Show More
@@ -1,4056 +1,4059 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 functools | |||
19 | import io |
|
20 | import io | |
20 | import os |
|
21 | import os | |
21 | import struct |
|
22 | import struct | |
22 | import weakref |
|
23 | import weakref | |
23 | import zlib |
|
24 | import zlib | |
24 |
|
25 | |||
25 | # import stuff from node for others to import from revlog |
|
26 | # import stuff from node for others to import from revlog | |
26 | from .node import ( |
|
27 | from .node import ( | |
27 | bin, |
|
28 | bin, | |
28 | hex, |
|
29 | hex, | |
29 | nullrev, |
|
30 | nullrev, | |
30 | sha1nodeconstants, |
|
31 | sha1nodeconstants, | |
31 | short, |
|
32 | short, | |
32 | wdirrev, |
|
33 | wdirrev, | |
33 | ) |
|
34 | ) | |
34 | from .i18n import _ |
|
35 | from .i18n import _ | |
35 | from .revlogutils.constants import ( |
|
36 | from .revlogutils.constants import ( | |
36 | ALL_KINDS, |
|
37 | ALL_KINDS, | |
37 | CHANGELOGV2, |
|
38 | CHANGELOGV2, | |
38 | COMP_MODE_DEFAULT, |
|
39 | COMP_MODE_DEFAULT, | |
39 | COMP_MODE_INLINE, |
|
40 | COMP_MODE_INLINE, | |
40 | COMP_MODE_PLAIN, |
|
41 | COMP_MODE_PLAIN, | |
41 | DELTA_BASE_REUSE_NO, |
|
42 | DELTA_BASE_REUSE_NO, | |
42 | DELTA_BASE_REUSE_TRY, |
|
43 | DELTA_BASE_REUSE_TRY, | |
43 | ENTRY_RANK, |
|
44 | ENTRY_RANK, | |
44 | FEATURES_BY_VERSION, |
|
45 | FEATURES_BY_VERSION, | |
45 | FLAG_GENERALDELTA, |
|
46 | FLAG_GENERALDELTA, | |
46 | FLAG_INLINE_DATA, |
|
47 | FLAG_INLINE_DATA, | |
47 | INDEX_HEADER, |
|
48 | INDEX_HEADER, | |
48 | KIND_CHANGELOG, |
|
49 | KIND_CHANGELOG, | |
49 | KIND_FILELOG, |
|
50 | KIND_FILELOG, | |
50 | RANK_UNKNOWN, |
|
51 | RANK_UNKNOWN, | |
51 | REVLOGV0, |
|
52 | REVLOGV0, | |
52 | REVLOGV1, |
|
53 | REVLOGV1, | |
53 | REVLOGV1_FLAGS, |
|
54 | REVLOGV1_FLAGS, | |
54 | REVLOGV2, |
|
55 | REVLOGV2, | |
55 | REVLOGV2_FLAGS, |
|
56 | REVLOGV2_FLAGS, | |
56 | REVLOG_DEFAULT_FLAGS, |
|
57 | REVLOG_DEFAULT_FLAGS, | |
57 | REVLOG_DEFAULT_FORMAT, |
|
58 | REVLOG_DEFAULT_FORMAT, | |
58 | REVLOG_DEFAULT_VERSION, |
|
59 | REVLOG_DEFAULT_VERSION, | |
59 | SUPPORTED_FLAGS, |
|
60 | SUPPORTED_FLAGS, | |
60 | ) |
|
61 | ) | |
61 | from .revlogutils.flagutil import ( |
|
62 | from .revlogutils.flagutil import ( | |
62 | REVIDX_DEFAULT_FLAGS, |
|
63 | REVIDX_DEFAULT_FLAGS, | |
63 | REVIDX_ELLIPSIS, |
|
64 | REVIDX_ELLIPSIS, | |
64 | REVIDX_EXTSTORED, |
|
65 | REVIDX_EXTSTORED, | |
65 | REVIDX_FLAGS_ORDER, |
|
66 | REVIDX_FLAGS_ORDER, | |
66 | REVIDX_HASCOPIESINFO, |
|
67 | REVIDX_HASCOPIESINFO, | |
67 | REVIDX_ISCENSORED, |
|
68 | REVIDX_ISCENSORED, | |
68 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
69 | REVIDX_RAWTEXT_CHANGING_FLAGS, | |
69 | ) |
|
70 | ) | |
70 | from .thirdparty import attr |
|
71 | from .thirdparty import attr | |
71 | from . import ( |
|
72 | from . import ( | |
72 | ancestor, |
|
73 | ancestor, | |
73 | dagop, |
|
74 | dagop, | |
74 | error, |
|
75 | error, | |
75 | mdiff, |
|
76 | mdiff, | |
76 | policy, |
|
77 | policy, | |
77 | pycompat, |
|
78 | pycompat, | |
78 | revlogutils, |
|
79 | revlogutils, | |
79 | templatefilters, |
|
80 | templatefilters, | |
80 | util, |
|
81 | util, | |
81 | ) |
|
82 | ) | |
82 | from .interfaces import ( |
|
83 | from .interfaces import ( | |
83 | repository, |
|
84 | repository, | |
84 | util as interfaceutil, |
|
85 | util as interfaceutil, | |
85 | ) |
|
86 | ) | |
86 | from .revlogutils import ( |
|
87 | from .revlogutils import ( | |
87 | deltas as deltautil, |
|
88 | deltas as deltautil, | |
88 | docket as docketutil, |
|
89 | docket as docketutil, | |
89 | flagutil, |
|
90 | flagutil, | |
90 | nodemap as nodemaputil, |
|
91 | nodemap as nodemaputil, | |
91 | randomaccessfile, |
|
92 | randomaccessfile, | |
92 | revlogv0, |
|
93 | revlogv0, | |
93 | rewrite, |
|
94 | rewrite, | |
94 | sidedata as sidedatautil, |
|
95 | sidedata as sidedatautil, | |
95 | ) |
|
96 | ) | |
96 | from .utils import ( |
|
97 | from .utils import ( | |
97 | storageutil, |
|
98 | storageutil, | |
98 | stringutil, |
|
99 | stringutil, | |
99 | ) |
|
100 | ) | |
100 |
|
101 | |||
101 | # blanked usage of all the name to prevent pyflakes constraints |
|
102 | # blanked usage of all the name to prevent pyflakes constraints | |
102 | # We need these name available in the module for extensions. |
|
103 | # We need these name available in the module for extensions. | |
103 |
|
104 | |||
104 | REVLOGV0 |
|
105 | REVLOGV0 | |
105 | REVLOGV1 |
|
106 | REVLOGV1 | |
106 | REVLOGV2 |
|
107 | REVLOGV2 | |
107 | CHANGELOGV2 |
|
108 | CHANGELOGV2 | |
108 | FLAG_INLINE_DATA |
|
109 | FLAG_INLINE_DATA | |
109 | FLAG_GENERALDELTA |
|
110 | FLAG_GENERALDELTA | |
110 | REVLOG_DEFAULT_FLAGS |
|
111 | REVLOG_DEFAULT_FLAGS | |
111 | REVLOG_DEFAULT_FORMAT |
|
112 | REVLOG_DEFAULT_FORMAT | |
112 | REVLOG_DEFAULT_VERSION |
|
113 | REVLOG_DEFAULT_VERSION | |
113 | REVLOGV1_FLAGS |
|
114 | REVLOGV1_FLAGS | |
114 | REVLOGV2_FLAGS |
|
115 | REVLOGV2_FLAGS | |
115 | REVIDX_ISCENSORED |
|
116 | REVIDX_ISCENSORED | |
116 | REVIDX_ELLIPSIS |
|
117 | REVIDX_ELLIPSIS | |
117 | REVIDX_HASCOPIESINFO |
|
118 | REVIDX_HASCOPIESINFO | |
118 | REVIDX_EXTSTORED |
|
119 | REVIDX_EXTSTORED | |
119 | REVIDX_DEFAULT_FLAGS |
|
120 | REVIDX_DEFAULT_FLAGS | |
120 | REVIDX_FLAGS_ORDER |
|
121 | REVIDX_FLAGS_ORDER | |
121 | REVIDX_RAWTEXT_CHANGING_FLAGS |
|
122 | REVIDX_RAWTEXT_CHANGING_FLAGS | |
122 |
|
123 | |||
123 | parsers = policy.importmod('parsers') |
|
124 | parsers = policy.importmod('parsers') | |
124 | rustancestor = policy.importrust('ancestor') |
|
125 | rustancestor = policy.importrust('ancestor') | |
125 | rustdagop = policy.importrust('dagop') |
|
126 | rustdagop = policy.importrust('dagop') | |
126 | rustrevlog = policy.importrust('revlog') |
|
127 | rustrevlog = policy.importrust('revlog') | |
127 |
|
128 | |||
128 | # Aliased for performance. |
|
129 | # Aliased for performance. | |
129 | _zlibdecompress = zlib.decompress |
|
130 | _zlibdecompress = zlib.decompress | |
130 |
|
131 | |||
131 | # max size of inline data embedded into a revlog |
|
132 | # max size of inline data embedded into a revlog | |
132 | _maxinline = 131072 |
|
133 | _maxinline = 131072 | |
133 |
|
134 | |||
134 | # Flag processors for REVIDX_ELLIPSIS. |
|
135 | # Flag processors for REVIDX_ELLIPSIS. | |
135 | def ellipsisreadprocessor(rl, text): |
|
136 | def ellipsisreadprocessor(rl, text): | |
136 | return text, False |
|
137 | return text, False | |
137 |
|
138 | |||
138 |
|
139 | |||
139 | def ellipsiswriteprocessor(rl, text): |
|
140 | def ellipsiswriteprocessor(rl, text): | |
140 | return text, False |
|
141 | return text, False | |
141 |
|
142 | |||
142 |
|
143 | |||
143 | def ellipsisrawprocessor(rl, text): |
|
144 | def ellipsisrawprocessor(rl, text): | |
144 | return False |
|
145 | return False | |
145 |
|
146 | |||
146 |
|
147 | |||
147 | ellipsisprocessor = ( |
|
148 | ellipsisprocessor = ( | |
148 | ellipsisreadprocessor, |
|
149 | ellipsisreadprocessor, | |
149 | ellipsiswriteprocessor, |
|
150 | ellipsiswriteprocessor, | |
150 | ellipsisrawprocessor, |
|
151 | ellipsisrawprocessor, | |
151 | ) |
|
152 | ) | |
152 |
|
153 | |||
153 |
|
154 | |||
154 | def _verify_revision(rl, skipflags, state, node): |
|
155 | def _verify_revision(rl, skipflags, state, node): | |
155 | """Verify the integrity of the given revlog ``node`` while providing a hook |
|
156 | """Verify the integrity of the given revlog ``node`` while providing a hook | |
156 | point for extensions to influence the operation.""" |
|
157 | point for extensions to influence the operation.""" | |
157 | if skipflags: |
|
158 | if skipflags: | |
158 | state[b'skipread'].add(node) |
|
159 | state[b'skipread'].add(node) | |
159 | else: |
|
160 | else: | |
160 | # Side-effect: read content and verify hash. |
|
161 | # Side-effect: read content and verify hash. | |
161 | rl.revision(node) |
|
162 | rl.revision(node) | |
162 |
|
163 | |||
163 |
|
164 | |||
164 | # True if a fast implementation for persistent-nodemap is available |
|
165 | # True if a fast implementation for persistent-nodemap is available | |
165 | # |
|
166 | # | |
166 | # We also consider we have a "fast" implementation in "pure" python because |
|
167 | # We also consider we have a "fast" implementation in "pure" python because | |
167 | # people using pure don't really have performance consideration (and a |
|
168 | # people using pure don't really have performance consideration (and a | |
168 | # wheelbarrow of other slowness source) |
|
169 | # wheelbarrow of other slowness source) | |
169 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or hasattr( |
|
170 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or hasattr( | |
170 | parsers, 'BaseIndexObject' |
|
171 | parsers, 'BaseIndexObject' | |
171 | ) |
|
172 | ) | |
172 |
|
173 | |||
173 |
|
174 | |||
174 | @interfaceutil.implementer(repository.irevisiondelta) |
|
175 | @interfaceutil.implementer(repository.irevisiondelta) | |
175 | @attr.s(slots=True) |
|
176 | @attr.s(slots=True) | |
176 | class revlogrevisiondelta: |
|
177 | class revlogrevisiondelta: | |
177 | node = attr.ib() |
|
178 | node = attr.ib() | |
178 | p1node = attr.ib() |
|
179 | p1node = attr.ib() | |
179 | p2node = attr.ib() |
|
180 | p2node = attr.ib() | |
180 | basenode = attr.ib() |
|
181 | basenode = attr.ib() | |
181 | flags = attr.ib() |
|
182 | flags = attr.ib() | |
182 | baserevisionsize = attr.ib() |
|
183 | baserevisionsize = attr.ib() | |
183 | revision = attr.ib() |
|
184 | revision = attr.ib() | |
184 | delta = attr.ib() |
|
185 | delta = attr.ib() | |
185 | sidedata = attr.ib() |
|
186 | sidedata = attr.ib() | |
186 | protocol_flags = attr.ib() |
|
187 | protocol_flags = attr.ib() | |
187 | linknode = attr.ib(default=None) |
|
188 | linknode = attr.ib(default=None) | |
188 |
|
189 | |||
189 |
|
190 | |||
190 | @interfaceutil.implementer(repository.iverifyproblem) |
|
191 | @interfaceutil.implementer(repository.iverifyproblem) | |
191 | @attr.s(frozen=True) |
|
192 | @attr.s(frozen=True) | |
192 | class revlogproblem: |
|
193 | class revlogproblem: | |
193 | warning = attr.ib(default=None) |
|
194 | warning = attr.ib(default=None) | |
194 | error = attr.ib(default=None) |
|
195 | error = attr.ib(default=None) | |
195 | node = attr.ib(default=None) |
|
196 | node = attr.ib(default=None) | |
196 |
|
197 | |||
197 |
|
198 | |||
198 | def parse_index_v1(data, inline): |
|
199 | def parse_index_v1(data, inline): | |
199 | # call the C implementation to parse the index data |
|
200 | # call the C implementation to parse the index data | |
200 | index, cache = parsers.parse_index2(data, inline) |
|
201 | index, cache = parsers.parse_index2(data, inline) | |
201 | return index, cache |
|
202 | return index, cache | |
202 |
|
203 | |||
203 |
|
204 | |||
204 | def parse_index_v2(data, inline): |
|
205 | def parse_index_v2(data, inline): | |
205 | # call the C implementation to parse the index data |
|
206 | # call the C implementation to parse the index data | |
206 | index, cache = parsers.parse_index2(data, inline, format=REVLOGV2) |
|
207 | index, cache = parsers.parse_index2(data, inline, format=REVLOGV2) | |
207 | return index, cache |
|
208 | return index, cache | |
208 |
|
209 | |||
209 |
|
210 | |||
210 | def parse_index_cl_v2(data, inline): |
|
211 | def parse_index_cl_v2(data, inline): | |
211 | # call the C implementation to parse the index data |
|
212 | # call the C implementation to parse the index data | |
212 | index, cache = parsers.parse_index2(data, inline, format=CHANGELOGV2) |
|
213 | index, cache = parsers.parse_index2(data, inline, format=CHANGELOGV2) | |
213 | return index, cache |
|
214 | return index, cache | |
214 |
|
215 | |||
215 |
|
216 | |||
216 | if hasattr(parsers, 'parse_index_devel_nodemap'): |
|
217 | if hasattr(parsers, 'parse_index_devel_nodemap'): | |
217 |
|
218 | |||
218 | def parse_index_v1_nodemap(data, inline): |
|
219 | def parse_index_v1_nodemap(data, inline): | |
219 | index, cache = parsers.parse_index_devel_nodemap(data, inline) |
|
220 | index, cache = parsers.parse_index_devel_nodemap(data, inline) | |
220 | return index, cache |
|
221 | return index, cache | |
221 |
|
222 | |||
222 |
|
223 | |||
223 | else: |
|
224 | else: | |
224 | parse_index_v1_nodemap = None |
|
225 | parse_index_v1_nodemap = None | |
225 |
|
226 | |||
226 |
|
227 | |||
227 | def parse_index_v1_mixed(data, inline): |
|
228 | def parse_index_v1_mixed(data, inline, default_header): | |
228 | index, cache = parse_index_v1(data, inline) |
|
229 | index, cache = parse_index_v1(data, inline) | |
229 | return rustrevlog.MixedIndex(index, data), cache |
|
230 | return rustrevlog.MixedIndex(index, data, default_header), cache | |
230 |
|
231 | |||
231 |
|
232 | |||
232 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte |
|
233 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte | |
233 | # signed integer) |
|
234 | # signed integer) | |
234 | _maxentrysize = 0x7FFFFFFF |
|
235 | _maxentrysize = 0x7FFFFFFF | |
235 |
|
236 | |||
236 | FILE_TOO_SHORT_MSG = _( |
|
237 | FILE_TOO_SHORT_MSG = _( | |
237 | b'cannot read from revlog %s;' |
|
238 | b'cannot read from revlog %s;' | |
238 | b' expected %d bytes from offset %d, data size is %d' |
|
239 | b' expected %d bytes from offset %d, data size is %d' | |
239 | ) |
|
240 | ) | |
240 |
|
241 | |||
241 | hexdigits = b'0123456789abcdefABCDEF' |
|
242 | hexdigits = b'0123456789abcdefABCDEF' | |
242 |
|
243 | |||
243 |
|
244 | |||
244 | class _Config: |
|
245 | class _Config: | |
245 | def copy(self): |
|
246 | def copy(self): | |
246 | return self.__class__(**self.__dict__) |
|
247 | return self.__class__(**self.__dict__) | |
247 |
|
248 | |||
248 |
|
249 | |||
249 | @attr.s() |
|
250 | @attr.s() | |
250 | class FeatureConfig(_Config): |
|
251 | class FeatureConfig(_Config): | |
251 | """Hold configuration values about the available revlog features""" |
|
252 | """Hold configuration values about the available revlog features""" | |
252 |
|
253 | |||
253 | # the default compression engine |
|
254 | # the default compression engine | |
254 | compression_engine = attr.ib(default=b'zlib') |
|
255 | compression_engine = attr.ib(default=b'zlib') | |
255 | # compression engines options |
|
256 | # compression engines options | |
256 | compression_engine_options = attr.ib(default=attr.Factory(dict)) |
|
257 | compression_engine_options = attr.ib(default=attr.Factory(dict)) | |
257 |
|
258 | |||
258 | # can we use censor on this revlog |
|
259 | # can we use censor on this revlog | |
259 | censorable = attr.ib(default=False) |
|
260 | censorable = attr.ib(default=False) | |
260 | # does this revlog use the "side data" feature |
|
261 | # does this revlog use the "side data" feature | |
261 | has_side_data = attr.ib(default=False) |
|
262 | has_side_data = attr.ib(default=False) | |
262 | # might remove rank configuration once the computation has no impact |
|
263 | # might remove rank configuration once the computation has no impact | |
263 | compute_rank = attr.ib(default=False) |
|
264 | compute_rank = attr.ib(default=False) | |
264 | # parent order is supposed to be semantically irrelevant, so we |
|
265 | # parent order is supposed to be semantically irrelevant, so we | |
265 | # normally resort parents to ensure that the first parent is non-null, |
|
266 | # normally resort parents to ensure that the first parent is non-null, | |
266 | # if there is a non-null parent at all. |
|
267 | # if there is a non-null parent at all. | |
267 | # filelog abuses the parent order as flag to mark some instances of |
|
268 | # filelog abuses the parent order as flag to mark some instances of | |
268 | # meta-encoded files, so allow it to disable this behavior. |
|
269 | # meta-encoded files, so allow it to disable this behavior. | |
269 | canonical_parent_order = attr.ib(default=False) |
|
270 | canonical_parent_order = attr.ib(default=False) | |
270 | # can ellipsis commit be used |
|
271 | # can ellipsis commit be used | |
271 | enable_ellipsis = attr.ib(default=False) |
|
272 | enable_ellipsis = attr.ib(default=False) | |
272 |
|
273 | |||
273 | def copy(self): |
|
274 | def copy(self): | |
274 | new = super().copy() |
|
275 | new = super().copy() | |
275 | new.compression_engine_options = self.compression_engine_options.copy() |
|
276 | new.compression_engine_options = self.compression_engine_options.copy() | |
276 | return new |
|
277 | return new | |
277 |
|
278 | |||
278 |
|
279 | |||
279 | @attr.s() |
|
280 | @attr.s() | |
280 | class DataConfig(_Config): |
|
281 | class DataConfig(_Config): | |
281 | """Hold configuration value about how the revlog data are read""" |
|
282 | """Hold configuration value about how the revlog data are read""" | |
282 |
|
283 | |||
283 | # should we try to open the "pending" version of the revlog |
|
284 | # should we try to open the "pending" version of the revlog | |
284 | try_pending = attr.ib(default=False) |
|
285 | try_pending = attr.ib(default=False) | |
285 | # should we try to open the "splitted" version of the revlog |
|
286 | # should we try to open the "splitted" version of the revlog | |
286 | try_split = attr.ib(default=False) |
|
287 | try_split = attr.ib(default=False) | |
287 | # When True, indexfile should be opened with checkambig=True at writing, |
|
288 | # When True, indexfile should be opened with checkambig=True at writing, | |
288 | # to avoid file stat ambiguity. |
|
289 | # to avoid file stat ambiguity. | |
289 | check_ambig = attr.ib(default=False) |
|
290 | check_ambig = attr.ib(default=False) | |
290 |
|
291 | |||
291 | # If true, use mmap instead of reading to deal with large index |
|
292 | # If true, use mmap instead of reading to deal with large index | |
292 | mmap_large_index = attr.ib(default=False) |
|
293 | mmap_large_index = attr.ib(default=False) | |
293 | # how much data is large |
|
294 | # how much data is large | |
294 | mmap_index_threshold = attr.ib(default=None) |
|
295 | mmap_index_threshold = attr.ib(default=None) | |
295 | # How much data to read and cache into the raw revlog data cache. |
|
296 | # How much data to read and cache into the raw revlog data cache. | |
296 | chunk_cache_size = attr.ib(default=65536) |
|
297 | chunk_cache_size = attr.ib(default=65536) | |
297 |
|
298 | |||
298 | # The size of the uncompressed cache compared to the largest revision seen. |
|
299 | # The size of the uncompressed cache compared to the largest revision seen. | |
299 | uncompressed_cache_factor = attr.ib(default=None) |
|
300 | uncompressed_cache_factor = attr.ib(default=None) | |
300 |
|
301 | |||
301 | # The number of chunk cached |
|
302 | # The number of chunk cached | |
302 | uncompressed_cache_count = attr.ib(default=None) |
|
303 | uncompressed_cache_count = attr.ib(default=None) | |
303 |
|
304 | |||
304 | # Allow sparse reading of the revlog data |
|
305 | # Allow sparse reading of the revlog data | |
305 | with_sparse_read = attr.ib(default=False) |
|
306 | with_sparse_read = attr.ib(default=False) | |
306 | # minimal density of a sparse read chunk |
|
307 | # minimal density of a sparse read chunk | |
307 | sr_density_threshold = attr.ib(default=0.50) |
|
308 | sr_density_threshold = attr.ib(default=0.50) | |
308 | # minimal size of data we skip when performing sparse read |
|
309 | # minimal size of data we skip when performing sparse read | |
309 | sr_min_gap_size = attr.ib(default=262144) |
|
310 | sr_min_gap_size = attr.ib(default=262144) | |
310 |
|
311 | |||
311 | # are delta encoded against arbitrary bases. |
|
312 | # are delta encoded against arbitrary bases. | |
312 | generaldelta = attr.ib(default=False) |
|
313 | generaldelta = attr.ib(default=False) | |
313 |
|
314 | |||
314 |
|
315 | |||
315 | @attr.s() |
|
316 | @attr.s() | |
316 | class DeltaConfig(_Config): |
|
317 | class DeltaConfig(_Config): | |
317 | """Hold configuration value about how new delta are computed |
|
318 | """Hold configuration value about how new delta are computed | |
318 |
|
319 | |||
319 | Some attributes are duplicated from DataConfig to help havign each object |
|
320 | Some attributes are duplicated from DataConfig to help havign each object | |
320 | self contained. |
|
321 | self contained. | |
321 | """ |
|
322 | """ | |
322 |
|
323 | |||
323 | # can delta be encoded against arbitrary bases. |
|
324 | # can delta be encoded against arbitrary bases. | |
324 | general_delta = attr.ib(default=False) |
|
325 | general_delta = attr.ib(default=False) | |
325 | # Allow sparse writing of the revlog data |
|
326 | # Allow sparse writing of the revlog data | |
326 | sparse_revlog = attr.ib(default=False) |
|
327 | sparse_revlog = attr.ib(default=False) | |
327 | # maximum length of a delta chain |
|
328 | # maximum length of a delta chain | |
328 | max_chain_len = attr.ib(default=None) |
|
329 | max_chain_len = attr.ib(default=None) | |
329 | # Maximum distance between delta chain base start and end |
|
330 | # Maximum distance between delta chain base start and end | |
330 | max_deltachain_span = attr.ib(default=-1) |
|
331 | max_deltachain_span = attr.ib(default=-1) | |
331 | # If `upper_bound_comp` is not None, this is the expected maximal gain from |
|
332 | # If `upper_bound_comp` is not None, this is the expected maximal gain from | |
332 | # compression for the data content. |
|
333 | # compression for the data content. | |
333 | upper_bound_comp = attr.ib(default=None) |
|
334 | upper_bound_comp = attr.ib(default=None) | |
334 | # Should we try a delta against both parent |
|
335 | # Should we try a delta against both parent | |
335 | delta_both_parents = attr.ib(default=True) |
|
336 | delta_both_parents = attr.ib(default=True) | |
336 | # Test delta base candidate group by chunk of this maximal size. |
|
337 | # Test delta base candidate group by chunk of this maximal size. | |
337 | candidate_group_chunk_size = attr.ib(default=0) |
|
338 | candidate_group_chunk_size = attr.ib(default=0) | |
338 | # Should we display debug information about delta computation |
|
339 | # Should we display debug information about delta computation | |
339 | debug_delta = attr.ib(default=False) |
|
340 | debug_delta = attr.ib(default=False) | |
340 | # trust incoming delta by default |
|
341 | # trust incoming delta by default | |
341 | lazy_delta = attr.ib(default=True) |
|
342 | lazy_delta = attr.ib(default=True) | |
342 | # trust the base of incoming delta by default |
|
343 | # trust the base of incoming delta by default | |
343 | lazy_delta_base = attr.ib(default=False) |
|
344 | lazy_delta_base = attr.ib(default=False) | |
344 |
|
345 | |||
345 |
|
346 | |||
346 | class _InnerRevlog: |
|
347 | class _InnerRevlog: | |
347 | """An inner layer of the revlog object |
|
348 | """An inner layer of the revlog object | |
348 |
|
349 | |||
349 | That layer exist to be able to delegate some operation to Rust, its |
|
350 | That layer exist to be able to delegate some operation to Rust, its | |
350 | boundaries are arbitrary and based on what we can delegate to Rust. |
|
351 | boundaries are arbitrary and based on what we can delegate to Rust. | |
351 | """ |
|
352 | """ | |
352 |
|
353 | |||
353 | def __init__( |
|
354 | def __init__( | |
354 | self, |
|
355 | self, | |
355 | opener, |
|
356 | opener, | |
356 | index, |
|
357 | index, | |
357 | index_file, |
|
358 | index_file, | |
358 | data_file, |
|
359 | data_file, | |
359 | sidedata_file, |
|
360 | sidedata_file, | |
360 | inline, |
|
361 | inline, | |
361 | data_config, |
|
362 | data_config, | |
362 | delta_config, |
|
363 | delta_config, | |
363 | feature_config, |
|
364 | feature_config, | |
364 | chunk_cache, |
|
365 | chunk_cache, | |
365 | default_compression_header, |
|
366 | default_compression_header, | |
366 | ): |
|
367 | ): | |
367 | self.opener = opener |
|
368 | self.opener = opener | |
368 | self.index = index |
|
369 | self.index = index | |
369 |
|
370 | |||
370 | self.index_file = index_file |
|
371 | self.index_file = index_file | |
371 | self.data_file = data_file |
|
372 | self.data_file = data_file | |
372 | self.sidedata_file = sidedata_file |
|
373 | self.sidedata_file = sidedata_file | |
373 | self.inline = inline |
|
374 | self.inline = inline | |
374 | self.data_config = data_config |
|
375 | self.data_config = data_config | |
375 | self.delta_config = delta_config |
|
376 | self.delta_config = delta_config | |
376 | self.feature_config = feature_config |
|
377 | self.feature_config = feature_config | |
377 |
|
378 | |||
378 | # used during diverted write. |
|
379 | # used during diverted write. | |
379 | self._orig_index_file = None |
|
380 | self._orig_index_file = None | |
380 |
|
381 | |||
381 | self._default_compression_header = default_compression_header |
|
382 | self._default_compression_header = default_compression_header | |
382 |
|
383 | |||
383 | # index |
|
384 | # index | |
384 |
|
385 | |||
385 | # 3-tuple of file handles being used for active writing. |
|
386 | # 3-tuple of file handles being used for active writing. | |
386 | self._writinghandles = None |
|
387 | self._writinghandles = None | |
387 |
|
388 | |||
388 | self._segmentfile = randomaccessfile.randomaccessfile( |
|
389 | self._segmentfile = randomaccessfile.randomaccessfile( | |
389 | self.opener, |
|
390 | self.opener, | |
390 | (self.index_file if self.inline else self.data_file), |
|
391 | (self.index_file if self.inline else self.data_file), | |
391 | self.data_config.chunk_cache_size, |
|
392 | self.data_config.chunk_cache_size, | |
392 | chunk_cache, |
|
393 | chunk_cache, | |
393 | ) |
|
394 | ) | |
394 | self._segmentfile_sidedata = randomaccessfile.randomaccessfile( |
|
395 | self._segmentfile_sidedata = randomaccessfile.randomaccessfile( | |
395 | self.opener, |
|
396 | self.opener, | |
396 | self.sidedata_file, |
|
397 | self.sidedata_file, | |
397 | self.data_config.chunk_cache_size, |
|
398 | self.data_config.chunk_cache_size, | |
398 | ) |
|
399 | ) | |
399 |
|
400 | |||
400 | # revlog header -> revlog compressor |
|
401 | # revlog header -> revlog compressor | |
401 | self._decompressors = {} |
|
402 | self._decompressors = {} | |
402 | # 3-tuple of (node, rev, text) for a raw revision. |
|
403 | # 3-tuple of (node, rev, text) for a raw revision. | |
403 | self._revisioncache = None |
|
404 | self._revisioncache = None | |
404 |
|
405 | |||
405 | # cache some uncompressed chunks |
|
406 | # cache some uncompressed chunks | |
406 | # rev → uncompressed_chunk |
|
407 | # rev → uncompressed_chunk | |
407 | # |
|
408 | # | |
408 | # the max cost is dynamically updated to be proportionnal to the |
|
409 | # the max cost is dynamically updated to be proportionnal to the | |
409 | # size of revision we actually encounter. |
|
410 | # size of revision we actually encounter. | |
410 | self._uncompressed_chunk_cache = None |
|
411 | self._uncompressed_chunk_cache = None | |
411 | if self.data_config.uncompressed_cache_factor is not None: |
|
412 | if self.data_config.uncompressed_cache_factor is not None: | |
412 | self._uncompressed_chunk_cache = util.lrucachedict( |
|
413 | self._uncompressed_chunk_cache = util.lrucachedict( | |
413 | self.data_config.uncompressed_cache_count, |
|
414 | self.data_config.uncompressed_cache_count, | |
414 | maxcost=65536, # some arbitrary initial value |
|
415 | maxcost=65536, # some arbitrary initial value | |
415 | ) |
|
416 | ) | |
416 |
|
417 | |||
417 | self._delay_buffer = None |
|
418 | self._delay_buffer = None | |
418 |
|
419 | |||
419 | def __len__(self): |
|
420 | def __len__(self): | |
420 | return len(self.index) |
|
421 | return len(self.index) | |
421 |
|
422 | |||
422 | def clear_cache(self): |
|
423 | def clear_cache(self): | |
423 | assert not self.is_delaying |
|
424 | assert not self.is_delaying | |
424 | self._revisioncache = None |
|
425 | self._revisioncache = None | |
425 | if self._uncompressed_chunk_cache is not None: |
|
426 | if self._uncompressed_chunk_cache is not None: | |
426 | self._uncompressed_chunk_cache.clear() |
|
427 | self._uncompressed_chunk_cache.clear() | |
427 | self._segmentfile.clear_cache() |
|
428 | self._segmentfile.clear_cache() | |
428 | self._segmentfile_sidedata.clear_cache() |
|
429 | self._segmentfile_sidedata.clear_cache() | |
429 |
|
430 | |||
430 | @property |
|
431 | @property | |
431 | def canonical_index_file(self): |
|
432 | def canonical_index_file(self): | |
432 | if self._orig_index_file is not None: |
|
433 | if self._orig_index_file is not None: | |
433 | return self._orig_index_file |
|
434 | return self._orig_index_file | |
434 | return self.index_file |
|
435 | return self.index_file | |
435 |
|
436 | |||
436 | @property |
|
437 | @property | |
437 | def is_delaying(self): |
|
438 | def is_delaying(self): | |
438 | """is the revlog is currently delaying the visibility of written data? |
|
439 | """is the revlog is currently delaying the visibility of written data? | |
439 |
|
440 | |||
440 | The delaying mechanism can be either in-memory or written on disk in a |
|
441 | The delaying mechanism can be either in-memory or written on disk in a | |
441 | side-file.""" |
|
442 | side-file.""" | |
442 | return (self._delay_buffer is not None) or ( |
|
443 | return (self._delay_buffer is not None) or ( | |
443 | self._orig_index_file is not None |
|
444 | self._orig_index_file is not None | |
444 | ) |
|
445 | ) | |
445 |
|
446 | |||
446 | # Derived from index values. |
|
447 | # Derived from index values. | |
447 |
|
448 | |||
448 | def start(self, rev): |
|
449 | def start(self, rev): | |
449 | """the offset of the data chunk for this revision""" |
|
450 | """the offset of the data chunk for this revision""" | |
450 | return int(self.index[rev][0] >> 16) |
|
451 | return int(self.index[rev][0] >> 16) | |
451 |
|
452 | |||
452 | def length(self, rev): |
|
453 | def length(self, rev): | |
453 | """the length of the data chunk for this revision""" |
|
454 | """the length of the data chunk for this revision""" | |
454 | return self.index[rev][1] |
|
455 | return self.index[rev][1] | |
455 |
|
456 | |||
456 | def end(self, rev): |
|
457 | def end(self, rev): | |
457 | """the end of the data chunk for this revision""" |
|
458 | """the end of the data chunk for this revision""" | |
458 | return self.start(rev) + self.length(rev) |
|
459 | return self.start(rev) + self.length(rev) | |
459 |
|
460 | |||
460 | def deltaparent(self, rev): |
|
461 | def deltaparent(self, rev): | |
461 | """return deltaparent of the given revision""" |
|
462 | """return deltaparent of the given revision""" | |
462 | base = self.index[rev][3] |
|
463 | base = self.index[rev][3] | |
463 | if base == rev: |
|
464 | if base == rev: | |
464 | return nullrev |
|
465 | return nullrev | |
465 | elif self.delta_config.general_delta: |
|
466 | elif self.delta_config.general_delta: | |
466 | return base |
|
467 | return base | |
467 | else: |
|
468 | else: | |
468 | return rev - 1 |
|
469 | return rev - 1 | |
469 |
|
470 | |||
470 | def issnapshot(self, rev): |
|
471 | def issnapshot(self, rev): | |
471 | """tells whether rev is a snapshot""" |
|
472 | """tells whether rev is a snapshot""" | |
472 | if not self.delta_config.sparse_revlog: |
|
473 | if not self.delta_config.sparse_revlog: | |
473 | return self.deltaparent(rev) == nullrev |
|
474 | return self.deltaparent(rev) == nullrev | |
474 | elif hasattr(self.index, 'issnapshot'): |
|
475 | elif hasattr(self.index, 'issnapshot'): | |
475 | # directly assign the method to cache the testing and access |
|
476 | # directly assign the method to cache the testing and access | |
476 | self.issnapshot = self.index.issnapshot |
|
477 | self.issnapshot = self.index.issnapshot | |
477 | return self.issnapshot(rev) |
|
478 | return self.issnapshot(rev) | |
478 | if rev == nullrev: |
|
479 | if rev == nullrev: | |
479 | return True |
|
480 | return True | |
480 | entry = self.index[rev] |
|
481 | entry = self.index[rev] | |
481 | base = entry[3] |
|
482 | base = entry[3] | |
482 | if base == rev: |
|
483 | if base == rev: | |
483 | return True |
|
484 | return True | |
484 | if base == nullrev: |
|
485 | if base == nullrev: | |
485 | return True |
|
486 | return True | |
486 | p1 = entry[5] |
|
487 | p1 = entry[5] | |
487 | while self.length(p1) == 0: |
|
488 | while self.length(p1) == 0: | |
488 | b = self.deltaparent(p1) |
|
489 | b = self.deltaparent(p1) | |
489 | if b == p1: |
|
490 | if b == p1: | |
490 | break |
|
491 | break | |
491 | p1 = b |
|
492 | p1 = b | |
492 | p2 = entry[6] |
|
493 | p2 = entry[6] | |
493 | while self.length(p2) == 0: |
|
494 | while self.length(p2) == 0: | |
494 | b = self.deltaparent(p2) |
|
495 | b = self.deltaparent(p2) | |
495 | if b == p2: |
|
496 | if b == p2: | |
496 | break |
|
497 | break | |
497 | p2 = b |
|
498 | p2 = b | |
498 | if base == p1 or base == p2: |
|
499 | if base == p1 or base == p2: | |
499 | return False |
|
500 | return False | |
500 | return self.issnapshot(base) |
|
501 | return self.issnapshot(base) | |
501 |
|
502 | |||
502 | def _deltachain(self, rev, stoprev=None): |
|
503 | def _deltachain(self, rev, stoprev=None): | |
503 | """Obtain the delta chain for a revision. |
|
504 | """Obtain the delta chain for a revision. | |
504 |
|
505 | |||
505 | ``stoprev`` specifies a revision to stop at. If not specified, we |
|
506 | ``stoprev`` specifies a revision to stop at. If not specified, we | |
506 | stop at the base of the chain. |
|
507 | stop at the base of the chain. | |
507 |
|
508 | |||
508 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of |
|
509 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of | |
509 | revs in ascending order and ``stopped`` is a bool indicating whether |
|
510 | revs in ascending order and ``stopped`` is a bool indicating whether | |
510 | ``stoprev`` was hit. |
|
511 | ``stoprev`` was hit. | |
511 | """ |
|
512 | """ | |
512 | generaldelta = self.delta_config.general_delta |
|
513 | generaldelta = self.delta_config.general_delta | |
513 | # Try C implementation. |
|
514 | # Try C implementation. | |
514 | try: |
|
515 | try: | |
515 | return self.index.deltachain(rev, stoprev, generaldelta) |
|
516 | return self.index.deltachain(rev, stoprev, generaldelta) | |
516 | except AttributeError: |
|
517 | except AttributeError: | |
517 | pass |
|
518 | pass | |
518 |
|
519 | |||
519 | chain = [] |
|
520 | chain = [] | |
520 |
|
521 | |||
521 | # Alias to prevent attribute lookup in tight loop. |
|
522 | # Alias to prevent attribute lookup in tight loop. | |
522 | index = self.index |
|
523 | index = self.index | |
523 |
|
524 | |||
524 | iterrev = rev |
|
525 | iterrev = rev | |
525 | e = index[iterrev] |
|
526 | e = index[iterrev] | |
526 | while iterrev != e[3] and iterrev != stoprev: |
|
527 | while iterrev != e[3] and iterrev != stoprev: | |
527 | chain.append(iterrev) |
|
528 | chain.append(iterrev) | |
528 | if generaldelta: |
|
529 | if generaldelta: | |
529 | iterrev = e[3] |
|
530 | iterrev = e[3] | |
530 | else: |
|
531 | else: | |
531 | iterrev -= 1 |
|
532 | iterrev -= 1 | |
532 | e = index[iterrev] |
|
533 | e = index[iterrev] | |
533 |
|
534 | |||
534 | if iterrev == stoprev: |
|
535 | if iterrev == stoprev: | |
535 | stopped = True |
|
536 | stopped = True | |
536 | else: |
|
537 | else: | |
537 | chain.append(iterrev) |
|
538 | chain.append(iterrev) | |
538 | stopped = False |
|
539 | stopped = False | |
539 |
|
540 | |||
540 | chain.reverse() |
|
541 | chain.reverse() | |
541 | return chain, stopped |
|
542 | return chain, stopped | |
542 |
|
543 | |||
543 | @util.propertycache |
|
544 | @util.propertycache | |
544 | def _compressor(self): |
|
545 | def _compressor(self): | |
545 | engine = util.compengines[self.feature_config.compression_engine] |
|
546 | engine = util.compengines[self.feature_config.compression_engine] | |
546 | return engine.revlogcompressor( |
|
547 | return engine.revlogcompressor( | |
547 | self.feature_config.compression_engine_options |
|
548 | self.feature_config.compression_engine_options | |
548 | ) |
|
549 | ) | |
549 |
|
550 | |||
550 | @util.propertycache |
|
551 | @util.propertycache | |
551 | def _decompressor(self): |
|
552 | def _decompressor(self): | |
552 | """the default decompressor""" |
|
553 | """the default decompressor""" | |
553 | if self._default_compression_header is None: |
|
554 | if self._default_compression_header is None: | |
554 | return None |
|
555 | return None | |
555 | t = self._default_compression_header |
|
556 | t = self._default_compression_header | |
556 | c = self._get_decompressor(t) |
|
557 | c = self._get_decompressor(t) | |
557 | return c.decompress |
|
558 | return c.decompress | |
558 |
|
559 | |||
559 | def _get_decompressor(self, t): |
|
560 | def _get_decompressor(self, t): | |
560 | try: |
|
561 | try: | |
561 | compressor = self._decompressors[t] |
|
562 | compressor = self._decompressors[t] | |
562 | except KeyError: |
|
563 | except KeyError: | |
563 | try: |
|
564 | try: | |
564 | engine = util.compengines.forrevlogheader(t) |
|
565 | engine = util.compengines.forrevlogheader(t) | |
565 | compressor = engine.revlogcompressor( |
|
566 | compressor = engine.revlogcompressor( | |
566 | self.feature_config.compression_engine_options |
|
567 | self.feature_config.compression_engine_options | |
567 | ) |
|
568 | ) | |
568 | self._decompressors[t] = compressor |
|
569 | self._decompressors[t] = compressor | |
569 | except KeyError: |
|
570 | except KeyError: | |
570 | raise error.RevlogError( |
|
571 | raise error.RevlogError( | |
571 | _(b'unknown compression type %s') % binascii.hexlify(t) |
|
572 | _(b'unknown compression type %s') % binascii.hexlify(t) | |
572 | ) |
|
573 | ) | |
573 | return compressor |
|
574 | return compressor | |
574 |
|
575 | |||
575 | def compress(self, data): |
|
576 | def compress(self, data): | |
576 | """Generate a possibly-compressed representation of data.""" |
|
577 | """Generate a possibly-compressed representation of data.""" | |
577 | if not data: |
|
578 | if not data: | |
578 | return b'', data |
|
579 | return b'', data | |
579 |
|
580 | |||
580 | compressed = self._compressor.compress(data) |
|
581 | compressed = self._compressor.compress(data) | |
581 |
|
582 | |||
582 | if compressed: |
|
583 | if compressed: | |
583 | # The revlog compressor added the header in the returned data. |
|
584 | # The revlog compressor added the header in the returned data. | |
584 | return b'', compressed |
|
585 | return b'', compressed | |
585 |
|
586 | |||
586 | if data[0:1] == b'\0': |
|
587 | if data[0:1] == b'\0': | |
587 | return b'', data |
|
588 | return b'', data | |
588 | return b'u', data |
|
589 | return b'u', data | |
589 |
|
590 | |||
590 | def decompress(self, data): |
|
591 | def decompress(self, data): | |
591 | """Decompress a revlog chunk. |
|
592 | """Decompress a revlog chunk. | |
592 |
|
593 | |||
593 | The chunk is expected to begin with a header identifying the |
|
594 | The chunk is expected to begin with a header identifying the | |
594 | format type so it can be routed to an appropriate decompressor. |
|
595 | format type so it can be routed to an appropriate decompressor. | |
595 | """ |
|
596 | """ | |
596 | if not data: |
|
597 | if not data: | |
597 | return data |
|
598 | return data | |
598 |
|
599 | |||
599 | # Revlogs are read much more frequently than they are written and many |
|
600 | # Revlogs are read much more frequently than they are written and many | |
600 | # chunks only take microseconds to decompress, so performance is |
|
601 | # chunks only take microseconds to decompress, so performance is | |
601 | # important here. |
|
602 | # important here. | |
602 | # |
|
603 | # | |
603 | # We can make a few assumptions about revlogs: |
|
604 | # We can make a few assumptions about revlogs: | |
604 | # |
|
605 | # | |
605 | # 1) the majority of chunks will be compressed (as opposed to inline |
|
606 | # 1) the majority of chunks will be compressed (as opposed to inline | |
606 | # raw data). |
|
607 | # raw data). | |
607 | # 2) decompressing *any* data will likely by at least 10x slower than |
|
608 | # 2) decompressing *any* data will likely by at least 10x slower than | |
608 | # returning raw inline data. |
|
609 | # returning raw inline data. | |
609 | # 3) we want to prioritize common and officially supported compression |
|
610 | # 3) we want to prioritize common and officially supported compression | |
610 | # engines |
|
611 | # engines | |
611 | # |
|
612 | # | |
612 | # It follows that we want to optimize for "decompress compressed data |
|
613 | # It follows that we want to optimize for "decompress compressed data | |
613 | # when encoded with common and officially supported compression engines" |
|
614 | # when encoded with common and officially supported compression engines" | |
614 | # case over "raw data" and "data encoded by less common or non-official |
|
615 | # case over "raw data" and "data encoded by less common or non-official | |
615 | # compression engines." That is why we have the inline lookup first |
|
616 | # compression engines." That is why we have the inline lookup first | |
616 | # followed by the compengines lookup. |
|
617 | # followed by the compengines lookup. | |
617 | # |
|
618 | # | |
618 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib |
|
619 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib | |
619 | # compressed chunks. And this matters for changelog and manifest reads. |
|
620 | # compressed chunks. And this matters for changelog and manifest reads. | |
620 | t = data[0:1] |
|
621 | t = data[0:1] | |
621 |
|
622 | |||
622 | if t == b'x': |
|
623 | if t == b'x': | |
623 | try: |
|
624 | try: | |
624 | return _zlibdecompress(data) |
|
625 | return _zlibdecompress(data) | |
625 | except zlib.error as e: |
|
626 | except zlib.error as e: | |
626 | raise error.RevlogError( |
|
627 | raise error.RevlogError( | |
627 | _(b'revlog decompress error: %s') |
|
628 | _(b'revlog decompress error: %s') | |
628 | % stringutil.forcebytestr(e) |
|
629 | % stringutil.forcebytestr(e) | |
629 | ) |
|
630 | ) | |
630 | # '\0' is more common than 'u' so it goes first. |
|
631 | # '\0' is more common than 'u' so it goes first. | |
631 | elif t == b'\0': |
|
632 | elif t == b'\0': | |
632 | return data |
|
633 | return data | |
633 | elif t == b'u': |
|
634 | elif t == b'u': | |
634 | return util.buffer(data, 1) |
|
635 | return util.buffer(data, 1) | |
635 |
|
636 | |||
636 | compressor = self._get_decompressor(t) |
|
637 | compressor = self._get_decompressor(t) | |
637 |
|
638 | |||
638 | return compressor.decompress(data) |
|
639 | return compressor.decompress(data) | |
639 |
|
640 | |||
640 | @contextlib.contextmanager |
|
641 | @contextlib.contextmanager | |
641 | def reading(self): |
|
642 | def reading(self): | |
642 | """Context manager that keeps data and sidedata files open for reading""" |
|
643 | """Context manager that keeps data and sidedata files open for reading""" | |
643 | if len(self.index) == 0: |
|
644 | if len(self.index) == 0: | |
644 | yield # nothing to be read |
|
645 | yield # nothing to be read | |
645 | elif self._delay_buffer is not None and self.inline: |
|
646 | elif self._delay_buffer is not None and self.inline: | |
646 | msg = "revlog with delayed write should not be inline" |
|
647 | msg = "revlog with delayed write should not be inline" | |
647 | raise error.ProgrammingError(msg) |
|
648 | raise error.ProgrammingError(msg) | |
648 | else: |
|
649 | else: | |
649 | with self._segmentfile.reading(): |
|
650 | with self._segmentfile.reading(): | |
650 | with self._segmentfile_sidedata.reading(): |
|
651 | with self._segmentfile_sidedata.reading(): | |
651 | yield |
|
652 | yield | |
652 |
|
653 | |||
653 | @property |
|
654 | @property | |
654 | def is_writing(self): |
|
655 | def is_writing(self): | |
655 | """True is a writing context is open""" |
|
656 | """True is a writing context is open""" | |
656 | return self._writinghandles is not None |
|
657 | return self._writinghandles is not None | |
657 |
|
658 | |||
658 | @property |
|
659 | @property | |
659 | def is_open(self): |
|
660 | def is_open(self): | |
660 | """True if any file handle is being held |
|
661 | """True if any file handle is being held | |
661 |
|
662 | |||
662 | Used for assert and debug in the python code""" |
|
663 | Used for assert and debug in the python code""" | |
663 | return self._segmentfile.is_open or self._segmentfile_sidedata.is_open |
|
664 | return self._segmentfile.is_open or self._segmentfile_sidedata.is_open | |
664 |
|
665 | |||
665 | @contextlib.contextmanager |
|
666 | @contextlib.contextmanager | |
666 | def writing(self, transaction, data_end=None, sidedata_end=None): |
|
667 | def writing(self, transaction, data_end=None, sidedata_end=None): | |
667 | """Open the revlog files for writing |
|
668 | """Open the revlog files for writing | |
668 |
|
669 | |||
669 | Add content to a revlog should be done within such context. |
|
670 | Add content to a revlog should be done within such context. | |
670 | """ |
|
671 | """ | |
671 | if self.is_writing: |
|
672 | if self.is_writing: | |
672 | yield |
|
673 | yield | |
673 | else: |
|
674 | else: | |
674 | ifh = dfh = sdfh = None |
|
675 | ifh = dfh = sdfh = None | |
675 | try: |
|
676 | try: | |
676 | r = len(self.index) |
|
677 | r = len(self.index) | |
677 | # opening the data file. |
|
678 | # opening the data file. | |
678 | dsize = 0 |
|
679 | dsize = 0 | |
679 | if r: |
|
680 | if r: | |
680 | dsize = self.end(r - 1) |
|
681 | dsize = self.end(r - 1) | |
681 | dfh = None |
|
682 | dfh = None | |
682 | if not self.inline: |
|
683 | if not self.inline: | |
683 | try: |
|
684 | try: | |
684 | dfh = self.opener(self.data_file, mode=b"r+") |
|
685 | dfh = self.opener(self.data_file, mode=b"r+") | |
685 | if data_end is None: |
|
686 | if data_end is None: | |
686 | dfh.seek(0, os.SEEK_END) |
|
687 | dfh.seek(0, os.SEEK_END) | |
687 | else: |
|
688 | else: | |
688 | dfh.seek(data_end, os.SEEK_SET) |
|
689 | dfh.seek(data_end, os.SEEK_SET) | |
689 | except FileNotFoundError: |
|
690 | except FileNotFoundError: | |
690 | dfh = self.opener(self.data_file, mode=b"w+") |
|
691 | dfh = self.opener(self.data_file, mode=b"w+") | |
691 | transaction.add(self.data_file, dsize) |
|
692 | transaction.add(self.data_file, dsize) | |
692 | if self.sidedata_file is not None: |
|
693 | if self.sidedata_file is not None: | |
693 | assert sidedata_end is not None |
|
694 | assert sidedata_end is not None | |
694 | # revlog-v2 does not inline, help Pytype |
|
695 | # revlog-v2 does not inline, help Pytype | |
695 | assert dfh is not None |
|
696 | assert dfh is not None | |
696 | try: |
|
697 | try: | |
697 | sdfh = self.opener(self.sidedata_file, mode=b"r+") |
|
698 | sdfh = self.opener(self.sidedata_file, mode=b"r+") | |
698 | dfh.seek(sidedata_end, os.SEEK_SET) |
|
699 | dfh.seek(sidedata_end, os.SEEK_SET) | |
699 | except FileNotFoundError: |
|
700 | except FileNotFoundError: | |
700 | sdfh = self.opener(self.sidedata_file, mode=b"w+") |
|
701 | sdfh = self.opener(self.sidedata_file, mode=b"w+") | |
701 | transaction.add(self.sidedata_file, sidedata_end) |
|
702 | transaction.add(self.sidedata_file, sidedata_end) | |
702 |
|
703 | |||
703 | # opening the index file. |
|
704 | # opening the index file. | |
704 | isize = r * self.index.entry_size |
|
705 | isize = r * self.index.entry_size | |
705 | ifh = self.__index_write_fp() |
|
706 | ifh = self.__index_write_fp() | |
706 | if self.inline: |
|
707 | if self.inline: | |
707 | transaction.add(self.index_file, dsize + isize) |
|
708 | transaction.add(self.index_file, dsize + isize) | |
708 | else: |
|
709 | else: | |
709 | transaction.add(self.index_file, isize) |
|
710 | transaction.add(self.index_file, isize) | |
710 | # exposing all file handle for writing. |
|
711 | # exposing all file handle for writing. | |
711 | self._writinghandles = (ifh, dfh, sdfh) |
|
712 | self._writinghandles = (ifh, dfh, sdfh) | |
712 | self._segmentfile.writing_handle = ifh if self.inline else dfh |
|
713 | self._segmentfile.writing_handle = ifh if self.inline else dfh | |
713 | self._segmentfile_sidedata.writing_handle = sdfh |
|
714 | self._segmentfile_sidedata.writing_handle = sdfh | |
714 | yield |
|
715 | yield | |
715 | finally: |
|
716 | finally: | |
716 | self._writinghandles = None |
|
717 | self._writinghandles = None | |
717 | self._segmentfile.writing_handle = None |
|
718 | self._segmentfile.writing_handle = None | |
718 | self._segmentfile_sidedata.writing_handle = None |
|
719 | self._segmentfile_sidedata.writing_handle = None | |
719 | if dfh is not None: |
|
720 | if dfh is not None: | |
720 | dfh.close() |
|
721 | dfh.close() | |
721 | if sdfh is not None: |
|
722 | if sdfh is not None: | |
722 | sdfh.close() |
|
723 | sdfh.close() | |
723 | # closing the index file last to avoid exposing referent to |
|
724 | # closing the index file last to avoid exposing referent to | |
724 | # potential unflushed data content. |
|
725 | # potential unflushed data content. | |
725 | if ifh is not None: |
|
726 | if ifh is not None: | |
726 | ifh.close() |
|
727 | ifh.close() | |
727 |
|
728 | |||
728 | def __index_write_fp(self, index_end=None): |
|
729 | def __index_write_fp(self, index_end=None): | |
729 | """internal method to open the index file for writing |
|
730 | """internal method to open the index file for writing | |
730 |
|
731 | |||
731 | You should not use this directly and use `_writing` instead |
|
732 | You should not use this directly and use `_writing` instead | |
732 | """ |
|
733 | """ | |
733 | try: |
|
734 | try: | |
734 | if self._delay_buffer is None: |
|
735 | if self._delay_buffer is None: | |
735 | f = self.opener( |
|
736 | f = self.opener( | |
736 | self.index_file, |
|
737 | self.index_file, | |
737 | mode=b"r+", |
|
738 | mode=b"r+", | |
738 | checkambig=self.data_config.check_ambig, |
|
739 | checkambig=self.data_config.check_ambig, | |
739 | ) |
|
740 | ) | |
740 | else: |
|
741 | else: | |
741 | # check_ambig affect we way we open file for writing, however |
|
742 | # check_ambig affect we way we open file for writing, however | |
742 | # here, we do not actually open a file for writting as write |
|
743 | # here, we do not actually open a file for writting as write | |
743 | # will appened to a delay_buffer. So check_ambig is not |
|
744 | # will appened to a delay_buffer. So check_ambig is not | |
744 | # meaningful and unneeded here. |
|
745 | # meaningful and unneeded here. | |
745 | f = randomaccessfile.appender( |
|
746 | f = randomaccessfile.appender( | |
746 | self.opener, self.index_file, b"r+", self._delay_buffer |
|
747 | self.opener, self.index_file, b"r+", self._delay_buffer | |
747 | ) |
|
748 | ) | |
748 | if index_end is None: |
|
749 | if index_end is None: | |
749 | f.seek(0, os.SEEK_END) |
|
750 | f.seek(0, os.SEEK_END) | |
750 | else: |
|
751 | else: | |
751 | f.seek(index_end, os.SEEK_SET) |
|
752 | f.seek(index_end, os.SEEK_SET) | |
752 | return f |
|
753 | return f | |
753 | except FileNotFoundError: |
|
754 | except FileNotFoundError: | |
754 | if self._delay_buffer is None: |
|
755 | if self._delay_buffer is None: | |
755 | return self.opener( |
|
756 | return self.opener( | |
756 | self.index_file, |
|
757 | self.index_file, | |
757 | mode=b"w+", |
|
758 | mode=b"w+", | |
758 | checkambig=self.data_config.check_ambig, |
|
759 | checkambig=self.data_config.check_ambig, | |
759 | ) |
|
760 | ) | |
760 | else: |
|
761 | else: | |
761 | return randomaccessfile.appender( |
|
762 | return randomaccessfile.appender( | |
762 | self.opener, self.index_file, b"w+", self._delay_buffer |
|
763 | self.opener, self.index_file, b"w+", self._delay_buffer | |
763 | ) |
|
764 | ) | |
764 |
|
765 | |||
765 | def __index_new_fp(self): |
|
766 | def __index_new_fp(self): | |
766 | """internal method to create a new index file for writing |
|
767 | """internal method to create a new index file for writing | |
767 |
|
768 | |||
768 | You should not use this unless you are upgrading from inline revlog |
|
769 | You should not use this unless you are upgrading from inline revlog | |
769 | """ |
|
770 | """ | |
770 | return self.opener( |
|
771 | return self.opener( | |
771 | self.index_file, |
|
772 | self.index_file, | |
772 | mode=b"w", |
|
773 | mode=b"w", | |
773 | checkambig=self.data_config.check_ambig, |
|
774 | checkambig=self.data_config.check_ambig, | |
774 | atomictemp=True, |
|
775 | atomictemp=True, | |
775 | ) |
|
776 | ) | |
776 |
|
777 | |||
777 | def split_inline(self, tr, header, new_index_file_path=None): |
|
778 | def split_inline(self, tr, header, new_index_file_path=None): | |
778 | """split the data of an inline revlog into an index and a data file""" |
|
779 | """split the data of an inline revlog into an index and a data file""" | |
779 | assert self._delay_buffer is None |
|
780 | assert self._delay_buffer is None | |
780 | existing_handles = False |
|
781 | existing_handles = False | |
781 | if self._writinghandles is not None: |
|
782 | if self._writinghandles is not None: | |
782 | existing_handles = True |
|
783 | existing_handles = True | |
783 | fp = self._writinghandles[0] |
|
784 | fp = self._writinghandles[0] | |
784 | fp.flush() |
|
785 | fp.flush() | |
785 | fp.close() |
|
786 | fp.close() | |
786 | # We can't use the cached file handle after close(). So prevent |
|
787 | # We can't use the cached file handle after close(). So prevent | |
787 | # its usage. |
|
788 | # its usage. | |
788 | self._writinghandles = None |
|
789 | self._writinghandles = None | |
789 | self._segmentfile.writing_handle = None |
|
790 | self._segmentfile.writing_handle = None | |
790 | # No need to deal with sidedata writing handle as it is only |
|
791 | # No need to deal with sidedata writing handle as it is only | |
791 | # relevant with revlog-v2 which is never inline, not reaching |
|
792 | # relevant with revlog-v2 which is never inline, not reaching | |
792 | # this code |
|
793 | # this code | |
793 |
|
794 | |||
794 | new_dfh = self.opener(self.data_file, mode=b"w+") |
|
795 | new_dfh = self.opener(self.data_file, mode=b"w+") | |
795 | new_dfh.truncate(0) # drop any potentially existing data |
|
796 | new_dfh.truncate(0) # drop any potentially existing data | |
796 | try: |
|
797 | try: | |
797 | with self.reading(): |
|
798 | with self.reading(): | |
798 | for r in range(len(self.index)): |
|
799 | for r in range(len(self.index)): | |
799 | new_dfh.write(self.get_segment_for_revs(r, r)[1]) |
|
800 | new_dfh.write(self.get_segment_for_revs(r, r)[1]) | |
800 | new_dfh.flush() |
|
801 | new_dfh.flush() | |
801 |
|
802 | |||
802 | if new_index_file_path is not None: |
|
803 | if new_index_file_path is not None: | |
803 | self.index_file = new_index_file_path |
|
804 | self.index_file = new_index_file_path | |
804 | with self.__index_new_fp() as fp: |
|
805 | with self.__index_new_fp() as fp: | |
805 | self.inline = False |
|
806 | self.inline = False | |
806 | for i in range(len(self.index)): |
|
807 | for i in range(len(self.index)): | |
807 | e = self.index.entry_binary(i) |
|
808 | e = self.index.entry_binary(i) | |
808 | if i == 0: |
|
809 | if i == 0: | |
809 | packed_header = self.index.pack_header(header) |
|
810 | packed_header = self.index.pack_header(header) | |
810 | e = packed_header + e |
|
811 | e = packed_header + e | |
811 | fp.write(e) |
|
812 | fp.write(e) | |
812 |
|
813 | |||
813 | # If we don't use side-write, the temp file replace the real |
|
814 | # If we don't use side-write, the temp file replace the real | |
814 | # index when we exit the context manager |
|
815 | # index when we exit the context manager | |
815 |
|
816 | |||
816 | self._segmentfile = randomaccessfile.randomaccessfile( |
|
817 | self._segmentfile = randomaccessfile.randomaccessfile( | |
817 | self.opener, |
|
818 | self.opener, | |
818 | self.data_file, |
|
819 | self.data_file, | |
819 | self.data_config.chunk_cache_size, |
|
820 | self.data_config.chunk_cache_size, | |
820 | ) |
|
821 | ) | |
821 |
|
822 | |||
822 | if existing_handles: |
|
823 | if existing_handles: | |
823 | # switched from inline to conventional reopen the index |
|
824 | # switched from inline to conventional reopen the index | |
824 | ifh = self.__index_write_fp() |
|
825 | ifh = self.__index_write_fp() | |
825 | self._writinghandles = (ifh, new_dfh, None) |
|
826 | self._writinghandles = (ifh, new_dfh, None) | |
826 | self._segmentfile.writing_handle = new_dfh |
|
827 | self._segmentfile.writing_handle = new_dfh | |
827 | new_dfh = None |
|
828 | new_dfh = None | |
828 | # No need to deal with sidedata writing handle as it is only |
|
829 | # No need to deal with sidedata writing handle as it is only | |
829 | # relevant with revlog-v2 which is never inline, not reaching |
|
830 | # relevant with revlog-v2 which is never inline, not reaching | |
830 | # this code |
|
831 | # this code | |
831 | finally: |
|
832 | finally: | |
832 | if new_dfh is not None: |
|
833 | if new_dfh is not None: | |
833 | new_dfh.close() |
|
834 | new_dfh.close() | |
834 | return self.index_file |
|
835 | return self.index_file | |
835 |
|
836 | |||
836 | def get_segment_for_revs(self, startrev, endrev): |
|
837 | def get_segment_for_revs(self, startrev, endrev): | |
837 | """Obtain a segment of raw data corresponding to a range of revisions. |
|
838 | """Obtain a segment of raw data corresponding to a range of revisions. | |
838 |
|
839 | |||
839 | Accepts the start and end revisions and an optional already-open |
|
840 | Accepts the start and end revisions and an optional already-open | |
840 | file handle to be used for reading. If the file handle is read, its |
|
841 | file handle to be used for reading. If the file handle is read, its | |
841 | seek position will not be preserved. |
|
842 | seek position will not be preserved. | |
842 |
|
843 | |||
843 | Requests for data may be satisfied by a cache. |
|
844 | Requests for data may be satisfied by a cache. | |
844 |
|
845 | |||
845 | Returns a 2-tuple of (offset, data) for the requested range of |
|
846 | Returns a 2-tuple of (offset, data) for the requested range of | |
846 | revisions. Offset is the integer offset from the beginning of the |
|
847 | revisions. Offset is the integer offset from the beginning of the | |
847 | revlog and data is a str or buffer of the raw byte data. |
|
848 | revlog and data is a str or buffer of the raw byte data. | |
848 |
|
849 | |||
849 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` |
|
850 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` | |
850 | to determine where each revision's data begins and ends. |
|
851 | to determine where each revision's data begins and ends. | |
851 |
|
852 | |||
852 | API: we should consider making this a private part of the InnerRevlog |
|
853 | API: we should consider making this a private part of the InnerRevlog | |
853 | at some point. |
|
854 | at some point. | |
854 | """ |
|
855 | """ | |
855 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons |
|
856 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons | |
856 | # (functions are expensive). |
|
857 | # (functions are expensive). | |
857 | index = self.index |
|
858 | index = self.index | |
858 | istart = index[startrev] |
|
859 | istart = index[startrev] | |
859 | start = int(istart[0] >> 16) |
|
860 | start = int(istart[0] >> 16) | |
860 | if startrev == endrev: |
|
861 | if startrev == endrev: | |
861 | end = start + istart[1] |
|
862 | end = start + istart[1] | |
862 | else: |
|
863 | else: | |
863 | iend = index[endrev] |
|
864 | iend = index[endrev] | |
864 | end = int(iend[0] >> 16) + iend[1] |
|
865 | end = int(iend[0] >> 16) + iend[1] | |
865 |
|
866 | |||
866 | if self.inline: |
|
867 | if self.inline: | |
867 | start += (startrev + 1) * self.index.entry_size |
|
868 | start += (startrev + 1) * self.index.entry_size | |
868 | end += (endrev + 1) * self.index.entry_size |
|
869 | end += (endrev + 1) * self.index.entry_size | |
869 | length = end - start |
|
870 | length = end - start | |
870 |
|
871 | |||
871 | return start, self._segmentfile.read_chunk(start, length) |
|
872 | return start, self._segmentfile.read_chunk(start, length) | |
872 |
|
873 | |||
873 | def _chunk(self, rev): |
|
874 | def _chunk(self, rev): | |
874 | """Obtain a single decompressed chunk for a revision. |
|
875 | """Obtain a single decompressed chunk for a revision. | |
875 |
|
876 | |||
876 | Accepts an integer revision and an optional already-open file handle |
|
877 | Accepts an integer revision and an optional already-open file handle | |
877 | to be used for reading. If used, the seek position of the file will not |
|
878 | to be used for reading. If used, the seek position of the file will not | |
878 | be preserved. |
|
879 | be preserved. | |
879 |
|
880 | |||
880 | Returns a str holding uncompressed data for the requested revision. |
|
881 | Returns a str holding uncompressed data for the requested revision. | |
881 | """ |
|
882 | """ | |
882 | if self._uncompressed_chunk_cache is not None: |
|
883 | if self._uncompressed_chunk_cache is not None: | |
883 | uncomp = self._uncompressed_chunk_cache.get(rev) |
|
884 | uncomp = self._uncompressed_chunk_cache.get(rev) | |
884 | if uncomp is not None: |
|
885 | if uncomp is not None: | |
885 | return uncomp |
|
886 | return uncomp | |
886 |
|
887 | |||
887 | compression_mode = self.index[rev][10] |
|
888 | compression_mode = self.index[rev][10] | |
888 | data = self.get_segment_for_revs(rev, rev)[1] |
|
889 | data = self.get_segment_for_revs(rev, rev)[1] | |
889 | if compression_mode == COMP_MODE_PLAIN: |
|
890 | if compression_mode == COMP_MODE_PLAIN: | |
890 | uncomp = data |
|
891 | uncomp = data | |
891 | elif compression_mode == COMP_MODE_DEFAULT: |
|
892 | elif compression_mode == COMP_MODE_DEFAULT: | |
892 | uncomp = self._decompressor(data) |
|
893 | uncomp = self._decompressor(data) | |
893 | elif compression_mode == COMP_MODE_INLINE: |
|
894 | elif compression_mode == COMP_MODE_INLINE: | |
894 | uncomp = self.decompress(data) |
|
895 | uncomp = self.decompress(data) | |
895 | else: |
|
896 | else: | |
896 | msg = b'unknown compression mode %d' |
|
897 | msg = b'unknown compression mode %d' | |
897 | msg %= compression_mode |
|
898 | msg %= compression_mode | |
898 | raise error.RevlogError(msg) |
|
899 | raise error.RevlogError(msg) | |
899 | if self._uncompressed_chunk_cache is not None: |
|
900 | if self._uncompressed_chunk_cache is not None: | |
900 | self._uncompressed_chunk_cache.insert(rev, uncomp, cost=len(uncomp)) |
|
901 | self._uncompressed_chunk_cache.insert(rev, uncomp, cost=len(uncomp)) | |
901 | return uncomp |
|
902 | return uncomp | |
902 |
|
903 | |||
903 | def _chunks(self, revs, targetsize=None): |
|
904 | def _chunks(self, revs, targetsize=None): | |
904 | """Obtain decompressed chunks for the specified revisions. |
|
905 | """Obtain decompressed chunks for the specified revisions. | |
905 |
|
906 | |||
906 | Accepts an iterable of numeric revisions that are assumed to be in |
|
907 | Accepts an iterable of numeric revisions that are assumed to be in | |
907 | ascending order. Also accepts an optional already-open file handle |
|
908 | ascending order. Also accepts an optional already-open file handle | |
908 | to be used for reading. If used, the seek position of the file will |
|
909 | to be used for reading. If used, the seek position of the file will | |
909 | not be preserved. |
|
910 | not be preserved. | |
910 |
|
911 | |||
911 | This function is similar to calling ``self._chunk()`` multiple times, |
|
912 | This function is similar to calling ``self._chunk()`` multiple times, | |
912 | but is faster. |
|
913 | but is faster. | |
913 |
|
914 | |||
914 | Returns a list with decompressed data for each requested revision. |
|
915 | Returns a list with decompressed data for each requested revision. | |
915 | """ |
|
916 | """ | |
916 | if not revs: |
|
917 | if not revs: | |
917 | return [] |
|
918 | return [] | |
918 | start = self.start |
|
919 | start = self.start | |
919 | length = self.length |
|
920 | length = self.length | |
920 | inline = self.inline |
|
921 | inline = self.inline | |
921 | iosize = self.index.entry_size |
|
922 | iosize = self.index.entry_size | |
922 | buffer = util.buffer |
|
923 | buffer = util.buffer | |
923 |
|
924 | |||
924 | fetched_revs = [] |
|
925 | fetched_revs = [] | |
925 | fadd = fetched_revs.append |
|
926 | fadd = fetched_revs.append | |
926 |
|
927 | |||
927 | chunks = [] |
|
928 | chunks = [] | |
928 | ladd = chunks.append |
|
929 | ladd = chunks.append | |
929 |
|
930 | |||
930 | if self._uncompressed_chunk_cache is None: |
|
931 | if self._uncompressed_chunk_cache is None: | |
931 | fetched_revs = revs |
|
932 | fetched_revs = revs | |
932 | else: |
|
933 | else: | |
933 | for rev in revs: |
|
934 | for rev in revs: | |
934 | cached_value = self._uncompressed_chunk_cache.get(rev) |
|
935 | cached_value = self._uncompressed_chunk_cache.get(rev) | |
935 | if cached_value is None: |
|
936 | if cached_value is None: | |
936 | fadd(rev) |
|
937 | fadd(rev) | |
937 | else: |
|
938 | else: | |
938 | ladd((rev, cached_value)) |
|
939 | ladd((rev, cached_value)) | |
939 |
|
940 | |||
940 | if not fetched_revs: |
|
941 | if not fetched_revs: | |
941 | slicedchunks = () |
|
942 | slicedchunks = () | |
942 | elif not self.data_config.with_sparse_read: |
|
943 | elif not self.data_config.with_sparse_read: | |
943 | slicedchunks = (fetched_revs,) |
|
944 | slicedchunks = (fetched_revs,) | |
944 | else: |
|
945 | else: | |
945 | slicedchunks = deltautil.slicechunk( |
|
946 | slicedchunks = deltautil.slicechunk( | |
946 | self, |
|
947 | self, | |
947 | fetched_revs, |
|
948 | fetched_revs, | |
948 | targetsize=targetsize, |
|
949 | targetsize=targetsize, | |
949 | ) |
|
950 | ) | |
950 |
|
951 | |||
951 | for revschunk in slicedchunks: |
|
952 | for revschunk in slicedchunks: | |
952 | firstrev = revschunk[0] |
|
953 | firstrev = revschunk[0] | |
953 | # Skip trailing revisions with empty diff |
|
954 | # Skip trailing revisions with empty diff | |
954 | for lastrev in revschunk[::-1]: |
|
955 | for lastrev in revschunk[::-1]: | |
955 | if length(lastrev) != 0: |
|
956 | if length(lastrev) != 0: | |
956 | break |
|
957 | break | |
957 |
|
958 | |||
958 | try: |
|
959 | try: | |
959 | offset, data = self.get_segment_for_revs(firstrev, lastrev) |
|
960 | offset, data = self.get_segment_for_revs(firstrev, lastrev) | |
960 | except OverflowError: |
|
961 | except OverflowError: | |
961 | # issue4215 - we can't cache a run of chunks greater than |
|
962 | # issue4215 - we can't cache a run of chunks greater than | |
962 | # 2G on Windows |
|
963 | # 2G on Windows | |
963 | for rev in revschunk: |
|
964 | for rev in revschunk: | |
964 | ladd((rev, self._chunk(rev))) |
|
965 | ladd((rev, self._chunk(rev))) | |
965 |
|
966 | |||
966 | decomp = self.decompress |
|
967 | decomp = self.decompress | |
967 | # self._decompressor might be None, but will not be used in that case |
|
968 | # self._decompressor might be None, but will not be used in that case | |
968 | def_decomp = self._decompressor |
|
969 | def_decomp = self._decompressor | |
969 | for rev in revschunk: |
|
970 | for rev in revschunk: | |
970 | chunkstart = start(rev) |
|
971 | chunkstart = start(rev) | |
971 | if inline: |
|
972 | if inline: | |
972 | chunkstart += (rev + 1) * iosize |
|
973 | chunkstart += (rev + 1) * iosize | |
973 | chunklength = length(rev) |
|
974 | chunklength = length(rev) | |
974 | comp_mode = self.index[rev][10] |
|
975 | comp_mode = self.index[rev][10] | |
975 | c = buffer(data, chunkstart - offset, chunklength) |
|
976 | c = buffer(data, chunkstart - offset, chunklength) | |
976 | if comp_mode == COMP_MODE_PLAIN: |
|
977 | if comp_mode == COMP_MODE_PLAIN: | |
977 | c = c |
|
978 | c = c | |
978 | elif comp_mode == COMP_MODE_INLINE: |
|
979 | elif comp_mode == COMP_MODE_INLINE: | |
979 | c = decomp(c) |
|
980 | c = decomp(c) | |
980 | elif comp_mode == COMP_MODE_DEFAULT: |
|
981 | elif comp_mode == COMP_MODE_DEFAULT: | |
981 | c = def_decomp(c) |
|
982 | c = def_decomp(c) | |
982 | else: |
|
983 | else: | |
983 | msg = b'unknown compression mode %d' |
|
984 | msg = b'unknown compression mode %d' | |
984 | msg %= comp_mode |
|
985 | msg %= comp_mode | |
985 | raise error.RevlogError(msg) |
|
986 | raise error.RevlogError(msg) | |
986 | ladd((rev, c)) |
|
987 | ladd((rev, c)) | |
987 | if self._uncompressed_chunk_cache is not None: |
|
988 | if self._uncompressed_chunk_cache is not None: | |
988 | self._uncompressed_chunk_cache.insert(rev, c, len(c)) |
|
989 | self._uncompressed_chunk_cache.insert(rev, c, len(c)) | |
989 |
|
990 | |||
990 | chunks.sort() |
|
991 | chunks.sort() | |
991 | return [x[1] for x in chunks] |
|
992 | return [x[1] for x in chunks] | |
992 |
|
993 | |||
993 | def raw_text(self, node, rev): |
|
994 | def raw_text(self, node, rev): | |
994 | """return the possibly unvalidated rawtext for a revision |
|
995 | """return the possibly unvalidated rawtext for a revision | |
995 |
|
996 | |||
996 | returns (rev, rawtext, validated) |
|
997 | returns (rev, rawtext, validated) | |
997 | """ |
|
998 | """ | |
998 |
|
999 | |||
999 | # revision in the cache (could be useful to apply delta) |
|
1000 | # revision in the cache (could be useful to apply delta) | |
1000 | cachedrev = None |
|
1001 | cachedrev = None | |
1001 | # An intermediate text to apply deltas to |
|
1002 | # An intermediate text to apply deltas to | |
1002 | basetext = None |
|
1003 | basetext = None | |
1003 |
|
1004 | |||
1004 | # Check if we have the entry in cache |
|
1005 | # Check if we have the entry in cache | |
1005 | # The cache entry looks like (node, rev, rawtext) |
|
1006 | # The cache entry looks like (node, rev, rawtext) | |
1006 | if self._revisioncache: |
|
1007 | if self._revisioncache: | |
1007 | cachedrev = self._revisioncache[1] |
|
1008 | cachedrev = self._revisioncache[1] | |
1008 |
|
1009 | |||
1009 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) |
|
1010 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) | |
1010 | if stopped: |
|
1011 | if stopped: | |
1011 | basetext = self._revisioncache[2] |
|
1012 | basetext = self._revisioncache[2] | |
1012 |
|
1013 | |||
1013 | # drop cache to save memory, the caller is expected to |
|
1014 | # drop cache to save memory, the caller is expected to | |
1014 | # update self._inner._revisioncache after validating the text |
|
1015 | # update self._inner._revisioncache after validating the text | |
1015 | self._revisioncache = None |
|
1016 | self._revisioncache = None | |
1016 |
|
1017 | |||
1017 | targetsize = None |
|
1018 | targetsize = None | |
1018 | rawsize = self.index[rev][2] |
|
1019 | rawsize = self.index[rev][2] | |
1019 | if 0 <= rawsize: |
|
1020 | if 0 <= rawsize: | |
1020 | targetsize = 4 * rawsize |
|
1021 | targetsize = 4 * rawsize | |
1021 |
|
1022 | |||
1022 | if self._uncompressed_chunk_cache is not None: |
|
1023 | if self._uncompressed_chunk_cache is not None: | |
1023 | # dynamically update the uncompressed_chunk_cache size to the |
|
1024 | # dynamically update the uncompressed_chunk_cache size to the | |
1024 | # largest revision we saw in this revlog. |
|
1025 | # largest revision we saw in this revlog. | |
1025 | factor = self.data_config.uncompressed_cache_factor |
|
1026 | factor = self.data_config.uncompressed_cache_factor | |
1026 | candidate_size = rawsize * factor |
|
1027 | candidate_size = rawsize * factor | |
1027 | if candidate_size > self._uncompressed_chunk_cache.maxcost: |
|
1028 | if candidate_size > self._uncompressed_chunk_cache.maxcost: | |
1028 | self._uncompressed_chunk_cache.maxcost = candidate_size |
|
1029 | self._uncompressed_chunk_cache.maxcost = candidate_size | |
1029 |
|
1030 | |||
1030 | bins = self._chunks(chain, targetsize=targetsize) |
|
1031 | bins = self._chunks(chain, targetsize=targetsize) | |
1031 | if basetext is None: |
|
1032 | if basetext is None: | |
1032 | basetext = bytes(bins[0]) |
|
1033 | basetext = bytes(bins[0]) | |
1033 | bins = bins[1:] |
|
1034 | bins = bins[1:] | |
1034 |
|
1035 | |||
1035 | rawtext = mdiff.patches(basetext, bins) |
|
1036 | rawtext = mdiff.patches(basetext, bins) | |
1036 | del basetext # let us have a chance to free memory early |
|
1037 | del basetext # let us have a chance to free memory early | |
1037 | return (rev, rawtext, False) |
|
1038 | return (rev, rawtext, False) | |
1038 |
|
1039 | |||
1039 | def sidedata(self, rev, sidedata_end): |
|
1040 | def sidedata(self, rev, sidedata_end): | |
1040 | """Return the sidedata for a given revision number.""" |
|
1041 | """Return the sidedata for a given revision number.""" | |
1041 | index_entry = self.index[rev] |
|
1042 | index_entry = self.index[rev] | |
1042 | sidedata_offset = index_entry[8] |
|
1043 | sidedata_offset = index_entry[8] | |
1043 | sidedata_size = index_entry[9] |
|
1044 | sidedata_size = index_entry[9] | |
1044 |
|
1045 | |||
1045 | if self.inline: |
|
1046 | if self.inline: | |
1046 | sidedata_offset += self.index.entry_size * (1 + rev) |
|
1047 | sidedata_offset += self.index.entry_size * (1 + rev) | |
1047 | if sidedata_size == 0: |
|
1048 | if sidedata_size == 0: | |
1048 | return {} |
|
1049 | return {} | |
1049 |
|
1050 | |||
1050 | if sidedata_end < sidedata_offset + sidedata_size: |
|
1051 | if sidedata_end < sidedata_offset + sidedata_size: | |
1051 | filename = self.sidedata_file |
|
1052 | filename = self.sidedata_file | |
1052 | end = sidedata_end |
|
1053 | end = sidedata_end | |
1053 | offset = sidedata_offset |
|
1054 | offset = sidedata_offset | |
1054 | length = sidedata_size |
|
1055 | length = sidedata_size | |
1055 | m = FILE_TOO_SHORT_MSG % (filename, length, offset, end) |
|
1056 | m = FILE_TOO_SHORT_MSG % (filename, length, offset, end) | |
1056 | raise error.RevlogError(m) |
|
1057 | raise error.RevlogError(m) | |
1057 |
|
1058 | |||
1058 | comp_segment = self._segmentfile_sidedata.read_chunk( |
|
1059 | comp_segment = self._segmentfile_sidedata.read_chunk( | |
1059 | sidedata_offset, sidedata_size |
|
1060 | sidedata_offset, sidedata_size | |
1060 | ) |
|
1061 | ) | |
1061 |
|
1062 | |||
1062 | comp = self.index[rev][11] |
|
1063 | comp = self.index[rev][11] | |
1063 | if comp == COMP_MODE_PLAIN: |
|
1064 | if comp == COMP_MODE_PLAIN: | |
1064 | segment = comp_segment |
|
1065 | segment = comp_segment | |
1065 | elif comp == COMP_MODE_DEFAULT: |
|
1066 | elif comp == COMP_MODE_DEFAULT: | |
1066 | segment = self._decompressor(comp_segment) |
|
1067 | segment = self._decompressor(comp_segment) | |
1067 | elif comp == COMP_MODE_INLINE: |
|
1068 | elif comp == COMP_MODE_INLINE: | |
1068 | segment = self.decompress(comp_segment) |
|
1069 | segment = self.decompress(comp_segment) | |
1069 | else: |
|
1070 | else: | |
1070 | msg = b'unknown compression mode %d' |
|
1071 | msg = b'unknown compression mode %d' | |
1071 | msg %= comp |
|
1072 | msg %= comp | |
1072 | raise error.RevlogError(msg) |
|
1073 | raise error.RevlogError(msg) | |
1073 |
|
1074 | |||
1074 | sidedata = sidedatautil.deserialize_sidedata(segment) |
|
1075 | sidedata = sidedatautil.deserialize_sidedata(segment) | |
1075 | return sidedata |
|
1076 | return sidedata | |
1076 |
|
1077 | |||
1077 | def write_entry( |
|
1078 | def write_entry( | |
1078 | self, |
|
1079 | self, | |
1079 | transaction, |
|
1080 | transaction, | |
1080 | entry, |
|
1081 | entry, | |
1081 | data, |
|
1082 | data, | |
1082 | link, |
|
1083 | link, | |
1083 | offset, |
|
1084 | offset, | |
1084 | sidedata, |
|
1085 | sidedata, | |
1085 | sidedata_offset, |
|
1086 | sidedata_offset, | |
1086 | index_end, |
|
1087 | index_end, | |
1087 | data_end, |
|
1088 | data_end, | |
1088 | sidedata_end, |
|
1089 | sidedata_end, | |
1089 | ): |
|
1090 | ): | |
1090 | # Files opened in a+ mode have inconsistent behavior on various |
|
1091 | # Files opened in a+ mode have inconsistent behavior on various | |
1091 | # platforms. Windows requires that a file positioning call be made |
|
1092 | # platforms. Windows requires that a file positioning call be made | |
1092 | # when the file handle transitions between reads and writes. See |
|
1093 | # when the file handle transitions between reads and writes. See | |
1093 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
1094 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other | |
1094 | # platforms, Python or the platform itself can be buggy. Some versions |
|
1095 | # platforms, Python or the platform itself can be buggy. Some versions | |
1095 | # of Solaris have been observed to not append at the end of the file |
|
1096 | # of Solaris have been observed to not append at the end of the file | |
1096 | # if the file was seeked to before the end. See issue4943 for more. |
|
1097 | # if the file was seeked to before the end. See issue4943 for more. | |
1097 | # |
|
1098 | # | |
1098 | # We work around this issue by inserting a seek() before writing. |
|
1099 | # We work around this issue by inserting a seek() before writing. | |
1099 | # Note: This is likely not necessary on Python 3. However, because |
|
1100 | # Note: This is likely not necessary on Python 3. However, because | |
1100 | # the file handle is reused for reads and may be seeked there, we need |
|
1101 | # the file handle is reused for reads and may be seeked there, we need | |
1101 | # to be careful before changing this. |
|
1102 | # to be careful before changing this. | |
1102 | if self._writinghandles is None: |
|
1103 | if self._writinghandles is None: | |
1103 | msg = b'adding revision outside `revlog._writing` context' |
|
1104 | msg = b'adding revision outside `revlog._writing` context' | |
1104 | raise error.ProgrammingError(msg) |
|
1105 | raise error.ProgrammingError(msg) | |
1105 | ifh, dfh, sdfh = self._writinghandles |
|
1106 | ifh, dfh, sdfh = self._writinghandles | |
1106 | if index_end is None: |
|
1107 | if index_end is None: | |
1107 | ifh.seek(0, os.SEEK_END) |
|
1108 | ifh.seek(0, os.SEEK_END) | |
1108 | else: |
|
1109 | else: | |
1109 | ifh.seek(index_end, os.SEEK_SET) |
|
1110 | ifh.seek(index_end, os.SEEK_SET) | |
1110 | if dfh: |
|
1111 | if dfh: | |
1111 | if data_end is None: |
|
1112 | if data_end is None: | |
1112 | dfh.seek(0, os.SEEK_END) |
|
1113 | dfh.seek(0, os.SEEK_END) | |
1113 | else: |
|
1114 | else: | |
1114 | dfh.seek(data_end, os.SEEK_SET) |
|
1115 | dfh.seek(data_end, os.SEEK_SET) | |
1115 | if sdfh: |
|
1116 | if sdfh: | |
1116 | sdfh.seek(sidedata_end, os.SEEK_SET) |
|
1117 | sdfh.seek(sidedata_end, os.SEEK_SET) | |
1117 |
|
1118 | |||
1118 | curr = len(self.index) - 1 |
|
1119 | curr = len(self.index) - 1 | |
1119 | if not self.inline: |
|
1120 | if not self.inline: | |
1120 | transaction.add(self.data_file, offset) |
|
1121 | transaction.add(self.data_file, offset) | |
1121 | if self.sidedata_file: |
|
1122 | if self.sidedata_file: | |
1122 | transaction.add(self.sidedata_file, sidedata_offset) |
|
1123 | transaction.add(self.sidedata_file, sidedata_offset) | |
1123 | transaction.add(self.canonical_index_file, curr * len(entry)) |
|
1124 | transaction.add(self.canonical_index_file, curr * len(entry)) | |
1124 | if data[0]: |
|
1125 | if data[0]: | |
1125 | dfh.write(data[0]) |
|
1126 | dfh.write(data[0]) | |
1126 | dfh.write(data[1]) |
|
1127 | dfh.write(data[1]) | |
1127 | if sidedata: |
|
1128 | if sidedata: | |
1128 | sdfh.write(sidedata) |
|
1129 | sdfh.write(sidedata) | |
1129 | if self._delay_buffer is None: |
|
1130 | if self._delay_buffer is None: | |
1130 | ifh.write(entry) |
|
1131 | ifh.write(entry) | |
1131 | else: |
|
1132 | else: | |
1132 | self._delay_buffer.append(entry) |
|
1133 | self._delay_buffer.append(entry) | |
1133 | elif self._delay_buffer is not None: |
|
1134 | elif self._delay_buffer is not None: | |
1134 | msg = b'invalid delayed write on inline revlog' |
|
1135 | msg = b'invalid delayed write on inline revlog' | |
1135 | raise error.ProgrammingError(msg) |
|
1136 | raise error.ProgrammingError(msg) | |
1136 | else: |
|
1137 | else: | |
1137 | offset += curr * self.index.entry_size |
|
1138 | offset += curr * self.index.entry_size | |
1138 | transaction.add(self.canonical_index_file, offset) |
|
1139 | transaction.add(self.canonical_index_file, offset) | |
1139 | assert not sidedata |
|
1140 | assert not sidedata | |
1140 | ifh.write(entry) |
|
1141 | ifh.write(entry) | |
1141 | ifh.write(data[0]) |
|
1142 | ifh.write(data[0]) | |
1142 | ifh.write(data[1]) |
|
1143 | ifh.write(data[1]) | |
1143 | return ( |
|
1144 | return ( | |
1144 | ifh.tell(), |
|
1145 | ifh.tell(), | |
1145 | dfh.tell() if dfh else None, |
|
1146 | dfh.tell() if dfh else None, | |
1146 | sdfh.tell() if sdfh else None, |
|
1147 | sdfh.tell() if sdfh else None, | |
1147 | ) |
|
1148 | ) | |
1148 |
|
1149 | |||
1149 | def _divert_index(self): |
|
1150 | def _divert_index(self): | |
1150 | return self.index_file + b'.a' |
|
1151 | return self.index_file + b'.a' | |
1151 |
|
1152 | |||
1152 | def delay(self): |
|
1153 | def delay(self): | |
1153 | assert not self.is_open |
|
1154 | assert not self.is_open | |
1154 | if self.inline: |
|
1155 | if self.inline: | |
1155 | msg = "revlog with delayed write should not be inline" |
|
1156 | msg = "revlog with delayed write should not be inline" | |
1156 | raise error.ProgrammingError(msg) |
|
1157 | raise error.ProgrammingError(msg) | |
1157 | if self._delay_buffer is not None or self._orig_index_file is not None: |
|
1158 | if self._delay_buffer is not None or self._orig_index_file is not None: | |
1158 | # delay or divert already in place |
|
1159 | # delay or divert already in place | |
1159 | return None |
|
1160 | return None | |
1160 | elif len(self.index) == 0: |
|
1161 | elif len(self.index) == 0: | |
1161 | self._orig_index_file = self.index_file |
|
1162 | self._orig_index_file = self.index_file | |
1162 | self.index_file = self._divert_index() |
|
1163 | self.index_file = self._divert_index() | |
1163 | assert self._orig_index_file is not None |
|
1164 | assert self._orig_index_file is not None | |
1164 | assert self.index_file is not None |
|
1165 | assert self.index_file is not None | |
1165 | if self.opener.exists(self.index_file): |
|
1166 | if self.opener.exists(self.index_file): | |
1166 | self.opener.unlink(self.index_file) |
|
