 | common = [nullrev] | |
1467 | if heads is None: |
|
1477 | if heads is None: | |
1468 | heads = self.headrevs() |
|
1478 | heads = self.headrevs() | |
1469 |
|
1479 | |||
1470 | inc = self.incrementalmissingrevs(common=common) |
|
1480 | inc = self.incrementalmissingrevs(common=common) | |
1471 | return inc.missingancestors(heads) |
|
1481 | return inc.missingancestors(heads) | |
1472 |
|
1482 | |||
1473 | def findmissing(self, common=None, heads=None): |
|
1483 | def findmissing(self, common=None, heads=None): | |
1474 | """Return the ancestors of heads that are not ancestors of common. |
|
1484 | """Return the ancestors of heads that are not ancestors of common. | |
1475 |
|
1485 | |||
1476 | More specifically, return a list of nodes N such that every N |
|
1486 | More specifically, return a list of nodes N such that every N | |
1477 | satisfies the following constraints: |
|
1487 | satisfies the following constraints: | |
1478 |
|
1488 | |||
1479 | 1. N is an ancestor of some node in 'heads' |
|
1489 | 1. N is an ancestor of some node in 'heads' | |
1480 | 2. N is not an ancestor of any node in 'common' |
|
1490 | 2. N is not an ancestor of any node in 'common' | |
1481 |
|
1491 | |||
1482 | The list is sorted by revision number, meaning it is |
|
1492 | The list is sorted by revision number, meaning it is | |
1483 | topologically sorted. |
|
1493 | topologically sorted. | |
1484 |
|
1494 | |||
1485 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1495 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1486 | not supplied, uses all of the revlog's heads. If common is not |
|
1496 | not supplied, uses all of the revlog's heads. If common is not | |
1487 | supplied, uses nullid.""" |
|
1497 | supplied, uses nullid.""" | |
1488 | if common is None: |
|
1498 | if common is None: | |
1489 | common = [self.nullid] |
|
1499 | common = [self.nullid] | |
1490 | if heads is None: |
|
1500 | if heads is None: | |
1491 | heads = self.heads() |
|
1501 | heads = self.heads() | |
1492 |
|
1502 | |||
1493 | common = [self.rev(n) for n in common] |
|
1503 | common = [self.rev(n) for n in common] | |
1494 | heads = [self.rev(n) for n in heads] |
|
1504 | heads = [self.rev(n) for n in heads] | |
1495 |
|
1505 | |||
1496 | inc = self.incrementalmissingrevs(common=common) |
|
1506 | inc = self.incrementalmissingrevs(common=common) | |
1497 | return [self.node(r) for r in inc.missingancestors(heads)] |
|
1507 | return [self.node(r) for r in inc.missingancestors(heads)] | |
1498 |
|
1508 | |||
1499 | def nodesbetween(self, roots=None, heads=None): |
|
1509 | def nodesbetween(self, roots=None, heads=None): | |
1500 | """Return a topological path from 'roots' to 'heads'. |
|
1510 | """Return a topological path from 'roots' to 'heads'. | |
1501 |
|
1511 | |||
1502 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a |
|
1512 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a | |
1503 | topologically sorted list of all nodes N that satisfy both of |
|
1513 | topologically sorted list of all nodes N that satisfy both of | |
1504 | these constraints: |
|
1514 | these constraints: | |
1505 |
|
1515 | |||
1506 | 1. N is a descendant of some node in 'roots' |
|
1516 | 1. N is a descendant of some node in 'roots' | |
1507 | 2. N is an ancestor of some node in 'heads' |
|
1517 | 2. N is an ancestor of some node in 'heads' | |
1508 |
|
1518 | |||
1509 | Every node is considered to be both a descendant and an ancestor |
|
1519 | Every node is considered to be both a descendant and an ancestor | |
1510 | of itself, so every reachable node in 'roots' and 'heads' will be |
|
1520 | of itself, so every reachable node in 'roots' and 'heads' will be | |
1511 | included in 'nodes'. |
|
1521 | included in 'nodes'. | |
1512 |
|
1522 | |||
1513 | 'outroots' is the list of reachable nodes in 'roots', i.e., the |
|
1523 | 'outroots' is the list of reachable nodes in 'roots', i.e., the | |
1514 | subset of 'roots' that is returned in 'nodes'. Likewise, |
|
1524 | subset of 'roots' that is returned in 'nodes'. Likewise, | |
1515 | 'outheads' is the subset of 'heads' that is also in 'nodes'. |
|
1525 | 'outheads' is the subset of 'heads' that is also in 'nodes'. | |
1516 |
|
1526 | |||
1517 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is |
|
1527 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is | |
1518 | unspecified, uses nullid as the only root. If 'heads' is |
|
1528 | unspecified, uses nullid as the only root. If 'heads' is | |
1519 | unspecified, uses list of all of the revlog's heads.""" |
|
1529 | unspecified, uses list of all of the revlog's heads.""" | |
1520 | nonodes = ([], [], []) |
|
1530 | nonodes = ([], [], []) | |
1521 | if roots is not None: |
|
1531 | if roots is not None: | |
1522 | roots = list(roots) |
|
1532 | roots = list(roots) | |
1523 | if not roots: |
|
1533 | if not roots: | |
1524 | return nonodes |
|
1534 | return nonodes | |
1525 | lowestrev = min([self.rev(n) for n in roots]) |
|
1535 | lowestrev = min([self.rev(n) for n in roots]) | |
1526 | else: |
|
1536 | else: | |
1527 | roots = [self.nullid] # Everybody's a descendant of nullid |
|
1537 | roots = [self.nullid] # Everybody's a descendant of nullid | |
1528 | lowestrev = nullrev |
|
1538 | lowestrev = nullrev | |
1529 | if (lowestrev == nullrev) and (heads is None): |
|
1539 | if (lowestrev == nullrev) and (heads is None): | |
1530 | # We want _all_ the nodes! |
|
1540 | # We want _all_ the nodes! | |
1531 | return ( |
|
1541 | return ( | |
1532 | [self.node(r) for r in self], |
|
1542 | [self.node(r) for r in self], | |
1533 | [self.nullid], |
|
1543 | [self.nullid], | |
1534 | list(self.heads()), |
|
1544 | list(self.heads()), | |
1535 | ) |
|
1545 | ) | |
1536 | if heads is None: |
|
1546 | if heads is None: | |
1537 | # All nodes are ancestors, so the latest ancestor is the last |
|
1547 | # All nodes are ancestors, so the latest ancestor is the last | |
1538 | # node. |
|
1548 | # node. | |
1539 | highestrev = len(self) - 1 |
|
1549 | highestrev = len(self) - 1 | |
1540 | # Set ancestors to None to signal that every node is an ancestor. |
|
1550 | # Set ancestors to None to signal that every node is an ancestor. | |
1541 | ancestors = None |
|
1551 | ancestors = None | |
1542 | # Set heads to an empty dictionary for later discovery of heads |
|
1552 | # Set heads to an empty dictionary for later discovery of heads | |
1543 | heads = {} |
|
1553 | heads = {} | |
1544 | else: |
|
1554 | else: | |
1545 | heads = list(heads) |
|
1555 | heads = list(heads) | |
1546 | if not heads: |
|
1556 | if not heads: | |
1547 | return nonodes |
|
1557 | return nonodes | |
1548 | ancestors = set() |
|
1558 | ancestors = set() | |
1549 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
1559 | # Turn heads into a dictionary so we can remove 'fake' heads. | |
1550 | # Also, later we will be using it to filter out the heads we can't |
|
1560 | # Also, later we will be using it to filter out the heads we can't | |
1551 | # find from roots. |
|
1561 | # find from roots. | |
1552 | heads = dict.fromkeys(heads, False) |
|
1562 | heads = dict.fromkeys(heads, False) | |
1553 | # Start at the top and keep marking parents until we're done. |
|
1563 | # Start at the top and keep marking parents until we're done. | |
1554 | nodestotag = set(heads) |
|
1564 | nodestotag = set(heads) | |
1555 | # Remember where the top was so we can use it as a limit later. |
|
1565 | # Remember where the top was so we can use it as a limit later. | |
1556 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
1566 | highestrev = max([self.rev(n) for n in nodestotag]) | |
1557 | while nodestotag: |
|
1567 | while nodestotag: | |
1558 | # grab a node to tag |
|
1568 | # grab a node to tag | |
1559 | n = nodestotag.pop() |
|
1569 | n = nodestotag.pop() | |
1560 | # Never tag nullid |
|
1570 | # Never tag nullid | |
1561 | if n == self.nullid: |
|
1571 | if n == self.nullid: | |
1562 | continue |
|
1572 | continue | |
1563 | # A node's revision number represents its place in a |
|
1573 | # A node's revision number represents its place in a | |
1564 | # topologically sorted list of nodes. |
|
1574 | # topologically sorted list of nodes. | |
1565 | r = self.rev(n) |
|
1575 | r = self.rev(n) | |
1566 | if r >= lowestrev: |
|
1576 | if r >= lowestrev: | |
1567 | if n not in ancestors: |
|
1577 | if n not in ancestors: | |
1568 | # If we are possibly a descendant of one of the roots |
|
1578 | # If we are possibly a descendant of one of the roots | |
1569 | # and we haven't already been marked as an ancestor |
|
1579 | # and we haven't already been marked as an ancestor | |
1570 | ancestors.add(n) # Mark as ancestor |
|
1580 | ancestors.add(n) # Mark as ancestor | |
1571 | # Add non-nullid parents to list of nodes to tag. |
|
1581 | # Add non-nullid parents to list of nodes to tag. | |
1572 | nodestotag.update( |
|
1582 | nodestotag.update( | |
1573 | [p for p in self.parents(n) if p != self.nullid] |
|
1583 | [p for p in self.parents(n) if p != self.nullid] | |
1574 | ) |
|
1584 | ) | |
1575 | elif n in heads: # We've seen it before, is it a fake head? |
|
1585 | elif n in heads: # We've seen it before, is it a fake head? | |
1576 | # So it is, real heads should not be the ancestors of |
|
1586 | # So it is, real heads should not be the ancestors of | |
1577 | # any other heads. |
|
1587 | # any other heads. | |
1578 | heads.pop(n) |
|
1588 | heads.pop(n) | |
1579 | if not ancestors: |
|
1589 | if not ancestors: | |
1580 | return nonodes |
|
1590 | return nonodes | |
1581 | # Now that we have our set of ancestors, we want to remove any |
|
1591 | # Now that we have our set of ancestors, we want to remove any | |
1582 | # roots that are not ancestors. |
|
1592 | # roots that are not ancestors. | |
1583 |
|
1593 | |||
1584 | # If one of the roots was nullid, everything is included anyway. |
|
1594 | # If one of the roots was nullid, everything is included anyway. | |
1585 | if lowestrev > nullrev: |
|
1595 | if lowestrev > nullrev: | |
1586 | # But, since we weren't, let's recompute the lowest rev to not |
|
1596 | # But, since we weren't, let's recompute the lowest rev to not | |
1587 | # include roots that aren't ancestors. |
|
1597 | # include roots that aren't ancestors. | |
1588 |
|
1598 | |||
1589 | # Filter out roots that aren't ancestors of heads |
|
1599 | # Filter out roots that aren't ancestors of heads | |
1590 | roots = [root for root in roots if root in ancestors] |
|
1600 | roots = [root for root in roots if root in ancestors] | |
1591 | # Recompute the lowest revision |
|
1601 | # Recompute the lowest revision | |
1592 | if roots: |
|
1602 | if roots: | |
1593 | lowestrev = min([self.rev(root) for root in roots]) |
|
1603 | lowestrev = min([self.rev(root) for root in roots]) | |
1594 | else: |
|
1604 | else: | |
1595 | # No more roots? Return empty list |
|
1605 | # No more roots? Return empty list | |
1596 | return nonodes |
|
1606 | return nonodes | |
1597 | else: |
|
1607 | else: | |
1598 | # We are descending from nullid, and don't need to care about |
|
1608 | # We are descending from nullid, and don't need to care about | |
1599 | # any other roots. |
|
1609 | # any other roots. | |
1600 | lowestrev = nullrev |
|
1610 | lowestrev = nullrev | |
1601 | roots = [self.nullid] |
|
1611 | roots = [self.nullid] | |
1602 | # Transform our roots list into a set. |
|
1612 | # Transform our roots list into a set. | |
1603 | descendants = set(roots) |
|
1613 | descendants = set(roots) | |
1604 | # Also, keep the original roots so we can filter out roots that aren't |
|
1614 | # Also, keep the original roots so we can filter out roots that aren't | |
1605 | # 'real' roots (i.e. are descended from other roots). |
|
1615 | # 'real' roots (i.e. are descended from other roots). | |
1606 | roots = descendants.copy() |
|
1616 | roots = descendants.copy() | |
1607 | # Our topologically sorted list of output nodes. |
|
1617 | # Our topologically sorted list of output nodes. | |
1608 | orderedout = [] |
|
1618 | orderedout = [] | |
1609 | # Don't start at nullid since we don't want nullid in our output list, |
|
1619 | # Don't start at nullid since we don't want nullid in our output list, | |
1610 | # and if nullid shows up in descendants, empty parents will look like |
|
1620 | # and if nullid shows up in descendants, empty parents will look like | |
1611 | # they're descendants. |
|
1621 | # they're descendants. | |
1612 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): |
|
1622 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): | |
1613 | n = self.node(r) |
|
1623 | n = self.node(r) | |
1614 | isdescendant = False |
|
1624 | isdescendant = False | |
1615 | if lowestrev == nullrev: # Everybody is a descendant of nullid |
|
1625 | if lowestrev == nullrev: # Everybody is a descendant of nullid | |
1616 | isdescendant = True |
|
1626 | isdescendant = True | |
1617 | elif n in descendants: |
|
1627 | elif n in descendants: | |
1618 | # n is already a descendant |
|
1628 | # n is already a descendant | |
1619 | isdescendant = True |
|
1629 | isdescendant = True | |
1620 | # This check only needs to be done here because all the roots |
|
1630 | # This check only needs to be done here because all the roots | |
1621 | # will start being marked is descendants before the loop. |
|
1631 | # will start being marked is descendants before the loop. | |
1622 | if n in roots: |
|
1632 | if n in roots: | |
1623 | # If n was a root, check if it's a 'real' root. |
|
1633 | # If n was a root, check if it's a 'real' root. | |
1624 | p = tuple(self.parents(n)) |
|
1634 | p = tuple(self.parents(n)) | |
1625 | # If any of its parents are descendants, it's not a root. |
|
1635 | # If any of its parents are descendants, it's not a root. | |
1626 | if (p[0] in descendants) or (p[1] in descendants): |
|
1636 | if (p[0] in descendants) or (p[1] in descendants): | |
1627 | roots.remove(n) |
|
1637 | roots.remove(n) | |
1628 | else: |
|
1638 | else: | |
1629 | p = tuple(self.parents(n)) |
|
1639 | p = tuple(self.parents(n)) | |
1630 | # A node is a descendant if either of its parents are |
|
1640 | # A node is a descendant if either of its parents are | |
1631 | # descendants. (We seeded the dependents list with the roots |
|
1641 | # descendants. (We seeded the dependents list with the roots | |
1632 | # up there, remember?) |
|
1642 | # up there, remember?) | |
1633 | if (p[0] in descendants) or (p[1] in descendants): |
|
1643 | if (p[0] in descendants) or (p[1] in descendants): | |
1634 | descendants.add(n) |
|
1644 | descendants.add(n) | |
1635 | isdescendant = True |
|
1645 | isdescendant = True | |
1636 | if isdescendant and ((ancestors is None) or (n in ancestors)): |
|
1646 | if isdescendant and ((ancestors is None) or (n in ancestors)): | |
1637 | # Only include nodes that are both descendants and ancestors. |
|
1647 | # Only include nodes that are both descendants and ancestors. | |
1638 | orderedout.append(n) |
|
1648 | orderedout.append(n) | |
1639 | if (ancestors is not None) and (n in heads): |
|
1649 | if (ancestors is not None) and (n in heads): | |
1640 | # We're trying to figure out which heads are reachable |
|
1650 | # We're trying to figure out which heads are reachable | |
1641 | # from roots. |
|
1651 | # from roots. | |
1642 | # Mark this head as having been reached |
|
1652 | # Mark this head as having been reached | |
1643 | heads[n] = True |
|
1653 | heads[n] = True | |
1644 | elif ancestors is None: |
|
1654 | elif ancestors is None: | |
1645 | # Otherwise, we're trying to discover the heads. |
|
1655 | # Otherwise, we're trying to discover the heads. | |
1646 | # Assume this is a head because if it isn't, the next step |
|
1656 | # Assume this is a head because if it isn't, the next step | |
1647 | # will eventually remove it. |
|
1657 | # will eventually remove it. | |
1648 | heads[n] = True |
|
1658 | heads[n] = True | |
1649 | # But, obviously its parents aren't. |
|
1659 | # But, obviously its parents aren't. | |
1650 | for p in self.parents(n): |
|
1660 | for p in self.parents(n): | |
1651 | heads.pop(p, None) |
|
1661 | heads.pop(p, None) | |
1652 | heads = [head for head, flag in heads.items() if flag] |
|
1662 | heads = [head for head, flag in heads.items() if flag] | |
1653 | roots = list(roots) |
|
1663 | roots = list(roots) | |
1654 | assert orderedout |
|
1664 | assert orderedout | |
1655 | assert roots |
|
1665 | assert roots | |
1656 | assert heads |
|
1666 | assert heads | |
1657 | return (orderedout, roots, heads) |
|
1667 | return (orderedout, roots, heads) | |
1658 |
|
1668 | |||
1659 | def headrevs(self, revs=None): |
|
1669 | def headrevs(self, revs=None): | |
1660 | if revs is None: |
|
1670 | if revs is None: | |
1661 | try: |
|
1671 | try: | |
1662 | return self.index.headrevs() |
|
1672 | return self.index.headrevs() | |
1663 | except AttributeError: |
|
1673 | except AttributeError: | |
1664 | return self._headrevs() |
|
1674 | return self._headrevs() | |
1665 | if rustdagop is not None and self.index.rust_ext_compat: |
|
1675 | if rustdagop is not None and self.index.rust_ext_compat: | |
1666 | return rustdagop.headrevs(self.index, revs) |
|
1676 | return rustdagop.headrevs(self.index, revs) | |
1667 | return dagop.headrevs(revs, self._uncheckedparentrevs) |
|
1677 | return dagop.headrevs(revs, self._uncheckedparentrevs) | |
1668 |
|
1678 | |||
1669 | def computephases(self, roots): |
|
1679 | def computephases(self, roots): | |
1670 | return self.index.computephasesmapsets(roots) |
|
1680 | return self.index.computephasesmapsets(roots) | |
1671 |
|
1681 | |||
1672 | def _headrevs(self): |
|
1682 | def _headrevs(self): | |
1673 | count = len(self) |
|
1683 | count = len(self) | |
1674 | if not count: |
|
1684 | if not count: | |
1675 | return [nullrev] |
|
1685 | return [nullrev] | |
1676 | # we won't iter over filtered rev so nobody is a head at start |
|
1686 | # we won't iter over filtered rev so nobody is a head at start | |
1677 | ishead = [0] * (count + 1) |
|
1687 | ishead = [0] * (count + 1) | |
1678 | index = self.index |
|
1688 | index = self.index | |
1679 | for r in self: |
|
1689 | for r in self: | |
1680 | ishead[r] = 1 # I may be an head |
|
1690 | ishead[r] = 1 # I may be an head | |
1681 | e = index[r] |
|
1691 | e = index[r] | |
1682 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not |
|
1692 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not | |
1683 | return [r for r, val in enumerate(ishead) if val] |
|
1693 | return [r for r, val in enumerate(ishead) if val] | |
1684 |
|
1694 | |||
1685 | def heads(self, start=None, stop=None): |
|
1695 | def heads(self, start=None, stop=None): | |
1686 | """return the list of all nodes that have no children |
|
1696 | """return the list of all nodes that have no children | |
1687 |
|
1697 | |||
1688 | if start is specified, only heads that are descendants of |
|
1698 | if start is specified, only heads that are descendants of | |
1689 | start will be returned |
|
1699 | start will be returned | |
1690 | if stop is specified, it will consider all the revs from stop |
|
1700 | if stop is specified, it will consider all the revs from stop | |
1691 | as if they had no children |
|
1701 | as if they had no children | |
1692 | """ |
|
1702 | """ | |
1693 | if start is None and stop is None: |
|
1703 | if start is None and stop is None: | |
1694 | if not len(self): |
|
1704 | if not len(self): | |
1695 | return [self.nullid] |
|
1705 | return [self.nullid] | |
1696 | return [self.node(r) for r in self.headrevs()] |
|
1706 | return [self.node(r) for r in self.headrevs()] | |
1697 |
|
1707 | |||
1698 | if start is None: |
|
1708 | if start is None: | |
1699 | start = nullrev |
|
1709 | start = nullrev | |
1700 | else: |
|
1710 | else: | |
1701 | start = self.rev(start) |
|
1711 | start = self.rev(start) | |
1702 |
|
1712 | |||
1703 | stoprevs = {self.rev(n) for n in stop or []} |
|
1713 | stoprevs = {self.rev(n) for n in stop or []} | |
1704 |
|
1714 | |||
1705 | revs = dagop.headrevssubset( |
|
1715 | revs = dagop.headrevssubset( | |
1706 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs |
|
1716 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs | |
1707 | ) |
|
1717 | ) | |
1708 |
|
1718 | |||
1709 | return [self.node(rev) for rev in revs] |
|
1719 | return [self.node(rev) for rev in revs] | |
1710 |
|
1720 | |||
1711 | def children(self, node): |
|
1721 | def children(self, node): | |
1712 | """find the children of a given node""" |
|
1722 | """find the children of a given node""" | |
1713 | c = [] |
|
1723 | c = [] | |
1714 | p = self.rev(node) |
|
1724 | p = self.rev(node) | |
1715 | for r in self.revs(start=p + 1): |
|
1725 | for r in self.revs(start=p + 1): | |
1716 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
1726 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
1717 | if prevs: |
|
1727 | if prevs: | |
1718 | for pr in prevs: |
|
1728 | for pr in prevs: | |
1719 | if pr == p: |
|
1729 | if pr == p: | |
1720 | c.append(self.node(r)) |
|
1730 | c.append(self.node(r)) | |
1721 | elif p == nullrev: |
|
1731 | elif p == nullrev: | |
1722 | c.append(self.node(r)) |
|
1732 | c.append(self.node(r)) | |
1723 | return c |
|
1733 | return c | |
1724 |
|
1734 | |||
1725 | def commonancestorsheads(self, a, b): |
|
1735 | def commonancestorsheads(self, a, b): | |
1726 | """calculate all the heads of the common ancestors of nodes a and b""" |
|
1736 | """calculate all the heads of the common ancestors of nodes a and b""" | |
1727 | a, b = self.rev(a), self.rev(b) |
|
1737 | a, b = self.rev(a), self.rev(b) | |
1728 | ancs = self._commonancestorsheads(a, b) |
|
1738 | ancs = self._commonancestorsheads(a, b) | |
1729 | return pycompat.maplist(self.node, ancs) |
|
1739 | return pycompat.maplist(self.node, ancs) | |
1730 |
|
1740 | |||
1731 | def _commonancestorsheads(self, *revs): |
|
1741 | def _commonancestorsheads(self, *revs): | |
1732 | """calculate all the heads of the common ancestors of revs""" |
|
1742 | """calculate all the heads of the common ancestors of revs""" | |
1733 | try: |
|
1743 | try: | |
1734 | ancs = self.index.commonancestorsheads(*revs) |
|
1744 | ancs = self.index.commonancestorsheads(*revs) | |
1735 | except (AttributeError, OverflowError): # C implementation failed |
|
1745 | except (AttributeError, OverflowError): # C implementation failed | |
1736 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) |
|
1746 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) | |
1737 | return ancs |
|
1747 | return ancs | |
1738 |
|
1748 | |||
1739 | def isancestor(self, a, b): |
|
1749 | def isancestor(self, a, b): | |
1740 | """return True if node a is an ancestor of node b |
|
1750 | """return True if node a is an ancestor of node b | |
1741 |
|
1751 | |||
1742 | A revision is considered an ancestor of itself.""" |
|
1752 | A revision is considered an ancestor of itself.""" | |
1743 | a, b = self.rev(a), self.rev(b) |
|
1753 | a, b = self.rev(a), self.rev(b) | |
1744 | return self.isancestorrev(a, b) |
|
1754 | return self.isancestorrev(a, b) | |
1745 |
|
1755 | |||
1746 | def isancestorrev(self, a, b): |
|
1756 | def isancestorrev(self, a, b): | |
1747 | """return True if revision a is an ancestor of revision b |
|
1757 | """return True if revision a is an ancestor of revision b | |
1748 |
|
1758 | |||
1749 | A revision is considered an ancestor of itself. |
|
1759 | A revision is considered an ancestor of itself. | |
1750 |
|
1760 | |||
1751 | The implementation of this is trivial but the use of |
|
1761 | The implementation of this is trivial but the use of | |
1752 | reachableroots is not.""" |
|
1762 | reachableroots is not.""" | |
1753 | if a == nullrev: |
|
1763 | if a == nullrev: | |
1754 | return True |
|
1764 | return True | |
1755 | elif a == b: |
|
1765 | elif a == b: | |
1756 | return True |
|
1766 | return True | |
1757 | elif a > b: |
|
1767 | elif a > b: | |
1758 | return False |
|
1768 | return False | |
1759 | return bool(self.reachableroots(a, [b], [a], includepath=False)) |
|
1769 | return bool(self.reachableroots(a, [b], [a], includepath=False)) | |
1760 |
|
1770 | |||
1761 | def reachableroots(self, minroot, heads, roots, includepath=False): |
|
1771 | def reachableroots(self, minroot, heads, roots, includepath=False): | |
1762 | """return (heads(::(<roots> and <roots>::<heads>))) |
|
1772 | """return (heads(::(<roots> and <roots>::<heads>))) | |
1763 |
|
1773 | |||
1764 | If includepath is True, return (<roots>::<heads>).""" |
|
1774 | If includepath is True, return (<roots>::<heads>).""" | |
1765 | try: |
|
1775 | try: | |
1766 | return self.index.reachableroots2( |
|
1776 | return self.index.reachableroots2( | |
1767 | minroot, heads, roots, includepath |
|
1777 | minroot, heads, roots, includepath | |
1768 | ) |
|
1778 | ) | |
1769 | except AttributeError: |
|
1779 | except AttributeError: | |
1770 | return dagop._reachablerootspure( |
|
1780 | return dagop._reachablerootspure( | |
1771 | self.parentrevs, minroot, roots, heads, includepath |
|
1781 | self.parentrevs, minroot, roots, heads, includepath | |
1772 | ) |
|
1782 | ) | |
1773 |
|
1783 | |||
1774 | def ancestor(self, a, b): |
|
1784 | def ancestor(self, a, b): | |
1775 | """calculate the "best" common ancestor of nodes a and b""" |
|
1785 | """calculate the "best" common ancestor of nodes a and b""" | |
1776 |
|
1786 | |||
1777 | a, b = self.rev(a), self.rev(b) |
|
1787 | a, b = self.rev(a), self.rev(b) | |
1778 | try: |
|
1788 | try: | |
1779 | ancs = self.index.ancestors(a, b) |
|
1789 | ancs = self.index.ancestors(a, b) | |
1780 | except (AttributeError, OverflowError): |
|
1790 | except (AttributeError, OverflowError): | |
1781 | ancs = ancestor.ancestors(self.parentrevs, a, b) |
|
1791 | ancs = ancestor.ancestors(self.parentrevs, a, b) | |
1782 | if ancs: |
|
1792 | if ancs: | |
1783 | # choose a consistent winner when there's a tie |
|
1793 | # choose a consistent winner when there's a tie | |
1784 | return min(map(self.node, ancs)) |
|
1794 | return min(map(self.node, ancs)) | |
1785 | return self.nullid |
|
1795 | return self.nullid | |
1786 |
|
1796 | |||
1787 | def _match(self, id): |
|
1797 | def _match(self, id): | |
1788 | if isinstance(id, int): |
|
1798 | if isinstance(id, int): | |
1789 | # rev |
|
1799 | # rev | |
1790 | return self.node(id) |
|
1800 | return self.node(id) | |
1791 | if len(id) == self.nodeconstants.nodelen: |
|
1801 | if len(id) == self.nodeconstants.nodelen: | |
1792 | # possibly a binary node |
|
1802 | # possibly a binary node | |
1793 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
1803 | # odds of a binary node being all hex in ASCII are 1 in 10**25 | |
1794 | try: |
|
1804 | try: | |
1795 | node = id |
|
1805 | node = id | |
1796 | self.rev(node) # quick search the index |
|
1806 | self.rev(node) # quick search the index | |
1797 | return node |
|
1807 | return node | |
1798 | except error.LookupError: |
|
1808 | except error.LookupError: | |
1799 | pass # may be partial hex id |
|
1809 | pass # may be partial hex id | |
1800 | try: |
|
1810 | try: | |
1801 | # str(rev) |
|
1811 | # str(rev) | |
1802 | rev = int(id) |
|
1812 | rev = int(id) | |
1803 | if b"%d" % rev != id: |
|
1813 | if b"%d" % rev != id: | |
1804 | raise ValueError |
|
1814 | raise ValueError | |
1805 | if rev < 0: |
|
1815 | if rev < 0: | |
1806 | rev = len(self) + rev |
|
1816 | rev = len(self) + rev | |
1807 | if rev < 0 or rev >= len(self): |
|
1817 | if rev < 0 or rev >= len(self): | |
1808 | raise ValueError |
|
1818 | raise ValueError | |
1809 | return self.node(rev) |
|
1819 | return self.node(rev) | |
1810 | except (ValueError, OverflowError): |
|
1820 | except (ValueError, OverflowError): | |
1811 | pass |
|
1821 | pass | |
1812 | if len(id) == 2 * self.nodeconstants.nodelen: |
|
1822 | if len(id) == 2 * self.nodeconstants.nodelen: | |
1813 | try: |
|
1823 | try: | |
1814 | # a full hex nodeid? |
|
1824 | # a full hex nodeid? | |
1815 | node = bin(id) |
|
1825 | node = bin(id) | |
1816 | self.rev(node) |
|
1826 | self.rev(node) | |
1817 | return node |
|
1827 | return node | |
1818 | except (binascii.Error, error.LookupError): |
|
1828 | except (binascii.Error, error.LookupError): | |
1819 | pass |
|
1829 | pass | |
1820 |
|
1830 | |||
1821 | def _partialmatch(self, id): |
|
1831 | def _partialmatch(self, id): | |
1822 | # we don't care wdirfilenodeids as they should be always full hash |
|
1832 | # we don't care wdirfilenodeids as they should be always full hash | |
1823 | maybewdir = self.nodeconstants.wdirhex.startswith(id) |
|
1833 | maybewdir = self.nodeconstants.wdirhex.startswith(id) | |
1824 | ambiguous = False |
|
1834 | ambiguous = False | |
1825 | try: |
|
1835 | try: | |
1826 | partial = self.index.partialmatch(id) |
|
1836 | partial = self.index.partialmatch(id) | |
1827 | if partial and self.hasnode(partial): |
|
1837 | if partial and self.hasnode(partial): | |
1828 | if maybewdir: |
|
1838 | if maybewdir: | |
1829 | # single 'ff...' match in radix tree, ambiguous with wdir |
|
1839 | # single 'ff...' match in radix tree, ambiguous with wdir | |
1830 | ambiguous = True |
|
1840 | ambiguous = True | |
1831 | else: |
|
1841 | else: | |
1832 | return partial |
|
1842 | return partial | |
1833 | elif maybewdir: |
|
1843 | elif maybewdir: | |
1834 | # no 'ff...' match in radix tree, wdir identified |
|
1844 | # no 'ff...' match in radix tree, wdir identified | |
1835 | raise error.WdirUnsupported |
|
1845 | raise error.WdirUnsupported | |
1836 | else: |
|
1846 | else: | |
1837 | return None |
|
1847 | return None | |
1838 | except error.RevlogError: |
|
1848 | except error.RevlogError: | |
1839 | # parsers.c radix tree lookup gave multiple matches |
|
1849 | # parsers.c radix tree lookup gave multiple matches | |
1840 | # fast path: for unfiltered changelog, radix tree is accurate |
|
1850 | # fast path: for unfiltered changelog, radix tree is accurate | |
1841 | if not getattr(self, 'filteredrevs', None): |
|
1851 | if not getattr(self, 'filteredrevs', None): | |
1842 | ambiguous = True |
|
1852 | ambiguous = True | |
1843 | # fall through to slow path that filters hidden revisions |
|
1853 | # fall through to slow path that filters hidden revisions | |
1844 | except (AttributeError, ValueError): |
|
1854 | except (AttributeError, ValueError): | |
1845 | # we are pure python, or key is not hex |
|
1855 | # we are pure python, or key is not hex | |
1846 | pass |
|
1856 | pass | |
1847 | if ambiguous: |
|
1857 | if ambiguous: | |
1848 | raise error.AmbiguousPrefixLookupError( |
|
1858 | raise error.AmbiguousPrefixLookupError( | |
1849 | id, self.display_id, _(b'ambiguous identifier') |
|
1859 | id, self.display_id, _(b'ambiguous identifier') | |
1850 | ) |
|
1860 | ) | |
1851 |
|
1861 | |||
1852 | if id in self._pcache: |
|
1862 | if id in self._pcache: | |
1853 | return self._pcache[id] |
|
1863 | return self._pcache[id] | |
1854 |
|
1864 | |||
1855 | if len(id) <= 40: |
|
1865 | if len(id) <= 40: | |
1856 | # hex(node)[:...] |
|
1866 | # hex(node)[:...] | |
1857 | l = len(id) // 2 * 2 # grab an even number of digits |
|
1867 | l = len(id) // 2 * 2 # grab an even number of digits | |
1858 | try: |
|
1868 | try: | |
1859 | # we're dropping the last digit, so let's check that it's hex, |
|
1869 | # we're dropping the last digit, so let's check that it's hex, | |
1860 | # to avoid the expensive computation below if it's not |
|
1870 | # to avoid the expensive computation below if it's not | |
1861 | if len(id) % 2 > 0: |
|
1871 | if len(id) % 2 > 0: | |
1862 | if not (id[-1] in hexdigits): |
|
1872 | if not (id[-1] in hexdigits): | |
1863 | return None |
|
1873 | return None | |
1864 | prefix = bin(id[:l]) |
|
1874 | prefix = bin(id[:l]) | |
1865 | except binascii.Error: |
|
1875 | except binascii.Error: | |
1866 | pass |
|
1876 | pass | |
1867 | else: |
|
1877 | else: | |
1868 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] |
|
1878 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] | |
1869 | nl = [ |
|
1879 | nl = [ | |
1870 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) |
|
1880 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) | |
1871 | ] |
|
1881 | ] | |
1872 | if self.nodeconstants.nullhex.startswith(id): |
|
1882 | if self.nodeconstants.nullhex.startswith(id): | |
1873 | nl.append(self.nullid) |
|
1883 | nl.append(self.nullid) | |
1874 | if len(nl) > 0: |
|
1884 | if len(nl) > 0: | |
1875 | if len(nl) == 1 and not maybewdir: |
|
1885 | if len(nl) == 1 and not maybewdir: | |
1876 | self._pcache[id] = nl[0] |
|
1886 | self._pcache[id] = nl[0] | |
1877 | return nl[0] |
|
1887 | return nl[0] | |
1878 | raise error.AmbiguousPrefixLookupError( |
|
1888 | raise error.AmbiguousPrefixLookupError( | |
1879 | id, self.display_id, _(b'ambiguous identifier') |
|
1889 | id, self.display_id, _(b'ambiguous identifier') | |
1880 | ) |
|
1890 | ) | |
1881 | if maybewdir: |
|
1891 | if maybewdir: | |
1882 | raise error.WdirUnsupported |
|
1892 | raise error.WdirUnsupported | |
1883 | return None |
|
1893 | return None | |
1884 |
|
1894 | |||
1885 | def lookup(self, id): |
|
1895 | def lookup(self, id): | |
1886 | """locate a node based on: |
|
1896 | """locate a node based on: | |
1887 | - revision number or str(revision number) |
|
1897 | - revision number or str(revision number) | |
1888 | - nodeid or subset of hex nodeid |
|
1898 | - nodeid or subset of hex nodeid | |
1889 | """ |
|
1899 | """ | |
1890 | n = self._match(id) |
|
1900 | n = self._match(id) | |
1891 | if n is not None: |
|
1901 | if n is not None: | |
1892 | return n |
|
1902 | return n | |
1893 | n = self._partialmatch(id) |
|
1903 | n = self._partialmatch(id) | |
1894 | if n: |
|
1904 | if n: | |
1895 | return n |
|
1905 | return n | |
1896 |
|
1906 | |||
1897 | raise error.LookupError(id, self.display_id, _(b'no match found')) |
|
1907 | raise error.LookupError(id, self.display_id, _(b'no match found')) | |
1898 |
|
1908 | |||
1899 | def shortest(self, node, minlength=1): |
|
1909 | def shortest(self, node, minlength=1): | |
1900 | """Find the shortest unambiguous prefix that matches node.""" |
|
1910 | """Find the shortest unambiguous prefix that matches node.""" | |
1901 |
|
1911 | |||
1902 | def isvalid(prefix): |
|
1912 | def isvalid(prefix): | |
1903 | try: |
|
1913 | try: | |
1904 | matchednode = self._partialmatch(prefix) |
|
1914 | matchednode = self._partialmatch(prefix) | |
1905 | except error.AmbiguousPrefixLookupError: |
|
1915 | except error.AmbiguousPrefixLookupError: | |
1906 | return False |
|
1916 | return False | |
1907 | except error.WdirUnsupported: |
|
1917 | except error.WdirUnsupported: | |
1908 | # single 'ff...' match |
|
1918 | # single 'ff...' match | |
1909 | return True |
|
1919 | return True | |
1910 | if matchednode is None: |
|
1920 | if matchednode is None: | |
1911 | raise error.LookupError(node, self.display_id, _(b'no node')) |
|
1921 | raise error.LookupError(node, self.display_id, _(b'no node')) | |
1912 | return True |
|
1922 | return True | |
1913 |
|
1923 | |||
1914 | def maybewdir(prefix): |
|
1924 | def maybewdir(prefix): | |
1915 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) |
|
1925 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) | |
1916 |
|
1926 | |||
1917 | hexnode = hex(node) |
|
1927 | hexnode = hex(node) | |
1918 |
|
1928 | |||
1919 | def disambiguate(hexnode, minlength): |
|
1929 | def disambiguate(hexnode, minlength): | |
1920 | """Disambiguate against wdirid.""" |
|
1930 | """Disambiguate against wdirid.""" | |
1921 | for length in range(minlength, len(hexnode) + 1): |
|
1931 | for length in range(minlength, len(hexnode) + 1): | |
1922 | prefix = hexnode[:length] |
|
1932 | prefix = hexnode[:length] | |
1923 | if not maybewdir(prefix): |
|
1933 | if not maybewdir(prefix): | |
1924 | return prefix |
|
1934 | return prefix | |
1925 |
|
1935 | |||
1926 | if not getattr(self, 'filteredrevs', None): |
|
1936 | if not getattr(self, 'filteredrevs', None): | |
1927 | try: |
|
1937 | try: | |
1928 | length = max(self.index.shortest(node), minlength) |
|
1938 | length = max(self.index.shortest(node), minlength) | |
1929 | return disambiguate(hexnode, length) |
|
1939 | return disambiguate(hexnode, length) | |
1930 | except error.RevlogError: |
|
1940 | except error.RevlogError: | |
1931 | if node != self.nodeconstants.wdirid: |
|
1941 | if node != self.nodeconstants.wdirid: | |
1932 | raise error.LookupError( |
|
1942 | raise error.LookupError( | |
1933 | node, self.display_id, _(b'no node') |
|
1943 | node, self.display_id, _(b'no node') | |
1934 | ) |
|
1944 | ) | |
1935 | except AttributeError: |
|
1945 | except AttributeError: | |
1936 | # Fall through to pure code |
|
1946 | # Fall through to pure code | |
1937 | pass |
|
1947 | pass | |
1938 |
|
1948 | |||
1939 | if node == self.nodeconstants.wdirid: |
|
1949 | if node == self.nodeconstants.wdirid: | |
1940 | for length in range(minlength, len(hexnode) + 1): |
|
1950 | for length in range(minlength, len(hexnode) + 1): | |
1941 | prefix = hexnode[:length] |
|
1951 | prefix = hexnode[:length] | |
1942 | if isvalid(prefix): |
|
1952 | if isvalid(prefix): | |
1943 | return prefix |
|
1953 | return prefix | |
1944 |
|
1954 | |||
1945 | for length in range(minlength, len(hexnode) + 1): |
|
1955 | for length in range(minlength, len(hexnode) + 1): | |
1946 | prefix = hexnode[:length] |
|
1956 | prefix = hexnode[:length] | |
1947 | if isvalid(prefix): |
|
1957 | if isvalid(prefix): | |
1948 | return disambiguate(hexnode, length) |
|
1958 | return disambiguate(hexnode, length) | |
1949 |
|
1959 | |||
1950 | def cmp(self, node, text): |
|
1960 | def cmp(self, node, text): | |
1951 | """compare text with a given file revision |
|
1961 | """compare text with a given file revision | |
1952 |
|
1962 | |||
1953 | returns True if text is different than what is stored. |
|
1963 | returns True if text is different than what is stored. | |
1954 | """ |
|
1964 | """ | |
1955 | p1, p2 = self.parents(node) |
|
1965 | p1, p2 = self.parents(node) | |
1956 | return storageutil.hashrevisionsha1(text, p1, p2) != node |
|
1966 | return storageutil.hashrevisionsha1(text, p1, p2) != node | |
1957 |
|
1967 | |||
1958 | def _getsegmentforrevs(self, startrev, endrev): |
|
1968 | def _getsegmentforrevs(self, startrev, endrev): | |
1959 | """Obtain a segment of raw data corresponding to a range of revisions. |
|
1969 | """Obtain a segment of raw data corresponding to a range of revisions. | |
1960 |
|
1970 | |||
1961 | Accepts the start and end revisions and an optional already-open |
|
1971 | Accepts the start and end revisions and an optional already-open | |
1962 | file handle to be used for reading. If the file handle is read, its |
|
1972 | file handle to be used for reading. If the file handle is read, its | |
1963 | seek position will not be preserved. |
|
1973 | seek position will not be preserved. | |
1964 |
|
1974 | |||
1965 | Requests for data may be satisfied by a cache. |
|
1975 | Requests for data may be satisfied by a cache. | |
1966 |
|
1976 | |||
1967 | Returns a 2-tuple of (offset, data) for the requested range of |
|
1977 | Returns a 2-tuple of (offset, data) for the requested range of | |
1968 | revisions. Offset is the integer offset from the beginning of the |
|
1978 | revisions. Offset is the integer offset from the beginning of the | |
1969 | revlog and data is a str or buffer of the raw byte data. |
|
1979 | revlog and data is a str or buffer of the raw byte data. | |
1970 |
|
1980 | |||
1971 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` |
|
1981 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` | |
1972 | to determine where each revision's data begins and ends. |
|
1982 | to determine where each revision's data begins and ends. | |
1973 | """ |
|
1983 | """ | |
1974 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons |
|
1984 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons | |
1975 | # (functions are expensive). |
|
1985 | # (functions are expensive). | |
1976 | index = self.index |
|
1986 | index = self.index | |
1977 | istart = index[startrev] |
|
1987 | istart = index[startrev] | |
1978 | start = int(istart[0] >> 16) |
|
1988 | start = int(istart[0] >> 16) | |
1979 | if startrev == endrev: |
|
1989 | if startrev == endrev: | |
1980 | end = start + istart[1] |
|
1990 | end = start + istart[1] | |
1981 | else: |
|
1991 | else: | |
1982 | iend = index[endrev] |
|
1992 | iend = index[endrev] | |
1983 | end = int(iend[0] >> 16) + iend[1] |
|
1993 | end = int(iend[0] >> 16) + iend[1] | |
1984 |
|
1994 | |||
1985 | if self._inline: |
|
1995 | if self._inline: | |
1986 | start += (startrev + 1) * self.index.entry_size |
|
1996 | start += (startrev + 1) * self.index.entry_size | |
1987 | end += (endrev + 1) * self.index.entry_size |
|
1997 | end += (endrev + 1) * self.index.entry_size | |
1988 | length = end - start |
|
1998 | length = end - start | |
1989 |
|
1999 | |||
1990 | return start, self._segmentfile.read_chunk(start, length) |
|
2000 | return start, self._segmentfile.read_chunk(start, length) | |
1991 |
|
2001 | |||
1992 | def _chunk(self, rev): |
|
2002 | def _chunk(self, rev): | |
1993 | """Obtain a single decompressed chunk for a revision. |
|
2003 | """Obtain a single decompressed chunk for a revision. | |
1994 |
|
2004 | |||
1995 | Accepts an integer revision and an optional already-open file handle |
|
2005 | Accepts an integer revision and an optional already-open file handle | |
1996 | to be used for reading. If used, the seek position of the file will not |
|
2006 | to be used for reading. If used, the seek position of the file will not | |
1997 | be preserved. |
|
2007 | be preserved. | |
1998 |
|
2008 | |||
1999 | Returns a str holding uncompressed data for the requested revision. |
|
2009 | Returns a str holding uncompressed data for the requested revision. | |
2000 | """ |
|
2010 | """ | |
2001 | compression_mode = self.index[rev][10] |
|
2011 | compression_mode = self.index[rev][10] | |
2002 | data = self._getsegmentforrevs(rev, rev)[1] |
|
2012 | data = self._getsegmentforrevs(rev, rev)[1] | |
2003 | if compression_mode == COMP_MODE_PLAIN: |
|
2013 | if compression_mode == COMP_MODE_PLAIN: | |
2004 | return data |
|
2014 | return data | |
2005 | elif compression_mode == COMP_MODE_DEFAULT: |
|
2015 | elif compression_mode == COMP_MODE_DEFAULT: | |
2006 | return self._decompressor(data) |
|
2016 | return self._decompressor(data) | |
2007 | elif compression_mode == COMP_MODE_INLINE: |
|
2017 | elif compression_mode == COMP_MODE_INLINE: | |
2008 | return self.decompress(data) |
|
2018 | return self.decompress(data) | |
2009 | else: |
|
2019 | else: | |
2010 | msg = b'unknown compression mode %d' |
|
2020 | msg = b'unknown compression mode %d' | |
2011 | msg %= compression_mode |
|
2021 | msg %= compression_mode | |
2012 | raise error.RevlogError(msg) |
|
2022 | raise error.RevlogError(msg) | |
2013 |
|
2023 | |||
2014 | def _chunks(self, revs, targetsize=None): |
|
2024 | def _chunks(self, revs, targetsize=None): | |
2015 | """Obtain decompressed chunks for the specified revisions. |
|
2025 | """Obtain decompressed chunks for the specified revisions. | |
2016 |
|
2026 | |||
2017 | Accepts an iterable of numeric revisions that are assumed to be in |
|
2027 | Accepts an iterable of numeric revisions that are assumed to be in | |
2018 | ascending order. Also accepts an optional already-open file handle |
|
2028 | ascending order. Also accepts an optional already-open file handle | |
2019 | to be used for reading. If used, the seek position of the file will |
|
2029 | to be used for reading. If used, the seek position of the file will | |
2020 | not be preserved. |
|
2030 | not be preserved. | |
2021 |
|
2031 | |||
2022 | This function is similar to calling ``self._chunk()`` multiple times, |
|
2032 | This function is similar to calling ``self._chunk()`` multiple times, | |
2023 | but is faster. |
|
2033 | but is faster. | |
2024 |
|
2034 | |||
2025 | Returns a list with decompressed data for each requested revision. |
|
2035 | Returns a list with decompressed data for each requested revision. | |
2026 | """ |
|
2036 | """ | |
2027 | if not revs: |
|
2037 | if not revs: | |
2028 | return [] |
|
2038 | return [] | |
2029 | start = self.start |
|
2039 | start = self.start | |
2030 | length = self.length |
|
2040 | length = self.length | |
2031 | inline = self._inline |
|
2041 | inline = self._inline | |
2032 | iosize = self.index.entry_size |
|
2042 | iosize = self.index.entry_size | |
2033 | buffer = util.buffer |
|
2043 | buffer = util.buffer | |
2034 |
|
2044 | |||
2035 | l = [] |
|
2045 | l = [] | |
2036 | ladd = l.append |
|
2046 | ladd = l.append | |
2037 |
|
2047 | |||
2038 | if not self._withsparseread: |
|
2048 | if not self._withsparseread: | |
2039 | slicedchunks = (revs,) |
|
2049 | slicedchunks = (revs,) | |
2040 | else: |
|
2050 | else: | |
2041 | slicedchunks = deltautil.slicechunk( |
|
2051 | slicedchunks = deltautil.slicechunk( | |
2042 | self, revs, targetsize=targetsize |
|
2052 | self, revs, targetsize=targetsize | |
2043 | ) |
|
2053 | ) | |
2044 |
|
2054 | |||
2045 | for revschunk in slicedchunks: |
|
2055 | for revschunk in slicedchunks: | |
2046 | firstrev = revschunk[0] |
|
2056 | firstrev = revschunk[0] | |
2047 | # Skip trailing revisions with empty diff |
|
2057 | # Skip trailing revisions with empty diff | |
2048 | for lastrev in revschunk[::-1]: |
|
2058 | for lastrev in revschunk[::-1]: | |
2049 | if length(lastrev) != 0: |
|
2059 | if length(lastrev) != 0: | |
2050 | break |
|
2060 | break | |
2051 |
|
2061 | |||
2052 | try: |
|
2062 | try: | |
2053 | offset, data = self._getsegmentforrevs(firstrev, lastrev) |
|
2063 | offset, data = self._getsegmentforrevs(firstrev, lastrev) | |
2054 | except OverflowError: |
|
2064 | except OverflowError: | |
2055 | # issue4215 - we can't cache a run of chunks greater than |
|
2065 | # issue4215 - we can't cache a run of chunks greater than | |
2056 | # 2G on Windows |
|
2066 | # 2G on Windows | |
2057 | return [self._chunk(rev) for rev in revschunk] |
|
2067 | return [self._chunk(rev) for rev in revschunk] | |
2058 |
|
2068 | |||
2059 | decomp = self.decompress |
|
2069 | decomp = self.decompress | |
2060 | # self._decompressor might be None, but will not be used in that case |
|
2070 | # self._decompressor might be None, but will not be used in that case | |
2061 | def_decomp = self._decompressor |
|
2071 | def_decomp = self._decompressor | |
2062 | for rev in revschunk: |
|
2072 | for rev in revschunk: | |
2063 | chunkstart = start(rev) |
|
2073 | chunkstart = start(rev) | |
2064 | if inline: |
|
2074 | if inline: | |
2065 | chunkstart += (rev + 1) * iosize |
|
2075 | chunkstart += (rev + 1) * iosize | |
2066 | chunklength = length(rev) |
|
2076 | chunklength = length(rev) | |
2067 | comp_mode = self.index[rev][10] |
|
2077 | comp_mode = self.index[rev][10] | |
2068 | c = buffer(data, chunkstart - offset, chunklength) |
|
2078 | c = buffer(data, chunkstart - offset, chunklength) | |
2069 | if comp_mode == COMP_MODE_PLAIN: |
|
2079 | if comp_mode == COMP_MODE_PLAIN: | |
2070 | ladd(c) |
|
2080 | ladd(c) | |
2071 | elif comp_mode == COMP_MODE_INLINE: |
|
2081 | elif comp_mode == COMP_MODE_INLINE: | |
2072 | ladd(decomp(c)) |
|
2082 | ladd(decomp(c)) | |
2073 | elif comp_mode == COMP_MODE_DEFAULT: |
|
2083 | elif comp_mode == COMP_MODE_DEFAULT: | |
2074 | ladd(def_decomp(c)) |
|
2084 | ladd(def_decomp(c)) | |
2075 | else: |
|
2085 | else: | |
2076 | msg = b'unknown compression mode %d' |
|
2086 | msg = b'unknown compression mode %d' | |
2077 | msg %= comp_mode |
|
2087 | msg %= comp_mode | |
2078 | raise error.RevlogError(msg) |
|
2088 | raise error.RevlogError(msg) | |
2079 |
|
2089 | |||
2080 | return l |
|
2090 | return l | |
2081 |
|
2091 | |||
2082 | def deltaparent(self, rev): |
|
2092 | def deltaparent(self, rev): | |
2083 | """return deltaparent of the given revision""" |
|
2093 | """return deltaparent of the given revision""" | |
2084 | base = self.index[rev][3] |
|
2094 | base = self.index[rev][3] | |
2085 | if base == rev: |
|
2095 | if base == rev: | |
2086 | return nullrev |
|
2096 | return nullrev | |
2087 | elif self._generaldelta: |
|
2097 | elif self._generaldelta: | |
2088 | return base |
|
2098 | return base | |
2089 | else: |
|
2099 | else: | |
2090 | return rev - 1 |
|
2100 | return rev - 1 | |
2091 |
|
2101 | |||
2092 | def issnapshot(self, rev): |
|
2102 | def issnapshot(self, rev): | |
2093 | """tells whether rev is a snapshot""" |
|
2103 | """tells whether rev is a snapshot""" | |
2094 | if not self._sparserevlog: |
|
2104 | if not self._sparserevlog: | |
2095 | return self.deltaparent(rev) == nullrev |
|
2105 | return self.deltaparent(rev) == nullrev | |
2096 | elif hasattr(self.index, 'issnapshot'): |
|
2106 | elif hasattr(self.index, 'issnapshot'): | |
2097 | # directly assign the method to cache the testing and access |
|
2107 | # directly assign the method to cache the testing and access | |
2098 | self.issnapshot = self.index.issnapshot |
|
2108 | self.issnapshot = self.index.issnapshot | |
2099 | return self.issnapshot(rev) |
|
2109 | return self.issnapshot(rev) | |
2100 | if rev == nullrev: |
|
2110 | if rev == nullrev: | |
2101 | return True |
|
2111 | return True | |
2102 | entry = self.index[rev] |
|
2112 | entry = self.index[rev] | |
2103 | base = entry[3] |
|
2113 | base = entry[3] | |
2104 | if base == rev: |
|
2114 | if base == rev: | |
2105 | return True |
|
2115 | return True | |
2106 | if base == nullrev: |
|
2116 | if base == nullrev: | |
2107 | return True |
|
2117 | return True | |
2108 | p1 = entry[5] |
|
2118 | p1 = entry[5] | |
2109 | while self.length(p1) == 0: |
|
2119 | while self.length(p1) == 0: | |
2110 | b = self.deltaparent(p1) |
|
2120 | b = self.deltaparent(p1) | |
2111 | if b == p1: |
|
2121 | if b == p1: | |
2112 | break |
|
2122 | break | |
2113 | p1 = b |
|
2123 | p1 = b | |
2114 | p2 = entry[6] |
|
2124 | p2 = entry[6] | |
2115 | while self.length(p2) == 0: |
|
2125 | while self.length(p2) == 0: | |
2116 | b = self.deltaparent(p2) |
|
2126 | b = self.deltaparent(p2) | |
2117 | if b == p2: |
|
2127 | if b == p2: | |
2118 | break |
|
2128 | break | |
2119 | p2 = b |
|
2129 | p2 = b | |
2120 | if base == p1 or base == p2: |
|
2130 | if base == p1 or base == p2: | |
2121 | return False |
|
2131 | return False | |
2122 | return self.issnapshot(base) |
|
2132 | return self.issnapshot(base) | |
2123 |
|
2133 | |||
2124 | def snapshotdepth(self, rev): |
|
2134 | def snapshotdepth(self, rev): | |
2125 | """number of snapshot in the chain before this one""" |
|
2135 | """number of snapshot in the chain before this one""" | |
2126 | if not self.issnapshot(rev): |
|
2136 | if not self.issnapshot(rev): | |
2127 | raise error.ProgrammingError(b'revision %d not a snapshot') |
|
2137 | raise error.ProgrammingError(b'revision %d not a snapshot') | |
2128 | return len(self._deltachain(rev)[0]) - 1 |
|
2138 | return len(self._deltachain(rev)[0]) - 1 | |
2129 |
|
2139 | |||
2130 | def revdiff(self, rev1, rev2): |
|
2140 | def revdiff(self, rev1, rev2): | |
2131 | """return or calculate a delta between two revisions |
|
2141 | """return or calculate a delta between two revisions | |
2132 |
|
2142 | |||
2133 | The delta calculated is in binary form and is intended to be written to |
|
2143 | The delta calculated is in binary form and is intended to be written to | |
2134 | revlog data directly. So this function needs raw revision data. |
|
2144 | revlog data directly. So this function needs raw revision data. | |
2135 | """ |
|
2145 | """ | |
2136 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: |