1167 | self.opener.unlink(self.index_file) | |
1167 | return self.index_file |
|
1168 | return self.index_file | |
1168 | else: |
|
1169 | else: | |
1169 | self._delay_buffer = [] |
|
1170 | self._delay_buffer = [] | |
1170 | return None |
|
1171 | return None | |
1171 |
|
1172 | |||
1172 | def write_pending(self): |
|
1173 | def write_pending(self): | |
1173 | assert not self.is_open |
|
1174 | assert not self.is_open | |
1174 | if self.inline: |
|
1175 | if self.inline: | |
1175 | msg = "revlog with delayed write should not be inline" |
|
1176 | msg = "revlog with delayed write should not be inline" | |
1176 | raise error.ProgrammingError(msg) |
|
1177 | raise error.ProgrammingError(msg) | |
1177 | if self._orig_index_file is not None: |
|
1178 | if self._orig_index_file is not None: | |
1178 | return None, True |
|
1179 | return None, True | |
1179 | any_pending = False |
|
1180 | any_pending = False | |
1180 | pending_index_file = self._divert_index() |
|
1181 | pending_index_file = self._divert_index() | |
1181 | if self.opener.exists(pending_index_file): |
|
1182 | if self.opener.exists(pending_index_file): | |
1182 | self.opener.unlink(pending_index_file) |
|
1183 | self.opener.unlink(pending_index_file) | |
1183 | util.copyfile( |
|
1184 | util.copyfile( | |
1184 | self.opener.join(self.index_file), |
|
1185 | self.opener.join(self.index_file), | |
1185 | self.opener.join(pending_index_file), |
|
1186 | self.opener.join(pending_index_file), | |
1186 | ) |
|
1187 | ) | |
1187 | if self._delay_buffer: |
|
1188 | if self._delay_buffer: | |
1188 | with self.opener(pending_index_file, b'r+') as ifh: |
|
1189 | with self.opener(pending_index_file, b'r+') as ifh: | |
1189 | ifh.seek(0, os.SEEK_END) |
|
1190 | ifh.seek(0, os.SEEK_END) | |
1190 | ifh.write(b"".join(self._delay_buffer)) |
|
1191 | ifh.write(b"".join(self._delay_buffer)) | |
1191 | any_pending = True |
|
1192 | any_pending = True | |
1192 | self._delay_buffer = None |
|
1193 | self._delay_buffer = None | |
1193 | self._orig_index_file = self.index_file |
|
1194 | self._orig_index_file = self.index_file | |
1194 | self.index_file = pending_index_file |
|
1195 | self.index_file = pending_index_file | |
1195 | return self.index_file, any_pending |
|
1196 | return self.index_file, any_pending | |
1196 |
|
1197 | |||
1197 | def finalize_pending(self): |
|
1198 | def finalize_pending(self): | |
1198 | assert not self.is_open |
|
1199 | assert not self.is_open | |
1199 | if self.inline: |
|
1200 | if self.inline: | |
1200 | msg = "revlog with delayed write should not be inline" |
|
1201 | msg = "revlog with delayed write should not be inline" | |
1201 | raise error.ProgrammingError(msg) |
|
1202 | raise error.ProgrammingError(msg) | |
1202 |
|
1203 | |||
1203 | delay = self._delay_buffer is not None |
|
1204 | delay = self._delay_buffer is not None | |
1204 | divert = self._orig_index_file is not None |
|
1205 | divert = self._orig_index_file is not None | |
1205 |
|
1206 | |||
1206 | if delay and divert: |
|
1207 | if delay and divert: | |
1207 | assert False, "unreachable" |
|
1208 | assert False, "unreachable" | |
1208 | elif delay: |
|
1209 | elif delay: | |
1209 | if self._delay_buffer: |
|
1210 | if self._delay_buffer: | |
1210 | with self.opener(self.index_file, b'r+') as ifh: |
|
1211 | with self.opener(self.index_file, b'r+') as ifh: | |
1211 | ifh.seek(0, os.SEEK_END) |
|
1212 | ifh.seek(0, os.SEEK_END) | |
1212 | ifh.write(b"".join(self._delay_buffer)) |
|
1213 | ifh.write(b"".join(self._delay_buffer)) | |
1213 | self._delay_buffer = None |
|
1214 | self._delay_buffer = None | |
1214 | elif divert: |
|
1215 | elif divert: | |
1215 | if self.opener.exists(self.index_file): |
|
1216 | if self.opener.exists(self.index_file): | |
1216 | self.opener.rename( |
|
1217 | self.opener.rename( | |
1217 | self.index_file, |
|
1218 | self.index_file, | |
1218 | self._orig_index_file, |
|
1219 | self._orig_index_file, | |
1219 | checkambig=True, |
|
1220 | checkambig=True, | |
1220 | ) |
|
1221 | ) | |
1221 | self.index_file = self._orig_index_file |
|
1222 | self.index_file = self._orig_index_file | |
1222 | self._orig_index_file = None |
|
1223 | self._orig_index_file = None | |
1223 | else: |
|
1224 | else: | |
1224 | msg = b"not delay or divert found on this revlog" |
|
1225 | msg = b"not delay or divert found on this revlog" | |
1225 | raise error.ProgrammingError(msg) |
|
1226 | raise error.ProgrammingError(msg) | |
1226 | return self.canonical_index_file |
|
1227 | return self.canonical_index_file | |
1227 |
|
1228 | |||
1228 |
|
1229 | |||
1229 | class revlog: |
|
1230 | class revlog: | |
1230 | """ |
|
1231 | """ | |
1231 | the underlying revision storage object |
|
1232 | the underlying revision storage object | |
1232 |
|
1233 | |||
1233 | A revlog consists of two parts, an index and the revision data. |
|
1234 | A revlog consists of two parts, an index and the revision data. | |
1234 |
|
1235 | |||
1235 | The index is a file with a fixed record size containing |
|
1236 | The index is a file with a fixed record size containing | |
1236 | information on each revision, including its nodeid (hash), the |
|
1237 | information on each revision, including its nodeid (hash), the | |
1237 | nodeids of its parents, the position and offset of its data within |
|
1238 | nodeids of its parents, the position and offset of its data within | |
1238 | the data file, and the revision it's based on. Finally, each entry |
|
1239 | the data file, and the revision it's based on. Finally, each entry | |
1239 | contains a linkrev entry that can serve as a pointer to external |
|
1240 | contains a linkrev entry that can serve as a pointer to external | |
1240 | data. |
|
1241 | data. | |
1241 |
|
1242 | |||
1242 | The revision data itself is a linear collection of data chunks. |
|
1243 | The revision data itself is a linear collection of data chunks. | |
1243 | Each chunk represents a revision and is usually represented as a |
|
1244 | Each chunk represents a revision and is usually represented as a | |
1244 | delta against the previous chunk. To bound lookup time, runs of |
|
1245 | delta against the previous chunk. To bound lookup time, runs of | |
1245 | deltas are limited to about 2 times the length of the original |
|
1246 | deltas are limited to about 2 times the length of the original | |
1246 | version data. This makes retrieval of a version proportional to |
|
1247 | version data. This makes retrieval of a version proportional to | |
1247 | its size, or O(1) relative to the number of revisions. |
|
1248 | its size, or O(1) relative to the number of revisions. | |
1248 |
|
1249 | |||
1249 | Both pieces of the revlog are written to in an append-only |
|
1250 | Both pieces of the revlog are written to in an append-only | |
1250 | fashion, which means we never need to rewrite a file to insert or |
|
1251 | fashion, which means we never need to rewrite a file to insert or | |
1251 | remove data, and can use some simple techniques to avoid the need |
|
1252 | remove data, and can use some simple techniques to avoid the need | |
1252 | for locking while reading. |
|
1253 | for locking while reading. | |
1253 |
|
1254 | |||
1254 | If checkambig, indexfile is opened with checkambig=True at |
|
1255 | If checkambig, indexfile is opened with checkambig=True at | |
1255 | writing, to avoid file stat ambiguity. |
|
1256 | writing, to avoid file stat ambiguity. | |
1256 |
|
1257 | |||
1257 | If mmaplargeindex is True, and an mmapindexthreshold is set, the |
|
1258 | If mmaplargeindex is True, and an mmapindexthreshold is set, the | |
1258 | index will be mmapped rather than read if it is larger than the |
|
1259 | index will be mmapped rather than read if it is larger than the | |
1259 | configured threshold. |
|
1260 | configured threshold. | |
1260 |
|
1261 | |||
1261 | If censorable is True, the revlog can have censored revisions. |
|
1262 | If censorable is True, the revlog can have censored revisions. | |
1262 |
|
1263 | |||
1263 | If `upperboundcomp` is not None, this is the expected maximal gain from |
|
1264 | If `upperboundcomp` is not None, this is the expected maximal gain from | |
1264 | compression for the data content. |
|
1265 | compression for the data content. | |
1265 |
|
1266 | |||
1266 | `concurrencychecker` is an optional function that receives 3 arguments: a |
|
1267 | `concurrencychecker` is an optional function that receives 3 arguments: a | |
1267 | file handle, a filename, and an expected position. It should check whether |
|
1268 | file handle, a filename, and an expected position. It should check whether | |
1268 | the current position in the file handle is valid, and log/warn/fail (by |
|
1269 | the current position in the file handle is valid, and log/warn/fail (by | |
1269 | raising). |
|
1270 | raising). | |
1270 |
|
1271 | |||
1271 | See mercurial/revlogutils/contants.py for details about the content of an |
|
1272 | See mercurial/revlogutils/contants.py for details about the content of an | |
1272 | index entry. |
|
1273 | index entry. | |
1273 | """ |
|
1274 | """ | |
1274 |
|
1275 | |||
1275 | _flagserrorclass = error.RevlogError |
|
1276 | _flagserrorclass = error.RevlogError | |
1276 |
|
1277 | |||
1277 | @staticmethod |
|
1278 | @staticmethod | |
1278 | def is_inline_index(header_bytes): |
|
1279 | def is_inline_index(header_bytes): | |
1279 | """Determine if a revlog is inline from the initial bytes of the index""" |
|
1280 | """Determine if a revlog is inline from the initial bytes of the index""" | |
1280 | if len(header_bytes) == 0: |
|
1281 | if len(header_bytes) == 0: | |
1281 | return True |
|
1282 | return True | |
1282 |
|
1283 | |||
1283 | header = INDEX_HEADER.unpack(header_bytes)[0] |
|
1284 | header = INDEX_HEADER.unpack(header_bytes)[0] | |
1284 |
|
1285 | |||
1285 | _format_flags = header & ~0xFFFF |
|
1286 | _format_flags = header & ~0xFFFF | |
1286 | _format_version = header & 0xFFFF |
|
1287 | _format_version = header & 0xFFFF | |
1287 |
|
1288 | |||
1288 | features = FEATURES_BY_VERSION[_format_version] |
|
1289 | features = FEATURES_BY_VERSION[_format_version] | |
1289 | return features[b'inline'](_format_flags) |
|
1290 | return features[b'inline'](_format_flags) | |
1290 |
|
1291 | |||
1291 | def __init__( |
|
1292 | def __init__( | |
1292 | self, |
|
1293 | self, | |
1293 | opener, |
|
1294 | opener, | |
1294 | target, |
|
1295 | target, | |
1295 | radix, |
|
1296 | radix, | |
1296 | postfix=None, # only exist for `tmpcensored` now |
|
1297 | postfix=None, # only exist for `tmpcensored` now | |
1297 | checkambig=False, |
|
1298 | checkambig=False, | |
1298 | mmaplargeindex=False, |
|
1299 | mmaplargeindex=False, | |
1299 | censorable=False, |
|
1300 | censorable=False, | |
1300 | upperboundcomp=None, |
|
1301 | upperboundcomp=None, | |
1301 | persistentnodemap=False, |
|
1302 | persistentnodemap=False, | |
1302 | concurrencychecker=None, |
|
1303 | concurrencychecker=None, | |
1303 | trypending=False, |
|
1304 | trypending=False, | |
1304 | try_split=False, |
|
1305 | try_split=False, | |
1305 | canonical_parent_order=True, |
|
1306 | canonical_parent_order=True, | |
1306 | data_config=None, |
|
1307 | data_config=None, | |
1307 | delta_config=None, |
|
1308 | delta_config=None, | |
1308 | feature_config=None, |
|
1309 | feature_config=None, | |
1309 | may_inline=True, # may inline new revlog |
|
1310 | may_inline=True, # may inline new revlog | |
1310 | ): |
|
1311 | ): | |
1311 | """ |
|
1312 | """ | |
1312 | create a revlog object |
|
1313 | create a revlog object | |
1313 |
|
1314 | |||
1314 | opener is a function that abstracts the file opening operation |
|
1315 | opener is a function that abstracts the file opening operation | |
1315 | and can be used to implement COW semantics or the like. |
|
1316 | and can be used to implement COW semantics or the like. | |
1316 |
|
1317 | |||
1317 | `target`: a (KIND, ID) tuple that identify the content stored in |
|
1318 | `target`: a (KIND, ID) tuple that identify the content stored in | |
1318 | this revlog. It help the rest of the code to understand what the revlog |
|
1319 | this revlog. It help the rest of the code to understand what the revlog | |
1319 | is about without having to resort to heuristic and index filename |
|
1320 | is about without having to resort to heuristic and index filename | |
1320 | analysis. Note: that this must be reliably be set by normal code, but |
|
1321 | analysis. Note: that this must be reliably be set by normal code, but | |
1321 | that test, debug, or performance measurement code might not set this to |
|
1322 | that test, debug, or performance measurement code might not set this to | |
1322 | accurate value. |
|
1323 | accurate value. | |
1323 | """ |
|
1324 | """ | |
1324 |
|
1325 | |||
1325 | self.radix = radix |
|
1326 | self.radix = radix | |
1326 |
|
1327 | |||
1327 | self._docket_file = None |
|
1328 | self._docket_file = None | |
1328 | self._indexfile = None |
|
1329 | self._indexfile = None | |
1329 | self._datafile = None |
|
1330 | self._datafile = None | |
1330 | self._sidedatafile = None |
|
1331 | self._sidedatafile = None | |
1331 | self._nodemap_file = None |
|
1332 | self._nodemap_file = None | |
1332 | self.postfix = postfix |
|
1333 | self.postfix = postfix | |
1333 | self._trypending = trypending |
|
1334 | self._trypending = trypending | |
1334 | self._try_split = try_split |
|
1335 | self._try_split = try_split | |
1335 | self._may_inline = may_inline |
|
1336 | self._may_inline = may_inline | |
1336 | self.opener = opener |
|
1337 | self.opener = opener | |
1337 | if persistentnodemap: |
|
1338 | if persistentnodemap: | |
1338 | self._nodemap_file = nodemaputil.get_nodemap_file(self) |
|
1339 | self._nodemap_file = nodemaputil.get_nodemap_file(self) | |
1339 |
|
1340 | |||
1340 | assert target[0] in ALL_KINDS |
|
1341 | assert target[0] in ALL_KINDS | |
1341 | assert len(target) == 2 |
|
1342 | assert len(target) == 2 | |
1342 | self.target = target |
|
1343 | self.target = target | |
1343 | if feature_config is not None: |
|
1344 | if feature_config is not None: | |
1344 | self.feature_config = feature_config.copy() |
|
1345 | self.feature_config = feature_config.copy() | |
1345 | elif b'feature-config' in self.opener.options: |
|
1346 | elif b'feature-config' in self.opener.options: | |
1346 | self.feature_config = self.opener.options[b'feature-config'].copy() |
|
1347 | self.feature_config = self.opener.options[b'feature-config'].copy() | |
1347 | else: |
|
1348 | else: | |
1348 | self.feature_config = FeatureConfig() |
|
1349 | self.feature_config = FeatureConfig() | |
1349 | self.feature_config.censorable = censorable |
|
1350 | self.feature_config.censorable = censorable | |
1350 | self.feature_config.canonical_parent_order = canonical_parent_order |
|
1351 | self.feature_config.canonical_parent_order = canonical_parent_order | |
1351 | if data_config is not None: |
|
1352 | if data_config is not None: | |
1352 | self.data_config = data_config.copy() |
|
1353 | self.data_config = data_config.copy() | |
1353 | elif b'data-config' in self.opener.options: |
|
1354 | elif b'data-config' in self.opener.options: | |
1354 | self.data_config = self.opener.options[b'data-config'].copy() |
|
1355 | self.data_config = self.opener.options[b'data-config'].copy() | |
1355 | else: |
|
1356 | else: | |
1356 | self.data_config = DataConfig() |
|
1357 | self.data_config = DataConfig() | |
1357 | self.data_config.check_ambig = checkambig |
|
1358 | self.data_config.check_ambig = checkambig | |
1358 | self.data_config.mmap_large_index = mmaplargeindex |
|
1359 | self.data_config.mmap_large_index = mmaplargeindex | |
1359 | if delta_config is not None: |
|
1360 | if delta_config is not None: | |
1360 | self.delta_config = delta_config.copy() |
|
1361 | self.delta_config = delta_config.copy() | |
1361 | elif b'delta-config' in self.opener.options: |
|
1362 | elif b'delta-config' in self.opener.options: | |
1362 | self.delta_config = self.opener.options[b'delta-config'].copy() |
|
1363 | self.delta_config = self.opener.options[b'delta-config'].copy() | |
1363 | else: |
|
1364 | else: | |
1364 | self.delta_config = DeltaConfig() |
|
1365 | self.delta_config = DeltaConfig() | |
1365 | self.delta_config.upper_bound_comp = upperboundcomp |
|
1366 | self.delta_config.upper_bound_comp = upperboundcomp | |
1366 |
|
1367 | |||
1367 | # Maps rev to chain base rev. |
|
1368 | # Maps rev to chain base rev. | |
1368 | self._chainbasecache = util.lrucachedict(100) |
|
1369 | self._chainbasecache = util.lrucachedict(100) | |
1369 |
|
1370 | |||
1370 | self.index = None |
|
1371 | self.index = None | |
1371 | self._docket = None |
|
1372 | self._docket = None | |
1372 | self._nodemap_docket = None |
|
1373 | self._nodemap_docket = None | |
1373 | # Mapping of partial identifiers to full nodes. |
|
1374 | # Mapping of partial identifiers to full nodes. | |
1374 | self._pcache = {} |
|
1375 | self._pcache = {} | |
1375 |
|
1376 | |||
1376 | # other optionnals features |
|
1377 | # other optionnals features | |
1377 |
|
1378 | |||
1378 | # Make copy of flag processors so each revlog instance can support |
|
1379 | # Make copy of flag processors so each revlog instance can support | |
1379 | # custom flags. |
|
1380 | # custom flags. | |
1380 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
1381 | self._flagprocessors = dict(flagutil.flagprocessors) | |
1381 | # prevent nesting of addgroup |
|
1382 | # prevent nesting of addgroup | |
1382 | self._adding_group = None |
|
1383 | self._adding_group = None | |
1383 |
|
1384 | |||
1384 | chunk_cache = self._loadindex() |
|
1385 | chunk_cache = self._loadindex() | |
1385 | self._load_inner(chunk_cache) |
|
1386 | self._load_inner(chunk_cache) | |
1386 | self._concurrencychecker = concurrencychecker |
|
1387 | self._concurrencychecker = concurrencychecker | |
1387 |
|
1388 | |||
1388 | def _init_opts(self): |
|
1389 | def _init_opts(self): | |
1389 | """process options (from above/config) to setup associated default revlog mode |
|
1390 | """process options (from above/config) to setup associated default revlog mode | |
1390 |
|
1391 | |||
1391 | These values might be affected when actually reading on disk information. |
|
1392 | These values might be affected when actually reading on disk information. | |
1392 |
|
1393 | |||
1393 | The relevant values are returned for use in _loadindex(). |
|
1394 | The relevant values are returned for use in _loadindex(). | |
1394 |
|
1395 | |||
1395 | * newversionflags: |
|
1396 | * newversionflags: | |
1396 | version header to use if we need to create a new revlog |
|
1397 | version header to use if we need to create a new revlog | |
1397 |
|
1398 | |||
1398 | * mmapindexthreshold: |
|
1399 | * mmapindexthreshold: | |
1399 | minimal index size for start to use mmap |
|
1400 | minimal index size for start to use mmap | |
1400 |
|
1401 | |||
1401 | * force_nodemap: |
|
1402 | * force_nodemap: | |
1402 | force the usage of a "development" version of the nodemap code |
|
1403 | force the usage of a "development" version of the nodemap code | |
1403 | """ |
|
1404 | """ | |
1404 | opts = self.opener.options |
|
1405 | opts = self.opener.options | |
1405 |
|
1406 | |||
1406 | if b'changelogv2' in opts and self.revlog_kind == KIND_CHANGELOG: |
|
1407 | if b'changelogv2' in opts and self.revlog_kind == KIND_CHANGELOG: | |
1407 | new_header = CHANGELOGV2 |
|
1408 | new_header = CHANGELOGV2 | |
1408 | compute_rank = opts.get(b'changelogv2.compute-rank', True) |
|
1409 | compute_rank = opts.get(b'changelogv2.compute-rank', True) | |
1409 | self.feature_config.compute_rank = compute_rank |
|
1410 | self.feature_config.compute_rank = compute_rank | |
1410 | elif b'revlogv2' in opts: |
|
1411 | elif b'revlogv2' in opts: | |
1411 | new_header = REVLOGV2 |
|
1412 | new_header = REVLOGV2 | |
1412 | elif b'revlogv1' in opts: |
|
1413 | elif b'revlogv1' in opts: | |
1413 | new_header = REVLOGV1 |
|
1414 | new_header = REVLOGV1 | |
1414 | if self._may_inline: |
|
1415 | if self._may_inline: | |
1415 | new_header |= FLAG_INLINE_DATA |
|
1416 | new_header |= FLAG_INLINE_DATA | |
1416 | if b'generaldelta' in opts: |
|
1417 | if b'generaldelta' in opts: | |
1417 | new_header |= FLAG_GENERALDELTA |
|
1418 | new_header |= FLAG_GENERALDELTA | |
1418 | elif b'revlogv0' in self.opener.options: |
|
1419 | elif b'revlogv0' in self.opener.options: | |
1419 | new_header = REVLOGV0 |
|
1420 | new_header = REVLOGV0 | |
1420 | else: |
|
1421 | else: | |
1421 | new_header = REVLOG_DEFAULT_VERSION |
|
1422 | new_header = REVLOG_DEFAULT_VERSION | |
1422 |
|
1423 | |||
1423 | mmapindexthreshold = None |
|
1424 | mmapindexthreshold = None | |
1424 | if self.data_config.mmap_large_index: |
|
1425 | if self.data_config.mmap_large_index: | |
1425 | mmapindexthreshold = self.data_config.mmap_index_threshold |
|
1426 | mmapindexthreshold = self.data_config.mmap_index_threshold | |
1426 | if self.feature_config.enable_ellipsis: |
|
1427 | if self.feature_config.enable_ellipsis: | |
1427 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor |
|
1428 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor | |
1428 |
|
1429 | |||
1429 | # revlog v0 doesn't have flag processors |
|
1430 | # revlog v0 doesn't have flag processors | |
1430 | for flag, processor in opts.get(b'flagprocessors', {}).items(): |
|
1431 | for flag, processor in opts.get(b'flagprocessors', {}).items(): | |
1431 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) |
|
1432 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) | |
1432 |
|
1433 | |||
1433 | chunk_cache_size = self.data_config.chunk_cache_size |
|
1434 | chunk_cache_size = self.data_config.chunk_cache_size | |
1434 | if chunk_cache_size <= 0: |
|
1435 | if chunk_cache_size <= 0: | |
1435 | raise error.RevlogError( |
|
1436 | raise error.RevlogError( | |
1436 | _(b'revlog chunk cache size %r is not greater than 0') |
|
1437 | _(b'revlog chunk cache size %r is not greater than 0') | |
1437 | % chunk_cache_size |
|
1438 | % chunk_cache_size | |
1438 | ) |
|
1439 | ) | |
1439 | elif chunk_cache_size & (chunk_cache_size - 1): |
|
1440 | elif chunk_cache_size & (chunk_cache_size - 1): | |
1440 | raise error.RevlogError( |
|
1441 | raise error.RevlogError( | |
1441 | _(b'revlog chunk cache size %r is not a power of 2') |
|
1442 | _(b'revlog chunk cache size %r is not a power of 2') | |
1442 | % chunk_cache_size |
|
1443 | % chunk_cache_size | |
1443 | ) |
|
1444 | ) | |
1444 | force_nodemap = opts.get(b'devel-force-nodemap', False) |
|
1445 | force_nodemap = opts.get(b'devel-force-nodemap', False) | |
1445 | return new_header, mmapindexthreshold, force_nodemap |
|
1446 | return new_header, mmapindexthreshold, force_nodemap | |
1446 |
|
1447 | |||
1447 | def _get_data(self, filepath, mmap_threshold, size=None): |
|
1448 | def _get_data(self, filepath, mmap_threshold, size=None): | |
1448 | """return a file content with or without mmap |
|
1449 | """return a file content with or without mmap | |
1449 |
|
1450 | |||
1450 | If the file is missing return the empty string""" |
|
1451 | If the file is missing return the empty string""" | |
1451 | try: |
|
1452 | try: | |
1452 | with self.opener(filepath) as fp: |
|
1453 | with self.opener(filepath) as fp: | |
1453 | if mmap_threshold is not None: |
|
1454 | if mmap_threshold is not None: | |
1454 | file_size = self.opener.fstat(fp).st_size |
|
1455 | file_size = self.opener.fstat(fp).st_size | |
1455 | if file_size >= mmap_threshold: |
|
1456 | if file_size >= mmap_threshold: | |
1456 | if size is not None: |
|
1457 | if size is not None: | |
1457 | # avoid potentiel mmap crash |
|
1458 | # avoid potentiel mmap crash | |
1458 | size = min(file_size, size) |
|
1459 | size = min(file_size, size) | |
1459 | # TODO: should .close() to release resources without |
|
1460 | # TODO: should .close() to release resources without | |
1460 | # relying on Python GC |
|
1461 | # relying on Python GC | |
1461 | if size is None: |
|
1462 | if size is None: | |
1462 | return util.buffer(util.mmapread(fp)) |
|
1463 | return util.buffer(util.mmapread(fp)) | |
1463 | else: |
|
1464 | else: | |
1464 | return util.buffer(util.mmapread(fp, size)) |
|
1465 | return util.buffer(util.mmapread(fp, size)) | |
1465 | if size is None: |
|
1466 | if size is None: | |
1466 | return fp.read() |
|
1467 | return fp.read() | |
1467 | else: |
|
1468 | else: | |
1468 | return fp.read(size) |
|
1469 | return fp.read(size) | |
1469 | except FileNotFoundError: |
|
1470 | except FileNotFoundError: | |
1470 | return b'' |
|
1471 | return b'' | |
1471 |
|
1472 | |||
1472 | def get_streams(self, max_linkrev, force_inline=False): |
|
1473 | def get_streams(self, max_linkrev, force_inline=False): | |
1473 | """return a list of streams that represent this revlog |
|
1474 | """return a list of streams that represent this revlog | |
1474 |
|
1475 | |||
1475 | This is used by stream-clone to do bytes to bytes copies of a repository. |
|
1476 | This is used by stream-clone to do bytes to bytes copies of a repository. | |
1476 |
|
1477 | |||
1477 | This streams data for all revisions that refer to a changelog revision up |
|
1478 | This streams data for all revisions that refer to a changelog revision up | |
1478 | to `max_linkrev`. |
|
1479 | to `max_linkrev`. | |
1479 |
|
1480 | |||
1480 | If `force_inline` is set, it enforces that the stream will represent an inline revlog. |
|
1481 | If `force_inline` is set, it enforces that the stream will represent an inline revlog. | |
1481 |
|
1482 | |||
1482 | It returns is a list of three-tuple: |
|
1483 | It returns is a list of three-tuple: | |
1483 |
|
1484 | |||
1484 | [ |
|
1485 | [ | |
1485 | (filename, bytes_stream, stream_size), |
|
1486 | (filename, bytes_stream, stream_size), | |
1486 | … |
|
1487 | … | |
1487 | ] |
|
1488 | ] | |
1488 | """ |
|
1489 | """ | |
1489 | n = len(self) |
|
1490 | n = len(self) | |
1490 | index = self.index |
|
1491 | index = self.index | |
1491 | while n > 0: |
|
1492 | while n > 0: | |
1492 | linkrev = index[n - 1][4] |
|
1493 | linkrev = index[n - 1][4] | |
1493 | if linkrev < max_linkrev: |
|
1494 | if linkrev < max_linkrev: | |
1494 | break |
|
1495 | break | |
1495 | # note: this loop will rarely go through multiple iterations, since |
|
1496 | # note: this loop will rarely go through multiple iterations, since | |
1496 | # it only traverses commits created during the current streaming |
|
1497 | # it only traverses commits created during the current streaming | |
1497 | # pull operation. |
|
1498 | # pull operation. | |
1498 | # |
|
1499 | # | |
1499 | # If this become a problem, using a binary search should cap the |
|
1500 | # If this become a problem, using a binary search should cap the | |
1500 | # runtime of this. |
|
1501 | # runtime of this. | |
1501 | n = n - 1 |
|
1502 | n = n - 1 | |
1502 | if n == 0: |
|
1503 | if n == 0: | |
1503 | # no data to send |
|
1504 | # no data to send | |
1504 | return [] |
|
1505 | return [] | |
1505 | index_size = n * index.entry_size |
|
1506 | index_size = n * index.entry_size | |
1506 | data_size = self.end(n - 1) |
|
1507 | data_size = self.end(n - 1) | |
1507 |
|
1508 | |||
1508 | # XXX we might have been split (or stripped) since the object |
|
1509 | # XXX we might have been split (or stripped) since the object | |
1509 | # initialization, We need to close this race too, but having a way to |
|
1510 | # initialization, We need to close this race too, but having a way to | |
1510 | # pre-open the file we feed to the revlog and never closing them before |
|
1511 | # pre-open the file we feed to the revlog and never closing them before | |
1511 | # we are done streaming. |
|
1512 | # we are done streaming. | |
1512 |
|
1513 | |||
1513 | if self._inline: |
|
1514 | if self._inline: | |
1514 |
|
1515 | |||
1515 | def get_stream(): |
|
1516 | def get_stream(): | |
1516 | with self.opener(self._indexfile, mode=b"r") as fp: |
|
1517 | with self.opener(self._indexfile, mode=b"r") as fp: | |
1517 | yield None |
|
1518 | yield None | |
1518 | size = index_size + data_size |
|
1519 | size = index_size + data_size | |
1519 | if size <= 65536: |
|
1520 | if size <= 65536: | |
1520 | yield fp.read(size) |
|
1521 | yield fp.read(size) | |
1521 | else: |
|
1522 | else: | |
1522 | yield from util.filechunkiter(fp, limit=size) |
|
1523 | yield from util.filechunkiter(fp, limit=size) | |
1523 |
|
1524 | |||
1524 | inline_stream = get_stream() |
|
1525 | inline_stream = get_stream() | |
1525 | next(inline_stream) |
|
1526 | next(inline_stream) | |
1526 | return [ |
|
1527 | return [ | |
1527 | (self._indexfile, inline_stream, index_size + data_size), |
|
1528 | (self._indexfile, inline_stream, index_size + data_size), | |
1528 | ] |
|
1529 | ] | |
1529 | elif force_inline: |
|
1530 | elif force_inline: | |
1530 |
|
1531 | |||
1531 | def get_stream(): |
|
1532 | def get_stream(): | |
1532 | with self.reading(): |
|
1533 | with self.reading(): | |
1533 | yield None |
|
1534 | yield None | |
1534 |
|
1535 | |||
1535 | for rev in range(n): |
|
1536 | for rev in range(n): | |
1536 | idx = self.index.entry_binary(rev) |
|
1537 | idx = self.index.entry_binary(rev) | |
1537 | if rev == 0 and self._docket is None: |
|
1538 | if rev == 0 and self._docket is None: | |
1538 | # re-inject the inline flag |
|
1539 | # re-inject the inline flag | |
1539 | header = self._format_flags |
|
1540 | header = self._format_flags | |
1540 | header |= self._format_version |
|
1541 | header |= self._format_version | |
1541 | header |= FLAG_INLINE_DATA |
|
1542 | header |= FLAG_INLINE_DATA | |
1542 | header = self.index.pack_header(header) |
|
1543 | header = self.index.pack_header(header) | |
1543 | idx = header + idx |
|
1544 | idx = header + idx | |
1544 | yield idx |
|
1545 | yield idx | |
1545 | yield self._inner.get_segment_for_revs(rev, rev)[1] |
|
1546 | yield self._inner.get_segment_for_revs(rev, rev)[1] | |
1546 |
|
1547 | |||
1547 | inline_stream = get_stream() |
|
1548 | inline_stream = get_stream() | |
1548 | next(inline_stream) |
|
1549 | next(inline_stream) | |
1549 | return [ |
|
1550 | return [ | |
1550 | (self._indexfile, inline_stream, index_size + data_size), |
|
1551 | (self._indexfile, inline_stream, index_size + data_size), | |
1551 | ] |
|
1552 | ] | |
1552 | else: |
|
1553 | else: | |
1553 |
|
1554 | |||
1554 | def get_index_stream(): |
|
1555 | def get_index_stream(): | |
1555 | with self.opener(self._indexfile, mode=b"r") as fp: |
|
1556 | with self.opener(self._indexfile, mode=b"r") as fp: | |
1556 | yield None |
|
1557 | yield None | |
1557 | if index_size <= 65536: |
|
1558 | if index_size <= 65536: | |
1558 | yield fp.read(index_size) |
|
1559 | yield fp.read(index_size) | |
1559 | else: |
|
1560 | else: | |
1560 | yield from util.filechunkiter(fp, limit=index_size) |
|
1561 | yield from util.filechunkiter(fp, limit=index_size) | |
1561 |
|
1562 | |||
1562 | def get_data_stream(): |
|
1563 | def get_data_stream(): | |
1563 | with self._datafp() as fp: |
|
1564 | with self._datafp() as fp: | |
1564 | yield None |
|
1565 | yield None | |
1565 | if data_size <= 65536: |
|
1566 | if data_size <= 65536: | |
1566 | yield fp.read(data_size) |
|
1567 | yield fp.read(data_size) | |
1567 | else: |
|
1568 | else: | |
1568 | yield from util.filechunkiter(fp, limit=data_size) |
|
1569 | yield from util.filechunkiter(fp, limit=data_size) | |
1569 |
|
1570 | |||
1570 | index_stream = get_index_stream() |
|
1571 | index_stream = get_index_stream() | |
1571 | next(index_stream) |
|
1572 | next(index_stream) | |
1572 | data_stream = get_data_stream() |
|
1573 | data_stream = get_data_stream() | |
1573 | next(data_stream) |
|
1574 | next(data_stream) | |
1574 | return [ |
|
1575 | return [ | |
1575 | (self._datafile, data_stream, data_size), |
|
1576 | (self._datafile, data_stream, data_size), | |
1576 | (self._indexfile, index_stream, index_size), |
|
1577 | (self._indexfile, index_stream, index_size), | |
1577 | ] |
|
1578 | ] | |
1578 |
|
1579 | |||
1579 | def _loadindex(self, docket=None): |
|
1580 | def _loadindex(self, docket=None): | |
1580 |
|
1581 | |||
1581 | new_header, mmapindexthreshold, force_nodemap = self._init_opts() |
|
1582 | new_header, mmapindexthreshold, force_nodemap = self._init_opts() | |
1582 |
|
1583 | |||
1583 | if self.postfix is not None: |
|
1584 | if self.postfix is not None: | |
1584 | entry_point = b'%s.i.%s' % (self.radix, self.postfix) |
|
1585 | entry_point = b'%s.i.%s' % (self.radix, self.postfix) | |
1585 | elif self._trypending and self.opener.exists(b'%s.i.a' % self.radix): |
|
1586 | elif self._trypending and self.opener.exists(b'%s.i.a' % self.radix): | |
1586 | entry_point = b'%s.i.a' % self.radix |
|
1587 | entry_point = b'%s.i.a' % self.radix | |
1587 | elif self._try_split and self.opener.exists(self._split_index_file): |
|
1588 | elif self._try_split and self.opener.exists(self._split_index_file): | |
1588 | entry_point = self._split_index_file |
|
1589 | entry_point = self._split_index_file | |
1589 | else: |
|
1590 | else: | |
1590 | entry_point = b'%s.i' % self.radix |
|
1591 | entry_point = b'%s.i' % self.radix | |
1591 |
|
1592 | |||
1592 | if docket is not None: |
|
1593 | if docket is not None: | |
1593 | self._docket = docket |
|
1594 | self._docket = docket | |
1594 | self._docket_file = entry_point |
|
1595 | self._docket_file = entry_point | |
1595 | else: |
|
1596 | else: | |
1596 | self._initempty = True |
|
1597 | self._initempty = True | |
1597 | entry_data = self._get_data(entry_point, mmapindexthreshold) |
|
1598 | entry_data = self._get_data(entry_point, mmapindexthreshold) | |
1598 | if len(entry_data) > 0: |
|
1599 | if len(entry_data) > 0: | |
1599 | header = INDEX_HEADER.unpack(entry_data[:4])[0] |
|
1600 | header = INDEX_HEADER.unpack(entry_data[:4])[0] | |
1600 | self._initempty = False |
|
1601 | self._initempty = False | |
1601 | else: |
|
1602 | else: | |
1602 | header = new_header |
|
1603 | header = new_header | |
1603 |
|
1604 | |||
1604 | self._format_flags = header & ~0xFFFF |
|
1605 | self._format_flags = header & ~0xFFFF | |
1605 | self._format_version = header & 0xFFFF |
|
1606 | self._format_version = header & 0xFFFF | |
1606 |
|
1607 | |||
1607 | supported_flags = SUPPORTED_FLAGS.get(self._format_version) |
|
1608 | supported_flags = SUPPORTED_FLAGS.get(self._format_version) | |
1608 | if supported_flags is None: |
|
1609 | if supported_flags is None: | |
1609 | msg = _(b'unknown version (%d) in revlog %s') |
|
1610 | msg = _(b'unknown version (%d) in revlog %s') | |
1610 | msg %= (self._format_version, self.display_id) |
|
1611 | msg %= (self._format_version, self.display_id) | |
1611 | raise error.RevlogError(msg) |
|
1612 | raise error.RevlogError(msg) | |
1612 | elif self._format_flags & ~supported_flags: |
|
1613 | elif self._format_flags & ~supported_flags: | |
1613 | msg = _(b'unknown flags (%#04x) in version %d revlog %s') |
|
1614 | msg = _(b'unknown flags (%#04x) in version %d revlog %s') | |
1614 | display_flag = self._format_flags >> 16 |
|
1615 | display_flag = self._format_flags >> 16 | |
1615 | msg %= (display_flag, self._format_version, self.display_id) |
|
1616 | msg %= (display_flag, self._format_version, self.display_id) | |
1616 | raise error.RevlogError(msg) |
|
1617 | raise error.RevlogError(msg) | |
1617 |
|
1618 | |||
1618 | features = FEATURES_BY_VERSION[self._format_version] |
|
1619 | features = FEATURES_BY_VERSION[self._format_version] | |
1619 | self._inline = features[b'inline'](self._format_flags) |
|
1620 | self._inline = features[b'inline'](self._format_flags) | |
1620 | self.delta_config.general_delta = features[b'generaldelta']( |
|
1621 | self.delta_config.general_delta = features[b'generaldelta']( | |
1621 | self._format_flags |
|
1622 | self._format_flags | |
1622 | ) |
|
1623 | ) | |
1623 | self.feature_config.has_side_data = features[b'sidedata'] |
|
1624 | self.feature_config.has_side_data = features[b'sidedata'] | |
1624 |
|
1625 | |||
1625 | if not features[b'docket']: |
|
1626 | if not features[b'docket']: | |
1626 | self._indexfile = entry_point |
|
1627 | self._indexfile = entry_point | |
1627 | index_data = entry_data |
|
1628 | index_data = entry_data | |
1628 | else: |
|
1629 | else: | |
1629 | self._docket_file = entry_point |
|
1630 | self._docket_file = entry_point | |
1630 | if self._initempty: |
|
1631 | if self._initempty: | |
1631 | self._docket = docketutil.default_docket(self, header) |
|
1632 | self._docket = docketutil.default_docket(self, header) | |
1632 | else: |
|
1633 | else: | |
1633 | self._docket = docketutil.parse_docket( |
|
1634 | self._docket = docketutil.parse_docket( | |
1634 | self, entry_data, use_pending=self._trypending |
|
1635 | self, entry_data, use_pending=self._trypending | |
1635 | ) |
|
1636 | ) | |
1636 |
|
1637 | |||
1637 | if self._docket is not None: |
|
1638 | if self._docket is not None: | |
1638 | self._indexfile = self._docket.index_filepath() |
|
1639 | self._indexfile = self._docket.index_filepath() | |
1639 | index_data = b'' |
|
1640 | index_data = b'' | |
1640 | index_size = self._docket.index_end |
|
1641 | index_size = self._docket.index_end | |
1641 | if index_size > 0: |
|
1642 | if index_size > 0: | |
1642 | index_data = self._get_data( |
|
1643 | index_data = self._get_data( | |
1643 | self._indexfile, mmapindexthreshold, size=index_size |
|
1644 | self._indexfile, mmapindexthreshold, size=index_size | |
1644 | ) |
|
1645 | ) | |
1645 | if len(index_data) < index_size: |
|
1646 | if len(index_data) < index_size: | |
1646 | msg = _(b'too few index data for %s: got %d, expected %d') |
|
1647 | msg = _(b'too few index data for %s: got %d, expected %d') | |
1647 | msg %= (self.display_id, len(index_data), index_size) |
|
1648 | msg %= (self.display_id, len(index_data), index_size) | |
1648 | raise error.RevlogError(msg) |
|
1649 | raise error.RevlogError(msg) | |
1649 |
|
1650 | |||
1650 | self._inline = False |
|
1651 | self._inline = False | |
1651 | # generaldelta implied by version 2 revlogs. |
|
1652 | # generaldelta implied by version 2 revlogs. | |
1652 | self.delta_config.general_delta = True |
|
1653 | self.delta_config.general_delta = True | |
1653 | # the logic for persistent nodemap will be dealt with within the |
|
1654 | # the logic for persistent nodemap will be dealt with within the | |
1654 | # main docket, so disable it for now. |
|
1655 | # main docket, so disable it for now. | |
1655 | self._nodemap_file = None |
|
1656 | self._nodemap_file = None | |
1656 |
|
1657 | |||
1657 | if self._docket is not None: |
|
1658 | if self._docket is not None: | |
1658 | self._datafile = self._docket.data_filepath() |
|
1659 | self._datafile = self._docket.data_filepath() | |
1659 | self._sidedatafile = self._docket.sidedata_filepath() |
|
1660 | self._sidedatafile = self._docket.sidedata_filepath() | |
1660 | elif self.postfix is None: |
|
1661 | elif self.postfix is None: | |
1661 | self._datafile = b'%s.d' % self.radix |
|
1662 | self._datafile = b'%s.d' % self.radix | |
1662 | else: |
|
1663 | else: | |
1663 | self._datafile = b'%s.d.%s' % (self.radix, self.postfix) |
|
1664 | self._datafile = b'%s.d.%s' % (self.radix, self.postfix) | |
1664 |
|
1665 | |||
1665 | self.nodeconstants = sha1nodeconstants |
|
1666 | self.nodeconstants = sha1nodeconstants | |
1666 | self.nullid = self.nodeconstants.nullid |
|
1667 | self.nullid = self.nodeconstants.nullid | |
1667 |
|
1668 | |||
1668 | # sparse-revlog can't be on without general-delta (issue6056) |
|
1669 | # sparse-revlog can't be on without general-delta (issue6056) | |
1669 | if not self.delta_config.general_delta: |
|
1670 | if not self.delta_config.general_delta: | |
1670 | self.delta_config.sparse_revlog = False |
|
1671 | self.delta_config.sparse_revlog = False | |
1671 |
|
1672 | |||
1672 | self._storedeltachains = True |
|
1673 | self._storedeltachains = True | |
1673 |
|
1674 | |||
1674 | devel_nodemap = ( |
|
1675 | devel_nodemap = ( | |
1675 | self._nodemap_file |
|
1676 | self._nodemap_file | |
1676 | and force_nodemap |
|
1677 | and force_nodemap | |
1677 | and parse_index_v1_nodemap is not None |
|
1678 | and parse_index_v1_nodemap is not None | |
1678 | ) |
|
1679 | ) | |
1679 |
|
1680 | |||
1680 | use_rust_index = False |
|
1681 | use_rust_index = False | |
1681 | if rustrevlog is not None: |
|
1682 | if rustrevlog is not None: | |
1682 | if self._nodemap_file is not None: |
|
1683 | if self._nodemap_file is not None: | |
1683 | use_rust_index = True |
|
1684 | use_rust_index = True | |
1684 | else: |
|
1685 | else: | |
1685 | use_rust_index = self.opener.options.get(b'rust.index') |
|
1686 | use_rust_index = self.opener.options.get(b'rust.index') | |
1686 |
|
1687 | |||
1687 | self._parse_index = parse_index_v1 |
|
1688 | self._parse_index = parse_index_v1 | |
1688 | if self._format_version == REVLOGV0: |
|
1689 | if self._format_version == REVLOGV0: | |
1689 | self._parse_index = revlogv0.parse_index_v0 |
|
1690 | self._parse_index = revlogv0.parse_index_v0 | |
1690 | elif self._format_version == REVLOGV2: |
|
1691 | elif self._format_version == REVLOGV2: | |
1691 | self._parse_index = parse_index_v2 |
|
1692 | self._parse_index = parse_index_v2 | |
1692 | elif self._format_version == CHANGELOGV2: |
|
1693 | elif self._format_version == CHANGELOGV2: | |
1693 | self._parse_index = parse_index_cl_v2 |
|
1694 | self._parse_index = parse_index_cl_v2 | |
1694 | elif devel_nodemap: |
|
1695 | elif devel_nodemap: | |
1695 | self._parse_index = parse_index_v1_nodemap |
|
1696 | self._parse_index = parse_index_v1_nodemap | |
1696 | elif use_rust_index: |
|
1697 | elif use_rust_index: | |
1697 |
self._parse_index = |
|
1698 | self._parse_index = functools.partial( | |
|
1699 | parse_index_v1_mixed, default_header=new_header | |||
|
1700 | ) | |||
1698 | try: |
|
1701 | try: | |
1699 | d = self._parse_index(index_data, self._inline) |
|
1702 | d = self._parse_index(index_data, self._inline) | |
1700 | index, chunkcache = d |
|
1703 | index, chunkcache = d | |
1701 | use_nodemap = ( |
|
1704 | use_nodemap = ( | |
1702 | not self._inline |
|
1705 | not self._inline | |
1703 | and self._nodemap_file is not None |
|
1706 | and self._nodemap_file is not None | |
1704 | and hasattr(index, 'update_nodemap_data') |
|
1707 | and hasattr(index, 'update_nodemap_data') | |
1705 | ) |
|
1708 | ) | |
1706 | if use_nodemap: |
|
1709 | if use_nodemap: | |
1707 | nodemap_data = nodemaputil.persisted_data(self) |
|
1710 | nodemap_data = nodemaputil.persisted_data(self) | |
1708 | if nodemap_data is not None: |
|
1711 | if nodemap_data is not None: | |
1709 | docket = nodemap_data[0] |
|
1712 | docket = nodemap_data[0] | |
1710 | if ( |
|
1713 | if ( | |
1711 | len(d[0]) > docket.tip_rev |
|
1714 | len(d[0]) > docket.tip_rev | |
1712 | and d[0][docket.tip_rev][7] == docket.tip_node |
|
1715 | and d[0][docket.tip_rev][7] == docket.tip_node | |
1713 | ): |
|
1716 | ): | |
1714 | # no changelog tampering |
|
1717 | # no changelog tampering | |
1715 | self._nodemap_docket = docket |
|
1718 | self._nodemap_docket = docket | |
1716 | index.update_nodemap_data(*nodemap_data) |
|
1719 | index.update_nodemap_data(*nodemap_data) | |
1717 | except (ValueError, IndexError): |
|
1720 | except (ValueError, IndexError): | |
1718 | raise error.RevlogError( |
|
1721 | raise error.RevlogError( | |
1719 | _(b"index %s is corrupted") % self.display_id |
|
1722 | _(b"index %s is corrupted") % self.display_id | |
1720 | ) |
|
1723 | ) | |
1721 | self.index = index |
|
1724 | self.index = index | |
1722 | # revnum -> (chain-length, sum-delta-length) |
|
1725 | # revnum -> (chain-length, sum-delta-length) | |
1723 | self._chaininfocache = util.lrucachedict(500) |
|
1726 | self._chaininfocache = util.lrucachedict(500) | |
1724 |
|
1727 | |||
1725 | return chunkcache |
|
1728 | return chunkcache | |
1726 |
|
1729 | |||
1727 | def _load_inner(self, chunk_cache): |
|
1730 | def _load_inner(self, chunk_cache): | |
1728 | if self._docket is None: |
|
1731 | if self._docket is None: | |
1729 | default_compression_header = None |
|
1732 | default_compression_header = None | |
1730 | else: |
|
1733 | else: | |
1731 | default_compression_header = self._docket.default_compression_header |
|
1734 | default_compression_header = self._docket.default_compression_header | |
1732 |
|
1735 | |||
1733 | self._inner = _InnerRevlog( |
|
1736 | self._inner = _InnerRevlog( | |
1734 | opener=self.opener, |
|
1737 | opener=self.opener, | |
1735 | index=self.index, |
|
1738 | index=self.index, | |
1736 | index_file=self._indexfile, |
|
1739 | index_file=self._indexfile, | |
1737 | data_file=self._datafile, |
|
1740 | data_file=self._datafile, | |
1738 | sidedata_file=self._sidedatafile, |
|
1741 | sidedata_file=self._sidedatafile, | |
1739 | inline=self._inline, |
|
1742 | inline=self._inline, | |
1740 | data_config=self.data_config, |
|
1743 | data_config=self.data_config, | |
1741 | delta_config=self.delta_config, |
|
1744 | delta_config=self.delta_config, | |
1742 | feature_config=self.feature_config, |
|
1745 | feature_config=self.feature_config, | |
1743 | chunk_cache=chunk_cache, |
|
1746 | chunk_cache=chunk_cache, | |
1744 | default_compression_header=default_compression_header, |
|
1747 | default_compression_header=default_compression_header, | |
1745 | ) |
|
1748 | ) | |
1746 |
|
1749 | |||
1747 | def get_revlog(self): |
|
1750 | def get_revlog(self): | |
1748 | """simple function to mirror API of other not-really-revlog API""" |
|
1751 | """simple function to mirror API of other not-really-revlog API""" | |
1749 | return self |
|
1752 | return self | |
1750 |
|
1753 | |||
1751 | @util.propertycache |
|
1754 | @util.propertycache | |
1752 | def revlog_kind(self): |
|
1755 | def revlog_kind(self): | |
1753 | return self.target[0] |
|
1756 | return self.target[0] | |
1754 |
|
1757 | |||
1755 | @util.propertycache |
|
1758 | @util.propertycache | |
1756 | def display_id(self): |
|
1759 | def display_id(self): | |
1757 | """The public facing "ID" of the revlog that we use in message""" |
|
1760 | """The public facing "ID" of the revlog that we use in message""" | |
1758 | if self.revlog_kind == KIND_FILELOG: |
|
1761 | if self.revlog_kind == KIND_FILELOG: | |
1759 | # Reference the file without the "data/" prefix, so it is familiar |
|
1762 | # Reference the file without the "data/" prefix, so it is familiar | |
1760 | # to the user. |
|
1763 | # to the user. | |
1761 | return self.target[1] |
|
1764 | return self.target[1] | |
1762 | else: |
|
1765 | else: | |
1763 | return self.radix |
|
1766 | return self.radix | |
1764 |
|
1767 | |||
1765 | def _datafp(self, mode=b'r'): |
|
1768 | def _datafp(self, mode=b'r'): | |
1766 | """file object for the revlog's data file""" |
|
1769 | """file object for the revlog's data file""" | |
1767 | return self.opener(self._datafile, mode=mode) |
|
1770 | return self.opener(self._datafile, mode=mode) | |
1768 |
|
1771 | |||
1769 | def tiprev(self): |
|
1772 | def tiprev(self): | |
1770 | return len(self.index) - 1 |
|
1773 | return len(self.index) - 1 | |
1771 |
|
1774 | |||
1772 | def tip(self): |
|
1775 | def tip(self): | |
1773 | return self.node(self.tiprev()) |
|
1776 | return self.node(self.tiprev()) | |
1774 |
|
1777 | |||
1775 | def __contains__(self, rev): |
|
1778 | def __contains__(self, rev): | |
1776 | return 0 <= rev < len(self) |
|
1779 | return 0 <= rev < len(self) | |
1777 |
|
1780 | |||
1778 | def __len__(self): |
|
1781 | def __len__(self): | |
1779 | return len(self.index) |
|
1782 | return len(self.index) | |
1780 |
|
1783 | |||
1781 | def __iter__(self): |
|
1784 | def __iter__(self): | |
1782 | return iter(range(len(self))) |
|
1785 | return iter(range(len(self))) | |
1783 |
|
1786 | |||
1784 | def revs(self, start=0, stop=None): |
|
1787 | def revs(self, start=0, stop=None): | |
1785 | """iterate over all rev in this revlog (from start to stop)""" |
|
1788 | """iterate over all rev in this revlog (from start to stop)""" | |
1786 | return storageutil.iterrevs(len(self), start=start, stop=stop) |
|
1789 | return storageutil.iterrevs(len(self), start=start, stop=stop) | |
1787 |
|
1790 | |||
1788 | def hasnode(self, node): |
|
1791 | def hasnode(self, node): | |
1789 | try: |
|
1792 | try: | |
1790 | self.rev(node) |
|
1793 | self.rev(node) | |
1791 | return True |
|
1794 | return True | |
1792 | except KeyError: |
|
1795 | except KeyError: | |
1793 | return False |
|
1796 | return False | |
1794 |
|
1797 | |||
1795 | def _candelta(self, baserev, rev): |
|
1798 | def _candelta(self, baserev, rev): | |
1796 | """whether two revisions (baserev, rev) can be delta-ed or not""" |
|
1799 | """whether two revisions (baserev, rev) can be delta-ed or not""" | |
1797 | # Disable delta if either rev requires a content-changing flag |
|
1800 | # Disable delta if either rev requires a content-changing flag | |
1798 | # processor (ex. LFS). This is because such flag processor can alter |
|
1801 | # processor (ex. LFS). This is because such flag processor can alter | |
1799 | # the rawtext content that the delta will be based on, and two clients |
|
1802 | # the rawtext content that the delta will be based on, and two clients | |
1800 | # could have a same revlog node with different flags (i.e. different |
|
1803 | # could have a same revlog node with different flags (i.e. different | |
1801 | # rawtext contents) and the delta could be incompatible. |
|
1804 | # rawtext contents) and the delta could be incompatible. | |
1802 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
1805 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( | |
1803 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS |
|
1806 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS | |
1804 | ): |
|
1807 | ): | |
1805 | return False |
|
1808 | return False | |
1806 | return True |
|
1809 | return True | |
1807 |
|
1810 | |||
1808 | def update_caches(self, transaction): |
|
1811 | def update_caches(self, transaction): | |
1809 | """update on disk cache |
|
1812 | """update on disk cache | |
1810 |
|
1813 | |||
1811 | If a transaction is passed, the update may be delayed to transaction |
|
1814 | If a transaction is passed, the update may be delayed to transaction | |
1812 | commit.""" |
|
1815 | commit.""" | |
1813 | if self._nodemap_file is not None: |
|
1816 | if self._nodemap_file is not None: | |
1814 | if transaction is None: |
|
1817 | if transaction is None: | |
1815 | nodemaputil.update_persistent_nodemap(self) |
|
1818 | nodemaputil.update_persistent_nodemap(self) | |
1816 | else: |
|
1819 | else: | |
1817 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
1820 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
1818 |
|
1821 | |||
1819 | def clearcaches(self): |
|
1822 | def clearcaches(self): | |
1820 | """Clear in-memory caches""" |
|
1823 | """Clear in-memory caches""" | |
1821 | self._chainbasecache.clear() |
|
1824 | self._chainbasecache.clear() | |
1822 | self._inner.clear_cache() |
|
1825 | self._inner.clear_cache() | |
1823 | self._pcache = {} |
|
1826 | self._pcache = {} | |
1824 | self._nodemap_docket = None |
|
1827 | self._nodemap_docket = None | |
1825 | self.index.clearcaches() |
|
1828 | self.index.clearcaches() | |
1826 | # The python code is the one responsible for validating the docket, we |
|
1829 | # The python code is the one responsible for validating the docket, we | |
1827 | # end up having to refresh it here. |
|
1830 | # end up having to refresh it here. | |
1828 | use_nodemap = ( |
|
1831 | use_nodemap = ( | |
1829 | not self._inline |
|
1832 | not self._inline | |
1830 | and self._nodemap_file is not None |
|
1833 | and self._nodemap_file is not None | |
1831 | and hasattr(self.index, 'update_nodemap_data') |
|
1834 | and hasattr(self.index, 'update_nodemap_data') | |
1832 | ) |
|
1835 | ) | |
1833 | if use_nodemap: |
|
1836 | if use_nodemap: | |
1834 | nodemap_data = nodemaputil.persisted_data(self) |
|
1837 | nodemap_data = nodemaputil.persisted_data(self) | |
1835 | if nodemap_data is not None: |
|
1838 | if nodemap_data is not None: | |
1836 | self._nodemap_docket = nodemap_data[0] |
|
1839 | self._nodemap_docket = nodemap_data[0] | |
1837 | self.index.update_nodemap_data(*nodemap_data) |
|
1840 | self.index.update_nodemap_data(*nodemap_data) | |
1838 |
|
1841 | |||
1839 | def rev(self, node): |
|
1842 | def rev(self, node): | |
1840 | """return the revision number associated with a <nodeid>""" |
|
1843 | """return the revision number associated with a <nodeid>""" | |
1841 | try: |
|
1844 | try: | |
1842 | return self.index.rev(node) |
|
1845 | return self.index.rev(node) | |
1843 | except TypeError: |
|
1846 | except TypeError: | |
1844 | raise |
|
1847 | raise | |
1845 | except error.RevlogError: |
|
1848 | except error.RevlogError: | |
1846 | # parsers.c radix tree lookup failed |
|
1849 | # parsers.c radix tree lookup failed | |
1847 | if ( |
|
1850 | if ( | |
1848 | node == self.nodeconstants.wdirid |
|
1851 | node == self.nodeconstants.wdirid | |
1849 | or node in self.nodeconstants.wdirfilenodeids |
|
1852 | or node in self.nodeconstants.wdirfilenodeids | |
1850 | ): |
|
1853 | ): | |
1851 | raise error.WdirUnsupported |
|
1854 | raise error.WdirUnsupported | |
1852 | raise error.LookupError(node, self.display_id, _(b'no node')) |
|
1855 | raise error.LookupError(node, self.display_id, _(b'no node')) | |
1853 |
|
1856 | |||
1854 | # Accessors for index entries. |
|
1857 | # Accessors for index entries. | |
1855 |
|
1858 | |||
1856 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes |
|
1859 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes | |
1857 | # are flags. |
|
1860 | # are flags. | |
1858 | def start(self, rev): |
|
1861 | def start(self, rev): | |
1859 | return int(self.index[rev][0] >> 16) |
|
1862 | return int(self.index[rev][0] >> 16) | |
1860 |
|
1863 | |||
1861 | def sidedata_cut_off(self, rev): |
|
1864 | def sidedata_cut_off(self, rev): | |
1862 | sd_cut_off = self.index[rev][8] |
|
1865 | sd_cut_off = self.index[rev][8] | |
1863 | if sd_cut_off != 0: |
|
1866 | if sd_cut_off != 0: | |
1864 | return sd_cut_off |
|
1867 | return sd_cut_off | |
1865 | # This is some annoying dance, because entries without sidedata |
|
1868 | # This is some annoying dance, because entries without sidedata | |
1866 | # currently use 0 as their ofsset. (instead of previous-offset + |
|
1869 | # currently use 0 as their ofsset. (instead of previous-offset + | |
1867 | # previous-size) |
|
1870 | # previous-size) | |
1868 | # |
|
1871 | # | |
1869 | # We should reconsider this sidedata → 0 sidata_offset policy. |
|
1872 | # We should reconsider this sidedata → 0 sidata_offset policy. | |
1870 | # In the meantime, we need this. |
|
1873 | # In the meantime, we need this. | |
1871 | while 0 <= rev: |
|
1874 | while 0 <= rev: | |
1872 | e = self.index[rev] |
|
1875 | e = self.index[rev] | |
1873 | if e[9] != 0: |
|
1876 | if e[9] != 0: | |
1874 | return e[8] + e[9] |
|
1877 | return e[8] + e[9] | |
1875 | rev -= 1 |
|
1878 | rev -= 1 | |
1876 | return 0 |
|
1879 | return 0 | |
1877 |
|
1880 | |||
1878 | def flags(self, rev): |
|
1881 | def flags(self, rev): | |
1879 | return self.index[rev][0] & 0xFFFF |
|
1882 | return self.index[rev][0] & 0xFFFF | |
1880 |
|
1883 | |||
1881 | def length(self, rev): |
|
1884 | def length(self, rev): | |
1882 | return self.index[rev][1] |
|
1885 | return self.index[rev][1] | |
1883 |
|
1886 | |||
1884 | def sidedata_length(self, rev): |
|
1887 | def sidedata_length(self, rev): | |
1885 | if not self.feature_config.has_side_data: |
|
1888 | if not self.feature_config.has_side_data: | |
1886 | return 0 |
|
1889 | return 0 | |
1887 | return self.index[rev][9] |
|
1890 | return self.index[rev][9] | |
1888 |
|
1891 | |||
1889 | def rawsize(self, rev): |
|
1892 | def rawsize(self, rev): | |
1890 | """return the length of the uncompressed text for a given revision""" |
|
1893 | """return the length of the uncompressed text for a given revision""" | |
1891 | l = self.index[rev][2] |
|
1894 | l = self.index[rev][2] | |
1892 | if l >= 0: |
|
1895 | if l >= 0: | |
1893 | return l |
|
1896 | return l | |
1894 |
|
1897 | |||
1895 | t = self.rawdata(rev) |
|
1898 | t = self.rawdata(rev) | |
1896 | return len(t) |
|
1899 | return len(t) | |
1897 |
|
1900 | |||
1898 | def size(self, rev): |
|
1901 | def size(self, rev): | |
1899 | """length of non-raw text (processed by a "read" flag processor)""" |
|
1902 | """length of non-raw text (processed by a "read" flag processor)""" | |
1900 | # fast path: if no "read" flag processor could change the content, |
|
1903 | # fast path: if no "read" flag processor could change the content, | |
1901 | # size is rawsize. note: ELLIPSIS is known to not change the content. |
|
1904 | # size is rawsize. note: ELLIPSIS is known to not change the content. | |
1902 | flags = self.flags(rev) |
|
1905 | flags = self.flags(rev) | |
1903 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: |
|
1906 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: | |
1904 | return self.rawsize(rev) |
|
1907 | return self.rawsize(rev) | |
1905 |
|
1908 | |||
1906 | return len(self.revision(rev)) |
|
1909 | return len(self.revision(rev)) | |
1907 |
|
1910 | |||
1908 | def fast_rank(self, rev): |
|
1911 | def fast_rank(self, rev): | |
1909 | """Return the rank of a revision if already known, or None otherwise. |
|
1912 | """Return the rank of a revision if already known, or None otherwise. | |
1910 |
|
1913 | |||
1911 | The rank of a revision is the size of the sub-graph it defines as a |
|
1914 | The rank of a revision is the size of the sub-graph it defines as a | |
1912 | head. Equivalently, the rank of a revision `r` is the size of the set |
|
1915 | head. Equivalently, the rank of a revision `r` is the size of the set | |
1913 | `ancestors(r)`, `r` included. |
|
1916 | `ancestors(r)`, `r` included. | |
1914 |
|
1917 | |||
1915 | This method returns the rank retrieved from the revlog in constant |
|
1918 | This method returns the rank retrieved from the revlog in constant | |
1916 | time. It makes no attempt at computing unknown values for versions of |
|
1919 | time. It makes no attempt at computing unknown values for versions of | |
1917 | the revlog which do not persist the rank. |
|
1920 | the revlog which do not persist the rank. | |
1918 | """ |
|
1921 | """ | |
1919 | rank = self.index[rev][ENTRY_RANK] |
|
1922 | rank = self.index[rev][ENTRY_RANK] | |
1920 | if self._format_version != CHANGELOGV2 or rank == RANK_UNKNOWN: |
|
1923 | if self._format_version != CHANGELOGV2 or rank == RANK_UNKNOWN: | |
1921 | return None |
|
1924 | return None | |
1922 | if rev == nullrev: |
|
1925 | if rev == nullrev: | |
1923 | return 0 # convention |
|
1926 | return 0 # convention | |
1924 | return rank |
|
1927 | return rank | |
1925 |
|
1928 | |||
1926 | def chainbase(self, rev): |
|
1929 | def chainbase(self, rev): | |
1927 | base = self._chainbasecache.get(rev) |
|
1930 | base = self._chainbasecache.get(rev) | |
1928 | if base is not None: |
|
1931 | if base is not None: | |
1929 | return base |
|
1932 | return base | |
1930 |
|
1933 | |||
1931 | index = self.index |
|
1934 | index = self.index | |
1932 | iterrev = rev |
|
1935 | iterrev = rev | |
1933 | base = index[iterrev][3] |
|
1936 | base = index[iterrev][3] | |
1934 | while base != iterrev: |
|
1937 | while base != iterrev: | |
1935 | iterrev = base |
|
1938 | iterrev = base | |
1936 | base = index[iterrev][3] |
|
1939 | base = index[iterrev][3] | |
1937 |
|
1940 | |||
1938 | self._chainbasecache[rev] = base |
|
1941 | self._chainbasecache[rev] = base | |
1939 | return base |
|
1942 | return base | |
1940 |
|
1943 | |||
1941 | def linkrev(self, rev): |
|
1944 | def linkrev(self, rev): | |
1942 | return self.index[rev][4] |
|
1945 | return self.index[rev][4] | |
1943 |
|
1946 | |||
1944 | def parentrevs(self, rev): |
|
1947 | def parentrevs(self, rev): | |
1945 | try: |
|
1948 | try: | |
1946 | entry = self.index[rev] |
|
1949 | entry = self.index[rev] | |
1947 | except IndexError: |
|
1950 | except IndexError: | |
1948 | if rev == wdirrev: |
|
1951 | if rev == wdirrev: | |
1949 | raise error.WdirUnsupported |
|
1952 | raise error.WdirUnsupported | |
1950 | raise |
|
1953 | raise | |
1951 |
|
1954 | |||
1952 | if self.feature_config.canonical_parent_order and entry[5] == nullrev: |
|
1955 | if self.feature_config.canonical_parent_order and entry[5] == nullrev: | |
1953 | return entry[6], entry[5] |
|
1956 | return entry[6], entry[5] | |
1954 | else: |
|
1957 | else: | |
1955 | return entry[5], entry[6] |
|
1958 | return entry[5], entry[6] | |
1956 |
|
1959 | |||
1957 | # fast parentrevs(rev) where rev isn't filtered |
|
1960 | # fast parentrevs(rev) where rev isn't filtered | |
1958 | _uncheckedparentrevs = parentrevs |
|
1961 | _uncheckedparentrevs = parentrevs | |
1959 |
|
1962 | |||
1960 | def node(self, rev): |
|
1963 | def node(self, rev): | |
1961 | try: |
|
1964 | try: | |
1962 | return self.index[rev][7] |
|
1965 | return self.index[rev][7] | |
1963 | except IndexError: |
|
1966 | except IndexError: | |
1964 | if rev == wdirrev: |
|
1967 | if rev == wdirrev: | |
1965 | raise error.WdirUnsupported |
|
1968 | raise error.WdirUnsupported | |
1966 | raise |
|
1969 | raise | |
1967 |
|
1970 | |||
1968 | # Derived from index values. |
|
1971 | # Derived from index values. | |
1969 |
|
1972 | |||
1970 | def end(self, rev): |
|
1973 | def end(self, rev): | |
1971 | return self.start(rev) + self.length(rev) |
|
1974 | return self.start(rev) + self.length(rev) | |
1972 |
|
1975 | |||
1973 | def parents(self, node): |
|
1976 | def parents(self, node): | |
1974 | i = self.index |
|
1977 | i = self.index | |
1975 | d = i[self.rev(node)] |
|
1978 | d = i[self.rev(node)] | |
1976 | # inline node() to avoid function call overhead |
|
1979 | # inline node() to avoid function call overhead | |
1977 | if self.feature_config.canonical_parent_order and d[5] == self.nullid: |
|
1980 | if self.feature_config.canonical_parent_order and d[5] == self.nullid: | |
1978 | return i[d[6]][7], i[d[5]][7] |
|
1981 | return i[d[6]][7], i[d[5]][7] | |
1979 | else: |
|
1982 | else: | |
1980 | return i[d[5]][7], i[d[6]][7] |
|
1983 | return i[d[5]][7], i[d[6]][7] | |
1981 |
|
1984 | |||
1982 | def chainlen(self, rev): |
|
1985 | def chainlen(self, rev): | |
1983 | return self._chaininfo(rev)[0] |
|
1986 | return self._chaininfo(rev)[0] | |
1984 |
|
1987 | |||
1985 | def _chaininfo(self, rev): |
|
1988 | def _chaininfo(self, rev): | |
1986 | chaininfocache = self._chaininfocache |
|
1989 | chaininfocache = self._chaininfocache | |
1987 | if rev in chaininfocache: |
|
1990 | if rev in chaininfocache: | |
1988 | return chaininfocache[rev] |
|
1991 | return chaininfocache[rev] | |
1989 | index = self.index |
|
1992 | index = self.index | |
1990 | generaldelta = self.delta_config.general_delta |
|
1993 | generaldelta = self.delta_config.general_delta | |
1991 | iterrev = rev |
|
1994 | iterrev = rev | |
1992 | e = index[iterrev] |
|
1995 | e = index[iterrev] | |
1993 | clen = 0 |
|
1996 | clen = 0 | |
1994 | compresseddeltalen = 0 |
|
1997 | compresseddeltalen = 0 | |
1995 | while iterrev != e[3]: |
|
1998 | while iterrev != e[3]: | |
1996 | clen += 1 |
|
1999 | clen += 1 | |
1997 | compresseddeltalen += e[1] |
|
2000 | compresseddeltalen += e[1] | |
1998 | if generaldelta: |
|
2001 | if generaldelta: | |
1999 | iterrev = e[3] |
|
2002 | iterrev = e[3] | |
2000 | else: |
|
2003 | else: | |
2001 | iterrev -= 1 |
|
2004 | iterrev -= 1 | |
2002 | if iterrev in chaininfocache: |
|
2005 | if iterrev in chaininfocache: | |
2003 | t = chaininfocache[iterrev] |
|
2006 | t = chaininfocache[iterrev] | |
2004 | clen += t[0] |
|
2007 | clen += t[0] | |
2005 | compresseddeltalen += t[1] |
|
2008 | compresseddeltalen += t[1] | |
2006 | break |
|
2009 | break | |
2007 | e = index[iterrev] |
|
2010 | e = index[iterrev] | |
2008 | else: |
|
2011 | else: | |
2009 | # Add text length of base since decompressing that also takes |
|
2012 | # Add text length of base since decompressing that also takes | |
2010 | # work. For cache hits the length is already included. |
|
2013 | # work. For cache hits the length is already included. | |
2011 | compresseddeltalen += e[1] |
|
2014 | compresseddeltalen += e[1] | |
2012 | r = (clen, compresseddeltalen) |
|
2015 | r = (clen, compresseddeltalen) | |
2013 | chaininfocache[rev] = r |
|
2016 | chaininfocache[rev] = r | |
2014 | return r |
|
2017 | return r | |
2015 |
|
2018 | |||
2016 | def _deltachain(self, rev, stoprev=None): |
|
2019 | def _deltachain(self, rev, stoprev=None): | |
2017 | return self._inner._deltachain(rev, stoprev=stoprev) |
|
2020 | return self._inner._deltachain(rev, stoprev=stoprev) | |
2018 |
|
2021 | |||
2019 | def ancestors(self, revs, stoprev=0, inclusive=False): |
|
2022 | def ancestors(self, revs, stoprev=0, inclusive=False): | |
2020 | """Generate the ancestors of 'revs' in reverse revision order. |
|
2023 | """Generate the ancestors of 'revs' in reverse revision order. | |
2021 | Does not generate revs lower than stoprev. |
|
2024 | Does not generate revs lower than stoprev. | |
2022 |
|
2025 | |||
2023 | See the documentation for ancestor.lazyancestors for more details.""" |
|
2026 | See the documentation for ancestor.lazyancestors for more details.""" | |
2024 |
|
2027 | |||
2025 | # first, make sure start revisions aren't filtered |
|
2028 | # first, make sure start revisions aren't filtered | |
2026 | revs = list(revs) |
|
2029 | revs = list(revs) | |
2027 | checkrev = self.node |
|
2030 | checkrev = self.node | |
2028 | for r in revs: |
|
2031 | for r in revs: | |
2029 | checkrev(r) |
|
2032 | checkrev(r) | |
2030 | # and we're sure ancestors aren't filtered as well |
|
2033 | # and we're sure ancestors aren't filtered as well | |
2031 |
|
2034 | |||
2032 | if rustancestor is not None and self.index.rust_ext_compat: |
|
2035 | if rustancestor is not None and self.index.rust_ext_compat: | |
2033 | lazyancestors = rustancestor.LazyAncestors |
|
2036 | lazyancestors = rustancestor.LazyAncestors | |
2034 | arg = self.index |
|
2037 | arg = self.index | |
2035 | else: |
|
2038 | else: | |
2036 | lazyancestors = ancestor.lazyancestors |
|
2039 | lazyancestors = ancestor.lazyancestors | |
2037 | arg = self._uncheckedparentrevs |
|
2040 | arg = self._uncheckedparentrevs | |
2038 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) |
|
2041 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) | |
2039 |
|
2042 | |||
2040 | def descendants(self, revs): |
|
2043 | def descendants(self, revs): | |
2041 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
2044 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) | |
2042 |
|
2045 | |||
2043 | def findcommonmissing(self, common=None, heads=None): |
|
2046 | def findcommonmissing(self, common=None, heads=None): | |
2044 | """Return a tuple of the ancestors of common and the ancestors of heads |
|
2047 | """Return a tuple of the ancestors of common and the ancestors of heads | |
2045 | that are not ancestors of common. In revset terminology, we return the |
|
2048 | that are not ancestors of common. In revset terminology, we return the | |
2046 | tuple: |
|
2049 | tuple: | |
2047 |
|
2050 | |||
2048 | ::common, (::heads) - (::common) |
|
2051 | ::common, (::heads) - (::common) | |
2049 |
|
2052 | |||
2050 | The list is sorted by revision number, meaning it is |
|
2053 | The list is sorted by revision number, meaning it is | |
2051 | topologically sorted. |
|
2054 | topologically sorted. | |
2052 |
|
2055 | |||
2053 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
2056 | 'heads' and 'common' are both lists of node IDs. If heads is | |
2054 | not supplied, uses all of the revlog's heads. If common is not |
|
2057 | not supplied, uses all of the revlog's heads. If common is not | |
2055 | supplied, uses nullid.""" |
|
2058 | supplied, uses nullid.""" | |
2056 | if common is None: |
|
2059 | if common is None: | |
2057 | common = [self.nullid] |
|
2060 | common = [self.nullid] | |
2058 | if heads is None: |
|
2061 | if heads is None: | |
2059 | heads = self.heads() |
|
2062 | heads = self.heads() | |
2060 |
|
2063 | |||
2061 | common = [self.rev(n) for n in common] |
|
2064 | common = [self.rev(n) for n in common] | |
2062 | heads = [self.rev(n) for n in heads] |
|
2065 | heads = [self.rev(n) for n in heads] | |
2063 |
|
2066 | |||
2064 | # we want the ancestors, but inclusive |
|
2067 | # we want the ancestors, but inclusive | |
2065 | class lazyset: |
|
2068 | class lazyset: | |
2066 | def __init__(self, lazyvalues): |
|
2069 | def __init__(self, lazyvalues): | |
2067 | self.addedvalues = set() |
|
2070 | self.addedvalues = set() | |
2068 | self.lazyvalues = lazyvalues |
|
2071 | self.lazyvalues = lazyvalues | |
2069 |
|
2072 | |||
2070 | def __contains__(self, value): |
|
2073 | def __contains__(self, value): | |
2071 | return value in self.addedvalues or value in self.lazyvalues |
|
2074 | return value in self.addedvalues or value in self.lazyvalues | |
2072 |
|
2075 | |||
2073 | def __iter__(self): |
|
2076 | def __iter__(self): | |
2074 | added = self.addedvalues |
|
2077 | added = self.addedvalues | |
2075 | for r in added: |
|
2078 | for r in added: | |
2076 | yield r |
|
2079 | yield r | |
2077 | for r in self.lazyvalues: |
|
2080 | for r in self.lazyvalues: | |
2078 | if not r in added: |
|
2081 | if not r in added: | |
2079 | yield r |
|
2082 | yield r | |
2080 |
|
2083 | |||
2081 | def add(self, value): |
|
2084 | def add(self, value): | |
2082 | self.addedvalues.add(value) |
|
2085 | self.addedvalues.add(value) | |
2083 |
|
2086 | |||
2084 | def update(self, values): |
|
2087 | def update(self, values): | |
2085 | self.addedvalues.update(values) |
|
2088 | self.addedvalues.update(values) | |
2086 |
|
2089 | |||
2087 | has = lazyset(self.ancestors(common)) |
|
2090 | has = lazyset(self.ancestors(common)) | |
2088 | has.add(nullrev) |
|
2091 | has.add(nullrev) | |
2089 | has.update(common) |
|
2092 | has.update(common) | |
2090 |
|
2093 | |||
2091 | # take all ancestors from heads that aren't in has |
|
2094 | # take all ancestors from heads that aren't in has | |
2092 | missing = set() |
|
2095 | missing = set() | |
2093 | visit = collections.deque(r for r in heads if r not in has) |
|
2096 | visit = collections.deque(r for r in heads if r not in has) | |
2094 | while visit: |
|
2097 | while visit: | |
2095 | r = visit.popleft() |
|
2098 | r = visit.popleft() | |
2096 | if r in missing: |
|
2099 | if r in missing: | |
2097 | continue |
|
2100 | continue | |
2098 | else: |
|
2101 | else: | |
2099 | missing.add(r) |
|
2102 | missing.add(r) | |
2100 | for p in self.parentrevs(r): |
|
2103 | for p in self.parentrevs(r): | |
2101 | if p not in has: |
|
2104 | if p not in has: | |
2102 | visit.append(p) |
|
2105 | visit.append(p) | |
2103 | missing = list(missing) |
|
2106 | missing = list(missing) | |
2104 | missing.sort() |
|
2107 | missing.sort() | |
2105 | return has, [self.node(miss) for miss in missing] |
|
2108 | return has, [self.node(miss) for miss in missing] | |
2106 |
|
2109 | |||
2107 | def incrementalmissingrevs(self, common=None): |
|
2110 | def incrementalmissingrevs(self, common=None): | |
2108 | """Return an object that can be used to incrementally compute the |
|
2111 | """Return an object that can be used to incrementally compute the | |
2109 | revision numbers of the ancestors of arbitrary sets that are not |
|
2112 | revision numbers of the ancestors of arbitrary sets that are not | |
2110 | ancestors of common. This is an ancestor.incrementalmissingancestors |
|
2113 | ancestors of common. This is an ancestor.incrementalmissingancestors | |
2111 | object. |
|
2114 | object. | |
2112 |
|
2115 | |||
2113 | 'common' is a list of revision numbers. If common is not supplied, uses |
|
2116 | 'common' is a list of revision numbers. If common is not supplied, uses | |
2114 | nullrev. |
|
2117 | nullrev. | |
2115 | """ |
|
2118 | """ | |
2116 | if common is None: |
|
2119 | if common is None: | |
2117 | common = [nullrev] |
|
2120 | common = [nullrev] | |
2118 |
|
2121 | |||
2119 | if rustancestor is not None and self.index.rust_ext_compat: |
|
2122 | if rustancestor is not None and self.index.rust_ext_compat: | |
2120 | return rustancestor.MissingAncestors(self.index, common) |
|
2123 | return rustancestor.MissingAncestors(self.index, common) | |
2121 | return ancestor.incrementalmissingancestors(self.parentrevs, common) |
|
2124 | return ancestor.incrementalmissingancestors(self.parentrevs, common) | |
2122 |
|
2125 | |||
2123 | def findmissingrevs(self, common=None, heads=None): |
|
2126 | def findmissingrevs(self, common=None, heads=None): | |
2124 | """Return the revision numbers of the ancestors of heads that |
|
2127 | """Return the revision numbers of the ancestors of heads that | |
2125 | are not ancestors of common. |
|
2128 | are not ancestors of common. | |
2126 |
|
2129 | |||
2127 | More specifically, return a list of revision numbers corresponding to |
|
2130 | More specifically, return a list of revision numbers corresponding to | |
2128 | nodes N such that every N satisfies the following constraints: |
|
2131 | nodes N such that every N satisfies the following constraints: | |
2129 |
|
2132 | |||
2130 | 1. N is an ancestor of some node in 'heads' |
|
2133 | 1. N is an ancestor of some node in 'heads' | |
2131 | 2. N is not an ancestor of any node in 'common' |
|
2134 | 2. N is not an ancestor of any node in 'common' | |
2132 |
|
2135 | |||
2133 | The list is sorted by revision number, meaning it is |
|
2136 | The list is sorted by revision number, meaning it is | |
2134 | topologically sorted. |
|
2137 | topologically sorted. | |
2135 |
|
2138 | |||
2136 | 'heads' and 'common' are both lists of revision numbers. If heads is |
|
2139 | 'heads' and 'common' are both lists of revision numbers. If heads is | |
2137 | not supplied, uses all of the revlog's heads. If common is not |
|
2140 | not supplied, uses all of the revlog's heads. If common is not | |
2138 | supplied, uses nullid.""" |
|
2141 | supplied, uses nullid.""" | |
2139 | if common is None: |
|
2142 | if common is None: | |
2140 | common = [nullrev] |
|
2143 | common = [nullrev] | |
2141 | if heads is None: |
|
2144 | if heads is None: | |
2142 | heads = self.headrevs() |
|
2145 | heads = self.headrevs() | |
2143 |
|
2146 | |||
2144 | inc = self.incrementalmissingrevs(common=common) |
|
2147 | inc = self.incrementalmissingrevs(common=common) | |
2145 | return inc.missingancestors(heads) |
|
2148 | return inc.missingancestors(heads) | |
2146 |
|
2149 | |||
2147 | def findmissing(self, common=None, heads=None): |
|
2150 | def findmissing(self, common=None, heads=None): | |
2148 | """Return the ancestors of heads that are not ancestors of common. |
|
2151 | """Return the ancestors of heads that are not ancestors of common. | |
2149 |
|
2152 | |||
2150 | More specifically, return a list of nodes N such that every N |
|
2153 | More specifically, return a list of nodes N such that every N | |
2151 | satisfies the following constraints: |
|
2154 | satisfies the following constraints: | |
2152 |
|
2155 | |||
2153 | 1. N is an ancestor of some node in 'heads' |
|
2156 | 1. N is an ancestor of some node in 'heads' | |
2154 | 2. N is not an ancestor of any node in 'common' |
|
2157 | 2. N is not an ancestor of any node in 'common' | |
2155 |
|
2158 | |||
2156 | The list is sorted by revision number, meaning it is |
|
2159 | The list is sorted by revision number, meaning it is | |
2157 | topologically sorted. |
|
2160 | topologically sorted. | |
2158 |
|
2161 | |||
2159 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
2162 | 'heads' and 'common' are both lists of node IDs. If heads is | |
2160 | not supplied, uses all of the revlog's heads. If common is not |
|
2163 | not supplied, uses all of the revlog's heads. If common is not | |
2161 | supplied, uses nullid.""" |
|
2164 | supplied, uses nullid.""" | |
2162 | if common is None: |
|
2165 | if common is None: | |
2163 | common = [self.nullid] |
|
2166 | common = [self.nullid] | |
2164 | if heads is None: |
|
2167 | if heads is None: | |
2165 | heads = self.heads() |
|
2168 | heads = self.heads() | |
2166 |
|
2169 | |||
2167 | common = [self.rev(n) for n in common] |
|
2170 | common = [self.rev(n) for n in common] | |
2168 | heads = [self.rev(n) for n in heads] |
|
2171 | heads = [self.rev(n) for n in heads] | |
2169 |
|
2172 | |||
2170 | inc = self.incrementalmissingrevs(common=common) |
|
2173 | inc = self.incrementalmissingrevs(common=common) | |
2171 | return [self.node(r) for r in inc.missingancestors(heads)] |
|
2174 | return [self.node(r) for r in inc.missingancestors(heads)] | |
2172 |
|
2175 | |||
2173 | def nodesbetween(self, roots=None, heads=None): |
|
2176 | def nodesbetween(self, roots=None, heads=None): | |
2174 | """Return a topological path from 'roots' to 'heads'. |
|
2177 | """Return a topological path from 'roots' to 'heads'. | |
2175 |
|
2178 | |||
2176 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a |
|
2179 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a | |
2177 | topologically sorted list of all nodes N that satisfy both of |
|
2180 | topologically sorted list of all nodes N that satisfy both of | |
2178 | these constraints: |
|
2181 | these constraints: | |
2179 |
|
2182 | |||
2180 | 1. N is a descendant of some node in 'roots' |
|
2183 | 1. N is a descendant of some node in 'roots' | |
2181 | 2. N is an ancestor of some node in 'heads' |
|
2184 | 2. N is an ancestor of some node in 'heads' | |
2182 |
|
2185 | |||
2183 | Every node is considered to be both a descendant and an ancestor |
|
2186 | Every node is considered to be both a descendant and an ancestor | |
2184 | of itself, so every reachable node in 'roots' and 'heads' will be |
|
2187 | of itself, so every reachable node in 'roots' and 'heads' will be | |
2185 | included in 'nodes'. |
|
2188 | included in 'nodes'. | |
2186 |
|
2189 | |||
2187 | 'outroots' is the list of reachable nodes in 'roots', i.e., the |
|
2190 | 'outroots' is the list of reachable nodes in 'roots', i.e., the | |
2188 | subset of 'roots' that is returned in 'nodes'. Likewise, |
|
2191 | subset of 'roots' that is returned in 'nodes'. Likewise, | |
2189 | 'outheads' is the subset of 'heads' that is also in 'nodes'. |
|
2192 | 'outheads' is the subset of 'heads' that is also in 'nodes'. | |
2190 |
|
2193 | |||
2191 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is |
|
2194 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is | |
2192 | unspecified, uses nullid as the only root. If 'heads' is |
|
2195 | unspecified, uses nullid as the only root. If 'heads' is | |
2193 | unspecified, uses list of all of the revlog's heads.""" |
|
2196 | unspecified, uses list of all of the revlog's heads.""" | |
2194 | nonodes = ([], [], []) |
|
2197 | nonodes = ([], [], []) | |
2195 | if roots is not None: |
|
2198 | if roots is not None: | |
2196 | roots = list(roots) |
|
2199 | roots = list(roots) | |
2197 | if not roots: |
|
2200 | if not roots: | |
2198 | return nonodes |
|
2201 | return nonodes | |
2199 | lowestrev = min([self.rev(n) for n in roots]) |
|
2202 | lowestrev = min([self.rev(n) for n in roots]) | |
2200 | else: |
|
2203 | else: | |
2201 | roots = [self.nullid] # Everybody's a descendant of nullid |
|
2204 | roots = [self.nullid] # Everybody's a descendant of nullid | |
2202 | lowestrev = nullrev |
|
2205 | lowestrev = nullrev | |
2203 | if (lowestrev == nullrev) and (heads is None): |
|
2206 | if (lowestrev == nullrev) and (heads is None): | |
2204 | # We want _all_ the nodes! |
|
2207 | # We want _all_ the nodes! | |
2205 | return ( |
|
2208 | return ( | |
2206 | [self.node(r) for r in self], |
|
2209 | [self.node(r) for r in self], | |
2207 | [self.nullid], |
|
2210 | [self.nullid], | |
2208 | list(self.heads()), |
|
2211 | list(self.heads()), | |
2209 | ) |
|
2212 | ) | |
2210 | if heads is None: |
|
2213 | if heads is None: | |
2211 | # All nodes are ancestors, so the latest ancestor is the last |
|
2214 | # All nodes are ancestors, so the latest ancestor is the last | |
2212 | # node. |
|
2215 | # node. | |
2213 | highestrev = len(self) - 1 |
|
2216 | highestrev = len(self) - 1 | |
2214 | # Set ancestors to None to signal that every node is an ancestor. |
|
2217 | # Set ancestors to None to signal that every node is an ancestor. | |
2215 | ancestors = None |
|
2218 | ancestors = None | |
2216 | # Set heads to an empty dictionary for later discovery of heads |
|
2219 | # Set heads to an empty dictionary for later discovery of heads | |
2217 | heads = {} |
|
2220 | heads = {} | |
2218 | else: |
|
2221 | else: | |
2219 | heads = list(heads) |
|
2222 | heads = list(heads) | |
2220 | if not heads: |
|
2223 | if not heads: | |
2221 | return nonodes |
|
2224 | return nonodes | |
2222 | ancestors = set() |
|
2225 | ancestors = set() | |
2223 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
2226 | # Turn heads into a dictionary so we can remove 'fake' heads. | |
2224 | # Also, later we will be using it to filter out the heads we can't |
|
2227 | # Also, later we will be using it to filter out the heads we can't | |
2225 | # find from roots. |
|
2228 | # find from roots. | |
2226 | heads = dict.fromkeys(heads, False) |
|
2229 | heads = dict.fromkeys(heads, False) | |
2227 | # Start at the top and keep marking parents until we're done. |
|
2230 | # Start at the top and keep marking parents until we're done. | |
2228 | nodestotag = set(heads) |
|
2231 | nodestotag = set(heads) | |
2229 | # Remember where the top was so we can use it as a limit later. |
|
2232 | # Remember where the top was so we can use it as a limit later. | |
2230 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
2233 | highestrev = max([self.rev(n) for n in nodestotag]) | |
2231 | while nodestotag: |
|
2234 | while nodestotag: | |
2232 | # grab a node to tag |
|
2235 | # grab a node to tag | |
2233 | n = nodestotag.pop() |
|
2236 | n = nodestotag.pop() | |
2234 | # Never tag nullid |
|
2237 | # Never tag nullid | |
2235 | if n == self.nullid: |
|
2238 | if n == self.nullid: | |
2236 | continue |
|
2239 | continue | |
2237 | # A node's revision number represents its place in a |
|