|
2146 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: | |
2137 | return bytes(self._chunk(rev2)) |
|
2147 | return bytes(self._chunk(rev2)) | |
2138 |
|
2148 | |||
2139 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) |
|
2149 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) | |
2140 |
|
2150 | |||
2141 | def revision(self, nodeorrev): |
|
2151 | def revision(self, nodeorrev): | |
2142 | """return an uncompressed revision of a given node or revision |
|
2152 | """return an uncompressed revision of a given node or revision | |
2143 | number. |
|
2153 | number. | |
2144 | """ |
|
2154 | """ | |
2145 | return self._revisiondata(nodeorrev) |
|
2155 | return self._revisiondata(nodeorrev) | |
2146 |
|
2156 | |||
2147 | def sidedata(self, nodeorrev): |
|
2157 | def sidedata(self, nodeorrev): | |
2148 | """a map of extra data related to the changeset but not part of the hash |
|
2158 | """a map of extra data related to the changeset but not part of the hash | |
2149 |
|
2159 | |||
2150 | This function currently return a dictionary. However, more advanced |
|
2160 | This function currently return a dictionary. However, more advanced | |
2151 | mapping object will likely be used in the future for a more |
|
2161 | mapping object will likely be used in the future for a more | |
2152 | efficient/lazy code. |
|
2162 | efficient/lazy code. | |
2153 | """ |
|
2163 | """ | |
2154 | # deal with <nodeorrev> argument type |
|
2164 | # deal with <nodeorrev> argument type | |
2155 | if isinstance(nodeorrev, int): |
|
2165 | if isinstance(nodeorrev, int): | |
2156 | rev = nodeorrev |
|
2166 | rev = nodeorrev | |
2157 | else: |
|
2167 | else: | |
2158 | rev = self.rev(nodeorrev) |
|
2168 | rev = self.rev(nodeorrev) | |
2159 | return self._sidedata(rev) |
|
2169 | return self._sidedata(rev) | |
2160 |
|
2170 | |||
2161 | def _revisiondata(self, nodeorrev, raw=False): |
|
2171 | def _revisiondata(self, nodeorrev, raw=False): | |
2162 | # deal with <nodeorrev> argument type |
|
2172 | # deal with <nodeorrev> argument type | |
2163 | if isinstance(nodeorrev, int): |
|
2173 | if isinstance(nodeorrev, int): | |
2164 | rev = nodeorrev |
|
2174 | rev = nodeorrev | |
2165 | node = self.node(rev) |
|
2175 | node = self.node(rev) | |
2166 | else: |
|
2176 | else: | |
2167 | node = nodeorrev |
|
2177 | node = nodeorrev | |
2168 | rev = None |
|
2178 | rev = None | |
2169 |
|
2179 | |||
2170 | # fast path the special `nullid` rev |
|
2180 | # fast path the special `nullid` rev | |
2171 | if node == self.nullid: |
|
2181 | if node == self.nullid: | |
2172 | return b"" |
|
2182 | return b"" | |
2173 |
|
2183 | |||
2174 | # ``rawtext`` is the text as stored inside the revlog. Might be the |
|
2184 | # ``rawtext`` is the text as stored inside the revlog. Might be the | |
2175 | # revision or might need to be processed to retrieve the revision. |
|
2185 | # revision or might need to be processed to retrieve the revision. | |
2176 | rev, rawtext, validated = self._rawtext(node, rev) |
|
2186 | rev, rawtext, validated = self._rawtext(node, rev) | |
2177 |
|
2187 | |||
2178 | if raw and validated: |
|
2188 | if raw and validated: | |
2179 | # if we don't want to process the raw text and that raw |
|
2189 | # if we don't want to process the raw text and that raw | |
2180 | # text is cached, we can exit early. |
|
2190 | # text is cached, we can exit early. | |
2181 | return rawtext |
|
2191 | return rawtext | |
2182 | if rev is None: |
|
2192 | if rev is None: | |
2183 | rev = self.rev(node) |
|
2193 | rev = self.rev(node) | |
2184 | # the revlog's flag for this revision |
|
2194 | # the revlog's flag for this revision | |
2185 | # (usually alter its state or content) |
|
2195 | # (usually alter its state or content) | |
2186 | flags = self.flags(rev) |
|
2196 | flags = self.flags(rev) | |
2187 |
|
2197 | |||
2188 | if validated and flags == REVIDX_DEFAULT_FLAGS: |
|
2198 | if validated and flags == REVIDX_DEFAULT_FLAGS: | |
2189 | # no extra flags set, no flag processor runs, text = rawtext |
|
2199 | # no extra flags set, no flag processor runs, text = rawtext | |
2190 | return rawtext |
|
2200 | return rawtext | |
2191 |
|
2201 | |||
2192 | if raw: |
|
2202 | if raw: | |
2193 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
2203 | validatehash = flagutil.processflagsraw(self, rawtext, flags) | |
2194 | text = rawtext |
|
2204 | text = rawtext | |
2195 | else: |
|
2205 | else: | |
2196 | r = flagutil.processflagsread(self, rawtext, flags) |
|
2206 | r = flagutil.processflagsread(self, rawtext, flags) | |
2197 | text, validatehash = r |
|
2207 | text, validatehash = r | |
2198 | if validatehash: |
|
2208 | if validatehash: | |
2199 | self.checkhash(text, node, rev=rev) |
|
2209 | self.checkhash(text, node, rev=rev) | |
2200 | if not validated: |
|
2210 | if not validated: | |
2201 | self._revisioncache = (node, rev, rawtext) |
|
2211 | self._revisioncache = (node, rev, rawtext) | |
2202 |
|
2212 | |||
2203 | return text |
|
2213 | return text | |
2204 |
|
2214 | |||
2205 | def _rawtext(self, node, rev): |
|
2215 | def _rawtext(self, node, rev): | |
2206 | """return the possibly unvalidated rawtext for a revision |
|
2216 | """return the possibly unvalidated rawtext for a revision | |
2207 |
|
2217 | |||
2208 | returns (rev, rawtext, validated) |
|
2218 | returns (rev, rawtext, validated) | |
2209 | """ |
|
2219 | """ | |
2210 |
|
2220 | |||
2211 | # revision in the cache (could be useful to apply delta) |
|
2221 | # revision in the cache (could be useful to apply delta) | |
2212 | cachedrev = None |
|
2222 | cachedrev = None | |
2213 | # An intermediate text to apply deltas to |
|
2223 | # An intermediate text to apply deltas to | |
2214 | basetext = None |
|
2224 | basetext = None | |
2215 |
|
2225 | |||
2216 | # Check if we have the entry in cache |
|
2226 | # Check if we have the entry in cache | |
2217 | # The cache entry looks like (node, rev, rawtext) |
|
2227 | # The cache entry looks like (node, rev, rawtext) | |
2218 | if self._revisioncache: |
|
2228 | if self._revisioncache: | |
2219 | if self._revisioncache[0] == node: |
|
2229 | if self._revisioncache[0] == node: | |
2220 | return (rev, self._revisioncache[2], True) |
|
2230 | return (rev, self._revisioncache[2], True) | |
2221 | cachedrev = self._revisioncache[1] |
|
2231 | cachedrev = self._revisioncache[1] | |
2222 |
|
2232 | |||
2223 | if rev is None: |
|
2233 | if rev is None: | |
2224 | rev = self.rev(node) |
|
2234 | rev = self.rev(node) | |
2225 |
|
2235 | |||
2226 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) |
|
2236 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) | |
2227 | if stopped: |
|
2237 | if stopped: | |
2228 | basetext = self._revisioncache[2] |
|
2238 | basetext = self._revisioncache[2] | |
2229 |
|
2239 | |||
2230 | # drop cache to save memory, the caller is expected to |
|
2240 | # drop cache to save memory, the caller is expected to | |
2231 | # update self._revisioncache after validating the text |
|
2241 | # update self._revisioncache after validating the text | |
2232 | self._revisioncache = None |
|
2242 | self._revisioncache = None | |
2233 |
|
2243 | |||
2234 | targetsize = None |
|
2244 | targetsize = None | |
2235 | rawsize = self.index[rev][2] |
|
2245 | rawsize = self.index[rev][2] | |
2236 | if 0 <= rawsize: |
|
2246 | if 0 <= rawsize: | |
2237 | targetsize = 4 * rawsize |
|
2247 | targetsize = 4 * rawsize | |
2238 |
|
2248 | |||
2239 | bins = self._chunks(chain, targetsize=targetsize) |
|
2249 | bins = self._chunks(chain, targetsize=targetsize) | |
2240 | if basetext is None: |
|
2250 | if basetext is None: | |
2241 | basetext = bytes(bins[0]) |
|
2251 | basetext = bytes(bins[0]) | |
2242 | bins = bins[1:] |
|
2252 | bins = bins[1:] | |
2243 |
|
2253 | |||
2244 | rawtext = mdiff.patches(basetext, bins) |
|
2254 | rawtext = mdiff.patches(basetext, bins) | |
2245 | del basetext # let us have a chance to free memory early |
|
2255 | del basetext # let us have a chance to free memory early | |
2246 | return (rev, rawtext, False) |
|
2256 | return (rev, rawtext, False) | |
2247 |
|
2257 | |||
2248 | def _sidedata(self, rev): |
|
2258 | def _sidedata(self, rev): | |
2249 | """Return the sidedata for a given revision number.""" |
|
2259 | """Return the sidedata for a given revision number.""" | |
2250 | index_entry = self.index[rev] |
|
2260 | index_entry = self.index[rev] | |
2251 | sidedata_offset = index_entry[8] |
|
2261 | sidedata_offset = index_entry[8] | |
2252 | sidedata_size = index_entry[9] |
|
2262 | sidedata_size = index_entry[9] | |
2253 |
|
2263 | |||
2254 | if self._inline: |
|
2264 | if self._inline: | |
2255 | sidedata_offset += self.index.entry_size * (1 + rev) |
|
2265 | sidedata_offset += self.index.entry_size * (1 + rev) | |
2256 | if sidedata_size == 0: |
|
2266 | if sidedata_size == 0: | |
2257 | return {} |
|
2267 | return {} | |
2258 |
|
2268 | |||
2259 | if self._docket.sidedata_end < sidedata_offset + sidedata_size: |
|
2269 | if self._docket.sidedata_end < sidedata_offset + sidedata_size: | |
2260 | filename = self._sidedatafile |
|
2270 | filename = self._sidedatafile | |
2261 | end = self._docket.sidedata_end |
|
2271 | end = self._docket.sidedata_end | |
2262 | offset = sidedata_offset |
|
2272 | offset = sidedata_offset | |
2263 | length = sidedata_size |
|
2273 | length = sidedata_size | |
2264 | m = FILE_TOO_SHORT_MSG % (filename, length, offset, end) |
|
2274 | m = FILE_TOO_SHORT_MSG % (filename, length, offset, end) | |
2265 | raise error.RevlogError(m) |
|
2275 | raise error.RevlogError(m) | |
2266 |
|
2276 | |||
2267 | comp_segment = self._segmentfile_sidedata.read_chunk( |
|
2277 | comp_segment = self._segmentfile_sidedata.read_chunk( | |
2268 | sidedata_offset, sidedata_size |
|
2278 | sidedata_offset, sidedata_size | |
2269 | ) |
|
2279 | ) | |
2270 |
|
2280 | |||
2271 | comp = self.index[rev][11] |
|
2281 | comp = self.index[rev][11] | |
2272 | if comp == COMP_MODE_PLAIN: |
|
2282 | if comp == COMP_MODE_PLAIN: | |
2273 | segment = comp_segment |
|
2283 | segment = comp_segment | |
2274 | elif comp == COMP_MODE_DEFAULT: |
|
2284 | elif comp == COMP_MODE_DEFAULT: | |
2275 | segment = self._decompressor(comp_segment) |
|
2285 | segment = self._decompressor(comp_segment) | |
2276 | elif comp == COMP_MODE_INLINE: |
|
2286 | elif comp == COMP_MODE_INLINE: | |
2277 | segment = self.decompress(comp_segment) |
|
2287 | segment = self.decompress(comp_segment) | |
2278 | else: |
|
2288 | else: | |
2279 | msg = b'unknown compression mode %d' |
|
2289 | msg = b'unknown compression mode %d' | |
2280 | msg %= comp |
|
2290 | msg %= comp | |
2281 | raise error.RevlogError(msg) |
|
2291 | raise error.RevlogError(msg) | |
2282 |
|
2292 | |||
2283 | sidedata = sidedatautil.deserialize_sidedata(segment) |
|
2293 | sidedata = sidedatautil.deserialize_sidedata(segment) | |
2284 | return sidedata |
|
2294 | return sidedata | |
2285 |
|
2295 | |||
2286 | def rawdata(self, nodeorrev): |
|
2296 | def rawdata(self, nodeorrev): | |
2287 | """return an uncompressed raw data of a given node or revision number.""" |
|
2297 | """return an uncompressed raw data of a given node or revision number.""" | |
2288 | return self._revisiondata(nodeorrev, raw=True) |
|
2298 | return self._revisiondata(nodeorrev, raw=True) | |
2289 |
|
2299 | |||
2290 | def hash(self, text, p1, p2): |
|
2300 | def hash(self, text, p1, p2): | |
2291 | """Compute a node hash. |
|
2301 | """Compute a node hash. | |
2292 |
|
2302 | |||
2293 | Available as a function so that subclasses can replace the hash |
|
2303 | Available as a function so that subclasses can replace the hash | |
2294 | as needed. |
|
2304 | as needed. | |
2295 | """ |
|
2305 | """ | |
2296 | return storageutil.hashrevisionsha1(text, p1, p2) |
|
2306 | return storageutil.hashrevisionsha1(text, p1, p2) | |
2297 |
|
2307 | |||
2298 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
2308 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
2299 | """Check node hash integrity. |
|
2309 | """Check node hash integrity. | |
2300 |
|
2310 | |||
2301 | Available as a function so that subclasses can extend hash mismatch |
|
2311 | Available as a function so that subclasses can extend hash mismatch | |
2302 | behaviors as needed. |
|
2312 | behaviors as needed. | |
2303 | """ |
|
2313 | """ | |
2304 | try: |
|
2314 | try: | |
2305 | if p1 is None and p2 is None: |
|
2315 | if p1 is None and p2 is None: | |
2306 | p1, p2 = self.parents(node) |
|
2316 | p1, p2 = self.parents(node) | |
2307 | if node != self.hash(text, p1, p2): |
|
2317 | if node != self.hash(text, p1, p2): | |
2308 | # Clear the revision cache on hash failure. The revision cache |
|
2318 | # Clear the revision cache on hash failure. The revision cache | |
2309 | # only stores the raw revision and clearing the cache does have |
|
2319 | # only stores the raw revision and clearing the cache does have | |
2310 | # the side-effect that we won't have a cache hit when the raw |
|
2320 | # the side-effect that we won't have a cache hit when the raw | |
2311 | # revision data is accessed. But this case should be rare and |
|
2321 | # revision data is accessed. But this case should be rare and | |
2312 | # it is extra work to teach the cache about the hash |
|
2322 | # it is extra work to teach the cache about the hash | |
2313 | # verification state. |
|
2323 | # verification state. | |
2314 | if self._revisioncache and self._revisioncache[0] == node: |
|
2324 | if self._revisioncache and self._revisioncache[0] == node: | |
2315 | self._revisioncache = None |
|
2325 | self._revisioncache = None | |
2316 |
|
2326 | |||
2317 | revornode = rev |
|
2327 | revornode = rev | |
2318 | if revornode is None: |
|
2328 | if revornode is None: | |
2319 | revornode = templatefilters.short(hex(node)) |
|
2329 | revornode = templatefilters.short(hex(node)) | |
2320 | raise error.RevlogError( |
|
2330 | raise error.RevlogError( | |
2321 | _(b"integrity check failed on %s:%s") |
|
2331 | _(b"integrity check failed on %s:%s") | |
2322 | % (self.display_id, pycompat.bytestr(revornode)) |
|
2332 | % (self.display_id, pycompat.bytestr(revornode)) | |
2323 | ) |
|
2333 | ) | |
2324 | except error.RevlogError: |
|
2334 | except error.RevlogError: | |
2325 | if self._censorable and storageutil.iscensoredtext(text): |
|
2335 | if self._censorable and storageutil.iscensoredtext(text): | |
2326 | raise error.CensoredNodeError(self.display_id, node, text) |
|
2336 | raise error.CensoredNodeError(self.display_id, node, text) | |
2327 | raise |
|
2337 | raise | |
2328 |
|
2338 | |||
2329 | @property |
|
2339 | @property | |
2330 | def _split_index_file(self): |
|
2340 | def _split_index_file(self): | |
2331 | """the path where to expect the index of an ongoing splitting operation |
|
2341 | """the path where to expect the index of an ongoing splitting operation | |
2332 |
|
2342 | |||
2333 | The file will only exist if a splitting operation is in progress, but |
|
2343 | The file will only exist if a splitting operation is in progress, but | |
2334 | it is always expected at the same location.""" |
|
2344 | it is always expected at the same location.""" | |
2335 | parts = self.radix.split(b'/') |
|
2345 | parts = self.radix.split(b'/') | |
2336 | if len(parts) > 1: |
|
2346 | if len(parts) > 1: | |
2337 | # adds a '-s' prefix to the ``data/` or `meta/` base |
|
2347 | # adds a '-s' prefix to the ``data/` or `meta/` base | |
2338 | head = parts[0] + b'-s' |
|
2348 | head = parts[0] + b'-s' | |
2339 | mids = parts[1:-1] |
|
2349 | mids = parts[1:-1] | |
2340 | tail = parts[-1] + b'.i' |
|
2350 | tail = parts[-1] + b'.i' | |
2341 | pieces = [head] + mids + [tail] |
|
2351 | pieces = [head] + mids + [tail] | |
2342 | return b'/'.join(pieces) |
|
2352 | return b'/'.join(pieces) | |
2343 | else: |
|
2353 | else: | |
2344 | # the revlog is stored at the root of the store (changelog or |
|
2354 | # the revlog is stored at the root of the store (changelog or | |
2345 | # manifest), no risk of collision. |
|
2355 | # manifest), no risk of collision. | |
2346 | return self.radix + b'.i.s' |
|
2356 | return self.radix + b'.i.s' | |
2347 |
|
2357 | |||
2348 | def _enforceinlinesize(self, tr, side_write=True): |
|
2358 | def _enforceinlinesize(self, tr, side_write=True): | |
2349 | """Check if the revlog is too big for inline and convert if so. |
|
2359 | """Check if the revlog is too big for inline and convert if so. | |
2350 |
|
2360 | |||
2351 | This should be called after revisions are added to the revlog. If the |
|
2361 | This should be called after revisions are added to the revlog. If the | |
2352 | revlog has grown too large to be an inline revlog, it will convert it |
|
2362 | revlog has grown too large to be an inline revlog, it will convert it | |
2353 | to use multiple index and data files. |
|
2363 | to use multiple index and data files. | |
2354 | """ |
|
2364 | """ | |
2355 | tiprev = len(self) - 1 |
|
2365 | tiprev = len(self) - 1 | |
2356 | total_size = self.start(tiprev) + self.length(tiprev) |
|
2366 | total_size = self.start(tiprev) + self.length(tiprev) | |
2357 | if not self._inline or total_size < _maxinline: |
|
2367 | if not self._inline or total_size < _maxinline: | |
2358 | return |
|
2368 | return | |
2359 |
|
2369 | |||
2360 | troffset = tr.findoffset(self._indexfile) |
|
2370 | troffset = tr.findoffset(self._indexfile) | |
2361 | if troffset is None: |
|
2371 | if troffset is None: | |
2362 | raise error.RevlogError( |
|
2372 | raise error.RevlogError( | |
2363 | _(b"%s not found in the transaction") % self._indexfile |
|
2373 | _(b"%s not found in the transaction") % self._indexfile | |
2364 | ) |
|
2374 | ) | |
2365 | if troffset: |
|
2375 | if troffset: | |
2366 | tr.addbackup(self._indexfile, for_offset=True) |
|
2376 | tr.addbackup(self._indexfile, for_offset=True) | |
2367 | tr.add(self._datafile, 0) |
|
2377 | tr.add(self._datafile, 0) | |
2368 |
|
2378 | |||
2369 | existing_handles = False |
|
2379 | existing_handles = False | |
2370 | if self._writinghandles is not None: |
|
2380 | if self._writinghandles is not None: | |
2371 | existing_handles = True |
|
2381 | existing_handles = True | |
2372 | fp = self._writinghandles[0] |
|
2382 | fp = self._writinghandles[0] | |
2373 | fp.flush() |
|
2383 | fp.flush() | |
2374 | fp.close() |
|
2384 | fp.close() | |
2375 | # We can't use the cached file handle after close(). So prevent |
|
2385 | # We can't use the cached file handle after close(). So prevent | |
2376 | # its usage. |
|
2386 | # its usage. | |
2377 | self._writinghandles = None |
|
2387 | self._writinghandles = None | |
2378 | self._segmentfile.writing_handle = None |
|
2388 | self._segmentfile.writing_handle = None | |
2379 | # No need to deal with sidedata writing handle as it is only |
|
2389 | # No need to deal with sidedata writing handle as it is only | |
2380 | # relevant with revlog-v2 which is never inline, not reaching |
|
2390 | # relevant with revlog-v2 which is never inline, not reaching | |
2381 | # this code |
|
2391 | # this code | |
2382 | if side_write: |
|
2392 | if side_write: | |
2383 | old_index_file_path = self._indexfile |
|
2393 | old_index_file_path = self._indexfile | |
2384 | new_index_file_path = self._split_index_file |
|
2394 | new_index_file_path = self._split_index_file | |
2385 | opener = self.opener |
|
2395 | opener = self.opener | |
2386 | weak_self = weakref.ref(self) |
|
2396 | weak_self = weakref.ref(self) | |
2387 |
|
2397 | |||
2388 | # the "split" index replace the real index when the transaction is finalized |
|
2398 | # the "split" index replace the real index when the transaction is finalized | |
2389 | def finalize_callback(tr): |
|
2399 | def finalize_callback(tr): | |
2390 | opener.rename( |
|
2400 | opener.rename( | |
2391 | new_index_file_path, |
|
2401 | new_index_file_path, | |
2392 | old_index_file_path, |
|
2402 | old_index_file_path, | |
2393 | checkambig=True, |
|
2403 | checkambig=True, | |
2394 | ) |
|
2404 | ) | |
2395 | maybe_self = weak_self() |
|
2405 | maybe_self = weak_self() | |
2396 | if maybe_self is not None: |
|
2406 | if maybe_self is not None: | |
2397 | maybe_self._indexfile = old_index_file_path |
|
2407 | maybe_self._indexfile = old_index_file_path | |
2398 |
|
2408 | |||
2399 | def abort_callback(tr): |
|
2409 | def abort_callback(tr): | |
2400 | maybe_self = weak_self() |
|
2410 | maybe_self = weak_self() | |
2401 | if maybe_self is not None: |
|
2411 | if maybe_self is not None: | |
2402 | maybe_self._indexfile = old_index_file_path |
|
2412 | maybe_self._indexfile = old_index_file_path | |
2403 |
|
2413 | |||
2404 | tr.registertmp(new_index_file_path) |
|
2414 | tr.registertmp(new_index_file_path) | |
2405 | if self.target[1] is not None: |
|
2415 | if self.target[1] is not None: | |
2406 | callback_id = b'000-revlog-split-%d-%s' % self.target |
|
2416 | callback_id = b'000-revlog-split-%d-%s' % self.target | |
2407 | else: |
|
2417 | else: | |
2408 | callback_id = b'000-revlog-split-%d' % self.target[0] |
|
2418 | callback_id = b'000-revlog-split-%d' % self.target[0] | |
2409 | tr.addfinalize(callback_id, finalize_callback) |
|
2419 | tr.addfinalize(callback_id, finalize_callback) | |
2410 | tr.addabort(callback_id, abort_callback) |
|
2420 | tr.addabort(callback_id, abort_callback) | |
2411 |
|
2421 | |||
2412 | new_dfh = self._datafp(b'w+') |
|
2422 | new_dfh = self._datafp(b'w+') | |
2413 | new_dfh.truncate(0) # drop any potentially existing data |
|
2423 | new_dfh.truncate(0) # drop any potentially existing data | |
2414 | try: |
|
2424 | try: | |
2415 | with self.reading(): |
|
2425 | with self.reading(): | |
2416 | for r in self: |
|
2426 | for r in self: | |
2417 | new_dfh.write(self._getsegmentforrevs(r, r)[1]) |
|
2427 | new_dfh.write(self._getsegmentforrevs(r, r)[1]) | |
2418 | new_dfh.flush() |
|
2428 | new_dfh.flush() | |
2419 |
|
2429 | |||
2420 | if side_write: |
|
2430 | if side_write: | |
2421 | self._indexfile = new_index_file_path |
|
2431 | self._indexfile = new_index_file_path | |
2422 | with self.__index_new_fp() as fp: |
|
2432 | with self.__index_new_fp() as fp: | |
2423 | self._format_flags &= ~FLAG_INLINE_DATA |
|
2433 | self._format_flags &= ~FLAG_INLINE_DATA | |
2424 | self._inline = False |
|
2434 | self._inline = False | |
2425 | for i in self: |
|
2435 | for i in self: | |
2426 | e = self.index.entry_binary(i) |
|
2436 | e = self.index.entry_binary(i) | |
2427 | if i == 0 and self._docket is None: |
|
2437 | if i == 0 and self._docket is None: | |
2428 | header = self._format_flags | self._format_version |
|
2438 | header = self._format_flags | self._format_version | |
2429 | header = self.index.pack_header(header) |
|
2439 | header = self.index.pack_header(header) | |
2430 | e = header + e |
|
2440 | e = header + e | |
2431 | fp.write(e) |
|
2441 | fp.write(e) | |
2432 | if self._docket is not None: |
|
2442 | if self._docket is not None: | |
2433 | self._docket.index_end = fp.tell() |
|
2443 | self._docket.index_end = fp.tell() | |
2434 |
|
2444 | |||
2435 | # If we don't use side-write, the temp file replace the real |
|
2445 | # If we don't use side-write, the temp file replace the real | |
2436 | # index when we exit the context manager |
|
2446 | # index when we exit the context manager | |
2437 |
|
2447 | |||
2438 | nodemaputil.setup_persistent_nodemap(tr, self) |
|
2448 | nodemaputil.setup_persistent_nodemap(tr, self) | |
2439 | self._segmentfile = randomaccessfile.randomaccessfile( |
|
2449 | self._segmentfile = randomaccessfile.randomaccessfile( | |
2440 | self.opener, |
|
2450 | self.opener, | |
2441 | self._datafile, |
|
2451 | self._datafile, | |
2442 | self._chunkcachesize, |
|
2452 | self._chunkcachesize, | |
2443 | ) |
|
2453 | ) | |
2444 |
|
2454 | |||
2445 | if existing_handles: |
|
2455 | if existing_handles: | |
2446 | # switched from inline to conventional reopen the index |
|
2456 | # switched from inline to conventional reopen the index | |
2447 | ifh = self.__index_write_fp() |
|
2457 | ifh = self.__index_write_fp() | |
2448 | self._writinghandles = (ifh, new_dfh, None) |
|
2458 | self._writinghandles = (ifh, new_dfh, None) | |
2449 | self._segmentfile.writing_handle = new_dfh |
|
2459 | self._segmentfile.writing_handle = new_dfh | |
2450 | new_dfh = None |
|
2460 | new_dfh = None | |
2451 | # No need to deal with sidedata writing handle as it is only |
|
2461 | # No need to deal with sidedata writing handle as it is only | |
2452 | # relevant with revlog-v2 which is never inline, not reaching |
|
2462 | # relevant with revlog-v2 which is never inline, not reaching | |
2453 | # this code |
|
2463 | # this code | |
2454 | finally: |
|
2464 | finally: | |
2455 | if new_dfh is not None: |
|
2465 | if new_dfh is not None: | |
2456 | new_dfh.close() |
|
2466 | new_dfh.close() | |
2457 |
|
2467 | |||
2458 | def _nodeduplicatecallback(self, transaction, node): |
|
2468 | def _nodeduplicatecallback(self, transaction, node): | |
2459 | """called when trying to add a node already stored.""" |
|
2469 | """called when trying to add a node already stored.""" | |
2460 |
|
2470 | |||
2461 | @contextlib.contextmanager |
|
2471 | @contextlib.contextmanager | |
2462 | def reading(self): |
|
2472 | def reading(self): | |
2463 | """Context manager that keeps data and sidedata files open for reading""" |
|
2473 | """Context manager that keeps data and sidedata files open for reading""" | |
2464 | if len(self.index) == 0: |
|
2474 | if len(self.index) == 0: | |
2465 | yield # nothing to be read |
|
2475 | yield # nothing to be read | |
2466 | else: |
|
2476 | else: | |
2467 | with self._segmentfile.reading(): |
|
2477 | with self._segmentfile.reading(): | |
2468 | with self._segmentfile_sidedata.reading(): |
|
2478 | with self._segmentfile_sidedata.reading(): | |
2469 | yield |
|
2479 | yield | |
2470 |
|
2480 | |||
2471 | @contextlib.contextmanager |
|
2481 | @contextlib.contextmanager | |
2472 | def _writing(self, transaction): |
|
2482 | def _writing(self, transaction): | |
2473 | if self._trypending: |
|
2483 | if self._trypending: | |
2474 | msg = b'try to write in a `trypending` revlog: %s' |
|
2484 | msg = b'try to write in a `trypending` revlog: %s' | |
2475 | msg %= self.display_id |
|
2485 | msg %= self.display_id | |
2476 | raise error.ProgrammingError(msg) |
|
2486 | raise error.ProgrammingError(msg) | |
2477 | if self._writinghandles is not None: |
|
2487 | if self._writinghandles is not None: | |
2478 | yield |
|
2488 | yield | |
2479 | else: |
|
2489 | else: | |
2480 | ifh = dfh = sdfh = None |
|
2490 | ifh = dfh = sdfh = None | |
2481 | try: |
|
2491 | try: | |
2482 | r = len(self) |
|
2492 | r = len(self) | |
2483 | # opening the data file. |
|
2493 | # opening the data file. | |
2484 | dsize = 0 |
|
2494 | dsize = 0 | |
2485 | if r: |
|
2495 | if r: | |
2486 | dsize = self.end(r - 1) |
|
2496 | dsize = self.end(r - 1) | |
2487 | dfh = None |
|
2497 | dfh = None | |
2488 | if not self._inline: |
|
2498 | if not self._inline: | |
2489 | try: |
|
2499 | try: | |
2490 | dfh = self._datafp(b"r+") |
|
2500 | dfh = self._datafp(b"r+") | |
2491 | if self._docket is None: |
|
2501 | if self._docket is None: | |
2492 | dfh.seek(0, os.SEEK_END) |
|
2502 | dfh.seek(0, os.SEEK_END) | |
2493 | else: |
|
2503 | else: | |
2494 | dfh.seek(self._docket.data_end, os.SEEK_SET) |
|
2504 | dfh.seek(self._docket.data_end, os.SEEK_SET) | |
2495 | except FileNotFoundError: |
|
2505 | except FileNotFoundError: | |
2496 | dfh = self._datafp(b"w+") |
|
2506 | dfh = self._datafp(b"w+") | |
2497 | transaction.add(self._datafile, dsize) |
|
2507 | transaction.add(self._datafile, dsize) | |
2498 | if self._sidedatafile is not None: |
|
2508 | if self._sidedatafile is not None: | |
2499 | # revlog-v2 does not inline, help Pytype |
|
2509 | # revlog-v2 does not inline, help Pytype | |
2500 | assert dfh is not None |
|
2510 | assert dfh is not None | |
2501 | try: |
|
2511 | try: | |
2502 | sdfh = self.opener(self._sidedatafile, mode=b"r+") |
|
2512 | sdfh = self.opener(self._sidedatafile, mode=b"r+") | |
2503 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) |
|
2513 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) | |
2504 | except FileNotFoundError: |
|
2514 | except FileNotFoundError: | |
2505 | sdfh = self.opener(self._sidedatafile, mode=b"w+") |
|
2515 | sdfh = self.opener(self._sidedatafile, mode=b"w+") | |
2506 | transaction.add( |
|
2516 | transaction.add( | |
2507 | self._sidedatafile, self._docket.sidedata_end |
|
2517 | self._sidedatafile, self._docket.sidedata_end | |
2508 | ) |
|
2518 | ) | |
2509 |
|
2519 | |||
2510 | # opening the index file. |
|
2520 | # opening the index file. | |
2511 | isize = r * self.index.entry_size |
|
2521 | isize = r * self.index.entry_size | |
2512 | ifh = self.__index_write_fp() |
|
2522 | ifh = self.__index_write_fp() | |
2513 | if self._inline: |
|
2523 | if self._inline: | |
2514 | transaction.add(self._indexfile, dsize + isize) |
|
2524 | transaction.add(self._indexfile, dsize + isize) | |
2515 | else: |
|
2525 | else: | |
2516 | transaction.add(self._indexfile, isize) |
|
2526 | transaction.add(self._indexfile, isize) | |
2517 | # exposing all file handle for writing. |
|
2527 | # exposing all file handle for writing. | |
2518 | self._writinghandles = (ifh, dfh, sdfh) |
|
2528 | self._writinghandles = (ifh, dfh, sdfh) | |
2519 | self._segmentfile.writing_handle = ifh if self._inline else dfh |
|
2529 | self._segmentfile.writing_handle = ifh if self._inline else dfh | |
2520 | self._segmentfile_sidedata.writing_handle = sdfh |
|
2530 | self._segmentfile_sidedata.writing_handle = sdfh | |
2521 | yield |
|
2531 | yield | |
2522 | if self._docket is not None: |
|
2532 | if self._docket is not None: | |
2523 | self._write_docket(transaction) |
|
2533 | self._write_docket(transaction) | |
2524 | finally: |
|
2534 | finally: | |
2525 | self._writinghandles = None |
|
2535 | self._writinghandles = None | |
2526 | self._segmentfile.writing_handle = None |
|
2536 | self._segmentfile.writing_handle = None | |
2527 | self._segmentfile_sidedata.writing_handle = None |
|
2537 | self._segmentfile_sidedata.writing_handle = None | |
2528 | if dfh is not None: |
|
2538 | if dfh is not None: | |
2529 | dfh.close() |
|
2539 | dfh.close() | |
2530 | if sdfh is not None: |
|
2540 | if sdfh is not None: | |
2531 | sdfh.close() |
|
2541 | sdfh.close() | |
2532 | # closing the index file last to avoid exposing referent to |
|
2542 | # closing the index file last to avoid exposing referent to | |
2533 | # potential unflushed data content. |
|
2543 | # potential unflushed data content. | |
2534 | if ifh is not None: |
|
2544 | if ifh is not None: | |
2535 | ifh.close() |
|
2545 | ifh.close() | |
2536 |
|
2546 | |||
2537 | def _write_docket(self, transaction): |
|
2547 | def _write_docket(self, transaction): | |
2538 | """write the current docket on disk |
|
2548 | """write the current docket on disk | |
2539 |
|
2549 | |||
2540 | Exist as a method to help changelog to implement transaction logic |
|
2550 | Exist as a method to help changelog to implement transaction logic | |
2541 |
|
2551 | |||
2542 | We could also imagine using the same transaction logic for all revlog |
|
2552 | We could also imagine using the same transaction logic for all revlog | |
2543 | since docket are cheap.""" |
|
2553 | since docket are cheap.""" | |
2544 | self._docket.write(transaction) |
|
2554 | self._docket.write(transaction) | |
2545 |
|
2555 | |||
2546 | def addrevision( |
|
2556 | def addrevision( | |
2547 | self, |
|
2557 | self, | |
2548 | text, |
|
2558 | text, | |
2549 | transaction, |
|
2559 | transaction, | |
2550 | link, |
|
2560 | link, | |
2551 | p1, |
|
2561 | p1, | |
2552 | p2, |
|
2562 | p2, | |
2553 | cachedelta=None, |
|
2563 | cachedelta=None, | |
2554 | node=None, |
|
2564 | node=None, | |
2555 | flags=REVIDX_DEFAULT_FLAGS, |
|
2565 | flags=REVIDX_DEFAULT_FLAGS, | |
2556 | deltacomputer=None, |
|
2566 | deltacomputer=None, | |
2557 | sidedata=None, |
|
2567 | sidedata=None, | |
2558 | ): |
|
2568 | ): | |
2559 | """add a revision to the log |
|
2569 | """add a revision to the log | |
2560 |
|
2570 | |||
2561 | text - the revision data to add |
|
2571 | text - the revision data to add | |
2562 | transaction - the transaction object used for rollback |
|
2572 | transaction - the transaction object used for rollback | |
2563 | link - the linkrev data to add |
|
2573 | link - the linkrev data to add | |
2564 | p1, p2 - the parent nodeids of the revision |
|
2574 | p1, p2 - the parent nodeids of the revision | |
2565 | cachedelta - an optional precomputed delta |
|
2575 | cachedelta - an optional precomputed delta | |
2566 | node - nodeid of revision; typically node is not specified, and it is |
|
2576 | node - nodeid of revision; typically node is not specified, and it is | |
2567 | computed by default as hash(text, p1, p2), however subclasses might |
|
2577 | computed by default as hash(text, p1, p2), however subclasses might | |
2568 | use different hashing method (and override checkhash() in such case) |
|
2578 | use different hashing method (and override checkhash() in such case) | |
2569 | flags - the known flags to set on the revision |
|
2579 | flags - the known flags to set on the revision | |
2570 | deltacomputer - an optional deltacomputer instance shared between |
|
2580 | deltacomputer - an optional deltacomputer instance shared between | |
2571 | multiple calls |
|
2581 | multiple calls | |
2572 | """ |
|
2582 | """ | |
2573 | if link == nullrev: |
|
2583 | if link == nullrev: | |
2574 | raise error.RevlogError( |
|
2584 | raise error.RevlogError( | |
2575 | _(b"attempted to add linkrev -1 to %s") % self.display_id |
|
2585 | _(b"attempted to add linkrev -1 to %s") % self.display_id | |
2576 | ) |
|
2586 | ) | |
2577 |
|
2587 | |||
2578 | if sidedata is None: |
|
2588 | if sidedata is None: | |
2579 | sidedata = {} |
|
2589 | sidedata = {} | |
2580 | elif sidedata and not self.hassidedata: |
|
2590 | elif sidedata and not self.hassidedata: | |
2581 | raise error.ProgrammingError( |
|
2591 | raise error.ProgrammingError( | |
2582 | _(b"trying to add sidedata to a revlog who don't support them") |
|
2592 | _(b"trying to add sidedata to a revlog who don't support them") | |
2583 | ) |
|
2593 | ) | |
2584 |
|
2594 | |||
2585 | if flags: |
|
2595 | if flags: | |
2586 | node = node or self.hash(text, p1, p2) |
|
2596 | node = node or self.hash(text, p1, p2) | |
2587 |
|
2597 | |||
2588 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) |
|
2598 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) | |
2589 |
|
2599 | |||
2590 | # If the flag processor modifies the revision data, ignore any provided |
|
2600 | # If the flag processor modifies the revision data, ignore any provided | |
2591 | # cachedelta. |
|
2601 | # cachedelta. | |
2592 | if rawtext != text: |
|
2602 | if rawtext != text: | |
2593 | cachedelta = None |
|
2603 | cachedelta = None | |
2594 |
|
2604 | |||
2595 | if len(rawtext) > _maxentrysize: |
|
2605 | if len(rawtext) > _maxentrysize: | |
2596 | raise error.RevlogError( |
|
2606 | raise error.RevlogError( | |
2597 | _( |
|
2607 | _( | |
2598 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" |
|
2608 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" | |
2599 | ) |
|
2609 | ) | |
2600 | % (self.display_id, len(rawtext)) |
|
2610 | % (self.display_id, len(rawtext)) | |
2601 | ) |
|
2611 | ) | |
2602 |
|
2612 | |||
2603 | node = node or self.hash(rawtext, p1, p2) |
|
2613 | node = node or self.hash(rawtext, p1, p2) | |
2604 | rev = self.index.get_rev(node) |
|
2614 | rev = self.index.get_rev(node) | |
2605 | if rev is not None: |
|
2615 | if rev is not None: | |
2606 | return rev |
|
2616 | return rev | |
2607 |
|
2617 | |||
2608 | if validatehash: |
|
2618 | if validatehash: | |
2609 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
2619 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
2610 |
|
2620 | |||
2611 | return self.addrawrevision( |
|
2621 | return self.addrawrevision( | |
2612 | rawtext, |
|
2622 | rawtext, | |
2613 | transaction, |
|
2623 | transaction, | |
2614 | link, |
|
2624 | link, | |
2615 | p1, |
|
2625 | p1, | |
2616 | p2, |
|
2626 | p2, | |
2617 | node, |
|
2627 | node, | |
2618 | flags, |
|
2628 | flags, | |
2619 | cachedelta=cachedelta, |
|
2629 | cachedelta=cachedelta, | |
2620 | deltacomputer=deltacomputer, |
|
2630 | deltacomputer=deltacomputer, | |
2621 | sidedata=sidedata, |
|
2631 | sidedata=sidedata, | |
2622 | ) |
|
2632 | ) | |
2623 |
|
2633 | |||
2624 | def addrawrevision( |
|
2634 | def addrawrevision( | |
2625 | self, |
|
2635 | self, | |
2626 | rawtext, |
|
2636 | rawtext, | |
2627 | transaction, |
|
2637 | transaction, | |
2628 | link, |
|
2638 | link, | |
2629 | p1, |
|
2639 | p1, | |
2630 | p2, |
|
2640 | p2, | |
2631 | node, |
|
2641 | node, | |
2632 | flags, |
|
2642 | flags, | |
2633 | cachedelta=None, |
|
2643 | cachedelta=None, | |
2634 | deltacomputer=None, |
|
2644 | deltacomputer=None, | |
2635 | sidedata=None, |
|
2645 | sidedata=None, | |
2636 | ): |
|
2646 | ): | |
2637 | """add a raw revision with known flags, node and parents |
|
2647 | """add a raw revision with known flags, node and parents | |
2638 | useful when reusing a revision not stored in this revlog (ex: received |
|
2648 | useful when reusing a revision not stored in this revlog (ex: received | |
2639 | over wire, or read from an external bundle). |
|
2649 | over wire, or read from an external bundle). | |
2640 | """ |
|
2650 | """ | |
2641 | with self._writing(transaction): |
|
2651 | with self._writing(transaction): | |
2642 | return self._addrevision( |
|
2652 | return self._addrevision( | |
2643 | node, |
|
2653 | node, | |
2644 | rawtext, |
|
2654 | rawtext, | |
2645 | transaction, |
|
2655 | transaction, | |
2646 | link, |
|
2656 | link, | |
2647 | p1, |
|
2657 | p1, | |
2648 | p2, |
|
2658 | p2, | |
2649 | flags, |
|
2659 | flags, | |
2650 | cachedelta, |
|
2660 | cachedelta, | |
2651 | deltacomputer=deltacomputer, |
|
2661 | deltacomputer=deltacomputer, | |
2652 | sidedata=sidedata, |
|
2662 | sidedata=sidedata, | |
2653 | ) |
|
2663 | ) | |
2654 |
|
2664 | |||
2655 | def compress(self, data): |
|
2665 | def compress(self, data): | |
2656 | """Generate a possibly-compressed representation of data.""" |
|
2666 | """Generate a possibly-compressed representation of data.""" | |
2657 | if not data: |
|
2667 | if not data: | |
2658 | return b'', data |
|
2668 | return b'', data | |
2659 |
|
2669 | |||
2660 | compressed = self._compressor.compress(data) |
|
2670 | compressed = self._compressor.compress(data) | |
2661 |
|
2671 | |||
2662 | if compressed: |
|
2672 | if compressed: | |
2663 | # The revlog compressor added the header in the returned data. |
|
2673 | # The revlog compressor added the header in the returned data. | |
2664 | return b'', compressed |
|
2674 | return b'', compressed | |
2665 |
|
2675 | |||
2666 | if data[0:1] == b'\0': |
|
2676 | if data[0:1] == b'\0': | |
2667 | return b'', data |
|
2677 | return b'', data | |
2668 | return b'u', data |
|
2678 | return b'u', data | |
2669 |
|
2679 | |||
2670 | def decompress(self, data): |
|
2680 | def decompress(self, data): | |
2671 | """Decompress a revlog chunk. |
|
2681 | """Decompress a revlog chunk. | |
2672 |
|
2682 | |||
2673 | The chunk is expected to begin with a header identifying the |
|
2683 | The chunk is expected to begin with a header identifying the | |
2674 | format type so it can be routed to an appropriate decompressor. |
|
2684 | format type so it can be routed to an appropriate decompressor. | |
2675 | """ |
|
2685 | """ | |
2676 | if not data: |
|
2686 | if not data: | |
2677 | return data |
|
2687 | return data | |
2678 |
|
2688 | |||
2679 | # Revlogs are read much more frequently than they are written and many |
|
2689 | # Revlogs are read much more frequently than they are written and many | |
2680 | # chunks only take microseconds to decompress, so performance is |
|
2690 | # chunks only take microseconds to decompress, so performance is | |
2681 | # important here. |
|
2691 | # important here. | |
2682 | # |
|
2692 | # | |
2683 | # We can make a few assumptions about revlogs: |
|
2693 | # We can make a few assumptions about revlogs: | |
2684 | # |
|
2694 | # | |
2685 | # 1) the majority of chunks will be compressed (as opposed to inline |
|
2695 | # 1) the majority of chunks will be compressed (as opposed to inline | |
2686 | # raw data). |
|
2696 | # raw data). | |
2687 | # 2) decompressing *any* data will likely by at least 10x slower than |
|
2697 | # 2) decompressing *any* data will likely by at least 10x slower than | |
2688 | # returning raw inline data. |
|
2698 | # returning raw inline data. | |
2689 | # 3) we want to prioritize common and officially supported compression |
|
2699 | # 3) we want to prioritize common and officially supported compression | |
2690 | # engines |
|
2700 | # engines | |
2691 | # |
|
2701 | # | |
2692 | # It follows that we want to optimize for "decompress compressed data |
|
2702 | # It follows that we want to optimize for "decompress compressed data | |
2693 | # when encoded with common and officially supported compression engines" |
|
2703 | # when encoded with common and officially supported compression engines" | |
2694 | # case over "raw data" and "data encoded by less common or non-official |
|
2704 | # case over "raw data" and "data encoded by less common or non-official | |
2695 | # compression engines." That is why we have the inline lookup first |
|
2705 | # compression engines." That is why we have the inline lookup first | |
2696 | # followed by the compengines lookup. |
|
2706 | # followed by the compengines lookup. | |
2697 | # |
|
2707 | # | |
2698 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib |
|
2708 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib | |
2699 | # compressed chunks. And this matters for changelog and manifest reads. |
|
2709 | # compressed chunks. And this matters for changelog and manifest reads. | |
2700 | t = data[0:1] |
|
2710 | t = data[0:1] | |
2701 |
|
2711 | |||
2702 | if t == b'x': |
|
2712 | if t == b'x': | |
2703 | try: |
|
2713 | try: | |
2704 | return _zlibdecompress(data) |
|
2714 | return _zlibdecompress(data) | |
2705 | except zlib.error as e: |
|
2715 | except zlib.error as e: | |
2706 | raise error.RevlogError( |
|
2716 | raise error.RevlogError( | |
2707 | _(b'revlog decompress error: %s') |
|
2717 | _(b'revlog decompress error: %s') | |
2708 | % stringutil.forcebytestr(e) |
|
2718 | % stringutil.forcebytestr(e) | |
2709 | ) |
|
2719 | ) | |
2710 | # '\0' is more common than 'u' so it goes first. |
|
2720 | # '\0' is more common than 'u' so it goes first. | |
2711 | elif t == b'\0': |
|
2721 | elif t == b'\0': | |
2712 | return data |
|
2722 | return data | |
2713 | elif t == b'u': |
|
2723 | elif t == b'u': | |
2714 | return util.buffer(data, 1) |
|
2724 | return util.buffer(data, 1) | |
2715 |
|
2725 | |||
2716 | compressor = self._get_decompressor(t) |
|
2726 | compressor = self._get_decompressor(t) | |
2717 |
|
2727 | |||
2718 | return compressor.decompress(data) |
|
2728 | return compressor.decompress(data) | |
2719 |
|
2729 | |||
2720 | def _addrevision( |
|
2730 | def _addrevision( | |
2721 | self, |
|
2731 | self, | |
2722 | node, |
|
2732 | node, | |
2723 | rawtext, |
|
2733 | rawtext, | |
2724 | transaction, |
|
2734 | transaction, | |
2725 | link, |
|
2735 | link, | |
2726 | p1, |
|
2736 | p1, | |
2727 | p2, |
|
2737 | p2, | |
2728 | flags, |
|
2738 | flags, | |
2729 | cachedelta, |
|
2739 | cachedelta, | |
2730 | alwayscache=False, |
|
2740 | alwayscache=False, | |
2731 | deltacomputer=None, |
|
2741 | deltacomputer=None, | |
2732 | sidedata=None, |
|
2742 | sidedata=None, | |
2733 | ): |
|
2743 | ): | |
2734 | """internal function to add revisions to the log |
|
2744 | """internal function to add revisions to the log | |
2735 |
|
2745 | |||
2736 | see addrevision for argument descriptions. |
|
2746 | see addrevision for argument descriptions. | |
2737 |
|
2747 | |||
2738 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. |
|
2748 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. | |
2739 |
|
2749 | |||
2740 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will |
|
2750 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will | |
2741 | be used. |
|
2751 | be used. | |
2742 |
|
2752 | |||
2743 | invariants: |
|
2753 | invariants: | |
2744 | - rawtext is optional (can be None); if not set, cachedelta must be set. |
|
2754 | - rawtext is optional (can be None); if not set, cachedelta must be set. | |
2745 | if both are set, they must correspond to each other. |
|
2755 | if both are set, they must correspond to each other. | |
2746 | """ |
|
2756 | """ | |
2747 | if node == self.nullid: |
|
2757 | if node == self.nullid: | |
2748 | raise error.RevlogError( |
|
2758 | raise error.RevlogError( | |
2749 | _(b"%s: attempt to add null revision") % self.display_id |
|
2759 | _(b"%s: attempt to add null revision") % self.display_id | |
2750 | ) |
|
2760 | ) | |
2751 | if ( |
|
2761 | if ( | |
2752 | node == self.nodeconstants.wdirid |
|
2762 | node == self.nodeconstants.wdirid | |
2753 | or node in self.nodeconstants.wdirfilenodeids |
|
2763 | or node in self.nodeconstants.wdirfilenodeids | |
2754 | ): |
|
2764 | ): | |
2755 | raise error.RevlogError( |
|
2765 | raise error.RevlogError( | |
2756 | _(b"%s: attempt to add wdir revision") % self.display_id |
|
2766 | _(b"%s: attempt to add wdir revision") % self.display_id | |
2757 | ) |
|
2767 | ) | |
2758 | if self._writinghandles is None: |
|
2768 | if self._writinghandles is None: | |
2759 | msg = b'adding revision outside `revlog._writing` context' |
|
2769 | msg = b'adding revision outside `revlog._writing` context' | |
2760 | raise error.ProgrammingError(msg) |
|
2770 | raise error.ProgrammingError(msg) | |
2761 |
|
2771 | |||
2762 | btext = [rawtext] |
|
2772 | btext = [rawtext] | |
2763 |
|
2773 | |||
2764 | curr = len(self) |
|
2774 | curr = len(self) | |
2765 | prev = curr - 1 |
|
2775 | prev = curr - 1 | |
2766 |
|
2776 | |||
2767 | offset = self._get_data_offset(prev) |
|
2777 | offset = self._get_data_offset(prev) | |
2768 |
|
2778 | |||
2769 | if self._concurrencychecker: |
|
2779 | if self._concurrencychecker: | |
2770 | ifh, dfh, sdfh = self._writinghandles |
|
2780 | ifh, dfh, sdfh = self._writinghandles | |
2771 | # XXX no checking for the sidedata file |
|
2781 | # XXX no checking for the sidedata file | |
2772 | if self._inline: |
|
2782 | if self._inline: | |
2773 | # offset is "as if" it were in the .d file, so we need to add on |
|
2783 | # offset is "as if" it were in the .d file, so we need to add on | |
2774 | # the size of the entry metadata. |
|
2784 | # the size of the entry metadata. | |
2775 | self._concurrencychecker( |
|
2785 | self._concurrencychecker( | |
2776 | ifh, self._indexfile, offset + curr * self.index.entry_size |
|
2786 | ifh, self._indexfile, offset + curr * self.index.entry_size | |
2777 | ) |
|
2787 | ) | |
2778 | else: |
|
2788 | else: | |
2779 | # Entries in the .i are a consistent size. |
|
2789 | # Entries in the .i are a consistent size. | |
2780 | self._concurrencychecker( |
|
2790 | self._concurrencychecker( | |
2781 | ifh, self._indexfile, curr * self.index.entry_size |
|
2791 | ifh, self._indexfile, curr * self.index.entry_size | |
2782 | ) |
|
2792 | ) | |
2783 | self._concurrencychecker(dfh, self._datafile, offset) |
|
2793 | self._concurrencychecker(dfh, self._datafile, offset) | |
2784 |
|
2794 | |||
2785 | p1r, p2r = self.rev(p1), self.rev(p2) |
|
2795 | p1r, p2r = self.rev(p1), self.rev(p2) | |
2786 |
|
2796 | |||
2787 | # full versions are inserted when the needed deltas |
|
2797 | # full versions are inserted when the needed deltas | |
2788 | # become comparable to the uncompressed text |
|
2798 | # become comparable to the uncompressed text | |
2789 | if rawtext is None: |
|
2799 | if rawtext is None: | |
2790 | # need rawtext size, before changed by flag processors, which is |
|
2800 | # need rawtext size, before changed by flag processors, which is | |
2791 | # the non-raw size. use revlog explicitly to avoid filelog's extra |
|
2801 | # the non-raw size. use revlog explicitly to avoid filelog's extra | |
2792 | # logic that might remove metadata size. |
|
2802 | # logic that might remove metadata size. | |
2793 | textlen = mdiff.patchedsize( |
|
2803 | textlen = mdiff.patchedsize( | |
2794 | revlog.size(self, cachedelta[0]), cachedelta[1] |
|
2804 | revlog.size(self, cachedelta[0]), cachedelta[1] | |
2795 | ) |
|
2805 | ) | |
2796 | else: |
|
2806 | else: | |
2797 | textlen = len(rawtext) |
|
2807 | textlen = len(rawtext) | |
2798 |
|
2808 | |||
2799 | if deltacomputer is None: |
|
2809 | if deltacomputer is None: | |
2800 | write_debug = None |
|
2810 | write_debug = None | |
2801 | if self._debug_delta: |
|
2811 | if self._debug_delta: | |
2802 | write_debug = transaction._report |
|
2812 | write_debug = transaction._report | |
2803 | deltacomputer = deltautil.deltacomputer( |
|
2813 | deltacomputer = deltautil.deltacomputer( | |
2804 | self, write_debug=write_debug |
|
2814 | self, write_debug=write_debug | |
2805 | ) |
|
2815 | ) | |
2806 |
|
2816 | |||
2807 | if cachedelta is not None and len(cachedelta) == 2: |
|
2817 | if cachedelta is not None and len(cachedelta) == 2: | |
2808 | # If the cached delta has no information about how it should be |
|
2818 | # If the cached delta has no information about how it should be | |
2809 | # reused, add the default reuse instruction according to the |
|
2819 | # reused, add the default reuse instruction according to the | |
2810 | # revlog's configuration. |
|
2820 | # revlog's configuration. | |
2811 | if self._generaldelta and self._lazydeltabase: |
|
2821 | if self._generaldelta and self._lazydeltabase: | |
2812 | delta_base_reuse = DELTA_BASE_REUSE_TRY |
|
2822 | delta_base_reuse = DELTA_BASE_REUSE_TRY | |
2813 | else: |
|
2823 | else: | |
2814 | delta_base_reuse = DELTA_BASE_REUSE_NO |
|
2824 | delta_base_reuse = DELTA_BASE_REUSE_NO | |
2815 | cachedelta = (cachedelta[0], cachedelta[1], delta_base_reuse) |
|
2825 | cachedelta = (cachedelta[0], cachedelta[1], delta_base_reuse) | |
2816 |
|
2826 | |||
2817 | revinfo = revlogutils.revisioninfo( |
|
2827 | revinfo = revlogutils.revisioninfo( | |
2818 | node, |
|
2828 | node, | |
2819 | p1, |
|
2829 | p1, | |
2820 | p2, |
|
2830 | p2, | |
2821 | btext, |
|
2831 | btext, | |
2822 | textlen, |
|
2832 | textlen, | |
2823 | cachedelta, |
|
2833 | cachedelta, | |
2824 | flags, |
|
2834 | flags, | |
2825 | ) |
|
2835 | ) | |
2826 |
|
2836 | |||
2827 | deltainfo = deltacomputer.finddeltainfo(revinfo) |
|
2837 | deltainfo = deltacomputer.finddeltainfo(revinfo) | |
2828 |
|
2838 | |||
2829 | compression_mode = COMP_MODE_INLINE |
|
2839 | compression_mode = COMP_MODE_INLINE | |
2830 | if self._docket is not None: |
|
2840 | if self._docket is not None: | |
2831 | default_comp = self._docket.default_compression_header |
|
2841 | default_comp = self._docket.default_compression_header | |
2832 | r = deltautil.delta_compression(default_comp, deltainfo) |
|
2842 | r = deltautil.delta_compression(default_comp, deltainfo) | |
2833 | compression_mode, deltainfo = r |
|
2843 | compression_mode, deltainfo = r | |
2834 |
|
2844 | |||
2835 | sidedata_compression_mode = COMP_MODE_INLINE |
|
2845 | sidedata_compression_mode = COMP_MODE_INLINE | |
2836 | if sidedata and self.hassidedata: |
|
2846 | if sidedata and self.hassidedata: | |
2837 | sidedata_compression_mode = COMP_MODE_PLAIN |
|