2240 | # A node's revision number represents its place in a | |
2238 | # topologically sorted list of nodes. |
|
2241 | # topologically sorted list of nodes. | |
2239 | r = self.rev(n) |
|
2242 | r = self.rev(n) | |
2240 | if r >= lowestrev: |
|
2243 | if r >= lowestrev: | |
2241 | if n not in ancestors: |
|
2244 | if n not in ancestors: | |
2242 | # If we are possibly a descendant of one of the roots |
|
2245 | # If we are possibly a descendant of one of the roots | |
2243 | # and we haven't already been marked as an ancestor |
|
2246 | # and we haven't already been marked as an ancestor | |
2244 | ancestors.add(n) # Mark as ancestor |
|
2247 | ancestors.add(n) # Mark as ancestor | |
2245 | # Add non-nullid parents to list of nodes to tag. |
|
2248 | # Add non-nullid parents to list of nodes to tag. | |
2246 | nodestotag.update( |
|
2249 | nodestotag.update( | |
2247 | [p for p in self.parents(n) if p != self.nullid] |
|
2250 | [p for p in self.parents(n) if p != self.nullid] | |
2248 | ) |
|
2251 | ) | |
2249 | elif n in heads: # We've seen it before, is it a fake head? |
|
2252 | elif n in heads: # We've seen it before, is it a fake head? | |
2250 | # So it is, real heads should not be the ancestors of |
|
2253 | # So it is, real heads should not be the ancestors of | |
2251 | # any other heads. |
|
2254 | # any other heads. | |
2252 | heads.pop(n) |
|
2255 | heads.pop(n) | |
2253 | if not ancestors: |
|
2256 | if not ancestors: | |
2254 | return nonodes |
|
2257 | return nonodes | |
2255 | # Now that we have our set of ancestors, we want to remove any |
|
2258 | # Now that we have our set of ancestors, we want to remove any | |
2256 | # roots that are not ancestors. |
|
2259 | # roots that are not ancestors. | |
2257 |
|
2260 | |||
2258 | # If one of the roots was nullid, everything is included anyway. |
|
2261 | # If one of the roots was nullid, everything is included anyway. | |
2259 | if lowestrev > nullrev: |
|
2262 | if lowestrev > nullrev: | |
2260 | # But, since we weren't, let's recompute the lowest rev to not |
|
2263 | # But, since we weren't, let's recompute the lowest rev to not | |
2261 | # include roots that aren't ancestors. |
|
2264 | # include roots that aren't ancestors. | |
2262 |
|
2265 | |||
2263 | # Filter out roots that aren't ancestors of heads |
|
2266 | # Filter out roots that aren't ancestors of heads | |
2264 | roots = [root for root in roots if root in ancestors] |
|
2267 | roots = [root for root in roots if root in ancestors] | |
2265 | # Recompute the lowest revision |
|
2268 | # Recompute the lowest revision | |
2266 | if roots: |
|
2269 | if roots: | |
2267 | lowestrev = min([self.rev(root) for root in roots]) |
|
2270 | lowestrev = min([self.rev(root) for root in roots]) | |
2268 | else: |
|
2271 | else: | |
2269 | # No more roots? Return empty list |
|
2272 | # No more roots? Return empty list | |
2270 | return nonodes |
|
2273 | return nonodes | |
2271 | else: |
|
2274 | else: | |
2272 | # We are descending from nullid, and don't need to care about |
|
2275 | # We are descending from nullid, and don't need to care about | |
2273 | # any other roots. |
|
2276 | # any other roots. | |
2274 | lowestrev = nullrev |
|
2277 | lowestrev = nullrev | |
2275 | roots = [self.nullid] |
|
2278 | roots = [self.nullid] | |
2276 | # Transform our roots list into a set. |
|
2279 | # Transform our roots list into a set. | |
2277 | descendants = set(roots) |
|
2280 | descendants = set(roots) | |
2278 | # Also, keep the original roots so we can filter out roots that aren't |
|
2281 | # Also, keep the original roots so we can filter out roots that aren't | |
2279 | # 'real' roots (i.e. are descended from other roots). |
|
2282 | # 'real' roots (i.e. are descended from other roots). | |
2280 | roots = descendants.copy() |
|
2283 | roots = descendants.copy() | |
2281 | # Our topologically sorted list of output nodes. |
|
2284 | # Our topologically sorted list of output nodes. | |
2282 | orderedout = [] |
|
2285 | orderedout = [] | |
2283 | # Don't start at nullid since we don't want nullid in our output list, |
|
2286 | # Don't start at nullid since we don't want nullid in our output list, | |
2284 | # and if nullid shows up in descendants, empty parents will look like |
|
2287 | # and if nullid shows up in descendants, empty parents will look like | |
2285 | # they're descendants. |
|
2288 | # they're descendants. | |
2286 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): |
|
2289 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): | |
2287 | n = self.node(r) |
|
2290 | n = self.node(r) | |
2288 | isdescendant = False |
|
2291 | isdescendant = False | |
2289 | if lowestrev == nullrev: # Everybody is a descendant of nullid |
|
2292 | if lowestrev == nullrev: # Everybody is a descendant of nullid | |
2290 | isdescendant = True |
|
2293 | isdescendant = True | |
2291 | elif n in descendants: |
|
2294 | elif n in descendants: | |
2292 | # n is already a descendant |
|
2295 | # n is already a descendant | |
2293 | isdescendant = True |
|
2296 | isdescendant = True | |
2294 | # This check only needs to be done here because all the roots |
|
2297 | # This check only needs to be done here because all the roots | |
2295 | # will start being marked is descendants before the loop. |
|
2298 | # will start being marked is descendants before the loop. | |
2296 | if n in roots: |
|
2299 | if n in roots: | |
2297 | # If n was a root, check if it's a 'real' root. |
|
2300 | # If n was a root, check if it's a 'real' root. | |
2298 | p = tuple(self.parents(n)) |
|
2301 | p = tuple(self.parents(n)) | |
2299 | # If any of its parents are descendants, it's not a root. |
|
2302 | # If any of its parents are descendants, it's not a root. | |
2300 | if (p[0] in descendants) or (p[1] in descendants): |
|
2303 | if (p[0] in descendants) or (p[1] in descendants): | |
2301 | roots.remove(n) |
|
2304 | roots.remove(n) | |
2302 | else: |
|
2305 | else: | |
2303 | p = tuple(self.parents(n)) |
|
2306 | p = tuple(self.parents(n)) | |
2304 | # A node is a descendant if either of its parents are |
|
2307 | # A node is a descendant if either of its parents are | |
2305 | # descendants. (We seeded the dependents list with the roots |
|
2308 | # descendants. (We seeded the dependents list with the roots | |
2306 | # up there, remember?) |
|
2309 | # up there, remember?) | |
2307 | if (p[0] in descendants) or (p[1] in descendants): |
|
2310 | if (p[0] in descendants) or (p[1] in descendants): | |
2308 | descendants.add(n) |
|
2311 | descendants.add(n) | |
2309 | isdescendant = True |
|
2312 | isdescendant = True | |
2310 | if isdescendant and ((ancestors is None) or (n in ancestors)): |
|
2313 | if isdescendant and ((ancestors is None) or (n in ancestors)): | |
2311 | # Only include nodes that are both descendants and ancestors. |
|
2314 | # Only include nodes that are both descendants and ancestors. | |
2312 | orderedout.append(n) |
|
2315 | orderedout.append(n) | |
2313 | if (ancestors is not None) and (n in heads): |
|
2316 | if (ancestors is not None) and (n in heads): | |
2314 | # We're trying to figure out which heads are reachable |
|
2317 | # We're trying to figure out which heads are reachable | |
2315 | # from roots. |
|
2318 | # from roots. | |
2316 | # Mark this head as having been reached |
|
2319 | # Mark this head as having been reached | |
2317 | heads[n] = True |
|
2320 | heads[n] = True | |
2318 | elif ancestors is None: |
|
2321 | elif ancestors is None: | |
2319 | # Otherwise, we're trying to discover the heads. |
|
2322 | # Otherwise, we're trying to discover the heads. | |
2320 | # Assume this is a head because if it isn't, the next step |
|
2323 | # Assume this is a head because if it isn't, the next step | |
2321 | # will eventually remove it. |
|
2324 | # will eventually remove it. | |
2322 | heads[n] = True |
|
2325 | heads[n] = True | |
2323 | # But, obviously its parents aren't. |
|
2326 | # But, obviously its parents aren't. | |
2324 | for p in self.parents(n): |
|
2327 | for p in self.parents(n): | |
2325 | heads.pop(p, None) |
|
2328 | heads.pop(p, None) | |
2326 | heads = [head for head, flag in heads.items() if flag] |
|
2329 | heads = [head for head, flag in heads.items() if flag] | |
2327 | roots = list(roots) |
|
2330 | roots = list(roots) | |
2328 | assert orderedout |
|
2331 | assert orderedout | |
2329 | assert roots |
|
2332 | assert roots | |
2330 | assert heads |
|
2333 | assert heads | |
2331 | return (orderedout, roots, heads) |
|
2334 | return (orderedout, roots, heads) | |
2332 |
|
2335 | |||
2333 | def headrevs(self, revs=None): |
|
2336 | def headrevs(self, revs=None): | |
2334 | if revs is None: |
|
2337 | if revs is None: | |
2335 | try: |
|
2338 | try: | |
2336 | return self.index.headrevs() |
|
2339 | return self.index.headrevs() | |
2337 | except AttributeError: |
|
2340 | except AttributeError: | |
2338 | return self._headrevs() |
|
2341 | return self._headrevs() | |
2339 | if rustdagop is not None and self.index.rust_ext_compat: |
|
2342 | if rustdagop is not None and self.index.rust_ext_compat: | |
2340 | return rustdagop.headrevs(self.index, revs) |
|
2343 | return rustdagop.headrevs(self.index, revs) | |
2341 | return dagop.headrevs(revs, self._uncheckedparentrevs) |
|
2344 | return dagop.headrevs(revs, self._uncheckedparentrevs) | |
2342 |
|
2345 | |||
2343 | def computephases(self, roots): |
|
2346 | def computephases(self, roots): | |
2344 | return self.index.computephasesmapsets(roots) |
|
2347 | return self.index.computephasesmapsets(roots) | |
2345 |
|
2348 | |||
2346 | def _headrevs(self): |
|
2349 | def _headrevs(self): | |
2347 | count = len(self) |
|
2350 | count = len(self) | |
2348 | if not count: |
|
2351 | if not count: | |
2349 | return [nullrev] |
|
2352 | return [nullrev] | |
2350 | # we won't iter over filtered rev so nobody is a head at start |
|
2353 | # we won't iter over filtered rev so nobody is a head at start | |
2351 | ishead = [0] * (count + 1) |
|
2354 | ishead = [0] * (count + 1) | |
2352 | index = self.index |
|
2355 | index = self.index | |
2353 | for r in self: |
|
2356 | for r in self: | |
2354 | ishead[r] = 1 # I may be an head |
|
2357 | ishead[r] = 1 # I may be an head | |
2355 | e = index[r] |
|
2358 | e = index[r] | |
2356 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not |
|
2359 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not | |
2357 | return [r for r, val in enumerate(ishead) if val] |
|
2360 | return [r for r, val in enumerate(ishead) if val] | |
2358 |
|
2361 | |||
2359 | def heads(self, start=None, stop=None): |
|
2362 | def heads(self, start=None, stop=None): | |
2360 | """return the list of all nodes that have no children |
|
2363 | """return the list of all nodes that have no children | |
2361 |
|
2364 | |||
2362 | if start is specified, only heads that are descendants of |
|
2365 | if start is specified, only heads that are descendants of | |
2363 | start will be returned |
|
2366 | start will be returned | |
2364 | if stop is specified, it will consider all the revs from stop |
|
2367 | if stop is specified, it will consider all the revs from stop | |
2365 | as if they had no children |
|
2368 | as if they had no children | |
2366 | """ |
|
2369 | """ | |
2367 | if start is None and stop is None: |
|
2370 | if start is None and stop is None: | |
2368 | if not len(self): |
|
2371 | if not len(self): | |
2369 | return [self.nullid] |
|
2372 | return [self.nullid] | |
2370 | return [self.node(r) for r in self.headrevs()] |
|
2373 | return [self.node(r) for r in self.headrevs()] | |
2371 |
|
2374 | |||
2372 | if start is None: |
|
2375 | if start is None: | |
2373 | start = nullrev |
|
2376 | start = nullrev | |
2374 | else: |
|
2377 | else: | |
2375 | start = self.rev(start) |
|
2378 | start = self.rev(start) | |
2376 |
|
2379 | |||
2377 | stoprevs = {self.rev(n) for n in stop or []} |
|
2380 | stoprevs = {self.rev(n) for n in stop or []} | |
2378 |
|
2381 | |||
2379 | revs = dagop.headrevssubset( |
|
2382 | revs = dagop.headrevssubset( | |
2380 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs |
|
2383 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs | |
2381 | ) |
|
2384 | ) | |
2382 |
|
2385 | |||
2383 | return [self.node(rev) for rev in revs] |
|
2386 | return [self.node(rev) for rev in revs] | |
2384 |
|
2387 | |||
2385 | def children(self, node): |
|
2388 | def children(self, node): | |
2386 | """find the children of a given node""" |
|
2389 | """find the children of a given node""" | |
2387 | c = [] |
|
2390 | c = [] | |
2388 | p = self.rev(node) |
|
2391 | p = self.rev(node) | |
2389 | for r in self.revs(start=p + 1): |
|
2392 | for r in self.revs(start=p + 1): | |
2390 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
2393 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
2391 | if prevs: |
|
2394 | if prevs: | |
2392 | for pr in prevs: |
|
2395 | for pr in prevs: | |
2393 | if pr == p: |
|
2396 | if pr == p: | |
2394 | c.append(self.node(r)) |
|
2397 | c.append(self.node(r)) | |
2395 | elif p == nullrev: |
|
2398 | elif p == nullrev: | |
2396 | c.append(self.node(r)) |
|
2399 | c.append(self.node(r)) | |
2397 | return c |
|
2400 | return c | |
2398 |
|
2401 | |||
2399 | def commonancestorsheads(self, a, b): |
|
2402 | def commonancestorsheads(self, a, b): | |
2400 | """calculate all the heads of the common ancestors of nodes a and b""" |
|
2403 | """calculate all the heads of the common ancestors of nodes a and b""" | |
2401 | a, b = self.rev(a), self.rev(b) |
|
2404 | a, b = self.rev(a), self.rev(b) | |
2402 | ancs = self._commonancestorsheads(a, b) |
|
2405 | ancs = self._commonancestorsheads(a, b) | |
2403 | return pycompat.maplist(self.node, ancs) |
|
2406 | return pycompat.maplist(self.node, ancs) | |
2404 |
|
2407 | |||
2405 | def _commonancestorsheads(self, *revs): |
|
2408 | def _commonancestorsheads(self, *revs): | |
2406 | """calculate all the heads of the common ancestors of revs""" |
|
2409 | """calculate all the heads of the common ancestors of revs""" | |
2407 | try: |
|
2410 | try: | |
2408 | ancs = self.index.commonancestorsheads(*revs) |
|
2411 | ancs = self.index.commonancestorsheads(*revs) | |
2409 | except (AttributeError, OverflowError): # C implementation failed |
|
2412 | except (AttributeError, OverflowError): # C implementation failed | |
2410 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) |
|
2413 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) | |
2411 | return ancs |
|
2414 | return ancs | |
2412 |
|
2415 | |||
2413 | def isancestor(self, a, b): |
|
2416 | def isancestor(self, a, b): | |
2414 | """return True if node a is an ancestor of node b |
|
2417 | """return True if node a is an ancestor of node b | |
2415 |
|
2418 | |||
2416 | A revision is considered an ancestor of itself.""" |
|
2419 | A revision is considered an ancestor of itself.""" | |
2417 | a, b = self.rev(a), self.rev(b) |
|
2420 | a, b = self.rev(a), self.rev(b) | |
2418 | return self.isancestorrev(a, b) |
|
2421 | return self.isancestorrev(a, b) | |
2419 |
|
2422 | |||
2420 | def isancestorrev(self, a, b): |
|
2423 | def isancestorrev(self, a, b): | |
2421 | """return True if revision a is an ancestor of revision b |
|
2424 | """return True if revision a is an ancestor of revision b | |
2422 |
|
2425 | |||
2423 | A revision is considered an ancestor of itself. |
|
2426 | A revision is considered an ancestor of itself. | |
2424 |
|
2427 | |||
2425 | The implementation of this is trivial but the use of |
|
2428 | The implementation of this is trivial but the use of | |
2426 | reachableroots is not.""" |
|
2429 | reachableroots is not.""" | |
2427 | if a == nullrev: |
|
2430 | if a == nullrev: | |
2428 | return True |
|
2431 | return True | |
2429 | elif a == b: |
|
2432 | elif a == b: | |
2430 | return True |
|
2433 | return True | |
2431 | elif a > b: |
|
2434 | elif a > b: | |
2432 | return False |
|
2435 | return False | |
2433 | return bool(self.reachableroots(a, [b], [a], includepath=False)) |
|
2436 | return bool(self.reachableroots(a, [b], [a], includepath=False)) | |
2434 |
|
2437 | |||
2435 | def reachableroots(self, minroot, heads, roots, includepath=False): |
|
2438 | def reachableroots(self, minroot, heads, roots, includepath=False): | |
2436 | """return (heads(::(<roots> and <roots>::<heads>))) |
|
2439 | """return (heads(::(<roots> and <roots>::<heads>))) | |
2437 |
|
2440 | |||
2438 | If includepath is True, return (<roots>::<heads>).""" |
|
2441 | If includepath is True, return (<roots>::<heads>).""" | |
2439 | try: |
|
2442 | try: | |
2440 | return self.index.reachableroots2( |
|
2443 | return self.index.reachableroots2( | |
2441 | minroot, heads, roots, includepath |
|
2444 | minroot, heads, roots, includepath | |
2442 | ) |
|
2445 | ) | |
2443 | except AttributeError: |
|
2446 | except AttributeError: | |
2444 | return dagop._reachablerootspure( |
|
2447 | return dagop._reachablerootspure( | |
2445 | self.parentrevs, minroot, roots, heads, includepath |
|
2448 | self.parentrevs, minroot, roots, heads, includepath | |
2446 | ) |
|
2449 | ) | |
2447 |
|
2450 | |||
2448 | def ancestor(self, a, b): |
|
2451 | def ancestor(self, a, b): | |
2449 | """calculate the "best" common ancestor of nodes a and b""" |
|
2452 | """calculate the "best" common ancestor of nodes a and b""" | |
2450 |
|
2453 | |||
2451 | a, b = self.rev(a), self.rev(b) |
|
2454 | a, b = self.rev(a), self.rev(b) | |
2452 | try: |
|
2455 | try: | |
2453 | ancs = self.index.ancestors(a, b) |
|
2456 | ancs = self.index.ancestors(a, b) | |
2454 | except (AttributeError, OverflowError): |
|
2457 | except (AttributeError, OverflowError): | |
2455 | ancs = ancestor.ancestors(self.parentrevs, a, b) |
|
2458 | ancs = ancestor.ancestors(self.parentrevs, a, b) | |
2456 | if ancs: |
|
2459 | if ancs: | |
2457 | # choose a consistent winner when there's a tie |
|
2460 | # choose a consistent winner when there's a tie | |
2458 | return min(map(self.node, ancs)) |
|
2461 | return min(map(self.node, ancs)) | |
2459 | return self.nullid |
|
2462 | return self.nullid | |
2460 |
|
2463 | |||
2461 | def _match(self, id): |
|
2464 | def _match(self, id): | |
2462 | if isinstance(id, int): |
|
2465 | if isinstance(id, int): | |
2463 | # rev |
|
2466 | # rev | |
2464 | return self.node(id) |
|
2467 | return self.node(id) | |
2465 | if len(id) == self.nodeconstants.nodelen: |
|
2468 | if len(id) == self.nodeconstants.nodelen: | |
2466 | # possibly a binary node |
|
2469 | # possibly a binary node | |
2467 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
2470 | # odds of a binary node being all hex in ASCII are 1 in 10**25 | |
2468 | try: |
|
2471 | try: | |
2469 | node = id |
|
2472 | node = id | |
2470 | self.rev(node) # quick search the index |
|
2473 | self.rev(node) # quick search the index | |
2471 | return node |
|
2474 | return node | |
2472 | except error.LookupError: |
|
2475 | except error.LookupError: | |
2473 | pass # may be partial hex id |
|
2476 | pass # may be partial hex id | |
2474 | try: |
|
2477 | try: | |
2475 | # str(rev) |
|
2478 | # str(rev) | |
2476 | rev = int(id) |
|
2479 | rev = int(id) | |
2477 | if b"%d" % rev != id: |
|
2480 | if b"%d" % rev != id: | |
2478 | raise ValueError |
|
2481 | raise ValueError | |
2479 | if rev < 0: |
|
2482 | if rev < 0: | |
2480 | rev = len(self) + rev |
|
2483 | rev = len(self) + rev | |
2481 | if rev < 0 or rev >= len(self): |
|
2484 | if rev < 0 or rev >= len(self): | |
2482 | raise ValueError |
|
2485 | raise ValueError | |
2483 | return self.node(rev) |
|
2486 | return self.node(rev) | |
2484 | except (ValueError, OverflowError): |
|
2487 | except (ValueError, OverflowError): | |
2485 | pass |
|
2488 | pass | |
2486 | if len(id) == 2 * self.nodeconstants.nodelen: |
|
2489 | if len(id) == 2 * self.nodeconstants.nodelen: | |
2487 | try: |
|
2490 | try: | |
2488 | # a full hex nodeid? |
|
2491 | # a full hex nodeid? | |
2489 | node = bin(id) |
|
2492 | node = bin(id) | |
2490 | self.rev(node) |
|
2493 | self.rev(node) | |
2491 | return node |
|
2494 | return node | |
2492 | except (binascii.Error, error.LookupError): |
|
2495 | except (binascii.Error, error.LookupError): | |
2493 | pass |
|
2496 | pass | |
2494 |
|
2497 | |||
2495 | def _partialmatch(self, id): |
|
2498 | def _partialmatch(self, id): | |
2496 | # we don't care wdirfilenodeids as they should be always full hash |
|
2499 | # we don't care wdirfilenodeids as they should be always full hash | |
2497 | maybewdir = self.nodeconstants.wdirhex.startswith(id) |
|
2500 | maybewdir = self.nodeconstants.wdirhex.startswith(id) | |
2498 | ambiguous = False |
|
2501 | ambiguous = False | |
2499 | try: |
|
2502 | try: | |
2500 | partial = self.index.partialmatch(id) |
|
2503 | partial = self.index.partialmatch(id) | |
2501 | if partial and self.hasnode(partial): |
|
2504 | if partial and self.hasnode(partial): | |
2502 | if maybewdir: |
|
2505 | if maybewdir: | |
2503 | # single 'ff...' match in radix tree, ambiguous with wdir |
|
2506 | # single 'ff...' match in radix tree, ambiguous with wdir | |
2504 | ambiguous = True |
|
2507 | ambiguous = True | |
2505 | else: |
|
2508 | else: | |
2506 | return partial |
|
2509 | return partial | |
2507 | elif maybewdir: |
|
2510 | elif maybewdir: | |
2508 | # no 'ff...' match in radix tree, wdir identified |
|
2511 | # no 'ff...' match in radix tree, wdir identified | |
2509 | raise error.WdirUnsupported |
|
2512 | raise error.WdirUnsupported | |
2510 | else: |
|
2513 | else: | |
2511 | return None |
|
2514 | return None | |
2512 | except error.RevlogError: |
|
2515 | except error.RevlogError: | |
2513 | # parsers.c radix tree lookup gave multiple matches |
|
2516 | # parsers.c radix tree lookup gave multiple matches | |
2514 | # fast path: for unfiltered changelog, radix tree is accurate |
|
2517 | # fast path: for unfiltered changelog, radix tree is accurate | |
2515 | if not getattr(self, 'filteredrevs', None): |
|
2518 | if not getattr(self, 'filteredrevs', None): | |
2516 | ambiguous = True |
|
2519 | ambiguous = True | |
2517 | # fall through to slow path that filters hidden revisions |
|
2520 | # fall through to slow path that filters hidden revisions | |
2518 | except (AttributeError, ValueError): |
|
2521 | except (AttributeError, ValueError): | |
2519 | # we are pure python, or key is not hex |
|
2522 | # we are pure python, or key is not hex | |
2520 | pass |
|
2523 | pass | |
2521 | if ambiguous: |
|
2524 | if ambiguous: | |
2522 | raise error.AmbiguousPrefixLookupError( |
|
2525 | raise error.AmbiguousPrefixLookupError( | |
2523 | id, self.display_id, _(b'ambiguous identifier') |
|
2526 | id, self.display_id, _(b'ambiguous identifier') | |
2524 | ) |
|
2527 | ) | |
2525 |
|
2528 | |||
2526 | if id in self._pcache: |
|
2529 | if id in self._pcache: | |
2527 | return self._pcache[id] |
|
2530 | return self._pcache[id] | |
2528 |
|
2531 | |||
2529 | if len(id) <= 40: |
|
2532 | if len(id) <= 40: | |
2530 | # hex(node)[:...] |
|
2533 | # hex(node)[:...] | |
2531 | l = len(id) // 2 * 2 # grab an even number of digits |
|
2534 | l = len(id) // 2 * 2 # grab an even number of digits | |
2532 | try: |
|
2535 | try: | |
2533 | # we're dropping the last digit, so let's check that it's hex, |
|
2536 | # we're dropping the last digit, so let's check that it's hex, | |
2534 | # to avoid the expensive computation below if it's not |
|
2537 | # to avoid the expensive computation below if it's not | |
2535 | if len(id) % 2 > 0: |
|
2538 | if len(id) % 2 > 0: | |
2536 | if not (id[-1] in hexdigits): |
|
2539 | if not (id[-1] in hexdigits): | |
2537 | return None |
|
2540 | return None | |
2538 | prefix = bin(id[:l]) |
|
2541 | prefix = bin(id[:l]) | |
2539 | except binascii.Error: |
|
2542 | except binascii.Error: | |
2540 | pass |
|
2543 | pass | |
2541 | else: |
|
2544 | else: | |
2542 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] |
|
2545 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] | |
2543 | nl = [ |
|
2546 | nl = [ | |
2544 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) |
|
2547 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) | |
2545 | ] |
|
2548 | ] | |
2546 | if self.nodeconstants.nullhex.startswith(id): |
|
2549 | if self.nodeconstants.nullhex.startswith(id): | |
2547 | nl.append(self.nullid) |
|
2550 | nl.append(self.nullid) | |
2548 | if len(nl) > 0: |
|
2551 | if len(nl) > 0: | |
2549 | if len(nl) == 1 and not maybewdir: |
|
2552 | if len(nl) == 1 and not maybewdir: | |
2550 | self._pcache[id] = nl[0] |
|
2553 | self._pcache[id] = nl[0] | |
2551 | return nl[0] |
|
2554 | return nl[0] | |
2552 | raise error.AmbiguousPrefixLookupError( |
|
2555 | raise error.AmbiguousPrefixLookupError( | |
2553 | id, self.display_id, _(b'ambiguous identifier') |
|
2556 | id, self.display_id, _(b'ambiguous identifier') | |
2554 | ) |
|
2557 | ) | |
2555 | if maybewdir: |
|
2558 | if maybewdir: | |
2556 | raise error.WdirUnsupported |
|
2559 | raise error.WdirUnsupported | |
2557 | return None |
|
2560 | return None | |
2558 |
|
2561 | |||
2559 | def lookup(self, id): |
|
2562 | def lookup(self, id): | |
2560 | """locate a node based on: |
|
2563 | """locate a node based on: | |
2561 | - revision number or str(revision number) |
|
2564 | - revision number or str(revision number) | |
2562 | - nodeid or subset of hex nodeid |
|
2565 | - nodeid or subset of hex nodeid | |
2563 | """ |
|
2566 | """ | |
2564 | n = self._match(id) |
|
2567 | n = self._match(id) | |
2565 | if n is not None: |
|
2568 | if n is not None: | |
2566 | return n |
|
2569 | return n | |
2567 | n = self._partialmatch(id) |
|
2570 | n = self._partialmatch(id) | |
2568 | if n: |
|
2571 | if n: | |
2569 | return n |
|
2572 | return n | |
2570 |
|
2573 | |||
2571 | raise error.LookupError(id, self.display_id, _(b'no match found')) |
|
2574 | raise error.LookupError(id, self.display_id, _(b'no match found')) | |
2572 |
|
2575 | |||
2573 | def shortest(self, node, minlength=1): |
|
2576 | def shortest(self, node, minlength=1): | |
2574 | """Find the shortest unambiguous prefix that matches node.""" |
|
2577 | """Find the shortest unambiguous prefix that matches node.""" | |
2575 |
|
2578 | |||
2576 | def isvalid(prefix): |
|
2579 | def isvalid(prefix): | |
2577 | try: |
|
2580 | try: | |
2578 | matchednode = self._partialmatch(prefix) |
|
2581 | matchednode = self._partialmatch(prefix) | |
2579 | except error.AmbiguousPrefixLookupError: |
|
2582 | except error.AmbiguousPrefixLookupError: | |
2580 | return False |
|
2583 | return False | |
2581 | except error.WdirUnsupported: |
|
2584 | except error.WdirUnsupported: | |
2582 | # single 'ff...' match |
|
2585 | # single 'ff...' match | |
2583 | return True |
|
2586 | return True | |
2584 | if matchednode is None: |
|
2587 | if matchednode is None: | |
2585 | raise error.LookupError(node, self.display_id, _(b'no node')) |
|
2588 | raise error.LookupError(node, self.display_id, _(b'no node')) | |
2586 | return True |
|
2589 | return True | |
2587 |
|
2590 | |||
2588 | def maybewdir(prefix): |
|
2591 | def maybewdir(prefix): | |
2589 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) |
|
2592 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) | |
2590 |
|
2593 | |||
2591 | hexnode = hex(node) |
|
2594 | hexnode = hex(node) | |
2592 |
|
2595 | |||
2593 | def disambiguate(hexnode, minlength): |
|
2596 | def disambiguate(hexnode, minlength): | |
2594 | """Disambiguate against wdirid.""" |
|
2597 | """Disambiguate against wdirid.""" | |
2595 | for length in range(minlength, len(hexnode) + 1): |
|
2598 | for length in range(minlength, len(hexnode) + 1): | |
2596 | prefix = hexnode[:length] |
|
2599 | prefix = hexnode[:length] | |
2597 | if not maybewdir(prefix): |
|
2600 | if not maybewdir(prefix): | |
2598 | return prefix |
|
2601 | return prefix | |
2599 |
|
2602 | |||
2600 | if not getattr(self, 'filteredrevs', None): |
|
2603 | if not getattr(self, 'filteredrevs', None): | |
2601 | try: |
|
2604 | try: | |
2602 | length = max(self.index.shortest(node), minlength) |
|
2605 | length = max(self.index.shortest(node), minlength) | |
2603 | return disambiguate(hexnode, length) |
|
2606 | return disambiguate(hexnode, length) | |
2604 | except error.RevlogError: |
|
2607 | except error.RevlogError: | |
2605 | if node != self.nodeconstants.wdirid: |
|
2608 | if node != self.nodeconstants.wdirid: | |
2606 | raise error.LookupError( |
|
2609 | raise error.LookupError( | |
2607 | node, self.display_id, _(b'no node') |
|
2610 | node, self.display_id, _(b'no node') | |
2608 | ) |
|
2611 | ) | |
2609 | except AttributeError: |
|
2612 | except AttributeError: | |
2610 | # Fall through to pure code |
|
2613 | # Fall through to pure code | |
2611 | pass |
|
2614 | pass | |
2612 |
|
2615 | |||
2613 | if node == self.nodeconstants.wdirid: |
|
2616 | if node == self.nodeconstants.wdirid: | |
2614 | for length in range(minlength, len(hexnode) + 1): |
|
2617 | for length in range(minlength, len(hexnode) + 1): | |
2615 | prefix = hexnode[:length] |
|
2618 | prefix = hexnode[:length] | |
2616 | if isvalid(prefix): |
|
2619 | if isvalid(prefix): | |
2617 | return prefix |
|
2620 | return prefix | |
2618 |
|
2621 | |||
2619 | for length in range(minlength, len(hexnode) + 1): |
|
2622 | for length in range(minlength, len(hexnode) + 1): | |
2620 | prefix = hexnode[:length] |
|
2623 | prefix = hexnode[:length] | |
2621 | if isvalid(prefix): |
|
2624 | if isvalid(prefix): | |
2622 | return disambiguate(hexnode, length) |
|
2625 | return disambiguate(hexnode, length) | |
2623 |
|
2626 | |||
2624 | def cmp(self, node, text): |
|
2627 | def cmp(self, node, text): | |
2625 | """compare text with a given file revision |
|
2628 | """compare text with a given file revision | |
2626 |
|
2629 | |||
2627 | returns True if text is different than what is stored. |
|
2630 | returns True if text is different than what is stored. | |
2628 | """ |
|
2631 | """ | |
2629 | p1, p2 = self.parents(node) |
|
2632 | p1, p2 = self.parents(node) | |
2630 | return storageutil.hashrevisionsha1(text, p1, p2) != node |
|
2633 | return storageutil.hashrevisionsha1(text, p1, p2) != node | |
2631 |
|
2634 | |||
2632 | def deltaparent(self, rev): |
|
2635 | def deltaparent(self, rev): | |
2633 | """return deltaparent of the given revision""" |
|
2636 | """return deltaparent of the given revision""" | |
2634 | base = self.index[rev][3] |
|
2637 | base = self.index[rev][3] | |
2635 | if base == rev: |
|
2638 | if base == rev: | |
2636 | return nullrev |
|
2639 | return nullrev | |
2637 | elif self.delta_config.general_delta: |
|
2640 | elif self.delta_config.general_delta: | |
2638 | return base |
|
2641 | return base | |
2639 | else: |
|
2642 | else: | |
2640 | return rev - 1 |
|
2643 | return rev - 1 | |
2641 |
|
2644 | |||
2642 | def issnapshot(self, rev): |
|
2645 | def issnapshot(self, rev): | |
2643 | """tells whether rev is a snapshot""" |
|
2646 | """tells whether rev is a snapshot""" | |
2644 | ret = self._inner.issnapshot(rev) |
|
2647 | ret = self._inner.issnapshot(rev) | |
2645 | self.issnapshot = self._inner.issnapshot |
|
2648 | self.issnapshot = self._inner.issnapshot | |
2646 | return ret |
|
2649 | return ret | |
2647 |
|
2650 | |||
2648 | def snapshotdepth(self, rev): |
|
2651 | def snapshotdepth(self, rev): | |
2649 | """number of snapshot in the chain before this one""" |
|
2652 | """number of snapshot in the chain before this one""" | |
2650 | if not self.issnapshot(rev): |
|
2653 | if not self.issnapshot(rev): | |
2651 | raise error.ProgrammingError(b'revision %d not a snapshot') |
|
2654 | raise error.ProgrammingError(b'revision %d not a snapshot') | |
2652 | return len(self._inner._deltachain(rev)[0]) - 1 |
|
2655 | return len(self._inner._deltachain(rev)[0]) - 1 | |
2653 |
|
2656 | |||
2654 | def revdiff(self, rev1, rev2): |
|
2657 | def revdiff(self, rev1, rev2): | |
2655 | """return or calculate a delta between two revisions |
|
2658 | """return or calculate a delta between two revisions | |
2656 |
|
2659 | |||
2657 | The delta calculated is in binary form and is intended to be written to |
|
2660 | The delta calculated is in binary form and is intended to be written to | |
2658 | revlog data directly. So this function needs raw revision data. |
|
2661 | revlog data directly. So this function needs raw revision data. | |
2659 | """ |
|
2662 | """ | |
2660 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: |
|
2663 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: | |
2661 | return bytes(self._inner._chunk(rev2)) |
|
2664 | return bytes(self._inner._chunk(rev2)) | |
2662 |
|
2665 | |||
2663 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) |
|
2666 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) | |
2664 |
|
2667 | |||
2665 | def revision(self, nodeorrev): |
|
2668 | def revision(self, nodeorrev): | |
2666 | """return an uncompressed revision of a given node or revision |
|
2669 | """return an uncompressed revision of a given node or revision | |
2667 | number. |
|
2670 | number. | |
2668 | """ |
|
2671 | """ | |
2669 | return self._revisiondata(nodeorrev) |
|
2672 | return self._revisiondata(nodeorrev) | |
2670 |
|
2673 | |||
2671 | def sidedata(self, nodeorrev): |
|
2674 | def sidedata(self, nodeorrev): | |
2672 | """a map of extra data related to the changeset but not part of the hash |
|
2675 | """a map of extra data related to the changeset but not part of the hash | |
2673 |
|
2676 | |||
2674 | This function currently return a dictionary. However, more advanced |
|
2677 | This function currently return a dictionary. However, more advanced | |
2675 | mapping object will likely be used in the future for a more |
|
2678 | mapping object will likely be used in the future for a more | |
2676 | efficient/lazy code. |
|
2679 | efficient/lazy code. | |
2677 | """ |
|
2680 | """ | |
2678 | # deal with <nodeorrev> argument type |
|
2681 | # deal with <nodeorrev> argument type | |
2679 | if isinstance(nodeorrev, int): |
|
2682 | if isinstance(nodeorrev, int): | |
2680 | rev = nodeorrev |
|
2683 | rev = nodeorrev | |
2681 | else: |
|
2684 | else: | |
2682 | rev = self.rev(nodeorrev) |
|
2685 | rev = self.rev(nodeorrev) | |
2683 | return self._sidedata(rev) |
|
2686 | return self._sidedata(rev) | |
2684 |
|
2687 | |||
2685 | def _rawtext(self, node, rev): |
|
2688 | def _rawtext(self, node, rev): | |
2686 | """return the possibly unvalidated rawtext for a revision |
|
2689 | """return the possibly unvalidated rawtext for a revision | |
2687 |
|
2690 | |||
2688 | returns (rev, rawtext, validated) |
|
2691 | returns (rev, rawtext, validated) | |
2689 | """ |
|
2692 | """ | |
2690 | # Check if we have the entry in cache |
|
2693 | # Check if we have the entry in cache | |
2691 | # The cache entry looks like (node, rev, rawtext) |
|
2694 | # The cache entry looks like (node, rev, rawtext) | |
2692 | if self._inner._revisioncache: |
|
2695 | if self._inner._revisioncache: | |
2693 | if self._inner._revisioncache[0] == node: |
|
2696 | if self._inner._revisioncache[0] == node: | |
2694 | return (rev, self._inner._revisioncache[2], True) |
|
2697 | return (rev, self._inner._revisioncache[2], True) | |
2695 |
|
2698 | |||
2696 | if rev is None: |
|
2699 | if rev is None: | |
2697 | rev = self.rev(node) |
|
2700 | rev = self.rev(node) | |
2698 |
|
2701 | |||
2699 | return self._inner.raw_text(node, rev) |
|
2702 | return self._inner.raw_text(node, rev) | |
2700 |
|
2703 | |||
2701 | def _revisiondata(self, nodeorrev, raw=False): |
|
2704 | def _revisiondata(self, nodeorrev, raw=False): | |
2702 | # deal with <nodeorrev> argument type |
|
2705 | # deal with <nodeorrev> argument type | |
2703 | if isinstance(nodeorrev, int): |
|
2706 | if isinstance(nodeorrev, int): | |
2704 | rev = nodeorrev |
|
2707 | rev = nodeorrev | |
2705 | node = self.node(rev) |
|
2708 | node = self.node(rev) | |
2706 | else: |
|
2709 | else: | |
2707 | node = nodeorrev |
|
2710 | node = nodeorrev | |
2708 | rev = None |
|
2711 | rev = None | |
2709 |
|
2712 | |||
2710 | # fast path the special `nullid` rev |
|
2713 | # fast path the special `nullid` rev | |
2711 | if node == self.nullid: |
|
2714 | if node == self.nullid: | |
2712 | return b"" |
|
2715 | return b"" | |
2713 |
|
2716 | |||
2714 | # ``rawtext`` is the text as stored inside the revlog. Might be the |
|
2717 | # ``rawtext`` is the text as stored inside the revlog. Might be the | |
2715 | # revision or might need to be processed to retrieve the revision. |
|
2718 | # revision or might need to be processed to retrieve the revision. | |
2716 | rev, rawtext, validated = self._rawtext(node, rev) |
|
2719 | rev, rawtext, validated = self._rawtext(node, rev) | |
2717 |
|
2720 | |||
2718 | if raw and validated: |
|
2721 | if raw and validated: | |
2719 | # if we don't want to process the raw text and that raw |
|
2722 | # if we don't want to process the raw text and that raw | |
2720 | # text is cached, we can exit early. |
|
2723 | # text is cached, we can exit early. | |
2721 | return rawtext |
|
2724 | return rawtext | |
2722 | if rev is None: |
|
2725 | if rev is None: | |
2723 | rev = self.rev(node) |
|
2726 | rev = self.rev(node) | |
2724 | # the revlog's flag for this revision |
|
2727 | # the revlog's flag for this revision | |
2725 | # (usually alter its state or content) |
|
2728 | # (usually alter its state or content) | |
2726 | flags = self.flags(rev) |
|
2729 | flags = self.flags(rev) | |
2727 |
|
2730 | |||
2728 | if validated and flags == REVIDX_DEFAULT_FLAGS: |
|
2731 | if validated and flags == REVIDX_DEFAULT_FLAGS: | |
2729 | # no extra flags set, no flag processor runs, text = rawtext |
|
2732 | # no extra flags set, no flag processor runs, text = rawtext | |
2730 | return rawtext |
|
2733 | return rawtext | |
2731 |
|
2734 | |||
2732 | if raw: |
|
2735 | if raw: | |
2733 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
2736 | validatehash = flagutil.processflagsraw(self, rawtext, flags) | |
2734 | text = rawtext |
|
2737 | text = rawtext | |
2735 | else: |
|
2738 | else: | |
2736 | r = flagutil.processflagsread(self, rawtext, flags) |
|
2739 | r = flagutil.processflagsread(self, rawtext, flags) | |
2737 | text, validatehash = r |
|
2740 | text, validatehash = r | |
2738 | if validatehash: |
|
2741 | if validatehash: | |
2739 | self.checkhash(text, node, rev=rev) |
|
2742 | self.checkhash(text, node, rev=rev) | |
2740 | if not validated: |
|
2743 | if not validated: | |
2741 | self._inner._revisioncache = (node, rev, rawtext) |
|
2744 | self._inner._revisioncache = (node, rev, rawtext) | |
2742 |
|
2745 | |||
2743 | return text |
|
2746 | return text | |
2744 |
|
2747 | |||
2745 | def _sidedata(self, rev): |
|
2748 | def _sidedata(self, rev): | |
2746 | """Return the sidedata for a given revision number.""" |
|
2749 | """Return the sidedata for a given revision number.""" | |
2747 | sidedata_end = None |
|
2750 | sidedata_end = None | |
2748 | if self._docket is not None: |
|
2751 | if self._docket is not None: | |
2749 | sidedata_end = self._docket.sidedata_end |
|
2752 | sidedata_end = self._docket.sidedata_end | |
2750 | return self._inner.sidedata(rev, sidedata_end) |
|
2753 | return self._inner.sidedata(rev, sidedata_end) | |
2751 |
|
2754 | |||
2752 | def rawdata(self, nodeorrev): |
|
2755 | def rawdata(self, nodeorrev): | |
2753 | """return an uncompressed raw data of a given node or revision number.""" |
|
2756 | """return an uncompressed raw data of a given node or revision number.""" | |
2754 | return self._revisiondata(nodeorrev, raw=True) |
|
2757 | return self._revisiondata(nodeorrev, raw=True) | |
2755 |
|
2758 | |||
2756 | def hash(self, text, p1, p2): |
|
2759 | def hash(self, text, p1, p2): | |
2757 | """Compute a node hash. |
|
2760 | """Compute a node hash. | |
2758 |
|
2761 | |||
2759 | Available as a function so that subclasses can replace the hash |
|
2762 | Available as a function so that subclasses can replace the hash | |
2760 | as needed. |
|
2763 | as needed. | |
2761 | """ |
|
2764 | """ | |
2762 | return storageutil.hashrevisionsha1(text, p1, p2) |
|
2765 | return storageutil.hashrevisionsha1(text, p1, p2) | |
2763 |
|
2766 | |||
2764 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
2767 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
2765 | """Check node hash integrity. |
|
2768 | """Check node hash integrity. | |
2766 |
|
2769 | |||
2767 | Available as a function so that subclasses can extend hash mismatch |
|
2770 | Available as a function so that subclasses can extend hash mismatch | |
2768 | behaviors as needed. |
|
2771 | behaviors as needed. | |
2769 | """ |
|
2772 | """ | |
2770 | try: |
|
2773 | try: | |
2771 | if p1 is None and p2 is None: |
|
2774 | if p1 is None and p2 is None: | |
2772 | p1, p2 = self.parents(node) |
|
2775 | p1, p2 = self.parents(node) | |
2773 | if node != self.hash(text, p1, p2): |
|
2776 | if node != self.hash(text, p1, p2): | |
2774 | # Clear the revision cache on hash failure. The revision cache |
|
2777 | # Clear the revision cache on hash failure. The revision cache | |
2775 | # only stores the raw revision and clearing the cache does have |
|
2778 | # only stores the raw revision and clearing the cache does have | |
2776 | # the side-effect that we won't have a cache hit when the raw |
|
2779 | # the side-effect that we won't have a cache hit when the raw | |
2777 | # revision data is accessed. But this case should be rare and |
|
2780 | # revision data is accessed. But this case should be rare and | |
2778 | # it is extra work to teach the cache about the hash |
|
2781 | # it is extra work to teach the cache about the hash | |
2779 | # verification state. |
|
2782 | # verification state. | |
2780 | if ( |
|
2783 | if ( | |
2781 | self._inner._revisioncache |
|
2784 | self._inner._revisioncache | |
2782 | and self._inner._revisioncache[0] == node |
|
2785 | and self._inner._revisioncache[0] == node | |
2783 | ): |
|
2786 | ): | |
2784 | self._inner._revisioncache = None |
|
2787 | self._inner._revisioncache = None | |
2785 |
|
2788 | |||
2786 | revornode = rev |
|
2789 | revornode = rev | |
2787 | if revornode is None: |
|
2790 | if revornode is None: | |
2788 | revornode = templatefilters.short(hex(node)) |
|
2791 | revornode = templatefilters.short(hex(node)) | |
2789 | raise error.RevlogError( |
|
2792 | raise error.RevlogError( | |
2790 | _(b"integrity check failed on %s:%s") |
|
2793 | _(b"integrity check failed on %s:%s") | |
2791 | % (self.display_id, pycompat.bytestr(revornode)) |
|
2794 | % (self.display_id, pycompat.bytestr(revornode)) | |
2792 | ) |
|
2795 | ) | |
2793 | except error.RevlogError: |
|
2796 | except error.RevlogError: | |
2794 | if self.feature_config.censorable and storageutil.iscensoredtext( |
|
2797 | if self.feature_config.censorable and storageutil.iscensoredtext( | |
2795 | text |
|
2798 | text | |
2796 | ): |
|
2799 | ): | |
2797 | raise error.CensoredNodeError(self.display_id, node, text) |
|
2800 | raise error.CensoredNodeError(self.display_id, node, text) | |
2798 | raise |
|
2801 | raise | |
2799 |
|
2802 | |||
2800 | @property |
|
2803 | @property | |
2801 | def _split_index_file(self): |
|
2804 | def _split_index_file(self): | |
2802 | """the path where to expect the index of an ongoing splitting operation |
|
2805 | """the path where to expect the index of an ongoing splitting operation | |
2803 |
|
2806 | |||
2804 | The file will only exist if a splitting operation is in progress, but |
|
2807 | The file will only exist if a splitting operation is in progress, but | |
2805 | it is always expected at the same location.""" |
|
2808 | it is always expected at the same location.""" | |
2806 | parts = self.radix.split(b'/') |
|
2809 | parts = self.radix.split(b'/') | |
2807 | if len(parts) > 1: |
|
2810 | if len(parts) > 1: | |
2808 | # adds a '-s' prefix to the ``data/` or `meta/` base |
|
2811 | # adds a '-s' prefix to the ``data/` or `meta/` base | |
2809 | head = parts[0] + b'-s' |
|
2812 | head = parts[0] + b'-s' | |
2810 | mids = parts[1:-1] |
|
2813 | mids = parts[1:-1] | |
2811 | tail = parts[-1] + b'.i' |
|
2814 | tail = parts[-1] + b'.i' | |
2812 | pieces = [head] + mids + [tail] |
|
2815 | pieces = [head] + mids + [tail] | |
2813 | return b'/'.join(pieces) |
|
2816 | return b'/'.join(pieces) | |
2814 | else: |
|
2817 | else: | |
2815 | # the revlog is stored at the root of the store (changelog or |
|
2818 | # the revlog is stored at the root of the store (changelog or | |
2816 | # manifest), no risk of collision. |
|
2819 | # manifest), no risk of collision. | |
2817 | return self.radix + b'.i.s' |
|
2820 | return self.radix + b'.i.s' | |
2818 |
|
2821 | |||
2819 | def _enforceinlinesize(self, tr, side_write=True): |
|
2822 | def _enforceinlinesize(self, tr, side_write=True): | |
2820 | """Check if the revlog is too big for inline and convert if so. |
|
2823 | """Check if the revlog is too big for inline and convert if so. | |
2821 |
|
2824 | |||
2822 | This should be called after revisions are added to the revlog. If the |
|
2825 | This should be called after revisions are added to the revlog. If the | |
2823 | revlog has grown too large to be an inline revlog, it will convert it |
|
2826 | revlog has grown too large to be an inline revlog, it will convert it | |
2824 | to use multiple index and data files. |
|
2827 | to use multiple index and data files. | |
2825 | """ |
|
2828 | """ | |
2826 | tiprev = len(self) - 1 |
|
2829 | tiprev = len(self) - 1 | |
2827 | total_size = self.start(tiprev) + self.length(tiprev) |
|
2830 | total_size = self.start(tiprev) + self.length(tiprev) | |
2828 | if not self._inline or (self._may_inline and total_size < _maxinline): |
|
2831 | if not self._inline or (self._may_inline and total_size < _maxinline): | |
2829 | return |
|
2832 | return | |
2830 |
|
2833 | |||
2831 | if self._docket is not None: |
|
2834 | if self._docket is not None: | |
2832 | msg = b"inline revlog should not have a docket" |
|
2835 | msg = b"inline revlog should not have a docket" | |
2833 | raise error.ProgrammingError(msg) |
|
2836 | raise error.ProgrammingError(msg) | |
2834 |
|
2837 | |||
2835 | # In the common case, we enforce inline size because the revlog has |
|
2838 | # In the common case, we enforce inline size because the revlog has | |
2836 | # been appened too. And in such case, it must have an initial offset |
|
2839 | # been appened too. And in such case, it must have an initial offset | |
2837 | # recorded in the transaction. |
|
2840 | # recorded in the transaction. | |
2838 | troffset = tr.findoffset(self._inner.canonical_index_file) |
|
2841 | troffset = tr.findoffset(self._inner.canonical_index_file) | |
2839 | pre_touched = troffset is not None |
|
2842 | pre_touched = troffset is not None | |
2840 | if not pre_touched and self.target[0] != KIND_CHANGELOG: |
|
2843 | if not pre_touched and self.target[0] != KIND_CHANGELOG: | |
2841 | raise error.RevlogError( |
|
2844 | raise error.RevlogError( | |
2842 | _(b"%s not found in the transaction") % self._indexfile |
|
2845 | _(b"%s not found in the transaction") % self._indexfile | |
2843 | ) |
|
2846 | ) | |
2844 |
|
2847 | |||
2845 | tr.addbackup(self._inner.canonical_index_file, for_offset=pre_touched) |
|
2848 | tr.addbackup(self._inner.canonical_index_file, for_offset=pre_touched) | |
2846 | tr.add(self._datafile, 0) |
|
2849 | tr.add(self._datafile, 0) | |
2847 |
|
2850 | |||
2848 | new_index_file_path = None |
|
2851 | new_index_file_path = None | |
2849 | if side_write: |
|
2852 | if side_write: | |
2850 | old_index_file_path = self._indexfile |
|
2853 | old_index_file_path = self._indexfile | |
2851 | new_index_file_path = self._split_index_file |
|
2854 | new_index_file_path = self._split_index_file | |
2852 | opener = self.opener |
|
2855 | opener = self.opener | |
2853 | weak_self = weakref.ref(self) |
|
2856 | weak_self = weakref.ref(self) | |
2854 |
|
2857 | |||
2855 | # the "split" index replace the real index when the transaction is |
|
2858 | # the "split" index replace the real index when the transaction is | |
2856 | # finalized |
|
2859 | # finalized | |
2857 | def finalize_callback(tr): |
|
2860 | def finalize_callback(tr): | |
2858 | opener.rename( |
|
2861 | opener.rename( | |
2859 | new_index_file_path, |
|
2862 | new_index_file_path, | |
2860 | old_index_file_path, |
|
2863 | old_index_file_path, | |
2861 | checkambig=True, |
|
2864 | checkambig=True, | |
2862 | ) |
|
2865 | ) | |
2863 | maybe_self = weak_self() |
|
2866 | maybe_self = weak_self() | |
2864 | if maybe_self is not None: |
|
2867 | if maybe_self is not None: | |
2865 | maybe_self._indexfile = old_index_file_path |
|
2868 | maybe_self._indexfile = old_index_file_path | |
2866 | maybe_self._inner.index_file = maybe_self._indexfile |
|
2869 | maybe_self._inner.index_file = maybe_self._indexfile | |
2867 |
|
2870 | |||
2868 | def abort_callback(tr): |
|
2871 | def abort_callback(tr): | |
2869 | maybe_self = weak_self() |
|
2872 | maybe_self = weak_self() | |
2870 | if maybe_self is not None: |
|
2873 | if maybe_self is not None: | |
2871 | maybe_self._indexfile = old_index_file_path |
|
2874 | maybe_self._indexfile = old_index_file_path | |
2872 | maybe_self._inner.inline = True |
|
2875 | maybe_self._inner.inline = True | |
2873 | maybe_self._inner.index_file = old_index_file_path |
|
2876 | maybe_self._inner.index_file = old_index_file_path | |
2874 |
|
2877 | |||
2875 | tr.registertmp(new_index_file_path) |
|
2878 | tr.registertmp(new_index_file_path) | |
2876 | if self.target[1] is not None: |
|
2879 | if self.target[1] is not None: | |
2877 | callback_id = b'000-revlog-split-%d-%s' % self.target |
|
2880 | callback_id = b'000-revlog-split-%d-%s' % self.target | |
2878 | else: |
|
2881 | else: | |
2879 | callback_id = b'000-revlog-split-%d' % self.target[0] |
|
2882 | callback_id = b'000-revlog-split-%d' % self.target[0] | |
2880 | tr.addfinalize(callback_id, finalize_callback) |
|
2883 | tr.addfinalize(callback_id, finalize_callback) | |
2881 | tr.addabort(callback_id, abort_callback) |
|
2884 | tr.addabort(callback_id, abort_callback) | |
2882 |
|
2885 | |||
2883 | self._format_flags &= ~FLAG_INLINE_DATA |
|
2886 | self._format_flags &= ~FLAG_INLINE_DATA | |
2884 | self._inner.split_inline( |
|
2887 | self._inner.split_inline( | |
2885 | tr, |
|
2888 | tr, | |
2886 | self._format_flags | self._format_version, |
|
2889 | self._format_flags | self._format_version, | |
2887 | new_index_file_path=new_index_file_path, |
|
2890 | new_index_file_path=new_index_file_path, | |
2888 | ) |
|
2891 | ) | |
2889 |
|
2892 | |||
2890 | self._inline = False |
|
2893 | self._inline = False | |
2891 | if new_index_file_path is not None: |
|
2894 | if new_index_file_path is not None: | |
2892 | self._indexfile = new_index_file_path |
|
2895 | self._indexfile = new_index_file_path | |
2893 |
|
2896 | |||
2894 | nodemaputil.setup_persistent_nodemap(tr, self) |
|
2897 | nodemaputil.setup_persistent_nodemap(tr, self) | |
2895 |
|
2898 | |||
2896 | def _nodeduplicatecallback(self, transaction, node): |
|
2899 | def _nodeduplicatecallback(self, transaction, node): | |
2897 | """called when trying to add a node already stored.""" |
|
2900 | """called when trying to add a node already stored.""" | |
2898 |
|
2901 | |||
2899 | @contextlib.contextmanager |
|
2902 | @contextlib.contextmanager | |
2900 | def reading(self): |
|
2903 | def reading(self): | |
2901 | with self._inner.reading(): |
|
2904 | with self._inner.reading(): | |
2902 | yield |
|
2905 | yield | |
2903 |
|
2906 | |||
2904 | @contextlib.contextmanager |
|
2907 | @contextlib.contextmanager | |
2905 | def _writing(self, transaction): |
|
2908 | def _writing(self, transaction): | |
2906 | if self._trypending: |
|
2909 | if self._trypending: | |
2907 | msg = b'try to write in a `trypending` revlog: %s' |
|
2910 | msg = b'try to write in a `trypending` revlog: %s' | |
2908 | msg %= self.display_id |
|
2911 | msg %= self.display_id | |
2909 | raise error.ProgrammingError(msg) |
|
2912 | raise error.ProgrammingError(msg) | |
2910 | if self._inner.is_writing: |
|
2913 | if self._inner.is_writing: | |
2911 | yield |
|
2914 | yield | |
2912 | else: |
|
2915 | else: | |
2913 | data_end = None |
|
2916 | data_end = None | |
2914 | sidedata_end = None |
|
2917 | sidedata_end = None | |
2915 | if self._docket is not None: |
|
2918 | if self._docket is not None: | |
2916 | data_end = self._docket.data_end |
|
2919 | data_end = self._docket.data_end | |
2917 | sidedata_end = self._docket.sidedata_end |
|
2920 | sidedata_end = self._docket.sidedata_end | |
2918 | with self._inner.writing( |
|
2921 | with self._inner.writing( | |
2919 | transaction, |
|
2922 | transaction, | |
2920 | data_end=data_end, |
|
2923 | data_end=data_end, | |
2921 | sidedata_end=sidedata_end, |
|
2924 | sidedata_end=sidedata_end, | |
2922 | ): |
|
2925 | ): | |
2923 | yield |
|
2926 | yield | |
2924 | if self._docket is not None: |
|
2927 | if self._docket is not None: | |
2925 | self._write_docket(transaction) |
|
2928 | self._write_docket(transaction) | |
2926 |
|
2929 | |||
2927 | @property |
|
2930 | @property | |
2928 | def is_delaying(self): |
|
2931 | def is_delaying(self): | |
2929 | return self._inner.is_delaying |
|
2932 | return self._inner.is_delaying | |
2930 |
|
2933 | |||
2931 | def _write_docket(self, transaction): |
|
2934 | def _write_docket(self, transaction): | |
2932 | """write the current docket on disk |
|
2935 | """write the current docket on disk | |
2933 |
|
2936 | |||
2934 | Exist as a method to help changelog to implement transaction logic |
|
2937 | Exist as a method to help changelog to implement transaction logic | |
2935 |
|
2938 | |||
2936 | We could also imagine using the same transaction logic for all revlog |
|
2939 | We could also imagine using the same transaction logic for all revlog | |
2937 | since docket are cheap.""" |
|
2940 | since docket are cheap.""" | |
2938 | self._docket.write(transaction) |
|
2941 | self._docket.write(transaction) | |
2939 |
|
2942 | |||
2940 | def addrevision( |
|
2943 | def addrevision( | |
2941 | self, |
|
2944 | self, | |
2942 | text, |
|
2945 | text, | |
2943 | transaction, |
|
2946 | transaction, | |
2944 | link, |
|
2947 | link, | |
2945 | p1, |
|
2948 | p1, | |
2946 | p2, |
|
2949 | p2, | |
2947 | cachedelta=None, |
|
2950 | cachedelta=None, | |
2948 | node=None, |
|
2951 | node=None, | |
2949 | flags=REVIDX_DEFAULT_FLAGS, |
|
2952 | flags=REVIDX_DEFAULT_FLAGS, | |
2950 | deltacomputer=None, |
|
2953 | deltacomputer=None, | |
2951 | sidedata=None, |
|
2954 | sidedata=None, | |
2952 | ): |
|
2955 | ): | |
2953 | """add a revision to the log |
|
2956 | """add a revision to the log | |
2954 |
|
2957 | |||
2955 | text - the revision data to add |
|
2958 | text - the revision data to add | |
2956 | transaction - the transaction object used for rollback |
|
2959 | transaction - the transaction object used for rollback | |
2957 | link - the linkrev data to add |
|
2960 | link - the linkrev data to add | |
2958 | p1, p2 - the parent nodeids of the revision |
|
2961 | p1, p2 - the parent nodeids of the revision | |
2959 | cachedelta - an optional precomputed delta |
|
2962 | cachedelta - an optional precomputed delta | |
2960 | node - nodeid of revision; typically node is not specified, and it is |
|
2963 | node - nodeid of revision; typically node is not specified, and it is | |
2961 | computed by default as hash(text, p1, p2), however subclasses might |
|
2964 | computed by default as hash(text, p1, p2), however subclasses might | |
2962 | use different hashing method (and override checkhash() in such case) |
|
2965 | use different hashing method (and override checkhash() in such case) | |
2963 | flags - the known flags to set on the revision |
|
2966 | flags - the known flags to set on the revision | |
2964 | deltacomputer - an optional deltacomputer instance shared between |
|
2967 | deltacomputer - an optional deltacomputer instance shared between | |
2965 | multiple calls |
|
2968 | multiple calls | |
2966 | """ |
|
2969 | """ | |
2967 | if link == nullrev: |
|
2970 | if link == nullrev: | |
2968 | raise error.RevlogError( |
|
2971 | raise error.RevlogError( | |
2969 | _(b"attempted to add linkrev -1 to %s") % self.display_id |
|
2972 | _(b"attempted to add linkrev -1 to %s") % self.display_id | |
2970 | ) |
|
2973 | ) | |
2971 |
|
2974 | |||
2972 | if sidedata is None: |
|
2975 | if sidedata is None: | |
2973 | sidedata = {} |
|
2976 | sidedata = {} | |
2974 | elif sidedata and not self.feature_config.has_side_data: |
|
2977 | elif sidedata and not self.feature_config.has_side_data: | |
2975 | raise error.ProgrammingError( |
|
2978 | raise error.ProgrammingError( | |
2976 | _(b"trying to add sidedata to a revlog who don't support them") |
|
2979 | _(b"trying to add sidedata to a revlog who don't support them") | |
2977 | ) |
|
2980 | ) | |
2978 |
|
2981 | |||
2979 | if flags: |
|
2982 | if flags: | |
2980 | node = node or self.hash(text, p1, p2) |
|
2983 | node = node or self.hash(text, p1, p2) | |
2981 |
|
2984 | |||
2982 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) |
|
2985 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) | |
2983 |
|
2986 | |||
2984 | # If the flag processor modifies the revision data, ignore any provided |
|
2987 | # If the flag processor modifies the revision data, ignore any provided | |
2985 | # cachedelta. |
|
2988 | # cachedelta. | |
2986 | if rawtext != text: |
|
2989 | if rawtext != text: | |
2987 | cachedelta = None |
|
2990 | cachedelta = None | |
2988 |
|
2991 | |||
2989 | if len(rawtext) > _maxentrysize: |
|
2992 | if len(rawtext) > _maxentrysize: | |
2990 | raise error.RevlogError( |
|
2993 | raise error.RevlogError( | |
2991 | _( |
|
2994 | _( | |
2992 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" |
|
2995 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" | |
2993 | ) |
|
2996 | ) | |
2994 | % (self.display_id, len(rawtext)) |
|
2997 | % (self.display_id, len(rawtext)) | |
2995 | ) |
|
2998 | ) | |
2996 |
|
2999 | |||
2997 | node = node or self.hash(rawtext, p1, p2) |
|
3000 | node = node or self.hash(rawtext, p1, p2) | |
2998 | rev = self.index.get_rev(node) |
|
3001 | rev = self.index.get_rev(node) | |
2999 | if rev is not None: |
|
3002 | if rev is not None: | |
3000 | return rev |
|
3003 | return rev | |
3001 |
|
3004 | |||
3002 | if validatehash: |
|
3005 | if validatehash: | |
3003 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
3006 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
3004 |
|
3007 | |||
3005 | return self.addrawrevision( |
|
3008 | return self.addrawrevision( | |
3006 | rawtext, |
|
3009 | rawtext, | |
3007 | transaction, |
|
3010 | transaction, | |
3008 | link, |
|
3011 | link, | |
3009 | p1, |
|
3012 | p1, | |
3010 | p2, |
|
3013 | p2, | |
3011 | node, |
|
3014 | node, | |
3012 | flags, |
|
3015 | flags, | |
3013 | cachedelta=cachedelta, |
|
3016 | cachedelta=cachedelta, | |
3014 | deltacomputer=deltacomputer, |
|
3017 | deltacomputer=deltacomputer, | |
3015 | sidedata=sidedata, |
|
3018 | sidedata=sidedata, | |
3016 | ) |
|
3019 | ) | |
3017 |
|
3020 | |||
3018 | def addrawrevision( |
|
3021 | def addrawrevision( | |
3019 | self, |
|
3022 | self, | |
3020 | rawtext, |
|
3023 | rawtext, | |
3021 | transaction, |
|
3024 | transaction, | |
3022 | link, |
|
3025 | link, | |
3023 | p1, |
|
3026 | p1, | |
3024 | p2, |
|
3027 | p2, | |
3025 | node, |
|
3028 | node, | |
3026 | flags, |
|
3029 | flags, | |
3027 | cachedelta=None, |
|
3030 | cachedelta=None, | |
3028 | deltacomputer=None, |
|
3031 | deltacomputer=None, | |
3029 | sidedata=None, |
|
3032 | sidedata=None, | |
3030 | ): |
|
3033 | ): | |
3031 | """add a raw revision with known flags, node and parents |
|
3034 | """add a raw revision with known flags, node and parents | |
3032 | useful when reusing a revision not stored in this revlog (ex: received |
|
3035 | useful when reusing a revision not stored in this revlog (ex: received | |
3033 | over wire, or read from an external bundle). |
|
3036 | over wire, or read from an external bundle). | |
3034 | """ |
|
3037 | """ | |
3035 | with self._writing(transaction): |
|
3038 | with self._writing(transaction): | |
3036 | return self._addrevision( |
|
3039 | return self._addrevision( | |
3037 | node, |
|
3040 | node, | |
3038 | rawtext, |
|
3041 | rawtext, | |
3039 | transaction, |
|
3042 | transaction, | |
3040 | link, |
|
3043 | link, | |
3041 | p1, |
|
3044 | p1, | |
3042 | p2, |
|
3045 | p2, | |
3043 | flags, |
|
3046 | flags, | |
3044 | cachedelta, |
|
3047 | cachedelta, | |
3045 | deltacomputer=deltacomputer, |
|
3048 | deltacomputer=deltacomputer, | |
3046 | sidedata=sidedata, |
|
3049 | sidedata=sidedata, | |
3047 | ) |
|
3050 | ) | |
3048 |
|
3051 | |||
3049 | def compress(self, data): |
|
3052 | def compress(self, data): | |
3050 | return self._inner.compress(data) |
|
3053 | return self._inner.compress(data) | |
3051 |
|
3054 | |||
3052 | def decompress(self, data): |
|
3055 | def decompress(self, data): | |
3053 | return self._inner.decompress(data) |
|
3056 | return self._inner.decompress(data) | |
3054 |
|
3057 | |||
3055 | def _addrevision( |
|
3058 | def _addrevision( | |
3056 | self, |
|
3059 | self, | |
3057 | node, |
|
3060 | node, | |
3058 | rawtext, |
|
3061 | rawtext, | |
3059 | transaction, |
|
3062 | transaction, | |
3060 | link, |
|
3063 | link, | |
3061 | p1, |
|
3064 | p1, | |
3062 | p2, |
|
3065 | p2, | |
3063 | flags, |
|
3066 | flags, | |
3064 | cachedelta, |
|
3067 | cachedelta, | |
3065 | alwayscache=False, |
|
3068 | alwayscache=False, | |
3066 | deltacomputer=None, |
|
3069 | deltacomputer=None, | |
3067 | sidedata=None, |
|
3070 | sidedata=None, | |
3068 | ): |
|
3071 | ): | |
3069 | """internal function to add revisions to the log |
|
3072 | """internal function to add revisions to the log | |
3070 |
|
3073 | |||
3071 | see addrevision for argument descriptions. |
|
3074 | see addrevision for argument descriptions. | |
3072 |
|
3075 | |||
3073 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. |
|
3076 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. | |
3074 |
|
3077 | |||
3075 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will |
|
3078 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will | |
3076 | be used. |
|
3079 | be used. | |
3077 |
|
3080 | |||
3078 | invariants: |
|
3081 | invariants: | |
3079 | - rawtext is optional (can be None); if not set, cachedelta must be set. |
|
3082 | - rawtext is optional (can be None); if not set, cachedelta must be set. | |
3080 | if both are set, they must correspond to each other. |
|
3083 | if both are set, they must correspond to each other. | |
3081 | """ |
|
3084 | """ | |
3082 | if node == self.nullid: |
|
3085 | if node == self.nullid: | |
3083 | raise error.RevlogError( |
|
3086 | raise error.RevlogError( | |
3084 | _(b"%s: attempt to add null revision") % self.display_id |
|
3087 | _(b"%s: attempt to add null revision") % self.display_id | |
3085 | ) |
|
3088 | ) | |
3086 | if ( |
|
3089 | if ( | |
3087 | node == self.nodeconstants.wdirid |
|
3090 | node == self.nodeconstants.wdirid | |
3088 | or node in self.nodeconstants.wdirfilenodeids |
|
3091 | or node in self.nodeconstants.wdirfilenodeids | |
3089 | ): |
|
3092 | ): | |
3090 | raise error.RevlogError( |
|
3093 | raise error.RevlogError( | |
3091 | _(b"%s: attempt to add wdir revision") % self.display_id |
|
3094 | _(b"%s: attempt to add wdir revision") % self.display_id | |
3092 | ) |
|
3095 | ) | |
3093 | if self._inner._writinghandles is None: |
|
3096 | if self._inner._writinghandles is None: | |
3094 | msg = b'adding revision outside `revlog._writing` context' |
|
3097 | msg = b'adding revision outside `revlog._writing` context' | |
3095 | raise error.ProgrammingError(msg) |
|
3098 | raise error.ProgrammingError(msg) | |
3096 |
|
3099 | |||
3097 | btext = [rawtext] |
|
3100 | btext = [rawtext] | |
3098 |
|
3101 | |||
3099 | curr = len(self) |
|
3102 | curr = len(self) | |
3100 | prev = curr - 1 |
|
3103 | prev = curr - 1 | |
3101 |
|
3104 | |||
3102 | offset = self._get_data_offset(prev) |
|
3105 | offset = self._get_data_offset(prev) | |
3103 |
|
3106 | |||
3104 | if self._concurrencychecker: |
|
3107 | if self._concurrencychecker: | |
3105 | ifh, dfh, sdfh = self._inner._writinghandles |
|
3108 | ifh, dfh, sdfh = self._inner._writinghandles | |
3106 | # XXX no checking for the sidedata file |
|
3109 | # XXX no checking for the sidedata file | |
3107 | if self._inline: |
|
3110 | if self._inline: | |
3108 | # offset is "as if" it were in the .d file, so we need to add on |
|
3111 | # offset is "as if" it were in the .d file, so we need to add on | |
3109 | # the size of the entry metadata. |
|
3112 | # the size of the entry metadata. | |
3110 | self._concurrencychecker( |
|
3113 | self._concurrencychecker( | |
3111 | ifh, self._indexfile, offset + curr * self.index.entry_size |
|
3114 | ifh, self._indexfile, offset + curr * self.index.entry_size | |
3112 | ) |
|
3115 | ) | |
3113 | else: |
|
3116 | else: | |
3114 | # Entries in the .i are a consistent size. |
|
3117 | # Entries in the .i are a consistent size. | |
3115 | self._concurrencychecker( |
|
3118 | self._concurrencychecker( | |
3116 | ifh, self._indexfile, curr * self.index.entry_size |
|
3119 | ifh, self._indexfile, curr * self.index.entry_size | |
3117 | ) |
|
3120 | ) | |
3118 | self._concurrencychecker(dfh, self._datafile, offset) |
|
3121 | self._concurrencychecker(dfh, self._datafile, offset) | |
3119 |
|
3122 | |||
3120 | p1r, p2r = self.rev(p1), self.rev(p2) |
|
3123 | p1r, p2r = self.rev(p1), self.rev(p2) | |
3121 |
|
3124 | |||
3122 | # full versions are inserted when the needed deltas |
|
3125 | # full versions are inserted when the needed deltas | |
3123 | # become comparable to the uncompressed text |
|
3126 | # become comparable to the uncompressed text | |
3124 | if rawtext is None: |
|
3127 | if rawtext is None: | |
3125 | # need rawtext size, before changed by flag processors, which is |
|
3128 | # need rawtext size, before changed by flag processors, which is | |
3126 | # the non-raw size. use revlog explicitly to avoid filelog's extra |
|
3129 | # the non-raw size. use revlog explicitly to avoid filelog's extra | |
3127 | # logic that might remove metadata size. |
|
3130 | # logic that might remove metadata size. | |
3128 | textlen = mdiff.patchedsize( |
|
3131 | textlen = mdiff.patchedsize( | |
3129 | revlog.size(self, cachedelta[0]), cachedelta[1] |
|
3132 | revlog.size(self, cachedelta[0]), cachedelta[1] | |
3130 | ) |
|
3133 | ) | |
3131 | else: |
|
3134 | else: | |
3132 | textlen = len(rawtext) |
|
3135 | textlen = len(rawtext) | |
3133 |
|
3136 | |||
3134 | if deltacomputer is None: |
|
3137 | if deltacomputer is None: | |
3135 | write_debug = None |
|
3138 | write_debug = None | |
3136 | if self.delta_config.debug_delta: |
|
3139 | if self.delta_config.debug_delta: | |
3137 | write_debug = transaction._report |
|
3140 | write_debug = transaction._report | |
3138 | deltacomputer = deltautil.deltacomputer( |
|
3141 | deltacomputer = deltautil.deltacomputer( | |
3139 | self, write_debug=write_debug |
|
3142 | self, write_debug=write_debug | |
3140 | ) |
|
3143 | ) | |
3141 |
|
3144 | |||
3142 | if cachedelta is not None and len(cachedelta) == 2: |
|
3145 | if cachedelta is not None and len(cachedelta) == 2: | |
3143 | # If the cached delta has no information about how it should be |
|
3146 | # If the cached delta has no information about how it should be | |
3144 | # reused, add the default reuse instruction according to the |
|
3147 | # reused, add the default reuse instruction according to the | |
3145 | # revlog's configuration. |
|
3148 | # revlog's configuration. | |
3146 | if ( |
|
3149 | if ( | |
3147 | self.delta_config.general_delta |
|
3150 | self.delta_config.general_delta | |
3148 | and self.delta_config.lazy_delta_base |
|
3151 | and self.delta_config.lazy_delta_base | |
3149 | ): |
|
3152 | ): | |
3150 | delta_base_reuse = DELTA_BASE_REUSE_TRY |
|
3153 | delta_base_reuse = DELTA_BASE_REUSE_TRY | |
3151 | else: |
|
3154 | else: | |
3152 | delta_base_reuse = DELTA_BASE_REUSE_NO |
|
3155 | delta_base_reuse = DELTA_BASE_REUSE_NO | |
3153 | cachedelta = (cachedelta[0], cachedelta[1], delta_base_reuse) |
|
3156 | cachedelta = (cachedelta[0], cachedelta[1], delta_base_reuse) | |
3154 |
|
3157 | |||
3155 | revinfo = revlogutils.revisioninfo( |
|
3158 | revinfo = revlogutils.revisioninfo( | |
3156 | node, |
|
3159 | node, | |
3157 | p1, |
|
3160 | p1, | |
3158 | p2, |
|
3161 | p2, | |
3159 | btext, |
|
3162 | btext, | |
3160 | textlen, |
|
3163 | textlen, | |
3161 | cachedelta, |
|
3164 | cachedelta, | |
3162 | flags, |
|
3165 | flags, | |
3163 | ) |
|
3166 | ) | |
3164 |
|
3167 | |||
3165 | deltainfo = deltacomputer.finddeltainfo(revinfo) |
|
3168 | deltainfo = deltacomputer.finddeltainfo(revinfo) | |
3166 |
|
3169 | |||
3167 | compression_mode = COMP_MODE_INLINE |
|
3170 | compression_mode = COMP_MODE_INLINE | |
3168 | if self._docket is not None: |
|
3171 | if self._docket is not None: | |
3169 | default_comp = self._docket.default_compression_header |
|
3172 | default_comp = self._docket.default_compression_header | |
3170 | r = deltautil.delta_compression(default_comp, deltainfo) |
|
3173 | r = deltautil.delta_compression(default_comp, deltainfo) | |
3171 | compression_mode, deltainfo = r |
|
3174 | compression_mode, deltainfo = r | |
3172 |
|
3175 | |||
3173 | sidedata_compression_mode = COMP_MODE_INLINE |
|
3176 | sidedata_compression_mode = COMP_MODE_INLINE | |
3174 | if sidedata and self.feature_config.has_side_data: |
|
3177 | if sidedata and self.feature_config.has_side_data: | |
3175 | sidedata_compression_mode = COMP_MODE_PLAIN |
|
3178 | sidedata_compression_mode = COMP_MODE_PLAIN | |
3176 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) |
|
3179 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) | |
3177 | sidedata_offset = self._docket.sidedata_end |
|
3180 | sidedata_offset = self._docket.sidedata_end | |
3178 | h, comp_sidedata = self._inner.compress(serialized_sidedata) |
|
3181 | h, comp_sidedata = self._inner.compress(serialized_sidedata) | |
3179 | if ( |
|
3182 | if ( | |
3180 | h != b'u' |
|
3183 | h != b'u' | |
3181 | and comp_sidedata[0:1] != b'\0' |
|
3184 | and comp_sidedata[0:1] != b'\0' | |
3182 | and len(comp_sidedata) < len(serialized_sidedata) |
|
3185 | and len(comp_sidedata) < len(serialized_sidedata) | |
3183 | ): |
|
3186 | ): | |
3184 | assert not h |
|
3187 | assert not h | |
3185 | if ( |
|
3188 | if ( | |
3186 | comp_sidedata[0:1] |
|
3189 | comp_sidedata[0:1] | |
3187 | == self._docket.default_compression_header |
|
3190 | == self._docket.default_compression_header | |
3188 | ): |
|
3191 | ): | |
3189 | sidedata_compression_mode = COMP_MODE_DEFAULT |
|
3192 | sidedata_compression_mode = COMP_MODE_DEFAULT | |
3190 | serialized_sidedata = comp_sidedata |
|
3193 | serialized_sidedata = comp_sidedata | |
3191 | else: |
|
3194 | else: | |
3192 | sidedata_compression_mode = COMP_MODE_INLINE |
|
3195 | sidedata_compression_mode = COMP_MODE_INLINE | |
3193 | serialized_sidedata = comp_sidedata |
|
3196 | serialized_sidedata = comp_sidedata | |
3194 | else: |
|
3197 | else: | |
3195 | serialized_sidedata = b"" |
|
3198 | serialized_sidedata = b"" | |
3196 | # Don't store the offset if the sidedata is empty, that way |
|
3199 | # Don't store the offset if the sidedata is empty, that way | |
3197 | # we can easily detect empty sidedata and they will be no different |
|
3200 | # we can easily detect empty sidedata and they will be no different | |
3198 | # than ones we manually add. |
|
3201 | # than ones we manually add. | |
3199 | sidedata_offset = 0 |
|
3202 | sidedata_offset = 0 | |
3200 |
|
3203 | |||
3201 | rank = RANK_UNKNOWN |
|
3204 | rank = RANK_UNKNOWN | |
3202 | if self.feature_config.compute_rank: |
|
3205 | if self.feature_config.compute_rank: | |
3203 | if (p1r, p2r) == (nullrev, nullrev): |
|
3206 | if (p1r, p2r) == (nullrev, nullrev): | |
3204 | rank = 1 |
|
3207 | rank = 1 | |
3205 | elif p1r != nullrev and p2r == nullrev: |
|
3208 | elif p1r != nullrev and p2r == nullrev: | |
3206 | rank = 1 + self.fast_rank(p1r) |
|
3209 | rank = 1 + self.fast_rank(p1r) | |
3207 | elif p1r == nullrev and p2r != nullrev: |
|
3210 | elif p1r == nullrev and p2r != nullrev: | |
3208 | rank = 1 + self.fast_rank(p2r) |
|
3211 | rank = 1 + self.fast_rank(p2r) | |
3209 | else: # merge node |
|
3212 | else: # merge node | |
3210 | if rustdagop is not None and self.index.rust_ext_compat: |
|
3213 | if rustdagop is not None and self.index.rust_ext_compat: | |
3211 | rank = rustdagop.rank(self.index, p1r, p2r) |
|
3214 | rank = rustdagop.rank(self.index, p1r, p2r) | |
3212 | else: |
|
3215 | else: | |
3213 | pmin, pmax = sorted((p1r, p2r)) |
|
3216 | pmin, pmax = sorted((p1r, p2r)) | |
3214 | rank = 1 + self.fast_rank(pmax) |
|
3217 | rank = 1 + self.fast_rank(pmax) | |
3215 | rank += sum(1 for _ in self.findmissingrevs([pmax], [pmin])) |
|
3218 | rank += sum(1 for _ in self.findmissingrevs([pmax], [pmin])) | |
3216 |
|
3219 | |||
3217 | e = revlogutils.entry( |
|
3220 | e = revlogutils.entry( | |
3218 | flags=flags, |
|
3221 | flags=flags, | |
3219 | data_offset=offset, |
|
3222 | data_offset=offset, | |
3220 | data_compressed_length=deltainfo.deltalen, |
|
3223 | data_compressed_length=deltainfo.deltalen, | |
3221 | data_uncompressed_length=textlen, |
|
3224 | data_uncompressed_length=textlen, | |
3222 | data_compression_mode=compression_mode, |
|
3225 | data_compression_mode=compression_mode, | |
3223 | data_delta_base=deltainfo.base, |
|
3226 | data_delta_base=deltainfo.base, | |
3224 | link_rev=link, |
|
3227 | link_rev=link, | |
3225 | parent_rev_1=p1r, |
|
3228 | parent_rev_1=p1r, | |
3226 | parent_rev_2=p2r, |
|
3229 | parent_rev_2=p2r, | |
3227 | node_id=node, |
|
3230 | node_id=node, | |
3228 | sidedata_offset=sidedata_offset, |
|
3231 | sidedata_offset=sidedata_offset, | |
3229 | sidedata_compressed_length=len(serialized_sidedata), |
|
3232 | sidedata_compressed_length=len(serialized_sidedata), | |
3230 | sidedata_compression_mode=sidedata_compression_mode, |
|
3233 | sidedata_compression_mode=sidedata_compression_mode, | |
3231 | rank=rank, |
|
3234 | rank=rank, | |
3232 | ) |
|
3235 | ) | |
3233 |
|
3236 | |||
3234 | self.index.append(e) |
|
3237 | self.index.append(e) | |
3235 | entry = self.index.entry_binary(curr) |
|
3238 | entry = self.index.entry_binary(curr) | |
3236 | if curr == 0 and self._docket is None: |
|
3239 | if curr == 0 and self._docket is None: | |
3237 | header = self._format_flags | self._format_version |
|
3240 | header = self._format_flags | self._format_version | |
3238 | header = self.index.pack_header(header) |
|
3241 | header = self.index.pack_header(header) | |
3239 | entry = header + entry |
|
3242 | entry = header + entry | |
3240 | self._writeentry( |
|
3243 | self._writeentry( | |
3241 | transaction, |
|
3244 | transaction, | |
3242 | entry, |
|
3245 | entry, | |
3243 | deltainfo.data, |
|
3246 | deltainfo.data, | |
3244 | link, |
|
3247 | link, | |
3245 | offset, |
|
3248 | offset, | |
3246 | serialized_sidedata, |
|
3249 | serialized_sidedata, | |
3247 | sidedata_offset, |
|
3250 | sidedata_offset, | |
3248 | ) |
|
3251 | ) | |
3249 |
|
3252 | |||
3250 | rawtext = btext[0] |
|
3253 | rawtext = btext[0] | |
3251 |
|
3254 | |||
3252 | if alwayscache and rawtext is None: |
|
3255 | if alwayscache and rawtext is None: | |
3253 | rawtext = deltacomputer.buildtext(revinfo) |
|
3256 | rawtext = deltacomputer.buildtext(revinfo) | |
3254 |
|
3257 | |||
3255 | if type(rawtext) == bytes: # only accept immutable objects |
|
3258 | if type(rawtext) == bytes: # only accept immutable objects | |
3256 | self._inner._revisioncache = (node, curr, rawtext) |
|
3259 | self._inner._revisioncache = (node, curr, rawtext) | |
3257 | self._chainbasecache[curr] = deltainfo.chainbase |
|
3260 | self._chainbasecache[curr] = deltainfo.chainbase | |
3258 | return curr |
|
3261 | return curr | |
3259 |
|
3262 | |||
3260 | def _get_data_offset(self, prev): |
|
3263 | def _get_data_offset(self, prev): | |
3261 | """Returns the current offset in the (in-transaction) data file. |
|
3264 | """Returns the current offset in the (in-transaction) data file. | |
3262 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket |
|
3265 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket | |
3263 | file to store that information: since sidedata can be rewritten to the |
|
3266 | file to store that information: since sidedata can be rewritten to the | |
3264 | end of the data file within a transaction, you can have cases where, for |
|
3267 | end of the data file within a transaction, you can have cases where, for | |
3265 | example, rev `n` does not have sidedata while rev `n - 1` does, leading |
|
3268 | example, rev `n` does not have sidedata while rev `n - 1` does, leading | |
3266 | to `n - 1`'s sidedata being written after `n`'s data. |
|
3269 | to `n - 1`'s sidedata being written after `n`'s data. | |
3267 |
|
3270 | |||
3268 | TODO cache this in a docket file before getting out of experimental.""" |
|
3271 | TODO cache this in a docket file before getting out of experimental.""" | |
3269 | if self._docket is None: |
|
3272 | if self._docket is None: | |
3270 | return self.end(prev) |
|
3273 | return self.end(prev) | |
3271 | else: |
|
3274 | else: | |
3272 | return self._docket.data_end |
|
3275 | return self._docket.data_end | |
3273 |
|
3276 | |||
3274 | def _writeentry( |
|
3277 | def _writeentry( | |
3275 | self, |
|
3278 | self, | |
3276 | transaction, |
|
3279 | transaction, | |
3277 | entry, |
|
3280 | entry, | |
3278 | data, |
|
3281 | data, | |
3279 | link, |
|
3282 | link, | |
3280 | offset, |
|
3283 | offset, | |
3281 | sidedata, |
|
3284 | sidedata, | |
3282 | sidedata_offset, |
|
3285 | sidedata_offset, | |
3283 | ): |
|
3286 | ): | |
3284 | # Files opened in a+ mode have inconsistent behavior on various |
|
3287 | # Files opened in a+ mode have inconsistent behavior on various | |
3285 | # platforms. Windows requires that a file positioning call be made |
|
3288 | # platforms. Windows requires that a file positioning call be made | |
3286 | # when the file handle transitions between reads and writes. See |
|
3289 | # when the file handle transitions between reads and writes. See | |
3287 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
3290 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other | |
3288 | # platforms, Python or the platform itself can be buggy. Some versions |
|
3291 | # platforms, Python or the platform itself can be buggy. Some versions | |
3289 | # of Solaris have been observed to not append at the end of the file |
|
3292 | # of Solaris have been observed to not append at the end of the file | |
3290 | # if the file was seeked to before the end. See issue4943 for more. |
|
3293 | # if the file was seeked to before the end. See issue4943 for more. | |
3291 | # |
|
3294 | # | |
3292 | # We work around this issue by inserting a seek() before writing. |
|
3295 | # We work around this issue by inserting a seek() before writing. | |
3293 | # Note: This is likely not necessary on Python 3. However, because |
|
3296 | # Note: This is likely not necessary on Python 3. However, because | |
3294 | # the file handle is reused for reads and may be seeked there, we need |
|
3297 | # the file handle is reused for reads and may be seeked there, we need | |
3295 | # to be careful before changing this. |
|
3298 | # to be careful before changing this. | |
3296 | index_end = data_end = sidedata_end = None |
|
3299 | index_end = data_end = sidedata_end = None | |
3297 | if self._docket is not None: |
|
3300 | if self._docket is not None: | |
3298 | index_end = self._docket.index_end |
|
3301 | index_end = self._docket.index_end | |
3299 | data_end = self._docket.data_end |
|
3302 | data_end = self._docket.data_end | |
3300 | sidedata_end = self._docket.sidedata_end |
|
3303 | sidedata_end = self._docket.sidedata_end | |
3301 |
|
3304 | |||
3302 | files_end = self._inner.write_entry( |
|
3305 | files_end = self._inner.write_entry( | |
3303 | transaction, |
|
3306 | transaction, | |
3304 | entry, |
|
3307 | entry, | |
3305 | data, |
|
3308 | data, | |
3306 | link, |
|
3309 | link, | |
3307 | offset, |
|
3310 | offset, | |
3308 | sidedata, |
|
3311 | sidedata, | |
3309 | sidedata_offset, |
|
3312 | sidedata_offset, | |
3310 | index_end, |
|
3313 | index_end, | |
3311 | data_end, |
|
3314 | data_end, | |
3312 | sidedata_end, |
|
3315 | sidedata_end, | |
3313 | ) |
|
3316 | ) | |
3314 | self._enforceinlinesize(transaction) |
|
3317 | self._enforceinlinesize(transaction) | |
3315 | if self._docket is not None: |
|
3318 | if self._docket is not None: | |
3316 | self._docket.index_end = files_end[0] |
|
3319 | self._docket.index_end = files_end[0] | |
3317 | self._docket.data_end = files_end[1] |
|
3320 | self._docket.data_end = files_end[1] | |
3318 | self._docket.sidedata_end = files_end[2] |
|
3321 | self._docket.sidedata_end = files_end[2] | |
3319 |
|
3322 | |||
3320 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
3323 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
3321 |
|
3324 | |||
3322 | def addgroup( |
|
3325 | def addgroup( | |
3323 | self, |
|
3326 | self, | |
3324 | deltas, |
|
3327 | deltas, | |
3325 | linkmapper, |
|
3328 | linkmapper, | |
3326 | transaction, |
|
3329 | transaction, | |
3327 | alwayscache=False, |
|
3330 | alwayscache=False, | |
3328 | addrevisioncb=None, |
|
3331 | addrevisioncb=None, | |
3329 | duplicaterevisioncb=None, |
|
3332 | duplicaterevisioncb=None, | |
3330 | debug_info=None, |
|
3333 | debug_info=None, | |
3331 | delta_base_reuse_policy=None, |
|
3334 | delta_base_reuse_policy=None, | |
3332 | ): |
|
3335 | ): | |
3333 | """ |
|
3336 | """ | |
3334 | add a delta group |
|
3337 | add a delta group | |
3335 |
|
3338 | |||
3336 | given a set of deltas, add them to the revision log. the |
|
3339 | given a set of deltas, add them to the revision log. the | |
3337 | first delta is against its parent, which should be in our |
|
3340 | first delta is against its parent, which should be in our | |
3338 | log, the rest are against the previous delta. |
|