2847 | sidedata_compression_mode = COMP_MODE_PLAIN | |
2838 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) |
|
2848 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) | |
2839 | sidedata_offset = self._docket.sidedata_end |
|
2849 | sidedata_offset = self._docket.sidedata_end | |
2840 | h, comp_sidedata = self.compress(serialized_sidedata) |
|
2850 | h, comp_sidedata = self.compress(serialized_sidedata) | |
2841 | if ( |
|
2851 | if ( | |
2842 | h != b'u' |
|
2852 | h != b'u' | |
2843 | and comp_sidedata[0:1] != b'\0' |
|
2853 | and comp_sidedata[0:1] != b'\0' | |
2844 | and len(comp_sidedata) < len(serialized_sidedata) |
|
2854 | and len(comp_sidedata) < len(serialized_sidedata) | |
2845 | ): |
|
2855 | ): | |
2846 | assert not h |
|
2856 | assert not h | |
2847 | if ( |
|
2857 | if ( | |
2848 | comp_sidedata[0:1] |
|
2858 | comp_sidedata[0:1] | |
2849 | == self._docket.default_compression_header |
|
2859 | == self._docket.default_compression_header | |
2850 | ): |
|
2860 | ): | |
2851 | sidedata_compression_mode = COMP_MODE_DEFAULT |
|
2861 | sidedata_compression_mode = COMP_MODE_DEFAULT | |
2852 | serialized_sidedata = comp_sidedata |
|
2862 | serialized_sidedata = comp_sidedata | |
2853 | else: |
|
2863 | else: | |
2854 | sidedata_compression_mode = COMP_MODE_INLINE |
|
2864 | sidedata_compression_mode = COMP_MODE_INLINE | |
2855 | serialized_sidedata = comp_sidedata |
|
2865 | serialized_sidedata = comp_sidedata | |
2856 | else: |
|
2866 | else: | |
2857 | serialized_sidedata = b"" |
|
2867 | serialized_sidedata = b"" | |
2858 | # Don't store the offset if the sidedata is empty, that way |
|
2868 | # Don't store the offset if the sidedata is empty, that way | |
2859 | # we can easily detect empty sidedata and they will be no different |
|
2869 | # we can easily detect empty sidedata and they will be no different | |
2860 | # than ones we manually add. |
|
2870 | # than ones we manually add. | |
2861 | sidedata_offset = 0 |
|
2871 | sidedata_offset = 0 | |
2862 |
|
2872 | |||
2863 | rank = RANK_UNKNOWN |
|
2873 | rank = RANK_UNKNOWN | |
2864 | if self._compute_rank: |
|
2874 | if self._compute_rank: | |
2865 | if (p1r, p2r) == (nullrev, nullrev): |
|
2875 | if (p1r, p2r) == (nullrev, nullrev): | |
2866 | rank = 1 |
|
2876 | rank = 1 | |
2867 | elif p1r != nullrev and p2r == nullrev: |
|
2877 | elif p1r != nullrev and p2r == nullrev: | |
2868 | rank = 1 + self.fast_rank(p1r) |
|
2878 | rank = 1 + self.fast_rank(p1r) | |
2869 | elif p1r == nullrev and p2r != nullrev: |
|
2879 | elif p1r == nullrev and p2r != nullrev: | |
2870 | rank = 1 + self.fast_rank(p2r) |
|
2880 | rank = 1 + self.fast_rank(p2r) | |
2871 | else: # merge node |
|
2881 | else: # merge node | |
2872 | if rustdagop is not None and self.index.rust_ext_compat: |
|
2882 | if rustdagop is not None and self.index.rust_ext_compat: | |
2873 | rank = rustdagop.rank(self.index, p1r, p2r) |
|
2883 | rank = rustdagop.rank(self.index, p1r, p2r) | |
2874 | else: |
|
2884 | else: | |
2875 | pmin, pmax = sorted((p1r, p2r)) |
|
2885 | pmin, pmax = sorted((p1r, p2r)) | |
2876 | rank = 1 + self.fast_rank(pmax) |
|
2886 | rank = 1 + self.fast_rank(pmax) | |
2877 | rank += sum(1 for _ in self.findmissingrevs([pmax], [pmin])) |
|
2887 | rank += sum(1 for _ in self.findmissingrevs([pmax], [pmin])) | |
2878 |
|
2888 | |||
2879 | e = revlogutils.entry( |
|
2889 | e = revlogutils.entry( | |
2880 | flags=flags, |
|
2890 | flags=flags, | |
2881 | data_offset=offset, |
|
2891 | data_offset=offset, | |
2882 | data_compressed_length=deltainfo.deltalen, |
|
2892 | data_compressed_length=deltainfo.deltalen, | |
2883 | data_uncompressed_length=textlen, |
|
2893 | data_uncompressed_length=textlen, | |
2884 | data_compression_mode=compression_mode, |
|
2894 | data_compression_mode=compression_mode, | |
2885 | data_delta_base=deltainfo.base, |
|
2895 | data_delta_base=deltainfo.base, | |
2886 | link_rev=link, |
|
2896 | link_rev=link, | |
2887 | parent_rev_1=p1r, |
|
2897 | parent_rev_1=p1r, | |
2888 | parent_rev_2=p2r, |
|
2898 | parent_rev_2=p2r, | |
2889 | node_id=node, |
|
2899 | node_id=node, | |
2890 | sidedata_offset=sidedata_offset, |
|
2900 | sidedata_offset=sidedata_offset, | |
2891 | sidedata_compressed_length=len(serialized_sidedata), |
|
2901 | sidedata_compressed_length=len(serialized_sidedata), | |
2892 | sidedata_compression_mode=sidedata_compression_mode, |
|
2902 | sidedata_compression_mode=sidedata_compression_mode, | |
2893 | rank=rank, |
|
2903 | rank=rank, | |
2894 | ) |
|
2904 | ) | |
2895 |
|
2905 | |||
2896 | self.index.append(e) |
|
2906 | self.index.append(e) | |
2897 | entry = self.index.entry_binary(curr) |
|
2907 | entry = self.index.entry_binary(curr) | |
2898 | if curr == 0 and self._docket is None: |
|
2908 | if curr == 0 and self._docket is None: | |
2899 | header = self._format_flags | self._format_version |
|
2909 | header = self._format_flags | self._format_version | |
2900 | header = self.index.pack_header(header) |
|
2910 | header = self.index.pack_header(header) | |
2901 | entry = header + entry |
|
2911 | entry = header + entry | |
2902 | self._writeentry( |
|
2912 | self._writeentry( | |
2903 | transaction, |
|
2913 | transaction, | |
2904 | entry, |
|
2914 | entry, | |
2905 | deltainfo.data, |
|
2915 | deltainfo.data, | |
2906 | link, |
|
2916 | link, | |
2907 | offset, |
|
2917 | offset, | |
2908 | serialized_sidedata, |
|
2918 | serialized_sidedata, | |
2909 | sidedata_offset, |
|
2919 | sidedata_offset, | |
2910 | ) |
|
2920 | ) | |
2911 |
|
2921 | |||
2912 | rawtext = btext[0] |
|
2922 | rawtext = btext[0] | |
2913 |
|
2923 | |||
2914 | if alwayscache and rawtext is None: |
|
2924 | if alwayscache and rawtext is None: | |
2915 | rawtext = deltacomputer.buildtext(revinfo) |
|
2925 | rawtext = deltacomputer.buildtext(revinfo) | |
2916 |
|
2926 | |||
2917 | if type(rawtext) == bytes: # only accept immutable objects |
|
2927 | if type(rawtext) == bytes: # only accept immutable objects | |
2918 | self._revisioncache = (node, curr, rawtext) |
|
2928 | self._revisioncache = (node, curr, rawtext) | |
2919 | self._chainbasecache[curr] = deltainfo.chainbase |
|
2929 | self._chainbasecache[curr] = deltainfo.chainbase | |
2920 | return curr |
|
2930 | return curr | |
2921 |
|
2931 | |||
2922 | def _get_data_offset(self, prev): |
|
2932 | def _get_data_offset(self, prev): | |
2923 | """Returns the current offset in the (in-transaction) data file. |
|
2933 | """Returns the current offset in the (in-transaction) data file. | |
2924 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket |
|
2934 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket | |
2925 | file to store that information: since sidedata can be rewritten to the |
|
2935 | file to store that information: since sidedata can be rewritten to the | |
2926 | end of the data file within a transaction, you can have cases where, for |
|
2936 | end of the data file within a transaction, you can have cases where, for | |
2927 | example, rev `n` does not have sidedata while rev `n - 1` does, leading |
|
2937 | example, rev `n` does not have sidedata while rev `n - 1` does, leading | |
2928 | to `n - 1`'s sidedata being written after `n`'s data. |
|
2938 | to `n - 1`'s sidedata being written after `n`'s data. | |
2929 |
|
2939 | |||
2930 | TODO cache this in a docket file before getting out of experimental.""" |
|
2940 | TODO cache this in a docket file before getting out of experimental.""" | |
2931 | if self._docket is None: |
|
2941 | if self._docket is None: | |
2932 | return self.end(prev) |
|
2942 | return self.end(prev) | |
2933 | else: |
|
2943 | else: | |
2934 | return self._docket.data_end |
|
2944 | return self._docket.data_end | |
2935 |
|
2945 | |||
2936 | def _writeentry( |
|
2946 | def _writeentry( | |
2937 | self, transaction, entry, data, link, offset, sidedata, sidedata_offset |
|
2947 | self, transaction, entry, data, link, offset, sidedata, sidedata_offset | |
2938 | ): |
|
2948 | ): | |
2939 | # Files opened in a+ mode have inconsistent behavior on various |
|
2949 | # Files opened in a+ mode have inconsistent behavior on various | |
2940 | # platforms. Windows requires that a file positioning call be made |
|
2950 | # platforms. Windows requires that a file positioning call be made | |
2941 | # when the file handle transitions between reads and writes. See |
|
2951 | # when the file handle transitions between reads and writes. See | |
2942 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
2952 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other | |
2943 | # platforms, Python or the platform itself can be buggy. Some versions |
|
2953 | # platforms, Python or the platform itself can be buggy. Some versions | |
2944 | # of Solaris have been observed to not append at the end of the file |
|
2954 | # of Solaris have been observed to not append at the end of the file | |
2945 | # if the file was seeked to before the end. See issue4943 for more. |
|
2955 | # if the file was seeked to before the end. See issue4943 for more. | |
2946 | # |
|
2956 | # | |
2947 | # We work around this issue by inserting a seek() before writing. |
|
2957 | # We work around this issue by inserting a seek() before writing. | |
2948 | # Note: This is likely not necessary on Python 3. However, because |
|
2958 | # Note: This is likely not necessary on Python 3. However, because | |
2949 | # the file handle is reused for reads and may be seeked there, we need |
|
2959 | # the file handle is reused for reads and may be seeked there, we need | |
2950 | # to be careful before changing this. |
|
2960 | # to be careful before changing this. | |
2951 | if self._writinghandles is None: |
|
2961 | if self._writinghandles is None: | |
2952 | msg = b'adding revision outside `revlog._writing` context' |
|
2962 | msg = b'adding revision outside `revlog._writing` context' | |
2953 | raise error.ProgrammingError(msg) |
|
2963 | raise error.ProgrammingError(msg) | |
2954 | ifh, dfh, sdfh = self._writinghandles |
|
2964 | ifh, dfh, sdfh = self._writinghandles | |
2955 | if self._docket is None: |
|
2965 | if self._docket is None: | |
2956 | ifh.seek(0, os.SEEK_END) |
|
2966 | ifh.seek(0, os.SEEK_END) | |
2957 | else: |
|
2967 | else: | |
2958 | ifh.seek(self._docket.index_end, os.SEEK_SET) |
|
2968 | ifh.seek(self._docket.index_end, os.SEEK_SET) | |
2959 | if dfh: |
|
2969 | if dfh: | |
2960 | if self._docket is None: |
|
2970 | if self._docket is None: | |
2961 | dfh.seek(0, os.SEEK_END) |
|
2971 | dfh.seek(0, os.SEEK_END) | |
2962 | else: |
|
2972 | else: | |
2963 | dfh.seek(self._docket.data_end, os.SEEK_SET) |
|
2973 | dfh.seek(self._docket.data_end, os.SEEK_SET) | |
2964 | if sdfh: |
|
2974 | if sdfh: | |
2965 | sdfh.seek(self._docket.sidedata_end, os.SEEK_SET) |
|
2975 | sdfh.seek(self._docket.sidedata_end, os.SEEK_SET) | |
2966 |
|
2976 | |||
2967 | curr = len(self) - 1 |
|
2977 | curr = len(self) - 1 | |
2968 | if not self._inline: |
|
2978 | if not self._inline: | |
2969 | transaction.add(self._datafile, offset) |
|
2979 | transaction.add(self._datafile, offset) | |
2970 | if self._sidedatafile: |
|
2980 | if self._sidedatafile: | |
2971 | transaction.add(self._sidedatafile, sidedata_offset) |
|
2981 | transaction.add(self._sidedatafile, sidedata_offset) | |
2972 | transaction.add(self._indexfile, curr * len(entry)) |
|
2982 | transaction.add(self._indexfile, curr * len(entry)) | |
2973 | if data[0]: |
|
2983 | if data[0]: | |
2974 | dfh.write(data[0]) |
|
2984 | dfh.write(data[0]) | |
2975 | dfh.write(data[1]) |
|
2985 | dfh.write(data[1]) | |
2976 | if sidedata: |
|
2986 | if sidedata: | |
2977 | sdfh.write(sidedata) |
|
2987 | sdfh.write(sidedata) | |
2978 | ifh.write(entry) |
|
2988 | ifh.write(entry) | |
2979 | else: |
|
2989 | else: | |
2980 | offset += curr * self.index.entry_size |
|
2990 | offset += curr * self.index.entry_size | |
2981 | transaction.add(self._indexfile, offset) |
|
2991 | transaction.add(self._indexfile, offset) | |
2982 | ifh.write(entry) |
|
2992 | ifh.write(entry) | |
2983 | ifh.write(data[0]) |
|
2993 | ifh.write(data[0]) | |
2984 | ifh.write(data[1]) |
|
2994 | ifh.write(data[1]) | |
2985 | assert not sidedata |
|
2995 | assert not sidedata | |
2986 | self._enforceinlinesize(transaction) |
|
2996 | self._enforceinlinesize(transaction) | |
2987 | if self._docket is not None: |
|
2997 | if self._docket is not None: | |
2988 | # revlog-v2 always has 3 writing handles, help Pytype |
|
2998 | # revlog-v2 always has 3 writing handles, help Pytype | |
2989 | wh1 = self._writinghandles[0] |
|
2999 | wh1 = self._writinghandles[0] | |
2990 | wh2 = self._writinghandles[1] |
|
3000 | wh2 = self._writinghandles[1] | |
2991 | wh3 = self._writinghandles[2] |
|
3001 | wh3 = self._writinghandles[2] | |
2992 | assert wh1 is not None |
|
3002 | assert wh1 is not None | |
2993 | assert wh2 is not None |
|
3003 | assert wh2 is not None | |
2994 | assert wh3 is not None |
|
3004 | assert wh3 is not None | |
2995 | self._docket.index_end = wh1.tell() |
|
3005 | self._docket.index_end = wh1.tell() | |
2996 | self._docket.data_end = wh2.tell() |
|
3006 | self._docket.data_end = wh2.tell() | |
2997 | self._docket.sidedata_end = wh3.tell() |
|
3007 | self._docket.sidedata_end = wh3.tell() | |
2998 |
|
3008 | |||
2999 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
3009 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
3000 |
|
3010 | |||
3001 | def addgroup( |
|
3011 | def addgroup( | |
3002 | self, |
|
3012 | self, | |
3003 | deltas, |
|
3013 | deltas, | |
3004 | linkmapper, |
|
3014 | linkmapper, | |
3005 | transaction, |
|
3015 | transaction, | |
3006 | alwayscache=False, |
|
3016 | alwayscache=False, | |
3007 | addrevisioncb=None, |
|
3017 | addrevisioncb=None, | |
3008 | duplicaterevisioncb=None, |
|
3018 | duplicaterevisioncb=None, | |
3009 | debug_info=None, |
|
3019 | debug_info=None, | |
3010 | delta_base_reuse_policy=None, |
|
3020 | delta_base_reuse_policy=None, | |
3011 | ): |
|
3021 | ): | |
3012 | """ |
|
3022 | """ | |
3013 | add a delta group |
|
3023 | add a delta group | |
3014 |
|
3024 | |||
3015 | given a set of deltas, add them to the revision log. the |
|
3025 | given a set of deltas, add them to the revision log. the | |
3016 | first delta is against its parent, which should be in our |
|
3026 | first delta is against its parent, which should be in our | |
3017 | log, the rest are against the previous delta. |
|
3027 | log, the rest are against the previous delta. | |
3018 |
|
3028 | |||
3019 | If ``addrevisioncb`` is defined, it will be called with arguments of |
|
3029 | If ``addrevisioncb`` is defined, it will be called with arguments of | |
3020 | this revlog and the node that was added. |
|
3030 | this revlog and the node that was added. | |
3021 | """ |
|
3031 | """ | |
3022 |
|
3032 | |||
3023 | if self._adding_group: |
|
3033 | if self._adding_group: | |
3024 | raise error.ProgrammingError(b'cannot nest addgroup() calls') |
|
3034 | raise error.ProgrammingError(b'cannot nest addgroup() calls') | |
3025 |
|
3035 | |||
3026 | # read the default delta-base reuse policy from revlog config if the |
|
3036 | # read the default delta-base reuse policy from revlog config if the | |
3027 | # group did not specify one. |
|
3037 | # group did not specify one. | |
3028 | if delta_base_reuse_policy is None: |
|
3038 | if delta_base_reuse_policy is None: | |
3029 | if self._generaldelta and self._lazydeltabase: |
|
3039 | if self._generaldelta and self._lazydeltabase: | |
3030 | delta_base_reuse_policy = DELTA_BASE_REUSE_TRY |
|
3040 | delta_base_reuse_policy = DELTA_BASE_REUSE_TRY | |
3031 | else: |
|
3041 | else: | |
3032 | delta_base_reuse_policy = DELTA_BASE_REUSE_NO |
|
3042 | delta_base_reuse_policy = DELTA_BASE_REUSE_NO | |
3033 |
|
3043 | |||
3034 | self._adding_group = True |
|
3044 | self._adding_group = True | |
3035 | empty = True |
|
3045 | empty = True | |
3036 | try: |
|
3046 | try: | |
3037 | with self._writing(transaction): |
|
3047 | with self._writing(transaction): | |
3038 | write_debug = None |
|
3048 | write_debug = None | |
3039 | if self._debug_delta: |
|
3049 | if self._debug_delta: | |
3040 | write_debug = transaction._report |
|
3050 | write_debug = transaction._report | |
3041 | deltacomputer = deltautil.deltacomputer( |
|
3051 | deltacomputer = deltautil.deltacomputer( | |
3042 | self, |
|
3052 | self, | |
3043 | write_debug=write_debug, |
|
3053 | write_debug=write_debug, | |
3044 | debug_info=debug_info, |
|
3054 | debug_info=debug_info, | |
3045 | ) |
|
3055 | ) | |
3046 | # loop through our set of deltas |
|
3056 | # loop through our set of deltas | |
3047 | for data in deltas: |
|
3057 | for data in deltas: | |
3048 | ( |
|
3058 | ( | |
3049 | node, |
|
3059 | node, | |
3050 | p1, |
|
3060 | p1, | |
3051 | p2, |
|
3061 | p2, | |
3052 | linknode, |
|
3062 | linknode, | |
3053 | deltabase, |
|
3063 | deltabase, | |
3054 | delta, |
|
3064 | delta, | |
3055 | flags, |
|
3065 | flags, | |
3056 | sidedata, |
|
3066 | sidedata, | |
3057 | ) = data |
|
3067 | ) = data | |
3058 | link = linkmapper(linknode) |
|
3068 | link = linkmapper(linknode) | |
3059 | flags = flags or REVIDX_DEFAULT_FLAGS |
|
3069 | flags = flags or REVIDX_DEFAULT_FLAGS | |
3060 |
|
3070 | |||
3061 | rev = self.index.get_rev(node) |
|
3071 | rev = self.index.get_rev(node) | |
3062 | if rev is not None: |
|
3072 | if rev is not None: | |
3063 | # this can happen if two branches make the same change |
|
3073 | # this can happen if two branches make the same change | |
3064 | self._nodeduplicatecallback(transaction, rev) |
|
3074 | self._nodeduplicatecallback(transaction, rev) | |
3065 | if duplicaterevisioncb: |
|
3075 | if duplicaterevisioncb: | |
3066 | duplicaterevisioncb(self, rev) |
|
3076 | duplicaterevisioncb(self, rev) | |
3067 | empty = False |
|
3077 | empty = False | |
3068 | continue |
|
3078 | continue | |
3069 |
|
3079 | |||
3070 | for p in (p1, p2): |
|
3080 | for p in (p1, p2): | |
3071 | if not self.index.has_node(p): |
|
3081 | if not self.index.has_node(p): | |
3072 | raise error.LookupError( |
|
3082 | raise error.LookupError( | |
3073 | p, self.radix, _(b'unknown parent') |
|
3083 | p, self.radix, _(b'unknown parent') | |
3074 | ) |
|
3084 | ) | |
3075 |
|
3085 | |||
3076 | if not self.index.has_node(deltabase): |
|
3086 | if not self.index.has_node(deltabase): | |
3077 | raise error.LookupError( |
|
3087 | raise error.LookupError( | |
3078 | deltabase, self.display_id, _(b'unknown delta base') |
|
3088 | deltabase, self.display_id, _(b'unknown delta base') | |
3079 | ) |
|
3089 | ) | |
3080 |
|
3090 | |||
3081 | baserev = self.rev(deltabase) |
|
3091 | baserev = self.rev(deltabase) | |
3082 |
|
3092 | |||
3083 | if baserev != nullrev and self.iscensored(baserev): |
|
3093 | if baserev != nullrev and self.iscensored(baserev): | |
3084 | # if base is censored, delta must be full replacement in a |
|
3094 | # if base is censored, delta must be full replacement in a | |
3085 | # single patch operation |
|
3095 | # single patch operation | |
3086 | hlen = struct.calcsize(b">lll") |
|
3096 | hlen = struct.calcsize(b">lll") | |
3087 | oldlen = self.rawsize(baserev) |
|
3097 | oldlen = self.rawsize(baserev) | |
3088 | newlen = len(delta) - hlen |
|
3098 | newlen = len(delta) - hlen | |
3089 | if delta[:hlen] != mdiff.replacediffheader( |
|
3099 | if delta[:hlen] != mdiff.replacediffheader( | |
3090 | oldlen, newlen |
|
3100 | oldlen, newlen | |
3091 | ): |
|
3101 | ): | |
3092 | raise error.CensoredBaseError( |
|
3102 | raise error.CensoredBaseError( | |
3093 | self.display_id, self.node(baserev) |
|
3103 | self.display_id, self.node(baserev) | |
3094 | ) |
|
3104 | ) | |
3095 |
|
3105 | |||
3096 | if not flags and self._peek_iscensored(baserev, delta): |
|
3106 | if not flags and self._peek_iscensored(baserev, delta): | |
3097 | flags |= REVIDX_ISCENSORED |
|
3107 | flags |= REVIDX_ISCENSORED | |
3098 |
|
3108 | |||
3099 | # We assume consumers of addrevisioncb will want to retrieve |
|
3109 | # We assume consumers of addrevisioncb will want to retrieve | |
3100 | # the added revision, which will require a call to |
|
3110 | # the added revision, which will require a call to | |
3101 | # revision(). revision() will fast path if there is a cache |
|
3111 | # revision(). revision() will fast path if there is a cache | |
3102 | # hit. So, we tell _addrevision() to always cache in this case. |
|
3112 | # hit. So, we tell _addrevision() to always cache in this case. | |
3103 | # We're only using addgroup() in the context of changegroup |
|
3113 | # We're only using addgroup() in the context of changegroup | |
3104 | # generation so the revision data can always be handled as raw |
|
3114 | # generation so the revision data can always be handled as raw | |
3105 | # by the flagprocessor. |
|
3115 | # by the flagprocessor. | |
3106 | rev = self._addrevision( |
|
3116 | rev = self._addrevision( | |
3107 | node, |
|
3117 | node, | |
3108 | None, |
|
3118 | None, | |
3109 | transaction, |
|
3119 | transaction, | |
3110 | link, |
|
3120 | link, | |
3111 | p1, |
|
3121 | p1, | |
3112 | p2, |
|
3122 | p2, | |
3113 | flags, |
|
3123 | flags, | |
3114 | (baserev, delta, delta_base_reuse_policy), |
|
3124 | (baserev, delta, delta_base_reuse_policy), | |
3115 | alwayscache=alwayscache, |
|
3125 | alwayscache=alwayscache, | |
3116 | deltacomputer=deltacomputer, |
|
3126 | deltacomputer=deltacomputer, | |
3117 | sidedata=sidedata, |
|
3127 | sidedata=sidedata, | |
3118 | ) |
|
3128 | ) | |
3119 |
|
3129 | |||
3120 | if addrevisioncb: |
|
3130 | if addrevisioncb: | |
3121 | addrevisioncb(self, rev) |
|
3131 | addrevisioncb(self, rev) | |
3122 | empty = False |
|
3132 | empty = False | |
3123 | finally: |
|
3133 | finally: | |
3124 | self._adding_group = False |
|
3134 | self._adding_group = False | |
3125 | return not empty |
|
3135 | return not empty | |
3126 |
|
3136 | |||
3127 | def iscensored(self, rev): |
|
3137 | def iscensored(self, rev): | |
3128 | """Check if a file revision is censored.""" |
|
3138 | """Check if a file revision is censored.""" | |
3129 | if not self._censorable: |
|
3139 | if not self._censorable: | |
3130 | return False |
|
3140 | return False | |
3131 |
|
3141 | |||
3132 | return self.flags(rev) & REVIDX_ISCENSORED |
|
3142 | return self.flags(rev) & REVIDX_ISCENSORED | |
3133 |
|
3143 | |||
3134 | def _peek_iscensored(self, baserev, delta): |
|
3144 | def _peek_iscensored(self, baserev, delta): | |
3135 | """Quickly check if a delta produces a censored revision.""" |
|
3145 | """Quickly check if a delta produces a censored revision.""" | |
3136 | if not self._censorable: |
|
3146 | if not self._censorable: | |
3137 | return False |
|
3147 | return False | |
3138 |
|
3148 | |||
3139 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) |
|
3149 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) | |
3140 |
|
3150 | |||
3141 | def getstrippoint(self, minlink): |
|
3151 | def getstrippoint(self, minlink): | |
3142 | """find the minimum rev that must be stripped to strip the linkrev |
|
3152 | """find the minimum rev that must be stripped to strip the linkrev | |
3143 |
|
3153 | |||
3144 | Returns a tuple containing the minimum rev and a set of all revs that |
|
3154 | Returns a tuple containing the minimum rev and a set of all revs that | |
3145 | have linkrevs that will be broken by this strip. |
|
3155 | have linkrevs that will be broken by this strip. | |
3146 | """ |
|
3156 | """ | |
3147 | return storageutil.resolvestripinfo( |
|
3157 | return storageutil.resolvestripinfo( | |
3148 | minlink, |
|
3158 | minlink, | |
3149 | len(self) - 1, |
|
3159 | len(self) - 1, | |
3150 | self.headrevs(), |
|
3160 | self.headrevs(), | |
3151 | self.linkrev, |
|
3161 | self.linkrev, | |
3152 | self.parentrevs, |
|
3162 | self.parentrevs, | |
3153 | ) |
|
3163 | ) | |
3154 |
|
3164 | |||
3155 | def strip(self, minlink, transaction): |
|
3165 | def strip(self, minlink, transaction): | |
3156 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
3166 | """truncate the revlog on the first revision with a linkrev >= minlink | |
3157 |
|
3167 | |||
3158 | This function is called when we're stripping revision minlink and |
|
3168 | This function is called when we're stripping revision minlink and | |
3159 | its descendants from the repository. |
|
3169 | its descendants from the repository. | |
3160 |
|
3170 | |||
3161 | We have to remove all revisions with linkrev >= minlink, because |
|
3171 | We have to remove all revisions with linkrev >= minlink, because | |
3162 | the equivalent changelog revisions will be renumbered after the |
|
3172 | the equivalent changelog revisions will be renumbered after the | |
3163 | strip. |
|
3173 | strip. | |
3164 |
|
3174 | |||
3165 | So we truncate the revlog on the first of these revisions, and |
|
3175 | So we truncate the revlog on the first of these revisions, and | |
3166 | trust that the caller has saved the revisions that shouldn't be |
|
3176 | trust that the caller has saved the revisions that shouldn't be | |
3167 | removed and that it'll re-add them after this truncation. |
|
3177 | removed and that it'll re-add them after this truncation. | |
3168 | """ |
|
3178 | """ | |
3169 | if len(self) == 0: |
|
3179 | if len(self) == 0: | |
3170 | return |
|
3180 | return | |
3171 |
|
3181 | |||
3172 | rev, _ = self.getstrippoint(minlink) |
|
3182 | rev, _ = self.getstrippoint(minlink) | |
3173 | if rev == len(self): |
|
3183 | if rev == len(self): | |
3174 | return |
|
3184 | return | |
3175 |
|
3185 | |||
3176 | # first truncate the files on disk |
|
3186 | # first truncate the files on disk | |
3177 | data_end = self.start(rev) |
|
3187 | data_end = self.start(rev) | |
3178 | if not self._inline: |
|
3188 | if not self._inline: | |
3179 | transaction.add(self._datafile, data_end) |
|
3189 | transaction.add(self._datafile, data_end) | |
3180 | end = rev * self.index.entry_size |
|
3190 | end = rev * self.index.entry_size | |
3181 | else: |
|
3191 | else: | |
3182 | end = data_end + (rev * self.index.entry_size) |
|
3192 | end = data_end + (rev * self.index.entry_size) | |
3183 |
|
3193 | |||
3184 | if self._sidedatafile: |
|
3194 | if self._sidedatafile: | |
3185 | sidedata_end = self.sidedata_cut_off(rev) |
|
3195 | sidedata_end = self.sidedata_cut_off(rev) | |
3186 | transaction.add(self._sidedatafile, sidedata_end) |
|
3196 | transaction.add(self._sidedatafile, sidedata_end) | |
3187 |
|
3197 | |||
3188 | transaction.add(self._indexfile, end) |
|
3198 | transaction.add(self._indexfile, end) | |
3189 | if self._docket is not None: |
|
3199 | if self._docket is not None: | |
3190 | # XXX we could, leverage the docket while stripping. However it is |
|
3200 | # XXX we could, leverage the docket while stripping. However it is | |
3191 | # not powerfull enough at the time of this comment |
|
3201 | # not powerfull enough at the time of this comment | |
3192 | self._docket.index_end = end |
|
3202 | self._docket.index_end = end | |
3193 | self._docket.data_end = data_end |
|
3203 | self._docket.data_end = data_end | |
3194 | self._docket.sidedata_end = sidedata_end |
|
3204 | self._docket.sidedata_end = sidedata_end | |
3195 | self._docket.write(transaction, stripping=True) |
|
3205 | self._docket.write(transaction, stripping=True) | |
3196 |
|
3206 | |||
3197 | # then reset internal state in memory to forget those revisions |
|
3207 | # then reset internal state in memory to forget those revisions | |
3198 | self._revisioncache = None |
|
3208 | self._revisioncache = None | |
3199 | self._chaininfocache = util.lrucachedict(500) |
|
3209 | self._chaininfocache = util.lrucachedict(500) | |
3200 | self._segmentfile.clear_cache() |
|
3210 | self._segmentfile.clear_cache() | |
3201 | self._segmentfile_sidedata.clear_cache() |
|
3211 | self._segmentfile_sidedata.clear_cache() | |
3202 |
|
3212 | |||
3203 | del self.index[rev:-1] |
|
3213 | del self.index[rev:-1] | |
3204 |
|
3214 | |||
3205 | def checksize(self): |
|
3215 | def checksize(self): | |
3206 | """Check size of index and data files |
|
3216 | """Check size of index and data files | |
3207 |
|
3217 | |||
3208 | return a (dd, di) tuple. |
|
3218 | return a (dd, di) tuple. | |
3209 | - dd: extra bytes for the "data" file |
|
3219 | - dd: extra bytes for the "data" file | |
3210 | - di: extra bytes for the "index" file |
|
3220 | - di: extra bytes for the "index" file | |
3211 |
|
3221 | |||
3212 | A healthy revlog will return (0, 0). |
|
3222 | A healthy revlog will return (0, 0). | |
3213 | """ |
|
3223 | """ | |
3214 | expected = 0 |
|
3224 | expected = 0 | |
3215 | if len(self): |
|
3225 | if len(self): | |
3216 | expected = max(0, self.end(len(self) - 1)) |
|
3226 | expected = max(0, self.end(len(self) - 1)) | |
3217 |
|
3227 | |||
3218 | try: |
|
3228 | try: | |
3219 | with self._datafp() as f: |
|
3229 | with self._datafp() as f: | |
3220 | f.seek(0, io.SEEK_END) |
|
3230 | f.seek(0, io.SEEK_END) | |
3221 | actual = f.tell() |
|
3231 | actual = f.tell() | |
3222 | dd = actual - expected |
|
3232 | dd = actual - expected | |
3223 | except FileNotFoundError: |
|
3233 | except FileNotFoundError: | |
3224 | dd = 0 |
|
3234 | dd = 0 | |
3225 |
|
3235 | |||
3226 | try: |
|
3236 | try: | |
3227 | f = self.opener(self._indexfile) |
|
3237 | f = self.opener(self._indexfile) | |
3228 | f.seek(0, io.SEEK_END) |
|
3238 | f.seek(0, io.SEEK_END) | |
3229 | actual = f.tell() |
|
3239 | actual = f.tell() | |
3230 | f.close() |
|
3240 | f.close() | |
3231 | s = self.index.entry_size |
|
3241 | s = self.index.entry_size | |
3232 | i = max(0, actual // s) |
|
3242 | i = max(0, actual // s) | |
3233 | di = actual - (i * s) |
|
3243 | di = actual - (i * s) | |
3234 | if self._inline: |
|
3244 | if self._inline: | |
3235 | databytes = 0 |
|
3245 | databytes = 0 | |
3236 | for r in self: |
|
3246 | for r in self: | |
3237 | databytes += max(0, self.length(r)) |
|
3247 | databytes += max(0, self.length(r)) | |
3238 | dd = 0 |
|
3248 | dd = 0 | |
3239 | di = actual - len(self) * s - databytes |
|
3249 | di = actual - len(self) * s - databytes | |
3240 | except FileNotFoundError: |
|
3250 | except FileNotFoundError: | |
3241 | di = 0 |
|
3251 | di = 0 | |
3242 |
|
3252 | |||
3243 | return (dd, di) |
|
3253 | return (dd, di) | |
3244 |
|
3254 | |||
3245 | def files(self): |
|
3255 | def files(self): | |
3246 | res = [self._indexfile] |
|
3256 | res = [self._indexfile] | |
3247 | if self._docket_file is None: |
|
3257 | if self._docket_file is None: | |
3248 | if not self._inline: |
|
3258 | if not self._inline: | |
3249 | res.append(self._datafile) |
|
3259 | res.append(self._datafile) | |
3250 | else: |
|
3260 | else: | |
3251 | res.append(self._docket_file) |
|
3261 | res.append(self._docket_file) | |
3252 | res.extend(self._docket.old_index_filepaths(include_empty=False)) |
|
3262 | res.extend(self._docket.old_index_filepaths(include_empty=False)) | |
3253 | if self._docket.data_end: |
|
3263 | if self._docket.data_end: | |
3254 | res.append(self._datafile) |
|
3264 | res.append(self._datafile) | |
3255 | res.extend(self._docket.old_data_filepaths(include_empty=False)) |
|
3265 | res.extend(self._docket.old_data_filepaths(include_empty=False)) | |
3256 | if self._docket.sidedata_end: |
|
3266 | if self._docket.sidedata_end: | |
3257 | res.append(self._sidedatafile) |
|
3267 | res.append(self._sidedatafile) | |
3258 | res.extend(self._docket.old_sidedata_filepaths(include_empty=False)) |
|
3268 | res.extend(self._docket.old_sidedata_filepaths(include_empty=False)) | |
3259 | return res |
|
3269 | return res | |
3260 |
|
3270 | |||
3261 | def emitrevisions( |
|
3271 | def emitrevisions( | |
3262 | self, |
|
3272 | self, | |
3263 | nodes, |
|
3273 | nodes, | |
3264 | nodesorder=None, |
|
3274 | nodesorder=None, | |
3265 | revisiondata=False, |
|
3275 | revisiondata=False, | |
3266 | assumehaveparentrevisions=False, |
|
3276 | assumehaveparentrevisions=False, | |
3267 | deltamode=repository.CG_DELTAMODE_STD, |
|
3277 | deltamode=repository.CG_DELTAMODE_STD, | |
3268 | sidedata_helpers=None, |
|
3278 | sidedata_helpers=None, | |
3269 | debug_info=None, |
|
3279 | debug_info=None, | |
3270 | ): |
|
3280 | ): | |
3271 | if nodesorder not in (b'nodes', b'storage', b'linear', None): |
|
3281 | if nodesorder not in (b'nodes', b'storage', b'linear', None): | |
3272 | raise error.ProgrammingError( |
|
3282 | raise error.ProgrammingError( | |
3273 | b'unhandled value for nodesorder: %s' % nodesorder |
|
3283 | b'unhandled value for nodesorder: %s' % nodesorder | |
3274 | ) |
|
3284 | ) | |
3275 |
|
3285 | |||
3276 | if nodesorder is None and not self._generaldelta: |
|
3286 | if nodesorder is None and not self._generaldelta: | |
3277 | nodesorder = b'storage' |
|
3287 | nodesorder = b'storage' | |
3278 |
|
3288 | |||
3279 | if ( |
|
3289 | if ( | |
3280 | not self._storedeltachains |
|
3290 | not self._storedeltachains | |
3281 | and deltamode != repository.CG_DELTAMODE_PREV |
|
3291 | and deltamode != repository.CG_DELTAMODE_PREV | |
3282 | ): |
|
3292 | ): | |
3283 | deltamode = repository.CG_DELTAMODE_FULL |
|
3293 | deltamode = repository.CG_DELTAMODE_FULL | |
3284 |
|
3294 | |||
3285 | return storageutil.emitrevisions( |
|
3295 | return storageutil.emitrevisions( | |
3286 | self, |
|
3296 | self, | |
3287 | nodes, |
|
3297 | nodes, | |
3288 | nodesorder, |
|
3298 | nodesorder, | |
3289 | revlogrevisiondelta, |
|
3299 | revlogrevisiondelta, | |
3290 | deltaparentfn=self.deltaparent, |
|
3300 | deltaparentfn=self.deltaparent, | |
3291 | candeltafn=self._candelta, |
|
3301 | candeltafn=self._candelta, | |
3292 | rawsizefn=self.rawsize, |
|
3302 | rawsizefn=self.rawsize, | |
3293 | revdifffn=self.revdiff, |
|
3303 | revdifffn=self.revdiff, | |
3294 | flagsfn=self.flags, |
|
3304 | flagsfn=self.flags, | |
3295 | deltamode=deltamode, |
|
3305 | deltamode=deltamode, | |
3296 | revisiondata=revisiondata, |
|
3306 | revisiondata=revisiondata, | |
3297 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
3307 | assumehaveparentrevisions=assumehaveparentrevisions, | |
3298 | sidedata_helpers=sidedata_helpers, |
|
3308 | sidedata_helpers=sidedata_helpers, | |
3299 | debug_info=debug_info, |
|
3309 | debug_info=debug_info, | |
3300 | ) |
|
3310 | ) | |
3301 |
|
3311 | |||
3302 | DELTAREUSEALWAYS = b'always' |
|
3312 | DELTAREUSEALWAYS = b'always' | |
3303 | DELTAREUSESAMEREVS = b'samerevs' |
|
3313 | DELTAREUSESAMEREVS = b'samerevs' | |
3304 | DELTAREUSENEVER = b'never' |
|
3314 | DELTAREUSENEVER = b'never' | |
3305 |
|
3315 | |||
3306 | DELTAREUSEFULLADD = b'fulladd' |
|
3316 | DELTAREUSEFULLADD = b'fulladd' | |
3307 |
|
3317 | |||
3308 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} |
|
3318 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} | |
3309 |
|
3319 | |||
3310 | def clone( |
|
3320 | def clone( | |
3311 | self, |
|
3321 | self, | |
3312 | tr, |
|
3322 | tr, | |
3313 | destrevlog, |
|
3323 | destrevlog, | |
3314 | addrevisioncb=None, |
|
3324 | addrevisioncb=None, | |
3315 | deltareuse=DELTAREUSESAMEREVS, |
|
3325 | deltareuse=DELTAREUSESAMEREVS, | |
3316 | forcedeltabothparents=None, |
|
3326 | forcedeltabothparents=None, | |
3317 | sidedata_helpers=None, |
|
3327 | sidedata_helpers=None, | |
3318 | ): |
|
3328 | ): | |
3319 | """Copy this revlog to another, possibly with format changes. |
|
3329 | """Copy this revlog to another, possibly with format changes. | |
3320 |
|
3330 | |||
3321 | The destination revlog will contain the same revisions and nodes. |
|
3331 | The destination revlog will contain the same revisions and nodes. | |
3322 | However, it may not be bit-for-bit identical due to e.g. delta encoding |
|
3332 | However, it may not be bit-for-bit identical due to e.g. delta encoding | |
3323 | differences. |
|
3333 | differences. | |
3324 |
|
3334 | |||
3325 | The ``deltareuse`` argument control how deltas from the existing revlog |
|
3335 | The ``deltareuse`` argument control how deltas from the existing revlog | |
3326 | are preserved in the destination revlog. The argument can have the |
|
3336 | are preserved in the destination revlog. The argument can have the | |
3327 | following values: |
|
3337 | following values: | |
3328 |
|
3338 | |||
3329 | DELTAREUSEALWAYS |
|
3339 | DELTAREUSEALWAYS | |
3330 | Deltas will always be reused (if possible), even if the destination |
|
3340 | Deltas will always be reused (if possible), even if the destination | |
3331 | revlog would not select the same revisions for the delta. This is the |
|
3341 | revlog would not select the same revisions for the delta. This is the | |
3332 | fastest mode of operation. |
|
3342 | fastest mode of operation. | |
3333 | DELTAREUSESAMEREVS |
|
3343 | DELTAREUSESAMEREVS | |
3334 | Deltas will be reused if the destination revlog would pick the same |
|
3344 | Deltas will be reused if the destination revlog would pick the same | |
3335 | revisions for the delta. This mode strikes a balance between speed |
|
3345 | revisions for the delta. This mode strikes a balance between speed | |
3336 | and optimization. |
|
3346 | and optimization. | |
3337 | DELTAREUSENEVER |
|
3347 | DELTAREUSENEVER | |
3338 | Deltas will never be reused. This is the slowest mode of execution. |
|
3348 | Deltas will never be reused. This is the slowest mode of execution. | |
3339 | This mode can be used to recompute deltas (e.g. if the diff/delta |
|
3349 | This mode can be used to recompute deltas (e.g. if the diff/delta | |
3340 | algorithm changes). |
|
3350 | algorithm changes). | |
3341 | DELTAREUSEFULLADD |
|
3351 | DELTAREUSEFULLADD | |
3342 | Revision will be re-added as if their were new content. This is |
|
3352 | Revision will be re-added as if their were new content. This is | |
3343 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. |
|
3353 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. | |
3344 | eg: large file detection and handling. |
|
3354 | eg: large file detection and handling. | |
3345 |
|
3355 | |||
3346 | Delta computation can be slow, so the choice of delta reuse policy can |
|
3356 | Delta computation can be slow, so the choice of delta reuse policy can | |
3347 | significantly affect run time. |
|
3357 | significantly affect run time. | |
3348 |
|
3358 | |||
3349 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between |
|
3359 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between | |
3350 | two extremes. Deltas will be reused if they are appropriate. But if the |
|
3360 | two extremes. Deltas will be reused if they are appropriate. But if the | |
3351 | delta could choose a better revision, it will do so. This means if you |
|
3361 | delta could choose a better revision, it will do so. This means if you | |
3352 | are converting a non-generaldelta revlog to a generaldelta revlog, |
|
3362 | are converting a non-generaldelta revlog to a generaldelta revlog, | |
3353 | deltas will be recomputed if the delta's parent isn't a parent of the |
|
3363 | deltas will be recomputed if the delta's parent isn't a parent of the | |
3354 | revision. |
|
3364 | revision. | |
3355 |
|
3365 | |||
3356 | In addition to the delta policy, the ``forcedeltabothparents`` |
|
3366 | In addition to the delta policy, the ``forcedeltabothparents`` | |
3357 | argument controls whether to force compute deltas against both parents |
|
3367 | argument controls whether to force compute deltas against both parents | |
3358 | for merges. By default, the current default is used. |
|
3368 | for merges. By default, the current default is used. | |
3359 |
|
3369 | |||
3360 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on |
|
3370 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on | |
3361 | `sidedata_helpers`. |
|
3371 | `sidedata_helpers`. | |
3362 | """ |
|
3372 | """ | |
3363 | if deltareuse not in self.DELTAREUSEALL: |
|
3373 | if deltareuse not in self.DELTAREUSEALL: | |
3364 | raise ValueError( |
|
3374 | raise ValueError( | |
3365 | _(b'value for deltareuse invalid: %s') % deltareuse |
|
3375 | _(b'value for deltareuse invalid: %s') % deltareuse | |
3366 | ) |
|
3376 | ) | |
3367 |
|
3377 | |||
3368 | if len(destrevlog): |
|
3378 | if len(destrevlog): | |
3369 | raise ValueError(_(b'destination revlog is not empty')) |
|
3379 | raise ValueError(_(b'destination revlog is not empty')) | |
3370 |
|
3380 | |||
3371 | if getattr(self, 'filteredrevs', None): |
|
3381 | if getattr(self, 'filteredrevs', None): | |
3372 | raise ValueError(_(b'source revlog has filtered revisions')) |
|
3382 | raise ValueError(_(b'source revlog has filtered revisions')) | |
3373 | if getattr(destrevlog, 'filteredrevs', None): |
|
3383 | if getattr(destrevlog, 'filteredrevs', None): | |
3374 | raise ValueError(_(b'destination revlog has filtered revisions')) |
|
3384 | raise ValueError(_(b'destination revlog has filtered revisions')) | |
3375 |
|
3385 | |||
3376 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, |
|
3386 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, | |
3377 | # if possible. |
|
3387 | # if possible. | |
3378 |
old |
|
3388 | old_delta_config = destrevlog.delta_config | |
3379 | oldlazydeltabase = destrevlog._lazydeltabase |
|
3389 | destrevlog.delta_config = destrevlog.delta_config.copy() | |
3380 | oldamd = destrevlog._deltabothparents |
|
|||
3381 |
|
3390 | |||
3382 | try: |
|
3391 | try: | |
3383 | if deltareuse == self.DELTAREUSEALWAYS: |
|
3392 | if deltareuse == self.DELTAREUSEALWAYS: | |
3384 | destrevlog.delta_config.lazy_delta_base = True |
|
3393 | destrevlog.delta_config.lazy_delta_base = True | |
3385 | destrevlog.delta_config.lazy_delta = True |
|
3394 | destrevlog.delta_config.lazy_delta = True | |
3386 | elif deltareuse == self.DELTAREUSESAMEREVS: |
|
3395 | elif deltareuse == self.DELTAREUSESAMEREVS: | |
3387 | destrevlog.delta_config.lazy_delta_base = False |
|
3396 | destrevlog.delta_config.lazy_delta_base = False | |
3388 | destrevlog.delta_config.lazy_delta = True |
|
3397 | destrevlog.delta_config.lazy_delta = True | |
3389 | elif deltareuse == self.DELTAREUSENEVER: |
|
3398 | elif deltareuse == self.DELTAREUSENEVER: | |
3390 | destrevlog.delta_config.lazy_delta_base = False |
|
3399 | destrevlog.delta_config.lazy_delta_base = False | |
3391 | destrevlog.delta_config.lazy_delta = False |
|
3400 | destrevlog.delta_config.lazy_delta = False | |
3392 |
|
3401 | |||
3393 |
delta_both_parents = |
|
3402 | delta_both_parents = ( | |
|
3403 | forcedeltabothparents or old_delta_config.delta_both_parents | |||
|
3404 | ) | |||
3394 | destrevlog.delta_config.delta_both_parents = delta_both_parents |
|
3405 | destrevlog.delta_config.delta_both_parents = delta_both_parents | |
3395 |
|
3406 | |||
3396 | with self.reading(): |
|
3407 | with self.reading(): | |
3397 | self._clone( |
|
3408 | self._clone( | |
3398 | tr, |
|
3409 | tr, | |
3399 | destrevlog, |
|
3410 | destrevlog, | |
3400 | addrevisioncb, |
|
3411 | addrevisioncb, | |
3401 | deltareuse, |
|
3412 | deltareuse, | |
3402 | forcedeltabothparents, |
|
3413 | forcedeltabothparents, | |
3403 | sidedata_helpers, |
|
3414 | sidedata_helpers, | |
3404 | ) |
|
3415 | ) | |
3405 |
|
3416 | |||
3406 | finally: |
|
3417 | finally: | |
3407 |
destrevlog.delta_config |
|
3418 | destrevlog.delta_config = old_delta_config | |
3408 | destrevlog.delta_config.lazy_delta_base = oldlazydeltabase |
|
|||
3409 | destrevlog.delta_config.delta_both_parents = oldamd |
|
|||
3410 |
|
3419 | |||
3411 | def _clone( |
|
3420 | def _clone( | |
3412 | self, |
|
3421 | self, | |
3413 | tr, |
|
3422 | tr, | |
3414 | destrevlog, |
|
3423 | destrevlog, | |
3415 | addrevisioncb, |
|
3424 | addrevisioncb, | |
3416 | deltareuse, |
|
3425 | deltareuse, | |
3417 | forcedeltabothparents, |
|
3426 | forcedeltabothparents, | |
3418 | sidedata_helpers, |
|
3427 | sidedata_helpers, | |
3419 | ): |
|
3428 | ): | |
3420 | """perform the core duty of `revlog.clone` after parameter processing""" |
|
3429 | """perform the core duty of `revlog.clone` after parameter processing""" | |
3421 | write_debug = None |
|
3430 | write_debug = None | |
3422 | if self._debug_delta: |
|
3431 | if self._debug_delta: | |
3423 | write_debug = tr._report |
|
3432 | write_debug = tr._report | |
3424 | deltacomputer = deltautil.deltacomputer( |
|
3433 | deltacomputer = deltautil.deltacomputer( | |
3425 | destrevlog, |
|
3434 | destrevlog, | |
3426 | write_debug=write_debug, |
|
3435 | write_debug=write_debug, | |
3427 | ) |
|
3436 | ) | |
3428 | index = self.index |
|
3437 | index = self.index | |
3429 | for rev in self: |
|
3438 | for rev in self: | |
3430 | entry = index[rev] |
|
3439 | entry = index[rev] | |
3431 |
|
3440 | |||
3432 | # Some classes override linkrev to take filtered revs into |
|
3441 | # Some classes override linkrev to take filtered revs into | |
3433 | # account. Use raw entry from index. |
|
3442 | # account. Use raw entry from index. | |
3434 | flags = entry[0] & 0xFFFF |
|
3443 | flags = entry[0] & 0xFFFF | |
3435 | linkrev = entry[4] |
|
3444 | linkrev = entry[4] | |
3436 | p1 = index[entry[5]][7] |
|
3445 | p1 = index[entry[5]][7] | |
3437 | p2 = index[entry[6]][7] |
|
3446 | p2 = index[entry[6]][7] | |
3438 | node = entry[7] |
|
3447 | node = entry[7] | |
3439 |
|
3448 | |||
3440 | # (Possibly) reuse the delta from the revlog if allowed and |
|
3449 | # (Possibly) reuse the delta from the revlog if allowed and | |
3441 | # the revlog chunk is a delta. |
|
3450 | # the revlog chunk is a delta. | |
3442 | cachedelta = None |
|
3451 | cachedelta = None | |
3443 | rawtext = None |
|
3452 | rawtext = None | |
3444 | if deltareuse == self.DELTAREUSEFULLADD: |
|
3453 | if deltareuse == self.DELTAREUSEFULLADD: | |
3445 | text = self._revisiondata(rev) |
|
3454 | text = self._revisiondata(rev) | |
3446 | sidedata = self.sidedata(rev) |
|
3455 | sidedata = self.sidedata(rev) | |
3447 |
|
3456 | |||
3448 | if sidedata_helpers is not None: |
|
3457 | if sidedata_helpers is not None: | |
3449 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
3458 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
3450 | self, sidedata_helpers, sidedata, rev |
|
3459 | self, sidedata_helpers, sidedata, rev | |
3451 | ) |
|
3460 | ) | |
3452 | flags = flags | new_flags[0] & ~new_flags[1] |
|
3461 | flags = flags | new_flags[0] & ~new_flags[1] | |
3453 |
|
3462 | |||
3454 | destrevlog.addrevision( |
|
3463 | destrevlog.addrevision( | |
3455 | text, |
|
3464 | text, | |
3456 | tr, |
|
3465 | tr, | |
3457 | linkrev, |
|
3466 | linkrev, | |
3458 | p1, |
|
3467 | p1, | |
3459 | p2, |
|
3468 | p2, | |
3460 | cachedelta=cachedelta, |
|
3469 | cachedelta=cachedelta, | |
3461 | node=node, |
|
3470 | node=node, | |
3462 | flags=flags, |
|
3471 | flags=flags, | |
3463 | deltacomputer=deltacomputer, |
|
3472 | deltacomputer=deltacomputer, | |
3464 | sidedata=sidedata, |
|
3473 | sidedata=sidedata, | |
3465 | ) |
|
3474 | ) | |
3466 | else: |
|
3475 | else: | |
3467 | if destrevlog._lazydelta: |
|
3476 | if destrevlog._lazydelta: | |
3468 | dp = self.deltaparent(rev) |
|
3477 | dp = self.deltaparent(rev) | |
3469 | if dp != nullrev: |
|
3478 | if dp != nullrev: | |
3470 | cachedelta = (dp, bytes(self._chunk(rev))) |
|
3479 | cachedelta = (dp, bytes(self._chunk(rev))) | |
3471 |
|
3480 | |||
3472 | sidedata = None |
|
3481 | sidedata = None | |
3473 | if not cachedelta: |
|
3482 | if not cachedelta: | |
3474 | rawtext = self._revisiondata(rev) |
|
3483 | rawtext = self._revisiondata(rev) | |
3475 | sidedata = self.sidedata(rev) |
|
3484 | sidedata = self.sidedata(rev) | |
3476 | if sidedata is None: |
|
3485 | if sidedata is None: | |
3477 | sidedata = self.sidedata(rev) |
|
3486 | sidedata = self.sidedata(rev) | |
3478 |
|
3487 | |||
3479 | if sidedata_helpers is not None: |
|
3488 | if sidedata_helpers is not None: | |
3480 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
3489 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
3481 | self, sidedata_helpers, sidedata, rev |
|
3490 | self, sidedata_helpers, sidedata, rev | |
3482 | ) |
|
3491 | ) | |
3483 | flags = flags | new_flags[0] & ~new_flags[1] |
|
3492 | flags = flags | new_flags[0] & ~new_flags[1] | |
3484 |
|
3493 | |||
3485 | with destrevlog._writing(tr): |
|
3494 | with destrevlog._writing(tr): | |
3486 | destrevlog._addrevision( |
|
3495 | destrevlog._addrevision( | |
3487 | node, |
|
3496 | node, | |
3488 | rawtext, |
|
3497 | rawtext, | |
3489 | tr, |
|
3498 | tr, | |
3490 | linkrev, |
|
3499 | linkrev, | |
3491 | p1, |
|
3500 | p1, | |
3492 | p2, |
|
3501 | p2, | |
3493 | flags, |
|
3502 | flags, | |
3494 | cachedelta, |
|
3503 | cachedelta, | |
3495 | deltacomputer=deltacomputer, |
|
3504 | deltacomputer=deltacomputer, | |
3496 | sidedata=sidedata, |
|
3505 | sidedata=sidedata, | |
3497 | ) |
|
3506 | ) | |
3498 |
|
3507 | |||
3499 | if addrevisioncb: |
|
3508 | if addrevisioncb: | |
3500 | addrevisioncb(self, rev, node) |
|
3509 | addrevisioncb(self, rev, node) | |
3501 |
|
3510 | |||
3502 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
3511 | def censorrevision(self, tr, censornode, tombstone=b''): | |
3503 | if self._format_version == REVLOGV0: |
|
3512 | if self._format_version == REVLOGV0: | |
3504 | raise error.RevlogError( |
|
3513 | raise error.RevlogError( | |
3505 | _(b'cannot censor with version %d revlogs') |
|
3514 | _(b'cannot censor with version %d revlogs') | |
3506 | % self._format_version |
|
3515 | % self._format_version | |
3507 | ) |
|
3516 | ) | |
3508 | elif self._format_version == REVLOGV1: |
|
3517 | elif self._format_version == REVLOGV1: | |
3509 | rewrite.v1_censor(self, tr, censornode, tombstone) |
|
3518 | rewrite.v1_censor(self, tr, censornode, tombstone) | |
3510 | else: |
|
3519 | else: | |
3511 | rewrite.v2_censor(self, tr, censornode, tombstone) |
|
3520 | rewrite.v2_censor(self, tr, censornode, tombstone) | |
3512 |
|
3521 | |||
3513 | def verifyintegrity(self, state): |
|
3522 | def verifyintegrity(self, state): | |
3514 | """Verifies the integrity of the revlog. |
|
3523 | """Verifies the integrity of the revlog. | |
3515 |
|
3524 | |||
3516 | Yields ``revlogproblem`` instances describing problems that are |
|
3525 | Yields ``revlogproblem`` instances describing problems that are | |
3517 | found. |
|
3526 | found. | |
3518 | """ |
|
3527 | """ | |
3519 | dd, di = self.checksize() |
|
3528 | dd, di = self.checksize() | |
3520 | if dd: |
|
3529 | if dd: | |
3521 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) |
|
3530 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) | |
3522 | if di: |
|
3531 | if di: | |
3523 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) |
|
3532 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) | |
3524 |
|
3533 | |||
3525 | version = self._format_version |
|
3534 | version = self._format_version | |
3526 |
|
3535 | |||
3527 | # The verifier tells us what version revlog we should be. |
|
3536 | # The verifier tells us what version revlog we should be. | |
3528 | if version != state[b'expectedversion']: |
|
3537 | if version != state[b'expectedversion']: | |
3529 | yield revlogproblem( |
|
3538 | yield revlogproblem( | |
3530 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") |
|
3539 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") | |
3531 | % (self.display_id, version, state[b'expectedversion']) |
|
3540 | % (self.display_id, version, state[b'expectedversion']) | |
3532 | ) |
|
3541 | ) | |
3533 |
|
3542 | |||
3534 | state[b'skipread'] = set() |
|
3543 | state[b'skipread'] = set() | |
3535 | state[b'safe_renamed'] = set() |
|
3544 | state[b'safe_renamed'] = set() | |
3536 |
|
3545 | |||
3537 | for rev in self: |
|
3546 | for rev in self: | |
3538 | node = self.node(rev) |
|