3341 | log, the rest are against the previous delta. | |
3339 |
|
3342 | |||
3340 | If ``addrevisioncb`` is defined, it will be called with arguments of |
|
3343 | If ``addrevisioncb`` is defined, it will be called with arguments of | |
3341 | this revlog and the node that was added. |
|
3344 | this revlog and the node that was added. | |
3342 | """ |
|
3345 | """ | |
3343 |
|
3346 | |||
3344 | if self._adding_group: |
|
3347 | if self._adding_group: | |
3345 | raise error.ProgrammingError(b'cannot nest addgroup() calls') |
|
3348 | raise error.ProgrammingError(b'cannot nest addgroup() calls') | |
3346 |
|
3349 | |||
3347 | # read the default delta-base reuse policy from revlog config if the |
|
3350 | # read the default delta-base reuse policy from revlog config if the | |
3348 | # group did not specify one. |
|
3351 | # group did not specify one. | |
3349 | if delta_base_reuse_policy is None: |
|
3352 | if delta_base_reuse_policy is None: | |
3350 | if ( |
|
3353 | if ( | |
3351 | self.delta_config.general_delta |
|
3354 | self.delta_config.general_delta | |
3352 | and self.delta_config.lazy_delta_base |
|
3355 | and self.delta_config.lazy_delta_base | |
3353 | ): |
|
3356 | ): | |
3354 | delta_base_reuse_policy = DELTA_BASE_REUSE_TRY |
|
3357 | delta_base_reuse_policy = DELTA_BASE_REUSE_TRY | |
3355 | else: |
|
3358 | else: | |
3356 | delta_base_reuse_policy = DELTA_BASE_REUSE_NO |
|
3359 | delta_base_reuse_policy = DELTA_BASE_REUSE_NO | |
3357 |
|
3360 | |||
3358 | self._adding_group = True |
|
3361 | self._adding_group = True | |
3359 | empty = True |
|
3362 | empty = True | |
3360 | try: |
|
3363 | try: | |
3361 | with self._writing(transaction): |
|
3364 | with self._writing(transaction): | |
3362 | write_debug = None |
|
3365 | write_debug = None | |
3363 | if self.delta_config.debug_delta: |
|
3366 | if self.delta_config.debug_delta: | |
3364 | write_debug = transaction._report |
|
3367 | write_debug = transaction._report | |
3365 | deltacomputer = deltautil.deltacomputer( |
|
3368 | deltacomputer = deltautil.deltacomputer( | |
3366 | self, |
|
3369 | self, | |
3367 | write_debug=write_debug, |
|
3370 | write_debug=write_debug, | |
3368 | debug_info=debug_info, |
|
3371 | debug_info=debug_info, | |
3369 | ) |
|
3372 | ) | |
3370 | # loop through our set of deltas |
|
3373 | # loop through our set of deltas | |
3371 | for data in deltas: |
|
3374 | for data in deltas: | |
3372 | ( |
|
3375 | ( | |
3373 | node, |
|
3376 | node, | |
3374 | p1, |
|
3377 | p1, | |
3375 | p2, |
|
3378 | p2, | |
3376 | linknode, |
|
3379 | linknode, | |
3377 | deltabase, |
|
3380 | deltabase, | |
3378 | delta, |
|
3381 | delta, | |
3379 | flags, |
|
3382 | flags, | |
3380 | sidedata, |
|
3383 | sidedata, | |
3381 | ) = data |
|
3384 | ) = data | |
3382 | link = linkmapper(linknode) |
|
3385 | link = linkmapper(linknode) | |
3383 | flags = flags or REVIDX_DEFAULT_FLAGS |
|
3386 | flags = flags or REVIDX_DEFAULT_FLAGS | |
3384 |
|
3387 | |||
3385 | rev = self.index.get_rev(node) |
|
3388 | rev = self.index.get_rev(node) | |
3386 | if rev is not None: |
|
3389 | if rev is not None: | |
3387 | # this can happen if two branches make the same change |
|
3390 | # this can happen if two branches make the same change | |
3388 | self._nodeduplicatecallback(transaction, rev) |
|
3391 | self._nodeduplicatecallback(transaction, rev) | |
3389 | if duplicaterevisioncb: |
|
3392 | if duplicaterevisioncb: | |
3390 | duplicaterevisioncb(self, rev) |
|
3393 | duplicaterevisioncb(self, rev) | |
3391 | empty = False |
|
3394 | empty = False | |
3392 | continue |
|
3395 | continue | |
3393 |
|
3396 | |||
3394 | for p in (p1, p2): |
|
3397 | for p in (p1, p2): | |
3395 | if not self.index.has_node(p): |
|
3398 | if not self.index.has_node(p): | |
3396 | raise error.LookupError( |
|
3399 | raise error.LookupError( | |
3397 | p, self.radix, _(b'unknown parent') |
|
3400 | p, self.radix, _(b'unknown parent') | |
3398 | ) |
|
3401 | ) | |
3399 |
|
3402 | |||
3400 | if not self.index.has_node(deltabase): |
|
3403 | if not self.index.has_node(deltabase): | |
3401 | raise error.LookupError( |
|
3404 | raise error.LookupError( | |
3402 | deltabase, self.display_id, _(b'unknown delta base') |
|
3405 | deltabase, self.display_id, _(b'unknown delta base') | |
3403 | ) |
|
3406 | ) | |
3404 |
|
3407 | |||
3405 | baserev = self.rev(deltabase) |
|
3408 | baserev = self.rev(deltabase) | |
3406 |
|
3409 | |||
3407 | if baserev != nullrev and self.iscensored(baserev): |
|
3410 | if baserev != nullrev and self.iscensored(baserev): | |
3408 | # if base is censored, delta must be full replacement in a |
|
3411 | # if base is censored, delta must be full replacement in a | |
3409 | # single patch operation |
|
3412 | # single patch operation | |
3410 | hlen = struct.calcsize(b">lll") |
|
3413 | hlen = struct.calcsize(b">lll") | |
3411 | oldlen = self.rawsize(baserev) |
|
3414 | oldlen = self.rawsize(baserev) | |
3412 | newlen = len(delta) - hlen |
|
3415 | newlen = len(delta) - hlen | |
3413 | if delta[:hlen] != mdiff.replacediffheader( |
|
3416 | if delta[:hlen] != mdiff.replacediffheader( | |
3414 | oldlen, newlen |
|
3417 | oldlen, newlen | |
3415 | ): |
|
3418 | ): | |
3416 | raise error.CensoredBaseError( |
|
3419 | raise error.CensoredBaseError( | |
3417 | self.display_id, self.node(baserev) |
|
3420 | self.display_id, self.node(baserev) | |
3418 | ) |
|
3421 | ) | |
3419 |
|
3422 | |||
3420 | if not flags and self._peek_iscensored(baserev, delta): |
|
3423 | if not flags and self._peek_iscensored(baserev, delta): | |
3421 | flags |= REVIDX_ISCENSORED |
|
3424 | flags |= REVIDX_ISCENSORED | |
3422 |
|
3425 | |||
3423 | # We assume consumers of addrevisioncb will want to retrieve |
|
3426 | # We assume consumers of addrevisioncb will want to retrieve | |
3424 | # the added revision, which will require a call to |
|
3427 | # the added revision, which will require a call to | |
3425 | # revision(). revision() will fast path if there is a cache |
|
3428 | # revision(). revision() will fast path if there is a cache | |
3426 | # hit. So, we tell _addrevision() to always cache in this case. |
|
3429 | # hit. So, we tell _addrevision() to always cache in this case. | |
3427 | # We're only using addgroup() in the context of changegroup |
|
3430 | # We're only using addgroup() in the context of changegroup | |
3428 | # generation so the revision data can always be handled as raw |
|
3431 | # generation so the revision data can always be handled as raw | |
3429 | # by the flagprocessor. |
|
3432 | # by the flagprocessor. | |
3430 | rev = self._addrevision( |
|
3433 | rev = self._addrevision( | |
3431 | node, |
|
3434 | node, | |
3432 | None, |
|
3435 | None, | |
3433 | transaction, |
|
3436 | transaction, | |
3434 | link, |
|
3437 | link, | |
3435 | p1, |
|
3438 | p1, | |
3436 | p2, |
|
3439 | p2, | |
3437 | flags, |
|
3440 | flags, | |
3438 | (baserev, delta, delta_base_reuse_policy), |
|
3441 | (baserev, delta, delta_base_reuse_policy), | |
3439 | alwayscache=alwayscache, |
|
3442 | alwayscache=alwayscache, | |
3440 | deltacomputer=deltacomputer, |
|
3443 | deltacomputer=deltacomputer, | |
3441 | sidedata=sidedata, |
|
3444 | sidedata=sidedata, | |
3442 | ) |
|
3445 | ) | |
3443 |
|
3446 | |||
3444 | if addrevisioncb: |
|
3447 | if addrevisioncb: | |
3445 | addrevisioncb(self, rev) |
|
3448 | addrevisioncb(self, rev) | |
3446 | empty = False |
|
3449 | empty = False | |
3447 | finally: |
|
3450 | finally: | |
3448 | self._adding_group = False |
|
3451 | self._adding_group = False | |
3449 | return not empty |
|
3452 | return not empty | |
3450 |
|
3453 | |||
3451 | def iscensored(self, rev): |
|
3454 | def iscensored(self, rev): | |
3452 | """Check if a file revision is censored.""" |
|
3455 | """Check if a file revision is censored.""" | |
3453 | if not self.feature_config.censorable: |
|
3456 | if not self.feature_config.censorable: | |
3454 | return False |
|
3457 | return False | |
3455 |
|
3458 | |||
3456 | return self.flags(rev) & REVIDX_ISCENSORED |
|
3459 | return self.flags(rev) & REVIDX_ISCENSORED | |
3457 |
|
3460 | |||
3458 | def _peek_iscensored(self, baserev, delta): |
|
3461 | def _peek_iscensored(self, baserev, delta): | |
3459 | """Quickly check if a delta produces a censored revision.""" |
|
3462 | """Quickly check if a delta produces a censored revision.""" | |
3460 | if not self.feature_config.censorable: |
|
3463 | if not self.feature_config.censorable: | |
3461 | return False |
|
3464 | return False | |
3462 |
|
3465 | |||
3463 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) |
|
3466 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) | |
3464 |
|
3467 | |||
3465 | def getstrippoint(self, minlink): |
|
3468 | def getstrippoint(self, minlink): | |
3466 | """find the minimum rev that must be stripped to strip the linkrev |
|
3469 | """find the minimum rev that must be stripped to strip the linkrev | |
3467 |
|
3470 | |||
3468 | Returns a tuple containing the minimum rev and a set of all revs that |
|
3471 | Returns a tuple containing the minimum rev and a set of all revs that | |
3469 | have linkrevs that will be broken by this strip. |
|
3472 | have linkrevs that will be broken by this strip. | |
3470 | """ |
|
3473 | """ | |
3471 | return storageutil.resolvestripinfo( |
|
3474 | return storageutil.resolvestripinfo( | |
3472 | minlink, |
|
3475 | minlink, | |
3473 | len(self) - 1, |
|
3476 | len(self) - 1, | |
3474 | self.headrevs(), |
|
3477 | self.headrevs(), | |
3475 | self.linkrev, |
|
3478 | self.linkrev, | |
3476 | self.parentrevs, |
|
3479 | self.parentrevs, | |
3477 | ) |
|
3480 | ) | |
3478 |
|
3481 | |||
3479 | def strip(self, minlink, transaction): |
|
3482 | def strip(self, minlink, transaction): | |
3480 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
3483 | """truncate the revlog on the first revision with a linkrev >= minlink | |
3481 |
|
3484 | |||
3482 | This function is called when we're stripping revision minlink and |
|
3485 | This function is called when we're stripping revision minlink and | |
3483 | its descendants from the repository. |
|
3486 | its descendants from the repository. | |
3484 |
|
3487 | |||
3485 | We have to remove all revisions with linkrev >= minlink, because |
|
3488 | We have to remove all revisions with linkrev >= minlink, because | |
3486 | the equivalent changelog revisions will be renumbered after the |
|
3489 | the equivalent changelog revisions will be renumbered after the | |
3487 | strip. |
|
3490 | strip. | |
3488 |
|
3491 | |||
3489 | So we truncate the revlog on the first of these revisions, and |
|
3492 | So we truncate the revlog on the first of these revisions, and | |
3490 | trust that the caller has saved the revisions that shouldn't be |
|
3493 | trust that the caller has saved the revisions that shouldn't be | |
3491 | removed and that it'll re-add them after this truncation. |
|
3494 | removed and that it'll re-add them after this truncation. | |
3492 | """ |
|
3495 | """ | |
3493 | if len(self) == 0: |
|
3496 | if len(self) == 0: | |
3494 | return |
|
3497 | return | |
3495 |
|
3498 | |||
3496 | rev, _ = self.getstrippoint(minlink) |
|
3499 | rev, _ = self.getstrippoint(minlink) | |
3497 | if rev == len(self): |
|
3500 | if rev == len(self): | |
3498 | return |
|
3501 | return | |
3499 |
|
3502 | |||
3500 | # first truncate the files on disk |
|
3503 | # first truncate the files on disk | |
3501 | data_end = self.start(rev) |
|
3504 | data_end = self.start(rev) | |
3502 | if not self._inline: |
|
3505 | if not self._inline: | |
3503 | transaction.add(self._datafile, data_end) |
|
3506 | transaction.add(self._datafile, data_end) | |
3504 | end = rev * self.index.entry_size |
|
3507 | end = rev * self.index.entry_size | |
3505 | else: |
|
3508 | else: | |
3506 | end = data_end + (rev * self.index.entry_size) |
|
3509 | end = data_end + (rev * self.index.entry_size) | |
3507 |
|
3510 | |||
3508 | if self._sidedatafile: |
|
3511 | if self._sidedatafile: | |
3509 | sidedata_end = self.sidedata_cut_off(rev) |
|
3512 | sidedata_end = self.sidedata_cut_off(rev) | |
3510 | transaction.add(self._sidedatafile, sidedata_end) |
|
3513 | transaction.add(self._sidedatafile, sidedata_end) | |
3511 |
|
3514 | |||
3512 | transaction.add(self._indexfile, end) |
|
3515 | transaction.add(self._indexfile, end) | |
3513 | if self._docket is not None: |
|
3516 | if self._docket is not None: | |
3514 | # XXX we could, leverage the docket while stripping. However it is |
|
3517 | # XXX we could, leverage the docket while stripping. However it is | |
3515 | # not powerfull enough at the time of this comment |
|
3518 | # not powerfull enough at the time of this comment | |
3516 | self._docket.index_end = end |
|
3519 | self._docket.index_end = end | |
3517 | self._docket.data_end = data_end |
|
3520 | self._docket.data_end = data_end | |
3518 | self._docket.sidedata_end = sidedata_end |
|
3521 | self._docket.sidedata_end = sidedata_end | |
3519 | self._docket.write(transaction, stripping=True) |
|
3522 | self._docket.write(transaction, stripping=True) | |
3520 |
|
3523 | |||
3521 | # then reset internal state in memory to forget those revisions |
|
3524 | # then reset internal state in memory to forget those revisions | |
3522 | self._chaininfocache = util.lrucachedict(500) |
|
3525 | self._chaininfocache = util.lrucachedict(500) | |
3523 | self._inner.clear_cache() |
|
3526 | self._inner.clear_cache() | |
3524 |
|
3527 | |||
3525 | del self.index[rev:-1] |
|
3528 | del self.index[rev:-1] | |
3526 |
|
3529 | |||
3527 | def checksize(self): |
|
3530 | def checksize(self): | |
3528 | """Check size of index and data files |
|
3531 | """Check size of index and data files | |
3529 |
|
3532 | |||
3530 | return a (dd, di) tuple. |
|
3533 | return a (dd, di) tuple. | |
3531 | - dd: extra bytes for the "data" file |
|
3534 | - dd: extra bytes for the "data" file | |
3532 | - di: extra bytes for the "index" file |
|
3535 | - di: extra bytes for the "index" file | |
3533 |
|
3536 | |||
3534 | A healthy revlog will return (0, 0). |
|
3537 | A healthy revlog will return (0, 0). | |
3535 | """ |
|
3538 | """ | |
3536 | expected = 0 |
|
3539 | expected = 0 | |
3537 | if len(self): |
|
3540 | if len(self): | |
3538 | expected = max(0, self.end(len(self) - 1)) |
|
3541 | expected = max(0, self.end(len(self) - 1)) | |
3539 |
|
3542 | |||
3540 | try: |
|
3543 | try: | |
3541 | with self._datafp() as f: |
|
3544 | with self._datafp() as f: | |
3542 | f.seek(0, io.SEEK_END) |
|
3545 | f.seek(0, io.SEEK_END) | |
3543 | actual = f.tell() |
|
3546 | actual = f.tell() | |
3544 | dd = actual - expected |
|
3547 | dd = actual - expected | |
3545 | except FileNotFoundError: |
|
3548 | except FileNotFoundError: | |
3546 | dd = 0 |
|
3549 | dd = 0 | |
3547 |
|
3550 | |||
3548 | try: |
|
3551 | try: | |
3549 | f = self.opener(self._indexfile) |
|
3552 | f = self.opener(self._indexfile) | |
3550 | f.seek(0, io.SEEK_END) |
|
3553 | f.seek(0, io.SEEK_END) | |
3551 | actual = f.tell() |
|
3554 | actual = f.tell() | |
3552 | f.close() |
|
3555 | f.close() | |
3553 | s = self.index.entry_size |
|
3556 | s = self.index.entry_size | |
3554 | i = max(0, actual // s) |
|
3557 | i = max(0, actual // s) | |
3555 | di = actual - (i * s) |
|
3558 | di = actual - (i * s) | |
3556 | if self._inline: |
|
3559 | if self._inline: | |
3557 | databytes = 0 |
|
3560 | databytes = 0 | |
3558 | for r in self: |
|
3561 | for r in self: | |
3559 | databytes += max(0, self.length(r)) |
|
3562 | databytes += max(0, self.length(r)) | |
3560 | dd = 0 |
|
3563 | dd = 0 | |
3561 | di = actual - len(self) * s - databytes |
|
3564 | di = actual - len(self) * s - databytes | |
3562 | except FileNotFoundError: |
|
3565 | except FileNotFoundError: | |
3563 | di = 0 |
|
3566 | di = 0 | |
3564 |
|
3567 | |||
3565 | return (dd, di) |
|
3568 | return (dd, di) | |
3566 |
|
3569 | |||
3567 | def files(self): |
|
3570 | def files(self): | |
3568 | """return list of files that compose this revlog""" |
|
3571 | """return list of files that compose this revlog""" | |
3569 | res = [self._indexfile] |
|
3572 | res = [self._indexfile] | |
3570 | if self._docket_file is None: |
|
3573 | if self._docket_file is None: | |
3571 | if not self._inline: |
|
3574 | if not self._inline: | |
3572 | res.append(self._datafile) |
|
3575 | res.append(self._datafile) | |
3573 | else: |
|
3576 | else: | |
3574 | res.append(self._docket_file) |
|
3577 | res.append(self._docket_file) | |
3575 | res.extend(self._docket.old_index_filepaths(include_empty=False)) |
|
3578 | res.extend(self._docket.old_index_filepaths(include_empty=False)) | |
3576 | if self._docket.data_end: |
|
3579 | if self._docket.data_end: | |
3577 | res.append(self._datafile) |
|
3580 | res.append(self._datafile) | |
3578 | res.extend(self._docket.old_data_filepaths(include_empty=False)) |
|
3581 | res.extend(self._docket.old_data_filepaths(include_empty=False)) | |
3579 | if self._docket.sidedata_end: |
|
3582 | if self._docket.sidedata_end: | |
3580 | res.append(self._sidedatafile) |
|
3583 | res.append(self._sidedatafile) | |
3581 | res.extend(self._docket.old_sidedata_filepaths(include_empty=False)) |
|
3584 | res.extend(self._docket.old_sidedata_filepaths(include_empty=False)) | |
3582 | return res |
|
3585 | return res | |
3583 |
|
3586 | |||
3584 | def emitrevisions( |
|
3587 | def emitrevisions( | |
3585 | self, |
|
3588 | self, | |
3586 | nodes, |
|
3589 | nodes, | |
3587 | nodesorder=None, |
|
3590 | nodesorder=None, | |
3588 | revisiondata=False, |
|
3591 | revisiondata=False, | |
3589 | assumehaveparentrevisions=False, |
|
3592 | assumehaveparentrevisions=False, | |
3590 | deltamode=repository.CG_DELTAMODE_STD, |
|
3593 | deltamode=repository.CG_DELTAMODE_STD, | |
3591 | sidedata_helpers=None, |
|
3594 | sidedata_helpers=None, | |
3592 | debug_info=None, |
|
3595 | debug_info=None, | |
3593 | ): |
|
3596 | ): | |
3594 | if nodesorder not in (b'nodes', b'storage', b'linear', None): |
|
3597 | if nodesorder not in (b'nodes', b'storage', b'linear', None): | |
3595 | raise error.ProgrammingError( |
|
3598 | raise error.ProgrammingError( | |
3596 | b'unhandled value for nodesorder: %s' % nodesorder |
|
3599 | b'unhandled value for nodesorder: %s' % nodesorder | |
3597 | ) |
|
3600 | ) | |
3598 |
|
3601 | |||
3599 | if nodesorder is None and not self.delta_config.general_delta: |
|
3602 | if nodesorder is None and not self.delta_config.general_delta: | |
3600 | nodesorder = b'storage' |
|
3603 | nodesorder = b'storage' | |
3601 |
|
3604 | |||
3602 | if ( |
|
3605 | if ( | |
3603 | not self._storedeltachains |
|
3606 | not self._storedeltachains | |
3604 | and deltamode != repository.CG_DELTAMODE_PREV |
|
3607 | and deltamode != repository.CG_DELTAMODE_PREV | |
3605 | ): |
|
3608 | ): | |
3606 | deltamode = repository.CG_DELTAMODE_FULL |
|
3609 | deltamode = repository.CG_DELTAMODE_FULL | |
3607 |
|
3610 | |||
3608 | return storageutil.emitrevisions( |
|
3611 | return storageutil.emitrevisions( | |
3609 | self, |
|
3612 | self, | |
3610 | nodes, |
|
3613 | nodes, | |
3611 | nodesorder, |
|
3614 | nodesorder, | |
3612 | revlogrevisiondelta, |
|
3615 | revlogrevisiondelta, | |
3613 | deltaparentfn=self.deltaparent, |
|
3616 | deltaparentfn=self.deltaparent, | |
3614 | candeltafn=self._candelta, |
|
3617 | candeltafn=self._candelta, | |
3615 | rawsizefn=self.rawsize, |
|
3618 | rawsizefn=self.rawsize, | |
3616 | revdifffn=self.revdiff, |
|
3619 | revdifffn=self.revdiff, | |
3617 | flagsfn=self.flags, |
|
3620 | flagsfn=self.flags, | |
3618 | deltamode=deltamode, |
|
3621 | deltamode=deltamode, | |
3619 | revisiondata=revisiondata, |
|
3622 | revisiondata=revisiondata, | |
3620 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
3623 | assumehaveparentrevisions=assumehaveparentrevisions, | |
3621 | sidedata_helpers=sidedata_helpers, |
|
3624 | sidedata_helpers=sidedata_helpers, | |
3622 | debug_info=debug_info, |
|
3625 | debug_info=debug_info, | |
3623 | ) |
|
3626 | ) | |
3624 |
|
3627 | |||
3625 | DELTAREUSEALWAYS = b'always' |
|
3628 | DELTAREUSEALWAYS = b'always' | |
3626 | DELTAREUSESAMEREVS = b'samerevs' |
|
3629 | DELTAREUSESAMEREVS = b'samerevs' | |
3627 | DELTAREUSENEVER = b'never' |
|
3630 | DELTAREUSENEVER = b'never' | |
3628 |
|
3631 | |||
3629 | DELTAREUSEFULLADD = b'fulladd' |
|
3632 | DELTAREUSEFULLADD = b'fulladd' | |
3630 |
|
3633 | |||
3631 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} |
|
3634 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} | |
3632 |
|
3635 | |||
3633 | def clone( |
|
3636 | def clone( | |
3634 | self, |
|
3637 | self, | |
3635 | tr, |
|
3638 | tr, | |
3636 | destrevlog, |
|
3639 | destrevlog, | |
3637 | addrevisioncb=None, |
|
3640 | addrevisioncb=None, | |
3638 | deltareuse=DELTAREUSESAMEREVS, |
|
3641 | deltareuse=DELTAREUSESAMEREVS, | |
3639 | forcedeltabothparents=None, |
|
3642 | forcedeltabothparents=None, | |
3640 | sidedata_helpers=None, |
|
3643 | sidedata_helpers=None, | |
3641 | ): |
|
3644 | ): | |
3642 | """Copy this revlog to another, possibly with format changes. |
|
3645 | """Copy this revlog to another, possibly with format changes. | |
3643 |
|
3646 | |||
3644 | The destination revlog will contain the same revisions and nodes. |
|
3647 | The destination revlog will contain the same revisions and nodes. | |
3645 | However, it may not be bit-for-bit identical due to e.g. delta encoding |
|
3648 | However, it may not be bit-for-bit identical due to e.g. delta encoding | |
3646 | differences. |
|
3649 | differences. | |
3647 |
|
3650 | |||
3648 | The ``deltareuse`` argument control how deltas from the existing revlog |
|
3651 | The ``deltareuse`` argument control how deltas from the existing revlog | |
3649 | are preserved in the destination revlog. The argument can have the |
|
3652 | are preserved in the destination revlog. The argument can have the | |
3650 | following values: |
|
3653 | following values: | |
3651 |
|
3654 | |||
3652 | DELTAREUSEALWAYS |
|
3655 | DELTAREUSEALWAYS | |
3653 | Deltas will always be reused (if possible), even if the destination |
|
3656 | Deltas will always be reused (if possible), even if the destination | |
3654 | revlog would not select the same revisions for the delta. This is the |
|
3657 | revlog would not select the same revisions for the delta. This is the | |
3655 | fastest mode of operation. |
|
3658 | fastest mode of operation. | |
3656 | DELTAREUSESAMEREVS |
|
3659 | DELTAREUSESAMEREVS | |
3657 | Deltas will be reused if the destination revlog would pick the same |
|
3660 | Deltas will be reused if the destination revlog would pick the same | |
3658 | revisions for the delta. This mode strikes a balance between speed |
|
3661 | revisions for the delta. This mode strikes a balance between speed | |
3659 | and optimization. |
|
3662 | and optimization. | |
3660 | DELTAREUSENEVER |
|
3663 | DELTAREUSENEVER | |
3661 | Deltas will never be reused. This is the slowest mode of execution. |
|
3664 | Deltas will never be reused. This is the slowest mode of execution. | |
3662 | This mode can be used to recompute deltas (e.g. if the diff/delta |
|
3665 | This mode can be used to recompute deltas (e.g. if the diff/delta | |
3663 | algorithm changes). |
|
3666 | algorithm changes). | |
3664 | DELTAREUSEFULLADD |
|
3667 | DELTAREUSEFULLADD | |
3665 | Revision will be re-added as if their were new content. This is |
|
3668 | Revision will be re-added as if their were new content. This is | |
3666 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. |
|
3669 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. | |
3667 | eg: large file detection and handling. |
|
3670 | eg: large file detection and handling. | |
3668 |
|
3671 | |||
3669 | Delta computation can be slow, so the choice of delta reuse policy can |
|
3672 | Delta computation can be slow, so the choice of delta reuse policy can | |
3670 | significantly affect run time. |
|
3673 | significantly affect run time. | |
3671 |
|
3674 | |||
3672 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between |
|
3675 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between | |
3673 | two extremes. Deltas will be reused if they are appropriate. But if the |
|
3676 | two extremes. Deltas will be reused if they are appropriate. But if the | |
3674 | delta could choose a better revision, it will do so. This means if you |
|
3677 | delta could choose a better revision, it will do so. This means if you | |
3675 | are converting a non-generaldelta revlog to a generaldelta revlog, |
|
3678 | are converting a non-generaldelta revlog to a generaldelta revlog, | |
3676 | deltas will be recomputed if the delta's parent isn't a parent of the |
|
3679 | deltas will be recomputed if the delta's parent isn't a parent of the | |
3677 | revision. |
|
3680 | revision. | |
3678 |
|
3681 | |||
3679 | In addition to the delta policy, the ``forcedeltabothparents`` |
|
3682 | In addition to the delta policy, the ``forcedeltabothparents`` | |
3680 | argument controls whether to force compute deltas against both parents |
|
3683 | argument controls whether to force compute deltas against both parents | |
3681 | for merges. By default, the current default is used. |
|
3684 | for merges. By default, the current default is used. | |
3682 |
|
3685 | |||
3683 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on |
|
3686 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on | |
3684 | `sidedata_helpers`. |
|
3687 | `sidedata_helpers`. | |
3685 | """ |
|
3688 | """ | |
3686 | if deltareuse not in self.DELTAREUSEALL: |
|
3689 | if deltareuse not in self.DELTAREUSEALL: | |
3687 | raise ValueError( |
|
3690 | raise ValueError( | |
3688 | _(b'value for deltareuse invalid: %s') % deltareuse |
|
3691 | _(b'value for deltareuse invalid: %s') % deltareuse | |
3689 | ) |
|
3692 | ) | |
3690 |
|
3693 | |||
3691 | if len(destrevlog): |
|
3694 | if len(destrevlog): | |
3692 | raise ValueError(_(b'destination revlog is not empty')) |
|
3695 | raise ValueError(_(b'destination revlog is not empty')) | |
3693 |
|
3696 | |||
3694 | if getattr(self, 'filteredrevs', None): |
|
3697 | if getattr(self, 'filteredrevs', None): | |
3695 | raise ValueError(_(b'source revlog has filtered revisions')) |
|
3698 | raise ValueError(_(b'source revlog has filtered revisions')) | |
3696 | if getattr(destrevlog, 'filteredrevs', None): |
|
3699 | if getattr(destrevlog, 'filteredrevs', None): | |
3697 | raise ValueError(_(b'destination revlog has filtered revisions')) |
|
3700 | raise ValueError(_(b'destination revlog has filtered revisions')) | |
3698 |
|
3701 | |||
3699 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, |
|
3702 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, | |
3700 | # if possible. |
|
3703 | # if possible. | |
3701 | old_delta_config = destrevlog.delta_config |
|
3704 | old_delta_config = destrevlog.delta_config | |
3702 | destrevlog.delta_config = destrevlog.delta_config.copy() |
|
3705 | destrevlog.delta_config = destrevlog.delta_config.copy() | |
3703 |
|
3706 | |||
3704 | try: |
|
3707 | try: | |
3705 | if deltareuse == self.DELTAREUSEALWAYS: |
|
3708 | if deltareuse == self.DELTAREUSEALWAYS: | |
3706 | destrevlog.delta_config.lazy_delta_base = True |
|
3709 | destrevlog.delta_config.lazy_delta_base = True | |
3707 | destrevlog.delta_config.lazy_delta = True |
|
3710 | destrevlog.delta_config.lazy_delta = True | |
3708 | elif deltareuse == self.DELTAREUSESAMEREVS: |
|
3711 | elif deltareuse == self.DELTAREUSESAMEREVS: | |
3709 | destrevlog.delta_config.lazy_delta_base = False |
|
3712 | destrevlog.delta_config.lazy_delta_base = False | |
3710 | destrevlog.delta_config.lazy_delta = True |
|
3713 | destrevlog.delta_config.lazy_delta = True | |
3711 | elif deltareuse == self.DELTAREUSENEVER: |
|
3714 | elif deltareuse == self.DELTAREUSENEVER: | |
3712 | destrevlog.delta_config.lazy_delta_base = False |
|
3715 | destrevlog.delta_config.lazy_delta_base = False | |
3713 | destrevlog.delta_config.lazy_delta = False |
|
3716 | destrevlog.delta_config.lazy_delta = False | |
3714 |
|
3717 | |||
3715 | delta_both_parents = ( |
|
3718 | delta_both_parents = ( | |
3716 | forcedeltabothparents or old_delta_config.delta_both_parents |
|
3719 | forcedeltabothparents or old_delta_config.delta_both_parents | |
3717 | ) |
|
3720 | ) | |
3718 | destrevlog.delta_config.delta_both_parents = delta_both_parents |
|
3721 | destrevlog.delta_config.delta_both_parents = delta_both_parents | |
3719 |
|
3722 | |||
3720 | with self.reading(), destrevlog._writing(tr): |
|
3723 | with self.reading(), destrevlog._writing(tr): | |
3721 | self._clone( |
|
3724 | self._clone( | |
3722 | tr, |
|
3725 | tr, | |
3723 | destrevlog, |
|
3726 | destrevlog, | |
3724 | addrevisioncb, |
|
3727 | addrevisioncb, | |
3725 | deltareuse, |
|
3728 | deltareuse, | |
3726 | forcedeltabothparents, |
|
3729 | forcedeltabothparents, | |
3727 | sidedata_helpers, |
|
3730 | sidedata_helpers, | |
3728 | ) |
|
3731 | ) | |
3729 |
|
3732 | |||
3730 | finally: |
|
3733 | finally: | |
3731 | destrevlog.delta_config = old_delta_config |
|
3734 | destrevlog.delta_config = old_delta_config | |
3732 |
|
3735 | |||
3733 | def _clone( |
|
3736 | def _clone( | |
3734 | self, |
|
3737 | self, | |
3735 | tr, |
|
3738 | tr, | |
3736 | destrevlog, |
|
3739 | destrevlog, | |
3737 | addrevisioncb, |
|
3740 | addrevisioncb, | |
3738 | deltareuse, |
|
3741 | deltareuse, | |
3739 | forcedeltabothparents, |
|
3742 | forcedeltabothparents, | |
3740 | sidedata_helpers, |
|
3743 | sidedata_helpers, | |
3741 | ): |
|
3744 | ): | |
3742 | """perform the core duty of `revlog.clone` after parameter processing""" |
|
3745 | """perform the core duty of `revlog.clone` after parameter processing""" | |
3743 | write_debug = None |
|
3746 | write_debug = None | |
3744 | if self.delta_config.debug_delta: |
|
3747 | if self.delta_config.debug_delta: | |
3745 | write_debug = tr._report |
|
3748 | write_debug = tr._report | |
3746 | deltacomputer = deltautil.deltacomputer( |
|
3749 | deltacomputer = deltautil.deltacomputer( | |
3747 | destrevlog, |
|
3750 | destrevlog, | |
3748 | write_debug=write_debug, |
|
3751 | write_debug=write_debug, | |
3749 | ) |
|
3752 | ) | |
3750 | index = self.index |
|
3753 | index = self.index | |
3751 | for rev in self: |
|
3754 | for rev in self: | |
3752 | entry = index[rev] |
|
3755 | entry = index[rev] | |
3753 |
|
3756 | |||
3754 | # Some classes override linkrev to take filtered revs into |
|
3757 | # Some classes override linkrev to take filtered revs into | |
3755 | # account. Use raw entry from index. |
|
3758 | # account. Use raw entry from index. | |
3756 | flags = entry[0] & 0xFFFF |
|
3759 | flags = entry[0] & 0xFFFF | |
3757 | linkrev = entry[4] |
|
3760 | linkrev = entry[4] | |
3758 | p1 = index[entry[5]][7] |
|
3761 | p1 = index[entry[5]][7] | |
3759 | p2 = index[entry[6]][7] |
|
3762 | p2 = index[entry[6]][7] | |
3760 | node = entry[7] |
|
3763 | node = entry[7] | |
3761 |
|
3764 | |||
3762 | # (Possibly) reuse the delta from the revlog if allowed and |
|
3765 | # (Possibly) reuse the delta from the revlog if allowed and | |
3763 | # the revlog chunk is a delta. |
|
3766 | # the revlog chunk is a delta. | |
3764 | cachedelta = None |
|
3767 | cachedelta = None | |
3765 | rawtext = None |
|
3768 | rawtext = None | |
3766 | if deltareuse == self.DELTAREUSEFULLADD: |
|
3769 | if deltareuse == self.DELTAREUSEFULLADD: | |
3767 | text = self._revisiondata(rev) |
|
3770 | text = self._revisiondata(rev) | |
3768 | sidedata = self.sidedata(rev) |
|
3771 | sidedata = self.sidedata(rev) | |
3769 |
|
3772 | |||
3770 | if sidedata_helpers is not None: |
|
3773 | if sidedata_helpers is not None: | |
3771 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
3774 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
3772 | self, sidedata_helpers, sidedata, rev |
|
3775 | self, sidedata_helpers, sidedata, rev | |
3773 | ) |
|
3776 | ) | |
3774 | flags = flags | new_flags[0] & ~new_flags[1] |
|
3777 | flags = flags | new_flags[0] & ~new_flags[1] | |
3775 |
|
3778 | |||
3776 | destrevlog.addrevision( |
|
3779 | destrevlog.addrevision( | |
3777 | text, |
|
3780 | text, | |
3778 | tr, |
|
3781 | tr, | |
3779 | linkrev, |
|
3782 | linkrev, | |
3780 | p1, |
|
3783 | p1, | |
3781 | p2, |
|
3784 | p2, | |
3782 | cachedelta=cachedelta, |
|
3785 | cachedelta=cachedelta, | |
3783 | node=node, |
|
3786 | node=node, | |
3784 | flags=flags, |
|
3787 | flags=flags, | |
3785 | deltacomputer=deltacomputer, |
|
3788 | deltacomputer=deltacomputer, | |
3786 | sidedata=sidedata, |
|
3789 | sidedata=sidedata, | |
3787 | ) |
|
3790 | ) | |
3788 | else: |
|
3791 | else: | |
3789 | if destrevlog.delta_config.lazy_delta: |
|
3792 | if destrevlog.delta_config.lazy_delta: | |
3790 | dp = self.deltaparent(rev) |
|
3793 | dp = self.deltaparent(rev) | |
3791 | if dp != nullrev: |
|
3794 | if dp != nullrev: | |
3792 | cachedelta = (dp, bytes(self._inner._chunk(rev))) |
|
3795 | cachedelta = (dp, bytes(self._inner._chunk(rev))) | |
3793 |
|
3796 | |||
3794 | sidedata = None |
|
3797 | sidedata = None | |
3795 | if not cachedelta: |
|
3798 | if not cachedelta: | |
3796 | try: |
|
3799 | try: | |
3797 | rawtext = self._revisiondata(rev) |
|
3800 | rawtext = self._revisiondata(rev) | |
3798 | except error.CensoredNodeError as censored: |
|
3801 | except error.CensoredNodeError as censored: | |
3799 | assert flags & REVIDX_ISCENSORED |
|
3802 | assert flags & REVIDX_ISCENSORED | |
3800 | rawtext = censored.tombstone |
|
3803 | rawtext = censored.tombstone | |
3801 | sidedata = self.sidedata(rev) |
|
3804 | sidedata = self.sidedata(rev) | |
3802 | if sidedata is None: |
|
3805 | if sidedata is None: | |
3803 | sidedata = self.sidedata(rev) |
|
3806 | sidedata = self.sidedata(rev) | |
3804 |
|
3807 | |||
3805 | if sidedata_helpers is not None: |
|
3808 | if sidedata_helpers is not None: | |
3806 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
3809 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
3807 | self, sidedata_helpers, sidedata, rev |
|
3810 | self, sidedata_helpers, sidedata, rev | |
3808 | ) |
|
3811 | ) | |
3809 | flags = flags | new_flags[0] & ~new_flags[1] |
|
3812 | flags = flags | new_flags[0] & ~new_flags[1] | |
3810 |
|
3813 | |||
3811 | destrevlog._addrevision( |
|
3814 | destrevlog._addrevision( | |
3812 | node, |
|
3815 | node, | |
3813 | rawtext, |
|
3816 | rawtext, | |
3814 | tr, |
|
3817 | tr, | |
3815 | linkrev, |
|
3818 | linkrev, | |
3816 | p1, |
|
3819 | p1, | |
3817 | p2, |
|
3820 | p2, | |
3818 | flags, |
|
3821 | flags, | |
3819 | cachedelta, |
|
3822 | cachedelta, | |
3820 | deltacomputer=deltacomputer, |
|
3823 | deltacomputer=deltacomputer, | |
3821 | sidedata=sidedata, |
|
3824 | sidedata=sidedata, | |
3822 | ) |
|
3825 | ) | |
3823 |
|
3826 | |||
3824 | if addrevisioncb: |
|
3827 | if addrevisioncb: | |
3825 | addrevisioncb(self, rev, node) |
|
3828 | addrevisioncb(self, rev, node) | |
3826 |
|
3829 | |||
3827 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
3830 | def censorrevision(self, tr, censornode, tombstone=b''): | |
3828 | if self._format_version == REVLOGV0: |
|
3831 | if self._format_version == REVLOGV0: | |
3829 | raise error.RevlogError( |
|
3832 | raise error.RevlogError( | |
3830 | _(b'cannot censor with version %d revlogs') |
|
3833 | _(b'cannot censor with version %d revlogs') | |
3831 | % self._format_version |
|
3834 | % self._format_version | |
3832 | ) |
|
3835 | ) | |
3833 | elif self._format_version == REVLOGV1: |
|
3836 | elif self._format_version == REVLOGV1: | |
3834 | rewrite.v1_censor(self, tr, censornode, tombstone) |
|
3837 | rewrite.v1_censor(self, tr, censornode, tombstone) | |
3835 | else: |
|
3838 | else: | |
3836 | rewrite.v2_censor(self, tr, censornode, tombstone) |
|
3839 | rewrite.v2_censor(self, tr, censornode, tombstone) | |
3837 |
|
3840 | |||
3838 | def verifyintegrity(self, state): |
|
3841 | def verifyintegrity(self, state): | |
3839 | """Verifies the integrity of the revlog. |
|
3842 | """Verifies the integrity of the revlog. | |
3840 |
|
3843 | |||
3841 | Yields ``revlogproblem`` instances describing problems that are |
|
3844 | Yields ``revlogproblem`` instances describing problems that are | |
3842 | found. |
|
3845 | found. | |
3843 | """ |
|
3846 | """ | |
3844 | dd, di = self.checksize() |
|
3847 | dd, di = self.checksize() | |
3845 | if dd: |
|
3848 | if dd: | |
3846 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) |
|
3849 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) | |
3847 | if di: |
|
3850 | if di: | |
3848 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) |
|
3851 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) | |
3849 |
|
3852 | |||
3850 | version = self._format_version |
|
3853 | version = self._format_version | |
3851 |
|
3854 | |||
3852 | # The verifier tells us what version revlog we should be. |
|
3855 | # The verifier tells us what version revlog we should be. | |
3853 | if version != state[b'expectedversion']: |
|
3856 | if version != state[b'expectedversion']: | |
3854 | yield revlogproblem( |
|
3857 | yield revlogproblem( | |
3855 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") |
|
3858 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") | |
3856 | % (self.display_id, version, state[b'expectedversion']) |
|
3859 | % (self.display_id, version, state[b'expectedversion']) | |
3857 | ) |
|
3860 | ) | |
3858 |
|
3861 | |||
3859 | state[b'skipread'] = set() |
|
3862 | state[b'skipread'] = set() | |
3860 | state[b'safe_renamed'] = set() |
|
3863 | state[b'safe_renamed'] = set() | |
3861 |
|
3864 | |||
3862 | for rev in self: |
|
3865 | for rev in self: | |
3863 | node = self.node(rev) |
|
3866 | node = self.node(rev) | |
3864 |
|
3867 | |||
3865 | # Verify contents. 4 cases to care about: |
|
3868 | # Verify contents. 4 cases to care about: | |
3866 | # |
|
3869 | # | |
3867 | # common: the most common case |
|
3870 | # common: the most common case | |
3868 | # rename: with a rename |
|
3871 | # rename: with a rename | |
3869 | # meta: file content starts with b'\1\n', the metadata |
|
3872 | # meta: file content starts with b'\1\n', the metadata | |
3870 | # header defined in filelog.py, but without a rename |
|
3873 | # header defined in filelog.py, but without a rename | |
3871 | # ext: content stored externally |
|
3874 | # ext: content stored externally | |
3872 | # |
|
3875 | # | |
3873 | # More formally, their differences are shown below: |
|
3876 | # More formally, their differences are shown below: | |
3874 | # |
|
3877 | # | |
3875 | # | common | rename | meta | ext |
|
3878 | # | common | rename | meta | ext | |
3876 | # ------------------------------------------------------- |
|
3879 | # ------------------------------------------------------- | |
3877 | # flags() | 0 | 0 | 0 | not 0 |
|
3880 | # flags() | 0 | 0 | 0 | not 0 | |
3878 | # renamed() | False | True | False | ? |
|
3881 | # renamed() | False | True | False | ? | |
3879 | # rawtext[0:2]=='\1\n'| False | True | True | ? |
|
3882 | # rawtext[0:2]=='\1\n'| False | True | True | ? | |
3880 | # |
|
3883 | # | |
3881 | # "rawtext" means the raw text stored in revlog data, which |
|
3884 | # "rawtext" means the raw text stored in revlog data, which | |
3882 | # could be retrieved by "rawdata(rev)". "text" |
|
3885 | # could be retrieved by "rawdata(rev)". "text" | |
3883 | # mentioned below is "revision(rev)". |
|
3886 | # mentioned below is "revision(rev)". | |
3884 | # |
|
3887 | # | |
3885 | # There are 3 different lengths stored physically: |
|
3888 | # There are 3 different lengths stored physically: | |
3886 | # 1. L1: rawsize, stored in revlog index |
|
3889 | # 1. L1: rawsize, stored in revlog index | |
3887 | # 2. L2: len(rawtext), stored in revlog data |
|
3890 | # 2. L2: len(rawtext), stored in revlog data | |
3888 | # 3. L3: len(text), stored in revlog data if flags==0, or |
|
3891 | # 3. L3: len(text), stored in revlog data if flags==0, or | |
3889 | # possibly somewhere else if flags!=0 |
|
3892 | # possibly somewhere else if flags!=0 | |
3890 | # |
|
3893 | # | |
3891 | # L1 should be equal to L2. L3 could be different from them. |
|
3894 | # L1 should be equal to L2. L3 could be different from them. | |
3892 | # "text" may or may not affect commit hash depending on flag |
|
3895 | # "text" may or may not affect commit hash depending on flag | |
3893 | # processors (see flagutil.addflagprocessor). |
|
3896 | # processors (see flagutil.addflagprocessor). | |
3894 | # |
|
3897 | # | |
3895 | # | common | rename | meta | ext |
|
3898 | # | common | rename | meta | ext | |
3896 | # ------------------------------------------------- |
|
3899 | # ------------------------------------------------- | |
3897 | # rawsize() | L1 | L1 | L1 | L1 |
|
3900 | # rawsize() | L1 | L1 | L1 | L1 | |
3898 | # size() | L1 | L2-LM | L1(*) | L1 (?) |
|
3901 | # size() | L1 | L2-LM | L1(*) | L1 (?) | |
3899 | # len(rawtext) | L2 | L2 | L2 | L2 |
|
3902 | # len(rawtext) | L2 | L2 | L2 | L2 | |
3900 | # len(text) | L2 | L2 | L2 | L3 |
|
3903 | # len(text) | L2 | L2 | L2 | L3 | |
3901 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) |
|
3904 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) | |
3902 | # |
|
3905 | # | |
3903 | # LM: length of metadata, depending on rawtext |
|
3906 | # LM: length of metadata, depending on rawtext | |
3904 | # (*): not ideal, see comment in filelog.size |
|
3907 | # (*): not ideal, see comment in filelog.size | |
3905 | # (?): could be "- len(meta)" if the resolved content has |
|
3908 | # (?): could be "- len(meta)" if the resolved content has | |
3906 | # rename metadata |
|
3909 | # rename metadata | |
3907 | # |
|
3910 | # | |
3908 | # Checks needed to be done: |
|
3911 | # Checks needed to be done: | |
3909 | # 1. length check: L1 == L2, in all cases. |
|
3912 | # 1. length check: L1 == L2, in all cases. | |
3910 | # 2. hash check: depending on flag processor, we may need to |
|
3913 | # 2. hash check: depending on flag processor, we may need to | |
3911 | # use either "text" (external), or "rawtext" (in revlog). |
|
3914 | # use either "text" (external), or "rawtext" (in revlog). | |
3912 |
|
3915 | |||
3913 | try: |
|
3916 | try: | |
3914 | skipflags = state.get(b'skipflags', 0) |
|
3917 | skipflags = state.get(b'skipflags', 0) | |
3915 | if skipflags: |
|
3918 | if skipflags: | |
3916 | skipflags &= self.flags(rev) |
|
3919 | skipflags &= self.flags(rev) | |
3917 |
|
3920 | |||
3918 | _verify_revision(self, skipflags, state, node) |
|
3921 | _verify_revision(self, skipflags, state, node) | |
3919 |
|
3922 | |||
3920 | l1 = self.rawsize(rev) |
|
3923 | l1 = self.rawsize(rev) | |
3921 | l2 = len(self.rawdata(node)) |
|
3924 | l2 = len(self.rawdata(node)) | |
3922 |
|
3925 | |||
3923 | if l1 != l2: |
|
3926 | if l1 != l2: | |
3924 | yield revlogproblem( |
|
3927 | yield revlogproblem( | |
3925 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), |
|
3928 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), | |
3926 | node=node, |
|
3929 | node=node, | |
3927 | ) |
|
3930 | ) | |
3928 |
|
3931 | |||
3929 | except error.CensoredNodeError: |
|
3932 | except error.CensoredNodeError: | |
3930 | if state[b'erroroncensored']: |
|
3933 | if state[b'erroroncensored']: | |
3931 | yield revlogproblem( |
|
3934 | yield revlogproblem( | |
3932 | error=_(b'censored file data'), node=node |
|
3935 | error=_(b'censored file data'), node=node | |
3933 | ) |
|
3936 | ) | |
3934 | state[b'skipread'].add(node) |
|
3937 | state[b'skipread'].add(node) | |
3935 | except Exception as e: |
|
3938 | except Exception as e: | |
3936 | yield revlogproblem( |
|
3939 | yield revlogproblem( | |
3937 | error=_(b'unpacking %s: %s') |
|
3940 | error=_(b'unpacking %s: %s') | |
3938 | % (short(node), stringutil.forcebytestr(e)), |
|
3941 | % (short(node), stringutil.forcebytestr(e)), | |
3939 | node=node, |
|
3942 | node=node, | |
3940 | ) |
|
3943 | ) | |
3941 | state[b'skipread'].add(node) |
|
3944 | state[b'skipread'].add(node) | |
3942 |
|
3945 | |||
3943 | def storageinfo( |
|
3946 | def storageinfo( | |
3944 | self, |
|
3947 | self, | |
3945 | exclusivefiles=False, |
|
3948 | exclusivefiles=False, | |
3946 | sharedfiles=False, |
|
3949 | sharedfiles=False, | |
3947 | revisionscount=False, |
|
3950 | revisionscount=False, | |
3948 | trackedsize=False, |
|
3951 | trackedsize=False, | |
3949 | storedsize=False, |
|
3952 | storedsize=False, | |
3950 | ): |
|
3953 | ): | |
3951 | d = {} |
|
3954 | d = {} | |
3952 |
|
3955 | |||
3953 | if exclusivefiles: |
|
3956 | if exclusivefiles: | |
3954 | d[b'exclusivefiles'] = [(self.opener, self._indexfile)] |
|
3957 | d[b'exclusivefiles'] = [(self.opener, self._indexfile)] | |
3955 | if not self._inline: |
|
3958 | if not self._inline: | |
3956 | d[b'exclusivefiles'].append((self.opener, self._datafile)) |
|
3959 | d[b'exclusivefiles'].append((self.opener, self._datafile)) | |
3957 |
|
3960 | |||
3958 | if sharedfiles: |
|
3961 | if sharedfiles: | |
3959 | d[b'sharedfiles'] = [] |
|
3962 | d[b'sharedfiles'] = [] | |
3960 |
|
3963 | |||
3961 | if revisionscount: |
|
3964 | if revisionscount: | |
3962 | d[b'revisionscount'] = len(self) |
|
3965 | d[b'revisionscount'] = len(self) | |
3963 |
|
3966 | |||
3964 | if trackedsize: |
|
3967 | if trackedsize: | |
3965 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) |
|
3968 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) | |
3966 |
|
3969 | |||
3967 | if storedsize: |
|
3970 | if storedsize: | |
3968 | d[b'storedsize'] = sum( |
|
3971 | d[b'storedsize'] = sum( | |
3969 | self.opener.stat(path).st_size for path in self.files() |
|
3972 | self.opener.stat(path).st_size for path in self.files() | |
3970 | ) |
|
3973 | ) | |
3971 |
|
3974 | |||
3972 | return d |
|
3975 | return d | |
3973 |
|
3976 | |||
3974 | def rewrite_sidedata(self, transaction, helpers, startrev, endrev): |
|
3977 | def rewrite_sidedata(self, transaction, helpers, startrev, endrev): | |
3975 | if not self.feature_config.has_side_data: |
|
3978 | if not self.feature_config.has_side_data: | |
3976 | return |
|
3979 | return | |
3977 | # revlog formats with sidedata support does not support inline |
|
3980 | # revlog formats with sidedata support does not support inline | |
3978 | assert not self._inline |
|
3981 | assert not self._inline | |
3979 | if not helpers[1] and not helpers[2]: |
|
3982 | if not helpers[1] and not helpers[2]: | |
3980 | # Nothing to generate or remove |
|
3983 | # Nothing to generate or remove | |
3981 | return |
|
3984 | return | |
3982 |
|
3985 | |||
3983 | new_entries = [] |
|
3986 | new_entries = [] | |
3984 | # append the new sidedata |
|
3987 | # append the new sidedata | |
3985 | with self._writing(transaction): |
|
3988 | with self._writing(transaction): | |
3986 | ifh, dfh, sdfh = self._inner._writinghandles |
|
3989 | ifh, dfh, sdfh = self._inner._writinghandles | |
3987 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) |
|
3990 | dfh.seek(self._docket.sidedata_end, os.SEEK_SET) | |
3988 |
|
3991 | |||
3989 | current_offset = sdfh.tell() |
|
3992 | current_offset = sdfh.tell() | |
3990 | for rev in range(startrev, endrev + 1): |
|
3993 | for rev in range(startrev, endrev + 1): | |
3991 | entry = self.index[rev] |
|
3994 | entry = self.index[rev] | |
3992 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( |
|
3995 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( | |
3993 | store=self, |
|
3996 | store=self, | |
3994 | sidedata_helpers=helpers, |
|
3997 | sidedata_helpers=helpers, | |
3995 | sidedata={}, |
|
3998 | sidedata={}, | |
3996 | rev=rev, |
|
3999 | rev=rev, | |
3997 | ) |
|
4000 | ) | |
3998 |
|
4001 | |||
3999 | serialized_sidedata = sidedatautil.serialize_sidedata( |
|
4002 | serialized_sidedata = sidedatautil.serialize_sidedata( | |
4000 | new_sidedata |
|
4003 | new_sidedata | |
4001 | ) |
|
4004 | ) | |
4002 |
|
4005 | |||
4003 | sidedata_compression_mode = COMP_MODE_INLINE |
|
4006 | sidedata_compression_mode = COMP_MODE_INLINE | |
4004 | if serialized_sidedata and self.feature_config.has_side_data: |
|
4007 | if serialized_sidedata and self.feature_config.has_side_data: | |
4005 | sidedata_compression_mode = COMP_MODE_PLAIN |
|
4008 | sidedata_compression_mode = COMP_MODE_PLAIN | |
4006 | h, comp_sidedata = self._inner.compress(serialized_sidedata) |
|
4009 | h, comp_sidedata = self._inner.compress(serialized_sidedata) | |
4007 | if ( |
|
4010 | if ( | |
4008 | h != b'u' |
|
4011 | h != b'u' | |
4009 | and comp_sidedata[0] != b'\0' |
|
4012 | and comp_sidedata[0] != b'\0' | |
4010 | and len(comp_sidedata) < len(serialized_sidedata) |
|
4013 | and len(comp_sidedata) < len(serialized_sidedata) | |
4011 | ): |
|
4014 | ): | |
4012 | assert not h |
|
4015 | assert not h | |
4013 | if ( |
|
4016 | if ( | |
4014 | comp_sidedata[0] |
|
4017 | comp_sidedata[0] | |
4015 | == self._docket.default_compression_header |
|
4018 | == self._docket.default_compression_header | |
4016 | ): |
|
4019 | ): | |
4017 | sidedata_compression_mode = COMP_MODE_DEFAULT |
|
4020 | sidedata_compression_mode = COMP_MODE_DEFAULT | |
4018 | serialized_sidedata = comp_sidedata |
|
4021 | serialized_sidedata = comp_sidedata | |
4019 | else: |
|
4022 | else: | |
4020 | sidedata_compression_mode = COMP_MODE_INLINE |
|
4023 | sidedata_compression_mode = COMP_MODE_INLINE | |
4021 | serialized_sidedata = comp_sidedata |
|
4024 | serialized_sidedata = comp_sidedata | |
4022 | if entry[8] != 0 or entry[9] != 0: |
|
4025 | if entry[8] != 0 or entry[9] != 0: | |
4023 | # rewriting entries that already have sidedata is not |
|
4026 | # rewriting entries that already have sidedata is not | |
4024 | # supported yet, because it introduces garbage data in the |
|
4027 | # supported yet, because it introduces garbage data in the | |
4025 | # revlog. |
|
4028 | # revlog. | |
4026 | msg = b"rewriting existing sidedata is not supported yet" |
|
4029 | msg = b"rewriting existing sidedata is not supported yet" | |
4027 | raise error.Abort(msg) |
|
4030 | raise error.Abort(msg) | |
4028 |
|
4031 | |||
4029 | # Apply (potential) flags to add and to remove after running |
|
4032 | # Apply (potential) flags to add and to remove after running | |
4030 | # the sidedata helpers |
|
4033 | # the sidedata helpers | |
4031 | new_offset_flags = entry[0] | flags[0] & ~flags[1] |
|
4034 | new_offset_flags = entry[0] | flags[0] & ~flags[1] | |
4032 | entry_update = ( |
|
4035 | entry_update = ( | |
4033 | current_offset, |
|
4036 | current_offset, | |
4034 | len(serialized_sidedata), |
|
4037 | len(serialized_sidedata), | |
4035 | new_offset_flags, |
|
4038 | new_offset_flags, | |
4036 | sidedata_compression_mode, |
|
4039 | sidedata_compression_mode, | |
4037 | ) |
|
4040 | ) | |
4038 |
|
4041 | |||
4039 | # the sidedata computation might have move the file cursors around |
|
4042 | # the sidedata computation might have move the file cursors around | |
4040 | sdfh.seek(current_offset, os.SEEK_SET) |
|
4043 | sdfh.seek(current_offset, os.SEEK_SET) | |
4041 | sdfh.write(serialized_sidedata) |
|
4044 | sdfh.write(serialized_sidedata) | |
4042 | new_entries.append(entry_update) |
|
4045 | new_entries.append(entry_update) | |
4043 | current_offset += len(serialized_sidedata) |
|
4046 | current_offset += len(serialized_sidedata) | |
4044 | self._docket.sidedata_end = sdfh.tell() |
|
4047 | self._docket.sidedata_end = sdfh.tell() | |
4045 |
|
4048 | |||
4046 | # rewrite the new index entries |
|
4049 | # rewrite the new index entries | |
4047 | ifh.seek(startrev * self.index.entry_size) |
|
4050 | ifh.seek(startrev * self.index.entry_size) | |
4048 | for i, e in enumerate(new_entries): |
|
4051 | for i, e in enumerate(new_entries): | |
4049 | rev = startrev + i |
|
4052 | rev = startrev + i | |
4050 | self.index.replace_sidedata_info(rev, *e) |
|
4053 | self.index.replace_sidedata_info(rev, *e) | |
4051 | packed = self.index.entry_binary(rev) |
|
4054 | packed = self.index.entry_binary(rev) | |
4052 | if rev == 0 and self._docket is None: |
|
4055 | if rev == 0 and self._docket is None: | |
4053 | header = self._format_flags | self._format_version |
|
4056 | header = self._format_flags | self._format_version | |
4054 | header = self.index.pack_header(header) |
|
4057 | header = self.index.pack_header(header) | |
4055 | packed = header + packed |
|
4058 | packed = header + packed | |
4056 | ifh.write(packed) |
|
4059 | ifh.write(packed) |
@@ -1,38 +1,38 b'' | |||||
1 | // debugdata.rs |
|
1 | // debugdata.rs | |
2 | // |
|
2 | // | |
3 | // Copyright 2020 Antoine Cezar <antoine.cezar@octobus.net> |
|
3 | // Copyright 2020 Antoine Cezar <antoine.cezar@octobus.net> | |
4 | // |
|
4 | // | |
5 | // This software may be used and distributed according to the terms of the |
|
5 | // This software may be used and distributed according to the terms of the | |
6 | // GNU General Public License version 2 or any later version. |
|
6 | // GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | use crate::repo::Repo; |
|
8 | use crate::repo::Repo; | |
9 | use crate::requirements; |
|
|||
10 | use crate::revlog::{Revlog, RevlogError}; |
|
9 | use crate::revlog::{Revlog, RevlogError}; | |
11 |
|
10 | |||
12 | /// Kind of data to debug |
|
11 | /// Kind of data to debug | |
13 | #[derive(Debug, Copy, Clone)] |
|
12 | #[derive(Debug, Copy, Clone, PartialEq, Eq)] | |
14 | pub enum DebugDataKind { |
|
13 | pub enum DebugDataKind { | |
15 | Changelog, |
|
14 | Changelog, | |
16 | Manifest, |
|
15 | Manifest, | |
17 | } |
|
16 | } | |
18 |
|
17 | |||
19 | /// Dump the contents data of a revision. |
|
18 | /// Dump the contents data of a revision. | |
20 | pub fn debug_data( |
|
19 | pub fn debug_data( | |
21 | repo: &Repo, |
|
20 | repo: &Repo, | |
22 | revset: &str, |
|
21 | revset: &str, | |
23 | kind: DebugDataKind, |
|
22 | kind: DebugDataKind, | |
24 | ) -> Result<Vec<u8>, RevlogError> { |
|
23 | ) -> Result<Vec<u8>, RevlogError> { | |
25 | let index_file = match kind { |
|
24 | let index_file = match kind { | |
26 | DebugDataKind::Changelog => "00changelog.i", |
|
25 | DebugDataKind::Changelog => "00changelog.i", | |
27 | DebugDataKind::Manifest => "00manifest.i", |
|
26 | DebugDataKind::Manifest => "00manifest.i", | |
28 | }; |
|
27 | }; | |
29 | let use_nodemap = repo |
|
28 | let revlog = Revlog::open( | |
30 |
. |
|
29 | &repo.store_vfs(), | |
31 | .contains(requirements::NODEMAP_REQUIREMENT); |
|
30 | index_file, | |
32 | let revlog = |
|
31 | None, | |
33 | Revlog::open(&repo.store_vfs(), index_file, None, use_nodemap)?; |
|
32 | repo.default_revlog_options(kind == DebugDataKind::Changelog)?, | |
|
33 | )?; | |||
34 | let rev = |
|
34 | let rev = | |
35 | crate::revset::resolve_rev_number_or_hex_prefix(revset, &revlog)?; |
|
35 | crate::revset::resolve_rev_number_or_hex_prefix(revset, &revlog)?; | |
36 | let data = revlog.get_rev_data_for_checked_rev(rev)?; |
|
36 | let data = revlog.get_rev_data_for_checked_rev(rev)?; | |
37 | Ok(data.into_owned()) |
|
37 | Ok(data.into_owned()) | |
38 | } |
|
38 | } |
@@ -1,782 +1,820 b'' | |||||
1 | use crate::changelog::Changelog; |
|
1 | use crate::changelog::Changelog; | |
2 | use crate::config::{Config, ConfigError, ConfigParseError}; |
|
2 | use crate::config::{Config, ConfigError, ConfigParseError}; | |
3 | use crate::dirstate::DirstateParents; |
|
3 | use crate::dirstate::DirstateParents; | |
4 | use crate::dirstate_tree::dirstate_map::DirstateMapWriteMode; |
|
4 | use crate::dirstate_tree::dirstate_map::DirstateMapWriteMode; | |
5 | use crate::dirstate_tree::on_disk::Docket as DirstateDocket; |
|
5 | use crate::dirstate_tree::on_disk::Docket as DirstateDocket; | |
6 | use crate::dirstate_tree::owning::OwningDirstateMap; |
|
6 | use crate::dirstate_tree::owning::OwningDirstateMap; | |
7 | use crate::errors::HgResultExt; |
|
7 | use crate::errors::HgResultExt; | |
8 | use crate::errors::{HgError, IoResultExt}; |
|
8 | use crate::errors::{HgError, IoResultExt}; | |
9 | use crate::lock::{try_with_lock_no_wait, LockError}; |
|
9 | use crate::lock::{try_with_lock_no_wait, LockError}; | |
10 | use crate::manifest::{Manifest, Manifestlog}; |
|
10 | use crate::manifest::{Manifest, Manifestlog}; | |
|
11 | use crate::requirements::{ | |||
|
12 | CHANGELOGV2_REQUIREMENT, GENERALDELTA_REQUIREMENT, NODEMAP_REQUIREMENT, | |||
|
13 | REVLOGV1_REQUIREMENT, REVLOGV2_REQUIREMENT, | |||
|
14 | }; | |||
11 | use crate::revlog::filelog::Filelog; |
|
15 | use crate::revlog::filelog::Filelog; | |
12 | use crate::revlog::RevlogError; |
|
16 | use crate::revlog::RevlogError; | |
13 | use crate::utils::debug::debug_wait_for_file_or_print; |
|
17 | use crate::utils::debug::debug_wait_for_file_or_print; | |
14 | use crate::utils::files::get_path_from_bytes; |
|
18 | use crate::utils::files::get_path_from_bytes; | |
15 | use crate::utils::hg_path::HgPath; |
|
19 | use crate::utils::hg_path::HgPath; | |
16 | use crate::utils::SliceExt; |
|
20 | use crate::utils::SliceExt; | |
17 | use crate::vfs::{is_dir, is_file, Vfs}; |
|
21 | use crate::vfs::{is_dir, is_file, Vfs}; | |
18 | use crate::DirstateError; |
|
22 | use crate::{ | |
19 |
|
|
23 | requirements, NodePrefix, RevlogVersionOptions, UncheckedRevision, | |
|
24 | }; | |||
|
25 | use crate::{DirstateError, RevlogOpenOptions}; | |||
20 | use std::cell::{Ref, RefCell, RefMut}; |
|
26 | use std::cell::{Ref, RefCell, RefMut}; | |
21 | use std::collections::HashSet; |
|
27 | use std::collections::HashSet; | |
22 | use std::io::Seek; |
|
28 | use std::io::Seek; | |
23 | use std::io::SeekFrom; |
|
29 | use std::io::SeekFrom; | |
24 | use std::io::Write as IoWrite; |
|
30 | use std::io::Write as IoWrite; | |
25 | use std::path::{Path, PathBuf}; |
|
31 | use std::path::{Path, PathBuf}; | |
26 |
|
32 | |||
27 | const V2_MAX_READ_ATTEMPTS: usize = 5; |
|
33 | const V2_MAX_READ_ATTEMPTS: usize = 5; | |
28 |
|
34 | |||
29 | type DirstateMapIdentity = (Option<u64>, Option<Vec<u8>>, usize); |
|
35 | type DirstateMapIdentity = (Option<u64>, Option<Vec<u8>>, usize); | |
30 |
|
36 | |||
31 | /// A repository on disk |
|
37 | /// A repository on disk | |
32 | pub struct Repo { |
|
38 | pub struct Repo { | |
33 | working_directory: PathBuf, |
|
39 | working_directory: PathBuf, | |
34 | dot_hg: PathBuf, |
|
40 | dot_hg: PathBuf, | |
35 | store: PathBuf, |
|
41 | store: PathBuf, | |
36 | requirements: HashSet<String>, |
|
42 | requirements: HashSet<String>, | |
37 | config: Config, |
|
43 | config: Config, | |
38 | dirstate_parents: LazyCell<DirstateParents>, |
|
44 | dirstate_parents: LazyCell<DirstateParents>, | |
39 | dirstate_map: LazyCell<OwningDirstateMap>, |
|
45 | dirstate_map: LazyCell<OwningDirstateMap>, | |
40 | changelog: LazyCell<Changelog>, |
|
46 | changelog: LazyCell<Changelog>, | |
41 | manifestlog: LazyCell<Manifestlog>, |
|
47 | manifestlog: LazyCell<Manifestlog>, | |
42 | } |
|
48 | } | |
43 |
|
49 | |||
44 | #[derive(Debug, derive_more::From)] |
|
50 | #[derive(Debug, derive_more::From)] | |
45 | pub enum RepoError { |
|
51 | pub enum RepoError { | |
46 | NotFound { |
|
52 | NotFound { | |
47 | at: PathBuf, |
|
53 | at: PathBuf, | |
48 | }, |
|
54 | }, | |
49 | #[from] |
|
55 | #[from] | |
50 | ConfigParseError(ConfigParseError), |
|
56 | ConfigParseError(ConfigParseError), | |
51 | #[from] |
|
57 | #[from] | |
52 | Other(HgError), |
|
58 | Other(HgError), | |
53 | } |
|
59 | } | |
54 |
|
60 | |||
55 | impl From<ConfigError> for RepoError { |
|
61 | impl From<ConfigError> for RepoError { | |
56 | fn from(error: ConfigError) -> Self { |
|
62 | fn from(error: ConfigError) -> Self { | |
57 | match error { |
|
63 | match error { | |
58 | ConfigError::Parse(error) => error.into(), |
|
64 | ConfigError::Parse(error) => error.into(), | |
59 | ConfigError::Other(error) => error.into(), |
|
65 | ConfigError::Other(error) => error.into(), | |
60 | } |
|
66 | } | |
61 | } |
|
67 | } | |
62 | } |
|
68 | } | |
63 |
|
69 | |||
64 | impl Repo { |
|
70 | impl Repo { | |
65 | /// tries to find nearest repository root in current working directory or |
|
71 | /// tries to find nearest repository root in current working directory or | |
66 | /// its ancestors |
|
72 | /// its ancestors | |
67 | pub fn find_repo_root() -> Result<PathBuf, RepoError> { |
|
73 | pub fn find_repo_root() -> Result<PathBuf, RepoError> { | |
68 | let current_directory = crate::utils::current_dir()?; |
|
74 | let current_directory = crate::utils::current_dir()?; | |
69 | // ancestors() is inclusive: it first yields `current_directory` |
|
75 | // ancestors() is inclusive: it first yields `current_directory` | |
70 | // as-is. |
|
76 | // as-is. | |
71 | for ancestor in current_directory.ancestors() { |
|
77 | for ancestor in current_directory.ancestors() { | |
72 | if is_dir(ancestor.join(".hg"))? { |
|
78 | if is_dir(ancestor.join(".hg"))? { | |
73 | return Ok(ancestor.to_path_buf()); |
|
79 | return Ok(ancestor.to_path_buf()); | |
74 | } |
|
80 | } | |
75 | } |
|
81 | } | |
76 | Err(RepoError::NotFound { |
|
82 | Err(RepoError::NotFound { | |
77 | at: current_directory, |
|
83 | at: current_directory, | |
78 | }) |
|
84 | }) | |
79 | } |
|
85 | } | |
80 |
|
86 | |||
81 | /// Find a repository, either at the given path (which must contain a `.hg` |
|
87 | /// Find a repository, either at the given path (which must contain a `.hg` | |
82 | /// sub-directory) or by searching the current directory and its |
|
88 | /// sub-directory) or by searching the current directory and its | |
83 | /// ancestors. |
|
89 | /// ancestors. | |
84 | /// |
|
90 | /// | |
85 | /// A method with two very different "modes" like this usually a code smell |
|
91 | /// A method with two very different "modes" like this usually a code smell | |
86 | /// to make two methods instead, but in this case an `Option` is what rhg |
|
92 | /// to make two methods instead, but in this case an `Option` is what rhg | |
87 | /// sub-commands get from Clap for the `-R` / `--repository` CLI argument. |
|
93 | /// sub-commands get from Clap for the `-R` / `--repository` CLI argument. | |
88 | /// Having two methods would just move that `if` to almost all callers. |
|
94 | /// Having two methods would just move that `if` to almost all callers. | |
89 | pub fn find( |
|
95 | pub fn find( | |
90 | config: &Config, |
|
96 | config: &Config, | |
91 | explicit_path: Option<PathBuf>, |
|
97 | explicit_path: Option<PathBuf>, | |
92 | ) -> Result<Self, RepoError> { |
|
98 | ) -> Result<Self, RepoError> { | |
93 | if let Some(root) = explicit_path { |
|
99 | if let Some(root) = explicit_path { | |
94 | if is_dir(root.join(".hg"))? { |
|
100 | if is_dir(root.join(".hg"))? { | |
95 | Self::new_at_path(root, config) |
|
101 | Self::new_at_path(root, config) | |
96 | } else if is_file(&root)? { |
|
102 | } else if is_file(&root)? { | |
97 | Err(HgError::unsupported("bundle repository").into()) |
|
103 | Err(HgError::unsupported("bundle repository").into()) | |
98 | } else { |
|
104 | } else { | |
99 | Err(RepoError::NotFound { at: root }) |
|
105 | Err(RepoError::NotFound { at: root }) | |
100 | } |
|
106 | } | |
101 | } else { |
|
107 | } else { | |
102 | let root = Self::find_repo_root()?; |
|
108 | let root = Self::find_repo_root()?; | |
103 | Self::new_at_path(root, config) |
|
109 | Self::new_at_path(root, config) | |
104 | } |
|
110 | } | |
105 | } |
|
111 | } | |
106 |
|
112 | |||
107 | /// To be called after checking that `.hg` is a sub-directory |
|
113 | /// To be called after checking that `.hg` is a sub-directory | |
108 | fn new_at_path( |
|
114 | fn new_at_path( | |
109 | working_directory: PathBuf, |
|
115 | working_directory: PathBuf, | |
110 | config: &Config, |
|
116 | config: &Config, | |
111 | ) -> Result<Self, RepoError> { |
|
117 | ) -> Result<Self, RepoError> { | |
112 | let dot_hg = working_directory.join(".hg"); |
|
118 | let dot_hg = working_directory.join(".hg"); | |
113 |
|
119 | |||
114 | let mut repo_config_files = |
|
120 | let mut repo_config_files = | |
115 | vec![dot_hg.join("hgrc"), dot_hg.join("hgrc-not-shared")]; |
|
121 | vec![dot_hg.join("hgrc"), dot_hg.join("hgrc-not-shared")]; | |
116 |
|
122 | |||
117 | let hg_vfs = Vfs { base: &dot_hg }; |
|
123 | let hg_vfs = Vfs { base: &dot_hg }; | |
118 | let mut reqs = requirements::load_if_exists(hg_vfs)?; |
|
124 | let mut reqs = requirements::load_if_exists(hg_vfs)?; | |
119 | let relative = |
|
125 | let relative = | |
120 | reqs.contains(requirements::RELATIVE_SHARED_REQUIREMENT); |
|
126 | reqs.contains(requirements::RELATIVE_SHARED_REQUIREMENT); | |
121 | let shared = |
|
127 | let shared = | |
122 | reqs.contains(requirements::SHARED_REQUIREMENT) || relative; |
|
128 | reqs.contains(requirements::SHARED_REQUIREMENT) || relative; | |
123 |
|
129 | |||
124 | // From `mercurial/localrepo.py`: |
|
130 | // From `mercurial/localrepo.py`: | |
125 | // |
|
131 | // | |
126 | // if .hg/requires contains the sharesafe requirement, it means |
|
132 | // if .hg/requires contains the sharesafe requirement, it means | |
127 | // there exists a `.hg/store/requires` too and we should read it |
|
133 | // there exists a `.hg/store/requires` too and we should read it | |
128 | // NOTE: presence of SHARESAFE_REQUIREMENT imply that store requirement |
|
134 | // NOTE: presence of SHARESAFE_REQUIREMENT imply that store requirement | |
129 | // is present. We never write SHARESAFE_REQUIREMENT for a repo if store |
|
135 | // is present. We never write SHARESAFE_REQUIREMENT for a repo if store | |
130 | // is not present, refer checkrequirementscompat() for that |
|
136 | // is not present, refer checkrequirementscompat() for that | |
131 | // |
|
137 | // | |
132 | // However, if SHARESAFE_REQUIREMENT is not present, it means that the |
|
138 | // However, if SHARESAFE_REQUIREMENT is not present, it means that the | |
133 | // repository was shared the old way. We check the share source |
|
139 | // repository was shared the old way. We check the share source | |
134 | // .hg/requires for SHARESAFE_REQUIREMENT to detect whether the |
|
140 | // .hg/requires for SHARESAFE_REQUIREMENT to detect whether the | |
135 | // current repository needs to be reshared |
|
141 | // current repository needs to be reshared | |
136 | let share_safe = reqs.contains(requirements::SHARESAFE_REQUIREMENT); |
|
142 | let share_safe = reqs.contains(requirements::SHARESAFE_REQUIREMENT); | |
137 |
|
143 | |||
138 | let store_path; |
|
144 | let store_path; | |
139 | if !shared { |
|
145 | if !shared { | |
140 | store_path = dot_hg.join("store"); |
|
146 | store_path = dot_hg.join("store"); | |
141 | } else { |
|
147 | } else { | |
142 | let bytes = hg_vfs.read("sharedpath")?; |
|
148 | let bytes = hg_vfs.read("sharedpath")?; | |
143 | let mut shared_path = |
|
149 | let mut shared_path = | |
144 | get_path_from_bytes(bytes.trim_end_matches(|b| b == b'\n')) |
|
150 | get_path_from_bytes(bytes.trim_end_matches(|b| b == b'\n')) | |
145 | .to_owned(); |
|
151 | .to_owned(); | |
146 | if relative { |
|
152 | if relative { | |
147 | shared_path = dot_hg.join(shared_path) |
|
153 | shared_path = dot_hg.join(shared_path) | |
148 | } |
|
154 | } | |
149 | if !is_dir(&shared_path)? { |
|
155 | if !is_dir(&shared_path)? { | |
150 | return Err(HgError::corrupted(format!( |
|
156 | return Err(HgError::corrupted(format!( | |
151 | ".hg/sharedpath points to nonexistent directory {}", |
|
157 | ".hg/sharedpath points to nonexistent directory {}", | |
152 | shared_path.display() |
|
158 | shared_path.display() | |
153 | )) |
|
159 | )) | |
154 | .into()); |
|
160 | .into()); | |
155 | } |
|
161 | } | |
156 |
|
162 | |||
157 | store_path = shared_path.join("store"); |
|
163 | store_path = shared_path.join("store"); | |
158 |
|
164 | |||
159 | let source_is_share_safe = |
|
165 | let source_is_share_safe = | |
160 | requirements::load(Vfs { base: &shared_path })? |
|
166 | requirements::load(Vfs { base: &shared_path })? | |
161 | .contains(requirements::SHARESAFE_REQUIREMENT); |
|
167 | .contains(requirements::SHARESAFE_REQUIREMENT); | |
162 |
|
168 | |||
163 | if share_safe != source_is_share_safe { |
|
169 | if share_safe != source_is_share_safe { | |
164 | return Err(HgError::unsupported("share-safe mismatch").into()); |
|
170 | return Err(HgError::unsupported("share-safe mismatch").into()); | |
165 | } |
|
171 | } | |
166 |
|
172 | |||
167 | if share_safe { |
|
173 | if share_safe { | |
168 | repo_config_files.insert(0, shared_path.join("hgrc")) |
|
174 | repo_config_files.insert(0, shared_path.join("hgrc")) | |
169 | } |
|
175 | } | |
170 | } |
|
176 | } | |
171 | if share_safe { |
|
177 | if share_safe { | |
172 | reqs.extend(requirements::load(Vfs { base: &store_path })?); |
|
178 | reqs.extend(requirements::load(Vfs { base: &store_path })?); | |
173 | } |
|
179 | } | |
174 |
|
180 | |||
175 | let repo_config = if std::env::var_os("HGRCSKIPREPO").is_none() { |
|
181 | let repo_config = if std::env::var_os("HGRCSKIPREPO").is_none() { | |
176 | config.combine_with_repo(&repo_config_files)? |
|
182 | config.combine_with_repo(&repo_config_files)? | |
177 | } else { |
|
183 | } else { | |
178 | config.clone() |
|
184 | config.clone() | |
179 | }; |
|
185 | }; | |
180 |
|
186 | |||
181 | let repo = Self { |
|
187 | let repo = Self { | |
182 | requirements: reqs, |
|
188 | requirements: reqs, | |
183 | working_directory, |
|
189 | working_directory, | |
184 | store: store_path, |
|
190 | store: store_path, | |
185 | dot_hg, |
|
191 | dot_hg, | |
186 | config: repo_config, |
|
192 | config: repo_config, | |
187 | dirstate_parents: LazyCell::new(), |
|
193 | dirstate_parents: LazyCell::new(), | |
188 | dirstate_map: LazyCell::new(), |
|
194 | dirstate_map: LazyCell::new(), | |
189 | changelog: LazyCell::new(), |
|
195 | changelog: LazyCell::new(), | |
190 | manifestlog: LazyCell::new(), |
|
196 | manifestlog: LazyCell::new(), | |
191 | }; |
|
197 | }; | |
192 |
|
198 | |||
193 | requirements::check(&repo)?; |
|
199 | requirements::check(&repo)?; | |
194 |
|
200 | |||
195 | Ok(repo) |
|
201 | Ok(repo) | |
196 | } |
|
202 | } | |
197 |
|
203 | |||
198 | pub fn working_directory_path(&self) -> &Path { |
|
204 | pub fn working_directory_path(&self) -> &Path { | |
199 | &self.working_directory |
|
205 | &self.working_directory | |
200 | } |
|
206 | } | |
201 |
|
207 | |||
202 | pub fn requirements(&self) -> &HashSet<String> { |
|
208 | pub fn requirements(&self) -> &HashSet<String> { | |
203 | &self.requirements |
|
209 | &self.requirements | |
204 | } |
|
210 | } | |
205 |
|
211 | |||
206 | pub fn config(&self) -> &Config { |
|
212 | pub fn config(&self) -> &Config { | |
207 | &self.config |
|
213 | &self.config | |
208 | } |
|
214 | } | |
209 |
|
215 | |||
210 | /// For accessing repository files (in `.hg`), except for the store |
|
216 | /// For accessing repository files (in `.hg`), except for the store | |
211 | /// (`.hg/store`). |
|
217 | /// (`.hg/store`). | |
212 | pub fn hg_vfs(&self) -> Vfs<'_> { |
|
218 | pub fn hg_vfs(&self) -> Vfs<'_> { | |
213 | Vfs { base: &self.dot_hg } |
|
219 | Vfs { base: &self.dot_hg } | |
214 | } |
|
220 | } | |
215 |
|
221 | |||
216 | /// For accessing repository store files (in `.hg/store`) |
|
222 | /// For accessing repository store files (in `.hg/store`) | |
217 | pub fn store_vfs(&self) -> Vfs<'_> { |
|
223 | pub fn store_vfs(&self) -> Vfs<'_> { | |
218 | Vfs { base: &self.store } |
|
224 | Vfs { base: &self.store } | |
219 | } |
|
225 | } | |
220 |
|
226 | |||
221 | /// For accessing the working copy |
|
227 | /// For accessing the working copy | |
222 | pub fn working_directory_vfs(&self) -> Vfs<'_> { |
|
228 | pub fn working_directory_vfs(&self) -> Vfs<'_> { | |
223 | Vfs { |
|
229 | Vfs { | |
224 | base: &self.working_directory, |
|
230 | base: &self.working_directory, | |
225 | } |
|
231 | } | |
226 | } |
|
232 | } | |
227 |
|
233 | |||
228 | pub fn try_with_wlock_no_wait<R>( |
|
234 | pub fn try_with_wlock_no_wait<R>( | |
229 | &self, |
|
235 | &self, | |
230 | f: impl FnOnce() -> R, |
|
236 | f: impl FnOnce() -> R, | |
231 | ) -> Result<R, LockError> { |
|
237 | ) -> Result<R, LockError> { | |
232 | try_with_lock_no_wait(self.hg_vfs(), "wlock", f) |
|
238 | try_with_lock_no_wait(self.hg_vfs(), "wlock", f) | |
233 | } |
|
239 | } | |
234 |
|
240 | |||
235 | /// Whether this repo should use dirstate-v2. |
|
241 | /// Whether this repo should use dirstate-v2. | |
236 | /// The presence of `dirstate-v2` in the requirements does not mean that |
|
242 | /// The presence of `dirstate-v2` in the requirements does not mean that | |
237 | /// the on-disk dirstate is necessarily in version 2. In most cases, |
|
243 | /// the on-disk dirstate is necessarily in version 2. In most cases, | |
238 | /// a dirstate-v2 file will indeed be found, but in rare cases (like the |
|
244 | /// a dirstate-v2 file will indeed be found, but in rare cases (like the | |
239 | /// upgrade mechanism being cut short), the on-disk version will be a |
|
245 | /// upgrade mechanism being cut short), the on-disk version will be a | |
240 | /// v1 file. |
|
246 | /// v1 file. | |
241 | /// Semantically, having a requirement only means that a client cannot |
|
247 | /// Semantically, having a requirement only means that a client cannot | |
242 | /// properly understand or properly update the repo if it lacks the support |
|
248 | /// properly understand or properly update the repo if it lacks the support | |
243 | /// for the required feature, but not that that feature is actually used |
|
249 | /// for the required feature, but not that that feature is actually used | |
244 | /// in all occasions. |
|
250 | /// in all occasions. | |
245 | pub fn use_dirstate_v2(&self) -> bool { |
|
251 | pub fn use_dirstate_v2(&self) -> bool { | |
246 | self.requirements |
|
252 | self.requirements | |
247 | .contains(requirements::DIRSTATE_V2_REQUIREMENT) |
|
253 | .contains(requirements::DIRSTATE_V2_REQUIREMENT) | |
248 | } |
|
254 | } | |
249 |
|
255 | |||
250 | pub fn has_sparse(&self) -> bool { |
|
256 | pub fn has_sparse(&self) -> bool { | |
251 | self.requirements.contains(requirements::SPARSE_REQUIREMENT) |
|
257 | self.requirements.contains(requirements::SPARSE_REQUIREMENT) | |
252 | } |
|
258 | } | |
253 |
|
259 | |||
254 | pub fn has_narrow(&self) -> bool { |
|
260 | pub fn has_narrow(&self) -> bool { | |
255 | self.requirements.contains(requirements::NARROW_REQUIREMENT) |
|
261 | self.requirements.contains(requirements::NARROW_REQUIREMENT) | |
256 | } |
|
262 | } | |
257 |
|
263 | |||
258 | pub fn has_nodemap(&self) -> bool { |
|
264 | pub fn has_nodemap(&self) -> bool { | |
259 | self.requirements |
|
265 | self.requirements | |
260 | .contains(requirements::NODEMAP_REQUIREMENT) |
|
266 | .contains(requirements::NODEMAP_REQUIREMENT) | |
261 | } |
|
267 | } | |
262 |
|
268 | |||
263 | fn dirstate_file_contents(&self) -> Result<Vec<u8>, HgError> { |
|
269 | fn dirstate_file_contents(&self) -> Result<Vec<u8>, HgError> { | |
264 | Ok(self |
|
270 | Ok(self | |
265 | .hg_vfs() |
|
271 | .hg_vfs() | |
266 | .read("dirstate") |
|
272 | .read("dirstate") | |
267 | .io_not_found_as_none()? |
|
273 | .io_not_found_as_none()? | |
268 | .unwrap_or_default()) |
|
274 | .unwrap_or_default()) | |
269 | } |
|
275 | } | |
270 |
|
276 | |||
271 | fn dirstate_identity(&self) -> Result<Option<u64>, HgError> { |
|
277 | fn dirstate_identity(&self) -> Result<Option<u64>, HgError> { | |
272 | use std::os::unix::fs::MetadataExt; |
|
278 | use std::os::unix::fs::MetadataExt; | |
273 | Ok(self |
|
279 | Ok(self | |
274 | .hg_vfs() |
|
280 | .hg_vfs() | |
275 | .symlink_metadata("dirstate") |
|
281 | .symlink_metadata("dirstate") | |
276 | .io_not_found_as_none()? |
|
282 | .io_not_found_as_none()? | |
277 | .map(|meta| meta.ino())) |
|
283 | .map(|meta| meta.ino())) | |
278 | } |
|
284 | } | |
279 |
|
285 | |||
280 | pub fn dirstate_parents(&self) -> Result<DirstateParents, HgError> { |
|
286 | pub fn dirstate_parents(&self) -> Result<DirstateParents, HgError> { | |
281 | Ok(*self |
|
287 | Ok(*self | |
282 | .dirstate_parents |
|
288 | .dirstate_parents | |
283 | .get_or_init(|| self.read_dirstate_parents())?) |
|
289 | .get_or_init(|| self.read_dirstate_parents())?) | |
284 | } |
|
290 | } | |
285 |
|
291 | |||
286 | fn read_dirstate_parents(&self) -> Result<DirstateParents, HgError> { |
|
292 | fn read_dirstate_parents(&self) -> Result<DirstateParents, HgError> { | |
287 | let dirstate = self.dirstate_file_contents()?; |
|
293 | let dirstate = self.dirstate_file_contents()?; | |
288 | let parents = if dirstate.is_empty() { |
|
294 | let parents = if dirstate.is_empty() { | |
289 | DirstateParents::NULL |
|
295 | DirstateParents::NULL | |
290 | } else if self.use_dirstate_v2() { |
|
296 | } else if self.use_dirstate_v2() { | |
291 | let docket_res = |
|
297 | let docket_res = | |
292 | crate::dirstate_tree::on_disk::read_docket(&dirstate); |
|
298 | crate::dirstate_tree::on_disk::read_docket(&dirstate); | |
293 | match docket_res { |
|
299 | match docket_res { | |
294 | Ok(docket) => docket.parents(), |
|
300 | Ok(docket) => docket.parents(), | |
295 | Err(_) => { |
|
301 | Err(_) => { | |
296 | log::info!( |
|
302 | log::info!( | |
297 | "Parsing dirstate docket failed, \ |
|
303 | "Parsing dirstate docket failed, \ | |
298 | falling back to dirstate-v1" |
|
304 | falling back to dirstate-v1" | |
299 | ); |
|
305 | ); | |
300 | *crate::dirstate::parsers::parse_dirstate_parents( |
|
306 | *crate::dirstate::parsers::parse_dirstate_parents( | |
301 | &dirstate, |
|
307 | &dirstate, | |
302 | )? |
|
308 | )? | |
303 | } |
|
309 | } | |
304 | } |
|
310 | } | |
305 | } else { |
|
311 | } else { | |
306 | *crate::dirstate::parsers::parse_dirstate_parents(&dirstate)? |
|
312 | *crate::dirstate::parsers::parse_dirstate_parents(&dirstate)? | |
307 | }; |
|
313 | }; | |
308 | self.dirstate_parents.set(parents); |
|
314 | self.dirstate_parents.set(parents); | |
309 | Ok(parents) |
|
315 | Ok(parents) | |
310 | } |
|
316 | } | |
311 |
|
317 | |||
312 | /// Returns the information read from the dirstate docket necessary to |
|
318 | /// Returns the information read from the dirstate docket necessary to | |
313 | /// check if the data file has been updated/deleted by another process |
|
319 | /// check if the data file has been updated/deleted by another process | |
314 | /// since we last read the dirstate. |
|
320 | /// since we last read the dirstate. | |
315 | /// Namely, the inode, data file uuid and the data size. |
|
321 | /// Namely, the inode, data file uuid and the data size. | |
316 | fn get_dirstate_data_file_integrity( |
|
322 | fn get_dirstate_data_file_integrity( | |
317 | &self, |
|
323 | &self, | |
318 | ) -> Result<DirstateMapIdentity, HgError> { |
|
324 | ) -> Result<DirstateMapIdentity, HgError> { | |
319 | assert!( |
|
325 | assert!( | |
320 | self.use_dirstate_v2(), |
|
326 | self.use_dirstate_v2(), | |
321 | "accessing dirstate data file ID without dirstate-v2" |
|
327 | "accessing dirstate data file ID without dirstate-v2" | |
322 | ); |
|
328 | ); | |
323 | // Get the identity before the contents since we could have a race |
|
329 | // Get the identity before the contents since we could have a race | |
324 | // between the two. Having an identity that is too old is fine, but |
|
330 | // between the two. Having an identity that is too old is fine, but | |
325 | // one that is younger than the content change is bad. |
|
331 | // one that is younger than the content change is bad. | |
326 | let identity = self.dirstate_identity()?; |
|
332 | let identity = self.dirstate_identity()?; | |
327 | let dirstate = self.dirstate_file_contents()?; |
|
333 | let dirstate = self.dirstate_file_contents()?; | |
328 | if dirstate.is_empty() { |
|
334 | if dirstate.is_empty() { | |
329 | self.dirstate_parents.set(DirstateParents::NULL); |
|
335 | self.dirstate_parents.set(DirstateParents::NULL); | |
330 | Ok((identity, None, 0)) |
|
336 | Ok((identity, None, 0)) | |
331 | } else { |
|
337 | } else { | |
332 | let docket_res = |
|
338 | let docket_res = | |
333 | crate::dirstate_tree::on_disk::read_docket(&dirstate); |
|
339 | crate::dirstate_tree::on_disk::read_docket(&dirstate); | |
334 | match docket_res { |
|
340 | match docket_res { | |
335 | Ok(docket) => { |
|
341 | Ok(docket) => { | |
336 | self.dirstate_parents.set(docket.parents()); |
|
342 | self.dirstate_parents.set(docket.parents()); | |
337 | Ok(( |
|
343 | Ok(( | |
338 | identity, |
|
344 | identity, | |
339 | Some(docket.uuid.to_owned()), |
|
345 | Some(docket.uuid.to_owned()), | |
340 | docket.data_size(), |
|
346 | docket.data_size(), | |
341 | )) |
|
347 | )) | |
342 | } |
|
348 | } | |
343 | Err(_) => { |
|
349 | Err(_) => { | |
344 | log::info!( |
|
350 | log::info!( | |
345 | "Parsing dirstate docket failed, \ |
|
351 | "Parsing dirstate docket failed, \ | |
346 | falling back to dirstate-v1" |
|
352 | falling back to dirstate-v1" | |
347 | ); |
|
353 | ); | |
348 | let parents = |
|
354 | let parents = | |
349 | *crate::dirstate::parsers::parse_dirstate_parents( |
|
355 | *crate::dirstate::parsers::parse_dirstate_parents( | |
350 | &dirstate, |
|
356 | &dirstate, | |
351 | )?; |
|
357 | )?; | |
352 | self.dirstate_parents.set(parents); |
|
358 | self.dirstate_parents.set(parents); | |
353 | Ok((identity, None, 0)) |
|
359 | Ok((identity, None, 0)) | |
354 | } |
|
360 | } | |
355 | } |
|
361 | } | |
356 | } |
|
362 | } | |
357 | } |
|
363 | } | |
358 |
|
364 | |||
359 | fn new_dirstate_map(&self) -> Result<OwningDirstateMap, DirstateError> { |
|
365 | fn new_dirstate_map(&self) -> Result<OwningDirstateMap, DirstateError> { | |
360 | if self.use_dirstate_v2() { |
|
366 | if self.use_dirstate_v2() { | |
361 | // The v2 dirstate is split into a docket and a data file. |
|
367 | // The v2 dirstate is split into a docket and a data file. | |
362 | // Since we don't always take the `wlock` to read it |
|
368 | // Since we don't always take the `wlock` to read it | |
363 | // (like in `hg status`), it is susceptible to races. |
|
369 | // (like in `hg status`), it is susceptible to races. | |
364 | // A simple retry method should be enough since full rewrites |
|
370 | // A simple retry method should be enough since full rewrites | |
365 | // only happen when too much garbage data is present and |
|
371 | // only happen when too much garbage data is present and | |
366 | // this race is unlikely. |
|
372 | // this race is unlikely. | |
367 | let mut tries = 0; |
|
373 | let mut tries = 0; | |
368 |
|
374 | |||
369 | while tries < V2_MAX_READ_ATTEMPTS { |
|
375 | while tries < V2_MAX_READ_ATTEMPTS { | |
370 | tries += 1; |
|
376 | tries += 1; | |
371 | match self.read_docket_and_data_file() { |
|
377 | match self.read_docket_and_data_file() { | |
372 | Ok(m) => { |
|
378 | Ok(m) => { | |
373 | return Ok(m); |
|
379 | return Ok(m); | |
374 | } |
|
380 | } | |
375 | Err(e) => match e { |
|
381 | Err(e) => match e { | |
376 | DirstateError::Common(HgError::RaceDetected( |
|
382 | DirstateError::Common(HgError::RaceDetected( | |
377 | context, |
|
383 | context, | |
378 | )) => { |
|
384 | )) => { | |
379 | log::info!( |
|
385 | log::info!( | |
380 | "dirstate read race detected {} (retry {}/{})", |
|
386 | "dirstate read race detected {} (retry {}/{})", | |
381 | context, |
|
387 | context, | |
382 | tries, |
|
388 | tries, | |
383 | V2_MAX_READ_ATTEMPTS, |
|
389 | V2_MAX_READ_ATTEMPTS, | |
384 | ); |
|
390 | ); | |
385 | continue; |
|
391 | continue; | |
386 | } |
|
392 | } | |
387 | _ => { |
|
393 | _ => { | |
388 | log::info!( |
|
394 | log::info!( | |
389 | "Reading dirstate v2 failed, \ |
|
395 | "Reading dirstate v2 failed, \ | |
390 | falling back to v1" |
|
396 | falling back to v1" | |
391 | ); |
|
397 | ); | |
392 | return self.new_dirstate_map_v1(); |
|
398 | return self.new_dirstate_map_v1(); | |
393 | } |
|
399 | } | |
394 | }, |
|
400 | }, | |
395 | } |
|
401 | } | |
396 | } |
|
402 | } | |
397 | let error = HgError::abort( |
|
403 | let error = HgError::abort( | |
398 | format!("dirstate read race happened {tries} times in a row"), |
|
404 | format!("dirstate read race happened {tries} times in a row"), | |
399 | 255, |
|
405 | 255, | |
400 | None, |
|
406 | None, | |
401 | ); |
|
407 | ); | |
402 | Err(DirstateError::Common(error)) |
|
408 | Err(DirstateError::Common(error)) | |
403 | } else { |
|
409 | } else { | |
404 | self.new_dirstate_map_v1() |
|
410 | self.new_dirstate_map_v1() | |
405 | } |
|
411 | } | |
406 | } |
|
412 | } | |
407 |
|
413 | |||
408 | fn new_dirstate_map_v1(&self) -> Result<OwningDirstateMap, DirstateError> { |
|
414 | fn new_dirstate_map_v1(&self) -> Result<OwningDirstateMap, DirstateError> { | |
409 | debug_wait_for_file_or_print(self.config(), "dirstate.pre-read-file"); |
|
415 | debug_wait_for_file_or_print(self.config(), "dirstate.pre-read-file"); | |
410 | let identity = self.dirstate_identity()?; |
|
416 | let identity = self.dirstate_identity()?; | |
411 | let dirstate_file_contents = self.dirstate_file_contents()?; |
|
417 | let dirstate_file_contents = self.dirstate_file_contents()?; | |
412 | if dirstate_file_contents.is_empty() { |
|
418 | if dirstate_file_contents.is_empty() { | |
413 | self.dirstate_parents.set(DirstateParents::NULL); |
|
419 | self.dirstate_parents.set(DirstateParents::NULL); | |
414 | Ok(OwningDirstateMap::new_empty(Vec::new())) |
|
420 | Ok(OwningDirstateMap::new_empty(Vec::new())) | |
415 | } else { |
|
421 | } else { | |
416 | let (map, parents) = |
|
422 | let (map, parents) = | |
417 | OwningDirstateMap::new_v1(dirstate_file_contents, identity)?; |
|
423 | OwningDirstateMap::new_v1(dirstate_file_contents, identity)?; | |
418 | self.dirstate_parents.set(parents); |
|
424 | self.dirstate_parents.set(parents); | |
419 | Ok(map) |
|
425 | Ok(map) | |
420 | } |
|
426 | } | |
421 | } |
|
427 | } | |
422 |
|
428 | |||
423 | fn read_docket_and_data_file( |
|
429 | fn read_docket_and_data_file( | |
424 | &self, |
|
430 | &self, | |
425 | ) -> Result<OwningDirstateMap, DirstateError> { |
|
431 | ) -> Result<OwningDirstateMap, DirstateError> { | |
426 | debug_wait_for_file_or_print(self.config(), "dirstate.pre-read-file"); |
|
432 | debug_wait_for_file_or_print(self.config(), "dirstate.pre-read-file"); | |
427 | let dirstate_file_contents = self.dirstate_file_contents()?; |
|
433 | let dirstate_file_contents = self.dirstate_file_contents()?; | |
428 | let identity = self.dirstate_identity()?; |
|
434 | let identity = self.dirstate_identity()?; | |
429 | if dirstate_file_contents.is_empty() { |
|
435 | if dirstate_file_contents.is_empty() { | |
430 | self.dirstate_parents.set(DirstateParents::NULL); |
|
436 | self.dirstate_parents.set(DirstateParents::NULL); | |
431 | return Ok(OwningDirstateMap::new_empty(Vec::new())); |
|
437 | return Ok(OwningDirstateMap::new_empty(Vec::new())); | |
432 | } |
|
438 | } | |
433 | let docket = crate::dirstate_tree::on_disk::read_docket( |
|
439 | let docket = crate::dirstate_tree::on_disk::read_docket( | |
434 | &dirstate_file_contents, |
|
440 | &dirstate_file_contents, | |
435 | )?; |
|
441 | )?; | |
436 | debug_wait_for_file_or_print( |
|
442 | debug_wait_for_file_or_print( | |
437 | self.config(), |
|
443 | self.config(), | |
438 | "dirstate.post-docket-read-file", |
|
444 | "dirstate.post-docket-read-file", | |
439 | ); |
|
445 | ); | |
440 | self.dirstate_parents.set(docket.parents()); |
|
446 | self.dirstate_parents.set(docket.parents()); | |
441 | let uuid = docket.uuid.to_owned(); |
|
447 | let uuid = docket.uuid.to_owned(); | |
442 | let data_size = docket.data_size(); |
|
448 | let data_size = docket.data_size(); | |
443 |
|
449 | |||
444 | let context = "between reading dirstate docket and data file"; |
|
450 | let context = "between reading dirstate docket and data file"; | |
445 | let race_error = HgError::RaceDetected(context.into()); |
|
451 | let race_error = HgError::RaceDetected(context.into()); | |
446 | let metadata = docket.tree_metadata(); |
|
452 | let metadata = docket.tree_metadata(); | |
447 |
|
453 | |||
448 | let mut map = if crate::vfs::is_on_nfs_mount(docket.data_filename()) { |
|
454 | let mut map = if crate::vfs::is_on_nfs_mount(docket.data_filename()) { | |
449 | // Don't mmap on NFS to prevent `SIGBUS` error on deletion |
|
455 | // Don't mmap on NFS to prevent `SIGBUS` error on deletion | |
450 | let contents = self.hg_vfs().read(docket.data_filename()); |
|
456 | let contents = self.hg_vfs().read(docket.data_filename()); | |
451 | let contents = match contents { |
|
457 | let contents = match contents { | |
452 | Ok(c) => c, |
|
458 | Ok(c) => c, | |
453 | Err(HgError::IoError { error, context }) => { |
|
459 | Err(HgError::IoError { error, context }) => { | |
454 | match error.raw_os_error().expect("real os error") { |
|
460 | match error.raw_os_error().expect("real os error") { | |
455 | // 2 = ENOENT, No such file or directory |
|
461 | // 2 = ENOENT, No such file or directory | |
456 | // 116 = ESTALE, Stale NFS file handle |
|
462 | // 116 = ESTALE, Stale NFS file handle | |
457 | // |
|
463 | // | |
458 | // TODO match on `error.kind()` when |
|
464 | // TODO match on `error.kind()` when | |
459 | // `ErrorKind::StaleNetworkFileHandle` is stable. |
|
465 | // `ErrorKind::StaleNetworkFileHandle` is stable. | |
460 | 2 | 116 => { |
|
466 | 2 | 116 => { | |
461 | // Race where the data file was deleted right after |
|
467 | // Race where the data file was deleted right after | |
462 | // we read the docket, try again |
|
468 | // we read the docket, try again | |
463 | return Err(race_error.into()); |
|
469 | return Err(race_error.into()); | |
464 | } |
|
470 | } | |
465 | _ => { |
|
471 | _ => { | |
466 | return Err( |
|
472 | return Err( | |
467 | HgError::IoError { error, context }.into() |
|
473 | HgError::IoError { error, context }.into() | |
468 | ) |
|
474 | ) | |
469 | } |
|
475 | } | |
470 | } |
|
476 | } | |
471 | } |
|
477 | } | |
472 | Err(e) => return Err(e.into()), |
|
478 | Err(e) => return Err(e.into()), | |
473 | }; |
|
479 | }; | |
474 | OwningDirstateMap::new_v2( |
|
480 | OwningDirstateMap::new_v2( | |
475 | contents, data_size, metadata, uuid, identity, |
|
481 | contents, data_size, metadata, uuid, identity, | |
476 | ) |
|
482 | ) | |
477 | } else { |
|
483 | } else { | |
478 | match self |
|
484 | match self | |
479 | .hg_vfs() |
|
485 | .hg_vfs() | |
480 | .mmap_open(docket.data_filename()) |
|
486 | .mmap_open(docket.data_filename()) | |
481 | .io_not_found_as_none() |
|
487 | .io_not_found_as_none() | |
482 | { |
|
488 | { | |
483 | Ok(Some(data_mmap)) => OwningDirstateMap::new_v2( |
|
489 | Ok(Some(data_mmap)) => OwningDirstateMap::new_v2( | |
484 | data_mmap, data_size, metadata, uuid, identity, |
|
490 | data_mmap, data_size, metadata, uuid, identity, | |
485 | ), |
|
491 | ), | |
486 | Ok(None) => { |
|
492 | Ok(None) => { | |
487 | // Race where the data file was deleted right after we |
|
493 | // Race where the data file was deleted right after we | |
488 | // read the docket, try again |
|
494 | // read the docket, try again | |
489 | return Err(race_error.into()); |
|
495 | return Err(race_error.into()); | |
490 | } |
|
496 | } | |
491 | Err(e) => return Err(e.into()), |
|
497 | Err(e) => return Err(e.into()), | |
492 | } |
|
498 | } | |
493 | }?; |
|
499 | }?; | |
494 |
|
500 | |||
495 | let write_mode_config = self |
|
501 | let write_mode_config = self | |
496 | .config() |
|
502 | .config() | |
497 | .get_str(b"devel", b"dirstate.v2.data_update_mode") |
|
503 | .get_str(b"devel", b"dirstate.v2.data_update_mode") | |
498 | .unwrap_or(Some("auto")) |
|
504 | .unwrap_or(Some("auto")) | |
499 | .unwrap_or("auto"); // don't bother for devel options |
|
505 | .unwrap_or("auto"); // don't bother for devel options | |
500 | let write_mode = match write_mode_config { |
|
506 | let write_mode = match write_mode_config { | |
501 | "auto" => DirstateMapWriteMode::Auto, |
|
507 | "auto" => DirstateMapWriteMode::Auto, | |
502 | "force-new" => DirstateMapWriteMode::ForceNewDataFile, |
|
508 | "force-new" => DirstateMapWriteMode::ForceNewDataFile, | |
503 | "force-append" => DirstateMapWriteMode::ForceAppend, |
|
509 | "force-append" => DirstateMapWriteMode::ForceAppend, | |
504 | _ => DirstateMapWriteMode::Auto, |
|
510 | _ => DirstateMapWriteMode::Auto, | |
505 | }; |
|
511 | }; | |
506 |
|
512 | |||
507 | map.with_dmap_mut(|m| m.set_write_mode(write_mode)); |
|
513 | map.with_dmap_mut(|m| m.set_write_mode(write_mode)); | |
508 |
|
514 | |||
509 | Ok(map) |
|
515 | Ok(map) | |
510 | } |
|
516 | } | |
511 |
|
517 | |||
512 | pub fn dirstate_map( |
|
518 | pub fn dirstate_map( | |
513 | &self, |
|
519 | &self, | |
514 | ) -> Result<Ref<OwningDirstateMap>, DirstateError> { |
|
520 | ) -> Result<Ref<OwningDirstateMap>, DirstateError> { | |
515 | self.dirstate_map.get_or_init(|| self.new_dirstate_map()) |
|
521 | self.dirstate_map.get_or_init(|| self.new_dirstate_map()) | |
516 | } |
|
522 | } | |
517 |
|
523 | |||
518 | pub fn dirstate_map_mut( |
|
524 | pub fn dirstate_map_mut( | |
519 | &self, |
|
525 | &self, | |
520 | ) -> Result<RefMut<OwningDirstateMap>, DirstateError> { |
|
526 | ) -> Result<RefMut<OwningDirstateMap>, DirstateError> { | |
521 | self.dirstate_map |
|
527 | self.dirstate_map | |
522 | .get_mut_or_init(|| self.new_dirstate_map()) |
|
528 | .get_mut_or_init(|| self.new_dirstate_map()) | |
523 | } |
|
529 | } | |
524 |
|
530 | |||
525 | fn new_changelog(&self) -> Result<Changelog, HgError> { |
|
531 | fn new_changelog(&self) -> Result<Changelog, HgError> { | |
526 |
Changelog::open(&self.store_vfs(), self. |
|
532 | Changelog::open(&self.store_vfs(), self.default_revlog_options(true)?) | |
527 | } |
|
533 | } | |
528 |
|
534 | |||
529 | pub fn changelog(&self) -> Result<Ref<Changelog>, HgError> { |
|
535 | pub fn changelog(&self) -> Result<Ref<Changelog>, HgError> { | |
530 | self.changelog.get_or_init(|| self.new_changelog()) |
|
536 | self.changelog.get_or_init(|| self.new_changelog()) | |
531 | } |
|
537 | } | |
532 |
|
538 | |||
533 | pub fn changelog_mut(&self) -> Result<RefMut<Changelog>, HgError> { |
|
539 | pub fn changelog_mut(&self) -> Result<RefMut<Changelog>, HgError> { | |
534 | self.changelog.get_mut_or_init(|| self.new_changelog()) |
|
540 | self.changelog.get_mut_or_init(|| self.new_changelog()) | |
535 | } |
|
541 | } | |
536 |
|
542 | |||
537 | fn new_manifestlog(&self) -> Result<Manifestlog, HgError> { |
|
543 | fn new_manifestlog(&self) -> Result<Manifestlog, HgError> { | |
538 | Manifestlog::open(&self.store_vfs(), self.has_nodemap()) |
|
544 | Manifestlog::open( | |
|
545 | &self.store_vfs(), | |||
|
546 | self.default_revlog_options(false)?, | |||
|
547 | ) | |||
539 | } |
|
548 | } | |
540 |
|
549 | |||
541 | pub fn manifestlog(&self) -> Result<Ref<Manifestlog>, HgError> { |
|
550 | pub fn manifestlog(&self) -> Result<Ref<Manifestlog>, HgError> { | |
542 | self.manifestlog.get_or_init(|| self.new_manifestlog()) |
|
551 | self.manifestlog.get_or_init(|| self.new_manifestlog()) | |
543 | } |
|
552 | } | |
544 |
|
553 | |||
545 | pub fn manifestlog_mut(&self) -> Result<RefMut<Manifestlog>, HgError> { |
|
554 | pub fn manifestlog_mut(&self) -> Result<RefMut<Manifestlog>, HgError> { | |
546 | self.manifestlog.get_mut_or_init(|| self.new_manifestlog()) |
|
555 | self.manifestlog.get_mut_or_init(|| self.new_manifestlog()) | |
547 | } |
|
556 | } | |
548 |
|
557 | |||
549 | /// Returns the manifest of the *changeset* with the given node ID |
|
558 | /// Returns the manifest of the *changeset* with the given node ID | |
550 | pub fn manifest_for_node( |
|
559 | pub fn manifest_for_node( | |
551 | &self, |
|
560 | &self, | |
552 | node: impl Into<NodePrefix>, |
|
561 | node: impl Into<NodePrefix>, | |
553 | ) -> Result<Manifest, RevlogError> { |
|
562 | ) -> Result<Manifest, RevlogError> { | |
554 | self.manifestlog()?.data_for_node( |
|
563 | self.manifestlog()?.data_for_node( | |
555 | self.changelog()? |
|
564 | self.changelog()? | |
556 | .data_for_node(node.into())? |
|
565 | .data_for_node(node.into())? | |
557 | .manifest_node()? |
|
566 | .manifest_node()? | |
558 | .into(), |
|
567 | .into(), | |
559 | ) |
|
568 | ) | |
560 | } |
|
569 | } | |
561 |
|
570 | |||
562 | /// Returns the manifest of the *changeset* with the given revision number |
|
571 | /// Returns the manifest of the *changeset* with the given revision number | |
563 | pub fn manifest_for_rev( |
|
572 | pub fn manifest_for_rev( | |
564 | &self, |
|
573 | &self, | |
565 | revision: UncheckedRevision, |
|
574 | revision: UncheckedRevision, | |
566 | ) -> Result<Manifest, RevlogError> { |
|
575 | ) -> Result<Manifest, RevlogError> { | |
567 | self.manifestlog()?.data_for_node( |
|
576 | self.manifestlog()?.data_for_node( | |
568 | self.changelog()? |
|
577 | self.changelog()? | |
569 | .data_for_rev(revision)? |
|
578 | .data_for_rev(revision)? | |
570 | .manifest_node()? |
|
579 | .manifest_node()? | |
571 | .into(), |
|
580 | .into(), | |
572 | ) |
|
581 | ) | |
573 | } |
|
582 | } | |
574 |
|
583 | |||
575 | pub fn has_subrepos(&self) -> Result<bool, DirstateError> { |
|
584 | pub fn has_subrepos(&self) -> Result<bool, DirstateError> { | |
576 | if let Some(entry) = self.dirstate_map()?.get(HgPath::new(".hgsub"))? { |
|
585 | if let Some(entry) = self.dirstate_map()?.get(HgPath::new(".hgsub"))? { | |
577 | Ok(entry.tracked()) |
|
586 | Ok(entry.tracked()) | |
578 | } else { |
|
587 | } else { | |
579 | Ok(false) |
|
588 | Ok(false) | |
580 | } |
|
589 | } | |
581 | } |
|
590 | } | |
582 |
|
591 | |||
583 | pub fn filelog(&self, path: &HgPath) -> Result<Filelog, HgError> { |
|
592 | pub fn filelog(&self, path: &HgPath) -> Result<Filelog, HgError> { | |
584 | Filelog::open(self, path) |
|
593 | Filelog::open(self, path, self.default_revlog_options(false)?) | |
585 | } |
|
594 | } | |
586 |
|
595 | |||
587 | /// Write to disk any updates that were made through `dirstate_map_mut`. |
|
596 | /// Write to disk any updates that were made through `dirstate_map_mut`. | |
588 | /// |
|
597 | /// | |
589 | /// The "wlock" must be held while calling this. |
|
598 | /// The "wlock" must be held while calling this. | |
590 | /// See for example `try_with_wlock_no_wait`. |
|
599 | /// See for example `try_with_wlock_no_wait`. | |
591 | /// |
|
600 | /// | |
592 | /// TODO: have a `WritableRepo` type only accessible while holding the |
|
601 | /// TODO: have a `WritableRepo` type only accessible while holding the | |
593 | /// lock? |
|
602 | /// lock? | |
594 | pub fn write_dirstate(&self) -> Result<(), DirstateError> { |
|
603 | pub fn write_dirstate(&self) -> Result<(), DirstateError> { | |
595 | let map = self.dirstate_map()?; |
|
604 | let map = self.dirstate_map()?; | |
596 | // TODO: Maintain a `DirstateMap::dirty` flag, and return early here if |
|
605 | // TODO: Maintain a `DirstateMap::dirty` flag, and return early here if | |
597 | // it’s unset |
|
606 | // it’s unset | |
598 | let parents = self.dirstate_parents()?; |
|
607 | let parents = self.dirstate_parents()?; | |
599 | let (packed_dirstate, old_uuid_to_remove) = if self.use_dirstate_v2() { |
|
608 | let (packed_dirstate, old_uuid_to_remove) = if self.use_dirstate_v2() { | |
600 | let (identity, uuid, data_size) = |
|
609 | let (identity, uuid, data_size) = | |
601 | self.get_dirstate_data_file_integrity()?; |
|
610 | self.get_dirstate_data_file_integrity()?; | |
602 | let identity_changed = identity != map.old_identity(); |
|
611 | let identity_changed = identity != map.old_identity(); | |
603 | let uuid_changed = uuid.as_deref() != map.old_uuid(); |
|
612 | let uuid_changed = uuid.as_deref() != map.old_uuid(); | |
604 | let data_length_changed = data_size != map.old_data_size(); |
|
613 | let data_length_changed = data_size != map.old_data_size(); | |
605 |
|
614 | |||
606 | if identity_changed || uuid_changed || data_length_changed { |
|
615 | if identity_changed || uuid_changed || data_length_changed { | |
607 | // If any of identity, uuid or length have changed since |
|
616 | // If any of identity, uuid or length have changed since | |
608 | // last disk read, don't write. |
|
617 | // last disk read, don't write. | |
609 | // This is fine because either we're in a command that doesn't |
|
618 | // This is fine because either we're in a command that doesn't | |
610 | // write anything too important (like `hg status`), or we're in |
|
619 | // write anything too important (like `hg status`), or we're in | |
611 | // `hg add` and we're supposed to have taken the lock before |
|
620 | // `hg add` and we're supposed to have taken the lock before | |
612 | // reading anyway. |
|
621 | // reading anyway. | |
613 | // |
|
622 | // | |
614 | // TODO complain loudly if we've changed anything important |
|
623 | // TODO complain loudly if we've changed anything important | |
615 | // without taking the lock. |
|
624 | // without taking the lock. | |
616 | // (see `hg help config.format.use-dirstate-tracked-hint`) |
|
625 | // (see `hg help config.format.use-dirstate-tracked-hint`) | |
617 | log::debug!( |
|
626 | log::debug!( | |
618 | "dirstate has changed since last read, not updating." |
|
627 | "dirstate has changed since last read, not updating." | |
619 | ); |
|
628 | ); | |
620 | return Ok(()); |
|
629 | return Ok(()); | |
621 | } |
|
630 | } | |
622 |
|
631 | |||
623 | let uuid_opt = map.old_uuid(); |
|
632 | let uuid_opt = map.old_uuid(); | |
624 | let write_mode = if uuid_opt.is_some() { |
|
633 | let write_mode = if uuid_opt.is_some() { | |
625 | DirstateMapWriteMode::Auto |
|
634 | DirstateMapWriteMode::Auto | |
626 | } else { |
|
635 | } else { | |
627 | DirstateMapWriteMode::ForceNewDataFile |
|
636 | DirstateMapWriteMode::ForceNewDataFile | |
628 | }; |
|
637 | }; | |
629 | let (data, tree_metadata, append, old_data_size) = |
|
638 | let (data, tree_metadata, append, old_data_size) = | |
630 | map.pack_v2(write_mode)?; |
|
639 | map.pack_v2(write_mode)?; | |
631 |
|
640 | |||
632 | // Reuse the uuid, or generate a new one, keeping the old for |
|
641 | // Reuse the uuid, or generate a new one, keeping the old for | |
633 | // deletion. |
|
642 | // deletion. | |
634 | let (uuid, old_uuid) = match uuid_opt { |
|
643 | let (uuid, old_uuid) = match uuid_opt { | |
635 | Some(uuid) => { |
|
644 | Some(uuid) => { | |
636 | let as_str = std::str::from_utf8(uuid) |
|
645 | let as_str = std::str::from_utf8(uuid) | |
637 | .map_err(|_| { |
|
646 | .map_err(|_| { | |
638 | HgError::corrupted( |
|
647 | HgError::corrupted( | |
639 | "non-UTF-8 dirstate data file ID", |
|
648 | "non-UTF-8 dirstate data file ID", | |
640 | ) |
|
649 | ) | |
641 | })? |
|
650 | })? | |
642 | .to_owned(); |
|
651 | .to_owned(); | |
643 | if append { |
|
652 | if append { | |
644 | (as_str, None) |
|
653 | (as_str, None) | |
645 | } else { |
|
654 | } else { | |
646 | (DirstateDocket::new_uid(), Some(as_str)) |
|
655 | (DirstateDocket::new_uid(), Some(as_str)) | |
647 | } |
|
656 | } | |
648 | } |
|
657 | } | |
649 | None => (DirstateDocket::new_uid(), None), |
|
658 | None => (DirstateDocket::new_uid(), None), | |
650 | }; |
|
659 | }; | |
651 |
|
660 | |||
652 | let data_filename = format!("dirstate.{}", uuid); |
|
661 | let data_filename = format!("dirstate.{}", uuid); | |
653 | let data_filename = self.hg_vfs().join(data_filename); |
|
662 | let data_filename = self.hg_vfs().join(data_filename); | |
654 | let mut options = std::fs::OpenOptions::new(); |
|
663 | let mut options = std::fs::OpenOptions::new(); | |
655 | options.write(true); |
|
664 | options.write(true); | |
656 |
|
665 | |||
657 | // Why are we not using the O_APPEND flag when appending? |
|
666 | // Why are we not using the O_APPEND flag when appending? | |
658 | // |
|
667 | // | |
659 | // - O_APPEND makes it trickier to deal with garbage at the end of |
|
668 | // - O_APPEND makes it trickier to deal with garbage at the end of | |
660 | // the file, left by a previous uncommitted transaction. By |
|
669 | // the file, left by a previous uncommitted transaction. By | |
661 | // starting the write at [old_data_size] we make sure we erase |
|
670 | // starting the write at [old_data_size] we make sure we erase | |
662 | // all such garbage. |
|
671 | // all such garbage. | |
663 | // |
|
672 | // | |
664 | // - O_APPEND requires to special-case 0-byte writes, whereas we |
|
673 | // - O_APPEND requires to special-case 0-byte writes, whereas we | |
665 | // don't need that. |
|
674 | // don't need that. | |
666 | // |
|
675 | // | |
667 | // - Some OSes have bugs in implementation O_APPEND: |
|
676 | // - Some OSes have bugs in implementation O_APPEND: | |
668 | // revlog.py talks about a Solaris bug, but we also saw some ZFS |
|
677 | // revlog.py talks about a Solaris bug, but we also saw some ZFS | |
669 | // bug: https://github.com/openzfs/zfs/pull/3124, |
|
678 | // bug: https://github.com/openzfs/zfs/pull/3124, | |
670 | // https://github.com/openzfs/zfs/issues/13370 |
|
679 | // https://github.com/openzfs/zfs/issues/13370 | |
671 | // |
|
680 | // | |
672 | if !append { |
|
681 | if !append { | |
673 | log::trace!("creating a new dirstate data file"); |
|
682 | log::trace!("creating a new dirstate data file"); | |
674 | options.create_new(true); |
|
683 | options.create_new(true); | |
675 | } else { |
|
684 | } else { | |
676 | log::trace!("appending to the dirstate data file"); |
|
685 | log::trace!("appending to the dirstate data file"); | |
677 | } |
|
686 | } | |
678 |
|
687 | |||
679 | let data_size = (|| { |
|
688 | let data_size = (|| { | |
680 | // TODO: loop and try another random ID if !append and this |
|
689 | // TODO: loop and try another random ID if !append and this | |
681 | // returns `ErrorKind::AlreadyExists`? Collision chance of two |
|
690 | // returns `ErrorKind::AlreadyExists`? Collision chance of two | |
682 | // random IDs is one in 2**32 |
|
691 | // random IDs is one in 2**32 | |
683 | let mut file = options.open(&data_filename)?; |
|
692 | let mut file = options.open(&data_filename)?; | |
684 | if append { |
|
693 | if append { | |
685 | file.seek(SeekFrom::Start(old_data_size as u64))?; |
|
694 | file.seek(SeekFrom::Start(old_data_size as u64))?; | |
686 | } |
|
695 | } | |
687 | file.write_all(&data)?; |
|
696 | file.write_all(&data)?; | |
688 | file.flush()?; |
|
697 | file.flush()?; | |
689 | file.stream_position() |
|
698 | file.stream_position() | |
690 | })() |
|
699 | })() | |
691 | .when_writing_file(&data_filename)?; |
|
700 | .when_writing_file(&data_filename)?; | |
692 |
|
701 | |||
693 | let packed_dirstate = DirstateDocket::serialize( |
|
702 | let packed_dirstate = DirstateDocket::serialize( | |
694 | parents, |
|
703 | parents, | |
695 | tree_metadata, |
|
704 | tree_metadata, | |
696 | data_size, |
|
705 | data_size, | |
697 | uuid.as_bytes(), |
|
706 | uuid.as_bytes(), | |
698 | ) |
|
707 | ) | |
699 | .map_err(|_: std::num::TryFromIntError| { |
|
708 | .map_err(|_: std::num::TryFromIntError| { | |
700 | HgError::corrupted("overflow in dirstate docket serialization") |
|
709 | HgError::corrupted("overflow in dirstate docket serialization") | |
701 | })?; |
|
710 | })?; | |
702 |
|
711 | |||
703 | (packed_dirstate, old_uuid) |
|
712 | (packed_dirstate, old_uuid) | |
704 | } else { |
|
713 | } else { | |
705 | let identity = self.dirstate_identity()?; |
|
714 | let identity = self.dirstate_identity()?; | |
706 | if identity != map.old_identity() { |
|
715 | if identity != map.old_identity() { | |
707 | // If identity changed since last disk read, don't write. |
|
716 | // If identity changed since last disk read, don't write. | |
708 | // This is fine because either we're in a command that doesn't |
|
717 | // This is fine because either we're in a command that doesn't | |
709 | // write anything too important (like `hg status`), or we're in |
|
718 | // write anything too important (like `hg status`), or we're in | |
710 | // `hg add` and we're supposed to have taken the lock before |
|
719 | // `hg add` and we're supposed to have taken the lock before | |
711 | // reading anyway. |
|
720 | // reading anyway. | |
712 | // |
|
721 | // | |
713 | // TODO complain loudly if we've changed anything important |
|
722 | // TODO complain loudly if we've changed anything important | |
714 | // without taking the lock. |
|
723 | // without taking the lock. | |
715 | // (see `hg help config.format.use-dirstate-tracked-hint`) |
|
724 | // (see `hg help config.format.use-dirstate-tracked-hint`) | |
716 | log::debug!( |
|
725 | log::debug!( | |
717 | "dirstate has changed since last read, not updating." |
|
726 | "dirstate has changed since last read, not updating." | |
718 | ); |
|
727 | ); | |
719 | return Ok(()); |
|
728 | return Ok(()); | |
720 | } |
|
729 | } | |
721 | (map.pack_v1(parents)?, None) |
|
730 | (map.pack_v1(parents)?, None) | |
722 | }; |
|
731 | }; | |
723 |
|
732 | |||
724 | let vfs = self.hg_vfs(); |
|
733 | let vfs = self.hg_vfs(); | |
725 | vfs.atomic_write("dirstate", &packed_dirstate)?; |
|
734 | vfs.atomic_write("dirstate", &packed_dirstate)?; | |
726 | if let Some(uuid) = old_uuid_to_remove { |
|
735 | if let Some(uuid) = old_uuid_to_remove { | |
727 | // Remove the old data file after the new docket pointing to the |
|
736 | // Remove the old data file after the new docket pointing to the | |
728 | // new data file was written. |
|
737 | // new data file was written. | |
729 | vfs.remove_file(format!("dirstate.{}", uuid))?; |
|
738 | vfs.remove_file(format!("dirstate.{}", uuid))?; | |
730 | } |
|
739 | } | |
731 | Ok(()) |
|
740 | Ok(()) | |
732 | } |
|
741 | } | |
|
742 | ||||
|
743 | pub fn default_revlog_options( | |||
|
744 | &self, | |||
|
745 | changelog: bool, | |||
|
746 | ) -> Result<RevlogOpenOptions, HgError> { | |||
|
747 | let requirements = self.requirements(); | |||
|
748 | let version = if changelog | |||
|
749 | && requirements.contains(CHANGELOGV2_REQUIREMENT) | |||
|
750 | { | |||
|
751 | let compute_rank = self | |||
|
752 | .config() | |||
|
753 | .get_bool(b"experimental", b"changelog-v2.compute-rank")?; | |||
|
754 | RevlogVersionOptions::ChangelogV2 { compute_rank } | |||
|
755 | } else if requirements.contains(REVLOGV2_REQUIREMENT) { | |||
|
756 | RevlogVersionOptions::V2 | |||
|
757 | } else if requirements.contains(REVLOGV1_REQUIREMENT) { | |||
|
758 | RevlogVersionOptions::V1 { | |||
|
759 | generaldelta: requirements.contains(GENERALDELTA_REQUIREMENT), | |||
|
760 | } | |||
|
761 | } else { | |||
|
762 | RevlogVersionOptions::V0 | |||
|
763 | }; | |||
|
764 | Ok(RevlogOpenOptions { | |||
|
765 | version, | |||
|
766 | // We don't need to dance around the slow path like in the Python | |||
|
767 | // implementation since we know we have access to the fast code. | |||
|
768 | use_nodemap: requirements.contains(NODEMAP_REQUIREMENT), | |||
|
769 | }) | |||
|
770 | } | |||
733 | } |
|
771 | } | |
734 |
|
772 | |||
735 | /// Lazily-initialized component of `Repo` with interior mutability |
|
773 | /// Lazily-initialized component of `Repo` with interior mutability | |
736 | /// |
|
774 | /// | |
737 | /// This differs from `OnceCell` in that the value can still be "deinitialized" |
|
775 | /// This differs from `OnceCell` in that the value can still be "deinitialized" | |
738 | /// later by setting its inner `Option` to `None`. It also takes the |
|
776 | /// later by setting its inner `Option` to `None`. It also takes the | |
739 | /// initialization function as an argument when the value is requested, not |
|
777 | /// initialization function as an argument when the value is requested, not | |
740 | /// when the instance is created. |
|
778 | /// when the instance is created. | |
741 | struct LazyCell<T> { |
|
779 | struct LazyCell<T> { | |
742 | value: RefCell<Option<T>>, |
|
780 | value: RefCell<Option<T>>, | |
743 | } |
|
781 | } | |
744 |
|
782 | |||
745 | impl<T> LazyCell<T> { |
|
783 | impl<T> LazyCell<T> { | |
746 | fn new() -> Self { |
|
784 | fn new() -> Self { | |
747 | Self { |
|
785 | Self { | |
748 | value: RefCell::new(None), |
|
786 | value: RefCell::new(None), | |
749 | } |
|
787 | } | |
750 | } |
|
788 | } | |
751 |
|
789 | |||
752 | fn set(&self, value: T) { |
|
790 | fn set(&self, value: T) { | |
753 | *self.value.borrow_mut() = Some(value) |
|
791 | *self.value.borrow_mut() = Some(value) | |
754 | } |
|
792 | } | |
755 |
|
793 | |||
756 | fn get_or_init<E>( |
|
794 | fn get_or_init<E>( | |
757 | &self, |
|
795 | &self, | |
758 | init: impl Fn() -> Result<T, E>, |
|
796 | init: impl Fn() -> Result<T, E>, | |
759 | ) -> Result<Ref<T>, E> { |
|
797 | ) -> Result<Ref<T>, E> { | |
760 | let mut borrowed = self.value.borrow(); |
|
798 | let mut borrowed = self.value.borrow(); | |
761 | if borrowed.is_none() { |
|
799 | if borrowed.is_none() { | |
762 | drop(borrowed); |
|
800 | drop(borrowed); | |
763 | // Only use `borrow_mut` if it is really needed to avoid panic in |
|
801 | // Only use `borrow_mut` if it is really needed to avoid panic in | |
764 | // case there is another outstanding borrow but mutation is not |
|
802 | // case there is another outstanding borrow but mutation is not | |
765 | // needed. |
|
803 | // needed. | |
766 | *self.value.borrow_mut() = Some(init()?); |
|
804 | *self.value.borrow_mut() = Some(init()?); | |
767 | borrowed = self.value.borrow() |
|
805 | borrowed = self.value.borrow() | |
768 | } |
|
806 | } | |
769 | Ok(Ref::map(borrowed, |option| option.as_ref().unwrap())) |
|
807 | Ok(Ref::map(borrowed, |option| option.as_ref().unwrap())) | |
770 | } |
|
808 | } | |
771 |
|
809 | |||
772 | fn get_mut_or_init<E>( |
|
810 | fn get_mut_or_init<E>( | |
773 | &self, |
|
811 | &self, | |
774 | init: impl Fn() -> Result<T, E>, |
|
812 | init: impl Fn() -> Result<T, E>, | |
775 | ) -> Result<RefMut<T>, E> { |
|
813 | ) -> Result<RefMut<T>, E> { | |
776 | let mut borrowed = self.value.borrow_mut(); |
|
814 | let mut borrowed = self.value.borrow_mut(); | |
777 | if borrowed.is_none() { |
|
815 | if borrowed.is_none() { | |
778 | *borrowed = Some(init()?); |
|
816 | *borrowed = Some(init()?); | |
779 | } |
|
817 | } | |
780 | Ok(RefMut::map(borrowed, |option| option.as_mut().unwrap())) |
|
818 | Ok(RefMut::map(borrowed, |option| option.as_mut().unwrap())) | |
781 | } |
|
819 | } | |
782 | } |
|
820 | } |
@@ -1,173 +1,183 b'' | |||||
1 | use crate::errors::{HgError, HgResultExt}; |
|
1 | use crate::errors::{HgError, HgResultExt}; | |
2 | use crate::repo::Repo; |
|
2 | use crate::repo::Repo; | |
3 | use crate::utils::join_display; |
|
3 | use crate::utils::join_display; | |
4 | use crate::vfs::Vfs; |
|
4 | use crate::vfs::Vfs; | |
5 | use std::collections::HashSet; |
|
5 | use std::collections::HashSet; | |
6 |
|
6 | |||
7 | fn parse(bytes: &[u8]) -> Result<HashSet<String>, HgError> { |
|
7 | fn parse(bytes: &[u8]) -> Result<HashSet<String>, HgError> { | |
8 | // The Python code reading this file uses `str.splitlines` |
|
8 | // The Python code reading this file uses `str.splitlines` | |
9 | // which looks for a number of line separators (even including a couple of |
|
9 | // which looks for a number of line separators (even including a couple of | |
10 | // non-ASCII ones), but Python code writing it always uses `\n`. |
|
10 | // non-ASCII ones), but Python code writing it always uses `\n`. | |
11 | let lines = bytes.split(|&byte| byte == b'\n'); |
|
11 | let lines = bytes.split(|&byte| byte == b'\n'); | |
12 |
|
12 | |||
13 | lines |
|
13 | lines | |
14 | .filter(|line| !line.is_empty()) |
|
14 | .filter(|line| !line.is_empty()) | |
15 | .map(|line| { |
|
15 | .map(|line| { | |
16 | // Python uses Unicode `str.isalnum` but feature names are all |
|
16 | // Python uses Unicode `str.isalnum` but feature names are all | |
17 | // ASCII |
|
17 | // ASCII | |
18 | if line[0].is_ascii_alphanumeric() && line.is_ascii() { |
|
18 | if line[0].is_ascii_alphanumeric() && line.is_ascii() { | |
19 | Ok(String::from_utf8(line.into()).unwrap()) |
|
19 | Ok(String::from_utf8(line.into()).unwrap()) | |
20 | } else { |
|
20 | } else { | |
21 | Err(HgError::corrupted("parse error in 'requires' file")) |
|
21 | Err(HgError::corrupted("parse error in 'requires' file")) | |
22 | } |
|
22 | } | |
23 | }) |
|
23 | }) | |
24 | .collect() |
|
24 | .collect() | |
25 | } |
|
25 | } | |
26 |
|
26 | |||
27 | pub(crate) fn load(hg_vfs: Vfs) -> Result<HashSet<String>, HgError> { |
|
27 | pub(crate) fn load(hg_vfs: Vfs) -> Result<HashSet<String>, HgError> { | |
28 | parse(&hg_vfs.read("requires")?) |
|
28 | parse(&hg_vfs.read("requires")?) | |
29 | } |
|
29 | } | |
30 |
|
30 | |||
31 | pub(crate) fn load_if_exists(hg_vfs: Vfs) -> Result<HashSet<String>, HgError> { |
|
31 | pub(crate) fn load_if_exists(hg_vfs: Vfs) -> Result<HashSet<String>, HgError> { | |
32 | if let Some(bytes) = hg_vfs.read("requires").io_not_found_as_none()? { |
|
32 | if let Some(bytes) = hg_vfs.read("requires").io_not_found_as_none()? { | |
33 | parse(&bytes) |
|
33 | parse(&bytes) | |
34 | } else { |
|
34 | } else { | |
35 | // Treat a missing file the same as an empty file. |
|
35 | // Treat a missing file the same as an empty file. | |
36 | // From `mercurial/localrepo.py`: |
|
36 | // From `mercurial/localrepo.py`: | |
37 | // > requires file contains a newline-delimited list of |
|
37 | // > requires file contains a newline-delimited list of | |
38 | // > features/capabilities the opener (us) must have in order to use |
|
38 | // > features/capabilities the opener (us) must have in order to use | |
39 | // > the repository. This file was introduced in Mercurial 0.9.2, |
|
39 | // > the repository. This file was introduced in Mercurial 0.9.2, | |
40 | // > which means very old repositories may not have one. We assume |
|
40 | // > which means very old repositories may not have one. We assume | |
41 | // > a missing file translates to no requirements. |
|
41 | // > a missing file translates to no requirements. | |
42 | Ok(HashSet::new()) |
|
42 | Ok(HashSet::new()) | |
43 | } |
|
43 | } | |
44 | } |
|
44 | } | |
45 |
|
45 | |||
46 | pub(crate) fn check(repo: &Repo) -> Result<(), HgError> { |
|
46 | pub(crate) fn check(repo: &Repo) -> Result<(), HgError> { | |
47 | let unknown: Vec<_> = repo |
|
47 | let unknown: Vec<_> = repo | |
48 | .requirements() |
|
48 | .requirements() | |
49 | .iter() |
|
49 | .iter() | |
50 | .map(String::as_str) |
|
50 | .map(String::as_str) | |
51 | // .filter(|feature| !ALL_SUPPORTED.contains(feature.as_str())) |
|
51 | // .filter(|feature| !ALL_SUPPORTED.contains(feature.as_str())) | |
52 | .filter(|feature| { |
|
52 | .filter(|feature| { | |
53 | !REQUIRED.contains(feature) && !SUPPORTED.contains(feature) |
|
53 | !REQUIRED.contains(feature) && !SUPPORTED.contains(feature) | |
54 | }) |
|
54 | }) | |
55 | .collect(); |
|
55 | .collect(); | |
56 | if !unknown.is_empty() { |
|
56 | if !unknown.is_empty() { | |
57 | return Err(HgError::unsupported(format!( |
|
57 | return Err(HgError::unsupported(format!( | |
58 | "repository requires feature unknown to this Mercurial: {}", |
|
58 | "repository requires feature unknown to this Mercurial: {}", | |
59 | join_display(&unknown, ", ") |
|
59 | join_display(&unknown, ", ") | |
60 | ))); |
|
60 | ))); | |
61 | } |
|
61 | } | |
62 | let missing: Vec<_> = REQUIRED |
|
62 | let missing: Vec<_> = REQUIRED | |
63 | .iter() |
|
63 | .iter() | |
64 | .filter(|&&feature| !repo.requirements().contains(feature)) |
|
64 | .filter(|&&feature| !repo.requirements().contains(feature)) | |
65 | .collect(); |
|
65 | .collect(); | |
66 | if !missing.is_empty() { |
|
66 | if !missing.is_empty() { | |
67 | return Err(HgError::unsupported(format!( |
|
67 | return Err(HgError::unsupported(format!( | |
68 | "repository is missing feature required by this Mercurial: {}", |
|
68 | "repository is missing feature required by this Mercurial: {}", | |
69 | join_display(&missing, ", ") |
|
69 | join_display(&missing, ", ") | |
70 | ))); |
|
70 | ))); | |
71 | } |
|
71 | } | |
72 | Ok(()) |
|
72 | Ok(()) | |
73 | } |
|
73 | } | |
74 |
|
74 | |||
75 | /// rhg does not support repositories that are *missing* any of these features |
|
75 | /// rhg does not support repositories that are *missing* any of these features | |
76 | const REQUIRED: &[&str] = &["revlogv1", "store", "fncache", "dotencode"]; |
|
76 | const REQUIRED: &[&str] = &["revlogv1", "store", "fncache", "dotencode"]; | |
77 |
|
77 | |||
78 | /// rhg supports repository with or without these |
|
78 | /// rhg supports repository with or without these | |
79 | const SUPPORTED: &[&str] = &[ |
|
79 | const SUPPORTED: &[&str] = &[ | |
80 | "generaldelta", |
|
80 | GENERALDELTA_REQUIREMENT, | |
81 | SHARED_REQUIREMENT, |
|
81 | SHARED_REQUIREMENT, | |
82 | SHARESAFE_REQUIREMENT, |
|
82 | SHARESAFE_REQUIREMENT, | |
83 | SPARSEREVLOG_REQUIREMENT, |
|
83 | SPARSEREVLOG_REQUIREMENT, | |
84 | RELATIVE_SHARED_REQUIREMENT, |
|
84 | RELATIVE_SHARED_REQUIREMENT, | |
85 | REVLOG_COMPRESSION_ZSTD, |
|
85 | REVLOG_COMPRESSION_ZSTD, | |
86 | DIRSTATE_V2_REQUIREMENT, |
|
86 | DIRSTATE_V2_REQUIREMENT, | |
87 | DIRSTATE_TRACKED_HINT_V1, |
|
87 | DIRSTATE_TRACKED_HINT_V1, | |
88 | // As of this writing everything rhg does is read-only. |
|
88 | // As of this writing everything rhg does is read-only. | |
89 | // When it starts writing to the repository, it’ll need to either keep the |
|
89 | // When it starts writing to the repository, it’ll need to either keep the | |
90 | // persistent nodemap up to date or remove this entry: |
|
90 | // persistent nodemap up to date or remove this entry: | |
91 | NODEMAP_REQUIREMENT, |
|
91 | NODEMAP_REQUIREMENT, | |
92 | // Not all commands support `sparse` and `narrow`. The commands that do |
|
92 | // Not all commands support `sparse` and `narrow`. The commands that do | |
93 | // not should opt out by checking `has_sparse` and `has_narrow`. |
|
93 | // not should opt out by checking `has_sparse` and `has_narrow`. | |
94 | SPARSE_REQUIREMENT, |
|
94 | SPARSE_REQUIREMENT, | |
95 | NARROW_REQUIREMENT, |
|
95 | NARROW_REQUIREMENT, | |
96 | // rhg doesn't care about bookmarks at all yet |
|
96 | // rhg doesn't care about bookmarks at all yet | |
97 | BOOKMARKS_IN_STORE_REQUIREMENT, |
|
97 | BOOKMARKS_IN_STORE_REQUIREMENT, | |
98 | ]; |
|
98 | ]; | |
99 |
|
99 | |||
100 | // Copied from mercurial/requirements.py: |
|
100 | // Copied from mercurial/requirements.py: | |
101 |
|
101 | |||
102 | pub const DIRSTATE_V2_REQUIREMENT: &str = "dirstate-v2"; |
|
102 | pub const DIRSTATE_V2_REQUIREMENT: &str = "dirstate-v2"; | |
|
103 | pub const GENERALDELTA_REQUIREMENT: &str = "generaldelta"; | |||
103 |
|
104 | |||
104 | /// A repository that uses the tracked hint dirstate file |
|
105 | /// A repository that uses the tracked hint dirstate file | |
105 | #[allow(unused)] |
|
106 | #[allow(unused)] | |
106 | pub const DIRSTATE_TRACKED_HINT_V1: &str = "dirstate-tracked-key-v1"; |
|
107 | pub const DIRSTATE_TRACKED_HINT_V1: &str = "dirstate-tracked-key-v1"; | |
107 |
|
108 | |||
108 | /// When narrowing is finalized and no longer subject to format changes, |
|
109 | /// When narrowing is finalized and no longer subject to format changes, | |
109 | /// we should move this to just "narrow" or similar. |
|
110 | /// we should move this to just "narrow" or similar. | |
110 | #[allow(unused)] |
|
111 | #[allow(unused)] | |
111 | pub const NARROW_REQUIREMENT: &str = "narrowhg-experimental"; |
|
112 | pub const NARROW_REQUIREMENT: &str = "narrowhg-experimental"; | |
112 |
|
113 | |||
113 | /// Bookmarks must be stored in the `store` part of the repository and will be |
|
114 | /// Bookmarks must be stored in the `store` part of the repository and will be | |
114 | /// share accross shares |
|
115 | /// share accross shares | |
115 | #[allow(unused)] |
|
116 | #[allow(unused)] | |
116 | pub const BOOKMARKS_IN_STORE_REQUIREMENT: &str = "bookmarksinstore"; |
|
117 | pub const BOOKMARKS_IN_STORE_REQUIREMENT: &str = "bookmarksinstore"; | |
117 |
|
118 | |||
118 | /// Enables sparse working directory usage |
|
119 | /// Enables sparse working directory usage | |
119 | #[allow(unused)] |
|
120 | #[allow(unused)] | |
120 | pub const SPARSE_REQUIREMENT: &str = "exp-sparse"; |
|
121 | pub const SPARSE_REQUIREMENT: &str = "exp-sparse"; | |
121 |
|
122 | |||
122 | /// Enables the internal phase which is used to hide changesets instead |
|
123 | /// Enables the internal phase which is used to hide changesets instead | |
123 | /// of stripping them |
|
124 | /// of stripping them | |
124 | #[allow(unused)] |
|
125 | #[allow(unused)] | |
125 | pub const INTERNAL_PHASE_REQUIREMENT: &str = "internal-phase"; |
|
126 | pub const INTERNAL_PHASE_REQUIREMENT: &str = "internal-phase"; | |
126 |
|
127 | |||
127 | /// Stores manifest in Tree structure |
|
128 | /// Stores manifest in Tree structure | |
128 | #[allow(unused)] |
|
129 | #[allow(unused)] | |
129 | pub const TREEMANIFEST_REQUIREMENT: &str = "treemanifest"; |
|
130 | pub const TREEMANIFEST_REQUIREMENT: &str = "treemanifest"; | |
130 |
|
131 | |||
|
132 | /// Whether to use the "RevlogNG" or V1 of the revlog format | |||
|
133 | #[allow(unused)] | |||
|
134 | pub const REVLOGV1_REQUIREMENT: &str = "revlogv1"; | |||
|
135 | ||||
131 | /// Increment the sub-version when the revlog v2 format changes to lock out old |
|
136 | /// Increment the sub-version when the revlog v2 format changes to lock out old | |
132 | /// clients. |
|
137 | /// clients. | |
133 | #[allow(unused)] |
|
138 | #[allow(unused)] | |
134 | pub const REVLOGV2_REQUIREMENT: &str = "exp-revlogv2.1"; |
|
139 | pub const REVLOGV2_REQUIREMENT: &str = "exp-revlogv2.1"; | |
135 |
|
140 | |||
|
141 | /// Increment the sub-version when the revlog v2 format changes to lock out old | |||
|
142 | /// clients. | |||
|
143 | #[allow(unused)] | |||
|
144 | pub const CHANGELOGV2_REQUIREMENT: &str = "exp-changelog-v2"; | |||
|
145 | ||||
136 | /// A repository with the sparserevlog feature will have delta chains that |
|
146 | /// A repository with the sparserevlog feature will have delta chains that | |
137 | /// can spread over a larger span. Sparse reading cuts these large spans into |
|
147 | /// can spread over a larger span. Sparse reading cuts these large spans into | |
138 | /// pieces, so that each piece isn't too big. |
|
148 | /// pieces, so that each piece isn't too big. | |
139 | /// Without the sparserevlog capability, reading from the repository could use |
|
149 | /// Without the sparserevlog capability, reading from the repository could use | |
140 | /// huge amounts of memory, because the whole span would be read at once, |
|
150 | /// huge amounts of memory, because the whole span would be read at once, | |
141 | /// including all the intermediate revisions that aren't pertinent for the |
|
151 | /// including all the intermediate revisions that aren't pertinent for the | |
142 | /// chain. This is why once a repository has enabled sparse-read, it becomes |
|
152 | /// chain. This is why once a repository has enabled sparse-read, it becomes | |
143 | /// required. |
|
153 | /// required. | |
144 | #[allow(unused)] |
|
154 | #[allow(unused)] | |
145 | pub const SPARSEREVLOG_REQUIREMENT: &str = "sparserevlog"; |
|
155 | pub const SPARSEREVLOG_REQUIREMENT: &str = "sparserevlog"; | |
146 |
|
156 | |||
147 | /// A repository with the the copies-sidedata-changeset requirement will store |
|
157 | /// A repository with the the copies-sidedata-changeset requirement will store | |
148 | /// copies related information in changeset's sidedata. |
|
158 | /// copies related information in changeset's sidedata. | |
149 | #[allow(unused)] |
|
159 | #[allow(unused)] | |
150 | pub const COPIESSDC_REQUIREMENT: &str = "exp-copies-sidedata-changeset"; |
|
160 | pub const COPIESSDC_REQUIREMENT: &str = "exp-copies-sidedata-changeset"; | |
151 |
|
161 | |||
152 | /// The repository use persistent nodemap for the changelog and the manifest. |
|
162 | /// The repository use persistent nodemap for the changelog and the manifest. | |
153 | #[allow(unused)] |
|
163 | #[allow(unused)] | |
154 | pub const NODEMAP_REQUIREMENT: &str = "persistent-nodemap"; |
|
164 | pub const NODEMAP_REQUIREMENT: &str = "persistent-nodemap"; | |
155 |
|
165 | |||
156 | /// Denotes that the current repository is a share |
|
166 | /// Denotes that the current repository is a share | |
157 | #[allow(unused)] |
|
167 | #[allow(unused)] | |
158 | pub const SHARED_REQUIREMENT: &str = "shared"; |
|
168 | pub const SHARED_REQUIREMENT: &str = "shared"; | |
159 |
|
169 | |||
160 | /// Denotes that current repository is a share and the shared source path is |
|
170 | /// Denotes that current repository is a share and the shared source path is | |
161 | /// relative to the current repository root path |
|
171 | /// relative to the current repository root path | |
162 | #[allow(unused)] |
|
172 | #[allow(unused)] | |
163 | pub const RELATIVE_SHARED_REQUIREMENT: &str = "relshared"; |
|
173 | pub const RELATIVE_SHARED_REQUIREMENT: &str = "relshared"; | |
164 |
|
174 | |||
165 | /// A repository with share implemented safely. The repository has different |
|
175 | /// A repository with share implemented safely. The repository has different | |
166 | /// store and working copy requirements i.e. both `.hg/requires` and |
|
176 | /// store and working copy requirements i.e. both `.hg/requires` and | |
167 | /// `.hg/store/requires` are present. |
|
177 | /// `.hg/store/requires` are present. | |
168 | #[allow(unused)] |
|
178 | #[allow(unused)] | |
169 | pub const SHARESAFE_REQUIREMENT: &str = "share-safe"; |
|
179 | pub const SHARESAFE_REQUIREMENT: &str = "share-safe"; | |
170 |
|
180 | |||
171 | /// A repository that use zstd compression inside its revlog |
|
181 | /// A repository that use zstd compression inside its revlog | |
172 | #[allow(unused)] |
|
182 | #[allow(unused)] | |
173 | pub const REVLOG_COMPRESSION_ZSTD: &str = "revlog-compression-zstd"; |
|
183 | pub const REVLOG_COMPRESSION_ZSTD: &str = "revlog-compression-zstd"; |
@@ -1,359 +1,363 b'' | |||||
1 | use crate::errors::HgError; |
|
1 | use crate::errors::HgError; | |
2 | use crate::revlog::Revision; |
|
2 | use crate::revlog::Revision; | |
3 | use crate::revlog::{Node, NodePrefix}; |
|
3 | use crate::revlog::{Node, NodePrefix}; | |
4 | use crate::revlog::{Revlog, RevlogEntry, RevlogError}; |
|
4 | use crate::revlog::{Revlog, RevlogEntry, RevlogError}; | |
5 | use crate::utils::hg_path::HgPath; |
|
5 | use crate::utils::hg_path::HgPath; | |
6 | use crate::vfs::Vfs; |
|
6 | use crate::vfs::Vfs; | |
7 | use crate::{Graph, GraphError, UncheckedRevision}; |
|
7 | use crate::{Graph, GraphError, RevlogOpenOptions, UncheckedRevision}; | |
8 | use itertools::Itertools; |
|
8 | use itertools::Itertools; | |
9 | use std::ascii::escape_default; |
|
9 | use std::ascii::escape_default; | |
10 | use std::borrow::Cow; |
|
10 | use std::borrow::Cow; | |
11 | use std::fmt::{Debug, Formatter}; |
|
11 | use std::fmt::{Debug, Formatter}; | |
12 |
|
12 | |||
13 | /// A specialized `Revlog` to work with changelog data format. |
|
13 | /// A specialized `Revlog` to work with changelog data format. | |
14 | pub struct Changelog { |
|
14 | pub struct Changelog { | |
15 | /// The generic `revlog` format. |
|
15 | /// The generic `revlog` format. | |
16 | pub(crate) revlog: Revlog, |
|
16 | pub(crate) revlog: Revlog, | |
17 | } |
|
17 | } | |
18 |
|
18 | |||
19 | impl Changelog { |
|
19 | impl Changelog { | |
20 | /// Open the `changelog` of a repository given by its root. |
|
20 | /// Open the `changelog` of a repository given by its root. | |
21 | pub fn open(store_vfs: &Vfs, use_nodemap: bool) -> Result<Self, HgError> { |
|
21 | pub fn open( | |
22 | let revlog = |
|
22 | store_vfs: &Vfs, | |
23 | Revlog::open(store_vfs, "00changelog.i", None, use_nodemap)?; |
|
23 | options: RevlogOpenOptions, | |
|
24 | ) -> Result<Self, HgError> { | |||
|
25 | let revlog = Revlog::open(store_vfs, "00changelog.i", None, options)?; | |||
24 | Ok(Self { revlog }) |
|
26 | Ok(Self { revlog }) | |
25 | } |
|
27 | } | |
26 |
|
28 | |||
27 | /// Return the `ChangelogRevisionData` for the given node ID. |
|
29 | /// Return the `ChangelogRevisionData` for the given node ID. | |
28 | pub fn data_for_node( |
|
30 | pub fn data_for_node( | |
29 | &self, |
|
31 | &self, | |
30 | node: NodePrefix, |
|
32 | node: NodePrefix, | |
31 | ) -> Result<ChangelogRevisionData, RevlogError> { |
|
33 | ) -> Result<ChangelogRevisionData, RevlogError> { | |
32 | let rev = self.revlog.rev_from_node(node)?; |
|
34 | let rev = self.revlog.rev_from_node(node)?; | |
33 | self.entry_for_checked_rev(rev)?.data() |
|
35 | self.entry_for_checked_rev(rev)?.data() | |
34 | } |
|
36 | } | |
35 |
|
37 | |||
36 | /// Return the [`ChangelogEntry`] for the given revision number. |
|
38 | /// Return the [`ChangelogEntry`] for the given revision number. | |
37 | pub fn entry_for_rev( |
|
39 | pub fn entry_for_rev( | |
38 | &self, |
|
40 | &self, | |
39 | rev: UncheckedRevision, |
|
41 | rev: UncheckedRevision, | |
40 | ) -> Result<ChangelogEntry, RevlogError> { |
|
42 | ) -> Result<ChangelogEntry, RevlogError> { | |
41 | let revlog_entry = self.revlog.get_entry(rev)?; |
|
43 | let revlog_entry = self.revlog.get_entry(rev)?; | |
42 | Ok(ChangelogEntry { revlog_entry }) |
|
44 | Ok(ChangelogEntry { revlog_entry }) | |
43 | } |
|
45 | } | |
44 |
|
46 | |||
45 | /// Same as [`Self::entry_for_rev`] for checked revisions. |
|
47 | /// Same as [`Self::entry_for_rev`] for checked revisions. | |
46 | fn entry_for_checked_rev( |
|
48 | fn entry_for_checked_rev( | |
47 | &self, |
|
49 | &self, | |
48 | rev: Revision, |
|
50 | rev: Revision, | |
49 | ) -> Result<ChangelogEntry, RevlogError> { |
|
51 | ) -> Result<ChangelogEntry, RevlogError> { | |
50 | let revlog_entry = self.revlog.get_entry_for_checked_rev(rev)?; |
|
52 | let revlog_entry = self.revlog.get_entry_for_checked_rev(rev)?; | |
51 | Ok(ChangelogEntry { revlog_entry }) |
|
53 | Ok(ChangelogEntry { revlog_entry }) | |
52 | } |
|
54 | } | |
53 |
|
55 | |||
54 | /// Return the [`ChangelogRevisionData`] for the given revision number. |
|
56 | /// Return the [`ChangelogRevisionData`] for the given revision number. | |
55 | /// |
|
57 | /// | |
56 | /// This is a useful shortcut in case the caller does not need the |
|
58 | /// This is a useful shortcut in case the caller does not need the | |
57 | /// generic revlog information (parents, hashes etc). Otherwise |
|
59 | /// generic revlog information (parents, hashes etc). Otherwise | |
58 | /// consider taking a [`ChangelogEntry`] with |
|
60 | /// consider taking a [`ChangelogEntry`] with | |
59 | /// [entry_for_rev](`Self::entry_for_rev`) and doing everything from there. |
|
61 | /// [entry_for_rev](`Self::entry_for_rev`) and doing everything from there. | |
60 | pub fn data_for_rev( |
|
62 | pub fn data_for_rev( | |
61 | &self, |
|
63 | &self, | |
62 | rev: UncheckedRevision, |
|
64 | rev: UncheckedRevision, | |
63 | ) -> Result<ChangelogRevisionData, RevlogError> { |
|
65 | ) -> Result<ChangelogRevisionData, RevlogError> { | |
64 | self.entry_for_rev(rev)?.data() |
|
66 | self.entry_for_rev(rev)?.data() | |
65 | } |
|
67 | } | |
66 |
|
68 | |||
67 | pub fn node_from_rev(&self, rev: UncheckedRevision) -> Option<&Node> { |
|
69 | pub fn node_from_rev(&self, rev: UncheckedRevision) -> Option<&Node> { | |
68 | self.revlog.node_from_rev(rev) |
|
70 | self.revlog.node_from_rev(rev) | |
69 | } |
|
71 | } | |
70 |
|
72 | |||
71 | pub fn rev_from_node( |
|
73 | pub fn rev_from_node( | |
72 | &self, |
|
74 | &self, | |
73 | node: NodePrefix, |
|
75 | node: NodePrefix, | |
74 | ) -> Result<Revision, RevlogError> { |
|
76 | ) -> Result<Revision, RevlogError> { | |
75 | self.revlog.rev_from_node(node) |
|
77 | self.revlog.rev_from_node(node) | |
76 | } |
|
78 | } | |
77 | } |
|
79 | } | |
78 |
|
80 | |||
79 | impl Graph for Changelog { |
|
81 | impl Graph for Changelog { | |
80 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { |
|
82 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { | |
81 | self.revlog.parents(rev) |
|
83 | self.revlog.parents(rev) | |
82 | } |
|
84 | } | |
83 | } |
|
85 | } | |
84 |
|
86 | |||
85 | /// A specialized `RevlogEntry` for `changelog` data format |
|
87 | /// A specialized `RevlogEntry` for `changelog` data format | |
86 | /// |
|
88 | /// | |
87 | /// This is a `RevlogEntry` with the added semantics that the associated |
|
89 | /// This is a `RevlogEntry` with the added semantics that the associated | |
88 | /// data should meet the requirements for `changelog`, materialized by |
|
90 | /// data should meet the requirements for `changelog`, materialized by | |
89 | /// the fact that `data()` constructs a `ChangelogRevisionData`. |
|
91 | /// the fact that `data()` constructs a `ChangelogRevisionData`. | |
90 | /// In case that promise would be broken, the `data` method returns an error. |
|
92 | /// In case that promise would be broken, the `data` method returns an error. | |
91 | #[derive(Clone)] |
|
93 | #[derive(Clone)] | |
92 | pub struct ChangelogEntry<'changelog> { |
|
94 | pub struct ChangelogEntry<'changelog> { | |
93 | /// Same data, as a generic `RevlogEntry`. |
|
95 | /// Same data, as a generic `RevlogEntry`. | |
94 | pub(crate) revlog_entry: RevlogEntry<'changelog>, |
|
96 | pub(crate) revlog_entry: RevlogEntry<'changelog>, | |
95 | } |
|
97 | } | |
96 |
|
98 | |||
97 | impl<'changelog> ChangelogEntry<'changelog> { |
|
99 | impl<'changelog> ChangelogEntry<'changelog> { | |
98 | pub fn data<'a>( |
|
100 | pub fn data<'a>( | |
99 | &'a self, |
|
101 | &'a self, | |
100 | ) -> Result<ChangelogRevisionData<'changelog>, RevlogError> { |
|
102 | ) -> Result<ChangelogRevisionData<'changelog>, RevlogError> { | |
101 | let bytes = self.revlog_entry.data()?; |
|
103 | let bytes = self.revlog_entry.data()?; | |
102 | if bytes.is_empty() { |
|
104 | if bytes.is_empty() { | |
103 | Ok(ChangelogRevisionData::null()) |
|
105 | Ok(ChangelogRevisionData::null()) | |
104 | } else { |
|
106 | } else { | |
105 | Ok(ChangelogRevisionData::new(bytes).map_err(|err| { |
|
107 | Ok(ChangelogRevisionData::new(bytes).map_err(|err| { | |
106 | RevlogError::Other(HgError::CorruptedRepository(format!( |
|
108 | RevlogError::Other(HgError::CorruptedRepository(format!( | |
107 | "Invalid changelog data for revision {}: {:?}", |
|
109 | "Invalid changelog data for revision {}: {:?}", | |
108 | self.revlog_entry.revision(), |
|
110 | self.revlog_entry.revision(), | |
109 | err |
|
111 | err | |
110 | ))) |
|
112 | ))) | |
111 | })?) |
|
113 | })?) | |
112 | } |
|
114 | } | |
113 | } |
|
115 | } | |
114 |
|
116 | |||
115 | /// Obtain a reference to the underlying `RevlogEntry`. |
|
117 | /// Obtain a reference to the underlying `RevlogEntry`. | |
116 | /// |
|
118 | /// | |
117 | /// This allows the caller to access the information that is common |
|
119 | /// This allows the caller to access the information that is common | |
118 | /// to all revlog entries: revision number, node id, parent revisions etc. |
|
120 | /// to all revlog entries: revision number, node id, parent revisions etc. | |
119 | pub fn as_revlog_entry(&self) -> &RevlogEntry { |
|
121 | pub fn as_revlog_entry(&self) -> &RevlogEntry { | |
120 | &self.revlog_entry |
|
122 | &self.revlog_entry | |
121 | } |
|
123 | } | |
122 |
|
124 | |||
123 | pub fn p1_entry(&self) -> Result<Option<ChangelogEntry>, RevlogError> { |
|
125 | pub fn p1_entry(&self) -> Result<Option<ChangelogEntry>, RevlogError> { | |
124 | Ok(self |
|
126 | Ok(self | |
125 | .revlog_entry |
|
127 | .revlog_entry | |
126 | .p1_entry()? |
|
128 | .p1_entry()? | |
127 | .map(|revlog_entry| Self { revlog_entry })) |
|
129 | .map(|revlog_entry| Self { revlog_entry })) | |
128 | } |
|
130 | } | |
129 |
|
131 | |||
130 | pub fn p2_entry(&self) -> Result<Option<ChangelogEntry>, RevlogError> { |
|
132 | pub fn p2_entry(&self) -> Result<Option<ChangelogEntry>, RevlogError> { | |
131 | Ok(self |
|
133 | Ok(self | |
132 | .revlog_entry |
|
134 | .revlog_entry | |
133 | .p2_entry()? |
|
135 | .p2_entry()? | |
134 | .map(|revlog_entry| Self { revlog_entry })) |
|
136 | .map(|revlog_entry| Self { revlog_entry })) | |
135 | } |
|
137 | } | |
136 | } |
|
138 | } | |
137 |
|
139 | |||
138 | /// `Changelog` entry which knows how to interpret the `changelog` data bytes. |
|
140 | /// `Changelog` entry which knows how to interpret the `changelog` data bytes. | |
139 | #[derive(PartialEq)] |
|
141 | #[derive(PartialEq)] | |
140 | pub struct ChangelogRevisionData<'changelog> { |
|
142 | pub struct ChangelogRevisionData<'changelog> { | |
141 | /// The data bytes of the `changelog` entry. |
|
143 | /// The data bytes of the `changelog` entry. | |
142 | bytes: Cow<'changelog, [u8]>, |
|
144 | bytes: Cow<'changelog, [u8]>, | |
143 | /// The end offset for the hex manifest (not including the newline) |
|
145 | /// The end offset for the hex manifest (not including the newline) | |
144 | manifest_end: usize, |
|
146 | manifest_end: usize, | |
145 | /// The end offset for the user+email (not including the newline) |
|
147 | /// The end offset for the user+email (not including the newline) | |
146 | user_end: usize, |
|
148 | user_end: usize, | |
147 | /// The end offset for the timestamp+timezone+extras (not including the |
|
149 | /// The end offset for the timestamp+timezone+extras (not including the | |
148 | /// newline) |
|
150 | /// newline) | |
149 | timestamp_end: usize, |
|
151 | timestamp_end: usize, | |
150 | /// The end offset for the file list (not including the newline) |
|
152 | /// The end offset for the file list (not including the newline) | |
151 | files_end: usize, |
|
153 | files_end: usize, | |
152 | } |
|
154 | } | |
153 |
|
155 | |||
154 | impl<'changelog> ChangelogRevisionData<'changelog> { |
|
156 | impl<'changelog> ChangelogRevisionData<'changelog> { | |
155 | fn new(bytes: Cow<'changelog, [u8]>) -> Result<Self, HgError> { |
|
157 | fn new(bytes: Cow<'changelog, [u8]>) -> Result<Self, HgError> { | |
156 | let mut line_iter = bytes.split(|b| b == &b'\n'); |
|
158 | let mut line_iter = bytes.split(|b| b == &b'\n'); | |
157 | let manifest_end = line_iter |
|
159 | let manifest_end = line_iter | |
158 | .next() |
|
160 | .next() | |
159 | .expect("Empty iterator from split()?") |
|
161 | .expect("Empty iterator from split()?") | |
160 | .len(); |
|
162 | .len(); | |
161 | let user_slice = line_iter.next().ok_or_else(|| { |
|
163 | let user_slice = line_iter.next().ok_or_else(|| { | |
162 | HgError::corrupted("Changeset data truncated after manifest line") |
|
164 | HgError::corrupted("Changeset data truncated after manifest line") | |
163 | })?; |
|
165 | })?; | |
164 | let user_end = manifest_end + 1 + user_slice.len(); |
|
166 | let user_end = manifest_end + 1 + user_slice.len(); | |
165 | let timestamp_slice = line_iter.next().ok_or_else(|| { |
|
167 | let timestamp_slice = line_iter.next().ok_or_else(|| { | |
166 | HgError::corrupted("Changeset data truncated after user line") |
|
168 | HgError::corrupted("Changeset data truncated after user line") | |
167 | })?; |
|
169 | })?; | |
168 | let timestamp_end = user_end + 1 + timestamp_slice.len(); |
|
170 | let timestamp_end = user_end + 1 + timestamp_slice.len(); | |
169 | let mut files_end = timestamp_end + 1; |
|
171 | let mut files_end = timestamp_end + 1; | |
170 | loop { |
|
172 | loop { | |
171 | let line = line_iter.next().ok_or_else(|| { |
|
173 | let line = line_iter.next().ok_or_else(|| { | |
172 | HgError::corrupted("Changeset data truncated in files list") |
|
174 | HgError::corrupted("Changeset data truncated in files list") | |
173 | })?; |
|
175 | })?; | |
174 | if line.is_empty() { |
|
176 | if line.is_empty() { | |
175 | if files_end == bytes.len() { |
|
177 | if files_end == bytes.len() { | |
176 | // The list of files ended with a single newline (there |
|
178 | // The list of files ended with a single newline (there | |
177 | // should be two) |
|
179 | // should be two) | |
178 | return Err(HgError::corrupted( |
|
180 | return Err(HgError::corrupted( | |
179 | "Changeset data truncated after files list", |
|
181 | "Changeset data truncated after files list", | |
180 | )); |
|
182 | )); | |
181 | } |
|
183 | } | |
182 | files_end -= 1; |
|
184 | files_end -= 1; | |
183 | break; |
|
185 | break; | |
184 | } |
|
186 | } | |
185 | files_end += line.len() + 1; |
|
187 | files_end += line.len() + 1; | |
186 | } |
|
188 | } | |
187 |
|
189 | |||
188 | Ok(Self { |
|
190 | Ok(Self { | |
189 | bytes, |
|
191 | bytes, | |
190 | manifest_end, |
|
192 | manifest_end, | |
191 | user_end, |
|
193 | user_end, | |
192 | timestamp_end, |
|
194 | timestamp_end, | |
193 | files_end, |
|
195 | files_end, | |
194 | }) |
|
196 | }) | |
195 | } |
|
197 | } | |
196 |
|
198 | |||
197 | fn null() -> Self { |
|
199 | fn null() -> Self { | |
198 | Self::new(Cow::Borrowed( |
|
200 | Self::new(Cow::Borrowed( | |
199 | b"0000000000000000000000000000000000000000\n\n0 0\n\n", |
|
201 | b"0000000000000000000000000000000000000000\n\n0 0\n\n", | |
200 | )) |
|
202 | )) | |
201 | .unwrap() |
|
203 | .unwrap() | |
202 | } |
|
204 | } | |
203 |
|
205 | |||
204 | /// Return an iterator over the lines of the entry. |
|
206 | /// Return an iterator over the lines of the entry. | |
205 | pub fn lines(&self) -> impl Iterator<Item = &[u8]> { |
|
207 | pub fn lines(&self) -> impl Iterator<Item = &[u8]> { | |
206 | self.bytes.split(|b| b == &b'\n') |
|
208 | self.bytes.split(|b| b == &b'\n') | |
207 | } |
|
209 | } | |
208 |
|
210 | |||
209 | /// Return the node id of the `manifest` referenced by this `changelog` |
|
211 | /// Return the node id of the `manifest` referenced by this `changelog` | |
210 | /// entry. |
|
212 | /// entry. | |
211 | pub fn manifest_node(&self) -> Result<Node, HgError> { |
|
213 | pub fn manifest_node(&self) -> Result<Node, HgError> { | |
212 | let manifest_node_hex = &self.bytes[..self.manifest_end]; |
|
214 | let manifest_node_hex = &self.bytes[..self.manifest_end]; | |
213 | Node::from_hex_for_repo(manifest_node_hex) |
|
215 | Node::from_hex_for_repo(manifest_node_hex) | |
214 | } |
|
216 | } | |
215 |
|
217 | |||
216 | /// The full user string (usually a name followed by an email enclosed in |
|
218 | /// The full user string (usually a name followed by an email enclosed in | |
217 | /// angle brackets) |
|
219 | /// angle brackets) | |
218 | pub fn user(&self) -> &[u8] { |
|
220 | pub fn user(&self) -> &[u8] { | |
219 | &self.bytes[self.manifest_end + 1..self.user_end] |
|
221 | &self.bytes[self.manifest_end + 1..self.user_end] | |
220 | } |
|
222 | } | |
221 |
|
223 | |||
222 | /// The full timestamp line (timestamp in seconds, offset in seconds, and |
|
224 | /// The full timestamp line (timestamp in seconds, offset in seconds, and | |
223 | /// possibly extras) |
|
225 | /// possibly extras) | |
224 | // TODO: We should expose this in a more useful way |
|
226 | // TODO: We should expose this in a more useful way | |
225 | pub fn timestamp_line(&self) -> &[u8] { |
|
227 | pub fn timestamp_line(&self) -> &[u8] { | |
226 | &self.bytes[self.user_end + 1..self.timestamp_end] |
|
228 | &self.bytes[self.user_end + 1..self.timestamp_end] | |
227 | } |
|
229 | } | |
228 |
|
230 | |||
229 | /// The files changed in this revision. |
|
231 | /// The files changed in this revision. | |
230 | pub fn files(&self) -> impl Iterator<Item = &HgPath> { |
|
232 | pub fn files(&self) -> impl Iterator<Item = &HgPath> { | |
231 | self.bytes[self.timestamp_end + 1..self.files_end] |
|
233 | self.bytes[self.timestamp_end + 1..self.files_end] | |
232 | .split(|b| b == &b'\n') |
|
234 | .split(|b| b == &b'\n') | |
233 | .map(HgPath::new) |
|
235 | .map(HgPath::new) | |
234 | } |
|
236 | } | |
235 |
|
237 | |||
236 | /// The change description. |
|
238 | /// The change description. | |
237 | pub fn description(&self) -> &[u8] { |
|
239 | pub fn description(&self) -> &[u8] { | |
238 | &self.bytes[self.files_end + 2..] |
|
240 | &self.bytes[self.files_end + 2..] | |
239 | } |
|
241 | } | |
240 | } |
|
242 | } | |
241 |
|
243 | |||
242 | impl Debug for ChangelogRevisionData<'_> { |
|
244 | impl Debug for ChangelogRevisionData<'_> { | |
243 | fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result { |
|
245 | fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result { | |
244 | f.debug_struct("ChangelogRevisionData") |
|
246 | f.debug_struct("ChangelogRevisionData") | |
245 | .field("bytes", &debug_bytes(&self.bytes)) |
|
247 | .field("bytes", &debug_bytes(&self.bytes)) | |
246 | .field("manifest", &debug_bytes(&self.bytes[..self.manifest_end])) |
|
248 | .field("manifest", &debug_bytes(&self.bytes[..self.manifest_end])) | |
247 | .field( |
|
249 | .field( | |
248 | "user", |
|
250 | "user", | |
249 | &debug_bytes( |
|
251 | &debug_bytes( | |
250 | &self.bytes[self.manifest_end + 1..self.user_end], |
|
252 | &self.bytes[self.manifest_end + 1..self.user_end], | |
251 | ), |
|
253 | ), | |
252 | ) |
|
254 | ) | |
253 | .field( |
|
255 | .field( | |
254 | "timestamp", |
|
256 | "timestamp", | |
255 | &debug_bytes( |
|
257 | &debug_bytes( | |
256 | &self.bytes[self.user_end + 1..self.timestamp_end], |
|
258 | &self.bytes[self.user_end + 1..self.timestamp_end], | |
257 | ), |
|
259 | ), | |
258 | ) |
|
260 | ) | |
259 | .field( |
|
261 | .field( | |
260 | "files", |
|
262 | "files", | |
261 | &debug_bytes( |
|
263 | &debug_bytes( | |
262 | &self.bytes[self.timestamp_end + 1..self.files_end], |
|
264 | &self.bytes[self.timestamp_end + 1..self.files_end], | |
263 | ), |
|
265 | ), | |
264 | ) |
|
266 | ) | |
265 | .field( |
|
267 | .field( | |
266 | "description", |
|
268 | "description", | |
267 | &debug_bytes(&self.bytes[self.files_end + 2..]), |
|
269 | &debug_bytes(&self.bytes[self.files_end + 2..]), | |
268 | ) |
|
270 | ) | |
269 | .finish() |
|
271 | .finish() | |
270 | } |
|
272 | } | |
271 | } |
|
273 | } | |
272 |
|
274 | |||
273 | fn debug_bytes(bytes: &[u8]) -> String { |
|
275 | fn debug_bytes(bytes: &[u8]) -> String { | |
274 | String::from_utf8_lossy( |
|
276 | String::from_utf8_lossy( | |
275 | &bytes.iter().flat_map(|b| escape_default(*b)).collect_vec(), |
|
277 | &bytes.iter().flat_map(|b| escape_default(*b)).collect_vec(), | |
276 | ) |
|
278 | ) | |
277 | .to_string() |
|
279 | .to_string() | |
278 | } |
|
280 | } | |
279 |
|
281 | |||
280 | #[cfg(test)] |
|
282 | #[cfg(test)] | |
281 | mod tests { |
|
283 | mod tests { | |
282 | use super::*; |
|
284 | use super::*; | |
283 | use crate::vfs::Vfs; |
|
285 | use crate::vfs::Vfs; | |
284 | use crate::NULL_REVISION; |
|
286 | use crate::NULL_REVISION; | |
285 | use pretty_assertions::assert_eq; |
|
287 | use pretty_assertions::assert_eq; | |
286 |
|
288 | |||
287 | #[test] |
|
289 | #[test] | |
288 | fn test_create_changelogrevisiondata_invalid() { |
|
290 | fn test_create_changelogrevisiondata_invalid() { | |
289 | // Completely empty |
|
291 | // Completely empty | |
290 | assert!(ChangelogRevisionData::new(Cow::Borrowed(b"abcd")).is_err()); |
|
292 | assert!(ChangelogRevisionData::new(Cow::Borrowed(b"abcd")).is_err()); | |
291 | // No newline after manifest |
|
293 | // No newline after manifest | |
292 | assert!(ChangelogRevisionData::new(Cow::Borrowed(b"abcd")).is_err()); |
|
294 | assert!(ChangelogRevisionData::new(Cow::Borrowed(b"abcd")).is_err()); | |
293 | // No newline after user |
|
295 | // No newline after user | |
294 | assert!(ChangelogRevisionData::new(Cow::Borrowed(b"abcd\n")).is_err()); |
|
296 | assert!(ChangelogRevisionData::new(Cow::Borrowed(b"abcd\n")).is_err()); | |
295 | // No newline after timestamp |
|
297 | // No newline after timestamp | |
296 | assert!( |
|
298 | assert!( | |
297 | ChangelogRevisionData::new(Cow::Borrowed(b"abcd\n\n0 0")).is_err() |
|
299 | ChangelogRevisionData::new(Cow::Borrowed(b"abcd\n\n0 0")).is_err() | |
298 | ); |
|
300 | ); | |
299 | // Missing newline after files |
|
301 | // Missing newline after files | |
300 | assert!(ChangelogRevisionData::new(Cow::Borrowed( |
|
302 | assert!(ChangelogRevisionData::new(Cow::Borrowed( | |
301 | b"abcd\n\n0 0\nfile1\nfile2" |
|
303 | b"abcd\n\n0 0\nfile1\nfile2" | |
302 | )) |
|
304 | )) | |
303 | .is_err(),); |
|
305 | .is_err(),); | |
304 | // Only one newline after files |
|
306 | // Only one newline after files | |
305 | assert!(ChangelogRevisionData::new(Cow::Borrowed( |
|
307 | assert!(ChangelogRevisionData::new(Cow::Borrowed( | |
306 | b"abcd\n\n0 0\nfile1\nfile2\n" |
|
308 | b"abcd\n\n0 0\nfile1\nfile2\n" | |
307 | )) |
|
309 | )) | |
308 | .is_err(),); |
|
310 | .is_err(),); | |
309 | } |
|
311 | } | |
310 |
|
312 | |||
311 | #[test] |
|
313 | #[test] | |
312 | fn test_create_changelogrevisiondata() { |
|
314 | fn test_create_changelogrevisiondata() { | |
313 | let data = ChangelogRevisionData::new(Cow::Borrowed( |
|
315 | let data = ChangelogRevisionData::new(Cow::Borrowed( | |
314 | b"0123456789abcdef0123456789abcdef01234567 |
|
316 | b"0123456789abcdef0123456789abcdef01234567 | |
315 | Some One <someone@example.com> |
|
317 | Some One <someone@example.com> | |
316 | 0 0 |
|
318 | 0 0 | |
317 | file1 |
|
319 | file1 | |
318 | file2 |
|
320 | file2 | |
319 |
|
321 | |||
320 | some |
|
322 | some | |
321 | commit |
|
323 | commit | |
322 | message", |
|
324 | message", | |
323 | )) |
|
325 | )) | |
324 | .unwrap(); |
|
326 | .unwrap(); | |
325 | assert_eq!( |
|
327 | assert_eq!( | |
326 | data.manifest_node().unwrap(), |
|
328 | data.manifest_node().unwrap(), | |
327 | Node::from_hex("0123456789abcdef0123456789abcdef01234567") |
|
329 | Node::from_hex("0123456789abcdef0123456789abcdef01234567") | |
328 | .unwrap() |
|
330 | .unwrap() | |
329 | ); |
|
331 | ); | |
330 | assert_eq!(data.user(), b"Some One <someone@example.com>"); |
|
332 | assert_eq!(data.user(), b"Some One <someone@example.com>"); | |
331 | assert_eq!(data.timestamp_line(), b"0 0"); |
|
333 | assert_eq!(data.timestamp_line(), b"0 0"); | |
332 | assert_eq!( |
|
334 | assert_eq!( | |
333 | data.files().collect_vec(), |
|
335 | data.files().collect_vec(), | |
334 | vec![HgPath::new("file1"), HgPath::new("file2")] |
|
336 | vec![HgPath::new("file1"), HgPath::new("file2")] | |
335 | ); |
|
337 | ); | |
336 | assert_eq!(data.description(), b"some\ncommit\nmessage"); |
|
338 | assert_eq!(data.description(), b"some\ncommit\nmessage"); | |
337 | } |
|
339 | } | |
338 |
|
340 | |||
339 | #[test] |
|
341 | #[test] | |
340 | fn test_data_from_rev_null() -> Result<(), RevlogError> { |
|
342 | fn test_data_from_rev_null() -> Result<(), RevlogError> { | |
341 | // an empty revlog will be enough for this case |
|
343 | // an empty revlog will be enough for this case | |
342 | let temp = tempfile::tempdir().unwrap(); |
|
344 | let temp = tempfile::tempdir().unwrap(); | |
343 | let vfs = Vfs { base: temp.path() }; |
|
345 | let vfs = Vfs { base: temp.path() }; | |
344 | std::fs::write(temp.path().join("foo.i"), b"").unwrap(); |
|
346 | std::fs::write(temp.path().join("foo.i"), b"").unwrap(); | |
345 | let revlog = Revlog::open(&vfs, "foo.i", None, false).unwrap(); |
|
347 | let revlog = | |
|
348 | Revlog::open(&vfs, "foo.i", None, RevlogOpenOptions::new()) | |||
|
349 | .unwrap(); | |||
346 |
|
350 | |||
347 | let changelog = Changelog { revlog }; |
|
351 | let changelog = Changelog { revlog }; | |
348 | assert_eq!( |
|
352 | assert_eq!( | |
349 | changelog.data_for_rev(NULL_REVISION.into())?, |
|
353 | changelog.data_for_rev(NULL_REVISION.into())?, | |
350 | ChangelogRevisionData::null() |
|
354 | ChangelogRevisionData::null() | |
351 | ); |
|
355 | ); | |
352 | // same with the intermediate entry object |
|
356 | // same with the intermediate entry object | |
353 | assert_eq!( |
|
357 | assert_eq!( | |
354 | changelog.entry_for_rev(NULL_REVISION.into())?.data()?, |
|
358 | changelog.entry_for_rev(NULL_REVISION.into())?.data()?, | |
355 | ChangelogRevisionData::null() |
|
359 | ChangelogRevisionData::null() | |
356 | ); |
|
360 | ); | |
357 | Ok(()) |
|
361 | Ok(()) | |
358 | } |
|
362 | } | |
359 | } |
|
363 | } |
@@ -1,239 +1,245 b'' | |||||
1 | use crate::errors::HgError; |
|
1 | use crate::errors::HgError; | |
2 | use crate::exit_codes; |
|
2 | use crate::exit_codes; | |
3 | use crate::repo::Repo; |
|
3 | use crate::repo::Repo; | |
4 | use crate::revlog::path_encode::path_encode; |
|
4 | use crate::revlog::path_encode::path_encode; | |
5 | use crate::revlog::NodePrefix; |
|
5 | use crate::revlog::NodePrefix; | |
6 | use crate::revlog::Revision; |
|
6 | use crate::revlog::Revision; | |
7 | use crate::revlog::RevlogEntry; |
|
7 | use crate::revlog::RevlogEntry; | |
8 | use crate::revlog::{Revlog, RevlogError}; |
|
8 | use crate::revlog::{Revlog, RevlogError}; | |
9 | use crate::utils::files::get_path_from_bytes; |
|
9 | use crate::utils::files::get_path_from_bytes; | |
10 | use crate::utils::hg_path::HgPath; |
|
10 | use crate::utils::hg_path::HgPath; | |
11 | use crate::utils::SliceExt; |
|
11 | use crate::utils::SliceExt; | |
12 | use crate::Graph; |
|
12 | use crate::Graph; | |
13 | use crate::GraphError; |
|
13 | use crate::GraphError; | |
|
14 | use crate::RevlogOpenOptions; | |||
14 | use crate::UncheckedRevision; |
|
15 | use crate::UncheckedRevision; | |
15 | use std::path::PathBuf; |
|
16 | use std::path::PathBuf; | |
16 |
|
17 | |||
17 | /// A specialized `Revlog` to work with file data logs. |
|
18 | /// A specialized `Revlog` to work with file data logs. | |
18 | pub struct Filelog { |
|
19 | pub struct Filelog { | |
19 | /// The generic `revlog` format. |
|
20 | /// The generic `revlog` format. | |
20 | revlog: Revlog, |
|
21 | revlog: Revlog, | |
21 | } |
|
22 | } | |
22 |
|
23 | |||
23 | impl Graph for Filelog { |
|
24 | impl Graph for Filelog { | |
24 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { |
|
25 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { | |
25 | self.revlog.parents(rev) |
|
26 | self.revlog.parents(rev) | |
26 | } |
|
27 | } | |
27 | } |
|
28 | } | |
28 |
|
29 | |||
29 | impl Filelog { |
|
30 | impl Filelog { | |
30 | pub fn open_vfs( |
|
31 | pub fn open_vfs( | |
31 | store_vfs: &crate::vfs::Vfs<'_>, |
|
32 | store_vfs: &crate::vfs::Vfs<'_>, | |
32 | file_path: &HgPath, |
|
33 | file_path: &HgPath, | |
|
34 | options: RevlogOpenOptions, | |||
33 | ) -> Result<Self, HgError> { |
|
35 | ) -> Result<Self, HgError> { | |
34 | let index_path = store_path(file_path, b".i"); |
|
36 | let index_path = store_path(file_path, b".i"); | |
35 | let data_path = store_path(file_path, b".d"); |
|
37 | let data_path = store_path(file_path, b".d"); | |
36 | let revlog = |
|
38 | let revlog = | |
37 |
Revlog::open(store_vfs, index_path, Some(&data_path), |
|
39 | Revlog::open(store_vfs, index_path, Some(&data_path), options)?; | |
38 | Ok(Self { revlog }) |
|
40 | Ok(Self { revlog }) | |
39 | } |
|
41 | } | |
40 |
|
42 | |||
41 | pub fn open(repo: &Repo, file_path: &HgPath) -> Result<Self, HgError> { |
|
43 | pub fn open( | |
42 | Self::open_vfs(&repo.store_vfs(), file_path) |
|
44 | repo: &Repo, | |
|
45 | file_path: &HgPath, | |||
|
46 | options: RevlogOpenOptions, | |||
|
47 | ) -> Result<Self, HgError> { | |||
|
48 | Self::open_vfs(&repo.store_vfs(), file_path, options) | |||
43 | } |
|
49 | } | |
44 |
|
50 | |||
45 | /// The given node ID is that of the file as found in a filelog, not of a |
|
51 | /// The given node ID is that of the file as found in a filelog, not of a | |
46 | /// changeset. |
|
52 | /// changeset. | |
47 | pub fn data_for_node( |
|
53 | pub fn data_for_node( | |
48 | &self, |
|
54 | &self, | |
49 | file_node: impl Into<NodePrefix>, |
|
55 | file_node: impl Into<NodePrefix>, | |
50 | ) -> Result<FilelogRevisionData, RevlogError> { |
|
56 | ) -> Result<FilelogRevisionData, RevlogError> { | |
51 | let file_rev = self.revlog.rev_from_node(file_node.into())?; |
|
57 | let file_rev = self.revlog.rev_from_node(file_node.into())?; | |
52 | self.data_for_rev(file_rev.into()) |
|
58 | self.data_for_rev(file_rev.into()) | |
53 | } |
|
59 | } | |
54 |
|
60 | |||
55 | /// The given revision is that of the file as found in a filelog, not of a |
|
61 | /// The given revision is that of the file as found in a filelog, not of a | |
56 | /// changeset. |
|
62 | /// changeset. | |
57 | pub fn data_for_rev( |
|
63 | pub fn data_for_rev( | |
58 | &self, |
|
64 | &self, | |
59 | file_rev: UncheckedRevision, |
|
65 | file_rev: UncheckedRevision, | |
60 | ) -> Result<FilelogRevisionData, RevlogError> { |
|
66 | ) -> Result<FilelogRevisionData, RevlogError> { | |
61 | let data: Vec<u8> = self.revlog.get_rev_data(file_rev)?.into_owned(); |
|
67 | let data: Vec<u8> = self.revlog.get_rev_data(file_rev)?.into_owned(); | |
62 | Ok(FilelogRevisionData(data)) |
|
68 | Ok(FilelogRevisionData(data)) | |
63 | } |
|
69 | } | |
64 |
|
70 | |||
65 | /// The given node ID is that of the file as found in a filelog, not of a |
|
71 | /// The given node ID is that of the file as found in a filelog, not of a | |
66 | /// changeset. |
|
72 | /// changeset. | |
67 | pub fn entry_for_node( |
|
73 | pub fn entry_for_node( | |
68 | &self, |
|
74 | &self, | |
69 | file_node: impl Into<NodePrefix>, |
|
75 | file_node: impl Into<NodePrefix>, | |
70 | ) -> Result<FilelogEntry, RevlogError> { |
|
76 | ) -> Result<FilelogEntry, RevlogError> { | |
71 | let file_rev = self.revlog.rev_from_node(file_node.into())?; |
|
77 | let file_rev = self.revlog.rev_from_node(file_node.into())?; | |
72 | self.entry_for_checked_rev(file_rev) |
|
78 | self.entry_for_checked_rev(file_rev) | |
73 | } |
|
79 | } | |
74 |
|
80 | |||
75 | /// The given revision is that of the file as found in a filelog, not of a |
|
81 | /// The given revision is that of the file as found in a filelog, not of a | |
76 | /// changeset. |
|
82 | /// changeset. | |
77 | pub fn entry_for_rev( |
|
83 | pub fn entry_for_rev( | |
78 | &self, |
|
84 | &self, | |
79 | file_rev: UncheckedRevision, |
|
85 | file_rev: UncheckedRevision, | |
80 | ) -> Result<FilelogEntry, RevlogError> { |
|
86 | ) -> Result<FilelogEntry, RevlogError> { | |
81 | Ok(FilelogEntry(self.revlog.get_entry(file_rev)?)) |
|
87 | Ok(FilelogEntry(self.revlog.get_entry(file_rev)?)) | |
82 | } |
|
88 | } | |
83 |
|
89 | |||
84 | fn entry_for_checked_rev( |
|
90 | fn entry_for_checked_rev( | |
85 | &self, |
|
91 | &self, | |
86 | file_rev: Revision, |
|
92 | file_rev: Revision, | |
87 | ) -> Result<FilelogEntry, RevlogError> { |
|
93 | ) -> Result<FilelogEntry, RevlogError> { | |
88 | Ok(FilelogEntry( |
|
94 | Ok(FilelogEntry( | |
89 | self.revlog.get_entry_for_checked_rev(file_rev)?, |
|
95 | self.revlog.get_entry_for_checked_rev(file_rev)?, | |
90 | )) |
|
96 | )) | |
91 | } |
|
97 | } | |
92 | } |
|
98 | } | |
93 |
|
99 | |||
94 | fn store_path(hg_path: &HgPath, suffix: &[u8]) -> PathBuf { |
|
100 | fn store_path(hg_path: &HgPath, suffix: &[u8]) -> PathBuf { | |
95 | let encoded_bytes = |
|
101 | let encoded_bytes = | |
96 | path_encode(&[b"data/", hg_path.as_bytes(), suffix].concat()); |
|
102 | path_encode(&[b"data/", hg_path.as_bytes(), suffix].concat()); | |
97 | get_path_from_bytes(&encoded_bytes).into() |
|
103 | get_path_from_bytes(&encoded_bytes).into() | |
98 | } |
|
104 | } | |
99 |
|
105 | |||
100 | pub struct FilelogEntry<'a>(RevlogEntry<'a>); |
|
106 | pub struct FilelogEntry<'a>(RevlogEntry<'a>); | |
101 |
|
107 | |||
102 | impl FilelogEntry<'_> { |
|
108 | impl FilelogEntry<'_> { | |
103 | /// `self.data()` can be expensive, with decompression and delta |
|
109 | /// `self.data()` can be expensive, with decompression and delta | |
104 | /// resolution. |
|
110 | /// resolution. | |
105 | /// |
|
111 | /// | |
106 | /// *Without* paying this cost, based on revlog index information |
|
112 | /// *Without* paying this cost, based on revlog index information | |
107 | /// including `RevlogEntry::uncompressed_len`: |
|
113 | /// including `RevlogEntry::uncompressed_len`: | |
108 | /// |
|
114 | /// | |
109 | /// * Returns `true` if the length that `self.data().file_data().len()` |
|
115 | /// * Returns `true` if the length that `self.data().file_data().len()` | |
110 | /// would return is definitely **not equal** to `other_len`. |
|
116 | /// would return is definitely **not equal** to `other_len`. | |
111 | /// * Returns `false` if available information is inconclusive. |
|
117 | /// * Returns `false` if available information is inconclusive. | |
112 | pub fn file_data_len_not_equal_to(&self, other_len: u64) -> bool { |
|
118 | pub fn file_data_len_not_equal_to(&self, other_len: u64) -> bool { | |
113 | // Relevant code that implement this behavior in Python code: |
|
119 | // Relevant code that implement this behavior in Python code: | |
114 | // basefilectx.cmp, filelog.size, storageutil.filerevisioncopied, |
|
120 | // basefilectx.cmp, filelog.size, storageutil.filerevisioncopied, | |
115 | // revlog.size, revlog.rawsize |