3547 | node = self.node(rev) | |
3539 |
|
3548 | |||
3540 | # Verify contents. 4 cases to care about: |
|
3549 | # Verify contents. 4 cases to care about: | |
3541 | # |
|
3550 | # | |
3542 | # common: the most common case |
|
3551 | # common: the most common case | |
3543 | # rename: with a rename |
|
3552 | # rename: with a rename | |
3544 | # meta: file content starts with b'\1\n', the metadata |
|
3553 | # meta: file content starts with b'\1\n', the metadata | |
3545 | # header defined in filelog.py, but without a rename |
|
3554 | # header defined in filelog.py, but without a rename | |
3546 | # ext: content stored externally |
|
3555 | # ext: content stored externally | |
3547 | # |
|
3556 | # | |
3548 | # More formally, their differences are shown below: |
|
3557 | # More formally, their differences are shown below: | |
3549 | # |
|
3558 | # | |
3550 | # | common | rename | meta | ext |
|
3559 | # | common | rename | meta | ext | |
3551 | # ------------------------------------------------------- |
|
3560 | # ------------------------------------------------------- | |
3552 | # flags() | 0 | 0 | 0 | not 0 |
|
3561 | # flags() | 0 | 0 | 0 | not 0 | |
3553 | # renamed() | False | True | False | ? |
|
3562 | # renamed() | False | True | False | ? | |
3554 | # rawtext[0:2]=='\1\n'| False | True | True | ? |
|
3563 | # rawtext[0:2]=='\1\n'| False | True | True | ? | |
3555 | # |
|
3564 | # | |
3556 | # "rawtext" means the raw text stored in revlog data, which |
|
3565 | # "rawtext" means the raw text stored in revlog data, which | |
3557 | # could be retrieved by "rawdata(rev)". "text" |
|
3566 | # could be retrieved by "rawdata(rev)". "text" | |
3558 | # mentioned below is "revision(rev)". |
|
3567 | # mentioned below is "revision(rev)". | |
3559 | # |
|
3568 | # | |
3560 | # There are 3 different lengths stored physically: |
|
3569 | # There are 3 different lengths stored physically: | |
3561 | # 1. L1: rawsize, stored in revlog index |
|
3570 | # 1. L1: rawsize, stored in revlog index | |
3562 | # 2. L2: len(rawtext), stored in revlog data |
|
3571 | # 2. L2: len(rawtext), stored in revlog data | |
3563 | # 3. L3: len(text), stored in revlog data if flags==0, or |
|
3572 | # 3. L3: len(text), stored in revlog data if flags==0, or | |
3564 | # possibly somewhere else if flags!=0 |
|
3573 | # possibly somewhere else if flags!=0 | |
3565 | # |
|
3574 | # | |
3566 | # L1 should be equal to L2. L3 could be different from them. |
|
3575 | # L1 should be equal to L2. L3 could be different from them. | |
3567 | # "text" may or may not affect commit hash depending on flag |
|
3576 | # "text" may or may not affect commit hash depending on flag | |
3568 | # processors (see flagutil.addflagprocessor). |
|
3577 | # processors (see flagutil.addflagprocessor). | |
3569 | # |
|
3578 | # | |
3570 | # | common | rename | meta | ext |
|
3579 | # | common | rename | meta | ext | |
3571 | # ------------------------------------------------- |
|
3580 | # ------------------------------------------------- | |
3572 | # rawsize() | L1 | L1 | L1 | L1 |
|
3581 | # rawsize() | L1 | L1 | L1 | L1 | |
3573 | # size() | L1 | L2-LM | L1(*) | L1 (?) |
|
3582 | # size() | L1 | L2-LM | L1(*) | L1 (?) | |
3574 | # len(rawtext) | L2 | L2 | L2 | L2 |
|
3583 | # len(rawtext) | L2 | L2 | L2 | L2 | |
3575 | # len(text) | L2 | L2 | L2 | L3 |
|
3584 | # len(text) | L2 | L2 | L2 | L3 | |
3576 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) |
|
3585 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) | |
3577 | # |
|
3586 | # | |
3578 | # LM: length of metadata, depending on rawtext |
|
3587 | # LM: length of metadata, depending on rawtext | |
3579 | # (*): not ideal, see comment in filelog.size |
|
3588 | # (*): not ideal, see comment in filelog.size | |
3580 | # (?): could be "- len(meta)" if the resolved content has |
|
3589 | # (?): could be "- len(meta)" if the resolved content has | |
3581 | # rename metadata |
|
3590 | # rename metadata | |
3582 | # |
|
3591 | # | |
3583 | # Checks needed to be done: |
|
3592 | # Checks needed to be done: | |
3584 | # 1. length check: L1 == L2, in all cases. |
|
3593 | # 1. length check: L1 == L2, in all cases. | |
3585 | # 2. hash check: depending on flag processor, we may need to |
|
3594 | # 2. hash check: depending on flag processor, we may need to | |
3586 | # use either "text" (external), or "rawtext" (in revlog). |
|
3595 | # use either "text" (external), or "rawtext" (in revlog). | |
3587 |
|
3596 | |||
3588 | try: |
|
3597 | try: | |
3589 | skipflags = state.get(b'skipflags', 0) |
|
3598 | skipflags = state.get(b'skipflags', 0) | |
3590 | if skipflags: |
|
3599 | if skipflags: | |
3591 | skipflags &= self.flags(rev) |
|
3600 | skipflags &= self.flags(rev) | |
3592 |
|
3601 | |||
3593 | _verify_revision(self, skipflags, state, node) |
|
3602 | _verify_revision(self, skipflags, state, node) | |
3594 |
|
3603 | |||
3595 | l1 = self.rawsize(rev) |
|
3604 | l1 = self.rawsize(rev) | |
3596 | l2 = len(self.rawdata(node)) |
|
3605 | l2 = len(self.rawdata(node)) | |
3597 |
|
3606 | |||
3598 | if l1 != l2: |
|
3607 | if l1 != l2: | |
3599 | yield revlogproblem( |
|
3608 | yield revlogproblem( | |
3600 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), |
|
3609 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), | |
3601 | node=node, |
|
3610 | node=node, | |
3602 | ) |
|
3611 | ) | |
3603 |
|
3612 | |||
3604 | except error.CensoredNodeError: |
|
3613 | except error.CensoredNodeError: | |
3605 | if state[b'erroroncensored']: |
|
3614 | if state[b'erroroncensored']: | |
3606 | yield revlogproblem( |
|
3615 | yield revlogproblem( | |
3607 | error=_(b'censored file data'), node=node |
|
3616 | error=_(b'censored file data'), node=node | |
3608 | ) |
|
3617 | ) | |
3609 | state[b'skipread'].add(node) |
|
3618 | state[b'skipread'].add(node) | |
3610 | except Exception as e: |
|
3619 | except Exception as e: | |
3611 | yield revlogproblem( |
|
3620 | yield revlogproblem( | |
3612 | error=_(b'unpacking %s: %s') |
|
3621 | error=_(b'unpacking %s: %s') | |
3613 | % (short(node), stringutil.forcebytestr(e)), |
|
3622 | % (short(node), stringutil.forcebytestr(e)), | |
3614 | node=node, |
|
3623 | node=node, | |
3615 | ) |
|
3624 | ) | |
3616 | state[b'skipread'].add(node) |
|
3625 | state[b'skipread'].add(node) | |
3617 |
|
3626 | |||
3618 | def storageinfo( |
|
3627 | def storageinfo( | |
3619 | self, |
|
3628 | self, | |
3620 | exclusivefiles=False, |
|
3629 | exclusivefiles=False, | |
3621 | sharedfiles=False, |
|
3630 | sharedfiles=False, | |
3622 | revisionscount=False, |
|
3631 | revisionscount=False, | |
3623 | trackedsize=False, |
|
3632 | trackedsize=False, | |
3624 | storedsize=False, |
|
3633 | storedsize=False, | |
3625 | ): |
|
3634 | ): | |
3626 | d = {} |
|
3635 | d = {} | |
3627 |
|
3636 | |||
3628 | if exclusivefiles: |
|
3637 | if exclusivefiles: | |
3629 | d[b'exclusivefiles'] = [(self.opener, self._indexfile)] |
|
3638 | d[b'exclusivefiles'] = [(self.opener, self._indexfile)] | |
3630 | if not self._inline: |
|
3639 | if not self._inline: | |
3631 | d[b'exclusivefiles'].append((self.opener, self._datafile)) |
|
3640 | d[b'exclusivefiles'].append((self.opener, self._datafile)) | |
3632 |
|
3641 | |||
3633 | if sharedfiles: |
|
3642 | if sharedfiles: | |
3634 | d[b'sharedfiles'] = [] |
|
3643 | d[b'sharedfiles'] = [] | |
3635 |
|
3644 | |||
3636 | if revisionscount: |
|
3645 | if revisionscount: | |
3637 | d[b'revisionscount'] = len(self) |
|
3646 | d[b'revisionscount'] = len(self) | |
3638 |
|
3647 | |||
3639 | if trackedsize: |
|
3648 | if trackedsize: | |
3640 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) |
|
3649 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) | |
3641 |
|
3650 | |||
3642 | if storedsize: |
|
3651 | if storedsize: | |
3643 | d[b'storedsize'] = sum( |
|
3652 | d[b'storedsize'] = sum( | |
3644 | self.opener.stat(path).st_size for path in self.files() |
|
3653 | self.opener.stat(path).st_size for path in self.files() | |
3645 | ) |
|
3654 | ) | |
3646 |
|
3655 | |||
3647 | return d |
|
3656 | return d | |
3648 |
|
3657 | |||
3649 | def rewrite_sidedata(self, transaction, helpers, startrev, endrev): |
|
3658 | def rewrite_sidedata(self, transaction, helpers, startrev, endrev): | |
3650 | if not self.hassidedata: |
|
3659 | if not self.hassidedata: | |
3651 | return |
|
3660 | return | |
3652 | # revlog formats with sidedata support does not support inline |
|
3661 | # revlog formats with sidedata support does not support inline | |
3653 | assert not self._inline |
|
3662 | assert not self._inline | |
3654 | if not helpers[1] and not helpers[2]: |
|
3663 | if not helpers[1] and not helpers[2]: | |
3655 | # Nothing to generate or remove |
|
3664 | # Nothing to generate or remove | |
3656 | return |
|
3665 | return | |
3657 |
|
3666 | |||
3658 | new_entries = [] |
|
3667 | new_entries = [] | |
3659 | # append the new sidedata |
|
3668 | # append the new sidedata | |
3660 | with self._writing(transaction): |
|
3669 | with self._writing(transaction): | |
3661 | ifh, dfh, sdfh = self._writinghandles |
|
3670 | ifh, dfh, sdfh = self._writinghandles | |
3662 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) |
|
3671 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) | |
3663 |
|
3672 | |||
3664 | current_offset = sdfh.tell() |
|
3673 | current_offset = sdfh.tell() | |
3665 | for rev in range(startrev, endrev + 1): |
|
3674 | for rev in range(startrev, endrev + 1): | |
3666 | entry = self.index[rev] |
|
3675 | entry = self.index[rev] | |
3667 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( |
|
3676 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( | |
3668 | store=self, |
|
3677 | store=self, | |
3669 | sidedata_helpers=helpers, |
|
3678 | sidedata_helpers=helpers, | |
3670 | sidedata={}, |
|
3679 | sidedata={}, | |
3671 | rev=rev, |
|
3680 | rev=rev, | |
3672 | ) |
|
3681 | ) | |
3673 |
|
3682 | |||
3674 | serialized_sidedata = sidedatautil.serialize_sidedata( |
|
3683 | serialized_sidedata = sidedatautil.serialize_sidedata( | |
3675 | new_sidedata |
|
3684 | new_sidedata | |
3676 | ) |
|
3685 | ) | |
3677 |
|
3686 | |||
3678 | sidedata_compression_mode = COMP_MODE_INLINE |
|
3687 | sidedata_compression_mode = COMP_MODE_INLINE | |
3679 | if serialized_sidedata and self.hassidedata: |
|
3688 | if serialized_sidedata and self.hassidedata: | |
3680 | sidedata_compression_mode = COMP_MODE_PLAIN |
|
3689 | sidedata_compression_mode = COMP_MODE_PLAIN | |
3681 | h, comp_sidedata = self.compress(serialized_sidedata) |
|
3690 | h, comp_sidedata = self.compress(serialized_sidedata) | |
3682 | if ( |
|
3691 | if ( | |
3683 | h != b'u' |
|
3692 | h != b'u' | |
3684 | and comp_sidedata[0] != b'\0' |
|
3693 | and comp_sidedata[0] != b'\0' | |
3685 | and len(comp_sidedata) < len(serialized_sidedata) |
|
3694 | and len(comp_sidedata) < len(serialized_sidedata) | |
3686 | ): |
|
3695 | ): | |
3687 | assert not h |
|
3696 | assert not h | |
3688 | if ( |
|
3697 | if ( | |
3689 | comp_sidedata[0] |
|
3698 | comp_sidedata[0] | |
3690 | == self._docket.default_compression_header |
|
3699 | == self._docket.default_compression_header | |
3691 | ): |
|
3700 | ): | |
3692 | sidedata_compression_mode = COMP_MODE_DEFAULT |
|
3701 | sidedata_compression_mode = COMP_MODE_DEFAULT | |
3693 | serialized_sidedata = comp_sidedata |
|
3702 | serialized_sidedata = comp_sidedata | |
3694 | else: |
|
3703 | else: | |
3695 | sidedata_compression_mode = COMP_MODE_INLINE |
|
3704 | sidedata_compression_mode = COMP_MODE_INLINE | |
3696 | serialized_sidedata = comp_sidedata |
|
3705 | serialized_sidedata = comp_sidedata | |
3697 | if entry[8] != 0 or entry[9] != 0: |
|
3706 | if entry[8] != 0 or entry[9] != 0: | |
3698 | # rewriting entries that already have sidedata is not |
|
3707 | # rewriting entries that already have sidedata is not | |
3699 | # supported yet, because it introduces garbage data in the |
|
3708 | # supported yet, because it introduces garbage data in the | |
3700 | # revlog. |
|
3709 | # revlog. | |
3701 | msg = b"rewriting existing sidedata is not supported yet" |
|
3710 | msg = b"rewriting existing sidedata is not supported yet" | |
3702 | raise error.Abort(msg) |
|
3711 | raise error.Abort(msg) | |
3703 |
|
3712 | |||
3704 | # Apply (potential) flags to add and to remove after running |
|
3713 | # Apply (potential) flags to add and to remove after running | |
3705 | # the sidedata helpers |
|
3714 | # the sidedata helpers | |
3706 | new_offset_flags = entry[0] | flags[0] & ~flags[1] |
|
3715 | new_offset_flags = entry[0] | flags[0] & ~flags[1] | |
3707 | entry_update = ( |
|
3716 | entry_update = ( | |
3708 | current_offset, |
|
3717 | current_offset, | |
3709 | len(serialized_sidedata), |
|
3718 | len(serialized_sidedata), | |
3710 | new_offset_flags, |
|
3719 | new_offset_flags, | |
3711 | sidedata_compression_mode, |
|
3720 | sidedata_compression_mode, | |
3712 | ) |
|
3721 | ) | |
3713 |
|
3722 | |||
3714 | # the sidedata computation might have move the file cursors around |
|
3723 | # the sidedata computation might have move the file cursors around | |
3715 | sdfh.seek(current_offset, os.SEEK_SET) |
|
3724 | sdfh.seek(current_offset, os.SEEK_SET) | |
3716 | sdfh.write(serialized_sidedata) |
|
3725 | sdfh.write(serialized_sidedata) | |
3717 | new_entries.append(entry_update) |
|
3726 | new_entries.append(entry_update) | |
3718 | current_offset += len(serialized_sidedata) |
|
3727 | current_offset += len(serialized_sidedata) | |
3719 | self._docket.sidedata_end = sdfh.tell() |
|
3728 | self._docket.sidedata_end = sdfh.tell() | |
3720 |
|
3729 | |||
3721 | # rewrite the new index entries |
|
3730 | # rewrite the new index entries | |
3722 | ifh.seek(startrev * self.index.entry_size) |
|
3731 | ifh.seek(startrev * self.index.entry_size) | |
3723 | for i, e in enumerate(new_entries): |
|
3732 | for i, e in enumerate(new_entries): | |
3724 | rev = startrev + i |
|
3733 | rev = startrev + i | |
3725 | self.index.replace_sidedata_info(rev, *e) |
|
3734 | self.index.replace_sidedata_info(rev, *e) | |
3726 | packed = self.index.entry_binary(rev) |
|
3735 | packed = self.index.entry_binary(rev) | |
3727 | if rev == 0 and self._docket is None: |
|
3736 | if rev == 0 and self._docket is None: | |
3728 | header = self._format_flags | self._format_version |
|
3737 | header = self._format_flags | self._format_version | |
3729 | header = self.index.pack_header(header) |
|
3738 | header = self.index.pack_header(header) | |
3730 | packed = header + packed |
|
3739 | packed = header + packed | |
3731 | ifh.write(packed) |
|
3740 | ifh.write(packed) |
@@ -1,525 +1,528 b'' | |||||
1 | # test revlog interaction about raw data (flagprocessor) |
|
1 | # test revlog interaction about raw data (flagprocessor) | |
2 |
|
2 | |||
3 |
|
3 | |||
4 | import hashlib |
|
4 | import hashlib | |
5 | import sys |
|
5 | import sys | |
6 |
|
6 | |||
7 | from mercurial import ( |
|
7 | from mercurial import ( | |
8 | encoding, |
|
8 | encoding, | |
9 | revlog, |
|
9 | revlog, | |
10 | transaction, |
|
10 | transaction, | |
11 | vfs, |
|
11 | vfs, | |
12 | ) |
|
12 | ) | |
13 |
|
13 | |||
14 | from mercurial.revlogutils import ( |
|
14 | from mercurial.revlogutils import ( | |
15 | constants, |
|
15 | constants, | |
16 | deltas, |
|
16 | deltas, | |
17 | flagutil, |
|
17 | flagutil, | |
18 | ) |
|
18 | ) | |
19 |
|
19 | |||
20 |
|
20 | |||
21 | class _NoTransaction: |
|
21 | class _NoTransaction: | |
22 | """transaction like object to update the nodemap outside a transaction""" |
|
22 | """transaction like object to update the nodemap outside a transaction""" | |
23 |
|
23 | |||
24 | def __init__(self): |
|
24 | def __init__(self): | |
25 | self._postclose = {} |
|
25 | self._postclose = {} | |
26 |
|
26 | |||
27 | def addpostclose(self, callback_id, callback_func): |
|
27 | def addpostclose(self, callback_id, callback_func): | |
28 | self._postclose[callback_id] = callback_func |
|
28 | self._postclose[callback_id] = callback_func | |
29 |
|
29 | |||
30 | def registertmp(self, *args, **kwargs): |
|
30 | def registertmp(self, *args, **kwargs): | |
31 | pass |
|
31 | pass | |
32 |
|
32 | |||
33 | def addbackup(self, *args, **kwargs): |
|
33 | def addbackup(self, *args, **kwargs): | |
34 | pass |
|
34 | pass | |
35 |
|
35 | |||
36 | def add(self, *args, **kwargs): |
|
36 | def add(self, *args, **kwargs): | |
37 | pass |
|
37 | pass | |
38 |
|
38 | |||
39 | def addabort(self, *args, **kwargs): |
|
39 | def addabort(self, *args, **kwargs): | |
40 | pass |
|
40 | pass | |
41 |
|
41 | |||
42 | def _report(self, *args): |
|
42 | def _report(self, *args): | |
43 | pass |
|
43 | pass | |
44 |
|
44 | |||
45 |
|
45 | |||
46 | # TESTTMP is optional. This makes it convenient to run without run-tests.py |
|
46 | # TESTTMP is optional. This makes it convenient to run without run-tests.py | |
47 | tvfs = vfs.vfs(encoding.environ.get(b'TESTTMP', b'/tmp')) |
|
47 | tvfs = vfs.vfs(encoding.environ.get(b'TESTTMP', b'/tmp')) | |
48 |
|
48 | |||
49 | # Enable generaldelta otherwise revlog won't use delta as expected by the test |
|
49 | # Enable generaldelta otherwise revlog won't use delta as expected by the test | |
50 | tvfs.options = { |
|
50 | tvfs.options = { | |
51 | b'generaldelta': True, |
|
51 | b'generaldelta': True, | |
52 | b'revlogv1': True, |
|
52 | b'revlogv1': True, | |
53 | b'sparse-revlog': True, |
|
53 | b'sparse-revlog': True, | |
54 | } |
|
54 | } | |
55 |
|
55 | |||
56 |
|
56 | |||
57 | def abort(msg): |
|
57 | def abort(msg): | |
58 | print('abort: %s' % msg) |
|
58 | print('abort: %s' % msg) | |
59 | # Return 0 so run-tests.py could compare the output. |
|
59 | # Return 0 so run-tests.py could compare the output. | |
60 | sys.exit() |
|
60 | sys.exit() | |
61 |
|
61 | |||
62 |
|
62 | |||
63 | # Register a revlog processor for flag EXTSTORED. |
|
63 | # Register a revlog processor for flag EXTSTORED. | |
64 | # |
|
64 | # | |
65 | # It simply prepends a fixed header, and replaces '1' to 'i'. So it has |
|
65 | # It simply prepends a fixed header, and replaces '1' to 'i'. So it has | |
66 | # insertion and replacement, and may be interesting to test revlog's line-based |
|
66 | # insertion and replacement, and may be interesting to test revlog's line-based | |
67 | # deltas. |
|
67 | # deltas. | |
68 | _extheader = b'E\n' |
|
68 | _extheader = b'E\n' | |
69 |
|
69 | |||
70 |
|
70 | |||
71 | def readprocessor(self, rawtext): |
|
71 | def readprocessor(self, rawtext): | |
72 | # True: the returned text could be used to verify hash |
|
72 | # True: the returned text could be used to verify hash | |
73 | text = rawtext[len(_extheader) :].replace(b'i', b'1') |
|
73 | text = rawtext[len(_extheader) :].replace(b'i', b'1') | |
74 | return text, True |
|
74 | return text, True | |
75 |
|
75 | |||
76 |
|
76 | |||
77 | def writeprocessor(self, text): |
|
77 | def writeprocessor(self, text): | |
78 | # False: the returned rawtext shouldn't be used to verify hash |
|
78 | # False: the returned rawtext shouldn't be used to verify hash | |
79 | rawtext = _extheader + text.replace(b'1', b'i') |
|
79 | rawtext = _extheader + text.replace(b'1', b'i') | |
80 | return rawtext, False |
|
80 | return rawtext, False | |
81 |
|
81 | |||
82 |
|
82 | |||
83 | def rawprocessor(self, rawtext): |
|
83 | def rawprocessor(self, rawtext): | |
84 | # False: do not verify hash. Only the content returned by "readprocessor" |
|
84 | # False: do not verify hash. Only the content returned by "readprocessor" | |
85 | # can be used to verify hash. |
|
85 | # can be used to verify hash. | |
86 | return False |
|
86 | return False | |
87 |
|
87 | |||
88 |
|
88 | |||
89 | flagutil.addflagprocessor( |
|
89 | flagutil.addflagprocessor( | |
90 | revlog.REVIDX_EXTSTORED, (readprocessor, writeprocessor, rawprocessor) |
|
90 | revlog.REVIDX_EXTSTORED, (readprocessor, writeprocessor, rawprocessor) | |
91 | ) |
|
91 | ) | |
92 |
|
92 | |||
93 | # Utilities about reading and appending revlog |
|
93 | # Utilities about reading and appending revlog | |
94 |
|
94 | |||
95 |
|
95 | |||
96 | def newtransaction(): |
|
96 | def newtransaction(): | |
97 | # A transaction is required to write revlogs |
|
97 | # A transaction is required to write revlogs | |
98 | report = lambda msg: None |
|
98 | report = lambda msg: None | |
99 | return transaction.transaction(report, tvfs, {'plain': tvfs}, b'journal') |
|
99 | return transaction.transaction(report, tvfs, {'plain': tvfs}, b'journal') | |
100 |
|
100 | |||
101 |
|
101 | |||
102 | def newrevlog(name=b'_testrevlog', recreate=False): |
|
102 | def newrevlog(name=b'_testrevlog', recreate=False): | |
103 | if recreate: |
|
103 | if recreate: | |
104 | tvfs.tryunlink(name + b'.i') |
|
104 | tvfs.tryunlink(name + b'.i') | |
105 | target = (constants.KIND_OTHER, b'test') |
|
105 | target = (constants.KIND_OTHER, b'test') | |
106 | rlog = revlog.revlog(tvfs, target=target, radix=name) |
|
106 | rlog = revlog.revlog(tvfs, target=target, radix=name) | |
107 | return rlog |
|
107 | return rlog | |
108 |
|
108 | |||
109 |
|
109 | |||
110 | def appendrev(rlog, text, tr, isext=False, isdelta=True): |
|
110 | def appendrev(rlog, text, tr, isext=False, isdelta=True): | |
111 | """Append a revision. If isext is True, set the EXTSTORED flag so flag |
|
111 | """Append a revision. If isext is True, set the EXTSTORED flag so flag | |
112 | processor will be used (and rawtext is different from text). If isdelta is |
|
112 | processor will be used (and rawtext is different from text). If isdelta is | |
113 | True, force the revision to be a delta, otherwise it's full text. |
|
113 | True, force the revision to be a delta, otherwise it's full text. | |
114 | """ |
|
114 | """ | |
115 | nextrev = len(rlog) |
|
115 | nextrev = len(rlog) | |
116 | p1 = rlog.node(nextrev - 1) |
|
116 | p1 = rlog.node(nextrev - 1) | |
117 | p2 = rlog.nullid |
|
117 | p2 = rlog.nullid | |
118 | if isext: |
|
118 | if isext: | |
119 | flags = revlog.REVIDX_EXTSTORED |
|
119 | flags = revlog.REVIDX_EXTSTORED | |
120 | else: |
|
120 | else: | |
121 | flags = revlog.REVIDX_DEFAULT_FLAGS |
|
121 | flags = revlog.REVIDX_DEFAULT_FLAGS | |
122 | # Change storedeltachains temporarily, to override revlog's delta decision |
|
122 | # Change storedeltachains temporarily, to override revlog's delta decision | |
123 | rlog._storedeltachains = isdelta |
|
123 | rlog._storedeltachains = isdelta | |
124 | try: |
|
124 | try: | |
125 | rlog.addrevision(text, tr, nextrev, p1, p2, flags=flags) |
|
125 | rlog.addrevision(text, tr, nextrev, p1, p2, flags=flags) | |
126 | return nextrev |
|
126 | return nextrev | |
127 | except Exception as ex: |
|
127 | except Exception as ex: | |
128 | abort('rev %d: failed to append: %s' % (nextrev, ex)) |
|
128 | abort('rev %d: failed to append: %s' % (nextrev, ex)) | |
129 | finally: |
|
129 | finally: | |
130 | # Restore storedeltachains. It is always True, see revlog.__init__ |
|
130 | # Restore storedeltachains. It is always True, see revlog.__init__ | |
131 | rlog._storedeltachains = True |
|
131 | rlog._storedeltachains = True | |
132 |
|
132 | |||
133 |
|
133 | |||
134 | def addgroupcopy(rlog, tr, destname=b'_destrevlog', optimaldelta=True): |
|
134 | def addgroupcopy(rlog, tr, destname=b'_destrevlog', optimaldelta=True): | |
135 | """Copy revlog to destname using revlog.addgroup. Return the copied revlog. |
|
135 | """Copy revlog to destname using revlog.addgroup. Return the copied revlog. | |
136 |
|
136 | |||
137 | This emulates push or pull. They use changegroup. Changegroup requires |
|
137 | This emulates push or pull. They use changegroup. Changegroup requires | |
138 | repo to work. We don't have a repo, so a dummy changegroup is used. |
|
138 | repo to work. We don't have a repo, so a dummy changegroup is used. | |
139 |
|
139 | |||
140 | If optimaldelta is True, use optimized delta parent, so the destination |
|
140 | If optimaldelta is True, use optimized delta parent, so the destination | |
141 | revlog could probably reuse it. Otherwise it builds sub-optimal delta, and |
|
141 | revlog could probably reuse it. Otherwise it builds sub-optimal delta, and | |
142 | the destination revlog needs more work to use it. |
|
142 | the destination revlog needs more work to use it. | |
143 |
|
143 | |||
144 | This exercises some revlog.addgroup (and revlog._addrevision(text=None)) |
|
144 | This exercises some revlog.addgroup (and revlog._addrevision(text=None)) | |
145 | code path, which is not covered by "appendrev" alone. |
|
145 | code path, which is not covered by "appendrev" alone. | |
146 | """ |
|
146 | """ | |
147 |
|
147 | |||
148 | class dummychangegroup: |
|
148 | class dummychangegroup: | |
149 | @staticmethod |
|
149 | @staticmethod | |
150 | def deltachunk(pnode): |
|
150 | def deltachunk(pnode): | |
151 | pnode = pnode or rlog.nullid |
|
151 | pnode = pnode or rlog.nullid | |
152 | parentrev = rlog.rev(pnode) |
|
152 | parentrev = rlog.rev(pnode) | |
153 | r = parentrev + 1 |
|
153 | r = parentrev + 1 | |
154 | if r >= len(rlog): |
|
154 | if r >= len(rlog): | |
155 | return {} |
|
155 | return {} | |
156 | if optimaldelta: |
|
156 | if optimaldelta: | |
157 | deltaparent = parentrev |
|
157 | deltaparent = parentrev | |
158 | else: |
|
158 | else: | |
159 | # suboptimal deltaparent |
|
159 | # suboptimal deltaparent | |
160 | deltaparent = min(0, parentrev) |
|
160 | deltaparent = min(0, parentrev) | |
161 | if not rlog._candelta(deltaparent, r): |
|
161 | if not rlog._candelta(deltaparent, r): | |
162 | deltaparent = -1 |
|
162 | deltaparent = -1 | |
163 | return { |
|
163 | return { | |
164 | b'node': rlog.node(r), |
|
164 | b'node': rlog.node(r), | |
165 | b'p1': pnode, |
|
165 | b'p1': pnode, | |
166 | b'p2': rlog.nullid, |
|
166 | b'p2': rlog.nullid, | |
167 | b'cs': rlog.node(rlog.linkrev(r)), |
|