|
121 | // revlog.size, revlog.rawsize | |
116 |
|
122 | |||
117 | // Let’s call `file_data_len` what would be returned by |
|
123 | // Let’s call `file_data_len` what would be returned by | |
118 | // `self.data().file_data().len()`. |
|
124 | // `self.data().file_data().len()`. | |
119 |
|
125 | |||
120 | if self.0.is_censored() { |
|
126 | if self.0.is_censored() { | |
121 | let file_data_len = 0; |
|
127 | let file_data_len = 0; | |
122 | return other_len != file_data_len; |
|
128 | return other_len != file_data_len; | |
123 | } |
|
129 | } | |
124 |
|
130 | |||
125 | if self.0.has_length_affecting_flag_processor() { |
|
131 | if self.0.has_length_affecting_flag_processor() { | |
126 | // We can’t conclude anything about `file_data_len`. |
|
132 | // We can’t conclude anything about `file_data_len`. | |
127 | return false; |
|
133 | return false; | |
128 | } |
|
134 | } | |
129 |
|
135 | |||
130 | // Revlog revisions (usually) have metadata for the size of |
|
136 | // Revlog revisions (usually) have metadata for the size of | |
131 | // their data after decompression and delta resolution |
|
137 | // their data after decompression and delta resolution | |
132 | // as would be returned by `Revlog::get_rev_data`. |
|
138 | // as would be returned by `Revlog::get_rev_data`. | |
133 | // |
|
139 | // | |
134 | // For filelogs this is the file’s contents preceded by an optional |
|
140 | // For filelogs this is the file’s contents preceded by an optional | |
135 | // metadata block. |
|
141 | // metadata block. | |
136 | let uncompressed_len = if let Some(l) = self.0.uncompressed_len() { |
|
142 | let uncompressed_len = if let Some(l) = self.0.uncompressed_len() { | |
137 | l as u64 |
|
143 | l as u64 | |
138 | } else { |
|
144 | } else { | |
139 | // The field was set to -1, the actual uncompressed len is unknown. |
|
145 | // The field was set to -1, the actual uncompressed len is unknown. | |
140 | // We need to decompress to say more. |
|
146 | // We need to decompress to say more. | |
141 | return false; |
|
147 | return false; | |
142 | }; |
|
148 | }; | |
143 | // `uncompressed_len = file_data_len + optional_metadata_len`, |
|
149 | // `uncompressed_len = file_data_len + optional_metadata_len`, | |
144 | // so `file_data_len <= uncompressed_len`. |
|
150 | // so `file_data_len <= uncompressed_len`. | |
145 | if uncompressed_len < other_len { |
|
151 | if uncompressed_len < other_len { | |
146 | // Transitively, `file_data_len < other_len`. |
|
152 | // Transitively, `file_data_len < other_len`. | |
147 | // So `other_len != file_data_len` definitely. |
|
153 | // So `other_len != file_data_len` definitely. | |
148 | return true; |
|
154 | return true; | |
149 | } |
|
155 | } | |
150 |
|
156 | |||
151 | if uncompressed_len == other_len + 4 { |
|
157 | if uncompressed_len == other_len + 4 { | |
152 | // It’s possible that `file_data_len == other_len` with an empty |
|
158 | // It’s possible that `file_data_len == other_len` with an empty | |
153 | // metadata block (2 start marker bytes + 2 end marker bytes). |
|
159 | // metadata block (2 start marker bytes + 2 end marker bytes). | |
154 | // This happens when there wouldn’t otherwise be metadata, but |
|
160 | // This happens when there wouldn’t otherwise be metadata, but | |
155 | // the first 2 bytes of file data happen to match a start marker |
|
161 | // the first 2 bytes of file data happen to match a start marker | |
156 | // and would be ambiguous. |
|
162 | // and would be ambiguous. | |
157 | return false; |
|
163 | return false; | |
158 | } |
|
164 | } | |
159 |
|
165 | |||
160 | if !self.0.has_p1() { |
|
166 | if !self.0.has_p1() { | |
161 | // There may or may not be copy metadata, so we can’t deduce more |
|
167 | // There may or may not be copy metadata, so we can’t deduce more | |
162 | // about `file_data_len` without computing file data. |
|
168 | // about `file_data_len` without computing file data. | |
163 | return false; |
|
169 | return false; | |
164 | } |
|
170 | } | |
165 |
|
171 | |||
166 | // Filelog ancestry is not meaningful in the way changelog ancestry is. |
|
172 | // Filelog ancestry is not meaningful in the way changelog ancestry is. | |
167 | // It only provides hints to delta generation. |
|
173 | // It only provides hints to delta generation. | |
168 | // p1 and p2 are set to null when making a copy or rename since |
|
174 | // p1 and p2 are set to null when making a copy or rename since | |
169 | // contents are likely unrelatedto what might have previously existed |
|
175 | // contents are likely unrelatedto what might have previously existed | |
170 | // at the destination path. |
|
176 | // at the destination path. | |
171 | // |
|
177 | // | |
172 | // Conversely, since here p1 is non-null, there is no copy metadata. |
|
178 | // Conversely, since here p1 is non-null, there is no copy metadata. | |
173 | // Note that this reasoning may be invalidated in the presence of |
|
179 | // Note that this reasoning may be invalidated in the presence of | |
174 | // merges made by some previous versions of Mercurial that |
|
180 | // merges made by some previous versions of Mercurial that | |
175 | // swapped p1 and p2. See <https://bz.mercurial-scm.org/show_bug.cgi?id=6528> |
|
181 | // swapped p1 and p2. See <https://bz.mercurial-scm.org/show_bug.cgi?id=6528> | |
176 | // and `tests/test-issue6528.t`. |
|
182 | // and `tests/test-issue6528.t`. | |
177 | // |
|
183 | // | |
178 | // Since copy metadata is currently the only kind of metadata |
|
184 | // Since copy metadata is currently the only kind of metadata | |
179 | // kept in revlog data of filelogs, |
|
185 | // kept in revlog data of filelogs, | |
180 | // this `FilelogEntry` does not have such metadata: |
|
186 | // this `FilelogEntry` does not have such metadata: | |
181 | let file_data_len = uncompressed_len; |
|
187 | let file_data_len = uncompressed_len; | |
182 |
|
188 | |||
183 | file_data_len != other_len |
|
189 | file_data_len != other_len | |
184 | } |
|
190 | } | |
185 |
|
191 | |||
186 | pub fn data(&self) -> Result<FilelogRevisionData, HgError> { |
|
192 | pub fn data(&self) -> Result<FilelogRevisionData, HgError> { | |
187 | let data = self.0.data(); |
|
193 | let data = self.0.data(); | |
188 | match data { |
|
194 | match data { | |
189 | Ok(data) => Ok(FilelogRevisionData(data.into_owned())), |
|
195 | Ok(data) => Ok(FilelogRevisionData(data.into_owned())), | |
190 | // Errors other than `HgError` should not happen at this point |
|
196 | // Errors other than `HgError` should not happen at this point | |
191 | Err(e) => match e { |
|
197 | Err(e) => match e { | |
192 | RevlogError::Other(hg_error) => Err(hg_error), |
|
198 | RevlogError::Other(hg_error) => Err(hg_error), | |
193 | revlog_error => Err(HgError::abort( |
|
199 | revlog_error => Err(HgError::abort( | |
194 | revlog_error.to_string(), |
|
200 | revlog_error.to_string(), | |
195 | exit_codes::ABORT, |
|
201 | exit_codes::ABORT, | |
196 | None, |
|
202 | None, | |
197 | )), |
|
203 | )), | |
198 | }, |
|
204 | }, | |
199 | } |
|
205 | } | |
200 | } |
|
206 | } | |
201 | } |
|
207 | } | |
202 |
|
208 | |||
203 | /// The data for one revision in a filelog, uncompressed and delta-resolved. |
|
209 | /// The data for one revision in a filelog, uncompressed and delta-resolved. | |
204 | pub struct FilelogRevisionData(Vec<u8>); |
|
210 | pub struct FilelogRevisionData(Vec<u8>); | |
205 |
|
211 | |||
206 | impl FilelogRevisionData { |
|
212 | impl FilelogRevisionData { | |
207 | /// Split into metadata and data |
|
213 | /// Split into metadata and data | |
208 | pub fn split(&self) -> Result<(Option<&[u8]>, &[u8]), HgError> { |
|
214 | pub fn split(&self) -> Result<(Option<&[u8]>, &[u8]), HgError> { | |
209 | const DELIMITER: &[u8; 2] = &[b'\x01', b'\n']; |
|
215 | const DELIMITER: &[u8; 2] = &[b'\x01', b'\n']; | |
210 |
|
216 | |||
211 | if let Some(rest) = self.0.drop_prefix(DELIMITER) { |
|
217 | if let Some(rest) = self.0.drop_prefix(DELIMITER) { | |
212 | if let Some((metadata, data)) = rest.split_2_by_slice(DELIMITER) { |
|
218 | if let Some((metadata, data)) = rest.split_2_by_slice(DELIMITER) { | |
213 | Ok((Some(metadata), data)) |
|
219 | Ok((Some(metadata), data)) | |
214 | } else { |
|
220 | } else { | |
215 | Err(HgError::corrupted( |
|
221 | Err(HgError::corrupted( | |
216 | "Missing metadata end delimiter in filelog entry", |
|
222 | "Missing metadata end delimiter in filelog entry", | |
217 | )) |
|
223 | )) | |
218 | } |
|
224 | } | |
219 | } else { |
|
225 | } else { | |
220 | Ok((None, &self.0)) |
|
226 | Ok((None, &self.0)) | |
221 | } |
|
227 | } | |
222 | } |
|
228 | } | |
223 |
|
229 | |||
224 | /// Returns the file contents at this revision, stripped of any metadata |
|
230 | /// Returns the file contents at this revision, stripped of any metadata | |
225 | pub fn file_data(&self) -> Result<&[u8], HgError> { |
|
231 | pub fn file_data(&self) -> Result<&[u8], HgError> { | |
226 | let (_metadata, data) = self.split()?; |
|
232 | let (_metadata, data) = self.split()?; | |
227 | Ok(data) |
|
233 | Ok(data) | |
228 | } |
|
234 | } | |
229 |
|
235 | |||
230 | /// Consume the entry, and convert it into data, discarding any metadata, |
|
236 | /// Consume the entry, and convert it into data, discarding any metadata, | |
231 | /// if present. |
|
237 | /// if present. | |
232 | pub fn into_file_data(self) -> Result<Vec<u8>, HgError> { |
|
238 | pub fn into_file_data(self) -> Result<Vec<u8>, HgError> { | |
233 | if let (Some(_metadata), data) = self.split()? { |
|
239 | if let (Some(_metadata), data) = self.split()? { | |
234 | Ok(data.to_owned()) |
|
240 | Ok(data.to_owned()) | |
235 | } else { |
|
241 | } else { | |
236 | Ok(self.0) |
|
242 | Ok(self.0) | |
237 | } |
|
243 | } | |
238 | } |
|
244 | } | |
239 | } |
|
245 | } |
@@ -1,775 +1,770 b'' | |||||
1 | use std::fmt::Debug; |
|
1 | use std::fmt::Debug; | |
2 | use std::ops::Deref; |
|
2 | use std::ops::Deref; | |
3 |
|
3 | |||
4 | use byteorder::{BigEndian, ByteOrder}; |
|
4 | use byteorder::{BigEndian, ByteOrder}; | |
5 | use bytes_cast::{unaligned, BytesCast}; |
|
5 | use bytes_cast::{unaligned, BytesCast}; | |
6 |
|
6 | |||
7 | use super::REVIDX_KNOWN_FLAGS; |
|
7 | use super::REVIDX_KNOWN_FLAGS; | |
8 | use crate::errors::HgError; |
|
8 | use crate::errors::HgError; | |
9 | use crate::node::{NODE_BYTES_LENGTH, STORED_NODE_ID_BYTES}; |
|
9 | use crate::node::{NODE_BYTES_LENGTH, STORED_NODE_ID_BYTES}; | |
10 | use crate::revlog::node::Node; |
|
10 | use crate::revlog::node::Node; | |
11 | use crate::revlog::{Revision, NULL_REVISION}; |
|
11 | use crate::revlog::{Revision, NULL_REVISION}; | |
12 | use crate::{Graph, GraphError, RevlogError, RevlogIndex, UncheckedRevision}; |
|
12 | use crate::{Graph, GraphError, RevlogError, RevlogIndex, UncheckedRevision}; | |
13 |
|
13 | |||
14 | pub const INDEX_ENTRY_SIZE: usize = 64; |
|
14 | pub const INDEX_ENTRY_SIZE: usize = 64; | |
15 | pub const COMPRESSION_MODE_INLINE: u8 = 2; |
|
15 | pub const COMPRESSION_MODE_INLINE: u8 = 2; | |
16 |
|
16 | |||
17 | pub struct IndexHeader { |
|
17 | pub struct IndexHeader { | |
18 | header_bytes: [u8; 4], |
|
18 | pub(super) header_bytes: [u8; 4], | |
19 | } |
|
19 | } | |
20 |
|
20 | |||
21 | #[derive(Copy, Clone)] |
|
21 | #[derive(Copy, Clone)] | |
22 | pub struct IndexHeaderFlags { |
|
22 | pub struct IndexHeaderFlags { | |
23 | flags: u16, |
|
23 | flags: u16, | |
24 | } |
|
24 | } | |
25 |
|
25 | |||
26 | /// Corresponds to the high bits of `_format_flags` in python |
|
26 | /// Corresponds to the high bits of `_format_flags` in python | |
27 | impl IndexHeaderFlags { |
|
27 | impl IndexHeaderFlags { | |
28 | /// Corresponds to FLAG_INLINE_DATA in python |
|
28 | /// Corresponds to FLAG_INLINE_DATA in python | |
29 | pub fn is_inline(self) -> bool { |
|
29 | pub fn is_inline(self) -> bool { | |
30 | self.flags & 1 != 0 |
|
30 | self.flags & 1 != 0 | |
31 | } |
|
31 | } | |
32 | /// Corresponds to FLAG_GENERALDELTA in python |
|
32 | /// Corresponds to FLAG_GENERALDELTA in python | |
33 | pub fn uses_generaldelta(self) -> bool { |
|
33 | pub fn uses_generaldelta(self) -> bool { | |
34 | self.flags & 2 != 0 |
|
34 | self.flags & 2 != 0 | |
35 | } |
|
35 | } | |
36 | } |
|
36 | } | |
37 |
|
37 | |||
38 | /// Corresponds to the INDEX_HEADER structure, |
|
38 | /// Corresponds to the INDEX_HEADER structure, | |
39 | /// which is parsed as a `header` variable in `_loadindex` in `revlog.py` |
|
39 | /// which is parsed as a `header` variable in `_loadindex` in `revlog.py` | |
40 | impl IndexHeader { |
|
40 | impl IndexHeader { | |
41 | fn format_flags(&self) -> IndexHeaderFlags { |
|
41 | fn format_flags(&self) -> IndexHeaderFlags { | |
42 | // No "unknown flags" check here, unlike in python. Maybe there should |
|
42 | // No "unknown flags" check here, unlike in python. Maybe there should | |
43 | // be. |
|
43 | // be. | |
44 | IndexHeaderFlags { |
|
44 | IndexHeaderFlags { | |
45 | flags: BigEndian::read_u16(&self.header_bytes[0..2]), |
|
45 | flags: BigEndian::read_u16(&self.header_bytes[0..2]), | |
46 | } |
|
46 | } | |
47 | } |
|
47 | } | |
48 |
|
48 | |||
49 | /// The only revlog version currently supported by rhg. |
|
49 | /// The only revlog version currently supported by rhg. | |
50 | const REVLOGV1: u16 = 1; |
|
50 | const REVLOGV1: u16 = 1; | |
51 |
|
51 | |||
52 | /// Corresponds to `_format_version` in Python. |
|
52 | /// Corresponds to `_format_version` in Python. | |
53 | fn format_version(&self) -> u16 { |
|
53 | fn format_version(&self) -> u16 { | |
54 | BigEndian::read_u16(&self.header_bytes[2..4]) |
|
54 | BigEndian::read_u16(&self.header_bytes[2..4]) | |
55 | } |
|
55 | } | |
56 |
|
56 | |||
57 | const EMPTY_INDEX_HEADER: IndexHeader = IndexHeader { |
|
57 | pub fn parse(index_bytes: &[u8]) -> Result<Option<IndexHeader>, HgError> { | |
58 | // We treat an empty file as a valid index with no entries. |
|
|||
59 | // Here we make an arbitrary choice of what we assume the format of the |
|
|||
60 | // index to be (V1, using generaldelta). |
|
|||
61 | // This doesn't matter too much, since we're only doing read-only |
|
|||
62 | // access. but the value corresponds to the `new_header` variable in |
|
|||
63 | // `revlog.py`, `_loadindex` |
|
|||
64 | header_bytes: [0, 3, 0, 1], |
|
|||
65 | }; |
|
|||
66 |
|
||||
67 | fn parse(index_bytes: &[u8]) -> Result<IndexHeader, HgError> { |
|
|||
68 | if index_bytes.is_empty() { |
|
58 | if index_bytes.is_empty() { | |
69 |
return Ok( |
|
59 | return Ok(None); | |
70 | } |
|
60 | } | |
71 | if index_bytes.len() < 4 { |
|
61 | if index_bytes.len() < 4 { | |
72 | return Err(HgError::corrupted( |
|
62 | return Err(HgError::corrupted( | |
73 | "corrupted revlog: can't read the index format header", |
|
63 | "corrupted revlog: can't read the index format header", | |
74 | )); |
|
64 | )); | |
75 | } |
|
65 | } | |
76 | Ok(IndexHeader { |
|
66 | Ok(Some(IndexHeader { | |
77 | header_bytes: { |
|
67 | header_bytes: { | |
78 | let bytes: [u8; 4] = |
|
68 | let bytes: [u8; 4] = | |
79 | index_bytes[0..4].try_into().expect("impossible"); |
|
69 | index_bytes[0..4].try_into().expect("impossible"); | |
80 | bytes |
|
70 | bytes | |
81 | }, |
|
71 | }, | |
82 | }) |
|
72 | })) | |
83 | } |
|
73 | } | |
84 | } |
|
74 | } | |
85 |
|
75 | |||
86 | /// Abstracts the access to the index bytes since they can be spread between |
|
76 | /// Abstracts the access to the index bytes since they can be spread between | |
87 | /// the immutable (bytes) part and the mutable (added) part if any appends |
|
77 | /// the immutable (bytes) part and the mutable (added) part if any appends | |
88 | /// happened. This makes it transparent for the callers. |
|
78 | /// happened. This makes it transparent for the callers. | |
89 | struct IndexData { |
|
79 | struct IndexData { | |
90 | /// Immutable bytes, most likely taken from disk |
|
80 | /// Immutable bytes, most likely taken from disk | |
91 | bytes: Box<dyn Deref<Target = [u8]> + Send>, |
|
81 | bytes: Box<dyn Deref<Target = [u8]> + Send>, | |
92 | /// Bytes that were added after reading the index |
|
82 | /// Bytes that were added after reading the index | |
93 | added: Vec<u8>, |
|
83 | added: Vec<u8>, | |
94 | } |
|
84 | } | |
95 |
|
85 | |||
96 | impl IndexData { |
|
86 | impl IndexData { | |
97 | pub fn new(bytes: Box<dyn Deref<Target = [u8]> + Send>) -> Self { |
|
87 | pub fn new(bytes: Box<dyn Deref<Target = [u8]> + Send>) -> Self { | |
98 | Self { |
|
88 | Self { | |
99 | bytes, |
|
89 | bytes, | |
100 | added: vec![], |
|
90 | added: vec![], | |
101 | } |
|
91 | } | |
102 | } |
|
92 | } | |
103 |
|
93 | |||
104 | pub fn len(&self) -> usize { |
|
94 | pub fn len(&self) -> usize { | |
105 | self.bytes.len() + self.added.len() |
|
95 | self.bytes.len() + self.added.len() | |
106 | } |
|
96 | } | |
107 | } |
|
97 | } | |
108 |
|
98 | |||
109 | impl std::ops::Index<std::ops::Range<usize>> for IndexData { |
|
99 | impl std::ops::Index<std::ops::Range<usize>> for IndexData { | |
110 | type Output = [u8]; |
|
100 | type Output = [u8]; | |
111 |
|
101 | |||
112 | fn index(&self, index: std::ops::Range<usize>) -> &Self::Output { |
|
102 | fn index(&self, index: std::ops::Range<usize>) -> &Self::Output { | |
113 | let start = index.start; |
|
103 | let start = index.start; | |
114 | let end = index.end; |
|
104 | let end = index.end; | |
115 | let immutable_len = self.bytes.len(); |
|
105 | let immutable_len = self.bytes.len(); | |
116 | if start < immutable_len { |
|
106 | if start < immutable_len { | |
117 | if end > immutable_len { |
|
107 | if end > immutable_len { | |
118 | panic!("index data cannot span existing and added ranges"); |
|
108 | panic!("index data cannot span existing and added ranges"); | |
119 | } |
|
109 | } | |
120 | &self.bytes[index] |
|
110 | &self.bytes[index] | |
121 | } else { |
|
111 | } else { | |
122 | &self.added[start - immutable_len..end - immutable_len] |
|
112 | &self.added[start - immutable_len..end - immutable_len] | |
123 | } |
|
113 | } | |
124 | } |
|
114 | } | |
125 | } |
|
115 | } | |
126 |
|
116 | |||
127 | pub struct RevisionDataParams { |
|
117 | pub struct RevisionDataParams { | |
128 | flags: u16, |
|
118 | flags: u16, | |
129 | data_offset: u64, |
|
119 | data_offset: u64, | |
130 | data_compressed_length: i32, |
|
120 | data_compressed_length: i32, | |
131 | data_uncompressed_length: i32, |
|
121 | data_uncompressed_length: i32, | |
132 | data_delta_base: i32, |
|
122 | data_delta_base: i32, | |
133 | link_rev: i32, |
|
123 | link_rev: i32, | |
134 | parent_rev_1: i32, |
|
124 | parent_rev_1: i32, | |
135 | parent_rev_2: i32, |
|
125 | parent_rev_2: i32, | |
136 | node_id: [u8; NODE_BYTES_LENGTH], |
|
126 | node_id: [u8; NODE_BYTES_LENGTH], | |
137 | _sidedata_offset: u64, |
|
127 | _sidedata_offset: u64, | |
138 | _sidedata_compressed_length: i32, |
|
128 | _sidedata_compressed_length: i32, | |
139 | data_compression_mode: u8, |
|
129 | data_compression_mode: u8, | |
140 | _sidedata_compression_mode: u8, |
|
130 | _sidedata_compression_mode: u8, | |
141 | _rank: i32, |
|
131 | _rank: i32, | |
142 | } |
|
132 | } | |
143 |
|
133 | |||
144 | #[derive(BytesCast)] |
|
134 | #[derive(BytesCast)] | |
145 | #[repr(C)] |
|
135 | #[repr(C)] | |
146 | pub struct RevisionDataV1 { |
|
136 | pub struct RevisionDataV1 { | |
147 | data_offset_or_flags: unaligned::U64Be, |
|
137 | data_offset_or_flags: unaligned::U64Be, | |
148 | data_compressed_length: unaligned::I32Be, |
|
138 | data_compressed_length: unaligned::I32Be, | |
149 | data_uncompressed_length: unaligned::I32Be, |
|
139 | data_uncompressed_length: unaligned::I32Be, | |
150 | data_delta_base: unaligned::I32Be, |
|
140 | data_delta_base: unaligned::I32Be, | |
151 | link_rev: unaligned::I32Be, |
|
141 | link_rev: unaligned::I32Be, | |
152 | parent_rev_1: unaligned::I32Be, |
|
142 | parent_rev_1: unaligned::I32Be, | |
153 | parent_rev_2: unaligned::I32Be, |
|
143 | parent_rev_2: unaligned::I32Be, | |
154 | node_id: [u8; STORED_NODE_ID_BYTES], |
|
144 | node_id: [u8; STORED_NODE_ID_BYTES], | |
155 | } |
|
145 | } | |
156 |
|
146 | |||
157 | fn _static_assert_size_of_revision_data_v1() { |
|
147 | fn _static_assert_size_of_revision_data_v1() { | |
158 | let _ = std::mem::transmute::<RevisionDataV1, [u8; 64]>; |
|
148 | let _ = std::mem::transmute::<RevisionDataV1, [u8; 64]>; | |
159 | } |
|
149 | } | |
160 |
|
150 | |||
161 | impl RevisionDataParams { |
|
151 | impl RevisionDataParams { | |
162 | pub fn validate(&self) -> Result<(), RevlogError> { |
|
152 | pub fn validate(&self) -> Result<(), RevlogError> { | |
163 | if self.flags & !REVIDX_KNOWN_FLAGS != 0 { |
|
153 | if self.flags & !REVIDX_KNOWN_FLAGS != 0 { | |
164 | return Err(RevlogError::corrupted(format!( |
|
154 | return Err(RevlogError::corrupted(format!( | |
165 | "unknown revlog index flags: {}", |
|
155 | "unknown revlog index flags: {}", | |
166 | self.flags |
|
156 | self.flags | |
167 | ))); |
|
157 | ))); | |
168 | } |
|
158 | } | |
169 | if self.data_compression_mode != COMPRESSION_MODE_INLINE { |
|
159 | if self.data_compression_mode != COMPRESSION_MODE_INLINE { | |
170 | return Err(RevlogError::corrupted(format!( |
|
160 | return Err(RevlogError::corrupted(format!( | |
171 | "invalid data compression mode: {}", |
|
161 | "invalid data compression mode: {}", | |
172 | self.data_compression_mode |
|
162 | self.data_compression_mode | |
173 | ))); |
|
163 | ))); | |
174 | } |
|
164 | } | |
175 | // FIXME isn't this only for v2 or changelog v2? |
|
165 | // FIXME isn't this only for v2 or changelog v2? | |
176 | if self._sidedata_compression_mode != COMPRESSION_MODE_INLINE { |
|
166 | if self._sidedata_compression_mode != COMPRESSION_MODE_INLINE { | |
177 | return Err(RevlogError::corrupted(format!( |
|
167 | return Err(RevlogError::corrupted(format!( | |
178 | "invalid sidedata compression mode: {}", |
|
168 | "invalid sidedata compression mode: {}", | |
179 | self._sidedata_compression_mode |
|
169 | self._sidedata_compression_mode | |
180 | ))); |
|
170 | ))); | |
181 | } |
|
171 | } | |
182 | Ok(()) |
|
172 | Ok(()) | |
183 | } |
|
173 | } | |
184 |
|
174 | |||
185 | pub fn into_v1(self) -> RevisionDataV1 { |
|
175 | pub fn into_v1(self) -> RevisionDataV1 { | |
186 | let data_offset_or_flags = self.data_offset << 16 | self.flags as u64; |
|
176 | let data_offset_or_flags = self.data_offset << 16 | self.flags as u64; | |
187 | let mut node_id = [0; STORED_NODE_ID_BYTES]; |
|
177 | let mut node_id = [0; STORED_NODE_ID_BYTES]; | |
188 | node_id[..NODE_BYTES_LENGTH].copy_from_slice(&self.node_id); |
|
178 | node_id[..NODE_BYTES_LENGTH].copy_from_slice(&self.node_id); | |
189 | RevisionDataV1 { |
|
179 | RevisionDataV1 { | |
190 | data_offset_or_flags: data_offset_or_flags.into(), |
|
180 | data_offset_or_flags: data_offset_or_flags.into(), | |
191 | data_compressed_length: self.data_compressed_length.into(), |
|
181 | data_compressed_length: self.data_compressed_length.into(), | |
192 | data_uncompressed_length: self.data_uncompressed_length.into(), |
|
182 | data_uncompressed_length: self.data_uncompressed_length.into(), | |
193 | data_delta_base: self.data_delta_base.into(), |
|
183 | data_delta_base: self.data_delta_base.into(), | |
194 | link_rev: self.link_rev.into(), |
|
184 | link_rev: self.link_rev.into(), | |
195 | parent_rev_1: self.parent_rev_1.into(), |
|
185 | parent_rev_1: self.parent_rev_1.into(), | |
196 | parent_rev_2: self.parent_rev_2.into(), |
|
186 | parent_rev_2: self.parent_rev_2.into(), | |
197 | node_id, |
|
187 | node_id, | |
198 | } |
|
188 | } | |
199 | } |
|
189 | } | |
200 | } |
|
190 | } | |
201 |
|
191 | |||
202 | /// A Revlog index |
|
192 | /// A Revlog index | |
203 | pub struct Index { |
|
193 | pub struct Index { | |
204 | bytes: IndexData, |
|
194 | bytes: IndexData, | |
205 | /// Offsets of starts of index blocks. |
|
195 | /// Offsets of starts of index blocks. | |
206 | /// Only needed when the index is interleaved with data. |
|
196 | /// Only needed when the index is interleaved with data. | |
207 | offsets: Option<Vec<usize>>, |
|
197 | offsets: Option<Vec<usize>>, | |
208 | uses_generaldelta: bool, |
|
198 | uses_generaldelta: bool, | |
209 | } |
|
199 | } | |
210 |
|
200 | |||
211 | impl Debug for Index { |
|
201 | impl Debug for Index { | |
212 | fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { |
|
202 | fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { | |
213 | f.debug_struct("Index") |
|
203 | f.debug_struct("Index") | |
214 | .field("offsets", &self.offsets) |
|
204 | .field("offsets", &self.offsets) | |
215 | .field("uses_generaldelta", &self.uses_generaldelta) |
|
205 | .field("uses_generaldelta", &self.uses_generaldelta) | |
216 | .finish() |
|
206 | .finish() | |
217 | } |
|
207 | } | |
218 | } |
|
208 | } | |
219 |
|
209 | |||
220 | impl Graph for Index { |
|
210 | impl Graph for Index { | |
221 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { |
|
211 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { | |
222 | let err = || GraphError::ParentOutOfRange(rev); |
|
212 | let err = || GraphError::ParentOutOfRange(rev); | |
223 | match self.get_entry(rev) { |
|
213 | match self.get_entry(rev) { | |
224 | Some(entry) => { |
|
214 | Some(entry) => { | |
225 | // The C implementation checks that the parents are valid |
|
215 | // The C implementation checks that the parents are valid | |
226 | // before returning |
|
216 | // before returning | |
227 | Ok([ |
|
217 | Ok([ | |
228 | self.check_revision(entry.p1()).ok_or_else(err)?, |
|
218 | self.check_revision(entry.p1()).ok_or_else(err)?, | |
229 | self.check_revision(entry.p2()).ok_or_else(err)?, |
|
219 | self.check_revision(entry.p2()).ok_or_else(err)?, | |
230 | ]) |
|
220 | ]) | |
231 | } |
|
221 | } | |
232 | None => Ok([NULL_REVISION, NULL_REVISION]), |
|
222 | None => Ok([NULL_REVISION, NULL_REVISION]), | |
233 | } |
|
223 | } | |
234 | } |
|
224 | } | |
235 | } |
|
225 | } | |
236 |
|
226 | |||
237 | impl Index { |
|
227 | impl Index { | |
238 | /// Create an index from bytes. |
|
228 | /// Create an index from bytes. | |
239 | /// Calculate the start of each entry when is_inline is true. |
|
229 | /// Calculate the start of each entry when is_inline is true. | |
240 | pub fn new( |
|
230 | pub fn new( | |
241 | bytes: Box<dyn Deref<Target = [u8]> + Send>, |
|
231 | bytes: Box<dyn Deref<Target = [u8]> + Send>, | |
|
232 | default_header: IndexHeader, | |||
242 | ) -> Result<Self, HgError> { |
|
233 | ) -> Result<Self, HgError> { | |
243 | let header = IndexHeader::parse(bytes.as_ref())?; |
|
234 | let header = | |
|
235 | IndexHeader::parse(bytes.as_ref())?.unwrap_or(default_header); | |||
244 |
|
236 | |||
245 | if header.format_version() != IndexHeader::REVLOGV1 { |
|
237 | if header.format_version() != IndexHeader::REVLOGV1 { | |
246 | // A proper new version should have had a repo/store |
|
238 | // A proper new version should have had a repo/store | |
247 | // requirement. |
|
239 | // requirement. | |
248 | return Err(HgError::corrupted("unsupported revlog version")); |
|
240 | return Err(HgError::corrupted("unsupported revlog version")); | |
249 | } |
|
241 | } | |
250 |
|
242 | |||
251 | // This is only correct because we know version is REVLOGV1. |
|
243 | // This is only correct because we know version is REVLOGV1. | |
252 | // In v2 we always use generaldelta, while in v0 we never use |
|
244 | // In v2 we always use generaldelta, while in v0 we never use | |
253 | // generaldelta. Similar for [is_inline] (it's only used in v1). |
|
245 | // generaldelta. Similar for [is_inline] (it's only used in v1). | |
254 | let uses_generaldelta = header.format_flags().uses_generaldelta(); |
|
246 | let uses_generaldelta = header.format_flags().uses_generaldelta(); | |
255 |
|
247 | |||
256 | if header.format_flags().is_inline() { |
|
248 | if header.format_flags().is_inline() { | |
257 | let mut offset: usize = 0; |
|
249 | let mut offset: usize = 0; | |
258 | let mut offsets = Vec::new(); |
|
250 | let mut offsets = Vec::new(); | |
259 |
|
251 | |||
260 | while offset + INDEX_ENTRY_SIZE <= bytes.len() { |
|
252 | while offset + INDEX_ENTRY_SIZE <= bytes.len() { | |
261 | offsets.push(offset); |
|
253 | offsets.push(offset); | |
262 | let end = offset + INDEX_ENTRY_SIZE; |
|
254 | let end = offset + INDEX_ENTRY_SIZE; | |
263 | let entry = IndexEntry { |
|
255 | let entry = IndexEntry { | |
264 | bytes: &bytes[offset..end], |
|
256 | bytes: &bytes[offset..end], | |
265 | offset_override: None, |
|
257 | offset_override: None, | |
266 | }; |
|
258 | }; | |
267 |
|
259 | |||
268 | offset += INDEX_ENTRY_SIZE + entry.compressed_len() as usize; |
|
260 | offset += INDEX_ENTRY_SIZE + entry.compressed_len() as usize; | |
269 | } |
|
261 | } | |
270 |
|
262 | |||
271 | if offset == bytes.len() { |
|
263 | if offset == bytes.len() { | |
272 | Ok(Self { |
|
264 | Ok(Self { | |
273 | bytes: IndexData::new(bytes), |
|
265 | bytes: IndexData::new(bytes), | |
274 | offsets: Some(offsets), |
|
266 | offsets: Some(offsets), | |
275 | uses_generaldelta, |
|
267 | uses_generaldelta, | |
276 | }) |
|
268 | }) | |
277 | } else { |
|
269 | } else { | |
278 | Err(HgError::corrupted("unexpected inline revlog length")) |
|
270 | Err(HgError::corrupted("unexpected inline revlog length")) | |
279 | } |
|
271 | } | |
280 | } else { |
|
272 | } else { | |
281 | Ok(Self { |
|
273 | Ok(Self { | |
282 | bytes: IndexData::new(bytes), |
|
274 | bytes: IndexData::new(bytes), | |
283 | offsets: None, |
|
275 | offsets: None, | |
284 | uses_generaldelta, |
|
276 | uses_generaldelta, | |
285 | }) |
|
277 | }) | |
286 | } |
|
278 | } | |
287 | } |
|
279 | } | |
288 |
|
280 | |||
289 | pub fn uses_generaldelta(&self) -> bool { |
|
281 | pub fn uses_generaldelta(&self) -> bool { | |
290 | self.uses_generaldelta |
|
282 | self.uses_generaldelta | |
291 | } |
|
283 | } | |
292 |
|
284 | |||
293 | /// Value of the inline flag. |
|
285 | /// Value of the inline flag. | |
294 | pub fn is_inline(&self) -> bool { |
|
286 | pub fn is_inline(&self) -> bool { | |
295 | self.offsets.is_some() |
|
287 | self.offsets.is_some() | |
296 | } |
|
288 | } | |
297 |
|
289 | |||
298 | /// Return a slice of bytes if `revlog` is inline. Panic if not. |
|
290 | /// Return a slice of bytes if `revlog` is inline. Panic if not. | |
299 | pub fn data(&self, start: usize, end: usize) -> &[u8] { |
|
291 | pub fn data(&self, start: usize, end: usize) -> &[u8] { | |
300 | if !self.is_inline() { |
|
292 | if !self.is_inline() { | |
301 | panic!("tried to access data in the index of a revlog that is not inline"); |
|
293 | panic!("tried to access data in the index of a revlog that is not inline"); | |
302 | } |
|
294 | } | |
303 | &self.bytes[start..end] |
|
295 | &self.bytes[start..end] | |
304 | } |
|
296 | } | |
305 |
|
297 | |||
306 | /// Return number of entries of the revlog index. |
|
298 | /// Return number of entries of the revlog index. | |
307 | pub fn len(&self) -> usize { |
|
299 | pub fn len(&self) -> usize { | |
308 | if let Some(offsets) = &self.offsets { |
|
300 | if let Some(offsets) = &self.offsets { | |
309 | offsets.len() |
|
301 | offsets.len() | |
310 | } else { |
|
302 | } else { | |
311 | self.bytes.len() / INDEX_ENTRY_SIZE |
|
303 | self.bytes.len() / INDEX_ENTRY_SIZE | |
312 | } |
|
304 | } | |
313 | } |
|
305 | } | |
314 |
|
306 | |||
315 | /// Returns `true` if the `Index` has zero `entries`. |
|
307 | /// Returns `true` if the `Index` has zero `entries`. | |
316 | pub fn is_empty(&self) -> bool { |
|
308 | pub fn is_empty(&self) -> bool { | |
317 | self.len() == 0 |
|
309 | self.len() == 0 | |
318 | } |
|
310 | } | |
319 |
|
311 | |||
320 | /// Return the index entry corresponding to the given revision if it |
|
312 | /// Return the index entry corresponding to the given revision if it | |
321 | /// exists. |
|
313 | /// exists. | |
322 | pub fn get_entry(&self, rev: Revision) -> Option<IndexEntry> { |
|
314 | pub fn get_entry(&self, rev: Revision) -> Option<IndexEntry> { | |
323 | if rev == NULL_REVISION { |
|
315 | if rev == NULL_REVISION { | |
324 | return None; |
|
316 | return None; | |
325 | } |
|
317 | } | |
326 | Some(if let Some(offsets) = &self.offsets { |
|
318 | Some(if let Some(offsets) = &self.offsets { | |
327 | self.get_entry_inline(rev, offsets) |
|
319 | self.get_entry_inline(rev, offsets) | |
328 | } else { |
|
320 | } else { | |
329 | self.get_entry_separated(rev) |
|
321 | self.get_entry_separated(rev) | |
330 | }) |
|
322 | }) | |
331 | } |
|
323 | } | |
332 |
|
324 | |||
333 | fn get_entry_inline( |
|
325 | fn get_entry_inline( | |
334 | &self, |
|
326 | &self, | |
335 | rev: Revision, |
|
327 | rev: Revision, | |
336 | offsets: &[usize], |
|
328 | offsets: &[usize], | |
337 | ) -> IndexEntry { |
|
329 | ) -> IndexEntry { | |
338 | let start = offsets[rev.0 as usize]; |
|
330 | let start = offsets[rev.0 as usize]; | |
339 | let end = start + INDEX_ENTRY_SIZE; |
|
331 | let end = start + INDEX_ENTRY_SIZE; | |
340 | let bytes = &self.bytes[start..end]; |
|
332 | let bytes = &self.bytes[start..end]; | |
341 |
|
333 | |||
342 | // See IndexEntry for an explanation of this override. |
|
334 | // See IndexEntry for an explanation of this override. | |
343 | let offset_override = Some(end); |
|
335 | let offset_override = Some(end); | |
344 |
|
336 | |||
345 | IndexEntry { |
|
337 | IndexEntry { | |
346 | bytes, |
|
338 | bytes, | |
347 | offset_override, |
|
339 | offset_override, | |
348 | } |
|
340 | } | |
349 | } |
|
341 | } | |
350 |
|
342 | |||
351 | fn get_entry_separated(&self, rev: Revision) -> IndexEntry { |
|
343 | fn get_entry_separated(&self, rev: Revision) -> IndexEntry { | |
352 | let start = rev.0 as usize * INDEX_ENTRY_SIZE; |
|
344 | let start = rev.0 as usize * INDEX_ENTRY_SIZE; | |
353 | let end = start + INDEX_ENTRY_SIZE; |
|
345 | let end = start + INDEX_ENTRY_SIZE; | |
354 | let bytes = &self.bytes[start..end]; |
|
346 | let bytes = &self.bytes[start..end]; | |
355 |
|
347 | |||
356 | // Override the offset of the first revision as its bytes are used |
|
348 | // Override the offset of the first revision as its bytes are used | |
357 | // for the index's metadata (saving space because it is always 0) |
|
349 | // for the index's metadata (saving space because it is always 0) | |
358 | let offset_override = if rev == Revision(0) { Some(0) } else { None }; |
|
350 | let offset_override = if rev == Revision(0) { Some(0) } else { None }; | |
359 |
|
351 | |||
360 | IndexEntry { |
|
352 | IndexEntry { | |
361 | bytes, |
|
353 | bytes, | |
362 | offset_override, |
|
354 | offset_override, | |
363 | } |
|
355 | } | |
364 | } |
|
356 | } | |
365 |
|
357 | |||
366 | /// TODO move this to the trait probably, along with other things |
|
358 | /// TODO move this to the trait probably, along with other things | |
367 | pub fn append( |
|
359 | pub fn append( | |
368 | &mut self, |
|
360 | &mut self, | |
369 | revision_data: RevisionDataParams, |
|
361 | revision_data: RevisionDataParams, | |
370 | ) -> Result<(), RevlogError> { |
|
362 | ) -> Result<(), RevlogError> { | |
371 | revision_data.validate()?; |
|
363 | revision_data.validate()?; | |
372 | let new_offset = self.bytes.len(); |
|
364 | let new_offset = self.bytes.len(); | |
373 | if let Some(offsets) = self.offsets.as_mut() { |
|
365 | if let Some(offsets) = self.offsets.as_mut() { | |
374 | offsets.push(new_offset) |
|
366 | offsets.push(new_offset) | |
375 | } |
|
367 | } | |
376 | self.bytes.added.extend(revision_data.into_v1().as_bytes()); |
|
368 | self.bytes.added.extend(revision_data.into_v1().as_bytes()); | |
377 | Ok(()) |
|
369 | Ok(()) | |
378 | } |
|
370 | } | |
379 | } |
|
371 | } | |
380 |
|
372 | |||
381 | impl super::RevlogIndex for Index { |
|
373 | impl super::RevlogIndex for Index { | |
382 | fn len(&self) -> usize { |
|
374 | fn len(&self) -> usize { | |
383 | self.len() |
|
375 | self.len() | |
384 | } |
|
376 | } | |
385 |
|
377 | |||
386 | fn node(&self, rev: Revision) -> Option<&Node> { |
|
378 | fn node(&self, rev: Revision) -> Option<&Node> { | |
387 | self.get_entry(rev).map(|entry| entry.hash()) |
|
379 | self.get_entry(rev).map(|entry| entry.hash()) | |
388 | } |
|
380 | } | |
389 | } |
|
381 | } | |
390 |
|
382 | |||
391 | #[derive(Debug)] |
|
383 | #[derive(Debug)] | |
392 | pub struct IndexEntry<'a> { |
|
384 | pub struct IndexEntry<'a> { | |
393 | bytes: &'a [u8], |
|
385 | bytes: &'a [u8], | |
394 | /// Allows to override the offset value of the entry. |
|
386 | /// Allows to override the offset value of the entry. | |
395 | /// |
|
387 | /// | |
396 | /// For interleaved index and data, the offset stored in the index |
|
388 | /// For interleaved index and data, the offset stored in the index | |
397 | /// corresponds to the separated data offset. |
|
389 | /// corresponds to the separated data offset. | |
398 | /// It has to be overridden with the actual offset in the interleaved |
|
390 | /// It has to be overridden with the actual offset in the interleaved | |
399 | /// index which is just after the index block. |
|
391 | /// index which is just after the index block. | |
400 | /// |
|
392 | /// | |
401 | /// For separated index and data, the offset stored in the first index |
|
393 | /// For separated index and data, the offset stored in the first index | |
402 | /// entry is mixed with the index headers. |
|
394 | /// entry is mixed with the index headers. | |
403 | /// It has to be overridden with 0. |
|
395 | /// It has to be overridden with 0. | |
404 | offset_override: Option<usize>, |
|
396 | offset_override: Option<usize>, | |
405 | } |
|
397 | } | |
406 |
|
398 | |||
407 | impl<'a> IndexEntry<'a> { |
|
399 | impl<'a> IndexEntry<'a> { | |
408 | /// Return the offset of the data. |
|
400 | /// Return the offset of the data. | |
409 | pub fn offset(&self) -> usize { |
|
401 | pub fn offset(&self) -> usize { | |
410 | if let Some(offset_override) = self.offset_override { |
|
402 | if let Some(offset_override) = self.offset_override { | |
411 | offset_override |
|
403 | offset_override | |
412 | } else { |
|
404 | } else { | |
413 | let mut bytes = [0; 8]; |
|
405 | let mut bytes = [0; 8]; | |
414 | bytes[2..8].copy_from_slice(&self.bytes[0..=5]); |
|
406 | bytes[2..8].copy_from_slice(&self.bytes[0..=5]); | |
415 | BigEndian::read_u64(&bytes[..]) as usize |
|
407 | BigEndian::read_u64(&bytes[..]) as usize | |
416 | } |
|
408 | } | |
417 | } |
|
409 | } | |
418 |
|
410 | |||
419 | pub fn flags(&self) -> u16 { |
|
411 | pub fn flags(&self) -> u16 { | |
420 | BigEndian::read_u16(&self.bytes[6..=7]) |
|
412 | BigEndian::read_u16(&self.bytes[6..=7]) | |
421 | } |
|
413 | } | |
422 |
|
414 | |||
423 | /// Return the compressed length of the data. |
|
415 | /// Return the compressed length of the data. | |
424 | pub fn compressed_len(&self) -> u32 { |
|
416 | pub fn compressed_len(&self) -> u32 { | |
425 | BigEndian::read_u32(&self.bytes[8..=11]) |
|
417 | BigEndian::read_u32(&self.bytes[8..=11]) | |
426 | } |
|
418 | } | |
427 |
|
419 | |||
428 | /// Return the uncompressed length of the data. |
|
420 | /// Return the uncompressed length of the data. | |
429 | pub fn uncompressed_len(&self) -> i32 { |
|
421 | pub fn uncompressed_len(&self) -> i32 { | |
430 | BigEndian::read_i32(&self.bytes[12..=15]) |
|
422 | BigEndian::read_i32(&self.bytes[12..=15]) | |
431 | } |
|
423 | } | |
432 |
|
424 | |||
433 | /// Return the revision upon which the data has been derived. |
|
425 | /// Return the revision upon which the data has been derived. | |
434 | pub fn base_revision_or_base_of_delta_chain(&self) -> UncheckedRevision { |
|
426 | pub fn base_revision_or_base_of_delta_chain(&self) -> UncheckedRevision { | |
435 | // TODO Maybe return an Option when base_revision == rev? |
|
427 | // TODO Maybe return an Option when base_revision == rev? | |
436 | // Requires to add rev to IndexEntry |
|
428 | // Requires to add rev to IndexEntry | |
437 |
|
429 | |||
438 | BigEndian::read_i32(&self.bytes[16..]).into() |
|
430 | BigEndian::read_i32(&self.bytes[16..]).into() | |
439 | } |
|
431 | } | |
440 |
|
432 | |||
441 | pub fn link_revision(&self) -> UncheckedRevision { |
|
433 | pub fn link_revision(&self) -> UncheckedRevision { | |
442 | BigEndian::read_i32(&self.bytes[20..]).into() |
|
434 | BigEndian::read_i32(&self.bytes[20..]).into() | |
443 | } |
|
435 | } | |
444 |
|
436 | |||
445 | pub fn p1(&self) -> UncheckedRevision { |
|
437 | pub fn p1(&self) -> UncheckedRevision { | |
446 | BigEndian::read_i32(&self.bytes[24..]).into() |
|
438 | BigEndian::read_i32(&self.bytes[24..]).into() | |
447 | } |
|
439 | } | |
448 |
|
440 | |||
449 | pub fn p2(&self) -> UncheckedRevision { |
|
441 | pub fn p2(&self) -> UncheckedRevision { | |
450 | BigEndian::read_i32(&self.bytes[28..]).into() |
|
442 | BigEndian::read_i32(&self.bytes[28..]).into() | |
451 | } |
|
443 | } | |
452 |
|
444 | |||
453 | /// Return the hash of revision's full text. |
|
445 | /// Return the hash of revision's full text. | |
454 | /// |
|
446 | /// | |
455 | /// Currently, SHA-1 is used and only the first 20 bytes of this field |
|
447 | /// Currently, SHA-1 is used and only the first 20 bytes of this field | |
456 | /// are used. |
|
448 | /// are used. | |
457 | pub fn hash(&self) -> &'a Node { |
|
449 | pub fn hash(&self) -> &'a Node { | |
458 | (&self.bytes[32..52]).try_into().unwrap() |
|
450 | (&self.bytes[32..52]).try_into().unwrap() | |
459 | } |
|
451 | } | |
460 | } |
|
452 | } | |
461 |
|
453 | |||
462 | #[cfg(test)] |
|
454 | #[cfg(test)] | |
463 | mod tests { |
|
455 | mod tests { | |
464 | use super::*; |
|
456 | use super::*; | |
465 | use crate::node::NULL_NODE; |
|
457 | use crate::node::NULL_NODE; | |
466 |
|
458 | |||
467 | #[cfg(test)] |
|
459 | #[cfg(test)] | |
468 | #[derive(Debug, Copy, Clone)] |
|
460 | #[derive(Debug, Copy, Clone)] | |
469 | pub struct IndexEntryBuilder { |
|
461 | pub struct IndexEntryBuilder { | |
470 | is_first: bool, |
|
462 | is_first: bool, | |
471 | is_inline: bool, |
|
463 | is_inline: bool, | |
472 | is_general_delta: bool, |
|
464 | is_general_delta: bool, | |
473 | version: u16, |
|
465 | version: u16, | |
474 | offset: usize, |
|
466 | offset: usize, | |
475 | compressed_len: usize, |
|
467 | compressed_len: usize, | |
476 | uncompressed_len: usize, |
|
468 | uncompressed_len: usize, | |
477 | base_revision_or_base_of_delta_chain: Revision, |
|
469 | base_revision_or_base_of_delta_chain: Revision, | |
478 | link_revision: Revision, |
|
470 | link_revision: Revision, | |
479 | p1: Revision, |
|
471 | p1: Revision, | |
480 | p2: Revision, |
|
472 | p2: Revision, | |
481 | node: Node, |
|
473 | node: Node, | |
482 | } |
|
474 | } | |
483 |
|
475 | |||
484 | #[cfg(test)] |
|
476 | #[cfg(test)] | |
485 | impl IndexEntryBuilder { |
|
477 | impl IndexEntryBuilder { | |
486 | #[allow(clippy::new_without_default)] |
|
478 | #[allow(clippy::new_without_default)] | |
487 | pub fn new() -> Self { |
|
479 | pub fn new() -> Self { | |
488 | Self { |
|
480 | Self { | |
489 | is_first: false, |
|
481 | is_first: false, | |
490 | is_inline: false, |
|
482 | is_inline: false, | |
491 | is_general_delta: true, |
|
483 | is_general_delta: true, | |
492 | version: 1, |
|
484 | version: 1, | |
493 | offset: 0, |
|
485 | offset: 0, | |
494 | compressed_len: 0, |
|
486 | compressed_len: 0, | |
495 | uncompressed_len: 0, |
|
487 | uncompressed_len: 0, | |
496 | base_revision_or_base_of_delta_chain: Revision(0), |
|
488 | base_revision_or_base_of_delta_chain: Revision(0), | |
497 | link_revision: Revision(0), |
|
489 | link_revision: Revision(0), | |
498 | p1: NULL_REVISION, |
|
490 | p1: NULL_REVISION, | |
499 | p2: NULL_REVISION, |
|
491 | p2: NULL_REVISION, | |
500 | node: NULL_NODE, |
|
492 | node: NULL_NODE, | |
501 | } |
|
493 | } | |
502 | } |
|
494 | } | |
503 |
|
495 | |||
504 | pub fn is_first(&mut self, value: bool) -> &mut Self { |
|
496 | pub fn is_first(&mut self, value: bool) -> &mut Self { | |
505 | self.is_first = value; |
|
497 | self.is_first = value; | |
506 | self |
|
498 | self | |
507 | } |
|
499 | } | |
508 |
|
500 | |||
509 | pub fn with_inline(&mut self, value: bool) -> &mut Self { |
|
501 | pub fn with_inline(&mut self, value: bool) -> &mut Self { | |
510 | self.is_inline = value; |
|
502 | self.is_inline = value; | |
511 | self |
|
503 | self | |
512 | } |
|
504 | } | |
513 |
|
505 | |||
514 | pub fn with_general_delta(&mut self, value: bool) -> &mut Self { |
|
506 | pub fn with_general_delta(&mut self, value: bool) -> &mut Self { | |
515 | self.is_general_delta = value; |
|
507 | self.is_general_delta = value; | |
516 | self |
|
508 | self | |
517 | } |
|
509 | } | |
518 |
|
510 | |||
519 | pub fn with_version(&mut self, value: u16) -> &mut Self { |
|
511 | pub fn with_version(&mut self, value: u16) -> &mut Self { | |
520 | self.version = value; |
|
512 | self.version = value; | |
521 | self |
|
513 | self | |
522 | } |
|
514 | } | |
523 |
|
515 | |||
524 | pub fn with_offset(&mut self, value: usize) -> &mut Self { |
|
516 | pub fn with_offset(&mut self, value: usize) -> &mut Self { | |
525 | self.offset = value; |
|
517 | self.offset = value; | |
526 | self |
|
518 | self | |
527 | } |
|
519 | } | |
528 |
|
520 | |||
529 | pub fn with_compressed_len(&mut self, value: usize) -> &mut Self { |
|
521 | pub fn with_compressed_len(&mut self, value: usize) -> &mut Self { | |
530 | self.compressed_len = value; |
|
522 | self.compressed_len = value; | |
531 | self |
|
523 | self | |
532 | } |
|
524 | } | |
533 |
|
525 | |||
534 | pub fn with_uncompressed_len(&mut self, value: usize) -> &mut Self { |
|
526 | pub fn with_uncompressed_len(&mut self, value: usize) -> &mut Self { | |
535 | self.uncompressed_len = value; |
|
527 | self.uncompressed_len = value; | |
536 | self |
|
528 | self | |
537 | } |
|
529 | } | |
538 |
|
530 | |||
539 | pub fn with_base_revision_or_base_of_delta_chain( |
|
531 | pub fn with_base_revision_or_base_of_delta_chain( | |
540 | &mut self, |
|
532 | &mut self, | |
541 | value: Revision, |
|
533 | value: Revision, | |
542 | ) -> &mut Self { |
|
534 | ) -> &mut Self { | |
543 | self.base_revision_or_base_of_delta_chain = value; |
|
535 | self.base_revision_or_base_of_delta_chain = value; | |
544 | self |
|
536 | self | |
545 | } |
|
537 | } | |
546 |
|
538 | |||
547 | pub fn with_link_revision(&mut self, value: Revision) -> &mut Self { |
|
539 | pub fn with_link_revision(&mut self, value: Revision) -> &mut Self { | |
548 | self.link_revision = value; |
|
540 | self.link_revision = value; | |
549 | self |
|
541 | self | |
550 | } |
|
542 | } | |
551 |
|
543 | |||
552 | pub fn with_p1(&mut self, value: Revision) -> &mut Self { |
|
544 | pub fn with_p1(&mut self, value: Revision) -> &mut Self { | |
553 | self.p1 = value; |
|
545 | self.p1 = value; | |
554 | self |
|
546 | self | |
555 | } |
|
547 | } | |
556 |
|
548 | |||
557 | pub fn with_p2(&mut self, value: Revision) -> &mut Self { |
|
549 | pub fn with_p2(&mut self, value: Revision) -> &mut Self { | |
558 | self.p2 = value; |
|
550 | self.p2 = value; | |
559 | self |
|
551 | self | |
560 | } |
|
552 | } | |
561 |
|
553 | |||
562 | pub fn with_node(&mut self, value: Node) -> &mut Self { |
|
554 | pub fn with_node(&mut self, value: Node) -> &mut Self { | |
563 | self.node = value; |
|
555 | self.node = value; | |
564 | self |
|
556 | self | |
565 | } |
|
557 | } | |
566 |
|
558 | |||
567 | pub fn build(&self) -> Vec<u8> { |
|
559 | pub fn build(&self) -> Vec<u8> { | |
568 | let mut bytes = Vec::with_capacity(INDEX_ENTRY_SIZE); |
|
560 | let mut bytes = Vec::with_capacity(INDEX_ENTRY_SIZE); | |
569 | if self.is_first { |
|
561 | if self.is_first { | |
570 | bytes.extend(&match (self.is_general_delta, self.is_inline) { |
|
562 | bytes.extend(&match (self.is_general_delta, self.is_inline) { | |
571 | (false, false) => [0u8, 0], |
|
563 | (false, false) => [0u8, 0], | |
572 | (false, true) => [0u8, 1], |
|
564 | (false, true) => [0u8, 1], | |
573 | (true, false) => [0u8, 2], |
|
565 | (true, false) => [0u8, 2], | |
574 | (true, true) => [0u8, 3], |
|
566 | (true, true) => [0u8, 3], | |
575 | }); |
|
567 | }); | |
576 | bytes.extend(&self.version.to_be_bytes()); |
|
568 | bytes.extend(&self.version.to_be_bytes()); | |
577 | // Remaining offset bytes. |
|
569 | // Remaining offset bytes. | |
578 | bytes.extend(&[0u8; 2]); |
|
570 | bytes.extend(&[0u8; 2]); | |
579 | } else { |
|
571 | } else { | |
580 | // Offset stored on 48 bits (6 bytes) |
|
572 | // Offset stored on 48 bits (6 bytes) | |
581 | bytes.extend(&(self.offset as u64).to_be_bytes()[2..]); |
|
573 | bytes.extend(&(self.offset as u64).to_be_bytes()[2..]); | |
582 | } |
|
574 | } | |
583 | bytes.extend(&[0u8; 2]); // Revision flags. |
|
575 | bytes.extend(&[0u8; 2]); // Revision flags. | |
584 | bytes.extend(&(self.compressed_len as u32).to_be_bytes()); |
|
576 | bytes.extend(&(self.compressed_len as u32).to_be_bytes()); | |
585 | bytes.extend(&(self.uncompressed_len as u32).to_be_bytes()); |
|
577 | bytes.extend(&(self.uncompressed_len as u32).to_be_bytes()); | |
586 | bytes.extend( |
|
578 | bytes.extend( | |
587 | &self.base_revision_or_base_of_delta_chain.0.to_be_bytes(), |
|
579 | &self.base_revision_or_base_of_delta_chain.0.to_be_bytes(), | |
588 | ); |
|
580 | ); | |
589 | bytes.extend(&self.link_revision.0.to_be_bytes()); |
|
581 | bytes.extend(&self.link_revision.0.to_be_bytes()); | |
590 | bytes.extend(&self.p1.0.to_be_bytes()); |
|
582 | bytes.extend(&self.p1.0.to_be_bytes()); | |
591 | bytes.extend(&self.p2.0.to_be_bytes()); |
|
583 | bytes.extend(&self.p2.0.to_be_bytes()); | |
592 | bytes.extend(self.node.as_bytes()); |
|
584 | bytes.extend(self.node.as_bytes()); | |
593 | bytes.extend(vec![0u8; 12]); |
|
585 | bytes.extend(vec![0u8; 12]); | |
594 | bytes |
|
586 | bytes | |
595 | } |
|
587 | } | |
596 | } |
|
588 | } | |
597 |
|
589 | |||
598 | pub fn is_inline(index_bytes: &[u8]) -> bool { |
|
590 | pub fn is_inline(index_bytes: &[u8]) -> bool { | |
599 | IndexHeader::parse(index_bytes) |
|
591 | IndexHeader::parse(index_bytes) | |
600 | .expect("too short") |
|
592 | .expect("too short") | |
|
593 | .unwrap() | |||
601 | .format_flags() |
|
594 | .format_flags() | |
602 | .is_inline() |
|
595 | .is_inline() | |
603 | } |
|
596 | } | |
604 |
|
597 | |||
605 | pub fn uses_generaldelta(index_bytes: &[u8]) -> bool { |
|
598 | pub fn uses_generaldelta(index_bytes: &[u8]) -> bool { | |
606 | IndexHeader::parse(index_bytes) |
|
599 | IndexHeader::parse(index_bytes) | |
607 | .expect("too short") |
|
600 | .expect("too short") | |
|
601 | .unwrap() | |||
608 | .format_flags() |
|
602 | .format_flags() | |
609 | .uses_generaldelta() |
|
603 | .uses_generaldelta() | |
610 | } |
|
604 | } | |
611 |
|
605 | |||
612 | pub fn get_version(index_bytes: &[u8]) -> u16 { |
|
606 | pub fn get_version(index_bytes: &[u8]) -> u16 { | |
613 | IndexHeader::parse(index_bytes) |
|
607 | IndexHeader::parse(index_bytes) | |
614 | .expect("too short") |
|
608 | .expect("too short") | |
|
609 | .unwrap() | |||
615 | .format_version() |
|
610 | .format_version() | |
616 | } |
|
611 | } | |
617 |
|
612 | |||
618 | #[test] |
|
613 | #[test] | |
619 | fn flags_when_no_inline_flag_test() { |
|
614 | fn flags_when_no_inline_flag_test() { | |
620 | let bytes = IndexEntryBuilder::new() |
|
615 | let bytes = IndexEntryBuilder::new() | |
621 | .is_first(true) |
|
616 | .is_first(true) | |
622 | .with_general_delta(false) |
|
617 | .with_general_delta(false) | |
623 | .with_inline(false) |
|
618 | .with_inline(false) | |
624 | .build(); |
|
619 | .build(); | |
625 |
|
620 | |||
626 | assert!(!is_inline(&bytes)); |
|
621 | assert!(!is_inline(&bytes)); | |
627 | assert!(!uses_generaldelta(&bytes)); |
|
622 | assert!(!uses_generaldelta(&bytes)); | |
628 | } |
|
623 | } | |
629 |
|
624 | |||
630 | #[test] |
|
625 | #[test] | |
631 | fn flags_when_inline_flag_test() { |
|
626 | fn flags_when_inline_flag_test() { | |
632 | let bytes = IndexEntryBuilder::new() |
|
627 | let bytes = IndexEntryBuilder::new() | |
633 | .is_first(true) |
|
628 | .is_first(true) | |
634 | .with_general_delta(false) |
|
629 | .with_general_delta(false) | |
635 | .with_inline(true) |
|
630 | .with_inline(true) | |
636 | .build(); |
|
631 | .build(); | |
637 |
|
632 | |||
638 | assert!(is_inline(&bytes)); |
|
633 | assert!(is_inline(&bytes)); | |
639 | assert!(!uses_generaldelta(&bytes)); |
|
634 | assert!(!uses_generaldelta(&bytes)); | |
640 | } |
|
635 | } | |
641 |
|
636 | |||
642 | #[test] |
|
637 | #[test] | |
643 | fn flags_when_inline_and_generaldelta_flags_test() { |
|
638 | fn flags_when_inline_and_generaldelta_flags_test() { | |
644 | let bytes = IndexEntryBuilder::new() |
|
639 | let bytes = IndexEntryBuilder::new() | |
645 | .is_first(true) |
|
640 | .is_first(true) | |
646 | .with_general_delta(true) |
|
641 | .with_general_delta(true) | |
647 | .with_inline(true) |
|
642 | .with_inline(true) | |
648 | .build(); |
|
643 | .build(); | |
649 |
|
644 | |||
650 | assert!(is_inline(&bytes)); |
|
645 | assert!(is_inline(&bytes)); | |
651 | assert!(uses_generaldelta(&bytes)); |
|
646 | assert!(uses_generaldelta(&bytes)); | |
652 | } |
|
647 | } | |
653 |
|
648 | |||
654 | #[test] |
|
649 | #[test] | |
655 | fn test_offset() { |
|
650 | fn test_offset() { | |
656 | let bytes = IndexEntryBuilder::new().with_offset(1).build(); |
|
651 | let bytes = IndexEntryBuilder::new().with_offset(1).build(); | |
657 | let entry = IndexEntry { |
|
652 | let entry = IndexEntry { | |
658 | bytes: &bytes, |
|
653 | bytes: &bytes, | |
659 | offset_override: None, |
|
654 | offset_override: None, | |
660 | }; |
|
655 | }; | |
661 |
|
656 | |||
662 | assert_eq!(entry.offset(), 1) |
|
657 | assert_eq!(entry.offset(), 1) | |
663 | } |
|
658 | } | |
664 |
|
659 | |||
665 | #[test] |
|
660 | #[test] | |
666 | fn test_with_overridden_offset() { |
|
661 | fn test_with_overridden_offset() { | |
667 | let bytes = IndexEntryBuilder::new().with_offset(1).build(); |
|
662 | let bytes = IndexEntryBuilder::new().with_offset(1).build(); | |
668 | let entry = IndexEntry { |
|
663 | let entry = IndexEntry { | |
669 | bytes: &bytes, |
|
664 | bytes: &bytes, | |
670 | offset_override: Some(2), |
|
665 | offset_override: Some(2), | |
671 | }; |
|
666 | }; | |
672 |
|
667 | |||
673 | assert_eq!(entry.offset(), 2) |
|
668 | assert_eq!(entry.offset(), 2) | |
674 | } |
|
669 | } | |
675 |
|
670 | |||
676 | #[test] |
|
671 | #[test] | |
677 | fn test_compressed_len() { |
|
672 | fn test_compressed_len() { | |
678 | let bytes = IndexEntryBuilder::new().with_compressed_len(1).build(); |
|
673 | let bytes = IndexEntryBuilder::new().with_compressed_len(1).build(); | |
679 | let entry = IndexEntry { |
|
674 | let entry = IndexEntry { | |
680 | bytes: &bytes, |
|
675 | bytes: &bytes, | |
681 | offset_override: None, |
|
676 | offset_override: None, | |
682 | }; |
|
677 | }; | |
683 |
|
678 | |||
684 | assert_eq!(entry.compressed_len(), 1) |
|
679 | assert_eq!(entry.compressed_len(), 1) | |
685 | } |
|
680 | } | |
686 |
|
681 | |||
687 | #[test] |
|
682 | #[test] | |
688 | fn test_uncompressed_len() { |
|
683 | fn test_uncompressed_len() { | |
689 | let bytes = IndexEntryBuilder::new().with_uncompressed_len(1).build(); |
|
684 | let bytes = IndexEntryBuilder::new().with_uncompressed_len(1).build(); | |
690 | let entry = IndexEntry { |
|
685 | let entry = IndexEntry { | |
691 | bytes: &bytes, |
|
686 | bytes: &bytes, | |
692 | offset_override: None, |
|
687 | offset_override: None, | |
693 | }; |
|
688 | }; | |
694 |
|
689 | |||
695 | assert_eq!(entry.uncompressed_len(), 1) |
|
690 | assert_eq!(entry.uncompressed_len(), 1) | |
696 | } |
|
691 | } | |
697 |
|
692 | |||
698 | #[test] |
|
693 | #[test] | |
699 | fn test_base_revision_or_base_of_delta_chain() { |
|
694 | fn test_base_revision_or_base_of_delta_chain() { | |
700 | let bytes = IndexEntryBuilder::new() |
|
695 | let bytes = IndexEntryBuilder::new() | |
701 | .with_base_revision_or_base_of_delta_chain(Revision(1)) |
|
696 | .with_base_revision_or_base_of_delta_chain(Revision(1)) | |
702 | .build(); |
|
697 | .build(); | |
703 | let entry = IndexEntry { |
|
698 | let entry = IndexEntry { | |
704 | bytes: &bytes, |
|
699 | bytes: &bytes, | |
705 | offset_override: None, |
|
700 | offset_override: None, | |
706 | }; |
|
701 | }; | |
707 |
|
702 | |||
708 | assert_eq!(entry.base_revision_or_base_of_delta_chain(), 1.into()) |
|
703 | assert_eq!(entry.base_revision_or_base_of_delta_chain(), 1.into()) | |
709 | } |
|
704 | } | |
710 |
|
705 | |||
711 | #[test] |
|
706 | #[test] | |
712 | fn link_revision_test() { |
|
707 | fn link_revision_test() { | |
713 | let bytes = IndexEntryBuilder::new() |
|
708 | let bytes = IndexEntryBuilder::new() | |
714 | .with_link_revision(Revision(123)) |
|
709 | .with_link_revision(Revision(123)) | |
715 | .build(); |
|
710 | .build(); | |
716 |
|
711 | |||
717 | let entry = IndexEntry { |
|
712 | let entry = IndexEntry { | |
718 | bytes: &bytes, |
|
713 | bytes: &bytes, | |
719 | offset_override: None, |
|
714 | offset_override: None, | |
720 | }; |
|
715 | }; | |
721 |
|
716 | |||
722 | assert_eq!(entry.link_revision(), 123.into()); |
|
717 | assert_eq!(entry.link_revision(), 123.into()); | |
723 | } |
|
718 | } | |
724 |
|
719 | |||
725 | #[test] |
|
720 | #[test] | |
726 | fn p1_test() { |
|
721 | fn p1_test() { | |
727 | let bytes = IndexEntryBuilder::new().with_p1(Revision(123)).build(); |
|
722 | let bytes = IndexEntryBuilder::new().with_p1(Revision(123)).build(); | |
728 |
|
723 | |||
729 | let entry = IndexEntry { |
|
724 | let entry = IndexEntry { | |
730 | bytes: &bytes, |
|
725 | bytes: &bytes, | |
731 | offset_override: None, |
|
726 | offset_override: None, | |
732 | }; |
|
727 | }; | |
733 |
|
728 | |||
734 | assert_eq!(entry.p1(), 123.into()); |
|
729 | assert_eq!(entry.p1(), 123.into()); | |
735 | } |
|
730 | } | |
736 |
|
731 | |||
737 | #[test] |
|
732 | #[test] | |
738 | fn p2_test() { |
|
733 | fn p2_test() { | |
739 | let bytes = IndexEntryBuilder::new().with_p2(Revision(123)).build(); |
|
734 | let bytes = IndexEntryBuilder::new().with_p2(Revision(123)).build(); | |
740 |
|
735 | |||
741 | let entry = IndexEntry { |
|
736 | let entry = IndexEntry { | |
742 | bytes: &bytes, |
|
737 | bytes: &bytes, | |
743 | offset_override: None, |
|
738 | offset_override: None, | |
744 | }; |
|
739 | }; | |
745 |
|
740 | |||
746 | assert_eq!(entry.p2(), 123.into()); |
|
741 | assert_eq!(entry.p2(), 123.into()); | |
747 | } |
|
742 | } | |
748 |
|
743 | |||
749 | #[test] |
|
744 | #[test] | |
750 | fn node_test() { |
|
745 | fn node_test() { | |
751 | let node = Node::from_hex("0123456789012345678901234567890123456789") |
|
746 | let node = Node::from_hex("0123456789012345678901234567890123456789") | |
752 | .unwrap(); |
|
747 | .unwrap(); | |
753 | let bytes = IndexEntryBuilder::new().with_node(node).build(); |
|
748 | let bytes = IndexEntryBuilder::new().with_node(node).build(); | |
754 |
|
749 | |||
755 | let entry = IndexEntry { |
|
750 | let entry = IndexEntry { | |
756 | bytes: &bytes, |
|
751 | bytes: &bytes, | |
757 | offset_override: None, |
|
752 | offset_override: None, | |
758 | }; |
|
753 | }; | |
759 |
|
754 | |||
760 | assert_eq!(*entry.hash(), node); |
|
755 | assert_eq!(*entry.hash(), node); | |
761 | } |
|
756 | } | |
762 |
|
757 | |||
763 | #[test] |
|
758 | #[test] | |
764 | fn version_test() { |
|
759 | fn version_test() { | |
765 | let bytes = IndexEntryBuilder::new() |
|
760 | let bytes = IndexEntryBuilder::new() | |
766 | .is_first(true) |
|
761 | .is_first(true) | |
767 | .with_version(2) |
|
762 | .with_version(2) | |
768 | .build(); |
|
763 | .build(); | |
769 |
|
764 | |||
770 | assert_eq!(get_version(&bytes), 2) |
|
765 | assert_eq!(get_version(&bytes), 2) | |
771 | } |
|
766 | } | |
772 | } |
|
767 | } | |
773 |
|
768 | |||
774 | #[cfg(test)] |
|
769 | #[cfg(test)] | |
775 | pub use tests::IndexEntryBuilder; |
|
770 | pub use tests::IndexEntryBuilder; |
@@ -1,209 +1,213 b'' | |||||
1 | use crate::errors::HgError; |
|
1 | use crate::errors::HgError; | |
2 | use crate::revlog::{Node, NodePrefix}; |
|
2 | use crate::revlog::{Node, NodePrefix}; | |
3 | use crate::revlog::{Revlog, RevlogError}; |
|
3 | use crate::revlog::{Revlog, RevlogError}; | |
4 | use crate::utils::hg_path::HgPath; |
|
4 | use crate::utils::hg_path::HgPath; | |
5 | use crate::utils::SliceExt; |
|
5 | use crate::utils::SliceExt; | |
6 | use crate::vfs::Vfs; |
|
6 | use crate::vfs::Vfs; | |
7 | use crate::{Graph, GraphError, Revision, UncheckedRevision}; |
|
7 | use crate::{ | |
|
8 | Graph, GraphError, Revision, RevlogOpenOptions, UncheckedRevision, | |||
|
9 | }; | |||
8 |
|
10 | |||
9 | /// A specialized `Revlog` to work with `manifest` data format. |
|
11 | /// A specialized `Revlog` to work with `manifest` data format. | |
10 | pub struct Manifestlog { |
|
12 | pub struct Manifestlog { | |
11 | /// The generic `revlog` format. |
|
13 | /// The generic `revlog` format. | |
12 | revlog: Revlog, |
|
14 | pub(crate) revlog: Revlog, | |
13 | } |
|
15 | } | |
14 |
|
16 | |||
15 | impl Graph for Manifestlog { |
|
17 | impl Graph for Manifestlog { | |
16 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { |
|
18 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { | |
17 | self.revlog.parents(rev) |
|
19 | self.revlog.parents(rev) | |
18 | } |
|
20 | } | |
19 | } |
|
21 | } | |
20 |
|
22 | |||
21 | impl Manifestlog { |
|
23 | impl Manifestlog { | |
22 | /// Open the `manifest` of a repository given by its root. |
|
24 | /// Open the `manifest` of a repository given by its root. | |
23 | pub fn open(store_vfs: &Vfs, use_nodemap: bool) -> Result<Self, HgError> { |
|
25 | pub fn open( | |
24 | let revlog = |
|
26 | store_vfs: &Vfs, | |
25 | Revlog::open(store_vfs, "00manifest.i", None, use_nodemap)?; |
|
27 | options: RevlogOpenOptions, | |
|
28 | ) -> Result<Self, HgError> { | |||
|
29 | let revlog = Revlog::open(store_vfs, "00manifest.i", None, options)?; | |||
26 | Ok(Self { revlog }) |
|
30 | Ok(Self { revlog }) | |
27 | } |
|
31 | } | |
28 |
|
32 | |||
29 | /// Return the `Manifest` for the given node ID. |
|
33 | /// Return the `Manifest` for the given node ID. | |
30 | /// |
|
34 | /// | |
31 | /// Note: this is a node ID in the manifestlog, typically found through |
|
35 | /// Note: this is a node ID in the manifestlog, typically found through | |
32 | /// `ChangelogEntry::manifest_node`. It is *not* the node ID of any |
|
36 | /// `ChangelogEntry::manifest_node`. It is *not* the node ID of any | |
33 | /// changeset. |
|
37 | /// changeset. | |
34 | /// |
|
38 | /// | |
35 | /// See also `Repo::manifest_for_node` |
|
39 | /// See also `Repo::manifest_for_node` | |
36 | pub fn data_for_node( |
|
40 | pub fn data_for_node( | |
37 | &self, |
|
41 | &self, | |
38 | node: NodePrefix, |
|
42 | node: NodePrefix, | |
39 | ) -> Result<Manifest, RevlogError> { |
|
43 | ) -> Result<Manifest, RevlogError> { | |
40 | let rev = self.revlog.rev_from_node(node)?; |
|
44 | let rev = self.revlog.rev_from_node(node)?; | |
41 | self.data_for_checked_rev(rev) |
|
45 | self.data_for_checked_rev(rev) | |
42 | } |
|
46 | } | |
43 |
|
47 | |||
44 | /// Return the `Manifest` of a given revision number. |
|
48 | /// Return the `Manifest` of a given revision number. | |
45 | /// |
|
49 | /// | |
46 | /// Note: this is a revision number in the manifestlog, *not* of any |
|
50 | /// Note: this is a revision number in the manifestlog, *not* of any | |
47 | /// changeset. |
|
51 | /// changeset. | |
48 | /// |
|
52 | /// | |
49 | /// See also `Repo::manifest_for_rev` |
|
53 | /// See also `Repo::manifest_for_rev` | |
50 | pub fn data_for_rev( |
|
54 | pub fn data_for_rev( | |
51 | &self, |
|
55 | &self, | |
52 | rev: UncheckedRevision, |
|
56 | rev: UncheckedRevision, | |
53 | ) -> Result<Manifest, RevlogError> { |
|
57 | ) -> Result<Manifest, RevlogError> { | |
54 | let bytes = self.revlog.get_rev_data(rev)?.into_owned(); |
|
58 | let bytes = self.revlog.get_rev_data(rev)?.into_owned(); | |
55 | Ok(Manifest { bytes }) |
|
59 | Ok(Manifest { bytes }) | |
56 | } |
|
60 | } | |
57 |
|
61 | |||
58 | pub fn data_for_checked_rev( |
|
62 | pub fn data_for_checked_rev( | |
59 | &self, |
|
63 | &self, | |
60 | rev: Revision, |
|
64 | rev: Revision, | |
61 | ) -> Result<Manifest, RevlogError> { |
|
65 | ) -> Result<Manifest, RevlogError> { | |
62 | let bytes = |
|
66 | let bytes = | |
63 | self.revlog.get_rev_data_for_checked_rev(rev)?.into_owned(); |
|
67 | self.revlog.get_rev_data_for_checked_rev(rev)?.into_owned(); | |
64 | Ok(Manifest { bytes }) |
|
68 | Ok(Manifest { bytes }) | |
65 | } |
|
69 | } | |
66 | } |
|
70 | } | |
67 |
|
71 | |||
68 | /// `Manifestlog` entry which knows how to interpret the `manifest` data bytes. |
|
72 | /// `Manifestlog` entry which knows how to interpret the `manifest` data bytes. | |
69 | #[derive(Debug)] |
|
73 | #[derive(Debug)] | |
70 | pub struct Manifest { |
|
74 | pub struct Manifest { | |
71 | /// Format for a manifest: flat sequence of variable-size entries, |
|
75 | /// Format for a manifest: flat sequence of variable-size entries, | |
72 | /// sorted by path, each as: |
|
76 | /// sorted by path, each as: | |
73 | /// |
|
77 | /// | |
74 | /// ```text |
|
78 | /// ```text | |
75 | /// <path> \0 <hex_node_id> <flags> \n |
|
79 | /// <path> \0 <hex_node_id> <flags> \n | |
76 | /// ``` |
|
80 | /// ``` | |
77 | /// |
|
81 | /// | |
78 | /// The last entry is also terminated by a newline character. |
|
82 | /// The last entry is also terminated by a newline character. | |
79 | /// Flags is one of `b""` (the empty string), `b"x"`, `b"l"`, or `b"t"`. |
|
83 | /// Flags is one of `b""` (the empty string), `b"x"`, `b"l"`, or `b"t"`. | |
80 | bytes: Vec<u8>, |
|
84 | bytes: Vec<u8>, | |
81 | } |
|
85 | } | |
82 |
|
86 | |||
83 | impl Manifest { |
|
87 | impl Manifest { | |
84 | pub fn iter( |
|
88 | pub fn iter( | |
85 | &self, |
|
89 | &self, | |
86 | ) -> impl Iterator<Item = Result<ManifestEntry, HgError>> { |
|
90 | ) -> impl Iterator<Item = Result<ManifestEntry, HgError>> { | |
87 | self.bytes |
|
91 | self.bytes | |
88 | .split(|b| b == &b'\n') |
|
92 | .split(|b| b == &b'\n') | |
89 | .filter(|line| !line.is_empty()) |
|
93 | .filter(|line| !line.is_empty()) | |
90 | .map(ManifestEntry::from_raw) |
|
94 | .map(ManifestEntry::from_raw) | |
91 | } |
|
95 | } | |
92 |
|
96 | |||
93 | /// If the given path is in this manifest, return its filelog node ID |
|
97 | /// If the given path is in this manifest, return its filelog node ID | |
94 | pub fn find_by_path( |
|
98 | pub fn find_by_path( | |
95 | &self, |
|
99 | &self, | |
96 | path: &HgPath, |
|
100 | path: &HgPath, | |
97 | ) -> Result<Option<ManifestEntry>, HgError> { |
|
101 | ) -> Result<Option<ManifestEntry>, HgError> { | |
98 | use std::cmp::Ordering::*; |
|
102 | use std::cmp::Ordering::*; | |
99 | let path = path.as_bytes(); |
|
103 | let path = path.as_bytes(); | |
100 | // Both boundaries of this `&[u8]` slice are always at the boundary of |
|
104 | // Both boundaries of this `&[u8]` slice are always at the boundary of | |
101 | // an entry |
|
105 | // an entry | |
102 | let mut bytes = &*self.bytes; |
|
106 | let mut bytes = &*self.bytes; | |
103 |
|
107 | |||
104 | // Binary search algorithm derived from `[T]::binary_search_by` |
|
108 | // Binary search algorithm derived from `[T]::binary_search_by` | |
105 | // <https://github.com/rust-lang/rust/blob/1.57.0/library/core/src/slice/mod.rs#L2221> |
|
109 | // <https://github.com/rust-lang/rust/blob/1.57.0/library/core/src/slice/mod.rs#L2221> | |
106 | // except we don’t have a slice of entries. Instead we jump to the |
|
110 | // except we don’t have a slice of entries. Instead we jump to the | |
107 | // middle of the byte slice and look around for entry delimiters |
|
111 | // middle of the byte slice and look around for entry delimiters | |
108 | // (newlines). |
|
112 | // (newlines). | |
109 | while let Some(entry_range) = Self::find_entry_near_middle_of(bytes)? { |
|
113 | while let Some(entry_range) = Self::find_entry_near_middle_of(bytes)? { | |
110 | let (entry_path, rest) = |
|
114 | let (entry_path, rest) = | |
111 | ManifestEntry::split_path(&bytes[entry_range.clone()])?; |
|
115 | ManifestEntry::split_path(&bytes[entry_range.clone()])?; | |
112 | let cmp = entry_path.cmp(path); |
|
116 | let cmp = entry_path.cmp(path); | |
113 | if cmp == Less { |
|
117 | if cmp == Less { | |
114 | let after_newline = entry_range.end + 1; |
|
118 | let after_newline = entry_range.end + 1; | |
115 | bytes = &bytes[after_newline..]; |
|
119 | bytes = &bytes[after_newline..]; | |
116 | } else if cmp == Greater { |
|
120 | } else if cmp == Greater { | |
117 | bytes = &bytes[..entry_range.start]; |
|
121 | bytes = &bytes[..entry_range.start]; | |
118 | } else { |
|
122 | } else { | |
119 | return Ok(Some(ManifestEntry::from_path_and_rest( |
|
123 | return Ok(Some(ManifestEntry::from_path_and_rest( | |
120 | entry_path, rest, |
|
124 | entry_path, rest, | |
121 | ))); |
|
125 | ))); | |
122 | } |
|
126 | } | |
123 | } |
|
127 | } | |
124 | Ok(None) |
|
128 | Ok(None) | |
125 | } |
|
129 | } | |
126 |
|
130 | |||
127 | /// If there is at least one, return the byte range of an entry *excluding* |
|
131 | /// If there is at least one, return the byte range of an entry *excluding* | |
128 | /// the final newline. |
|
132 | /// the final newline. | |
129 | fn find_entry_near_middle_of( |
|
133 | fn find_entry_near_middle_of( | |
130 | bytes: &[u8], |
|
134 | bytes: &[u8], | |
131 | ) -> Result<Option<std::ops::Range<usize>>, HgError> { |
|
135 | ) -> Result<Option<std::ops::Range<usize>>, HgError> { | |
132 | let len = bytes.len(); |
|
136 | let len = bytes.len(); | |
133 | if len > 0 { |
|
137 | if len > 0 { | |
134 | let middle = bytes.len() / 2; |
|
138 | let middle = bytes.len() / 2; | |
135 | // Integer division rounds down, so `middle < len`. |
|
139 | // Integer division rounds down, so `middle < len`. | |
136 | let (before, after) = bytes.split_at(middle); |
|
140 | let (before, after) = bytes.split_at(middle); | |
137 | let is_newline = |&byte: &u8| byte == b'\n'; |
|
141 | let is_newline = |&byte: &u8| byte == b'\n'; | |
138 | let entry_start = match before.iter().rposition(is_newline) { |
|
142 | let entry_start = match before.iter().rposition(is_newline) { | |
139 | Some(i) => i + 1, |
|
143 | Some(i) => i + 1, | |
140 | None => 0, // We choose the first entry in `bytes` |
|
144 | None => 0, // We choose the first entry in `bytes` | |
141 | }; |
|
145 | }; | |
142 | let entry_end = match after.iter().position(is_newline) { |
|
146 | let entry_end = match after.iter().position(is_newline) { | |
143 | Some(i) => { |
|
147 | Some(i) => { | |
144 | // No `+ 1` here to exclude this newline from the range |
|
148 | // No `+ 1` here to exclude this newline from the range | |
145 | middle + i |
|
149 | middle + i | |
146 | } |
|
150 | } | |
147 | None => { |
|
151 | None => { | |
148 | // In a well-formed manifest: |
|
152 | // In a well-formed manifest: | |
149 | // |
|
153 | // | |
150 | // * Since `len > 0`, `bytes` contains at least one entry |
|
154 | // * Since `len > 0`, `bytes` contains at least one entry | |
151 | // * Every entry ends with a newline |
|
155 | // * Every entry ends with a newline | |
152 | // * Since `middle < len`, `after` contains at least the |
|
156 | // * Since `middle < len`, `after` contains at least the | |
153 | // newline at the end of the last entry of `bytes`. |
|
157 | // newline at the end of the last entry of `bytes`. | |
154 | // |
|
158 | // | |
155 | // We didn’t find a newline, so this manifest is not |
|
159 | // We didn’t find a newline, so this manifest is not | |
156 | // well-formed. |
|
160 | // well-formed. | |
157 | return Err(HgError::corrupted( |
|
161 | return Err(HgError::corrupted( | |
158 | "manifest entry without \\n delimiter", |
|
162 | "manifest entry without \\n delimiter", | |
159 | )); |
|
163 | )); | |
160 | } |
|
164 | } | |
161 | }; |
|
165 | }; | |
162 | Ok(Some(entry_start..entry_end)) |
|
166 | Ok(Some(entry_start..entry_end)) | |
163 | } else { |
|
167 | } else { | |
164 | // len == 0 |
|
168 | // len == 0 | |
165 | Ok(None) |
|
169 | Ok(None) | |
166 | } |
|
170 | } | |
167 | } |
|
171 | } | |
168 | } |
|
172 | } | |
169 |
|
173 | |||
170 | /// `Manifestlog` entry which knows how to interpret the `manifest` data bytes. |
|
174 | /// `Manifestlog` entry which knows how to interpret the `manifest` data bytes. | |
171 | #[derive(Debug)] |
|
175 | #[derive(Debug)] | |
172 | pub struct ManifestEntry<'manifest> { |
|
176 | pub struct ManifestEntry<'manifest> { | |
173 | pub path: &'manifest HgPath, |
|
177 | pub path: &'manifest HgPath, | |
174 | pub hex_node_id: &'manifest [u8], |
|
178 | pub hex_node_id: &'manifest [u8], | |
175 |
|
179 | |||
176 | /// `Some` values are b'x', b'l', or 't' |
|
180 | /// `Some` values are b'x', b'l', or 't' | |
177 | pub flags: Option<u8>, |
|
181 | pub flags: Option<u8>, | |
178 | } |
|
182 | } | |
179 |
|
183 | |||
180 | impl<'a> ManifestEntry<'a> { |
|
184 | impl<'a> ManifestEntry<'a> { | |
181 | fn split_path(bytes: &[u8]) -> Result<(&[u8], &[u8]), HgError> { |
|
185 | fn split_path(bytes: &[u8]) -> Result<(&[u8], &[u8]), HgError> { | |
182 | bytes.split_2(b'\0').ok_or_else(|| { |
|
186 | bytes.split_2(b'\0').ok_or_else(|| { | |
183 | HgError::corrupted("manifest entry without \\0 delimiter") |
|
187 | HgError::corrupted("manifest entry without \\0 delimiter") | |
184 | }) |
|
188 | }) | |
185 | } |
|
189 | } | |
186 |
|
190 | |||
187 | fn from_path_and_rest(path: &'a [u8], rest: &'a [u8]) -> Self { |
|
191 | fn from_path_and_rest(path: &'a [u8], rest: &'a [u8]) -> Self { | |
188 | let (hex_node_id, flags) = match rest.split_last() { |
|
192 | let (hex_node_id, flags) = match rest.split_last() { | |
189 | Some((&b'x', rest)) => (rest, Some(b'x')), |
|
193 | Some((&b'x', rest)) => (rest, Some(b'x')), | |
190 | Some((&b'l', rest)) => (rest, Some(b'l')), |
|
194 | Some((&b'l', rest)) => (rest, Some(b'l')), | |
191 | Some((&b't', rest)) => (rest, Some(b't')), |
|
195 | Some((&b't', rest)) => (rest, Some(b't')), | |
192 | _ => (rest, None), |
|
196 | _ => (rest, None), | |
193 | }; |
|
197 | }; | |
194 | Self { |
|
198 | Self { | |
195 | path: HgPath::new(path), |
|
199 | path: HgPath::new(path), | |
196 | hex_node_id, |
|
200 | hex_node_id, | |
197 | flags, |
|
201 | flags, | |
198 | } |
|
202 | } | |
199 | } |
|
203 | } | |
200 |
|
204 | |||
201 | fn from_raw(bytes: &'a [u8]) -> Result<Self, HgError> { |
|
205 | fn from_raw(bytes: &'a [u8]) -> Result<Self, HgError> { | |
202 | let (path, rest) = Self::split_path(bytes)?; |
|
206 | let (path, rest) = Self::split_path(bytes)?; | |
203 | Ok(Self::from_path_and_rest(path, rest)) |
|
207 | Ok(Self::from_path_and_rest(path, rest)) | |
204 | } |
|
208 | } | |
205 |
|
209 | |||
206 | pub fn node_id(&self) -> Result<Node, HgError> { |
|
210 | pub fn node_id(&self) -> Result<Node, HgError> { | |
207 | Node::from_hex_for_repo(self.hex_node_id) |
|
211 | Node::from_hex_for_repo(self.hex_node_id) | |
208 | } |
|
212 | } | |
209 | } |
|
213 | } |
@@ -1,965 +1,1030 b'' | |||||
1 | // Copyright 2018-2023 Georges Racinet <georges.racinet@octobus.net> |
|
1 | // Copyright 2018-2023 Georges Racinet <georges.racinet@octobus.net> | |
2 | // and Mercurial contributors |
|
2 | // and Mercurial contributors | |
3 | // |
|
3 | // | |
4 | // This software may be used and distributed according to the terms of the |
|
4 | // This software may be used and distributed according to the terms of the | |
5 | // GNU General Public License version 2 or any later version. |
|
5 | // GNU General Public License version 2 or any later version. | |
6 | //! Mercurial concepts for handling revision history |
|
6 | //! Mercurial concepts for handling revision history | |
7 |
|
7 | |||
8 | pub mod node; |
|
8 | pub mod node; | |
9 | pub mod nodemap; |
|
9 | pub mod nodemap; | |
10 | mod nodemap_docket; |
|
10 | mod nodemap_docket; | |
11 | pub mod path_encode; |
|
11 | pub mod path_encode; | |
12 | pub use node::{FromHexError, Node, NodePrefix}; |
|
12 | pub use node::{FromHexError, Node, NodePrefix}; | |
13 | pub mod changelog; |
|
13 | pub mod changelog; | |
14 | pub mod filelog; |
|
14 | pub mod filelog; | |
15 | pub mod index; |
|
15 | pub mod index; | |
16 | pub mod manifest; |
|
16 | pub mod manifest; | |
17 | pub mod patch; |
|
17 | pub mod patch; | |
18 |
|
18 | |||
19 | use std::borrow::Cow; |
|
19 | use std::borrow::Cow; | |
20 | use std::io::Read; |
|
20 | use std::io::Read; | |
21 | use std::ops::Deref; |
|
21 | use std::ops::Deref; | |
22 | use std::path::Path; |
|
22 | use std::path::Path; | |
23 |
|
23 | |||
24 | use flate2::read::ZlibDecoder; |
|
24 | use flate2::read::ZlibDecoder; | |
25 | use sha1::{Digest, Sha1}; |
|
25 | use sha1::{Digest, Sha1}; | |
26 | use std::cell::RefCell; |
|
26 | use std::cell::RefCell; | |
27 | use zstd; |
|
27 | use zstd; | |
28 |
|
28 | |||
29 | use self::node::{NODE_BYTES_LENGTH, NULL_NODE}; |
|
29 | use self::node::{NODE_BYTES_LENGTH, NULL_NODE}; | |
30 | use self::nodemap_docket::NodeMapDocket; |
|
30 | use self::nodemap_docket::NodeMapDocket; | |
31 | use super::index::Index; |
|
31 | use super::index::Index; | |
32 | use super::nodemap::{NodeMap, NodeMapError}; |
|
32 | use super::nodemap::{NodeMap, NodeMapError}; | |
33 | use crate::errors::HgError; |
|
33 | use crate::errors::HgError; | |
34 | use crate::vfs::Vfs; |
|
34 | use crate::vfs::Vfs; | |
35 |
|
35 | |||
36 | /// As noted in revlog.c, revision numbers are actually encoded in |
|
36 | /// As noted in revlog.c, revision numbers are actually encoded in | |
37 | /// 4 bytes, and are liberally converted to ints, whence the i32 |
|
37 | /// 4 bytes, and are liberally converted to ints, whence the i32 | |
38 | pub type BaseRevision = i32; |
|
38 | pub type BaseRevision = i32; | |
39 |
|
39 | |||
40 | /// Mercurial revision numbers |
|
40 | /// Mercurial revision numbers | |
41 | /// In contrast to the more general [`UncheckedRevision`], these are "checked" |
|
41 | /// In contrast to the more general [`UncheckedRevision`], these are "checked" | |
42 | /// in the sense that they should only be used for revisions that are |
|
42 | /// in the sense that they should only be used for revisions that are | |
43 | /// valid for a given index (i.e. in bounds). |
|
43 | /// valid for a given index (i.e. in bounds). | |
44 | #[derive( |
|
44 | #[derive( | |
45 | Debug, |
|
45 | Debug, | |
46 | derive_more::Display, |
|
46 | derive_more::Display, | |
47 | Clone, |
|
47 | Clone, | |
48 | Copy, |
|
48 | Copy, | |
49 | Hash, |
|
49 | Hash, | |
50 | PartialEq, |
|
50 | PartialEq, | |
51 | Eq, |
|
51 | Eq, | |
52 | PartialOrd, |
|
52 | PartialOrd, | |
53 | Ord, |
|
53 | Ord, | |
54 | )] |
|
54 | )] | |
55 | pub struct Revision(pub BaseRevision); |
|
55 | pub struct Revision(pub BaseRevision); | |
56 |
|
56 | |||
57 | impl format_bytes::DisplayBytes for Revision { |
|
57 | impl format_bytes::DisplayBytes for Revision { | |
58 | fn display_bytes( |
|
58 | fn display_bytes( | |
59 | &self, |
|
59 | &self, | |
60 | output: &mut dyn std::io::Write, |
|
60 | output: &mut dyn std::io::Write, | |
61 | ) -> std::io::Result<()> { |
|
61 | ) -> std::io::Result<()> { | |
62 | self.0.display_bytes(output) |
|
62 | self.0.display_bytes(output) | |
63 | } |
|
63 | } | |
64 | } |
|
64 | } | |
65 |
|
65 | |||
66 | /// Unchecked Mercurial revision numbers. |
|
66 | /// Unchecked Mercurial revision numbers. | |
67 | /// |
|
67 | /// | |
68 | /// Values of this type have no guarantee of being a valid revision number |
|
68 | /// Values of this type have no guarantee of being a valid revision number | |
69 | /// in any context. Use method `check_revision` to get a valid revision within |
|
69 | /// in any context. Use method `check_revision` to get a valid revision within | |
70 | /// the appropriate index object. |
|
70 | /// the appropriate index object. | |
71 | #[derive( |
|
71 | #[derive( | |
72 | Debug, |
|
72 | Debug, | |
73 | derive_more::Display, |
|
73 | derive_more::Display, | |
74 | Clone, |
|
74 | Clone, | |
75 | Copy, |
|
75 | Copy, | |
76 | Hash, |
|
76 | Hash, | |
77 | PartialEq, |
|
77 | PartialEq, | |
78 | Eq, |
|
78 | Eq, | |
79 | PartialOrd, |
|
79 | PartialOrd, | |
80 | Ord, |
|
80 | Ord, | |
81 | )] |
|
81 | )] | |
82 | pub struct UncheckedRevision(pub BaseRevision); |
|
82 | pub struct UncheckedRevision(pub BaseRevision); | |
83 |
|
83 | |||
84 | impl format_bytes::DisplayBytes for UncheckedRevision { |
|
84 | impl format_bytes::DisplayBytes for UncheckedRevision { | |
85 | fn display_bytes( |
|
85 | fn display_bytes( | |
86 | &self, |
|
86 | &self, | |
87 | output: &mut dyn std::io::Write, |
|
87 | output: &mut dyn std::io::Write, | |
88 | ) -> std::io::Result<()> { |
|
88 | ) -> std::io::Result<()> { | |
89 | self.0.display_bytes(output) |
|
89 | self.0.display_bytes(output) | |
90 | } |
|
90 | } | |
91 | } |
|
91 | } | |
92 |
|
92 | |||
93 | impl From<Revision> for UncheckedRevision { |
|
93 | impl From<Revision> for UncheckedRevision { | |
94 | fn from(value: Revision) -> Self { |
|
94 | fn from(value: Revision) -> Self { | |
95 | Self(value.0) |
|
95 | Self(value.0) | |
96 | } |
|
96 | } | |
97 | } |
|
97 | } | |
98 |
|
98 | |||
99 | impl From<BaseRevision> for UncheckedRevision { |
|
99 | impl From<BaseRevision> for UncheckedRevision { | |
100 | fn from(value: BaseRevision) -> Self { |
|
100 | fn from(value: BaseRevision) -> Self { | |
101 | Self(value) |
|
101 | Self(value) | |
102 | } |
|
102 | } | |
103 | } |
|
103 | } | |
104 |
|
104 | |||
105 | /// Marker expressing the absence of a parent |
|
105 | /// Marker expressing the absence of a parent | |
106 | /// |
|
106 | /// | |
107 | /// Independently of the actual representation, `NULL_REVISION` is guaranteed |
|
107 | /// Independently of the actual representation, `NULL_REVISION` is guaranteed | |
108 | /// to be smaller than all existing revisions. |
|
108 | /// to be smaller than all existing revisions. | |
109 | pub const NULL_REVISION: Revision = Revision(-1); |
|
109 | pub const NULL_REVISION: Revision = Revision(-1); | |
110 |
|
110 | |||
111 | /// Same as `mercurial.node.wdirrev` |
|
111 | /// Same as `mercurial.node.wdirrev` | |
112 | /// |
|
112 | /// | |
113 | /// This is also equal to `i32::max_value()`, but it's better to spell |
|
113 | /// This is also equal to `i32::max_value()`, but it's better to spell | |
114 | /// it out explicitely, same as in `mercurial.node` |
|
114 | /// it out explicitely, same as in `mercurial.node` | |
115 | #[allow(clippy::unreadable_literal)] |
|
115 | #[allow(clippy::unreadable_literal)] | |
116 | pub const WORKING_DIRECTORY_REVISION: UncheckedRevision = |
|
116 | pub const WORKING_DIRECTORY_REVISION: UncheckedRevision = | |
117 | UncheckedRevision(0x7fffffff); |
|
117 | UncheckedRevision(0x7fffffff); | |
118 |
|
118 | |||
119 | pub const WORKING_DIRECTORY_HEX: &str = |
|
119 | pub const WORKING_DIRECTORY_HEX: &str = | |
120 | "ffffffffffffffffffffffffffffffffffffffff"; |
|
120 | "ffffffffffffffffffffffffffffffffffffffff"; | |
121 |
|
121 | |||
122 | /// The simplest expression of what we need of Mercurial DAGs. |
|
122 | /// The simplest expression of what we need of Mercurial DAGs. | |
123 | pub trait Graph { |
|
123 | pub trait Graph { | |
124 | /// Return the two parents of the given `Revision`. |
|
124 | /// Return the two parents of the given `Revision`. | |
125 | /// |
|
125 | /// | |
126 | /// Each of the parents can be independently `NULL_REVISION` |
|
126 | /// Each of the parents can be independently `NULL_REVISION` | |
127 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError>; |
|
127 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError>; | |
128 | } |
|
128 | } | |
129 |
|
129 | |||
130 | #[derive(Clone, Debug, PartialEq)] |
|
130 | #[derive(Clone, Debug, PartialEq)] | |
131 | pub enum GraphError { |
|
131 | pub enum GraphError { | |
132 | ParentOutOfRange(Revision), |
|
132 | ParentOutOfRange(Revision), | |
133 | } |
|
133 | } | |
134 |
|
134 | |||
135 | /// The Mercurial Revlog Index |
|
135 | /// The Mercurial Revlog Index | |
136 | /// |
|
136 | /// | |
137 | /// This is currently limited to the minimal interface that is needed for |
|
137 | /// This is currently limited to the minimal interface that is needed for | |
138 | /// the [`nodemap`](nodemap/index.html) module |
|
138 | /// the [`nodemap`](nodemap/index.html) module | |
139 | pub trait RevlogIndex { |
|
139 | pub trait RevlogIndex { | |
140 | /// Total number of Revisions referenced in this index |
|
140 | /// Total number of Revisions referenced in this index | |
141 | fn len(&self) -> usize; |
|
141 | fn len(&self) -> usize; | |
142 |
|
142 | |||
143 | fn is_empty(&self) -> bool { |
|
143 | fn is_empty(&self) -> bool { | |
144 | self.len() == 0 |
|
144 | self.len() == 0 | |
145 | } |
|
145 | } | |
146 |
|
146 | |||
147 | /// Return a reference to the Node or `None` for `NULL_REVISION` |
|
147 | /// Return a reference to the Node or `None` for `NULL_REVISION` | |
148 | fn node(&self, rev: Revision) -> Option<&Node>; |
|
148 | fn node(&self, rev: Revision) -> Option<&Node>; | |
149 |
|
149 | |||
150 | /// Return a [`Revision`] if `rev` is a valid revision number for this |
|
150 | /// Return a [`Revision`] if `rev` is a valid revision number for this | |
151 | /// index |
|
151 | /// index | |
152 | fn check_revision(&self, rev: UncheckedRevision) -> Option<Revision> { |
|
152 | fn check_revision(&self, rev: UncheckedRevision) -> Option<Revision> { | |
153 | let rev = rev.0; |
|
153 | let rev = rev.0; | |
154 |
|
154 | |||
155 | if rev == NULL_REVISION.0 || (rev >= 0 && (rev as usize) < self.len()) |
|
155 | if rev == NULL_REVISION.0 || (rev >= 0 && (rev as usize) < self.len()) | |
156 | { |
|
156 | { | |
157 | Some(Revision(rev)) |
|
157 | Some(Revision(rev)) | |
158 | } else { |
|
158 | } else { | |
159 | None |
|
159 | None | |
160 | } |
|
160 | } | |
161 | } |
|
161 | } | |
162 | } |
|
162 | } | |
163 |
|
163 | |||
164 | const REVISION_FLAG_CENSORED: u16 = 1 << 15; |
|
164 | const REVISION_FLAG_CENSORED: u16 = 1 << 15; | |
165 | const REVISION_FLAG_ELLIPSIS: u16 = 1 << 14; |
|
165 | const REVISION_FLAG_ELLIPSIS: u16 = 1 << 14; | |
166 | const REVISION_FLAG_EXTSTORED: u16 = 1 << 13; |
|
166 | const REVISION_FLAG_EXTSTORED: u16 = 1 << 13; | |
167 | const REVISION_FLAG_HASCOPIESINFO: u16 = 1 << 12; |
|
167 | const REVISION_FLAG_HASCOPIESINFO: u16 = 1 << 12; | |
168 |
|
168 | |||
169 | // Keep this in sync with REVIDX_KNOWN_FLAGS in |
|
169 | // Keep this in sync with REVIDX_KNOWN_FLAGS in | |
170 | // mercurial/revlogutils/flagutil.py |
|
170 | // mercurial/revlogutils/flagutil.py | |
171 | const REVIDX_KNOWN_FLAGS: u16 = REVISION_FLAG_CENSORED |
|
171 | const REVIDX_KNOWN_FLAGS: u16 = REVISION_FLAG_CENSORED | |
172 | | REVISION_FLAG_ELLIPSIS |
|
172 | | REVISION_FLAG_ELLIPSIS | |
173 | | REVISION_FLAG_EXTSTORED |
|
173 | | REVISION_FLAG_EXTSTORED | |
174 | | REVISION_FLAG_HASCOPIESINFO; |
|
174 | | REVISION_FLAG_HASCOPIESINFO; | |
175 |
|
175 | |||
176 | const NULL_REVLOG_ENTRY_FLAGS: u16 = 0; |
|
176 | const NULL_REVLOG_ENTRY_FLAGS: u16 = 0; | |
177 |
|
177 | |||
178 | #[derive(Debug, derive_more::From, derive_more::Display)] |
|
178 | #[derive(Debug, derive_more::From, derive_more::Display)] | |
179 | pub enum RevlogError { |
|
179 | pub enum RevlogError { | |
180 | InvalidRevision, |
|
180 | InvalidRevision, | |
181 | /// Working directory is not supported |
|
181 | /// Working directory is not supported | |
182 | WDirUnsupported, |
|
182 | WDirUnsupported, | |
183 | /// Found more than one entry whose ID match the requested prefix |
|
183 | /// Found more than one entry whose ID match the requested prefix | |
184 | AmbiguousPrefix, |
|
184 | AmbiguousPrefix, | |
185 | #[from] |
|
185 | #[from] | |
186 | Other(HgError), |
|
186 | Other(HgError), | |
187 | } |
|
187 | } | |
188 |
|
188 | |||
189 | impl From<NodeMapError> for RevlogError { |
|
189 | impl From<NodeMapError> for RevlogError { | |
190 | fn from(error: NodeMapError) -> Self { |
|
190 | fn from(error: NodeMapError) -> Self { | |
191 | match error { |
|
191 | match error { | |
192 | NodeMapError::MultipleResults => RevlogError::AmbiguousPrefix, |
|
192 | NodeMapError::MultipleResults => RevlogError::AmbiguousPrefix, | |
193 | NodeMapError::RevisionNotInIndex(rev) => RevlogError::corrupted( |
|
193 | NodeMapError::RevisionNotInIndex(rev) => RevlogError::corrupted( | |
194 | format!("nodemap point to revision {} not in index", rev), |
|
194 | format!("nodemap point to revision {} not in index", rev), | |
195 | ), |
|
195 | ), | |
196 | } |
|
196 | } | |
197 | } |
|
197 | } | |
198 | } |
|
198 | } | |
199 |
|
199 | |||
200 | fn corrupted<S: AsRef<str>>(context: S) -> HgError { |
|
200 | fn corrupted<S: AsRef<str>>(context: S) -> HgError { | |
201 | HgError::corrupted(format!("corrupted revlog, {}", context.as_ref())) |
|
201 | HgError::corrupted(format!("corrupted revlog, {}", context.as_ref())) | |
202 | } |
|
202 | } | |
203 |
|
203 | |||
204 | impl RevlogError { |
|
204 | impl RevlogError { | |
205 | fn corrupted<S: AsRef<str>>(context: S) -> Self { |
|
205 | fn corrupted<S: AsRef<str>>(context: S) -> Self { | |
206 | RevlogError::Other(corrupted(context)) |
|
206 | RevlogError::Other(corrupted(context)) | |
207 | } |
|
207 | } | |
208 | } |
|
208 | } | |
209 |
|
209 | |||
210 | /// Read only implementation of revlog. |
|
210 | /// Read only implementation of revlog. | |
211 | pub struct Revlog { |
|
211 | pub struct Revlog { | |
212 | /// When index and data are not interleaved: bytes of the revlog index. |
|
212 | /// When index and data are not interleaved: bytes of the revlog index. | |
213 | /// When index and data are interleaved: bytes of the revlog index and |
|
213 | /// When index and data are interleaved: bytes of the revlog index and | |
214 | /// data. |
|
214 | /// data. | |
215 | index: Index, |
|
215 | index: Index, | |
216 | /// When index and data are not interleaved: bytes of the revlog data |
|
216 | /// When index and data are not interleaved: bytes of the revlog data | |
217 | data_bytes: Option<Box<dyn Deref<Target = [u8]> + Send>>, |
|
217 | data_bytes: Option<Box<dyn Deref<Target = [u8]> + Send>>, | |
218 | /// When present on disk: the persistent nodemap for this revlog |
|
218 | /// When present on disk: the persistent nodemap for this revlog | |
219 | nodemap: Option<nodemap::NodeTree>, |
|
219 | nodemap: Option<nodemap::NodeTree>, | |
220 | } |
|
220 | } | |
221 |
|
221 | |||
222 | impl Graph for Revlog { |
|
222 | impl Graph for Revlog { | |
223 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { |
|
223 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { | |
224 | self.index.parents(rev) |
|
224 | self.index.parents(rev) | |
225 | } |
|
225 | } | |
226 | } |
|
226 | } | |
227 |
|
227 | |||
|
228 | #[derive(Debug, Copy, Clone)] | |||
|
229 | pub enum RevlogVersionOptions { | |||
|
230 | V0, | |||
|
231 | V1 { generaldelta: bool }, | |||
|
232 | V2, | |||
|
233 | ChangelogV2 { compute_rank: bool }, | |||
|
234 | } | |||
|
235 | ||||
|
236 | /// Options to govern how a revlog should be opened, usually from the | |||
|
237 | /// repository configuration or requirements. | |||
|
238 | #[derive(Debug, Copy, Clone)] | |||
|
239 | pub struct RevlogOpenOptions { | |||
|
240 | /// The revlog version, along with any option specific to this version | |||
|
241 | pub version: RevlogVersionOptions, | |||
|
242 | /// Whether the revlog uses a persistent nodemap. | |||
|
243 | pub use_nodemap: bool, | |||
|
244 | // TODO other non-header/version options, | |||
|
245 | } | |||
|
246 | ||||
|
247 | impl RevlogOpenOptions { | |||
|
248 | pub fn new() -> Self { | |||
|
249 | Self { | |||
|
250 | version: RevlogVersionOptions::V1 { generaldelta: true }, | |||
|
251 | use_nodemap: false, | |||
|
252 | } | |||
|
253 | } | |||
|
254 | ||||
|
255 | fn default_index_header(&self) -> index::IndexHeader { | |||
|
256 | index::IndexHeader { | |||
|
257 | header_bytes: match self.version { | |||
|
258 | RevlogVersionOptions::V0 => [0, 0, 0, 0], | |||
|
259 | RevlogVersionOptions::V1 { generaldelta } => { | |||
|
260 | [0, if generaldelta { 3 } else { 1 }, 0, 1] | |||
|
261 | } | |||
|
262 | RevlogVersionOptions::V2 => 0xDEADu32.to_be_bytes(), | |||
|
263 | RevlogVersionOptions::ChangelogV2 { compute_rank: _ } => { | |||
|
264 | 0xD34Du32.to_be_bytes() | |||
|
265 | } | |||
|
266 | }, | |||
|
267 | } | |||
|
268 | } | |||
|
269 | } | |||
|
270 | ||||
|
271 | impl Default for RevlogOpenOptions { | |||
|
272 | fn default() -> Self { | |||
|
273 | Self::new() | |||
|
274 | } | |||
|
275 | } | |||
|
276 | ||||
228 | impl Revlog { |
|
277 | impl Revlog { | |
229 | /// Open a revlog index file. |
|
278 | /// Open a revlog index file. | |
230 | /// |
|
279 | /// | |
231 | /// It will also open the associated data file if index and data are not |
|
280 | /// It will also open the associated data file if index and data are not | |
232 | /// interleaved. |
|
281 | /// interleaved. | |
233 | pub fn open( |
|
282 | pub fn open( | |
234 | store_vfs: &Vfs, |
|
283 | store_vfs: &Vfs, | |
235 | index_path: impl AsRef<Path>, |
|
284 | index_path: impl AsRef<Path>, | |
236 | data_path: Option<&Path>, |
|
285 | data_path: Option<&Path>, | |
237 | use_nodemap: bool, |
|
286 | options: RevlogOpenOptions, | |
238 | ) -> Result<Self, HgError> { |
|
287 | ) -> Result<Self, HgError> { | |
239 |
Self::open_gen(store_vfs, index_path, data_path, |
|
288 | Self::open_gen(store_vfs, index_path, data_path, options, None) | |
240 | } |
|
289 | } | |
241 |
|
290 | |||
242 | fn open_gen( |
|
291 | fn open_gen( | |
243 | store_vfs: &Vfs, |
|
292 | store_vfs: &Vfs, | |
244 | index_path: impl AsRef<Path>, |
|
293 | index_path: impl AsRef<Path>, | |
245 | data_path: Option<&Path>, |
|
294 | data_path: Option<&Path>, | |
246 | use_nodemap: bool, |
|
295 | options: RevlogOpenOptions, | |
247 | nodemap_for_test: Option<nodemap::NodeTree>, |
|
296 | nodemap_for_test: Option<nodemap::NodeTree>, | |
248 | ) -> Result<Self, HgError> { |
|
297 | ) -> Result<Self, HgError> { | |
249 | let index_path = index_path.as_ref(); |
|
298 | let index_path = index_path.as_ref(); | |
250 | let index = { |
|
299 | let index = { | |
251 | match store_vfs.mmap_open_opt(index_path)? { |
|
300 | match store_vfs.mmap_open_opt(index_path)? { | |
252 |
None => Index::new( |
|
301 | None => Index::new( | |
|
302 | Box::<Vec<_>>::default(), | |||
|
303 | options.default_index_header(), | |||
|
304 | ), | |||
253 | Some(index_mmap) => { |
|
305 | Some(index_mmap) => { | |
254 |
let index = Index::new( |
|
306 | let index = Index::new( | |
|
307 | Box::new(index_mmap), | |||
|
308 | options.default_index_header(), | |||
|
309 | )?; | |||
255 | Ok(index) |
|
310 | Ok(index) | |
256 | } |
|
311 | } | |
257 | } |
|
312 | } | |
258 | }?; |
|
313 | }?; | |
259 |
|
314 | |||
260 | let default_data_path = index_path.with_extension("d"); |
|
315 | let default_data_path = index_path.with_extension("d"); | |
261 |
|
316 | |||
262 | // type annotation required |
|
317 | // type annotation required | |
263 | // won't recognize Mmap as Deref<Target = [u8]> |
|
318 | // won't recognize Mmap as Deref<Target = [u8]> | |
264 | let data_bytes: Option<Box<dyn Deref<Target = [u8]> + Send>> = |
|
319 | let data_bytes: Option<Box<dyn Deref<Target = [u8]> + Send>> = | |
265 | if index.is_inline() { |
|
320 | if index.is_inline() { | |
266 | None |
|
321 | None | |
267 | } else { |
|
322 | } else { | |
268 | let data_path = data_path.unwrap_or(&default_data_path); |
|
323 | let data_path = data_path.unwrap_or(&default_data_path); | |
269 | let data_mmap = store_vfs.mmap_open(data_path)?; |
|
324 | let data_mmap = store_vfs.mmap_open(data_path)?; | |
270 | Some(Box::new(data_mmap)) |
|
325 | Some(Box::new(data_mmap)) | |
271 | }; |
|
326 | }; | |
272 |
|
327 | |||
273 | let nodemap = if index.is_inline() || !use_nodemap { |
|
328 | let nodemap = if index.is_inline() || !options.use_nodemap { | |
274 | None |
|
329 | None | |
275 | } else { |
|
330 | } else { | |
276 | NodeMapDocket::read_from_file(store_vfs, index_path)?.map( |
|
331 | NodeMapDocket::read_from_file(store_vfs, index_path)?.map( | |
277 | |(docket, data)| { |
|
332 | |(docket, data)| { | |
278 | nodemap::NodeTree::load_bytes( |
|
333 | nodemap::NodeTree::load_bytes( | |
279 | Box::new(data), |
|
334 | Box::new(data), | |
280 | docket.data_length, |
|
335 | docket.data_length, | |
281 | ) |
|
336 | ) | |
282 | }, |
|
337 | }, | |
283 | ) |
|
338 | ) | |
284 | }; |
|
339 | }; | |
285 |
|
340 | |||
286 | let nodemap = nodemap_for_test.or(nodemap); |
|
341 | let nodemap = nodemap_for_test.or(nodemap); | |
287 |
|
342 | |||
288 | Ok(Revlog { |
|
343 | Ok(Revlog { | |
289 | index, |
|
344 | index, | |
290 | data_bytes, |
|
345 | data_bytes, | |
291 | nodemap, |
|
346 | nodemap, | |
292 | }) |
|
347 | }) | |
293 | } |
|
348 | } | |
294 |
|
349 | |||
295 | /// Return number of entries of the `Revlog`. |
|
350 | /// Return number of entries of the `Revlog`. | |
296 | pub fn len(&self) -> usize { |
|
351 | pub fn len(&self) -> usize { | |
297 | self.index.len() |
|
352 | self.index.len() | |
298 | } |
|
353 | } | |
299 |
|
354 | |||
300 | /// Returns `true` if the `Revlog` has zero `entries`. |
|
355 | /// Returns `true` if the `Revlog` has zero `entries`. | |
301 | pub fn is_empty(&self) -> bool { |
|
356 | pub fn is_empty(&self) -> bool { | |
302 | self.index.is_empty() |
|
357 | self.index.is_empty() | |
303 | } |
|
358 | } | |
304 |
|
359 | |||
305 | /// Returns the node ID for the given revision number, if it exists in this |
|
360 | /// Returns the node ID for the given revision number, if it exists in this | |
306 | /// revlog |
|
361 | /// revlog | |
307 | pub fn node_from_rev(&self, rev: UncheckedRevision) -> Option<&Node> { |
|
362 | pub fn node_from_rev(&self, rev: UncheckedRevision) -> Option<&Node> { | |
308 | if rev == NULL_REVISION.into() { |
|
363 | if rev == NULL_REVISION.into() { | |
309 | return Some(&NULL_NODE); |
|
364 | return Some(&NULL_NODE); | |
310 | } |
|
365 | } | |
311 | let rev = self.index.check_revision(rev)?; |
|
366 | let rev = self.index.check_revision(rev)?; | |
312 | Some(self.index.get_entry(rev)?.hash()) |
|
367 | Some(self.index.get_entry(rev)?.hash()) | |
313 | } |
|
368 | } | |
314 |
|
369 | |||
315 | /// Return the revision number for the given node ID, if it exists in this |
|
370 | /// Return the revision number for the given node ID, if it exists in this | |
316 | /// revlog |
|
371 | /// revlog | |
317 | pub fn rev_from_node( |
|
372 | pub fn rev_from_node( | |
318 | &self, |
|
373 | &self, | |
319 | node: NodePrefix, |
|
374 | node: NodePrefix, | |
320 | ) -> Result<Revision, RevlogError> { |
|
375 | ) -> Result<Revision, RevlogError> { | |
321 | if let Some(nodemap) = &self.nodemap { |
|
376 | if let Some(nodemap) = &self.nodemap { | |
322 | nodemap |
|
377 | nodemap | |
323 | .find_bin(&self.index, node)? |
|
378 | .find_bin(&self.index, node)? | |
324 | .ok_or(RevlogError::InvalidRevision) |
|
379 | .ok_or(RevlogError::InvalidRevision) | |
325 | } else { |
|
380 | } else { | |
326 | self.rev_from_node_no_persistent_nodemap(node) |
|
381 | self.rev_from_node_no_persistent_nodemap(node) | |
327 | } |
|
382 | } | |
328 | } |
|
383 | } | |
329 |
|
384 | |||
330 | /// Same as `rev_from_node`, without using a persistent nodemap |
|
385 | /// Same as `rev_from_node`, without using a persistent nodemap | |
331 | /// |
|
386 | /// | |
332 | /// This is used as fallback when a persistent nodemap is not present. |
|
387 | /// This is used as fallback when a persistent nodemap is not present. | |
333 | /// This happens when the persistent-nodemap experimental feature is not |
|
388 | /// This happens when the persistent-nodemap experimental feature is not | |
334 | /// enabled, or for small revlogs. |
|
389 | /// enabled, or for small revlogs. | |
335 | fn rev_from_node_no_persistent_nodemap( |
|
390 | fn rev_from_node_no_persistent_nodemap( | |
336 | &self, |
|
391 | &self, | |
337 | node: NodePrefix, |
|
392 | node: NodePrefix, | |
338 | ) -> Result<Revision, RevlogError> { |
|
393 | ) -> Result<Revision, RevlogError> { | |
339 | // Linear scan of the revlog |
|
394 | // Linear scan of the revlog | |
340 | // TODO: consider building a non-persistent nodemap in memory to |
|
395 | // TODO: consider building a non-persistent nodemap in memory to | |
341 | // optimize these cases. |
|
396 | // optimize these cases. | |
342 | let mut found_by_prefix = None; |
|
397 | let mut found_by_prefix = None; | |
343 | for rev in (-1..self.len() as BaseRevision).rev() { |
|
398 | for rev in (-1..self.len() as BaseRevision).rev() { | |
344 | let rev = Revision(rev as BaseRevision); |
|
399 | let rev = Revision(rev as BaseRevision); | |
345 | let candidate_node = if rev == Revision(-1) { |
|
400 | let candidate_node = if rev == Revision(-1) { | |
346 | NULL_NODE |
|
401 | NULL_NODE | |
347 | } else { |
|
402 | } else { | |
348 | let index_entry = |
|
403 | let index_entry = | |
349 | self.index.get_entry(rev).ok_or_else(|| { |
|
404 | self.index.get_entry(rev).ok_or_else(|| { | |
350 | HgError::corrupted( |
|
405 | HgError::corrupted( | |
351 | "revlog references a revision not in the index", |
|
406 | "revlog references a revision not in the index", | |
352 | ) |
|
407 | ) | |
353 | })?; |
|
408 | })?; | |
354 | *index_entry.hash() |
|
409 | *index_entry.hash() | |
355 | }; |
|
410 | }; | |
356 | if node == candidate_node { |
|
411 | if node == candidate_node { | |
357 | return Ok(rev); |
|
412 | return Ok(rev); | |
358 | } |
|
413 | } | |
359 | if node.is_prefix_of(&candidate_node) { |
|
414 | if node.is_prefix_of(&candidate_node) { | |
360 | if found_by_prefix.is_some() { |
|
415 | if found_by_prefix.is_some() { | |
361 | return Err(RevlogError::AmbiguousPrefix); |
|
416 | return Err(RevlogError::AmbiguousPrefix); | |
362 | } |
|
417 | } | |
363 | found_by_prefix = Some(rev) |
|
418 | found_by_prefix = Some(rev) | |
364 | } |
|
419 | } | |
365 | } |
|
420 | } | |
366 | found_by_prefix.ok_or(RevlogError::InvalidRevision) |
|
421 | found_by_prefix.ok_or(RevlogError::InvalidRevision) | |
367 | } |
|
422 | } | |
368 |
|
423 | |||
369 | /// Returns whether the given revision exists in this revlog. |
|
424 | /// Returns whether the given revision exists in this revlog. | |
370 | pub fn has_rev(&self, rev: UncheckedRevision) -> bool { |
|
425 | pub fn has_rev(&self, rev: UncheckedRevision) -> bool { | |
371 | self.index.check_revision(rev).is_some() |
|
426 | self.index.check_revision(rev).is_some() | |
372 | } |
|
427 | } | |
373 |
|
428 | |||
374 | /// Return the full data associated to a revision. |
|
429 | /// Return the full data associated to a revision. | |
375 | /// |
|
430 | /// | |
376 | /// All entries required to build the final data out of deltas will be |
|
431 | /// All entries required to build the final data out of deltas will be | |
377 | /// retrieved as needed, and the deltas will be applied to the inital |
|
432 | /// retrieved as needed, and the deltas will be applied to the inital | |
378 | /// snapshot to rebuild the final data. |
|
433 | /// snapshot to rebuild the final data. | |
379 | pub fn get_rev_data( |
|
434 | pub fn get_rev_data( | |
380 | &self, |
|
435 | &self, | |
381 | rev: UncheckedRevision, |
|
436 | rev: UncheckedRevision, | |
382 | ) -> Result<Cow<[u8]>, RevlogError> { |
|
437 | ) -> Result<Cow<[u8]>, RevlogError> { | |
383 | if rev == NULL_REVISION.into() { |
|
438 | if rev == NULL_REVISION.into() { | |
384 | return Ok(Cow::Borrowed(&[])); |
|
439 | return Ok(Cow::Borrowed(&[])); | |
385 | }; |
|
440 | }; | |
386 | self.get_entry(rev)?.data() |
|
441 | self.get_entry(rev)?.data() | |
387 | } |
|
442 | } | |
388 |
|
443 | |||
389 | /// [`Self::get_rev_data`] for checked revisions. |
|
444 | /// [`Self::get_rev_data`] for checked revisions. | |
390 | pub fn get_rev_data_for_checked_rev( |
|
445 | pub fn get_rev_data_for_checked_rev( | |
391 | &self, |
|
446 | &self, | |
392 | rev: Revision, |
|
447 | rev: Revision, | |
393 | ) -> Result<Cow<[u8]>, RevlogError> { |
|
448 | ) -> Result<Cow<[u8]>, RevlogError> { | |
394 | if rev == NULL_REVISION { |
|
449 | if rev == NULL_REVISION { | |
395 | return Ok(Cow::Borrowed(&[])); |
|
450 | return Ok(Cow::Borrowed(&[])); | |
396 | }; |
|
451 | }; | |
397 | self.get_entry_for_checked_rev(rev)?.data() |
|
452 | self.get_entry_for_checked_rev(rev)?.data() | |
398 | } |
|
453 | } | |
399 |
|
454 | |||
400 | /// Check the hash of some given data against the recorded hash. |
|
455 | /// Check the hash of some given data against the recorded hash. | |
401 | pub fn check_hash( |
|
456 | pub fn check_hash( | |
402 | &self, |
|
457 | &self, | |
403 | p1: Revision, |
|
458 | p1: Revision, | |
404 | p2: Revision, |
|
459 | p2: Revision, | |
405 | expected: &[u8], |
|
460 | expected: &[u8], | |
406 | data: &[u8], |
|
461 | data: &[u8], | |
407 | ) -> bool { |
|
462 | ) -> bool { | |
408 | let e1 = self.index.get_entry(p1); |
|
463 | let e1 = self.index.get_entry(p1); | |
409 | let h1 = match e1 { |
|
464 | let h1 = match e1 { | |
410 | Some(ref entry) => entry.hash(), |
|
465 | Some(ref entry) => entry.hash(), | |
411 | None => &NULL_NODE, |
|
466 | None => &NULL_NODE, | |
412 | }; |
|
467 | }; | |
413 | let e2 = self.index.get_entry(p2); |
|
468 | let e2 = self.index.get_entry(p2); | |
414 | let h2 = match e2 { |
|
469 | let h2 = match e2 { | |
415 | Some(ref entry) => entry.hash(), |
|
470 | Some(ref entry) => entry.hash(), | |
416 | None => &NULL_NODE, |
|
471 | None => &NULL_NODE, | |
417 | }; |
|
472 | }; | |
418 |
|
473 | |||
419 | hash(data, h1.as_bytes(), h2.as_bytes()) == expected |
|
474 | hash(data, h1.as_bytes(), h2.as_bytes()) == expected | |
420 | } |
|
475 | } | |
421 |
|
476 | |||
422 | /// Build the full data of a revision out its snapshot |
|
477 | /// Build the full data of a revision out its snapshot | |
423 | /// and its deltas. |
|
478 | /// and its deltas. | |
424 | fn build_data_from_deltas( |
|
479 | fn build_data_from_deltas( | |
425 | snapshot: RevlogEntry, |
|
480 | snapshot: RevlogEntry, | |
426 | deltas: &[RevlogEntry], |
|
481 | deltas: &[RevlogEntry], | |
427 | ) -> Result<Vec<u8>, HgError> { |
|
482 | ) -> Result<Vec<u8>, HgError> { | |
428 | let snapshot = snapshot.data_chunk()?; |
|
483 | let snapshot = snapshot.data_chunk()?; | |
429 | let deltas = deltas |
|
484 | let deltas = deltas | |
430 | .iter() |
|
485 | .iter() | |
431 | .rev() |
|
486 | .rev() | |
432 | .map(RevlogEntry::data_chunk) |
|
487 | .map(RevlogEntry::data_chunk) | |
433 | .collect::<Result<Vec<_>, _>>()?; |
|
488 | .collect::<Result<Vec<_>, _>>()?; | |
434 | let patches: Vec<_> = |
|
489 | let patches: Vec<_> = | |
435 | deltas.iter().map(|d| patch::PatchList::new(d)).collect(); |
|
490 | deltas.iter().map(|d| patch::PatchList::new(d)).collect(); | |
436 | let patch = patch::fold_patch_lists(&patches); |
|
491 | let patch = patch::fold_patch_lists(&patches); | |
437 | Ok(patch.apply(&snapshot)) |
|
492 | Ok(patch.apply(&snapshot)) | |
438 | } |
|
493 | } | |
439 |
|
494 | |||
440 | /// Return the revlog data. |
|
495 | /// Return the revlog data. | |
441 | fn data(&self) -> &[u8] { |
|
496 | fn data(&self) -> &[u8] { | |
442 | match &self.data_bytes { |
|
497 | match &self.data_bytes { | |
443 | Some(data_bytes) => data_bytes, |
|
498 | Some(data_bytes) => data_bytes, | |
444 | None => panic!( |
|
499 | None => panic!( | |
445 | "forgot to load the data or trying to access inline data" |
|
500 | "forgot to load the data or trying to access inline data" | |
446 | ), |
|
501 | ), | |
447 | } |
|
502 | } | |
448 | } |
|
503 | } | |
449 |
|
504 | |||
450 | pub fn make_null_entry(&self) -> RevlogEntry { |
|
505 | pub fn make_null_entry(&self) -> RevlogEntry { | |
451 | RevlogEntry { |
|
506 | RevlogEntry { | |
452 | revlog: self, |
|
507 | revlog: self, | |
453 | rev: NULL_REVISION, |
|
508 | rev: NULL_REVISION, | |
454 | bytes: b"", |
|
509 | bytes: b"", | |
455 | compressed_len: 0, |
|
510 | compressed_len: 0, | |
456 | uncompressed_len: 0, |
|
511 | uncompressed_len: 0, | |
457 | base_rev_or_base_of_delta_chain: None, |
|
512 | base_rev_or_base_of_delta_chain: None, | |
458 | p1: NULL_REVISION, |
|
513 | p1: NULL_REVISION, | |
459 | p2: NULL_REVISION, |
|
514 | p2: NULL_REVISION, | |
460 | flags: NULL_REVLOG_ENTRY_FLAGS, |
|
515 | flags: NULL_REVLOG_ENTRY_FLAGS, | |
461 | hash: NULL_NODE, |
|
516 | hash: NULL_NODE, | |
462 | } |
|
517 | } | |
463 | } |
|
518 | } | |
464 |
|
519 | |||
465 | fn get_entry_for_checked_rev( |
|
520 | fn get_entry_for_checked_rev( | |
466 | &self, |
|
521 | &self, | |
467 | rev: Revision, |
|
522 | rev: Revision, | |
468 | ) -> Result<RevlogEntry, RevlogError> { |
|
523 | ) -> Result<RevlogEntry, RevlogError> { | |
469 | if rev == NULL_REVISION { |
|
524 | if rev == NULL_REVISION { | |
470 | return Ok(self.make_null_entry()); |
|
525 | return Ok(self.make_null_entry()); | |
471 | } |
|
526 | } | |
472 | let index_entry = self |
|
527 | let index_entry = self | |
473 | .index |
|
528 | .index | |
474 | .get_entry(rev) |
|
529 | .get_entry(rev) | |
475 | .ok_or(RevlogError::InvalidRevision)?; |
|
530 | .ok_or(RevlogError::InvalidRevision)?; | |
476 | let start = index_entry.offset(); |
|
531 | let start = index_entry.offset(); | |
477 | let end = start + index_entry.compressed_len() as usize; |
|
532 | let end = start + index_entry.compressed_len() as usize; | |
478 | let data = if self.index.is_inline() { |
|
533 | let data = if self.index.is_inline() { | |
479 | self.index.data(start, end) |
|
534 | self.index.data(start, end) | |
480 | } else { |
|
535 | } else { | |
481 | &self.data()[start..end] |
|
536 | &self.data()[start..end] | |
482 | }; |
|
537 | }; | |
483 | let base_rev = self |
|
538 | let base_rev = self | |
484 | .index |
|
539 | .index | |
485 | .check_revision(index_entry.base_revision_or_base_of_delta_chain()) |
|
540 | .check_revision(index_entry.base_revision_or_base_of_delta_chain()) | |
486 | .ok_or_else(|| { |
|
541 | .ok_or_else(|| { | |
487 | RevlogError::corrupted(format!( |
|
542 | RevlogError::corrupted(format!( | |
488 | "base revision for rev {} is invalid", |
|
543 | "base revision for rev {} is invalid", | |
489 | rev |
|
544 | rev | |
490 | )) |
|
545 | )) | |
491 | })?; |
|
546 | })?; | |
492 | let p1 = |
|
547 | let p1 = | |
493 | self.index.check_revision(index_entry.p1()).ok_or_else(|| { |
|
548 | self.index.check_revision(index_entry.p1()).ok_or_else(|| { | |
494 | RevlogError::corrupted(format!( |
|
549 | RevlogError::corrupted(format!( | |
495 | "p1 for rev {} is invalid", |
|
550 | "p1 for rev {} is invalid", | |
496 | rev |
|
551 | rev | |
497 | )) |
|
552 | )) | |
498 | })?; |
|
553 | })?; | |
499 | let p2 = |
|
554 | let p2 = | |
500 | self.index.check_revision(index_entry.p2()).ok_or_else(|| { |
|
555 | self.index.check_revision(index_entry.p2()).ok_or_else(|| { | |
501 | RevlogError::corrupted(format!( |
|
556 | RevlogError::corrupted(format!( | |
502 | "p2 for rev {} is invalid", |
|
557 | "p2 for rev {} is invalid", | |
503 | rev |
|
558 | rev | |
504 | )) |
|
559 | )) | |
505 | })?; |
|
560 | })?; | |
506 | let entry = RevlogEntry { |
|
561 | let entry = RevlogEntry { | |
507 | revlog: self, |
|
562 | revlog: self, | |
508 | rev, |
|
563 | rev, | |
509 | bytes: data, |
|
564 | bytes: data, | |
510 | compressed_len: index_entry.compressed_len(), |
|
565 | compressed_len: index_entry.compressed_len(), | |
511 | uncompressed_len: index_entry.uncompressed_len(), |
|
566 | uncompressed_len: index_entry.uncompressed_len(), | |
512 | base_rev_or_base_of_delta_chain: if base_rev == rev { |
|
567 | base_rev_or_base_of_delta_chain: if base_rev == rev { | |
513 | None |
|
568 | None | |
514 | } else { |
|
569 | } else { | |
515 | Some(base_rev) |
|
570 | Some(base_rev) | |
516 | }, |
|
571 | }, | |
517 | p1, |
|
572 | p1, | |
518 | p2, |
|
573 | p2, | |
519 | flags: index_entry.flags(), |
|
574 | flags: index_entry.flags(), | |
520 | hash: *index_entry.hash(), |
|
575 | hash: *index_entry.hash(), | |
521 | }; |
|
576 | }; | |
522 | Ok(entry) |
|
577 | Ok(entry) | |
523 | } |
|
578 | } | |
524 |
|
579 | |||
525 | /// Get an entry of the revlog. |
|
580 | /// Get an entry of the revlog. | |
526 | pub fn get_entry( |
|
581 | pub fn get_entry( | |
527 | &self, |
|
582 | &self, | |
528 | rev: UncheckedRevision, |
|
583 | rev: UncheckedRevision, | |
529 | ) -> Result<RevlogEntry, RevlogError> { |
|
584 | ) -> Result<RevlogEntry, RevlogError> { | |
530 | if rev == NULL_REVISION.into() { |
|
585 | if rev == NULL_REVISION.into() { | |
531 | return Ok(self.make_null_entry()); |
|
586 | return Ok(self.make_null_entry()); | |
532 | } |
|
587 | } | |
533 | let rev = self.index.check_revision(rev).ok_or_else(|| { |
|
588 | let rev = self.index.check_revision(rev).ok_or_else(|| { | |
534 | RevlogError::corrupted(format!("rev {} is invalid", rev)) |
|
589 | RevlogError::corrupted(format!("rev {} is invalid", rev)) | |
535 | })?; |
|
590 | })?; | |
536 | self.get_entry_for_checked_rev(rev) |
|
591 | self.get_entry_for_checked_rev(rev) | |
537 | } |
|
592 | } | |
538 | } |
|
593 | } | |
539 |
|
594 | |||
540 | /// The revlog entry's bytes and the necessary informations to extract |
|
595 | /// The revlog entry's bytes and the necessary informations to extract | |
541 | /// the entry's data. |
|
596 | /// the entry's data. | |
542 | #[derive(Clone)] |
|
597 | #[derive(Clone)] | |
543 | pub struct RevlogEntry<'revlog> { |
|
598 | pub struct RevlogEntry<'revlog> { | |
544 | revlog: &'revlog Revlog, |
|
599 | revlog: &'revlog Revlog, | |
545 | rev: Revision, |
|
600 | rev: Revision, | |
546 | bytes: &'revlog [u8], |
|
601 | bytes: &'revlog [u8], | |
547 | compressed_len: u32, |
|
602 | compressed_len: u32, | |
548 | uncompressed_len: i32, |
|
603 | uncompressed_len: i32, | |
549 | base_rev_or_base_of_delta_chain: Option<Revision>, |
|
604 | base_rev_or_base_of_delta_chain: Option<Revision>, | |
550 | p1: Revision, |
|
605 | p1: Revision, | |
551 | p2: Revision, |
|
606 | p2: Revision, | |
552 | flags: u16, |
|
607 | flags: u16, | |
553 | hash: Node, |
|
608 | hash: Node, | |
554 | } |
|
609 | } | |
555 |
|
610 | |||
556 | thread_local! { |
|
611 | thread_local! { | |
557 | // seems fine to [unwrap] here: this can only fail due to memory allocation |
|
612 | // seems fine to [unwrap] here: this can only fail due to memory allocation | |
558 | // failing, and it's normal for that to cause panic. |
|
613 | // failing, and it's normal for that to cause panic. | |
559 | static ZSTD_DECODER : RefCell<zstd::bulk::Decompressor<'static>> = |
|
614 | static ZSTD_DECODER : RefCell<zstd::bulk::Decompressor<'static>> = | |
560 | RefCell::new(zstd::bulk::Decompressor::new().ok().unwrap()); |
|
615 | RefCell::new(zstd::bulk::Decompressor::new().ok().unwrap()); | |
561 | } |
|
616 | } | |
562 |
|
617 | |||
563 | fn zstd_decompress_to_buffer( |
|
618 | fn zstd_decompress_to_buffer( | |
564 | bytes: &[u8], |
|
619 | bytes: &[u8], | |
565 | buf: &mut Vec<u8>, |
|
620 | buf: &mut Vec<u8>, | |
566 | ) -> Result<usize, std::io::Error> { |
|
621 | ) -> Result<usize, std::io::Error> { | |
567 | ZSTD_DECODER |
|
622 | ZSTD_DECODER | |
568 | .with(|decoder| decoder.borrow_mut().decompress_to_buffer(bytes, buf)) |
|
623 | .with(|decoder| decoder.borrow_mut().decompress_to_buffer(bytes, buf)) | |
569 | } |
|
624 | } | |
570 |
|
625 | |||
571 | impl<'revlog> RevlogEntry<'revlog> { |
|
626 | impl<'revlog> RevlogEntry<'revlog> { | |
572 | pub fn revision(&self) -> Revision { |
|
627 | pub fn revision(&self) -> Revision { | |
573 | self.rev |
|
628 | self.rev | |
574 | } |
|
629 | } | |
575 |
|
630 | |||
576 | pub fn node(&self) -> &Node { |
|
631 | pub fn node(&self) -> &Node { | |
577 | &self.hash |
|
632 | &self.hash | |
578 | } |
|
633 | } | |
579 |
|
634 | |||
580 | pub fn uncompressed_len(&self) -> Option<u32> { |
|
635 | pub fn uncompressed_len(&self) -> Option<u32> { | |
581 | u32::try_from(self.uncompressed_len).ok() |
|
636 | u32::try_from(self.uncompressed_len).ok() | |
582 | } |
|
637 | } | |
583 |
|
638 | |||
584 | pub fn has_p1(&self) -> bool { |
|
639 | pub fn has_p1(&self) -> bool { | |
585 | self.p1 != NULL_REVISION |
|
640 | self.p1 != NULL_REVISION | |
586 | } |
|
641 | } | |
587 |
|
642 | |||
588 | pub fn p1_entry( |
|
643 | pub fn p1_entry( | |
589 | &self, |
|
644 | &self, | |
590 | ) -> Result<Option<RevlogEntry<'revlog>>, RevlogError> { |
|
645 | ) -> Result<Option<RevlogEntry<'revlog>>, RevlogError> { | |
591 | if self.p1 == NULL_REVISION { |
|
646 | if self.p1 == NULL_REVISION { | |
592 | Ok(None) |
|
647 | Ok(None) | |
593 | } else { |
|
648 | } else { | |
594 | Ok(Some(self.revlog.get_entry_for_checked_rev(self.p1)?)) |
|
649 | Ok(Some(self.revlog.get_entry_for_checked_rev(self.p1)?)) | |
595 | } |
|
650 | } | |
596 | } |
|
651 | } | |
597 |
|
652 | |||
598 | pub fn p2_entry( |
|
653 | pub fn p2_entry( | |
599 | &self, |
|
654 | &self, | |
600 | ) -> Result<Option<RevlogEntry<'revlog>>, RevlogError> { |
|
655 | ) -> Result<Option<RevlogEntry<'revlog>>, RevlogError> { | |
601 | if self.p2 == NULL_REVISION { |
|
656 | if self.p2 == NULL_REVISION { | |
602 | Ok(None) |
|
657 | Ok(None) | |
603 | } else { |
|
658 | } else { | |
604 | Ok(Some(self.revlog.get_entry_for_checked_rev(self.p2)?)) |
|
659 | Ok(Some(self.revlog.get_entry_for_checked_rev(self.p2)?)) | |
605 | } |
|
660 | } | |
606 | } |
|
661 | } | |
607 |
|
662 | |||
608 | pub fn p1(&self) -> Option<Revision> { |
|
663 | pub fn p1(&self) -> Option<Revision> { | |
609 | if self.p1 == NULL_REVISION { |
|
664 | if self.p1 == NULL_REVISION { | |
610 | None |
|
665 | None | |
611 | } else { |
|
666 | } else { | |
612 | Some(self.p1) |
|
667 | Some(self.p1) | |
613 | } |
|
668 | } | |
614 | } |
|
669 | } | |
615 |
|
670 | |||
616 | pub fn p2(&self) -> Option<Revision> { |
|
671 | pub fn p2(&self) -> Option<Revision> { | |
617 | if self.p2 == NULL_REVISION { |
|
672 | if self.p2 == NULL_REVISION { | |
618 | None |
|
673 | None | |
619 | } else { |
|
674 | } else { | |
620 | Some(self.p2) |
|
675 | Some(self.p2) | |
621 | } |
|
676 | } | |
622 | } |
|
677 | } | |
623 |
|
678 | |||
624 | pub fn is_censored(&self) -> bool { |
|
679 | pub fn is_censored(&self) -> bool { | |
625 | (self.flags & REVISION_FLAG_CENSORED) != 0 |
|
680 | (self.flags & REVISION_FLAG_CENSORED) != 0 | |
626 | } |
|
681 | } | |
627 |
|
682 | |||
628 | pub fn has_length_affecting_flag_processor(&self) -> bool { |
|
683 | pub fn has_length_affecting_flag_processor(&self) -> bool { | |
629 | // Relevant Python code: revlog.size() |
|
684 | // Relevant Python code: revlog.size() | |
630 | // note: ELLIPSIS is known to not change the content |
|
685 | // note: ELLIPSIS is known to not change the content | |
631 | (self.flags & (REVIDX_KNOWN_FLAGS ^ REVISION_FLAG_ELLIPSIS)) != 0 |
|
686 | (self.flags & (REVIDX_KNOWN_FLAGS ^ REVISION_FLAG_ELLIPSIS)) != 0 | |
632 | } |
|
687 | } | |
633 |
|
688 | |||
634 | /// The data for this entry, after resolving deltas if any. |
|
689 | /// The data for this entry, after resolving deltas if any. | |
635 | pub fn rawdata(&self) -> Result<Cow<'revlog, [u8]>, RevlogError> { |
|
690 | pub fn rawdata(&self) -> Result<Cow<'revlog, [u8]>, RevlogError> { | |
636 | let mut entry = self.clone(); |
|
691 | let mut entry = self.clone(); | |
637 | let mut delta_chain = vec![]; |
|
692 | let mut delta_chain = vec![]; | |
638 |
|
693 | |||
639 | // The meaning of `base_rev_or_base_of_delta_chain` depends on |
|
694 | // The meaning of `base_rev_or_base_of_delta_chain` depends on | |
640 | // generaldelta. See the doc on `ENTRY_DELTA_BASE` in |
|
695 | // generaldelta. See the doc on `ENTRY_DELTA_BASE` in | |
641 | // `mercurial/revlogutils/constants.py` and the code in |
|
696 | // `mercurial/revlogutils/constants.py` and the code in | |
642 | // [_chaininfo] and in [index_deltachain]. |
|
697 | // [_chaininfo] and in [index_deltachain]. | |
643 | let uses_generaldelta = self.revlog.index.uses_generaldelta(); |
|
698 | let uses_generaldelta = self.revlog.index.uses_generaldelta(); | |
644 | while let Some(base_rev) = entry.base_rev_or_base_of_delta_chain { |
|
699 | while let Some(base_rev) = entry.base_rev_or_base_of_delta_chain { | |
645 | entry = if uses_generaldelta { |
|
700 | entry = if uses_generaldelta { | |
646 | delta_chain.push(entry); |
|
701 | delta_chain.push(entry); | |
647 | self.revlog.get_entry_for_checked_rev(base_rev)? |
|
702 | self.revlog.get_entry_for_checked_rev(base_rev)? | |
648 | } else { |
|
703 | } else { | |
649 | let base_rev = UncheckedRevision(entry.rev.0 - 1); |
|
704 | let base_rev = UncheckedRevision(entry.rev.0 - 1); | |
650 | delta_chain.push(entry); |
|
705 | delta_chain.push(entry); | |
651 | self.revlog.get_entry(base_rev)? |
|
706 | self.revlog.get_entry(base_rev)? | |
652 | }; |
|
707 | }; | |
653 | } |
|
708 | } | |
654 |
|
709 | |||
655 | let data = if delta_chain.is_empty() { |
|
710 | let data = if delta_chain.is_empty() { | |
656 | entry.data_chunk()? |
|
711 | entry.data_chunk()? | |
657 | } else { |
|
712 | } else { | |
658 | Revlog::build_data_from_deltas(entry, &delta_chain)?.into() |
|
713 | Revlog::build_data_from_deltas(entry, &delta_chain)?.into() | |
659 | }; |
|
714 | }; | |
660 |
|
715 | |||
661 | Ok(data) |
|
716 | Ok(data) | |
662 | } |
|
717 | } | |
663 |
|
718 | |||
664 | fn check_data( |
|
719 | fn check_data( | |
665 | &self, |
|
720 | &self, | |
666 | data: Cow<'revlog, [u8]>, |
|
721 | data: Cow<'revlog, [u8]>, | |
667 | ) -> Result<Cow<'revlog, [u8]>, RevlogError> { |
|
722 | ) -> Result<Cow<'revlog, [u8]>, RevlogError> { | |
668 | if self.revlog.check_hash( |
|
723 | if self.revlog.check_hash( | |
669 | self.p1, |
|
724 | self.p1, | |
670 | self.p2, |
|
725 | self.p2, | |
671 | self.hash.as_bytes(), |
|
726 | self.hash.as_bytes(), | |
672 | &data, |
|
727 | &data, | |
673 | ) { |
|
728 | ) { | |
674 | Ok(data) |
|
729 | Ok(data) | |
675 | } else { |
|
730 | } else { | |
676 | if (self.flags & REVISION_FLAG_ELLIPSIS) != 0 { |
|
731 | if (self.flags & REVISION_FLAG_ELLIPSIS) != 0 { | |
677 | return Err(HgError::unsupported( |
|
732 | return Err(HgError::unsupported( | |
678 | "ellipsis revisions are not supported by rhg", |
|
733 | "ellipsis revisions are not supported by rhg", | |
679 | ) |
|
734 | ) | |
680 | .into()); |
|
735 | .into()); | |
681 | } |
|
736 | } | |
682 | Err(corrupted(format!( |
|
737 | Err(corrupted(format!( | |
683 | "hash check failed for revision {}", |
|
738 | "hash check failed for revision {}", | |
684 | self.rev |
|
739 | self.rev | |
685 | )) |
|
740 | )) | |
686 | .into()) |
|
741 | .into()) | |
687 | } |
|
742 | } | |
688 | } |
|
743 | } | |
689 |
|
744 | |||
690 | pub fn data(&self) -> Result<Cow<'revlog, [u8]>, RevlogError> { |
|
745 | pub fn data(&self) -> Result<Cow<'revlog, [u8]>, RevlogError> { | |
691 | let data = self.rawdata()?; |
|
746 | let data = self.rawdata()?; | |
692 | if self.rev == NULL_REVISION { |
|
747 | if self.rev == NULL_REVISION { | |
693 | return Ok(data); |
|
748 | return Ok(data); | |
694 | } |
|
749 | } | |
695 | if self.is_censored() { |
|
750 | if self.is_censored() { | |
696 | return Err(HgError::CensoredNodeError.into()); |
|
751 | return Err(HgError::CensoredNodeError.into()); | |
697 | } |
|
752 | } | |
698 | self.check_data(data) |
|
753 | self.check_data(data) | |
699 | } |
|
754 | } | |
700 |
|
755 | |||
701 | /// Extract the data contained in the entry. |
|
756 | /// Extract the data contained in the entry. | |
702 | /// This may be a delta. (See `is_delta`.) |
|
757 | /// This may be a delta. (See `is_delta`.) | |
703 | fn data_chunk(&self) -> Result<Cow<'revlog, [u8]>, HgError> { |
|
758 | fn data_chunk(&self) -> Result<Cow<'revlog, [u8]>, HgError> { | |
704 | if self.bytes.is_empty() { |
|
759 | if self.bytes.is_empty() { | |
705 | return Ok(Cow::Borrowed(&[])); |
|
760 | return Ok(Cow::Borrowed(&[])); | |
706 | } |
|
761 | } | |
707 | match self.bytes[0] { |
|
762 | match self.bytes[0] { | |
708 | // Revision data is the entirety of the entry, including this |
|
763 | // Revision data is the entirety of the entry, including this | |
709 | // header. |
|
764 | // header. | |
710 | b'\0' => Ok(Cow::Borrowed(self.bytes)), |
|
765 | b'\0' => Ok(Cow::Borrowed(self.bytes)), | |
711 | // Raw revision data follows. |
|
766 | // Raw revision data follows. | |
712 | b'u' => Ok(Cow::Borrowed(&self.bytes[1..])), |
|
767 | b'u' => Ok(Cow::Borrowed(&self.bytes[1..])), | |
713 | // zlib (RFC 1950) data. |
|
768 | // zlib (RFC 1950) data. | |
714 | b'x' => Ok(Cow::Owned(self.uncompressed_zlib_data()?)), |
|
769 | b'x' => Ok(Cow::Owned(self.uncompressed_zlib_data()?)), | |
715 | // zstd data. |
|
770 | // zstd data. | |
716 | b'\x28' => Ok(Cow::Owned(self.uncompressed_zstd_data()?)), |
|
771 | b'\x28' => Ok(Cow::Owned(self.uncompressed_zstd_data()?)), | |
717 | // A proper new format should have had a repo/store requirement. |
|
772 | // A proper new format should have had a repo/store requirement. | |
718 | format_type => Err(corrupted(format!( |
|
773 | format_type => Err(corrupted(format!( | |
719 | "unknown compression header '{}'", |
|
774 | "unknown compression header '{}'", | |
720 | format_type |
|
775 | format_type | |
721 | ))), |
|
776 | ))), | |
722 | } |
|
777 | } | |
723 | } |
|
778 | } | |
724 |
|
779 | |||
725 | fn uncompressed_zlib_data(&self) -> Result<Vec<u8>, HgError> { |
|
780 | fn uncompressed_zlib_data(&self) -> Result<Vec<u8>, HgError> { | |
726 | let mut decoder = ZlibDecoder::new(self.bytes); |
|
781 | let mut decoder = ZlibDecoder::new(self.bytes); | |
727 | if self.is_delta() { |
|
782 | if self.is_delta() { | |
728 | let mut buf = Vec::with_capacity(self.compressed_len as usize); |
|
783 | let mut buf = Vec::with_capacity(self.compressed_len as usize); | |
729 | decoder |
|
784 | decoder | |
730 | .read_to_end(&mut buf) |
|
785 | .read_to_end(&mut buf) | |
731 | .map_err(|e| corrupted(e.to_string()))?; |
|
786 | .map_err(|e| corrupted(e.to_string()))?; | |
732 | Ok(buf) |
|
787 | Ok(buf) | |
733 | } else { |
|
788 | } else { | |
734 | let cap = self.uncompressed_len.max(0) as usize; |
|
789 | let cap = self.uncompressed_len.max(0) as usize; | |
735 | let mut buf = vec![0; cap]; |
|
790 | let mut buf = vec![0; cap]; | |
736 | decoder |
|
791 | decoder | |
737 | .read_exact(&mut buf) |
|
792 | .read_exact(&mut buf) | |
738 | .map_err(|e| corrupted(e.to_string()))?; |
|
793 | .map_err(|e| corrupted(e.to_string()))?; | |
739 | Ok(buf) |
|
794 | Ok(buf) | |
740 | } |
|
795 | } | |
741 | } |
|
796 | } | |
742 |
|
797 | |||
743 | fn uncompressed_zstd_data(&self) -> Result<Vec<u8>, HgError> { |
|
798 | fn uncompressed_zstd_data(&self) -> Result<Vec<u8>, HgError> { | |
744 | let cap = self.uncompressed_len.max(0) as usize; |
|
799 | let cap = self.uncompressed_len.max(0) as usize; | |
745 | if self.is_delta() { |
|
800 | if self.is_delta() { | |
746 | // [cap] is usually an over-estimate of the space needed because |
|
801 | // [cap] is usually an over-estimate of the space needed because | |
747 | // it's the length of delta-decoded data, but we're interested |
|
802 | // it's the length of delta-decoded data, but we're interested | |
748 | // in the size of the delta. |
|
803 | // in the size of the delta. | |
749 | // This means we have to [shrink_to_fit] to avoid holding on |
|
804 | // This means we have to [shrink_to_fit] to avoid holding on | |
750 | // to a large chunk of memory, but it also means we must have a |
|
805 | // to a large chunk of memory, but it also means we must have a | |
751 | // fallback branch, for the case when the delta is longer than |
|
806 | // fallback branch, for the case when the delta is longer than | |
752 | // the original data (surprisingly, this does happen in practice) |
|
807 | // the original data (surprisingly, this does happen in practice) | |
753 | let mut buf = Vec::with_capacity(cap); |
|
808 | let mut buf = Vec::with_capacity(cap); | |
754 | match zstd_decompress_to_buffer(self.bytes, &mut buf) { |
|
809 | match zstd_decompress_to_buffer(self.bytes, &mut buf) { | |
755 | Ok(_) => buf.shrink_to_fit(), |
|
810 | Ok(_) => buf.shrink_to_fit(), | |
756 | Err(_) => { |
|
811 | Err(_) => { | |
757 | buf.clear(); |
|
812 | buf.clear(); | |
758 | zstd::stream::copy_decode(self.bytes, &mut buf) |
|
813 | zstd::stream::copy_decode(self.bytes, &mut buf) | |
759 | .map_err(|e| corrupted(e.to_string()))?; |
|
814 | .map_err(|e| corrupted(e.to_string()))?; | |
760 | } |
|
815 | } | |
761 | }; |
|
816 | }; | |
762 | Ok(buf) |
|
817 | Ok(buf) | |
763 | } else { |
|
818 | } else { | |
764 | let mut buf = Vec::with_capacity(cap); |
|
819 | let mut buf = Vec::with_capacity(cap); | |
765 | let len = zstd_decompress_to_buffer(self.bytes, &mut buf) |
|
820 | let len = zstd_decompress_to_buffer(self.bytes, &mut buf) | |
766 | .map_err(|e| corrupted(e.to_string()))?; |
|
821 | .map_err(|e| corrupted(e.to_string()))?; | |
767 | if len != self.uncompressed_len as usize { |
|
822 | if len != self.uncompressed_len as usize { | |
768 | Err(corrupted("uncompressed length does not match")) |
|
823 | Err(corrupted("uncompressed length does not match")) | |
769 | } else { |
|
824 | } else { | |
770 | Ok(buf) |
|
825 | Ok(buf) | |
771 | } |
|
826 | } | |
772 | } |
|
827 | } | |
773 | } |
|
828 | } | |
774 |
|
829 | |||
775 | /// Tell if the entry is a snapshot or a delta |
|
830 | /// Tell if the entry is a snapshot or a delta | |
776 | /// (influences on decompression). |
|
831 | /// (influences on decompression). | |
777 | fn is_delta(&self) -> bool { |
|
832 | fn is_delta(&self) -> bool { | |
778 | self.base_rev_or_base_of_delta_chain.is_some() |
|
833 | self.base_rev_or_base_of_delta_chain.is_some() | |
779 | } |
|
834 | } | |
780 | } |
|
835 | } | |
781 |
|
836 | |||
782 | /// Calculate the hash of a revision given its data and its parents. |
|
837 | /// Calculate the hash of a revision given its data and its parents. | |
783 | fn hash( |
|
838 | fn hash( | |
784 | data: &[u8], |
|
839 | data: &[u8], | |
785 | p1_hash: &[u8], |
|
840 | p1_hash: &[u8], | |
786 | p2_hash: &[u8], |
|
841 | p2_hash: &[u8], | |
787 | ) -> [u8; NODE_BYTES_LENGTH] { |
|
842 | ) -> [u8; NODE_BYTES_LENGTH] { | |
788 | let mut hasher = Sha1::new(); |
|
843 | let mut hasher = Sha1::new(); | |
789 | let (a, b) = (p1_hash, p2_hash); |
|
844 | let (a, b) = (p1_hash, p2_hash); | |
790 | if a > b { |
|
845 | if a > b { | |
791 | hasher.update(b); |
|
846 | hasher.update(b); | |
792 | hasher.update(a); |
|
847 | hasher.update(a); | |
793 | } else { |
|
848 | } else { | |
794 | hasher.update(a); |
|
849 | hasher.update(a); | |
795 | hasher.update(b); |
|
850 | hasher.update(b); | |
796 | } |
|
851 | } | |
797 | hasher.update(data); |
|
852 | hasher.update(data); | |
798 | *hasher.finalize().as_ref() |
|
853 | *hasher.finalize().as_ref() | |
799 | } |
|
854 | } | |
800 |
|
855 | |||
801 | #[cfg(test)] |
|
856 | #[cfg(test)] | |
802 | mod tests { |
|
857 | mod tests { | |
803 | use super::*; |
|
858 | use super::*; | |
804 | use crate::index::{IndexEntryBuilder, INDEX_ENTRY_SIZE}; |
|
859 | use crate::index::{IndexEntryBuilder, INDEX_ENTRY_SIZE}; | |
805 | use itertools::Itertools; |
|
860 | use itertools::Itertools; | |
806 |
|
861 | |||
807 | #[test] |
|
862 | #[test] | |
808 | fn test_empty() { |
|
863 | fn test_empty() { | |
809 | let temp = tempfile::tempdir().unwrap(); |
|
864 | let temp = tempfile::tempdir().unwrap(); | |
810 | let vfs = Vfs { base: temp.path() }; |
|
865 | let vfs = Vfs { base: temp.path() }; | |
811 | std::fs::write(temp.path().join("foo.i"), b"").unwrap(); |
|
866 | std::fs::write(temp.path().join("foo.i"), b"").unwrap(); | |
812 | let revlog = Revlog::open(&vfs, "foo.i", None, false).unwrap(); |
|
867 | let revlog = | |
|
868 | Revlog::open(&vfs, "foo.i", None, RevlogOpenOptions::new()) | |||
|
869 | .unwrap(); | |||
813 | assert!(revlog.is_empty()); |
|
870 | assert!(revlog.is_empty()); | |
814 | assert_eq!(revlog.len(), 0); |
|
871 | assert_eq!(revlog.len(), 0); | |
815 | assert!(revlog.get_entry(0.into()).is_err()); |
|
872 | assert!(revlog.get_entry(0.into()).is_err()); | |
816 | assert!(!revlog.has_rev(0.into())); |
|
873 | assert!(!revlog.has_rev(0.into())); | |
817 | assert_eq!( |
|
874 | assert_eq!( | |
818 | revlog.rev_from_node(NULL_NODE.into()).unwrap(), |
|
875 | revlog.rev_from_node(NULL_NODE.into()).unwrap(), | |
819 | NULL_REVISION |
|
876 | NULL_REVISION | |
820 | ); |
|
877 | ); | |
821 | let null_entry = revlog.get_entry(NULL_REVISION.into()).ok().unwrap(); |
|
878 | let null_entry = revlog.get_entry(NULL_REVISION.into()).ok().unwrap(); | |
822 | assert_eq!(null_entry.revision(), NULL_REVISION); |
|
879 | assert_eq!(null_entry.revision(), NULL_REVISION); | |
823 | assert!(null_entry.data().unwrap().is_empty()); |
|
880 | assert!(null_entry.data().unwrap().is_empty()); | |
824 | } |
|
881 | } | |
825 |
|
882 | |||
826 | #[test] |
|
883 | #[test] | |
827 | fn test_inline() { |
|
884 | fn test_inline() { | |
828 | let temp = tempfile::tempdir().unwrap(); |
|
885 | let temp = tempfile::tempdir().unwrap(); | |
829 | let vfs = Vfs { base: temp.path() }; |
|
886 | let vfs = Vfs { base: temp.path() }; | |
830 | let node0 = Node::from_hex("2ed2a3912a0b24502043eae84ee4b279c18b90dd") |
|
887 | let node0 = Node::from_hex("2ed2a3912a0b24502043eae84ee4b279c18b90dd") | |
831 | .unwrap(); |
|
888 | .unwrap(); | |
832 | let node1 = Node::from_hex("b004912a8510032a0350a74daa2803dadfb00e12") |
|
889 | let node1 = Node::from_hex("b004912a8510032a0350a74daa2803dadfb00e12") | |
833 | .unwrap(); |
|
890 | .unwrap(); | |
834 | let node2 = Node::from_hex("dd6ad206e907be60927b5a3117b97dffb2590582") |
|
891 | let node2 = Node::from_hex("dd6ad206e907be60927b5a3117b97dffb2590582") | |
835 | .unwrap(); |
|
892 | .unwrap(); | |
836 | let entry0_bytes = IndexEntryBuilder::new() |
|
893 | let entry0_bytes = IndexEntryBuilder::new() | |
837 | .is_first(true) |
|
894 | .is_first(true) | |
838 | .with_version(1) |
|
895 | .with_version(1) | |
839 | .with_inline(true) |
|
896 | .with_inline(true) | |
840 | .with_offset(INDEX_ENTRY_SIZE) |
|
897 | .with_offset(INDEX_ENTRY_SIZE) | |
841 | .with_node(node0) |
|
898 | .with_node(node0) | |
842 | .build(); |
|
899 | .build(); | |
843 | let entry1_bytes = IndexEntryBuilder::new() |
|
900 | let entry1_bytes = IndexEntryBuilder::new() | |
844 | .with_offset(INDEX_ENTRY_SIZE) |
|
901 | .with_offset(INDEX_ENTRY_SIZE) | |
845 | .with_node(node1) |
|
902 | .with_node(node1) | |
846 | .build(); |
|
903 | .build(); | |
847 | let entry2_bytes = IndexEntryBuilder::new() |
|
904 | let entry2_bytes = IndexEntryBuilder::new() | |
848 | .with_offset(INDEX_ENTRY_SIZE) |
|
905 | .with_offset(INDEX_ENTRY_SIZE) | |
849 | .with_p1(Revision(0)) |
|
906 | .with_p1(Revision(0)) | |
850 | .with_p2(Revision(1)) |
|
907 | .with_p2(Revision(1)) | |
851 | .with_node(node2) |
|
908 | .with_node(node2) | |
852 | .build(); |
|
909 | .build(); | |
853 | let contents = vec![entry0_bytes, entry1_bytes, entry2_bytes] |
|
910 | let contents = vec![entry0_bytes, entry1_bytes, entry2_bytes] | |
854 | .into_iter() |
|
911 | .into_iter() | |
855 | .flatten() |
|
912 | .flatten() | |
856 | .collect_vec(); |
|
913 | .collect_vec(); | |
857 | std::fs::write(temp.path().join("foo.i"), contents).unwrap(); |
|
914 | std::fs::write(temp.path().join("foo.i"), contents).unwrap(); | |
858 | let revlog = Revlog::open(&vfs, "foo.i", None, false).unwrap(); |
|
915 | let revlog = | |
|
916 | Revlog::open(&vfs, "foo.i", None, RevlogOpenOptions::new()) | |||
|
917 | .unwrap(); | |||
859 |
|
918 | |||
860 | let entry0 = revlog.get_entry(0.into()).ok().unwrap(); |
|
919 | let entry0 = revlog.get_entry(0.into()).ok().unwrap(); | |
861 | assert_eq!(entry0.revision(), Revision(0)); |
|
920 | assert_eq!(entry0.revision(), Revision(0)); | |
862 | assert_eq!(*entry0.node(), node0); |
|
921 | assert_eq!(*entry0.node(), node0); | |
863 | assert!(!entry0.has_p1()); |
|
922 | assert!(!entry0.has_p1()); | |
864 | assert_eq!(entry0.p1(), None); |
|
923 | assert_eq!(entry0.p1(), None); | |
865 | assert_eq!(entry0.p2(), None); |
|
924 | assert_eq!(entry0.p2(), None); | |
866 | let p1_entry = entry0.p1_entry().unwrap(); |
|
925 | let p1_entry = entry0.p1_entry().unwrap(); | |
867 | assert!(p1_entry.is_none()); |
|
926 | assert!(p1_entry.is_none()); | |
868 | let p2_entry = entry0.p2_entry().unwrap(); |
|
927 | let p2_entry = entry0.p2_entry().unwrap(); | |
869 | assert!(p2_entry.is_none()); |
|
928 | assert!(p2_entry.is_none()); | |
870 |
|
929 | |||
871 | let entry1 = revlog.get_entry(1.into()).ok().unwrap(); |
|
930 | let entry1 = revlog.get_entry(1.into()).ok().unwrap(); | |
872 | assert_eq!(entry1.revision(), Revision(1)); |
|
931 | assert_eq!(entry1.revision(), Revision(1)); | |
873 | assert_eq!(*entry1.node(), node1); |
|
932 | assert_eq!(*entry1.node(), node1); | |
874 | assert!(!entry1.has_p1()); |
|
933 | assert!(!entry1.has_p1()); | |
875 | assert_eq!(entry1.p1(), None); |
|
934 | assert_eq!(entry1.p1(), None); | |
876 | assert_eq!(entry1.p2(), None); |
|
935 | assert_eq!(entry1.p2(), None); | |
877 | let p1_entry = entry1.p1_entry().unwrap(); |
|
936 | let p1_entry = entry1.p1_entry().unwrap(); | |
878 | assert!(p1_entry.is_none()); |
|
937 | assert!(p1_entry.is_none()); | |
879 | let p2_entry = entry1.p2_entry().unwrap(); |
|
938 | let p2_entry = entry1.p2_entry().unwrap(); | |
880 | assert!(p2_entry.is_none()); |
|
939 | assert!(p2_entry.is_none()); | |
881 |
|
940 | |||
882 | let entry2 = revlog.get_entry(2.into()).ok().unwrap(); |
|
941 | let entry2 = revlog.get_entry(2.into()).ok().unwrap(); | |
883 | assert_eq!(entry2.revision(), Revision(2)); |
|
942 | assert_eq!(entry2.revision(), Revision(2)); | |
884 | assert_eq!(*entry2.node(), node2); |
|
943 | assert_eq!(*entry2.node(), node2); | |
885 | assert!(entry2.has_p1()); |
|
944 | assert!(entry2.has_p1()); | |
886 | assert_eq!(entry2.p1(), Some(Revision(0))); |
|
945 | assert_eq!(entry2.p1(), Some(Revision(0))); | |
887 | assert_eq!(entry2.p2(), Some(Revision(1))); |
|
946 | assert_eq!(entry2.p2(), Some(Revision(1))); | |
888 | let p1_entry = entry2.p1_entry().unwrap(); |
|
947 | let p1_entry = entry2.p1_entry().unwrap(); | |
889 | assert!(p1_entry.is_some()); |
|
948 | assert!(p1_entry.is_some()); | |
890 | assert_eq!(p1_entry.unwrap().revision(), Revision(0)); |
|
949 | assert_eq!(p1_entry.unwrap().revision(), Revision(0)); | |
891 | let p2_entry = entry2.p2_entry().unwrap(); |
|
950 | let p2_entry = entry2.p2_entry().unwrap(); | |
892 | assert!(p2_entry.is_some()); |
|
951 | assert!(p2_entry.is_some()); | |
893 | assert_eq!(p2_entry.unwrap().revision(), Revision(1)); |
|
952 | assert_eq!(p2_entry.unwrap().revision(), Revision(1)); | |
894 | } |
|
953 | } | |
895 |
|
954 | |||
896 | #[test] |
|
955 | #[test] | |
897 | fn test_nodemap() { |
|
956 | fn test_nodemap() { | |
898 | let temp = tempfile::tempdir().unwrap(); |
|
957 | let temp = tempfile::tempdir().unwrap(); | |
899 | let vfs = Vfs { base: temp.path() }; |
|
958 | let vfs = Vfs { base: temp.path() }; | |
900 |
|
959 | |||
901 | // building a revlog with a forced Node starting with zeros |
|
960 | // building a revlog with a forced Node starting with zeros | |
902 | // This is a corruption, but it does not preclude using the nodemap |
|
961 | // This is a corruption, but it does not preclude using the nodemap | |
903 | // if we don't try and access the data |
|
962 | // if we don't try and access the data | |
904 | let node0 = Node::from_hex("00d2a3912a0b24502043eae84ee4b279c18b90dd") |
|
963 | let node0 = Node::from_hex("00d2a3912a0b24502043eae84ee4b279c18b90dd") | |
905 | .unwrap(); |
|
964 | .unwrap(); | |
906 | let node1 = Node::from_hex("b004912a8510032a0350a74daa2803dadfb00e12") |
|
965 | let node1 = Node::from_hex("b004912a8510032a0350a74daa2803dadfb00e12") | |
907 | .unwrap(); |
|
966 | .unwrap(); | |
908 | let entry0_bytes = IndexEntryBuilder::new() |
|
967 | let entry0_bytes = IndexEntryBuilder::new() | |
909 | .is_first(true) |
|
968 | .is_first(true) | |
910 | .with_version(1) |
|
969 | .with_version(1) | |
911 | .with_inline(true) |
|
970 | .with_inline(true) | |
912 | .with_offset(INDEX_ENTRY_SIZE) |
|
971 | .with_offset(INDEX_ENTRY_SIZE) | |
913 | .with_node(node0) |
|
972 | .with_node(node0) | |
914 | .build(); |
|
973 | .build(); | |
915 | let entry1_bytes = IndexEntryBuilder::new() |
|
974 | let entry1_bytes = IndexEntryBuilder::new() | |
916 | .with_offset(INDEX_ENTRY_SIZE) |
|
975 | .with_offset(INDEX_ENTRY_SIZE) | |
917 | .with_node(node1) |
|
976 | .with_node(node1) | |
918 | .build(); |
|
977 | .build(); | |
919 | let contents = vec![entry0_bytes, entry1_bytes] |
|
978 | let contents = vec![entry0_bytes, entry1_bytes] | |
920 | .into_iter() |
|
979 | .into_iter() | |
921 | .flatten() |
|
980 | .flatten() | |
922 | .collect_vec(); |
|
981 | .collect_vec(); | |
923 | std::fs::write(temp.path().join("foo.i"), contents).unwrap(); |
|
982 | std::fs::write(temp.path().join("foo.i"), contents).unwrap(); | |
924 |
|
983 | |||
925 | let mut idx = nodemap::tests::TestNtIndex::new(); |
|
984 | let mut idx = nodemap::tests::TestNtIndex::new(); | |
926 | idx.insert_node(Revision(0), node0).unwrap(); |
|
985 | idx.insert_node(Revision(0), node0).unwrap(); | |
927 | idx.insert_node(Revision(1), node1).unwrap(); |
|
986 | idx.insert_node(Revision(1), node1).unwrap(); | |
928 |
|
987 | |||
929 | let revlog = |
|
988 | let revlog = Revlog::open_gen( | |
930 | Revlog::open_gen(&vfs, "foo.i", None, true, Some(idx.nt)).unwrap(); |
|
989 | &vfs, | |
|
990 | "foo.i", | |||
|
991 | None, | |||
|
992 | RevlogOpenOptions::new(), | |||
|
993 | Some(idx.nt), | |||
|
994 | ) | |||
|
995 | .unwrap(); | |||
931 |
|
996 | |||
932 | // accessing the data shows the corruption |
|
997 | // accessing the data shows the corruption | |
933 | revlog.get_entry(0.into()).unwrap().data().unwrap_err(); |
|
998 | revlog.get_entry(0.into()).unwrap().data().unwrap_err(); | |
934 |
|
999 | |||
935 | assert_eq!( |
|
1000 | assert_eq!( | |
936 | revlog.rev_from_node(NULL_NODE.into()).unwrap(), |
|
1001 | revlog.rev_from_node(NULL_NODE.into()).unwrap(), | |
937 | Revision(-1) |
|
1002 | Revision(-1) | |
938 | ); |
|
1003 | ); | |
939 | assert_eq!(revlog.rev_from_node(node0.into()).unwrap(), Revision(0)); |
|
1004 | assert_eq!(revlog.rev_from_node(node0.into()).unwrap(), Revision(0)); | |
940 | assert_eq!(revlog.rev_from_node(node1.into()).unwrap(), Revision(1)); |
|
1005 | assert_eq!(revlog.rev_from_node(node1.into()).unwrap(), Revision(1)); | |
941 | assert_eq!( |
|
1006 | assert_eq!( | |
942 | revlog |
|
1007 | revlog | |
943 | .rev_from_node(NodePrefix::from_hex("000").unwrap()) |
|
1008 | .rev_from_node(NodePrefix::from_hex("000").unwrap()) | |
944 | .unwrap(), |
|
1009 | .unwrap(), | |
945 | Revision(-1) |
|
1010 | Revision(-1) | |
946 | ); |
|
1011 | ); | |
947 | assert_eq!( |
|
1012 | assert_eq!( | |
948 | revlog |
|
1013 | revlog | |
949 | .rev_from_node(NodePrefix::from_hex("b00").unwrap()) |
|
1014 | .rev_from_node(NodePrefix::from_hex("b00").unwrap()) | |
950 | .unwrap(), |
|
1015 | .unwrap(), | |
951 | Revision(1) |
|
1016 | Revision(1) | |
952 | ); |
|
1017 | ); | |
953 | // RevlogError does not implement PartialEq |
|
1018 | // RevlogError does not implement PartialEq | |
954 | // (ultimately because io::Error does not) |
|
1019 | // (ultimately because io::Error does not) | |
955 | match revlog |
|
1020 | match revlog | |
956 | .rev_from_node(NodePrefix::from_hex("00").unwrap()) |
|
1021 | .rev_from_node(NodePrefix::from_hex("00").unwrap()) | |
957 | .expect_err("Expected to give AmbiguousPrefix error") |
|
1022 | .expect_err("Expected to give AmbiguousPrefix error") | |
958 | { |
|
1023 | { | |
959 | RevlogError::AmbiguousPrefix => (), |
|
1024 | RevlogError::AmbiguousPrefix => (), | |
960 | e => { |
|
1025 | e => { | |
961 | panic!("Got another error than AmbiguousPrefix: {:?}", e); |
|
1026 | panic!("Got another error than AmbiguousPrefix: {:?}", e); | |
962 | } |
|
1027 | } | |
963 | }; |
|
1028 | }; | |
964 | } |
|
1029 | } | |
965 | } |
|
1030 | } |
@@ -1,561 +1,572 b'' | |||||
1 | // revlog.rs |
|
1 | // revlog.rs | |
2 | // |
|
2 | // | |
3 | // Copyright 2019-2020 Georges Racinet <georges.racinet@octobus.net> |
|
3 | // Copyright 2019-2020 Georges Racinet <georges.racinet@octobus.net> | |
4 | // |
|
4 | // | |
5 | // This software may be used and distributed according to the terms of the |
|
5 | // This software may be used and distributed according to the terms of the | |
6 | // GNU General Public License version 2 or any later version. |
|
6 | // GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | use crate::{ |
|
8 | use crate::{ | |
9 | cindex, |
|
9 | cindex, | |
10 | utils::{node_from_py_bytes, node_from_py_object}, |
|
10 | utils::{node_from_py_bytes, node_from_py_object}, | |
11 | PyRevision, |
|
11 | PyRevision, | |
12 | }; |
|
12 | }; | |
13 | use cpython::{ |
|
13 | use cpython::{ | |
14 | buffer::{Element, PyBuffer}, |
|
14 | buffer::{Element, PyBuffer}, | |
15 | exc::{IndexError, ValueError}, |
|
15 | exc::{IndexError, ValueError}, | |
16 | ObjectProtocol, PyBytes, PyClone, PyDict, PyErr, PyInt, PyModule, |
|
16 | ObjectProtocol, PyBytes, PyClone, PyDict, PyErr, PyInt, PyModule, | |
17 | PyObject, PyResult, PyString, PyTuple, Python, PythonObject, ToPyObject, |
|
17 | PyObject, PyResult, PyString, PyTuple, Python, PythonObject, ToPyObject, | |
18 | }; |
|
18 | }; | |
19 | use hg::{ |
|
19 | use hg::{ | |
|
20 | index::IndexHeader, | |||
20 | nodemap::{Block, NodeMapError, NodeTree}, |
|
21 | nodemap::{Block, NodeMapError, NodeTree}, | |
21 | revlog::{nodemap::NodeMap, NodePrefix, RevlogIndex}, |
|
22 | revlog::{nodemap::NodeMap, NodePrefix, RevlogIndex}, | |
22 | BaseRevision, Revision, UncheckedRevision, |
|
23 | BaseRevision, Revision, UncheckedRevision, | |
23 | }; |
|
24 | }; | |
24 | use std::cell::RefCell; |
|
25 | use std::cell::RefCell; | |
25 |
|
26 | |||
26 | /// Return a Struct implementing the Graph trait |
|
27 | /// Return a Struct implementing the Graph trait | |
27 | pub(crate) fn pyindex_to_graph( |
|
28 | pub(crate) fn pyindex_to_graph( | |
28 | py: Python, |
|
29 | py: Python, | |
29 | index: PyObject, |
|
30 | index: PyObject, | |
30 | ) -> PyResult<cindex::Index> { |
|
31 | ) -> PyResult<cindex::Index> { | |
31 | match index.extract::<MixedIndex>(py) { |
|
32 | match index.extract::<MixedIndex>(py) { | |
32 | Ok(midx) => Ok(midx.clone_cindex(py)), |
|
33 | Ok(midx) => Ok(midx.clone_cindex(py)), | |
33 | Err(_) => cindex::Index::new(py, index), |
|
34 | Err(_) => cindex::Index::new(py, index), | |
34 | } |
|
35 | } | |
35 | } |
|
36 | } | |
36 |
|
37 | |||
37 | py_class!(pub class MixedIndex |py| { |
|
38 | py_class!(pub class MixedIndex |py| { | |
38 | data cindex: RefCell<cindex::Index>; |
|
39 | data cindex: RefCell<cindex::Index>; | |
39 | data index: RefCell<hg::index::Index>; |
|
40 | data index: RefCell<hg::index::Index>; | |
40 | data nt: RefCell<Option<NodeTree>>; |
|
41 | data nt: RefCell<Option<NodeTree>>; | |
41 | data docket: RefCell<Option<PyObject>>; |
|
42 | data docket: RefCell<Option<PyObject>>; | |
42 | // Holds a reference to the mmap'ed persistent nodemap data |
|
43 | // Holds a reference to the mmap'ed persistent nodemap data | |
43 | data nodemap_mmap: RefCell<Option<PyBuffer>>; |
|
44 | data nodemap_mmap: RefCell<Option<PyBuffer>>; | |
44 | // Holds a reference to the mmap'ed persistent index data |
|
45 | // Holds a reference to the mmap'ed persistent index data | |
45 | data index_mmap: RefCell<Option<PyBuffer>>; |
|
46 | data index_mmap: RefCell<Option<PyBuffer>>; | |
46 |
|
47 | |||
47 | def __new__( |
|
48 | def __new__( | |
48 | _cls, |
|
49 | _cls, | |
49 | cindex: PyObject, |
|
50 | cindex: PyObject, | |
50 | data: PyObject |
|
51 | data: PyObject, | |
|
52 | default_header: u32, | |||
51 | ) -> PyResult<MixedIndex> { |
|
53 | ) -> PyResult<MixedIndex> { | |
52 | Self::new(py, cindex, data) |
|
54 | Self::new(py, cindex, data, default_header) | |
53 | } |
|
55 | } | |
54 |
|
56 | |||
55 | /// Compatibility layer used for Python consumers needing access to the C index |
|
57 | /// Compatibility layer used for Python consumers needing access to the C index | |
56 | /// |
|
58 | /// | |
57 | /// Only use case so far is `scmutil.shortesthexnodeidprefix`, |
|
59 | /// Only use case so far is `scmutil.shortesthexnodeidprefix`, | |
58 | /// that may need to build a custom `nodetree`, based on a specified revset. |
|
60 | /// that may need to build a custom `nodetree`, based on a specified revset. | |
59 | /// With a Rust implementation of the nodemap, we will be able to get rid of |
|
61 | /// With a Rust implementation of the nodemap, we will be able to get rid of | |
60 | /// this, by exposing our own standalone nodemap class, |
|
62 | /// this, by exposing our own standalone nodemap class, | |
61 | /// ready to accept `MixedIndex`. |
|
63 | /// ready to accept `MixedIndex`. | |
62 | def get_cindex(&self) -> PyResult<PyObject> { |
|
64 | def get_cindex(&self) -> PyResult<PyObject> { | |
63 | Ok(self.cindex(py).borrow().inner().clone_ref(py)) |
|
65 | Ok(self.cindex(py).borrow().inner().clone_ref(py)) | |
64 | } |
|
66 | } | |
65 |
|
67 | |||
66 | // Index API involving nodemap, as defined in mercurial/pure/parsers.py |
|
68 | // Index API involving nodemap, as defined in mercurial/pure/parsers.py | |
67 |
|
69 | |||
68 | /// Return Revision if found, raises a bare `error.RevlogError` |
|
70 | /// Return Revision if found, raises a bare `error.RevlogError` | |
69 | /// in case of ambiguity, same as C version does |
|
71 | /// in case of ambiguity, same as C version does | |
70 | def get_rev(&self, node: PyBytes) -> PyResult<Option<PyRevision>> { |
|
72 | def get_rev(&self, node: PyBytes) -> PyResult<Option<PyRevision>> { | |
71 | let opt = self.get_nodetree(py)?.borrow(); |
|
73 | let opt = self.get_nodetree(py)?.borrow(); | |
72 | let nt = opt.as_ref().unwrap(); |
|
74 | let nt = opt.as_ref().unwrap(); | |
73 | let idx = &*self.cindex(py).borrow(); |
|
75 | let idx = &*self.cindex(py).borrow(); | |
74 | let node = node_from_py_bytes(py, &node)?; |
|
76 | let node = node_from_py_bytes(py, &node)?; | |
75 | let res = nt.find_bin(idx, node.into()); |
|
77 | let res = nt.find_bin(idx, node.into()); | |
76 | Ok(res.map_err(|e| nodemap_error(py, e))?.map(Into::into)) |
|
78 | Ok(res.map_err(|e| nodemap_error(py, e))?.map(Into::into)) | |
77 | } |
|
79 | } | |
78 |
|
80 | |||
79 | /// same as `get_rev()` but raises a bare `error.RevlogError` if node |
|
81 | /// same as `get_rev()` but raises a bare `error.RevlogError` if node | |
80 | /// is not found. |
|
82 | /// is not found. | |
81 | /// |
|
83 | /// | |
82 | /// No need to repeat `node` in the exception, `mercurial/revlog.py` |
|
84 | /// No need to repeat `node` in the exception, `mercurial/revlog.py` | |
83 | /// will catch and rewrap with it |
|
85 | /// will catch and rewrap with it | |
84 | def rev(&self, node: PyBytes) -> PyResult<PyRevision> { |
|
86 | def rev(&self, node: PyBytes) -> PyResult<PyRevision> { | |
85 | self.get_rev(py, node)?.ok_or_else(|| revlog_error(py)) |
|
87 | self.get_rev(py, node)?.ok_or_else(|| revlog_error(py)) | |
86 | } |
|
88 | } | |
87 |
|
89 | |||
88 | /// return True if the node exist in the index |
|
90 | /// return True if the node exist in the index | |
89 | def has_node(&self, node: PyBytes) -> PyResult<bool> { |
|
91 | def has_node(&self, node: PyBytes) -> PyResult<bool> { | |
90 | self.get_rev(py, node).map(|opt| opt.is_some()) |
|
92 | self.get_rev(py, node).map(|opt| opt.is_some()) | |
91 | } |
|
93 | } | |
92 |
|
94 | |||
93 | /// find length of shortest hex nodeid of a binary ID |
|
95 | /// find length of shortest hex nodeid of a binary ID | |
94 | def shortest(&self, node: PyBytes) -> PyResult<usize> { |
|
96 | def shortest(&self, node: PyBytes) -> PyResult<usize> { | |
95 | let opt = self.get_nodetree(py)?.borrow(); |
|
97 | let opt = self.get_nodetree(py)?.borrow(); | |
96 | let nt = opt.as_ref().unwrap(); |
|
98 | let nt = opt.as_ref().unwrap(); | |
97 | let idx = &*self.cindex(py).borrow(); |
|
99 | let idx = &*self.cindex(py).borrow(); | |
98 | match nt.unique_prefix_len_node(idx, &node_from_py_bytes(py, &node)?) |
|
100 | match nt.unique_prefix_len_node(idx, &node_from_py_bytes(py, &node)?) | |
99 | { |
|
101 | { | |
100 | Ok(Some(l)) => Ok(l), |
|
102 | Ok(Some(l)) => Ok(l), | |
101 | Ok(None) => Err(revlog_error(py)), |
|
103 | Ok(None) => Err(revlog_error(py)), | |
102 | Err(e) => Err(nodemap_error(py, e)), |
|
104 | Err(e) => Err(nodemap_error(py, e)), | |
103 | } |
|
105 | } | |
104 | } |
|
106 | } | |
105 |
|
107 | |||
106 | def partialmatch(&self, node: PyObject) -> PyResult<Option<PyBytes>> { |
|
108 | def partialmatch(&self, node: PyObject) -> PyResult<Option<PyBytes>> { | |
107 | let opt = self.get_nodetree(py)?.borrow(); |
|
109 | let opt = self.get_nodetree(py)?.borrow(); | |
108 | let nt = opt.as_ref().unwrap(); |
|
110 | let nt = opt.as_ref().unwrap(); | |
109 | let idx = &*self.cindex(py).borrow(); |
|
111 | let idx = &*self.cindex(py).borrow(); | |
110 |
|
112 | |||
111 | let node_as_string = if cfg!(feature = "python3-sys") { |
|
113 | let node_as_string = if cfg!(feature = "python3-sys") { | |
112 | node.cast_as::<PyString>(py)?.to_string(py)?.to_string() |
|
114 | node.cast_as::<PyString>(py)?.to_string(py)?.to_string() | |
113 | } |
|
115 | } | |
114 | else { |
|
116 | else { | |
115 | let node = node.extract::<PyBytes>(py)?; |
|
117 | let node = node.extract::<PyBytes>(py)?; | |
116 | String::from_utf8_lossy(node.data(py)).to_string() |
|
118 | String::from_utf8_lossy(node.data(py)).to_string() | |
117 | }; |
|
119 | }; | |
118 |
|
120 | |||
119 | let prefix = NodePrefix::from_hex(&node_as_string) |
|
121 | let prefix = NodePrefix::from_hex(&node_as_string) | |
120 | .map_err(|_| PyErr::new::<ValueError, _>( |
|
122 | .map_err(|_| PyErr::new::<ValueError, _>( | |
121 | py, format!("Invalid node or prefix '{}'", node_as_string)) |
|
123 | py, format!("Invalid node or prefix '{}'", node_as_string)) | |
122 | )?; |
|
124 | )?; | |
123 |
|
125 | |||
124 | nt.find_bin(idx, prefix) |
|
126 | nt.find_bin(idx, prefix) | |
125 | // TODO make an inner API returning the node directly |
|
127 | // TODO make an inner API returning the node directly | |
126 | .map(|opt| opt.map( |
|
128 | .map(|opt| opt.map( | |
127 | |rev| PyBytes::new(py, idx.node(rev).unwrap().as_bytes()))) |
|
129 | |rev| PyBytes::new(py, idx.node(rev).unwrap().as_bytes()))) | |
128 | .map_err(|e| nodemap_error(py, e)) |
|
130 | .map_err(|e| nodemap_error(py, e)) | |
129 |
|
131 | |||
130 | } |
|
132 | } | |
131 |
|
133 | |||
132 | /// append an index entry |
|
134 | /// append an index entry | |
133 | def append(&self, tup: PyTuple) -> PyResult<PyObject> { |
|
135 | def append(&self, tup: PyTuple) -> PyResult<PyObject> { | |
134 | if tup.len(py) < 8 { |
|
136 | if tup.len(py) < 8 { | |
135 | // this is better than the panic promised by tup.get_item() |
|
137 | // this is better than the panic promised by tup.get_item() | |
136 | return Err( |
|
138 | return Err( | |
137 | PyErr::new::<IndexError, _>(py, "tuple index out of range")) |
|
139 | PyErr::new::<IndexError, _>(py, "tuple index out of range")) | |
138 | } |
|
140 | } | |
139 | let node_bytes = tup.get_item(py, 7).extract(py)?; |
|
141 | let node_bytes = tup.get_item(py, 7).extract(py)?; | |
140 | let node = node_from_py_object(py, &node_bytes)?; |
|
142 | let node = node_from_py_object(py, &node_bytes)?; | |
141 |
|
143 | |||
142 | let mut idx = self.cindex(py).borrow_mut(); |
|
144 | let mut idx = self.cindex(py).borrow_mut(); | |
143 |
|
145 | |||
144 | // This is ok since we will just add the revision to the index |
|
146 | // This is ok since we will just add the revision to the index | |
145 | let rev = Revision(idx.len() as BaseRevision); |
|
147 | let rev = Revision(idx.len() as BaseRevision); | |
146 | idx.append(py, tup)?; |
|
148 | idx.append(py, tup)?; | |
147 |
|
149 | |||
148 | self.get_nodetree(py)?.borrow_mut().as_mut().unwrap() |
|
150 | self.get_nodetree(py)?.borrow_mut().as_mut().unwrap() | |
149 | .insert(&*idx, &node, rev) |
|
151 | .insert(&*idx, &node, rev) | |
150 | .map_err(|e| nodemap_error(py, e))?; |
|
152 | .map_err(|e| nodemap_error(py, e))?; | |
151 | Ok(py.None()) |
|
153 | Ok(py.None()) | |
152 | } |
|
154 | } | |
153 |
|
155 | |||
154 | def __delitem__(&self, key: PyObject) -> PyResult<()> { |
|
156 | def __delitem__(&self, key: PyObject) -> PyResult<()> { | |
155 | // __delitem__ is both for `del idx[r]` and `del idx[r1:r2]` |
|
157 | // __delitem__ is both for `del idx[r]` and `del idx[r1:r2]` | |
156 | self.cindex(py).borrow().inner().del_item(py, key)?; |
|
158 | self.cindex(py).borrow().inner().del_item(py, key)?; | |
157 | let mut opt = self.get_nodetree(py)?.borrow_mut(); |
|
159 | let mut opt = self.get_nodetree(py)?.borrow_mut(); | |
158 | let nt = opt.as_mut().unwrap(); |
|
160 | let nt = opt.as_mut().unwrap(); | |
159 | nt.invalidate_all(); |
|
161 | nt.invalidate_all(); | |
160 | self.fill_nodemap(py, nt)?; |
|
162 | self.fill_nodemap(py, nt)?; | |
161 | Ok(()) |
|
163 | Ok(()) | |
162 | } |
|
164 | } | |
163 |
|
165 | |||
164 | // |
|
166 | // | |
165 | // Reforwarded C index API |
|
167 | // Reforwarded C index API | |
166 | // |
|
168 | // | |
167 |
|
169 | |||
168 | // index_methods (tp_methods). Same ordering as in revlog.c |
|
170 | // index_methods (tp_methods). Same ordering as in revlog.c | |
169 |
|
171 | |||
170 | /// return the gca set of the given revs |
|
172 | /// return the gca set of the given revs | |
171 | def ancestors(&self, *args, **kw) -> PyResult<PyObject> { |
|
173 | def ancestors(&self, *args, **kw) -> PyResult<PyObject> { | |
172 | self.call_cindex(py, "ancestors", args, kw) |
|
174 | self.call_cindex(py, "ancestors", args, kw) | |
173 | } |
|
175 | } | |
174 |
|
176 | |||
175 | /// return the heads of the common ancestors of the given revs |
|
177 | /// return the heads of the common ancestors of the given revs | |
176 | def commonancestorsheads(&self, *args, **kw) -> PyResult<PyObject> { |
|
178 | def commonancestorsheads(&self, *args, **kw) -> PyResult<PyObject> { | |
177 | self.call_cindex(py, "commonancestorsheads", args, kw) |
|
179 | self.call_cindex(py, "commonancestorsheads", args, kw) | |
178 | } |
|
180 | } | |
179 |
|
181 | |||
180 | /// Clear the index caches and inner py_class data. |
|
182 | /// Clear the index caches and inner py_class data. | |
181 | /// It is Python's responsibility to call `update_nodemap_data` again. |
|
183 | /// It is Python's responsibility to call `update_nodemap_data` again. | |
182 | def clearcaches(&self, *args, **kw) -> PyResult<PyObject> { |
|
184 | def clearcaches(&self, *args, **kw) -> PyResult<PyObject> { | |
183 | self.nt(py).borrow_mut().take(); |
|
185 | self.nt(py).borrow_mut().take(); | |
184 | self.docket(py).borrow_mut().take(); |
|
186 | self.docket(py).borrow_mut().take(); | |
185 | self.nodemap_mmap(py).borrow_mut().take(); |
|
187 | self.nodemap_mmap(py).borrow_mut().take(); | |
186 | self.call_cindex(py, "clearcaches", args, kw) |
|
188 | self.call_cindex(py, "clearcaches", args, kw) | |
187 | } |
|
189 | } | |
188 |
|
190 | |||
189 | /// return the raw binary string representing a revision |
|
191 | /// return the raw binary string representing a revision | |
190 | def entry_binary(&self, *args, **kw) -> PyResult<PyObject> { |
|
192 | def entry_binary(&self, *args, **kw) -> PyResult<PyObject> { | |
191 | self.call_cindex(py, "entry_binary", args, kw) |
|
193 | self.call_cindex(py, "entry_binary", args, kw) | |
192 | } |
|
194 | } | |
193 |
|
195 | |||
194 | /// return a binary packed version of the header |
|
196 | /// return a binary packed version of the header | |
195 | def pack_header(&self, *args, **kw) -> PyResult<PyObject> { |
|
197 | def pack_header(&self, *args, **kw) -> PyResult<PyObject> { | |
196 | self.call_cindex(py, "pack_header", args, kw) |
|
198 | self.call_cindex(py, "pack_header", args, kw) | |
197 | } |
|
199 | } | |
198 |
|
200 | |||
199 | /// get an index entry |
|
201 | /// get an index entry | |
200 | def get(&self, *args, **kw) -> PyResult<PyObject> { |
|
202 | def get(&self, *args, **kw) -> PyResult<PyObject> { | |
201 | self.call_cindex(py, "get", args, kw) |
|
203 | self.call_cindex(py, "get", args, kw) | |
202 | } |
|
204 | } | |
203 |
|
205 | |||
204 | /// compute phases |
|
206 | /// compute phases | |
205 | def computephasesmapsets(&self, *args, **kw) -> PyResult<PyObject> { |
|
207 | def computephasesmapsets(&self, *args, **kw) -> PyResult<PyObject> { | |
206 | self.call_cindex(py, "computephasesmapsets", args, kw) |
|
208 | self.call_cindex(py, "computephasesmapsets", args, kw) | |
207 | } |
|
209 | } | |
208 |
|
210 | |||
209 | /// reachableroots |
|
211 | /// reachableroots | |
210 | def reachableroots2(&self, *args, **kw) -> PyResult<PyObject> { |
|
212 | def reachableroots2(&self, *args, **kw) -> PyResult<PyObject> { | |
211 | self.call_cindex(py, "reachableroots2", args, kw) |
|
213 | self.call_cindex(py, "reachableroots2", args, kw) | |
212 | } |
|
214 | } | |
213 |
|
215 | |||
214 | /// get head revisions |
|
216 | /// get head revisions | |
215 | def headrevs(&self, *args, **kw) -> PyResult<PyObject> { |
|
217 | def headrevs(&self, *args, **kw) -> PyResult<PyObject> { | |
216 | self.call_cindex(py, "headrevs", args, kw) |
|
218 | self.call_cindex(py, "headrevs", args, kw) | |
217 | } |
|
219 | } | |
218 |
|
220 | |||
219 | /// get filtered head revisions |
|
221 | /// get filtered head revisions | |
220 | def headrevsfiltered(&self, *args, **kw) -> PyResult<PyObject> { |
|
222 | def headrevsfiltered(&self, *args, **kw) -> PyResult<PyObject> { | |
221 | self.call_cindex(py, "headrevsfiltered", args, kw) |
|
223 | self.call_cindex(py, "headrevsfiltered", args, kw) | |
222 | } |
|
224 | } | |
223 |
|
225 | |||
224 | /// True if the object is a snapshot |
|
226 | /// True if the object is a snapshot | |
225 | def issnapshot(&self, *args, **kw) -> PyResult<PyObject> { |
|
227 | def issnapshot(&self, *args, **kw) -> PyResult<PyObject> { | |
226 | self.call_cindex(py, "issnapshot", args, kw) |
|
228 | self.call_cindex(py, "issnapshot", args, kw) | |
227 | } |
|
229 | } | |
228 |
|
230 | |||
229 | /// Gather snapshot data in a cache dict |
|
231 | /// Gather snapshot data in a cache dict | |
230 | def findsnapshots(&self, *args, **kw) -> PyResult<PyObject> { |
|
232 | def findsnapshots(&self, *args, **kw) -> PyResult<PyObject> { | |
231 | self.call_cindex(py, "findsnapshots", args, kw) |
|
233 | self.call_cindex(py, "findsnapshots", args, kw) | |
232 | } |
|
234 | } | |
233 |
|
235 | |||
234 | /// determine revisions with deltas to reconstruct fulltext |
|
236 | /// determine revisions with deltas to reconstruct fulltext | |
235 | def deltachain(&self, *args, **kw) -> PyResult<PyObject> { |
|
237 | def deltachain(&self, *args, **kw) -> PyResult<PyObject> { | |
236 | self.call_cindex(py, "deltachain", args, kw) |
|
238 | self.call_cindex(py, "deltachain", args, kw) | |
237 | } |
|
239 | } | |
238 |
|
240 | |||
239 | /// slice planned chunk read to reach a density threshold |
|
241 | /// slice planned chunk read to reach a density threshold | |
240 | def slicechunktodensity(&self, *args, **kw) -> PyResult<PyObject> { |
|
242 | def slicechunktodensity(&self, *args, **kw) -> PyResult<PyObject> { | |
241 | self.call_cindex(py, "slicechunktodensity", args, kw) |
|
243 | self.call_cindex(py, "slicechunktodensity", args, kw) | |
242 | } |
|
244 | } | |
243 |
|
245 | |||
244 | /// stats for the index |
|
246 | /// stats for the index | |
245 | def stats(&self, *args, **kw) -> PyResult<PyObject> { |
|
247 | def stats(&self, *args, **kw) -> PyResult<PyObject> { | |
246 | self.call_cindex(py, "stats", args, kw) |
|
248 | self.call_cindex(py, "stats", args, kw) | |
247 | } |
|
249 | } | |
248 |
|
250 | |||
249 | // index_sequence_methods and index_mapping_methods. |
|
251 | // index_sequence_methods and index_mapping_methods. | |
250 | // |
|
252 | // | |
251 | // Since we call back through the high level Python API, |
|
253 | // Since we call back through the high level Python API, | |
252 | // there's no point making a distinction between index_get |
|
254 | // there's no point making a distinction between index_get | |
253 | // and index_getitem. |
|
255 | // and index_getitem. | |
254 |
|
256 | |||
255 | def __len__(&self) -> PyResult<usize> { |