167 | b'cs': rlog.node(rlog.linkrev(r)), | |
168 | b'flags': rlog.flags(r), |
|
168 | b'flags': rlog.flags(r), | |
169 | b'deltabase': rlog.node(deltaparent), |
|
169 | b'deltabase': rlog.node(deltaparent), | |
170 | b'delta': rlog.revdiff(deltaparent, r), |
|
170 | b'delta': rlog.revdiff(deltaparent, r), | |
171 | b'sidedata': rlog.sidedata(r), |
|
171 | b'sidedata': rlog.sidedata(r), | |
172 | } |
|
172 | } | |
173 |
|
173 | |||
174 | def deltaiter(self): |
|
174 | def deltaiter(self): | |
175 | chain = None |
|
175 | chain = None | |
176 | for chunkdata in iter(lambda: self.deltachunk(chain), {}): |
|
176 | for chunkdata in iter(lambda: self.deltachunk(chain), {}): | |
177 | node = chunkdata[b'node'] |
|
177 | node = chunkdata[b'node'] | |
178 | p1 = chunkdata[b'p1'] |
|
178 | p1 = chunkdata[b'p1'] | |
179 | p2 = chunkdata[b'p2'] |
|
179 | p2 = chunkdata[b'p2'] | |
180 | cs = chunkdata[b'cs'] |
|
180 | cs = chunkdata[b'cs'] | |
181 | deltabase = chunkdata[b'deltabase'] |
|
181 | deltabase = chunkdata[b'deltabase'] | |
182 | delta = chunkdata[b'delta'] |
|
182 | delta = chunkdata[b'delta'] | |
183 | flags = chunkdata[b'flags'] |
|
183 | flags = chunkdata[b'flags'] | |
184 | sidedata = chunkdata[b'sidedata'] |
|
184 | sidedata = chunkdata[b'sidedata'] | |
185 |
|
185 | |||
186 | chain = node |
|
186 | chain = node | |
187 |
|
187 | |||
188 | yield (node, p1, p2, cs, deltabase, delta, flags, sidedata) |
|
188 | yield (node, p1, p2, cs, deltabase, delta, flags, sidedata) | |
189 |
|
189 | |||
190 | def linkmap(lnode): |
|
190 | def linkmap(lnode): | |
191 | return rlog.rev(lnode) |
|
191 | return rlog.rev(lnode) | |
192 |
|
192 | |||
193 | dlog = newrevlog(destname, recreate=True) |
|
193 | dlog = newrevlog(destname, recreate=True) | |
194 | dummydeltas = dummychangegroup().deltaiter() |
|
194 | dummydeltas = dummychangegroup().deltaiter() | |
195 | dlog.addgroup(dummydeltas, linkmap, tr) |
|
195 | dlog.addgroup(dummydeltas, linkmap, tr) | |
196 | return dlog |
|
196 | return dlog | |
197 |
|
197 | |||
198 |
|
198 | |||
199 | def lowlevelcopy(rlog, tr, destname=b'_destrevlog'): |
|
199 | def lowlevelcopy(rlog, tr, destname=b'_destrevlog'): | |
200 | """Like addgroupcopy, but use the low level revlog._addrevision directly. |
|
200 | """Like addgroupcopy, but use the low level revlog._addrevision directly. | |
201 |
|
201 | |||
202 | It exercises some code paths that are hard to reach easily otherwise. |
|
202 | It exercises some code paths that are hard to reach easily otherwise. | |
203 | """ |
|
203 | """ | |
204 | dlog = newrevlog(destname, recreate=True) |
|
204 | dlog = newrevlog(destname, recreate=True) | |
205 | for r in rlog: |
|
205 | for r in rlog: | |
206 | p1 = rlog.node(r - 1) |
|
206 | p1 = rlog.node(r - 1) | |
207 | p2 = rlog.nullid |
|
207 | p2 = rlog.nullid | |
208 | if r == 0 or (rlog.flags(r) & revlog.REVIDX_EXTSTORED): |
|
208 | if r == 0 or (rlog.flags(r) & revlog.REVIDX_EXTSTORED): | |
209 | text = rlog.rawdata(r) |
|
209 | text = rlog.rawdata(r) | |
210 | cachedelta = None |
|
210 | cachedelta = None | |
211 | else: |
|
211 | else: | |
212 | # deltaparent cannot have EXTSTORED flag. |
|
212 | # deltaparent cannot have EXTSTORED flag. | |
213 | deltaparent = max( |
|
213 | deltaparent = max( | |
214 | [-1] |
|
214 | [-1] | |
215 | + [ |
|
215 | + [ | |
216 | p |
|
216 | p | |
217 | for p in range(r) |
|
217 | for p in range(r) | |
218 | if rlog.flags(p) & revlog.REVIDX_EXTSTORED == 0 |
|
218 | if rlog.flags(p) & revlog.REVIDX_EXTSTORED == 0 | |
219 | ] |
|
219 | ] | |
220 | ) |
|
220 | ) | |
221 | text = None |
|
221 | text = None | |
222 | cachedelta = (deltaparent, rlog.revdiff(deltaparent, r)) |
|
222 | cachedelta = (deltaparent, rlog.revdiff(deltaparent, r)) | |
223 | flags = rlog.flags(r) |
|
223 | flags = rlog.flags(r) | |
224 | with dlog._writing(_NoTransaction()): |
|
224 | with dlog._writing(_NoTransaction()): | |
225 | dlog._addrevision( |
|
225 | dlog._addrevision( | |
226 | rlog.node(r), |
|
226 | rlog.node(r), | |
227 | text, |
|
227 | text, | |
228 | tr, |
|
228 | tr, | |
229 | r, |
|
229 | r, | |
230 | p1, |
|
230 | p1, | |
231 | p2, |
|
231 | p2, | |
232 | flags, |
|
232 | flags, | |
233 | cachedelta, |
|
233 | cachedelta, | |
234 | ) |
|
234 | ) | |
235 | return dlog |
|
235 | return dlog | |
236 |
|
236 | |||
237 |
|
237 | |||
238 | # Utilities to generate revisions for testing |
|
238 | # Utilities to generate revisions for testing | |
239 |
|
239 | |||
240 |
|
240 | |||
241 | def genbits(n): |
|
241 | def genbits(n): | |
242 | """Given a number n, generate (2 ** (n * 2) + 1) numbers in range(2 ** n). |
|
242 | """Given a number n, generate (2 ** (n * 2) + 1) numbers in range(2 ** n). | |
243 | i.e. the generated numbers have a width of n bits. |
|
243 | i.e. the generated numbers have a width of n bits. | |
244 |
|
244 | |||
245 | The combination of two adjacent numbers will cover all possible cases. |
|
245 | The combination of two adjacent numbers will cover all possible cases. | |
246 | That is to say, given any x, y where both x, and y are in range(2 ** n), |
|
246 | That is to say, given any x, y where both x, and y are in range(2 ** n), | |
247 | there is an x followed immediately by y in the generated sequence. |
|
247 | there is an x followed immediately by y in the generated sequence. | |
248 | """ |
|
248 | """ | |
249 | m = 2 ** n |
|
249 | m = 2 ** n | |
250 |
|
250 | |||
251 | # Gray Code. See https://en.wikipedia.org/wiki/Gray_code |
|
251 | # Gray Code. See https://en.wikipedia.org/wiki/Gray_code | |
252 | gray = lambda x: x ^ (x >> 1) |
|
252 | gray = lambda x: x ^ (x >> 1) | |
253 | reversegray = {gray(i): i for i in range(m)} |
|
253 | reversegray = {gray(i): i for i in range(m)} | |
254 |
|
254 | |||
255 | # Generate (n * 2) bit gray code, yield lower n bits as X, and look for |
|
255 | # Generate (n * 2) bit gray code, yield lower n bits as X, and look for | |
256 | # the next unused gray code where higher n bits equal to X. |
|
256 | # the next unused gray code where higher n bits equal to X. | |
257 |
|
257 | |||
258 | # For gray codes whose higher bits are X, a[X] of them have been used. |
|
258 | # For gray codes whose higher bits are X, a[X] of them have been used. | |
259 | a = [0] * m |
|
259 | a = [0] * m | |
260 |
|
260 | |||
261 | # Iterate from 0. |
|
261 | # Iterate from 0. | |
262 | x = 0 |
|
262 | x = 0 | |
263 | yield x |
|
263 | yield x | |
264 | for i in range(m * m): |
|
264 | for i in range(m * m): | |
265 | x = reversegray[x] |
|
265 | x = reversegray[x] | |
266 | y = gray(a[x] + x * m) & (m - 1) |
|
266 | y = gray(a[x] + x * m) & (m - 1) | |
267 | assert a[x] < m |
|
267 | assert a[x] < m | |
268 | a[x] += 1 |
|
268 | a[x] += 1 | |
269 | x = y |
|
269 | x = y | |
270 | yield x |
|
270 | yield x | |
271 |
|
271 | |||
272 |
|
272 | |||
273 | def gentext(rev): |
|
273 | def gentext(rev): | |
274 | '''Given a revision number, generate dummy text''' |
|
274 | '''Given a revision number, generate dummy text''' | |
275 | return b''.join(b'%d\n' % j for j in range(-1, rev % 5)) |
|
275 | return b''.join(b'%d\n' % j for j in range(-1, rev % 5)) | |
276 |
|
276 | |||
277 |
|
277 | |||
278 | def writecases(rlog, tr): |
|
278 | def writecases(rlog, tr): | |
279 | """Write some revisions interested to the test. |
|
279 | """Write some revisions interested to the test. | |
280 |
|
280 | |||
281 | The test is interested in 3 properties of a revision: |
|
281 | The test is interested in 3 properties of a revision: | |
282 |
|
282 | |||
283 | - Is it a delta or a full text? (isdelta) |
|
283 | - Is it a delta or a full text? (isdelta) | |
284 | This is to catch some delta application issues. |
|
284 | This is to catch some delta application issues. | |
285 | - Does it have a flag of EXTSTORED? (isext) |
|
285 | - Does it have a flag of EXTSTORED? (isext) | |
286 | This is to catch some flag processor issues. Especially when |
|
286 | This is to catch some flag processor issues. Especially when | |
287 | interacted with revlog deltas. |
|
287 | interacted with revlog deltas. | |
288 | - Is its text empty? (isempty) |
|
288 | - Is its text empty? (isempty) | |
289 | This is less important. It is intended to try to catch some careless |
|
289 | This is less important. It is intended to try to catch some careless | |
290 | checks like "if text" instead of "if text is None". Note: if flag |
|
290 | checks like "if text" instead of "if text is None". Note: if flag | |
291 | processor is involved, raw text may be not empty. |
|
291 | processor is involved, raw text may be not empty. | |
292 |
|
292 | |||
293 | Write 65 revisions. So that all combinations of the above flags for |
|
293 | Write 65 revisions. So that all combinations of the above flags for | |
294 | adjacent revisions are covered. That is to say, |
|
294 | adjacent revisions are covered. That is to say, | |
295 |
|
295 | |||
296 | len(set( |
|
296 | len(set( | |
297 | (r.delta, r.ext, r.empty, (r+1).delta, (r+1).ext, (r+1).empty) |
|
297 | (r.delta, r.ext, r.empty, (r+1).delta, (r+1).ext, (r+1).empty) | |
298 | for r in range(len(rlog) - 1) |
|
298 | for r in range(len(rlog) - 1) | |
299 | )) is 64. |
|
299 | )) is 64. | |
300 |
|
300 | |||
301 | Where "r.delta", "r.ext", and "r.empty" are booleans matching properties |
|
301 | Where "r.delta", "r.ext", and "r.empty" are booleans matching properties | |
302 | mentioned above. |
|
302 | mentioned above. | |
303 |
|
303 | |||
304 | Return expected [(text, rawtext)]. |
|
304 | Return expected [(text, rawtext)]. | |
305 | """ |
|
305 | """ | |
306 | result = [] |
|
306 | result = [] | |
307 | for i, x in enumerate(genbits(3)): |
|
307 | for i, x in enumerate(genbits(3)): | |
308 | isdelta, isext, isempty = bool(x & 1), bool(x & 2), bool(x & 4) |
|
308 | isdelta, isext, isempty = bool(x & 1), bool(x & 2), bool(x & 4) | |
309 | if isempty: |
|
309 | if isempty: | |
310 | text = b'' |
|
310 | text = b'' | |
311 | else: |
|
311 | else: | |
312 | text = gentext(i) |
|
312 | text = gentext(i) | |
313 | rev = appendrev(rlog, text, tr, isext=isext, isdelta=isdelta) |
|
313 | rev = appendrev(rlog, text, tr, isext=isext, isdelta=isdelta) | |
314 |
|
314 | |||
315 | # Verify text, rawtext, and rawsize |
|
315 | # Verify text, rawtext, and rawsize | |
316 | if isext: |
|
316 | if isext: | |
317 | rawtext = writeprocessor(None, text)[0] |
|
317 | rawtext = writeprocessor(None, text)[0] | |
318 | else: |
|
318 | else: | |
319 | rawtext = text |
|
319 | rawtext = text | |
320 | if rlog.rawsize(rev) != len(rawtext): |
|
320 | if rlog.rawsize(rev) != len(rawtext): | |
321 | abort('rev %d: wrong rawsize' % rev) |
|
321 | abort('rev %d: wrong rawsize' % rev) | |
322 | if rlog.revision(rev) != text: |
|
322 | if rlog.revision(rev) != text: | |
323 | abort('rev %d: wrong text' % rev) |
|
323 | abort('rev %d: wrong text' % rev) | |
324 | if rlog.rawdata(rev) != rawtext: |
|
324 | if rlog.rawdata(rev) != rawtext: | |
325 | abort('rev %d: wrong rawtext' % rev) |
|
325 | abort('rev %d: wrong rawtext' % rev) | |
326 | result.append((text, rawtext)) |
|
326 | result.append((text, rawtext)) | |
327 |
|
327 | |||
328 | # Verify flags like isdelta, isext work as expected |
|
328 | # Verify flags like isdelta, isext work as expected | |
329 | # isdelta can be overridden to False if this or p1 has isext set |
|
329 | # isdelta can be overridden to False if this or p1 has isext set | |
330 | if bool(rlog.deltaparent(rev) > -1) and not isdelta: |
|
330 | if bool(rlog.deltaparent(rev) > -1) and not isdelta: | |
331 | abort('rev %d: isdelta is unexpected' % rev) |
|
331 | abort('rev %d: isdelta is unexpected' % rev) | |
332 | if bool(rlog.flags(rev)) != isext: |
|
332 | if bool(rlog.flags(rev)) != isext: | |
333 | abort('rev %d: isext is ineffective' % rev) |
|
333 | abort('rev %d: isext is ineffective' % rev) | |
334 | return result |
|
334 | return result | |
335 |
|
335 | |||
336 |
|
336 | |||
337 | # Main test and checking |
|
337 | # Main test and checking | |
338 |
|
338 | |||
339 |
|
339 | |||
340 | def checkrevlog(rlog, expected): |
|
340 | def checkrevlog(rlog, expected): | |
341 | '''Check if revlog has expected contents. expected is [(text, rawtext)]''' |
|
341 | '''Check if revlog has expected contents. expected is [(text, rawtext)]''' | |
342 | # Test using different access orders. This could expose some issues |
|
342 | # Test using different access orders. This could expose some issues | |
343 | # depending on revlog caching (see revlog._cache). |
|
343 | # depending on revlog caching (see revlog._cache). | |
344 | for r0 in range(len(rlog) - 1): |
|
344 | for r0 in range(len(rlog) - 1): | |
345 | r1 = r0 + 1 |
|
345 | r1 = r0 + 1 | |
346 | for revorder in [[r0, r1], [r1, r0]]: |
|
346 | for revorder in [[r0, r1], [r1, r0]]: | |
347 | for raworder in [[True], [False], [True, False], [False, True]]: |
|
347 | for raworder in [[True], [False], [True, False], [False, True]]: | |
348 | nlog = newrevlog() |
|
348 | nlog = newrevlog() | |
349 | for rev in revorder: |
|
349 | for rev in revorder: | |
350 | for raw in raworder: |
|
350 | for raw in raworder: | |
351 | if raw: |
|
351 | if raw: | |
352 | t = nlog.rawdata(rev) |
|
352 | t = nlog.rawdata(rev) | |
353 | else: |
|
353 | else: | |
354 | t = nlog.revision(rev) |
|
354 | t = nlog.revision(rev) | |
355 | if t != expected[rev][int(raw)]: |
|
355 | if t != expected[rev][int(raw)]: | |
356 | abort( |
|
356 | abort( | |
357 | 'rev %d: corrupted %stext' |
|
357 | 'rev %d: corrupted %stext' | |
358 | % (rev, raw and 'raw' or '') |
|
358 | % (rev, raw and 'raw' or '') | |
359 | ) |
|
359 | ) | |
360 |
|
360 | |||
361 |
|
361 | |||
362 | slicingdata = [ |
|
362 | slicingdata = [ | |
363 | ([0, 1, 2, 3, 55, 56, 58, 59, 60], [[0, 1], [2], [58], [59, 60]], 10), |
|
363 | ([0, 1, 2, 3, 55, 56, 58, 59, 60], [[0, 1], [2], [58], [59, 60]], 10), | |
364 | ([0, 1, 2, 3, 55, 56, 58, 59, 60], [[0, 1], [2], [58], [59, 60]], 10), |
|
364 | ([0, 1, 2, 3, 55, 56, 58, 59, 60], [[0, 1], [2], [58], [59, 60]], 10), | |
365 | ( |
|
365 | ( | |
366 | [-1, 0, 1, 2, 3, 55, 56, 58, 59, 60], |
|
366 | [-1, 0, 1, 2, 3, 55, 56, 58, 59, 60], | |
367 | [[-1, 0, 1], [2], [58], [59, 60]], |
|
367 | [[-1, 0, 1], [2], [58], [59, 60]], | |
368 | 10, |
|
368 | 10, | |
369 | ), |
|
369 | ), | |
370 | ] |
|
370 | ] | |
371 |
|
371 | |||
372 |
|
372 | |||
373 | def slicingtest(rlog): |
|
373 | def slicingtest(rlog): | |
374 | oldmin = rlog._srmingapsize |
|
374 | old_delta_config = rlog.delta_config | |
|
375 | old_data_config = rlog.data_config | |||
|
376 | rlog.delta_config = rlog.delta_config.copy() | |||
|
377 | rlog.data_config = rlog.data_config.copy() | |||
375 | try: |
|
378 | try: | |
376 | # the test revlog is small, we remove the floor under which we |
|
379 | # the test revlog is small, we remove the floor under which we | |
377 | # slicing is diregarded. |
|
380 | # slicing is diregarded. | |
378 | rlog.data_config.sr_min_gap_size = 0 |
|
381 | rlog.data_config.sr_min_gap_size = 0 | |
379 | rlog.delta_config.sr_min_gap_size = 0 |
|
382 | rlog.delta_config.sr_min_gap_size = 0 | |
380 | for item in slicingdata: |
|
383 | for item in slicingdata: | |
381 | chain, expected, target = item |
|
384 | chain, expected, target = item | |
382 | result = deltas.slicechunk(rlog, chain, targetsize=target) |
|
385 | result = deltas.slicechunk(rlog, chain, targetsize=target) | |
383 | result = list(result) |
|
386 | result = list(result) | |
384 | if result != expected: |
|
387 | if result != expected: | |
385 | print('slicing differ:') |
|
388 | print('slicing differ:') | |
386 | print(' chain: %s' % chain) |
|
389 | print(' chain: %s' % chain) | |
387 | print(' target: %s' % target) |
|
390 | print(' target: %s' % target) | |
388 | print(' expected: %s' % expected) |
|
391 | print(' expected: %s' % expected) | |
389 | print(' result: %s' % result) |
|
392 | print(' result: %s' % result) | |
390 | finally: |
|
393 | finally: | |
391 |
rlog.d |
|
394 | rlog.delta_config = old_delta_config | |
392 |
rlog.d |
|
395 | rlog.data_config = old_data_config | |
393 |
|
396 | |||
394 |
|
397 | |||
395 | def md5sum(s): |
|
398 | def md5sum(s): | |
396 | return hashlib.md5(s).digest() |
|
399 | return hashlib.md5(s).digest() | |
397 |
|
400 | |||
398 |
|
401 | |||
399 | def _maketext(*coord): |
|
402 | def _maketext(*coord): | |
400 | """create piece of text according to range of integers |
|
403 | """create piece of text according to range of integers | |
401 |
|
404 | |||
402 | The test returned use a md5sum of the integer to make it less |
|
405 | The test returned use a md5sum of the integer to make it less | |
403 | compressible""" |
|
406 | compressible""" | |
404 | pieces = [] |
|
407 | pieces = [] | |
405 | for start, size in coord: |
|
408 | for start, size in coord: | |
406 | num = range(start, start + size) |
|
409 | num = range(start, start + size) | |
407 | p = [md5sum(b'%d' % r) for r in num] |
|
410 | p = [md5sum(b'%d' % r) for r in num] | |
408 | pieces.append(b'\n'.join(p)) |
|
411 | pieces.append(b'\n'.join(p)) | |
409 | return b'\n'.join(pieces) + b'\n' |
|
412 | return b'\n'.join(pieces) + b'\n' | |
410 |
|
413 | |||
411 |
|
414 | |||
412 | data = [ |
|
415 | data = [ | |
413 | _maketext((0, 120), (456, 60)), |
|
416 | _maketext((0, 120), (456, 60)), | |
414 | _maketext((0, 120), (345, 60)), |
|
417 | _maketext((0, 120), (345, 60)), | |
415 | _maketext((0, 120), (734, 60)), |
|
418 | _maketext((0, 120), (734, 60)), | |
416 | _maketext((0, 120), (734, 60), (923, 45)), |
|
419 | _maketext((0, 120), (734, 60), (923, 45)), | |
417 | _maketext((0, 120), (734, 60), (234, 45)), |
|
420 | _maketext((0, 120), (734, 60), (234, 45)), | |
418 | _maketext((0, 120), (734, 60), (564, 45)), |
|
421 | _maketext((0, 120), (734, 60), (564, 45)), | |
419 | _maketext((0, 120), (734, 60), (361, 45)), |
|
422 | _maketext((0, 120), (734, 60), (361, 45)), | |
420 | _maketext((0, 120), (734, 60), (489, 45)), |
|
423 | _maketext((0, 120), (734, 60), (489, 45)), | |
421 | _maketext((0, 120), (123, 60)), |
|
424 | _maketext((0, 120), (123, 60)), | |
422 | _maketext((0, 120), (145, 60)), |
|
425 | _maketext((0, 120), (145, 60)), | |
423 | _maketext((0, 120), (104, 60)), |
|
426 | _maketext((0, 120), (104, 60)), | |
424 | _maketext((0, 120), (430, 60)), |
|
427 | _maketext((0, 120), (430, 60)), | |
425 | _maketext((0, 120), (430, 60), (923, 45)), |
|
428 | _maketext((0, 120), (430, 60), (923, 45)), | |
426 | _maketext((0, 120), (430, 60), (234, 45)), |
|
429 | _maketext((0, 120), (430, 60), (234, 45)), | |
427 | _maketext((0, 120), (430, 60), (564, 45)), |
|
430 | _maketext((0, 120), (430, 60), (564, 45)), | |
428 | _maketext((0, 120), (430, 60), (361, 45)), |
|
431 | _maketext((0, 120), (430, 60), (361, 45)), | |
429 | _maketext((0, 120), (430, 60), (489, 45)), |
|
432 | _maketext((0, 120), (430, 60), (489, 45)), | |
430 | _maketext((0, 120), (249, 60)), |
|
433 | _maketext((0, 120), (249, 60)), | |
431 | _maketext((0, 120), (832, 60)), |
|
434 | _maketext((0, 120), (832, 60)), | |
432 | _maketext((0, 120), (891, 60)), |
|
435 | _maketext((0, 120), (891, 60)), | |
433 | _maketext((0, 120), (543, 60)), |
|
436 | _maketext((0, 120), (543, 60)), | |
434 | _maketext((0, 120), (120, 60)), |
|
437 | _maketext((0, 120), (120, 60)), | |
435 | _maketext((0, 120), (60, 60), (768, 30)), |
|
438 | _maketext((0, 120), (60, 60), (768, 30)), | |
436 | _maketext((0, 120), (60, 60), (260, 30)), |
|
439 | _maketext((0, 120), (60, 60), (260, 30)), | |
437 | _maketext((0, 120), (60, 60), (450, 30)), |
|
440 | _maketext((0, 120), (60, 60), (450, 30)), | |
438 | _maketext((0, 120), (60, 60), (361, 30)), |
|
441 | _maketext((0, 120), (60, 60), (361, 30)), | |
439 | _maketext((0, 120), (60, 60), (886, 30)), |
|
442 | _maketext((0, 120), (60, 60), (886, 30)), | |
440 | _maketext((0, 120), (60, 60), (116, 30)), |
|
443 | _maketext((0, 120), (60, 60), (116, 30)), | |
441 | _maketext((0, 120), (60, 60), (567, 30), (629, 40)), |
|
444 | _maketext((0, 120), (60, 60), (567, 30), (629, 40)), | |
442 | _maketext((0, 120), (60, 60), (569, 30), (745, 40)), |
|
445 | _maketext((0, 120), (60, 60), (569, 30), (745, 40)), | |
443 | _maketext((0, 120), (60, 60), (777, 30), (700, 40)), |
|
446 | _maketext((0, 120), (60, 60), (777, 30), (700, 40)), | |
444 | _maketext((0, 120), (60, 60), (618, 30), (398, 40), (158, 10)), |
|
447 | _maketext((0, 120), (60, 60), (618, 30), (398, 40), (158, 10)), | |
445 | ] |
|
448 | ] | |
446 |
|
449 | |||
447 |
|
450 | |||
448 | def makesnapshot(tr): |
|
451 | def makesnapshot(tr): | |
449 | rl = newrevlog(name=b'_snaprevlog3', recreate=True) |
|
452 | rl = newrevlog(name=b'_snaprevlog3', recreate=True) | |
450 | for i in data: |
|
453 | for i in data: | |
451 | appendrev(rl, i, tr) |
|
454 | appendrev(rl, i, tr) | |
452 | return rl |
|
455 | return rl | |
453 |
|
456 | |||
454 |
|
457 | |||
455 | snapshots = [-1, 0, 6, 8, 11, 17, 19, 21, 25, 30] |
|
458 | snapshots = [-1, 0, 6, 8, 11, 17, 19, 21, 25, 30] | |
456 |
|
459 | |||
457 |
|
460 | |||
458 | def issnapshottest(rlog): |
|
461 | def issnapshottest(rlog): | |
459 | result = [] |
|
462 | result = [] | |
460 | if rlog.issnapshot(-1): |
|
463 | if rlog.issnapshot(-1): | |
461 | result.append(-1) |
|
464 | result.append(-1) | |
462 | for rev in rlog: |
|
465 | for rev in rlog: | |
463 | if rlog.issnapshot(rev): |
|
466 | if rlog.issnapshot(rev): | |
464 | result.append(rev) |
|
467 | result.append(rev) | |
465 | if snapshots != result: |
|
468 | if snapshots != result: | |
466 | print('snapshot differ:') |
|
469 | print('snapshot differ:') | |
467 | print(' expected: %s' % snapshots) |
|
470 | print(' expected: %s' % snapshots) | |
468 | print(' got: %s' % result) |
|
471 | print(' got: %s' % result) | |
469 |
|
472 | |||
470 |
|
473 | |||
471 | snapshotmapall = {0: {6, 8, 11, 17, 19, 25}, 8: {21}, -1: {0, 30}} |
|
474 | snapshotmapall = {0: {6, 8, 11, 17, 19, 25}, 8: {21}, -1: {0, 30}} | |
472 | snapshotmap15 = {0: {17, 19, 25}, 8: {21}, -1: {30}} |
|
475 | snapshotmap15 = {0: {17, 19, 25}, 8: {21}, -1: {30}} | |
473 |
|
476 | |||
474 |
|
477 | |||
475 | def findsnapshottest(rlog): |
|
478 | def findsnapshottest(rlog): | |
476 | cache = deltas.SnapshotCache() |
|
479 | cache = deltas.SnapshotCache() | |
477 | cache.update(rlog) |
|
480 | cache.update(rlog) | |
478 | resultall = dict(cache.snapshots) |
|
481 | resultall = dict(cache.snapshots) | |
479 | if resultall != snapshotmapall: |
|
482 | if resultall != snapshotmapall: | |
480 | print('snapshot map differ:') |
|
483 | print('snapshot map differ:') | |
481 | print(' expected: %s' % snapshotmapall) |
|
484 | print(' expected: %s' % snapshotmapall) | |
482 | print(' got: %s' % resultall) |
|
485 | print(' got: %s' % resultall) | |
483 | cache15 = deltas.SnapshotCache() |
|
486 | cache15 = deltas.SnapshotCache() | |
484 | cache15.update(rlog, 15) |
|
487 | cache15.update(rlog, 15) | |
485 | result15 = dict(cache15.snapshots) |
|
488 | result15 = dict(cache15.snapshots) | |
486 | if result15 != snapshotmap15: |
|
489 | if result15 != snapshotmap15: | |
487 | print('snapshot map differ:') |
|
490 | print('snapshot map differ:') | |
488 | print(' expected: %s' % snapshotmap15) |
|
491 | print(' expected: %s' % snapshotmap15) | |
489 | print(' got: %s' % result15) |
|
492 | print(' got: %s' % result15) | |
490 |
|
493 | |||
491 |
|
494 | |||
492 | def maintest(): |
|
495 | def maintest(): | |
493 | with newtransaction() as tr: |
|
496 | with newtransaction() as tr: | |
494 | rl = newrevlog(recreate=True) |
|
497 | rl = newrevlog(recreate=True) | |
495 | expected = writecases(rl, tr) |
|
498 | expected = writecases(rl, tr) | |
496 | checkrevlog(rl, expected) |
|
499 | checkrevlog(rl, expected) | |
497 | print('local test passed') |
|
500 | print('local test passed') | |
498 | # Copy via revlog.addgroup |
|
501 | # Copy via revlog.addgroup | |
499 | rl1 = addgroupcopy(rl, tr) |
|
502 | rl1 = addgroupcopy(rl, tr) | |
500 | checkrevlog(rl1, expected) |
|
503 | checkrevlog(rl1, expected) | |
501 | rl2 = addgroupcopy(rl, tr, optimaldelta=False) |
|
504 | rl2 = addgroupcopy(rl, tr, optimaldelta=False) | |
502 | checkrevlog(rl2, expected) |
|
505 | checkrevlog(rl2, expected) | |
503 | print('addgroupcopy test passed') |
|
506 | print('addgroupcopy test passed') | |
504 | # Copy via revlog.clone |
|
507 | # Copy via revlog.clone | |
505 | rl3 = newrevlog(name=b'_destrevlog3', recreate=True) |
|
508 | rl3 = newrevlog(name=b'_destrevlog3', recreate=True) | |
506 | rl.clone(tr, rl3) |
|
509 | rl.clone(tr, rl3) | |
507 | checkrevlog(rl3, expected) |
|
510 | checkrevlog(rl3, expected) | |
508 | print('clone test passed') |
|
511 | print('clone test passed') | |
509 | # Copy via low-level revlog._addrevision |
|
512 | # Copy via low-level revlog._addrevision | |
510 | rl4 = lowlevelcopy(rl, tr) |
|
513 | rl4 = lowlevelcopy(rl, tr) | |
511 | checkrevlog(rl4, expected) |
|
514 | checkrevlog(rl4, expected) | |
512 | print('lowlevelcopy test passed') |
|
515 | print('lowlevelcopy test passed') | |
513 | slicingtest(rl) |
|
516 | slicingtest(rl) | |
514 | print('slicing test passed') |
|
517 | print('slicing test passed') | |
515 | rl5 = makesnapshot(tr) |
|
518 | rl5 = makesnapshot(tr) | |
516 | issnapshottest(rl5) |
|
519 | issnapshottest(rl5) | |
517 | print('issnapshot test passed') |
|
520 | print('issnapshot test passed') | |
518 | findsnapshottest(rl5) |
|
521 | findsnapshottest(rl5) | |
519 | print('findsnapshot test passed') |
|
522 | print('findsnapshot test passed') | |
520 |
|
523 | |||
521 |
|
524 | |||
522 | try: |
|
525 | try: | |
523 | maintest() |
|
526 | maintest() | |
524 | except Exception as ex: |
|
527 | except Exception as ex: | |
525 | abort('crashed: %s' % ex) |
|
528 | abort('crashed: %s' % ex) |
General Comments 0
You need to be logged in to leave comments.
Login now