|
257 | def __len__(&self) -> PyResult<usize> { | |
256 | self.cindex(py).borrow().inner().len(py) |
|
258 | self.cindex(py).borrow().inner().len(py) | |
257 | } |
|
259 | } | |
258 |
|
260 | |||
259 | def __getitem__(&self, key: PyObject) -> PyResult<PyObject> { |
|
261 | def __getitem__(&self, key: PyObject) -> PyResult<PyObject> { | |
260 | // this conversion seems needless, but that's actually because |
|
262 | // this conversion seems needless, but that's actually because | |
261 | // `index_getitem` does not handle conversion from PyLong, |
|
263 | // `index_getitem` does not handle conversion from PyLong, | |
262 | // which expressions such as [e for e in index] internally use. |
|
264 | // which expressions such as [e for e in index] internally use. | |
263 | // Note that we don't seem to have a direct way to call |
|
265 | // Note that we don't seem to have a direct way to call | |
264 | // PySequence_GetItem (does the job), which would possibly be better |
|
266 | // PySequence_GetItem (does the job), which would possibly be better | |
265 | // for performance |
|
267 | // for performance | |
266 | let key = match key.extract::<i32>(py) { |
|
268 | let key = match key.extract::<i32>(py) { | |
267 | Ok(rev) => rev.to_py_object(py).into_object(), |
|
269 | Ok(rev) => rev.to_py_object(py).into_object(), | |
268 | Err(_) => key, |
|
270 | Err(_) => key, | |
269 | }; |
|
271 | }; | |
270 | self.cindex(py).borrow().inner().get_item(py, key) |
|
272 | self.cindex(py).borrow().inner().get_item(py, key) | |
271 | } |
|
273 | } | |
272 |
|
274 | |||
273 | def __setitem__(&self, key: PyObject, value: PyObject) -> PyResult<()> { |
|
275 | def __setitem__(&self, key: PyObject, value: PyObject) -> PyResult<()> { | |
274 | self.cindex(py).borrow().inner().set_item(py, key, value) |
|
276 | self.cindex(py).borrow().inner().set_item(py, key, value) | |
275 | } |
|
277 | } | |
276 |
|
278 | |||
277 | def __contains__(&self, item: PyObject) -> PyResult<bool> { |
|
279 | def __contains__(&self, item: PyObject) -> PyResult<bool> { | |
278 | // ObjectProtocol does not seem to provide contains(), so |
|
280 | // ObjectProtocol does not seem to provide contains(), so | |
279 | // this is an equivalent implementation of the index_contains() |
|
281 | // this is an equivalent implementation of the index_contains() | |
280 | // defined in revlog.c |
|
282 | // defined in revlog.c | |
281 | let cindex = self.cindex(py).borrow(); |
|
283 | let cindex = self.cindex(py).borrow(); | |
282 | match item.extract::<i32>(py) { |
|
284 | match item.extract::<i32>(py) { | |
283 | Ok(rev) => { |
|
285 | Ok(rev) => { | |
284 | Ok(rev >= -1 && rev < cindex.inner().len(py)? as BaseRevision) |
|
286 | Ok(rev >= -1 && rev < cindex.inner().len(py)? as BaseRevision) | |
285 | } |
|
287 | } | |
286 | Err(_) => { |
|
288 | Err(_) => { | |
287 | cindex.inner().call_method( |
|
289 | cindex.inner().call_method( | |
288 | py, |
|
290 | py, | |
289 | "has_node", |
|
291 | "has_node", | |
290 | PyTuple::new(py, &[item]), |
|
292 | PyTuple::new(py, &[item]), | |
291 | None)? |
|
293 | None)? | |
292 | .extract(py) |
|
294 | .extract(py) | |
293 | } |
|
295 | } | |
294 | } |
|
296 | } | |
295 | } |
|
297 | } | |
296 |
|
298 | |||
297 | def nodemap_data_all(&self) -> PyResult<PyBytes> { |
|
299 | def nodemap_data_all(&self) -> PyResult<PyBytes> { | |
298 | self.inner_nodemap_data_all(py) |
|
300 | self.inner_nodemap_data_all(py) | |
299 | } |
|
301 | } | |
300 |
|
302 | |||
301 | def nodemap_data_incremental(&self) -> PyResult<PyObject> { |
|
303 | def nodemap_data_incremental(&self) -> PyResult<PyObject> { | |
302 | self.inner_nodemap_data_incremental(py) |
|
304 | self.inner_nodemap_data_incremental(py) | |
303 | } |
|
305 | } | |
304 | def update_nodemap_data( |
|
306 | def update_nodemap_data( | |
305 | &self, |
|
307 | &self, | |
306 | docket: PyObject, |
|
308 | docket: PyObject, | |
307 | nm_data: PyObject |
|
309 | nm_data: PyObject | |
308 | ) -> PyResult<PyObject> { |
|
310 | ) -> PyResult<PyObject> { | |
309 | self.inner_update_nodemap_data(py, docket, nm_data) |
|
311 | self.inner_update_nodemap_data(py, docket, nm_data) | |
310 | } |
|
312 | } | |
311 |
|
313 | |||
312 | @property |
|
314 | @property | |
313 | def entry_size(&self) -> PyResult<PyInt> { |
|
315 | def entry_size(&self) -> PyResult<PyInt> { | |
314 | self.cindex(py).borrow().inner().getattr(py, "entry_size")?.extract::<PyInt>(py) |
|
316 | self.cindex(py).borrow().inner().getattr(py, "entry_size")?.extract::<PyInt>(py) | |
315 | } |
|
317 | } | |
316 |
|
318 | |||
317 | @property |
|
319 | @property | |
318 | def rust_ext_compat(&self) -> PyResult<PyInt> { |
|
320 | def rust_ext_compat(&self) -> PyResult<PyInt> { | |
319 | self.cindex(py).borrow().inner().getattr(py, "rust_ext_compat")?.extract::<PyInt>(py) |
|
321 | self.cindex(py).borrow().inner().getattr(py, "rust_ext_compat")?.extract::<PyInt>(py) | |
320 | } |
|
322 | } | |
321 |
|
323 | |||
322 | }); |
|
324 | }); | |
323 |
|
325 | |||
324 | /// Take a (potentially) mmap'ed buffer, and return the underlying Python |
|
326 | /// Take a (potentially) mmap'ed buffer, and return the underlying Python | |
325 | /// buffer along with the Rust slice into said buffer. We need to keep the |
|
327 | /// buffer along with the Rust slice into said buffer. We need to keep the | |
326 | /// Python buffer around, otherwise we'd get a dangling pointer once the buffer |
|
328 | /// Python buffer around, otherwise we'd get a dangling pointer once the buffer | |
327 | /// is freed from Python's side. |
|
329 | /// is freed from Python's side. | |
328 | /// |
|
330 | /// | |
329 | /// # Safety |
|
331 | /// # Safety | |
330 | /// |
|
332 | /// | |
331 | /// The caller must make sure that the buffer is kept around for at least as |
|
333 | /// The caller must make sure that the buffer is kept around for at least as | |
332 | /// long as the slice. |
|
334 | /// long as the slice. | |
333 | #[deny(unsafe_op_in_unsafe_fn)] |
|
335 | #[deny(unsafe_op_in_unsafe_fn)] | |
334 | unsafe fn mmap_keeparound( |
|
336 | unsafe fn mmap_keeparound( | |
335 | py: Python, |
|
337 | py: Python, | |
336 | data: PyObject, |
|
338 | data: PyObject, | |
337 | ) -> PyResult<( |
|
339 | ) -> PyResult<( | |
338 | PyBuffer, |
|
340 | PyBuffer, | |
339 | Box<dyn std::ops::Deref<Target = [u8]> + Send + 'static>, |
|
341 | Box<dyn std::ops::Deref<Target = [u8]> + Send + 'static>, | |
340 | )> { |
|
342 | )> { | |
341 | let buf = PyBuffer::get(py, &data)?; |
|
343 | let buf = PyBuffer::get(py, &data)?; | |
342 | let len = buf.item_count(); |
|
344 | let len = buf.item_count(); | |
343 |
|
345 | |||
344 | // Build a slice from the mmap'ed buffer data |
|
346 | // Build a slice from the mmap'ed buffer data | |
345 | let cbuf = buf.buf_ptr(); |
|
347 | let cbuf = buf.buf_ptr(); | |
346 | let bytes = if std::mem::size_of::<u8>() == buf.item_size() |
|
348 | let bytes = if std::mem::size_of::<u8>() == buf.item_size() | |
347 | && buf.is_c_contiguous() |
|
349 | && buf.is_c_contiguous() | |
348 | && u8::is_compatible_format(buf.format()) |
|
350 | && u8::is_compatible_format(buf.format()) | |
349 | { |
|
351 | { | |
350 | unsafe { std::slice::from_raw_parts(cbuf as *const u8, len) } |
|
352 | unsafe { std::slice::from_raw_parts(cbuf as *const u8, len) } | |
351 | } else { |
|
353 | } else { | |
352 | return Err(PyErr::new::<ValueError, _>( |
|
354 | return Err(PyErr::new::<ValueError, _>( | |
353 | py, |
|
355 | py, | |
354 | "Nodemap data buffer has an invalid memory representation" |
|
356 | "Nodemap data buffer has an invalid memory representation" | |
355 | .to_string(), |
|
357 | .to_string(), | |
356 | )); |
|
358 | )); | |
357 | }; |
|
359 | }; | |
358 |
|
360 | |||
359 | Ok((buf, Box::new(bytes))) |
|
361 | Ok((buf, Box::new(bytes))) | |
360 | } |
|
362 | } | |
361 |
|
363 | |||
362 | impl MixedIndex { |
|
364 | impl MixedIndex { | |
363 | fn new( |
|
365 | fn new( | |
364 | py: Python, |
|
366 | py: Python, | |
365 | cindex: PyObject, |
|
367 | cindex: PyObject, | |
366 | data: PyObject, |
|
368 | data: PyObject, | |
|
369 | header: u32, | |||
367 | ) -> PyResult<MixedIndex> { |
|
370 | ) -> PyResult<MixedIndex> { | |
368 | // Safety: we keep the buffer around inside the class as `index_mmap` |
|
371 | // Safety: we keep the buffer around inside the class as `index_mmap` | |
369 | let (buf, bytes) = unsafe { mmap_keeparound(py, data)? }; |
|
372 | let (buf, bytes) = unsafe { mmap_keeparound(py, data)? }; | |
370 |
|
373 | |||
371 | Self::create_instance( |
|
374 | Self::create_instance( | |
372 | py, |
|
375 | py, | |
373 | RefCell::new(cindex::Index::new(py, cindex)?), |
|
376 | RefCell::new(cindex::Index::new(py, cindex)?), | |
374 | RefCell::new(hg::index::Index::new(bytes).unwrap()), |
|
377 | RefCell::new( | |
|
378 | hg::index::Index::new( | |||
|
379 | bytes, | |||
|
380 | IndexHeader::parse(&header.to_be_bytes()) | |||
|
381 | .expect("default header is broken") | |||
|
382 | .unwrap(), | |||
|
383 | ) | |||
|
384 | .unwrap(), | |||
|
385 | ), | |||
375 | RefCell::new(None), |
|
386 | RefCell::new(None), | |
376 | RefCell::new(None), |
|
387 | RefCell::new(None), | |
377 | RefCell::new(None), |
|
388 | RefCell::new(None), | |
378 | RefCell::new(Some(buf)), |
|
389 | RefCell::new(Some(buf)), | |
379 | ) |
|
390 | ) | |
380 | } |
|
391 | } | |
381 |
|
392 | |||
382 | /// This is scaffolding at this point, but it could also become |
|
393 | /// This is scaffolding at this point, but it could also become | |
383 | /// a way to start a persistent nodemap or perform a |
|
394 | /// a way to start a persistent nodemap or perform a | |
384 | /// vacuum / repack operation |
|
395 | /// vacuum / repack operation | |
385 | fn fill_nodemap( |
|
396 | fn fill_nodemap( | |
386 | &self, |
|
397 | &self, | |
387 | py: Python, |
|
398 | py: Python, | |
388 | nt: &mut NodeTree, |
|
399 | nt: &mut NodeTree, | |
389 | ) -> PyResult<PyObject> { |
|
400 | ) -> PyResult<PyObject> { | |
390 | let index = self.cindex(py).borrow(); |
|
401 | let index = self.cindex(py).borrow(); | |
391 | for r in 0..index.len() { |
|
402 | for r in 0..index.len() { | |
392 | let rev = Revision(r as BaseRevision); |
|
403 | let rev = Revision(r as BaseRevision); | |
393 | // in this case node() won't ever return None |
|
404 | // in this case node() won't ever return None | |
394 | nt.insert(&*index, index.node(rev).unwrap(), rev) |
|
405 | nt.insert(&*index, index.node(rev).unwrap(), rev) | |
395 | .map_err(|e| nodemap_error(py, e))? |
|
406 | .map_err(|e| nodemap_error(py, e))? | |
396 | } |
|
407 | } | |
397 | Ok(py.None()) |
|
408 | Ok(py.None()) | |
398 | } |
|
409 | } | |
399 |
|
410 | |||
400 | fn get_nodetree<'a>( |
|
411 | fn get_nodetree<'a>( | |
401 | &'a self, |
|
412 | &'a self, | |
402 | py: Python<'a>, |
|
413 | py: Python<'a>, | |
403 | ) -> PyResult<&'a RefCell<Option<NodeTree>>> { |
|
414 | ) -> PyResult<&'a RefCell<Option<NodeTree>>> { | |
404 | if self.nt(py).borrow().is_none() { |
|
415 | if self.nt(py).borrow().is_none() { | |
405 | let readonly = Box::<Vec<_>>::default(); |
|
416 | let readonly = Box::<Vec<_>>::default(); | |
406 | let mut nt = NodeTree::load_bytes(readonly, 0); |
|
417 | let mut nt = NodeTree::load_bytes(readonly, 0); | |
407 | self.fill_nodemap(py, &mut nt)?; |
|
418 | self.fill_nodemap(py, &mut nt)?; | |
408 | self.nt(py).borrow_mut().replace(nt); |
|
419 | self.nt(py).borrow_mut().replace(nt); | |
409 | } |
|
420 | } | |
410 | Ok(self.nt(py)) |
|
421 | Ok(self.nt(py)) | |
411 | } |
|
422 | } | |
412 |
|
423 | |||
413 | /// forward a method call to the underlying C index |
|
424 | /// forward a method call to the underlying C index | |
414 | fn call_cindex( |
|
425 | fn call_cindex( | |
415 | &self, |
|
426 | &self, | |
416 | py: Python, |
|
427 | py: Python, | |
417 | name: &str, |
|
428 | name: &str, | |
418 | args: &PyTuple, |
|
429 | args: &PyTuple, | |
419 | kwargs: Option<&PyDict>, |
|
430 | kwargs: Option<&PyDict>, | |
420 | ) -> PyResult<PyObject> { |
|
431 | ) -> PyResult<PyObject> { | |
421 | self.cindex(py) |
|
432 | self.cindex(py) | |
422 | .borrow() |
|
433 | .borrow() | |
423 | .inner() |
|
434 | .inner() | |
424 | .call_method(py, name, args, kwargs) |
|
435 | .call_method(py, name, args, kwargs) | |
425 | } |
|
436 | } | |
426 |
|
437 | |||
427 | pub fn clone_cindex(&self, py: Python) -> cindex::Index { |
|
438 | pub fn clone_cindex(&self, py: Python) -> cindex::Index { | |
428 | self.cindex(py).borrow().clone_ref(py) |
|
439 | self.cindex(py).borrow().clone_ref(py) | |
429 | } |
|
440 | } | |
430 |
|
441 | |||
431 | /// Returns the full nodemap bytes to be written as-is to disk |
|
442 | /// Returns the full nodemap bytes to be written as-is to disk | |
432 | fn inner_nodemap_data_all(&self, py: Python) -> PyResult<PyBytes> { |
|
443 | fn inner_nodemap_data_all(&self, py: Python) -> PyResult<PyBytes> { | |
433 | let nodemap = self.get_nodetree(py)?.borrow_mut().take().unwrap(); |
|
444 | let nodemap = self.get_nodetree(py)?.borrow_mut().take().unwrap(); | |
434 | let (readonly, bytes) = nodemap.into_readonly_and_added_bytes(); |
|
445 | let (readonly, bytes) = nodemap.into_readonly_and_added_bytes(); | |
435 |
|
446 | |||
436 | // If there's anything readonly, we need to build the data again from |
|
447 | // If there's anything readonly, we need to build the data again from | |
437 | // scratch |
|
448 | // scratch | |
438 | let bytes = if readonly.len() > 0 { |
|
449 | let bytes = if readonly.len() > 0 { | |
439 | let mut nt = NodeTree::load_bytes(Box::<Vec<_>>::default(), 0); |
|
450 | let mut nt = NodeTree::load_bytes(Box::<Vec<_>>::default(), 0); | |
440 | self.fill_nodemap(py, &mut nt)?; |
|
451 | self.fill_nodemap(py, &mut nt)?; | |
441 |
|
452 | |||
442 | let (readonly, bytes) = nt.into_readonly_and_added_bytes(); |
|
453 | let (readonly, bytes) = nt.into_readonly_and_added_bytes(); | |
443 | assert_eq!(readonly.len(), 0); |
|
454 | assert_eq!(readonly.len(), 0); | |
444 |
|
455 | |||
445 | bytes |
|
456 | bytes | |
446 | } else { |
|
457 | } else { | |
447 | bytes |
|
458 | bytes | |
448 | }; |
|
459 | }; | |
449 |
|
460 | |||
450 | let bytes = PyBytes::new(py, &bytes); |
|
461 | let bytes = PyBytes::new(py, &bytes); | |
451 | Ok(bytes) |
|
462 | Ok(bytes) | |
452 | } |
|
463 | } | |
453 |
|
464 | |||
454 | /// Returns the last saved docket along with the size of any changed data |
|
465 | /// Returns the last saved docket along with the size of any changed data | |
455 | /// (in number of blocks), and said data as bytes. |
|
466 | /// (in number of blocks), and said data as bytes. | |
456 | fn inner_nodemap_data_incremental( |
|
467 | fn inner_nodemap_data_incremental( | |
457 | &self, |
|
468 | &self, | |
458 | py: Python, |
|
469 | py: Python, | |
459 | ) -> PyResult<PyObject> { |
|
470 | ) -> PyResult<PyObject> { | |
460 | let docket = self.docket(py).borrow(); |
|
471 | let docket = self.docket(py).borrow(); | |
461 | let docket = match docket.as_ref() { |
|
472 | let docket = match docket.as_ref() { | |
462 | Some(d) => d, |
|
473 | Some(d) => d, | |
463 | None => return Ok(py.None()), |
|
474 | None => return Ok(py.None()), | |
464 | }; |
|
475 | }; | |
465 |
|
476 | |||
466 | let node_tree = self.get_nodetree(py)?.borrow_mut().take().unwrap(); |
|
477 | let node_tree = self.get_nodetree(py)?.borrow_mut().take().unwrap(); | |
467 | let masked_blocks = node_tree.masked_readonly_blocks(); |
|
478 | let masked_blocks = node_tree.masked_readonly_blocks(); | |
468 | let (_, data) = node_tree.into_readonly_and_added_bytes(); |
|
479 | let (_, data) = node_tree.into_readonly_and_added_bytes(); | |
469 | let changed = masked_blocks * std::mem::size_of::<Block>(); |
|
480 | let changed = masked_blocks * std::mem::size_of::<Block>(); | |
470 |
|
481 | |||
471 | Ok((docket, changed, PyBytes::new(py, &data)) |
|
482 | Ok((docket, changed, PyBytes::new(py, &data)) | |
472 | .to_py_object(py) |
|
483 | .to_py_object(py) | |
473 | .into_object()) |
|
484 | .into_object()) | |
474 | } |
|
485 | } | |
475 |
|
486 | |||
476 | /// Update the nodemap from the new (mmaped) data. |
|
487 | /// Update the nodemap from the new (mmaped) data. | |
477 | /// The docket is kept as a reference for later incremental calls. |
|
488 | /// The docket is kept as a reference for later incremental calls. | |
478 | fn inner_update_nodemap_data( |
|
489 | fn inner_update_nodemap_data( | |
479 | &self, |
|
490 | &self, | |
480 | py: Python, |
|
491 | py: Python, | |
481 | docket: PyObject, |
|
492 | docket: PyObject, | |
482 | nm_data: PyObject, |
|
493 | nm_data: PyObject, | |
483 | ) -> PyResult<PyObject> { |
|
494 | ) -> PyResult<PyObject> { | |
484 | // Safety: we keep the buffer around inside the class as `nodemap_mmap` |
|
495 | // Safety: we keep the buffer around inside the class as `nodemap_mmap` | |
485 | let (buf, bytes) = unsafe { mmap_keeparound(py, nm_data)? }; |
|
496 | let (buf, bytes) = unsafe { mmap_keeparound(py, nm_data)? }; | |
486 | let len = buf.item_count(); |
|
497 | let len = buf.item_count(); | |
487 | self.nodemap_mmap(py).borrow_mut().replace(buf); |
|
498 | self.nodemap_mmap(py).borrow_mut().replace(buf); | |
488 |
|
499 | |||
489 | let mut nt = NodeTree::load_bytes(bytes, len); |
|
500 | let mut nt = NodeTree::load_bytes(bytes, len); | |
490 |
|
501 | |||
491 | let data_tip = docket |
|
502 | let data_tip = docket | |
492 | .getattr(py, "tip_rev")? |
|
503 | .getattr(py, "tip_rev")? | |
493 | .extract::<BaseRevision>(py)? |
|
504 | .extract::<BaseRevision>(py)? | |
494 | .into(); |
|
505 | .into(); | |
495 | self.docket(py).borrow_mut().replace(docket.clone_ref(py)); |
|
506 | self.docket(py).borrow_mut().replace(docket.clone_ref(py)); | |
496 | let idx = self.cindex(py).borrow(); |
|
507 | let idx = self.cindex(py).borrow(); | |
497 | let data_tip = idx.check_revision(data_tip).ok_or_else(|| { |
|
508 | let data_tip = idx.check_revision(data_tip).ok_or_else(|| { | |
498 | nodemap_error(py, NodeMapError::RevisionNotInIndex(data_tip)) |
|
509 | nodemap_error(py, NodeMapError::RevisionNotInIndex(data_tip)) | |
499 | })?; |
|
510 | })?; | |
500 | let current_tip = idx.len(); |
|
511 | let current_tip = idx.len(); | |
501 |
|
512 | |||
502 | for r in (data_tip.0 + 1)..current_tip as BaseRevision { |
|
513 | for r in (data_tip.0 + 1)..current_tip as BaseRevision { | |
503 | let rev = Revision(r); |
|
514 | let rev = Revision(r); | |
504 | // in this case node() won't ever return None |
|
515 | // in this case node() won't ever return None | |
505 | nt.insert(&*idx, idx.node(rev).unwrap(), rev) |
|
516 | nt.insert(&*idx, idx.node(rev).unwrap(), rev) | |
506 | .map_err(|e| nodemap_error(py, e))? |
|
517 | .map_err(|e| nodemap_error(py, e))? | |
507 | } |
|
518 | } | |
508 |
|
519 | |||
509 | *self.nt(py).borrow_mut() = Some(nt); |
|
520 | *self.nt(py).borrow_mut() = Some(nt); | |
510 |
|
521 | |||
511 | Ok(py.None()) |
|
522 | Ok(py.None()) | |
512 | } |
|
523 | } | |
513 | } |
|
524 | } | |
514 |
|
525 | |||
515 | fn revlog_error(py: Python) -> PyErr { |
|
526 | fn revlog_error(py: Python) -> PyErr { | |
516 | match py |
|
527 | match py | |
517 | .import("mercurial.error") |
|
528 | .import("mercurial.error") | |
518 | .and_then(|m| m.get(py, "RevlogError")) |
|
529 | .and_then(|m| m.get(py, "RevlogError")) | |
519 | { |
|
530 | { | |
520 | Err(e) => e, |
|
531 | Err(e) => e, | |
521 | Ok(cls) => PyErr::from_instance( |
|
532 | Ok(cls) => PyErr::from_instance( | |
522 | py, |
|
533 | py, | |
523 | cls.call(py, (py.None(),), None).ok().into_py_object(py), |
|
534 | cls.call(py, (py.None(),), None).ok().into_py_object(py), | |
524 | ), |
|
535 | ), | |
525 | } |
|
536 | } | |
526 | } |
|
537 | } | |
527 |
|
538 | |||
528 | fn rev_not_in_index(py: Python, rev: UncheckedRevision) -> PyErr { |
|
539 | fn rev_not_in_index(py: Python, rev: UncheckedRevision) -> PyErr { | |
529 | PyErr::new::<ValueError, _>( |
|
540 | PyErr::new::<ValueError, _>( | |
530 | py, |
|
541 | py, | |
531 | format!( |
|
542 | format!( | |
532 | "Inconsistency: Revision {} found in nodemap \ |
|
543 | "Inconsistency: Revision {} found in nodemap \ | |
533 | is not in revlog index", |
|
544 | is not in revlog index", | |
534 | rev |
|
545 | rev | |
535 | ), |
|
546 | ), | |
536 | ) |
|
547 | ) | |
537 | } |
|
548 | } | |
538 |
|
549 | |||
539 | /// Standard treatment of NodeMapError |
|
550 | /// Standard treatment of NodeMapError | |
540 | fn nodemap_error(py: Python, err: NodeMapError) -> PyErr { |
|
551 | fn nodemap_error(py: Python, err: NodeMapError) -> PyErr { | |
541 | match err { |
|
552 | match err { | |
542 | NodeMapError::MultipleResults => revlog_error(py), |
|
553 | NodeMapError::MultipleResults => revlog_error(py), | |
543 | NodeMapError::RevisionNotInIndex(r) => rev_not_in_index(py, r), |
|
554 | NodeMapError::RevisionNotInIndex(r) => rev_not_in_index(py, r), | |
544 | } |
|
555 | } | |
545 | } |
|
556 | } | |
546 |
|
557 | |||
547 | /// Create the module, with __package__ given from parent |
|
558 | /// Create the module, with __package__ given from parent | |
548 | pub fn init_module(py: Python, package: &str) -> PyResult<PyModule> { |
|
559 | pub fn init_module(py: Python, package: &str) -> PyResult<PyModule> { | |
549 | let dotted_name = &format!("{}.revlog", package); |
|
560 | let dotted_name = &format!("{}.revlog", package); | |
550 | let m = PyModule::new(py, dotted_name)?; |
|
561 | let m = PyModule::new(py, dotted_name)?; | |
551 | m.add(py, "__package__", package)?; |
|
562 | m.add(py, "__package__", package)?; | |
552 | m.add(py, "__doc__", "RevLog - Rust implementations")?; |
|
563 | m.add(py, "__doc__", "RevLog - Rust implementations")?; | |
553 |
|
564 | |||
554 | m.add_class::<MixedIndex>(py)?; |
|
565 | m.add_class::<MixedIndex>(py)?; | |
555 |
|
566 | |||
556 | let sys = PyModule::import(py, "sys")?; |
|
567 | let sys = PyModule::import(py, "sys")?; | |
557 | let sys_modules: PyDict = sys.get(py, "modules")?.extract(py)?; |
|
568 | let sys_modules: PyDict = sys.get(py, "modules")?.extract(py)?; | |
558 | sys_modules.set_item(py, dotted_name, &m)?; |
|
569 | sys_modules.set_item(py, dotted_name, &m)?; | |
559 |
|
570 | |||
560 | Ok(m) |
|
571 | Ok(m) | |
561 | } |
|
572 | } |
@@ -1,815 +1,822 b'' | |||||
1 | // status.rs |
|
1 | // status.rs | |
2 | // |
|
2 | // | |
3 | // Copyright 2020, Georges Racinet <georges.racinets@octobus.net> |
|
3 | // Copyright 2020, Georges Racinet <georges.racinets@octobus.net> | |
4 | // |
|
4 | // | |
5 | // This software may be used and distributed according to the terms of the |
|
5 | // This software may be used and distributed according to the terms of the | |
6 | // GNU General Public License version 2 or any later version. |
|
6 | // GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | use crate::error::CommandError; |
|
8 | use crate::error::CommandError; | |
9 | use crate::ui::{ |
|
9 | use crate::ui::{ | |
10 | format_pattern_file_warning, print_narrow_sparse_warnings, relative_paths, |
|
10 | format_pattern_file_warning, print_narrow_sparse_warnings, relative_paths, | |
11 | RelativePaths, Ui, |
|
11 | RelativePaths, Ui, | |
12 | }; |
|
12 | }; | |
13 | use crate::utils::path_utils::RelativizePaths; |
|
13 | use crate::utils::path_utils::RelativizePaths; | |
14 | use clap::Arg; |
|
14 | use clap::Arg; | |
15 | use format_bytes::format_bytes; |
|
15 | use format_bytes::format_bytes; | |
16 | use hg::config::Config; |
|
16 | use hg::config::Config; | |
17 | use hg::dirstate::has_exec_bit; |
|
17 | use hg::dirstate::has_exec_bit; | |
18 | use hg::dirstate::status::StatusPath; |
|
18 | use hg::dirstate::status::StatusPath; | |
19 | use hg::dirstate::TruncatedTimestamp; |
|
19 | use hg::dirstate::TruncatedTimestamp; | |
20 | use hg::errors::{HgError, IoResultExt}; |
|
20 | use hg::errors::{HgError, IoResultExt}; | |
21 | use hg::filepatterns::parse_pattern_args; |
|
21 | use hg::filepatterns::parse_pattern_args; | |
22 | use hg::lock::LockError; |
|
22 | use hg::lock::LockError; | |
23 | use hg::manifest::Manifest; |
|
23 | use hg::manifest::Manifest; | |
24 | use hg::matchers::{AlwaysMatcher, IntersectionMatcher}; |
|
24 | use hg::matchers::{AlwaysMatcher, IntersectionMatcher}; | |
25 | use hg::repo::Repo; |
|
25 | use hg::repo::Repo; | |
26 | use hg::utils::debug::debug_wait_for_file; |
|
26 | use hg::utils::debug::debug_wait_for_file; | |
27 | use hg::utils::files::{ |
|
27 | use hg::utils::files::{ | |
28 | get_bytes_from_os_str, get_bytes_from_os_string, get_path_from_bytes, |
|
28 | get_bytes_from_os_str, get_bytes_from_os_string, get_path_from_bytes, | |
29 | }; |
|
29 | }; | |
30 | use hg::utils::hg_path::{hg_path_to_path_buf, HgPath}; |
|
30 | use hg::utils::hg_path::{hg_path_to_path_buf, HgPath}; | |
31 | use hg::DirstateStatus; |
|
|||
32 | use hg::PatternFileWarning; |
|
31 | use hg::PatternFileWarning; | |
33 | use hg::Revision; |
|
32 | use hg::Revision; | |
34 | use hg::StatusError; |
|
33 | use hg::StatusError; | |
35 | use hg::StatusOptions; |
|
34 | use hg::StatusOptions; | |
36 | use hg::{self, narrow, sparse}; |
|
35 | use hg::{self, narrow, sparse}; | |
|
36 | use hg::{DirstateStatus, RevlogOpenOptions}; | |||
37 | use log::info; |
|
37 | use log::info; | |
38 | use rayon::prelude::*; |
|
38 | use rayon::prelude::*; | |
39 | use std::borrow::Cow; |
|
39 | use std::borrow::Cow; | |
40 | use std::io; |
|
40 | use std::io; | |
41 | use std::mem::take; |
|
41 | use std::mem::take; | |
42 | use std::path::PathBuf; |
|
42 | use std::path::PathBuf; | |
43 |
|
43 | |||
44 | pub const HELP_TEXT: &str = " |
|
44 | pub const HELP_TEXT: &str = " | |
45 | Show changed files in the working directory |
|
45 | Show changed files in the working directory | |
46 |
|
46 | |||
47 | This is a pure Rust version of `hg status`. |
|
47 | This is a pure Rust version of `hg status`. | |
48 |
|
48 | |||
49 | Some options might be missing, check the list below. |
|
49 | Some options might be missing, check the list below. | |
50 | "; |
|
50 | "; | |
51 |
|
51 | |||
52 | pub fn args() -> clap::Command { |
|
52 | pub fn args() -> clap::Command { | |
53 | clap::command!("status") |
|
53 | clap::command!("status") | |
54 | .alias("st") |
|
54 | .alias("st") | |
55 | .about(HELP_TEXT) |
|
55 | .about(HELP_TEXT) | |
56 | .arg( |
|
56 | .arg( | |
57 | Arg::new("file") |
|
57 | Arg::new("file") | |
58 | .value_parser(clap::value_parser!(std::ffi::OsString)) |
|
58 | .value_parser(clap::value_parser!(std::ffi::OsString)) | |
59 | .help("show only these files") |
|
59 | .help("show only these files") | |
60 | .action(clap::ArgAction::Append), |
|
60 | .action(clap::ArgAction::Append), | |
61 | ) |
|
61 | ) | |
62 | .arg( |
|
62 | .arg( | |
63 | Arg::new("all") |
|
63 | Arg::new("all") | |
64 | .help("show status of all files") |
|
64 | .help("show status of all files") | |
65 | .short('A') |
|
65 | .short('A') | |
66 | .action(clap::ArgAction::SetTrue) |
|
66 | .action(clap::ArgAction::SetTrue) | |
67 | .long("all"), |
|
67 | .long("all"), | |
68 | ) |
|
68 | ) | |
69 | .arg( |
|
69 | .arg( | |
70 | Arg::new("modified") |
|
70 | Arg::new("modified") | |
71 | .help("show only modified files") |
|
71 | .help("show only modified files") | |
72 | .short('m') |
|
72 | .short('m') | |
73 | .action(clap::ArgAction::SetTrue) |
|
73 | .action(clap::ArgAction::SetTrue) | |
74 | .long("modified"), |
|
74 | .long("modified"), | |
75 | ) |
|
75 | ) | |
76 | .arg( |
|
76 | .arg( | |
77 | Arg::new("added") |
|
77 | Arg::new("added") | |
78 | .help("show only added files") |
|
78 | .help("show only added files") | |
79 | .short('a') |
|
79 | .short('a') | |
80 | .action(clap::ArgAction::SetTrue) |
|
80 | .action(clap::ArgAction::SetTrue) | |
81 | .long("added"), |
|
81 | .long("added"), | |
82 | ) |
|
82 | ) | |
83 | .arg( |
|
83 | .arg( | |
84 | Arg::new("removed") |
|
84 | Arg::new("removed") | |
85 | .help("show only removed files") |
|
85 | .help("show only removed files") | |
86 | .short('r') |
|
86 | .short('r') | |
87 | .action(clap::ArgAction::SetTrue) |
|
87 | .action(clap::ArgAction::SetTrue) | |
88 | .long("removed"), |
|
88 | .long("removed"), | |
89 | ) |
|
89 | ) | |
90 | .arg( |
|
90 | .arg( | |
91 | Arg::new("clean") |
|
91 | Arg::new("clean") | |
92 | .help("show only clean files") |
|
92 | .help("show only clean files") | |
93 | .short('c') |
|
93 | .short('c') | |
94 | .action(clap::ArgAction::SetTrue) |
|
94 | .action(clap::ArgAction::SetTrue) | |
95 | .long("clean"), |
|
95 | .long("clean"), | |
96 | ) |
|
96 | ) | |
97 | .arg( |
|
97 | .arg( | |
98 | Arg::new("deleted") |
|
98 | Arg::new("deleted") | |
99 | .help("show only deleted files") |
|
99 | .help("show only deleted files") | |
100 | .short('d') |
|
100 | .short('d') | |
101 | .action(clap::ArgAction::SetTrue) |
|
101 | .action(clap::ArgAction::SetTrue) | |
102 | .long("deleted"), |
|
102 | .long("deleted"), | |
103 | ) |
|
103 | ) | |
104 | .arg( |
|
104 | .arg( | |
105 | Arg::new("unknown") |
|
105 | Arg::new("unknown") | |
106 | .help("show only unknown (not tracked) files") |
|
106 | .help("show only unknown (not tracked) files") | |
107 | .short('u') |
|
107 | .short('u') | |
108 | .action(clap::ArgAction::SetTrue) |
|
108 | .action(clap::ArgAction::SetTrue) | |
109 | .long("unknown"), |
|
109 | .long("unknown"), | |
110 | ) |
|
110 | ) | |
111 | .arg( |
|
111 | .arg( | |
112 | Arg::new("ignored") |
|
112 | Arg::new("ignored") | |
113 | .help("show only ignored files") |
|
113 | .help("show only ignored files") | |
114 | .short('i') |
|
114 | .short('i') | |
115 | .action(clap::ArgAction::SetTrue) |
|
115 | .action(clap::ArgAction::SetTrue) | |
116 | .long("ignored"), |
|
116 | .long("ignored"), | |
117 | ) |
|
117 | ) | |
118 | .arg( |
|
118 | .arg( | |
119 | Arg::new("copies") |
|
119 | Arg::new("copies") | |
120 | .help("show source of copied files (DEFAULT: ui.statuscopies)") |
|
120 | .help("show source of copied files (DEFAULT: ui.statuscopies)") | |
121 | .short('C') |
|
121 | .short('C') | |
122 | .action(clap::ArgAction::SetTrue) |
|
122 | .action(clap::ArgAction::SetTrue) | |
123 | .long("copies"), |
|
123 | .long("copies"), | |
124 | ) |
|
124 | ) | |
125 | .arg( |
|
125 | .arg( | |
126 | Arg::new("print0") |
|
126 | Arg::new("print0") | |
127 | .help("end filenames with NUL, for use with xargs") |
|
127 | .help("end filenames with NUL, for use with xargs") | |
128 | .short('0') |
|
128 | .short('0') | |
129 | .action(clap::ArgAction::SetTrue) |
|
129 | .action(clap::ArgAction::SetTrue) | |
130 | .long("print0"), |
|
130 | .long("print0"), | |
131 | ) |
|
131 | ) | |
132 | .arg( |
|
132 | .arg( | |
133 | Arg::new("no-status") |
|
133 | Arg::new("no-status") | |
134 | .help("hide status prefix") |
|
134 | .help("hide status prefix") | |
135 | .short('n') |
|
135 | .short('n') | |
136 | .action(clap::ArgAction::SetTrue) |
|
136 | .action(clap::ArgAction::SetTrue) | |
137 | .long("no-status"), |
|
137 | .long("no-status"), | |
138 | ) |
|
138 | ) | |
139 | .arg( |
|
139 | .arg( | |
140 | Arg::new("verbose") |
|
140 | Arg::new("verbose") | |
141 | .help("enable additional output") |
|
141 | .help("enable additional output") | |
142 | .short('v') |
|
142 | .short('v') | |
143 | .action(clap::ArgAction::SetTrue) |
|
143 | .action(clap::ArgAction::SetTrue) | |
144 | .long("verbose"), |
|
144 | .long("verbose"), | |
145 | ) |
|
145 | ) | |
146 | .arg( |
|
146 | .arg( | |
147 | Arg::new("rev") |
|
147 | Arg::new("rev") | |
148 | .help("show difference from/to revision") |
|
148 | .help("show difference from/to revision") | |
149 | .long("rev") |
|
149 | .long("rev") | |
150 | .num_args(1) |
|
150 | .num_args(1) | |
151 | .action(clap::ArgAction::Append) |
|
151 | .action(clap::ArgAction::Append) | |
152 | .value_name("REV"), |
|
152 | .value_name("REV"), | |
153 | ) |
|
153 | ) | |
154 | } |
|
154 | } | |
155 |
|
155 | |||
156 | fn parse_revpair( |
|
156 | fn parse_revpair( | |
157 | repo: &Repo, |
|
157 | repo: &Repo, | |
158 | revs: Option<Vec<String>>, |
|
158 | revs: Option<Vec<String>>, | |
159 | ) -> Result<Option<(Revision, Revision)>, CommandError> { |
|
159 | ) -> Result<Option<(Revision, Revision)>, CommandError> { | |
160 | let revs = match revs { |
|
160 | let revs = match revs { | |
161 | None => return Ok(None), |
|
161 | None => return Ok(None), | |
162 | Some(revs) => revs, |
|
162 | Some(revs) => revs, | |
163 | }; |
|
163 | }; | |
164 | if revs.is_empty() { |
|
164 | if revs.is_empty() { | |
165 | return Ok(None); |
|
165 | return Ok(None); | |
166 | } |
|
166 | } | |
167 | if revs.len() != 2 { |
|
167 | if revs.len() != 2 { | |
168 | return Err(CommandError::unsupported("expected 0 or 2 --rev flags")); |
|
168 | return Err(CommandError::unsupported("expected 0 or 2 --rev flags")); | |
169 | } |
|
169 | } | |
170 |
|
170 | |||
171 | let rev1 = &revs[0]; |
|
171 | let rev1 = &revs[0]; | |
172 | let rev2 = &revs[1]; |
|
172 | let rev2 = &revs[1]; | |
173 | let rev1 = hg::revset::resolve_single(rev1, repo) |
|
173 | let rev1 = hg::revset::resolve_single(rev1, repo) | |
174 | .map_err(|e| (e, rev1.as_str()))?; |
|
174 | .map_err(|e| (e, rev1.as_str()))?; | |
175 | let rev2 = hg::revset::resolve_single(rev2, repo) |
|
175 | let rev2 = hg::revset::resolve_single(rev2, repo) | |
176 | .map_err(|e| (e, rev2.as_str()))?; |
|
176 | .map_err(|e| (e, rev2.as_str()))?; | |
177 | Ok(Some((rev1, rev2))) |
|
177 | Ok(Some((rev1, rev2))) | |
178 | } |
|
178 | } | |
179 |
|
179 | |||
180 | /// Pure data type allowing the caller to specify file states to display |
|
180 | /// Pure data type allowing the caller to specify file states to display | |
181 | #[derive(Copy, Clone, Debug)] |
|
181 | #[derive(Copy, Clone, Debug)] | |
182 | pub struct DisplayStates { |
|
182 | pub struct DisplayStates { | |
183 | pub modified: bool, |
|
183 | pub modified: bool, | |
184 | pub added: bool, |
|
184 | pub added: bool, | |
185 | pub removed: bool, |
|
185 | pub removed: bool, | |
186 | pub clean: bool, |
|
186 | pub clean: bool, | |
187 | pub deleted: bool, |
|
187 | pub deleted: bool, | |
188 | pub unknown: bool, |
|
188 | pub unknown: bool, | |
189 | pub ignored: bool, |
|
189 | pub ignored: bool, | |
190 | } |
|
190 | } | |
191 |
|
191 | |||
192 | pub const DEFAULT_DISPLAY_STATES: DisplayStates = DisplayStates { |
|
192 | pub const DEFAULT_DISPLAY_STATES: DisplayStates = DisplayStates { | |
193 | modified: true, |
|
193 | modified: true, | |
194 | added: true, |
|
194 | added: true, | |
195 | removed: true, |
|
195 | removed: true, | |
196 | clean: false, |
|
196 | clean: false, | |
197 | deleted: true, |
|
197 | deleted: true, | |
198 | unknown: true, |
|
198 | unknown: true, | |
199 | ignored: false, |
|
199 | ignored: false, | |
200 | }; |
|
200 | }; | |
201 |
|
201 | |||
202 | pub const ALL_DISPLAY_STATES: DisplayStates = DisplayStates { |
|
202 | pub const ALL_DISPLAY_STATES: DisplayStates = DisplayStates { | |
203 | modified: true, |
|
203 | modified: true, | |
204 | added: true, |
|
204 | added: true, | |
205 | removed: true, |
|
205 | removed: true, | |
206 | clean: true, |
|
206 | clean: true, | |
207 | deleted: true, |
|
207 | deleted: true, | |
208 | unknown: true, |
|
208 | unknown: true, | |
209 | ignored: true, |
|
209 | ignored: true, | |
210 | }; |
|
210 | }; | |
211 |
|
211 | |||
212 | impl DisplayStates { |
|
212 | impl DisplayStates { | |
213 | pub fn is_empty(&self) -> bool { |
|
213 | pub fn is_empty(&self) -> bool { | |
214 | !(self.modified |
|
214 | !(self.modified | |
215 | || self.added |
|
215 | || self.added | |
216 | || self.removed |
|
216 | || self.removed | |
217 | || self.clean |
|
217 | || self.clean | |
218 | || self.deleted |
|
218 | || self.deleted | |
219 | || self.unknown |
|
219 | || self.unknown | |
220 | || self.ignored) |
|
220 | || self.ignored) | |
221 | } |
|
221 | } | |
222 | } |
|
222 | } | |
223 |
|
223 | |||
224 | fn has_unfinished_merge(repo: &Repo) -> Result<bool, CommandError> { |
|
224 | fn has_unfinished_merge(repo: &Repo) -> Result<bool, CommandError> { | |
225 | Ok(repo.dirstate_parents()?.is_merge()) |
|
225 | Ok(repo.dirstate_parents()?.is_merge()) | |
226 | } |
|
226 | } | |
227 |
|
227 | |||
228 | fn has_unfinished_state(repo: &Repo) -> Result<bool, CommandError> { |
|
228 | fn has_unfinished_state(repo: &Repo) -> Result<bool, CommandError> { | |
229 | // These are all the known values for the [fname] argument of |
|
229 | // These are all the known values for the [fname] argument of | |
230 | // [addunfinished] function in [state.py] |
|
230 | // [addunfinished] function in [state.py] | |
231 | let known_state_files: &[&str] = &[ |
|
231 | let known_state_files: &[&str] = &[ | |
232 | "bisect.state", |
|
232 | "bisect.state", | |
233 | "graftstate", |
|
233 | "graftstate", | |
234 | "histedit-state", |
|
234 | "histedit-state", | |
235 | "rebasestate", |
|
235 | "rebasestate", | |
236 | "shelvedstate", |
|
236 | "shelvedstate", | |
237 | "transplant/journal", |
|
237 | "transplant/journal", | |
238 | "updatestate", |
|
238 | "updatestate", | |
239 | ]; |
|
239 | ]; | |
240 | if has_unfinished_merge(repo)? { |
|
240 | if has_unfinished_merge(repo)? { | |
241 | return Ok(true); |
|
241 | return Ok(true); | |
242 | }; |
|
242 | }; | |
243 | for f in known_state_files { |
|
243 | for f in known_state_files { | |
244 | if repo.hg_vfs().join(f).exists() { |
|
244 | if repo.hg_vfs().join(f).exists() { | |
245 | return Ok(true); |
|
245 | return Ok(true); | |
246 | } |
|
246 | } | |
247 | } |
|
247 | } | |
248 | Ok(false) |
|
248 | Ok(false) | |
249 | } |
|
249 | } | |
250 |
|
250 | |||
251 | pub fn run(invocation: &crate::CliInvocation) -> Result<(), CommandError> { |
|
251 | pub fn run(invocation: &crate::CliInvocation) -> Result<(), CommandError> { | |
252 | // TODO: lift these limitations |
|
252 | // TODO: lift these limitations | |
253 | if invocation |
|
253 | if invocation | |
254 | .config |
|
254 | .config | |
255 | .get(b"commands", b"status.terse") |
|
255 | .get(b"commands", b"status.terse") | |
256 | .is_some() |
|
256 | .is_some() | |
257 | { |
|
257 | { | |
258 | return Err(CommandError::unsupported( |
|
258 | return Err(CommandError::unsupported( | |
259 | "status.terse is not yet supported with rhg status", |
|
259 | "status.terse is not yet supported with rhg status", | |
260 | )); |
|
260 | )); | |
261 | } |
|
261 | } | |
262 |
|
262 | |||
263 | let ui = invocation.ui; |
|
263 | let ui = invocation.ui; | |
264 | let config = invocation.config; |
|
264 | let config = invocation.config; | |
265 | let args = invocation.subcommand_args; |
|
265 | let args = invocation.subcommand_args; | |
266 |
|
266 | |||
267 | let revs = args.get_many::<String>("rev"); |
|
267 | let revs = args.get_many::<String>("rev"); | |
268 | let print0 = args.get_flag("print0"); |
|
268 | let print0 = args.get_flag("print0"); | |
269 | let verbose = args.get_flag("verbose") |
|
269 | let verbose = args.get_flag("verbose") | |
270 | || config.get_bool(b"ui", b"verbose")? |
|
270 | || config.get_bool(b"ui", b"verbose")? | |
271 | || config.get_bool(b"commands", b"status.verbose")?; |
|
271 | || config.get_bool(b"commands", b"status.verbose")?; | |
272 | let verbose = verbose && !print0; |
|
272 | let verbose = verbose && !print0; | |
273 |
|
273 | |||
274 | let all = args.get_flag("all"); |
|
274 | let all = args.get_flag("all"); | |
275 | let display_states = if all { |
|
275 | let display_states = if all { | |
276 | // TODO when implementing `--quiet`: it excludes clean files |
|
276 | // TODO when implementing `--quiet`: it excludes clean files | |
277 | // from `--all` |
|
277 | // from `--all` | |
278 | ALL_DISPLAY_STATES |
|
278 | ALL_DISPLAY_STATES | |
279 | } else { |
|
279 | } else { | |
280 | let requested = DisplayStates { |
|
280 | let requested = DisplayStates { | |
281 | modified: args.get_flag("modified"), |
|
281 | modified: args.get_flag("modified"), | |
282 | added: args.get_flag("added"), |
|
282 | added: args.get_flag("added"), | |
283 | removed: args.get_flag("removed"), |
|
283 | removed: args.get_flag("removed"), | |
284 | clean: args.get_flag("clean"), |
|
284 | clean: args.get_flag("clean"), | |
285 | deleted: args.get_flag("deleted"), |
|
285 | deleted: args.get_flag("deleted"), | |
286 | unknown: args.get_flag("unknown"), |
|
286 | unknown: args.get_flag("unknown"), | |
287 | ignored: args.get_flag("ignored"), |
|
287 | ignored: args.get_flag("ignored"), | |
288 | }; |
|
288 | }; | |
289 | if requested.is_empty() { |
|
289 | if requested.is_empty() { | |
290 | DEFAULT_DISPLAY_STATES |
|
290 | DEFAULT_DISPLAY_STATES | |
291 | } else { |
|
291 | } else { | |
292 | requested |
|
292 | requested | |
293 | } |
|
293 | } | |
294 | }; |
|
294 | }; | |
295 | let no_status = args.get_flag("no-status"); |
|
295 | let no_status = args.get_flag("no-status"); | |
296 | let list_copies = all |
|
296 | let list_copies = all | |
297 | || args.get_flag("copies") |
|
297 | || args.get_flag("copies") | |
298 | || config.get_bool(b"ui", b"statuscopies")?; |
|
298 | || config.get_bool(b"ui", b"statuscopies")?; | |
299 |
|
299 | |||
300 | let repo = invocation.repo?; |
|
300 | let repo = invocation.repo?; | |
301 | let revpair = parse_revpair(repo, revs.map(|i| i.cloned().collect()))?; |
|
301 | let revpair = parse_revpair(repo, revs.map(|i| i.cloned().collect()))?; | |
302 |
|
302 | |||
303 | if verbose && has_unfinished_state(repo)? { |
|
303 | if verbose && has_unfinished_state(repo)? { | |
304 | return Err(CommandError::unsupported( |
|
304 | return Err(CommandError::unsupported( | |
305 | "verbose status output is not supported by rhg (and is needed because we're in an unfinished operation)", |
|
305 | "verbose status output is not supported by rhg (and is needed because we're in an unfinished operation)", | |
306 | )); |
|
306 | )); | |
307 | } |
|
307 | } | |
308 |
|
308 | |||
309 | let mut dmap = repo.dirstate_map_mut()?; |
|
309 | let mut dmap = repo.dirstate_map_mut()?; | |
310 |
|
310 | |||
311 | let check_exec = hg::checkexec::check_exec(repo.working_directory_path()); |
|
311 | let check_exec = hg::checkexec::check_exec(repo.working_directory_path()); | |
312 |
|
312 | |||
313 | let options = StatusOptions { |
|
313 | let options = StatusOptions { | |
314 | check_exec, |
|
314 | check_exec, | |
315 | list_clean: display_states.clean, |
|
315 | list_clean: display_states.clean, | |
316 | list_unknown: display_states.unknown, |
|
316 | list_unknown: display_states.unknown, | |
317 | list_ignored: display_states.ignored, |
|
317 | list_ignored: display_states.ignored, | |
318 | list_copies, |
|
318 | list_copies, | |
319 | collect_traversed_dirs: false, |
|
319 | collect_traversed_dirs: false, | |
320 | }; |
|
320 | }; | |
321 |
|
321 | |||
322 | type StatusResult<'a> = |
|
322 | type StatusResult<'a> = | |
323 | Result<(DirstateStatus<'a>, Vec<PatternFileWarning>), StatusError>; |
|
323 | Result<(DirstateStatus<'a>, Vec<PatternFileWarning>), StatusError>; | |
324 |
|
324 | |||
325 | let relative_status = config |
|
325 | let relative_status = config | |
326 | .get_option(b"commands", b"status.relative")? |
|
326 | .get_option(b"commands", b"status.relative")? | |
327 | .expect("commands.status.relative should have a default value"); |
|
327 | .expect("commands.status.relative should have a default value"); | |
328 |
|
328 | |||
329 | let relativize_paths = relative_status || { |
|
329 | let relativize_paths = relative_status || { | |
330 | // See in Python code with `getuipathfn` usage in `commands.py`. |
|
330 | // See in Python code with `getuipathfn` usage in `commands.py`. | |
331 | let legacy_relative_behavior = args.contains_id("file"); |
|
331 | let legacy_relative_behavior = args.contains_id("file"); | |
332 | match relative_paths(invocation.config)? { |
|
332 | match relative_paths(invocation.config)? { | |
333 | RelativePaths::Legacy => legacy_relative_behavior, |
|
333 | RelativePaths::Legacy => legacy_relative_behavior, | |
334 | RelativePaths::Bool(v) => v, |
|
334 | RelativePaths::Bool(v) => v, | |
335 | } |
|
335 | } | |
336 | }; |
|
336 | }; | |
337 |
|
337 | |||
338 | let mut output = DisplayStatusPaths { |
|
338 | let mut output = DisplayStatusPaths { | |
339 | ui, |
|
339 | ui, | |
340 | no_status, |
|
340 | no_status, | |
341 | relativize: if relativize_paths { |
|
341 | relativize: if relativize_paths { | |
342 | Some(RelativizePaths::new(repo)?) |
|
342 | Some(RelativizePaths::new(repo)?) | |
343 | } else { |
|
343 | } else { | |
344 | None |
|
344 | None | |
345 | }, |
|
345 | }, | |
346 | print0, |
|
346 | print0, | |
347 | }; |
|
347 | }; | |
348 |
|
348 | |||
349 | let after_status = |res: StatusResult| -> Result<_, CommandError> { |
|
349 | let after_status = |res: StatusResult| -> Result<_, CommandError> { | |
350 | let (mut ds_status, pattern_warnings) = res?; |
|
350 | let (mut ds_status, pattern_warnings) = res?; | |
351 | for warning in pattern_warnings { |
|
351 | for warning in pattern_warnings { | |
352 | ui.write_stderr(&format_pattern_file_warning(&warning, repo))?; |
|
352 | ui.write_stderr(&format_pattern_file_warning(&warning, repo))?; | |
353 | } |
|
353 | } | |
354 |
|
354 | |||
355 | for (path, error) in take(&mut ds_status.bad) { |
|
355 | for (path, error) in take(&mut ds_status.bad) { | |
356 | let error = match error { |
|
356 | let error = match error { | |
357 | hg::BadMatch::OsError(code) => { |
|
357 | hg::BadMatch::OsError(code) => { | |
358 | std::io::Error::from_raw_os_error(code).to_string() |
|
358 | std::io::Error::from_raw_os_error(code).to_string() | |
359 | } |
|
359 | } | |
360 | hg::BadMatch::BadType(ty) => { |
|
360 | hg::BadMatch::BadType(ty) => { | |
361 | format!("unsupported file type (type is {})", ty) |
|
361 | format!("unsupported file type (type is {})", ty) | |
362 | } |
|
362 | } | |
363 | }; |
|
363 | }; | |
364 | ui.write_stderr(&format_bytes!( |
|
364 | ui.write_stderr(&format_bytes!( | |
365 | b"{}: {}\n", |
|
365 | b"{}: {}\n", | |
366 | path.as_bytes(), |
|
366 | path.as_bytes(), | |
367 | error.as_bytes() |
|
367 | error.as_bytes() | |
368 | ))? |
|
368 | ))? | |
369 | } |
|
369 | } | |
370 | if !ds_status.unsure.is_empty() { |
|
370 | if !ds_status.unsure.is_empty() { | |
371 | info!( |
|
371 | info!( | |
372 | "Files to be rechecked by retrieval from filelog: {:?}", |
|
372 | "Files to be rechecked by retrieval from filelog: {:?}", | |
373 | ds_status.unsure.iter().map(|s| &s.path).collect::<Vec<_>>() |
|
373 | ds_status.unsure.iter().map(|s| &s.path).collect::<Vec<_>>() | |
374 | ); |
|
374 | ); | |
375 | } |
|
375 | } | |
376 | let mut fixup = Vec::new(); |
|
376 | let mut fixup = Vec::new(); | |
377 | if !ds_status.unsure.is_empty() |
|
377 | if !ds_status.unsure.is_empty() | |
378 | && (display_states.modified || display_states.clean) |
|
378 | && (display_states.modified || display_states.clean) | |
379 | { |
|
379 | { | |
380 | let p1 = repo.dirstate_parents()?.p1; |
|
380 | let p1 = repo.dirstate_parents()?.p1; | |
381 | let manifest = repo.manifest_for_node(p1).map_err(|e| { |
|
381 | let manifest = repo.manifest_for_node(p1).map_err(|e| { | |
382 | CommandError::from((e, &*format!("{:x}", p1.short()))) |
|
382 | CommandError::from((e, &*format!("{:x}", p1.short()))) | |
383 | })?; |
|
383 | })?; | |
384 | let working_directory_vfs = repo.working_directory_vfs(); |
|
384 | let working_directory_vfs = repo.working_directory_vfs(); | |
385 | let store_vfs = repo.store_vfs(); |
|
385 | let store_vfs = repo.store_vfs(); | |
|
386 | let revlog_open_options = repo.default_revlog_options(false)?; | |||
386 | let res: Vec<_> = take(&mut ds_status.unsure) |
|
387 | let res: Vec<_> = take(&mut ds_status.unsure) | |
387 | .into_par_iter() |
|
388 | .into_par_iter() | |
388 | .map(|to_check| { |
|
389 | .map(|to_check| { | |
389 | // The compiler seems to get a bit confused with complex |
|
390 | // The compiler seems to get a bit confused with complex | |
390 | // inference when using a parallel iterator + map |
|
391 | // inference when using a parallel iterator + map | |
391 | // + map_err + collect, so let's just inline some of the |
|
392 | // + map_err + collect, so let's just inline some of the | |
392 | // logic. |
|
393 | // logic. | |
393 | match unsure_is_modified( |
|
394 | match unsure_is_modified( | |
394 | working_directory_vfs, |
|
395 | working_directory_vfs, | |
395 | store_vfs, |
|
396 | store_vfs, | |
396 | check_exec, |
|
397 | check_exec, | |
397 | &manifest, |
|
398 | &manifest, | |
398 | &to_check.path, |
|
399 | &to_check.path, | |
|
400 | revlog_open_options, | |||
399 | ) { |
|
401 | ) { | |
400 | Err(HgError::IoError { .. }) => { |
|
402 | Err(HgError::IoError { .. }) => { | |
401 | // IO errors most likely stem from the file being |
|
403 | // IO errors most likely stem from the file being | |
402 | // deleted even though we know it's in the |
|
404 | // deleted even though we know it's in the | |
403 | // dirstate. |
|
405 | // dirstate. | |
404 | Ok((to_check, UnsureOutcome::Deleted)) |
|
406 | Ok((to_check, UnsureOutcome::Deleted)) | |
405 | } |
|
407 | } | |
406 | Ok(outcome) => Ok((to_check, outcome)), |
|
408 | Ok(outcome) => Ok((to_check, outcome)), | |
407 | Err(e) => Err(e), |
|
409 | Err(e) => Err(e), | |
408 | } |
|
410 | } | |
409 | }) |
|
411 | }) | |
410 | .collect::<Result<_, _>>()?; |
|
412 | .collect::<Result<_, _>>()?; | |
411 | for (status_path, outcome) in res.into_iter() { |
|
413 | for (status_path, outcome) in res.into_iter() { | |
412 | match outcome { |
|
414 | match outcome { | |
413 | UnsureOutcome::Clean => { |
|
415 | UnsureOutcome::Clean => { | |
414 | if display_states.clean { |
|
416 | if display_states.clean { | |
415 | ds_status.clean.push(status_path.clone()); |
|
417 | ds_status.clean.push(status_path.clone()); | |
416 | } |
|
418 | } | |
417 | fixup.push(status_path.path.into_owned()) |
|
419 | fixup.push(status_path.path.into_owned()) | |
418 | } |
|
420 | } | |
419 | UnsureOutcome::Modified => { |
|
421 | UnsureOutcome::Modified => { | |
420 | if display_states.modified { |
|
422 | if display_states.modified { | |
421 | ds_status.modified.push(status_path); |
|
423 | ds_status.modified.push(status_path); | |
422 | } |
|
424 | } | |
423 | } |
|
425 | } | |
424 | UnsureOutcome::Deleted => { |
|
426 | UnsureOutcome::Deleted => { | |
425 | if display_states.deleted { |
|
427 | if display_states.deleted { | |
426 | ds_status.deleted.push(status_path); |
|
428 | ds_status.deleted.push(status_path); | |
427 | } |
|
429 | } | |
428 | } |
|
430 | } | |
429 | } |
|
431 | } | |
430 | } |
|
432 | } | |
431 | } |
|
433 | } | |
432 |
|
434 | |||
433 | let dirstate_write_needed = ds_status.dirty; |
|
435 | let dirstate_write_needed = ds_status.dirty; | |
434 | let filesystem_time_at_status_start = |
|
436 | let filesystem_time_at_status_start = | |
435 | ds_status.filesystem_time_at_status_start; |
|
437 | ds_status.filesystem_time_at_status_start; | |
436 |
|
438 | |||
437 | output.output(display_states, ds_status)?; |
|
439 | output.output(display_states, ds_status)?; | |
438 |
|
440 | |||
439 | Ok(( |
|
441 | Ok(( | |
440 | fixup, |
|
442 | fixup, | |
441 | dirstate_write_needed, |
|
443 | dirstate_write_needed, | |
442 | filesystem_time_at_status_start, |
|
444 | filesystem_time_at_status_start, | |
443 | )) |
|
445 | )) | |
444 | }; |
|
446 | }; | |
445 | let (narrow_matcher, narrow_warnings) = narrow::matcher(repo)?; |
|
447 | let (narrow_matcher, narrow_warnings) = narrow::matcher(repo)?; | |
446 |
|
448 | |||
447 | match revpair { |
|
449 | match revpair { | |
448 | Some((rev1, rev2)) => { |
|
450 | Some((rev1, rev2)) => { | |
449 | let mut ds_status = DirstateStatus::default(); |
|
451 | let mut ds_status = DirstateStatus::default(); | |
450 | if list_copies { |
|
452 | if list_copies { | |
451 | return Err(CommandError::unsupported( |
|
453 | return Err(CommandError::unsupported( | |
452 | "status --rev --rev with copy information is not implemented yet", |
|
454 | "status --rev --rev with copy information is not implemented yet", | |
453 | )); |
|
455 | )); | |
454 | } |
|
456 | } | |
455 |
|
457 | |||
456 | let stat = hg::operations::status_rev_rev_no_copies( |
|
458 | let stat = hg::operations::status_rev_rev_no_copies( | |
457 | repo, |
|
459 | repo, | |
458 | rev1, |
|
460 | rev1, | |
459 | rev2, |
|
461 | rev2, | |
460 | narrow_matcher, |
|
462 | narrow_matcher, | |
461 | )?; |
|
463 | )?; | |
462 | for entry in stat.iter() { |
|
464 | for entry in stat.iter() { | |
463 | let (path, status) = entry?; |
|
465 | let (path, status) = entry?; | |
464 | let path = StatusPath { |
|
466 | let path = StatusPath { | |
465 | path: Cow::Borrowed(path), |
|
467 | path: Cow::Borrowed(path), | |
466 | copy_source: None, |
|
468 | copy_source: None, | |
467 | }; |
|
469 | }; | |
468 | match status { |
|
470 | match status { | |
469 | hg::operations::DiffStatus::Removed => { |
|
471 | hg::operations::DiffStatus::Removed => { | |
470 | if display_states.removed { |
|
472 | if display_states.removed { | |
471 | ds_status.removed.push(path) |
|
473 | ds_status.removed.push(path) | |
472 | } |
|
474 | } | |
473 | } |
|
475 | } | |
474 | hg::operations::DiffStatus::Added => { |
|
476 | hg::operations::DiffStatus::Added => { | |
475 | if display_states.added { |
|
477 | if display_states.added { | |
476 | ds_status.added.push(path) |
|
478 | ds_status.added.push(path) | |
477 | } |
|
479 | } | |
478 | } |
|
480 | } | |
479 | hg::operations::DiffStatus::Modified => { |
|
481 | hg::operations::DiffStatus::Modified => { | |
480 | if display_states.modified { |
|
482 | if display_states.modified { | |
481 | ds_status.modified.push(path) |
|
483 | ds_status.modified.push(path) | |
482 | } |
|
484 | } | |
483 | } |
|
485 | } | |
484 | hg::operations::DiffStatus::Matching => { |
|
486 | hg::operations::DiffStatus::Matching => { | |
485 | if display_states.clean { |
|
487 | if display_states.clean { | |
486 | ds_status.clean.push(path) |
|
488 | ds_status.clean.push(path) | |
487 | } |
|
489 | } | |
488 | } |
|
490 | } | |
489 | } |
|
491 | } | |
490 | } |
|
492 | } | |
491 | output.output(display_states, ds_status)?; |
|
493 | output.output(display_states, ds_status)?; | |
492 | return Ok(()); |
|
494 | return Ok(()); | |
493 | } |
|
495 | } | |
494 | None => (), |
|
496 | None => (), | |
495 | } |
|
497 | } | |
496 |
|
498 | |||
497 | let (sparse_matcher, sparse_warnings) = sparse::matcher(repo)?; |
|
499 | let (sparse_matcher, sparse_warnings) = sparse::matcher(repo)?; | |
498 | let matcher = match (repo.has_narrow(), repo.has_sparse()) { |
|
500 | let matcher = match (repo.has_narrow(), repo.has_sparse()) { | |
499 | (true, true) => { |
|
501 | (true, true) => { | |
500 | Box::new(IntersectionMatcher::new(narrow_matcher, sparse_matcher)) |
|
502 | Box::new(IntersectionMatcher::new(narrow_matcher, sparse_matcher)) | |
501 | } |
|
503 | } | |
502 | (true, false) => narrow_matcher, |
|
504 | (true, false) => narrow_matcher, | |
503 | (false, true) => sparse_matcher, |
|
505 | (false, true) => sparse_matcher, | |
504 | (false, false) => Box::new(AlwaysMatcher), |
|
506 | (false, false) => Box::new(AlwaysMatcher), | |
505 | }; |
|
507 | }; | |
506 | let matcher = match args.get_many::<std::ffi::OsString>("file") { |
|
508 | let matcher = match args.get_many::<std::ffi::OsString>("file") { | |
507 | None => matcher, |
|
509 | None => matcher, | |
508 | Some(files) => { |
|
510 | Some(files) => { | |
509 | let patterns: Vec<Vec<u8>> = files |
|
511 | let patterns: Vec<Vec<u8>> = files | |
510 | .filter(|s| !s.is_empty()) |
|
512 | .filter(|s| !s.is_empty()) | |
511 | .map(get_bytes_from_os_str) |
|
513 | .map(get_bytes_from_os_str) | |
512 | .collect(); |
|
514 | .collect(); | |
513 | for file in &patterns { |
|
515 | for file in &patterns { | |
514 | if file.starts_with(b"set:") { |
|
516 | if file.starts_with(b"set:") { | |
515 | return Err(CommandError::unsupported("fileset")); |
|
517 | return Err(CommandError::unsupported("fileset")); | |
516 | } |
|
518 | } | |
517 | } |
|
519 | } | |
518 | let cwd = hg::utils::current_dir()?; |
|
520 | let cwd = hg::utils::current_dir()?; | |
519 | let root = repo.working_directory_path(); |
|
521 | let root = repo.working_directory_path(); | |
520 | let ignore_patterns = parse_pattern_args(patterns, &cwd, root)?; |
|
522 | let ignore_patterns = parse_pattern_args(patterns, &cwd, root)?; | |
521 | let files_matcher = |
|
523 | let files_matcher = | |
522 | hg::matchers::PatternMatcher::new(ignore_patterns)?; |
|
524 | hg::matchers::PatternMatcher::new(ignore_patterns)?; | |
523 | Box::new(IntersectionMatcher::new( |
|
525 | Box::new(IntersectionMatcher::new( | |
524 | Box::new(files_matcher), |
|
526 | Box::new(files_matcher), | |
525 | matcher, |
|
527 | matcher, | |
526 | )) |
|
528 | )) | |
527 | } |
|
529 | } | |
528 | }; |
|
530 | }; | |
529 |
|
531 | |||
530 | print_narrow_sparse_warnings( |
|
532 | print_narrow_sparse_warnings( | |
531 | &narrow_warnings, |
|
533 | &narrow_warnings, | |
532 | &sparse_warnings, |
|
534 | &sparse_warnings, | |
533 | ui, |
|
535 | ui, | |
534 | repo, |
|
536 | repo, | |
535 | )?; |
|
537 | )?; | |
536 | let (fixup, mut dirstate_write_needed, filesystem_time_at_status_start) = |
|
538 | let (fixup, mut dirstate_write_needed, filesystem_time_at_status_start) = | |
537 | dmap.with_status( |
|
539 | dmap.with_status( | |
538 | matcher.as_ref(), |
|
540 | matcher.as_ref(), | |
539 | repo.working_directory_path().to_owned(), |
|
541 | repo.working_directory_path().to_owned(), | |
540 | ignore_files(repo, config), |
|
542 | ignore_files(repo, config), | |
541 | options, |
|
543 | options, | |
542 | after_status, |
|
544 | after_status, | |
543 | )?; |
|
545 | )?; | |
544 |
|
546 | |||
545 | // Development config option to test write races |
|
547 | // Development config option to test write races | |
546 | if let Err(e) = |
|
548 | if let Err(e) = | |
547 | debug_wait_for_file(config, "status.pre-dirstate-write-file") |
|
549 | debug_wait_for_file(config, "status.pre-dirstate-write-file") | |
548 | { |
|
550 | { | |
549 | ui.write_stderr(e.as_bytes()).ok(); |
|
551 | ui.write_stderr(e.as_bytes()).ok(); | |
550 | } |
|
552 | } | |
551 |
|
553 | |||
552 | if (fixup.is_empty() || filesystem_time_at_status_start.is_none()) |
|
554 | if (fixup.is_empty() || filesystem_time_at_status_start.is_none()) | |
553 | && !dirstate_write_needed |
|
555 | && !dirstate_write_needed | |
554 | { |
|
556 | { | |
555 | // Nothing to update |
|
557 | // Nothing to update | |
556 | return Ok(()); |
|
558 | return Ok(()); | |
557 | } |
|
559 | } | |
558 |
|
560 | |||
559 | // Update the dirstate on disk if we can |
|
561 | // Update the dirstate on disk if we can | |
560 | let with_lock_result = |
|
562 | let with_lock_result = | |
561 | repo.try_with_wlock_no_wait(|| -> Result<(), CommandError> { |
|
563 | repo.try_with_wlock_no_wait(|| -> Result<(), CommandError> { | |
562 | if let Some(mtime_boundary) = filesystem_time_at_status_start { |
|
564 | if let Some(mtime_boundary) = filesystem_time_at_status_start { | |
563 | for hg_path in fixup { |
|
565 | for hg_path in fixup { | |
564 | use std::os::unix::fs::MetadataExt; |
|
566 | use std::os::unix::fs::MetadataExt; | |
565 | let fs_path = hg_path_to_path_buf(&hg_path) |
|
567 | let fs_path = hg_path_to_path_buf(&hg_path) | |
566 | .expect("HgPath conversion"); |
|
568 | .expect("HgPath conversion"); | |
567 | // Specifically do not reuse `fs_metadata` from |
|
569 | // Specifically do not reuse `fs_metadata` from | |
568 | // `unsure_is_clean` which was needed before reading |
|
570 | // `unsure_is_clean` which was needed before reading | |
569 | // contents. Here we access metadata again after reading |
|
571 | // contents. Here we access metadata again after reading | |
570 | // content, in case it changed in the meantime. |
|
572 | // content, in case it changed in the meantime. | |
571 | let metadata_res = repo |
|
573 | let metadata_res = repo | |
572 | .working_directory_vfs() |
|
574 | .working_directory_vfs() | |
573 | .symlink_metadata(&fs_path); |
|
575 | .symlink_metadata(&fs_path); | |
574 | let fs_metadata = match metadata_res { |
|
576 | let fs_metadata = match metadata_res { | |
575 | Ok(meta) => meta, |
|
577 | Ok(meta) => meta, | |
576 | Err(err) => match err { |
|
578 | Err(err) => match err { | |
577 | HgError::IoError { .. } => { |
|
579 | HgError::IoError { .. } => { | |
578 | // The file has probably been deleted. In any |
|
580 | // The file has probably been deleted. In any | |
579 | // case, it was in the dirstate before, so |
|
581 | // case, it was in the dirstate before, so | |
580 | // let's ignore the error. |
|
582 | // let's ignore the error. | |
581 | continue; |
|
583 | continue; | |
582 | } |
|
584 | } | |
583 | _ => return Err(err.into()), |
|
585 | _ => return Err(err.into()), | |
584 | }, |
|
586 | }, | |
585 | }; |
|
587 | }; | |
586 | if let Some(mtime) = |
|
588 | if let Some(mtime) = | |
587 | TruncatedTimestamp::for_reliable_mtime_of( |
|
589 | TruncatedTimestamp::for_reliable_mtime_of( | |
588 | &fs_metadata, |
|
590 | &fs_metadata, | |
589 | &mtime_boundary, |
|
591 | &mtime_boundary, | |
590 | ) |
|
592 | ) | |
591 | .when_reading_file(&fs_path)? |
|
593 | .when_reading_file(&fs_path)? | |
592 | { |
|
594 | { | |
593 | let mode = fs_metadata.mode(); |
|
595 | let mode = fs_metadata.mode(); | |
594 | let size = fs_metadata.len(); |
|
596 | let size = fs_metadata.len(); | |
595 | dmap.set_clean(&hg_path, mode, size as u32, mtime)?; |
|
597 | dmap.set_clean(&hg_path, mode, size as u32, mtime)?; | |
596 | dirstate_write_needed = true |
|
598 | dirstate_write_needed = true | |
597 | } |
|
599 | } | |
598 | } |
|
600 | } | |
599 | } |
|
601 | } | |
600 | drop(dmap); // Avoid "already mutably borrowed" RefCell panics |
|
602 | drop(dmap); // Avoid "already mutably borrowed" RefCell panics | |
601 | if dirstate_write_needed { |
|
603 | if dirstate_write_needed { | |
602 | repo.write_dirstate()? |
|
604 | repo.write_dirstate()? | |
603 | } |
|
605 | } | |
604 | Ok(()) |
|
606 | Ok(()) | |
605 | }); |
|
607 | }); | |
606 | match with_lock_result { |
|
608 | match with_lock_result { | |
607 | Ok(closure_result) => closure_result?, |
|
609 | Ok(closure_result) => closure_result?, | |
608 | Err(LockError::AlreadyHeld) => { |
|
610 | Err(LockError::AlreadyHeld) => { | |
609 | // Not updating the dirstate is not ideal but not critical: |
|
611 | // Not updating the dirstate is not ideal but not critical: | |
610 | // don’t keep our caller waiting until some other Mercurial |
|
612 | // don’t keep our caller waiting until some other Mercurial | |
611 | // process releases the lock. |
|
613 | // process releases the lock. | |
612 | log::info!("not writing dirstate from `status`: lock is held") |
|
614 | log::info!("not writing dirstate from `status`: lock is held") | |
613 | } |
|
615 | } | |
614 | Err(LockError::Other(HgError::IoError { error, .. })) |
|
616 | Err(LockError::Other(HgError::IoError { error, .. })) | |
615 | if error.kind() == io::ErrorKind::PermissionDenied => |
|
617 | if error.kind() == io::ErrorKind::PermissionDenied => | |
616 | { |
|
618 | { | |
617 | // `hg status` on a read-only repository is fine |
|
619 | // `hg status` on a read-only repository is fine | |
618 | } |
|
620 | } | |
619 | Err(LockError::Other(error)) => { |
|
621 | Err(LockError::Other(error)) => { | |
620 | // Report other I/O errors |
|
622 | // Report other I/O errors | |
621 | Err(error)? |
|
623 | Err(error)? | |
622 | } |
|
624 | } | |
623 | } |
|
625 | } | |
624 | Ok(()) |
|
626 | Ok(()) | |
625 | } |
|
627 | } | |
626 |
|
628 | |||
627 | fn ignore_files(repo: &Repo, config: &Config) -> Vec<PathBuf> { |
|
629 | fn ignore_files(repo: &Repo, config: &Config) -> Vec<PathBuf> { | |
628 | let mut ignore_files = Vec::new(); |
|
630 | let mut ignore_files = Vec::new(); | |
629 | let repo_ignore = repo.working_directory_vfs().join(".hgignore"); |
|
631 | let repo_ignore = repo.working_directory_vfs().join(".hgignore"); | |
630 | if repo_ignore.exists() { |
|
632 | if repo_ignore.exists() { | |
631 | ignore_files.push(repo_ignore) |
|
633 | ignore_files.push(repo_ignore) | |
632 | } |
|
634 | } | |
633 | for (key, value) in config.iter_section(b"ui") { |
|
635 | for (key, value) in config.iter_section(b"ui") { | |
634 | if key == b"ignore" || key.starts_with(b"ignore.") { |
|
636 | if key == b"ignore" || key.starts_with(b"ignore.") { | |
635 | let path = get_path_from_bytes(value); |
|
637 | let path = get_path_from_bytes(value); | |
636 | // TODO: expand "~/" and environment variable here, like Python |
|
638 | // TODO: expand "~/" and environment variable here, like Python | |
637 | // does with `os.path.expanduser` and `os.path.expandvars` |
|
639 | // does with `os.path.expanduser` and `os.path.expandvars` | |
638 |
|
640 | |||
639 | let joined = repo.working_directory_path().join(path); |
|
641 | let joined = repo.working_directory_path().join(path); | |
640 | ignore_files.push(joined); |
|
642 | ignore_files.push(joined); | |
641 | } |
|
643 | } | |
642 | } |
|
644 | } | |
643 | ignore_files |
|
645 | ignore_files | |
644 | } |
|
646 | } | |
645 |
|
647 | |||
646 | struct DisplayStatusPaths<'a> { |
|
648 | struct DisplayStatusPaths<'a> { | |
647 | ui: &'a Ui, |
|
649 | ui: &'a Ui, | |
648 | no_status: bool, |
|
650 | no_status: bool, | |
649 | relativize: Option<RelativizePaths>, |
|
651 | relativize: Option<RelativizePaths>, | |
650 | print0: bool, |
|
652 | print0: bool, | |
651 | } |
|
653 | } | |
652 |
|
654 | |||
653 | impl DisplayStatusPaths<'_> { |
|
655 | impl DisplayStatusPaths<'_> { | |
654 | // Probably more elegant to use a Deref or Borrow trait rather than |
|
656 | // Probably more elegant to use a Deref or Borrow trait rather than | |
655 | // harcode HgPathBuf, but probably not really useful at this point |
|
657 | // harcode HgPathBuf, but probably not really useful at this point | |
656 | fn display( |
|
658 | fn display( | |
657 | &self, |
|
659 | &self, | |
658 | status_prefix: &[u8], |
|
660 | status_prefix: &[u8], | |
659 | label: &'static str, |
|
661 | label: &'static str, | |
660 | mut paths: Vec<StatusPath<'_>>, |
|
662 | mut paths: Vec<StatusPath<'_>>, | |
661 | ) -> Result<(), CommandError> { |
|
663 | ) -> Result<(), CommandError> { | |
662 | paths.sort_unstable(); |
|
664 | paths.sort_unstable(); | |
663 | // TODO: get the stdout lock once for the whole loop |
|
665 | // TODO: get the stdout lock once for the whole loop | |
664 | // instead of in each write |
|
666 | // instead of in each write | |
665 | for StatusPath { path, copy_source } in paths { |
|
667 | for StatusPath { path, copy_source } in paths { | |
666 | let relative_path; |
|
668 | let relative_path; | |
667 | let relative_source; |
|
669 | let relative_source; | |
668 | let (path, copy_source) = if let Some(relativize) = |
|
670 | let (path, copy_source) = if let Some(relativize) = | |
669 | &self.relativize |
|
671 | &self.relativize | |
670 | { |
|
672 | { | |
671 | relative_path = relativize.relativize(&path); |
|
673 | relative_path = relativize.relativize(&path); | |
672 | relative_source = |
|
674 | relative_source = | |
673 | copy_source.as_ref().map(|s| relativize.relativize(s)); |
|
675 | copy_source.as_ref().map(|s| relativize.relativize(s)); | |
674 | (&*relative_path, relative_source.as_deref()) |
|
676 | (&*relative_path, relative_source.as_deref()) | |
675 | } else { |
|
677 | } else { | |
676 | (path.as_bytes(), copy_source.as_ref().map(|s| s.as_bytes())) |
|
678 | (path.as_bytes(), copy_source.as_ref().map(|s| s.as_bytes())) | |
677 | }; |
|
679 | }; | |
678 | // TODO: Add a way to use `write_bytes!` instead of `format_bytes!` |
|
680 | // TODO: Add a way to use `write_bytes!` instead of `format_bytes!` | |
679 | // in order to stream to stdout instead of allocating an |
|
681 | // in order to stream to stdout instead of allocating an | |
680 | // itermediate `Vec<u8>`. |
|
682 | // itermediate `Vec<u8>`. | |
681 | if !self.no_status { |
|
683 | if !self.no_status { | |
682 | self.ui.write_stdout_labelled(status_prefix, label)? |
|
684 | self.ui.write_stdout_labelled(status_prefix, label)? | |
683 | } |
|
685 | } | |
684 | let linebreak = if self.print0 { b"\x00" } else { b"\n" }; |
|
686 | let linebreak = if self.print0 { b"\x00" } else { b"\n" }; | |
685 | self.ui.write_stdout_labelled( |
|
687 | self.ui.write_stdout_labelled( | |
686 | &format_bytes!(b"{}{}", path, linebreak), |
|
688 | &format_bytes!(b"{}{}", path, linebreak), | |
687 | label, |
|
689 | label, | |
688 | )?; |
|
690 | )?; | |
689 | if let Some(source) = copy_source.filter(|_| !self.no_status) { |
|
691 | if let Some(source) = copy_source.filter(|_| !self.no_status) { | |
690 | let label = "status.copied"; |
|
692 | let label = "status.copied"; | |
691 | self.ui.write_stdout_labelled( |
|
693 | self.ui.write_stdout_labelled( | |
692 | &format_bytes!(b" {}{}", source, linebreak), |
|
694 | &format_bytes!(b" {}{}", source, linebreak), | |
693 | label, |
|
695 | label, | |
694 | )? |
|
696 | )? | |
695 | } |
|
697 | } | |
696 | } |
|
698 | } | |
697 | Ok(()) |
|
699 | Ok(()) | |
698 | } |
|
700 | } | |
699 |
|
701 | |||
700 | fn output( |
|
702 | fn output( | |
701 | &mut self, |
|
703 | &mut self, | |
702 | display_states: DisplayStates, |
|
704 | display_states: DisplayStates, | |
703 | ds_status: DirstateStatus, |
|
705 | ds_status: DirstateStatus, | |
704 | ) -> Result<(), CommandError> { |
|
706 | ) -> Result<(), CommandError> { | |
705 | if display_states.modified { |
|
707 | if display_states.modified { | |
706 | self.display(b"M ", "status.modified", ds_status.modified)?; |
|
708 | self.display(b"M ", "status.modified", ds_status.modified)?; | |
707 | } |
|
709 | } | |
708 | if display_states.added { |
|
710 | if display_states.added { | |
709 | self.display(b"A ", "status.added", ds_status.added)?; |
|
711 | self.display(b"A ", "status.added", ds_status.added)?; | |
710 | } |
|
712 | } | |
711 | if display_states.removed { |
|
713 | if display_states.removed { | |
712 | self.display(b"R ", "status.removed", ds_status.removed)?; |
|
714 | self.display(b"R ", "status.removed", ds_status.removed)?; | |
713 | } |
|
715 | } | |
714 | if display_states.deleted { |
|
716 | if display_states.deleted { | |
715 | self.display(b"! ", "status.deleted", ds_status.deleted)?; |
|
717 | self.display(b"! ", "status.deleted", ds_status.deleted)?; | |
716 | } |
|
718 | } | |
717 | if display_states.unknown { |
|
719 | if display_states.unknown { | |
718 | self.display(b"? ", "status.unknown", ds_status.unknown)?; |
|
720 | self.display(b"? ", "status.unknown", ds_status.unknown)?; | |
719 | } |
|
721 | } | |
720 | if display_states.ignored { |
|
722 | if display_states.ignored { | |
721 | self.display(b"I ", "status.ignored", ds_status.ignored)?; |
|
723 | self.display(b"I ", "status.ignored", ds_status.ignored)?; | |
722 | } |
|
724 | } | |
723 | if display_states.clean { |
|
725 | if display_states.clean { | |
724 | self.display(b"C ", "status.clean", ds_status.clean)?; |
|
726 | self.display(b"C ", "status.clean", ds_status.clean)?; | |
725 | } |
|
727 | } | |
726 | Ok(()) |
|
728 | Ok(()) | |
727 | } |
|
729 | } | |
728 | } |
|
730 | } | |
729 |
|
731 | |||
730 | /// Outcome of the additional check for an ambiguous tracked file |
|
732 | /// Outcome of the additional check for an ambiguous tracked file | |
731 | enum UnsureOutcome { |
|
733 | enum UnsureOutcome { | |
732 | /// The file is actually clean |
|
734 | /// The file is actually clean | |
733 | Clean, |
|
735 | Clean, | |
734 | /// The file has been modified |
|
736 | /// The file has been modified | |
735 | Modified, |
|
737 | Modified, | |
736 | /// The file was deleted on disk (or became another type of fs entry) |
|
738 | /// The file was deleted on disk (or became another type of fs entry) | |
737 | Deleted, |
|
739 | Deleted, | |
738 | } |
|
740 | } | |
739 |
|
741 | |||
740 | /// Check if a file is modified by comparing actual repo store and file system. |
|
742 | /// Check if a file is modified by comparing actual repo store and file system. | |
741 | /// |
|
743 | /// | |
742 | /// This meant to be used for those that the dirstate cannot resolve, due |
|
744 | /// This meant to be used for those that the dirstate cannot resolve, due | |
743 | /// to time resolution limits. |
|
745 | /// to time resolution limits. | |
744 | fn unsure_is_modified( |
|
746 | fn unsure_is_modified( | |
745 | working_directory_vfs: hg::vfs::Vfs, |
|
747 | working_directory_vfs: hg::vfs::Vfs, | |
746 | store_vfs: hg::vfs::Vfs, |
|
748 | store_vfs: hg::vfs::Vfs, | |
747 | check_exec: bool, |
|
749 | check_exec: bool, | |
748 | manifest: &Manifest, |
|
750 | manifest: &Manifest, | |
749 | hg_path: &HgPath, |
|
751 | hg_path: &HgPath, | |
|
752 | revlog_open_options: RevlogOpenOptions, | |||
750 | ) -> Result<UnsureOutcome, HgError> { |
|
753 | ) -> Result<UnsureOutcome, HgError> { | |
751 | let vfs = working_directory_vfs; |
|
754 | let vfs = working_directory_vfs; | |
752 | let fs_path = hg_path_to_path_buf(hg_path).expect("HgPath conversion"); |
|
755 | let fs_path = hg_path_to_path_buf(hg_path).expect("HgPath conversion"); | |
753 | let fs_metadata = vfs.symlink_metadata(&fs_path)?; |
|
756 | let fs_metadata = vfs.symlink_metadata(&fs_path)?; | |
754 | let is_symlink = fs_metadata.file_type().is_symlink(); |
|
757 | let is_symlink = fs_metadata.file_type().is_symlink(); | |
755 |
|
758 | |||
756 | let entry = manifest |
|
759 | let entry = manifest | |
757 | .find_by_path(hg_path)? |
|
760 | .find_by_path(hg_path)? | |
758 | .expect("ambgious file not in p1"); |
|
761 | .expect("ambgious file not in p1"); | |
759 |
|
762 | |||
760 | // TODO: Also account for `FALLBACK_SYMLINK` and `FALLBACK_EXEC` from the |
|
763 | // TODO: Also account for `FALLBACK_SYMLINK` and `FALLBACK_EXEC` from the | |
761 | // dirstate |
|
764 | // dirstate | |
762 | let fs_flags = if is_symlink { |
|
765 | let fs_flags = if is_symlink { | |
763 | Some(b'l') |
|
766 | Some(b'l') | |
764 | } else if check_exec && has_exec_bit(&fs_metadata) { |
|
767 | } else if check_exec && has_exec_bit(&fs_metadata) { | |
765 | Some(b'x') |
|
768 | Some(b'x') | |
766 | } else { |
|
769 | } else { | |
767 | None |
|
770 | None | |
768 | }; |
|
771 | }; | |
769 |
|
772 | |||
770 | let entry_flags = if check_exec { |
|
773 | let entry_flags = if check_exec { | |
771 | entry.flags |
|
774 | entry.flags | |
772 | } else if entry.flags == Some(b'x') { |
|
775 | } else if entry.flags == Some(b'x') { | |
773 | None |
|
776 | None | |
774 | } else { |
|
777 | } else { | |
775 | entry.flags |
|
778 | entry.flags | |
776 | }; |
|
779 | }; | |
777 |
|
780 | |||
778 | if entry_flags != fs_flags { |
|
781 | if entry_flags != fs_flags { | |
779 | return Ok(UnsureOutcome::Modified); |
|
782 | return Ok(UnsureOutcome::Modified); | |
780 | } |
|
783 | } | |
781 |
let filelog = hg::filelog::Filelog::open_vfs( |
|
784 | let filelog = hg::filelog::Filelog::open_vfs( | |
|
785 | &store_vfs, | |||
|
786 | hg_path, | |||
|
787 | revlog_open_options, | |||
|
788 | )?; | |||
782 | let fs_len = fs_metadata.len(); |
|
789 | let fs_len = fs_metadata.len(); | |
783 | let file_node = entry.node_id()?; |
|
790 | let file_node = entry.node_id()?; | |
784 | let filelog_entry = filelog.entry_for_node(file_node).map_err(|_| { |
|
791 | let filelog_entry = filelog.entry_for_node(file_node).map_err(|_| { | |
785 | HgError::corrupted(format!( |
|
792 | HgError::corrupted(format!( | |
786 | "filelog {:?} missing node {:?} from manifest", |
|
793 | "filelog {:?} missing node {:?} from manifest", | |
787 | hg_path, file_node |
|
794 | hg_path, file_node | |
788 | )) |
|
795 | )) | |
789 | })?; |
|
796 | })?; | |
790 | if filelog_entry.file_data_len_not_equal_to(fs_len) { |
|
797 | if filelog_entry.file_data_len_not_equal_to(fs_len) { | |
791 | // No need to read file contents: |
|
798 | // No need to read file contents: | |
792 | // it cannot be equal if it has a different length. |
|
799 | // it cannot be equal if it has a different length. | |
793 | return Ok(UnsureOutcome::Modified); |
|
800 | return Ok(UnsureOutcome::Modified); | |
794 | } |
|
801 | } | |
795 |
|
802 | |||
796 | let p1_filelog_data = filelog_entry.data()?; |
|
803 | let p1_filelog_data = filelog_entry.data()?; | |
797 | let p1_contents = p1_filelog_data.file_data()?; |
|
804 | let p1_contents = p1_filelog_data.file_data()?; | |
798 | if p1_contents.len() as u64 != fs_len { |
|
805 | if p1_contents.len() as u64 != fs_len { | |
799 | // No need to read file contents: |
|
806 | // No need to read file contents: | |
800 | // it cannot be equal if it has a different length. |
|
807 | // it cannot be equal if it has a different length. | |
801 | return Ok(UnsureOutcome::Modified); |
|
808 | return Ok(UnsureOutcome::Modified); | |
802 | } |
|
809 | } | |
803 |
|
810 | |||
804 | let fs_contents = if is_symlink { |
|
811 | let fs_contents = if is_symlink { | |
805 | get_bytes_from_os_string(vfs.read_link(fs_path)?.into_os_string()) |
|
812 | get_bytes_from_os_string(vfs.read_link(fs_path)?.into_os_string()) | |
806 | } else { |
|
813 | } else { | |
807 | vfs.read(fs_path)? |
|
814 | vfs.read(fs_path)? | |
808 | }; |
|
815 | }; | |
809 |
|
816 | |||
810 | Ok(if p1_contents != &*fs_contents { |
|
817 | Ok(if p1_contents != &*fs_contents { | |
811 | UnsureOutcome::Modified |
|
818 | UnsureOutcome::Modified | |
812 | } else { |
|
819 | } else { | |
813 | UnsureOutcome::Clean |
|
820 | UnsureOutcome::Clean | |
814 | }) |
|
821 | }) | |
815 | } |
|
822 | } |
@@ -1,60 +1,63 b'' | |||||
|
1 | import struct | |||
1 | import unittest |
|
2 | import unittest | |
2 |
|
3 | |||
3 | try: |
|
4 | try: | |
4 | from mercurial import rustext |
|
5 | from mercurial import rustext | |
5 |
|
6 | |||
6 | rustext.__name__ # trigger immediate actual import |
|
7 | rustext.__name__ # trigger immediate actual import | |
7 | except ImportError: |
|
8 | except ImportError: | |
8 | rustext = None |
|
9 | rustext = None | |
9 | else: |
|
10 | else: | |
10 | from mercurial.rustext import revlog |
|
11 | from mercurial.rustext import revlog | |
11 |
|
12 | |||
12 | # this would fail already without appropriate ancestor.__package__ |
|
13 | # this would fail already without appropriate ancestor.__package__ | |
13 | from mercurial.rustext.ancestor import LazyAncestors |
|
14 | from mercurial.rustext.ancestor import LazyAncestors | |
14 |
|
15 | |||
15 | from mercurial.testing import revlog as revlogtesting |
|
16 | from mercurial.testing import revlog as revlogtesting | |
16 |
|
17 | |||
|
18 | header = struct.unpack(">I", revlogtesting.data_non_inlined[:4])[0] | |||
|
19 | ||||
17 |
|
20 | |||
18 | @unittest.skipIf( |
|
21 | @unittest.skipIf( | |
19 | rustext is None, |
|
22 | rustext is None, | |
20 | "rustext module revlog relies on is not available", |
|
23 | "rustext module revlog relies on is not available", | |
21 | ) |
|
24 | ) | |
22 | class RustRevlogIndexTest(revlogtesting.RevlogBasedTestBase): |
|
25 | class RustRevlogIndexTest(revlogtesting.RevlogBasedTestBase): | |
23 | def test_heads(self): |
|
26 | def test_heads(self): | |
24 | idx = self.parseindex() |
|
27 | idx = self.parseindex() | |
25 | rustidx = revlog.MixedIndex(idx, revlogtesting.data_non_inlined) |
|
28 | rustidx = revlog.MixedIndex(idx, revlogtesting.data_non_inlined, header) | |
26 | self.assertEqual(rustidx.headrevs(), idx.headrevs()) |
|
29 | self.assertEqual(rustidx.headrevs(), idx.headrevs()) | |
27 |
|
30 | |||
28 | def test_get_cindex(self): |
|
31 | def test_get_cindex(self): | |
29 | # drop me once we no longer need the method for shortest node |
|
32 | # drop me once we no longer need the method for shortest node | |
30 | idx = self.parseindex() |
|
33 | idx = self.parseindex() | |
31 | rustidx = revlog.MixedIndex(idx, revlogtesting.data_non_inlined) |
|
34 | rustidx = revlog.MixedIndex(idx, revlogtesting.data_non_inlined, header) | |
32 | cidx = rustidx.get_cindex() |
|
35 | cidx = rustidx.get_cindex() | |
33 | self.assertTrue(idx is cidx) |
|
36 | self.assertTrue(idx is cidx) | |
34 |
|
37 | |||
35 | def test_len(self): |
|
38 | def test_len(self): | |
36 | idx = self.parseindex() |
|
39 | idx = self.parseindex() | |
37 | rustidx = revlog.MixedIndex(idx, revlogtesting.data_non_inlined) |
|
40 | rustidx = revlog.MixedIndex(idx, revlogtesting.data_non_inlined, header) | |
38 | self.assertEqual(len(rustidx), len(idx)) |
|
41 | self.assertEqual(len(rustidx), len(idx)) | |
39 |
|
42 | |||
40 | def test_ancestors(self): |
|
43 | def test_ancestors(self): | |
41 | idx = self.parseindex() |
|
44 | idx = self.parseindex() | |
42 | rustidx = revlog.MixedIndex(idx, revlogtesting.data_non_inlined) |
|
45 | rustidx = revlog.MixedIndex(idx, revlogtesting.data_non_inlined, header) | |
43 | lazy = LazyAncestors(rustidx, [3], 0, True) |
|
46 | lazy = LazyAncestors(rustidx, [3], 0, True) | |
44 | # we have two more references to the index: |
|
47 | # we have two more references to the index: | |
45 | # - in its inner iterator for __contains__ and __bool__ |
|
48 | # - in its inner iterator for __contains__ and __bool__ | |
46 | # - in the LazyAncestors instance itself (to spawn new iterators) |
|
49 | # - in the LazyAncestors instance itself (to spawn new iterators) | |
47 | self.assertTrue(2 in lazy) |
|
50 | self.assertTrue(2 in lazy) | |
48 | self.assertTrue(bool(lazy)) |
|
51 | self.assertTrue(bool(lazy)) | |
49 | self.assertEqual(list(lazy), [3, 2, 1, 0]) |
|
52 | self.assertEqual(list(lazy), [3, 2, 1, 0]) | |
50 | # a second time to validate that we spawn new iterators |
|
53 | # a second time to validate that we spawn new iterators | |
51 | self.assertEqual(list(lazy), [3, 2, 1, 0]) |
|
54 | self.assertEqual(list(lazy), [3, 2, 1, 0]) | |
52 |
|
55 | |||
53 | # let's check bool for an empty one |
|
56 | # let's check bool for an empty one | |
54 | self.assertFalse(LazyAncestors(idx, [0], 0, False)) |
|
57 | self.assertFalse(LazyAncestors(idx, [0], 0, False)) | |
55 |
|
58 | |||
56 |
|
59 | |||
57 | if __name__ == '__main__': |
|
60 | if __name__ == '__main__': | |
58 | import silenttestrunner |
|
61 | import silenttestrunner | |
59 |
|
62 | |||
60 | silenttestrunner.main(__name__) |
|
63 | silenttestrunner.main(__name__) |
General Comments 0
You need to be logged in to leave comments.
Login now