Show More
@@ -1,2619 +1,2623 b'' | |||||
1 | # revlog.py - storage back-end for mercurial |
|
1 | # revlog.py - storage back-end for mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | """Storage back-end for Mercurial. |
|
8 | """Storage back-end for Mercurial. | |
9 |
|
9 | |||
10 | This provides efficient delta storage with O(1) retrieve and append |
|
10 | This provides efficient delta storage with O(1) retrieve and append | |
11 | and O(changes) merge between branches. |
|
11 | and O(changes) merge between branches. | |
12 | """ |
|
12 | """ | |
13 |
|
13 | |||
14 | from __future__ import absolute_import |
|
14 | from __future__ import absolute_import | |
15 |
|
15 | |||
16 | import collections |
|
16 | import collections | |
17 | import contextlib |
|
17 | import contextlib | |
18 | import errno |
|
18 | import errno | |
19 | import io |
|
19 | import io | |
20 | import os |
|
20 | import os | |
21 | import struct |
|
21 | import struct | |
22 | import zlib |
|
22 | import zlib | |
23 |
|
23 | |||
24 | # import stuff from node for others to import from revlog |
|
24 | # import stuff from node for others to import from revlog | |
25 | from .node import ( |
|
25 | from .node import ( | |
26 | bin, |
|
26 | bin, | |
27 | hex, |
|
27 | hex, | |
28 | nullhex, |
|
28 | nullhex, | |
29 | nullid, |
|
29 | nullid, | |
30 | nullrev, |
|
30 | nullrev, | |
31 | short, |
|
31 | short, | |
32 | wdirfilenodeids, |
|
32 | wdirfilenodeids, | |
33 | wdirhex, |
|
33 | wdirhex, | |
34 | wdirid, |
|
34 | wdirid, | |
35 | wdirrev, |
|
35 | wdirrev, | |
36 | ) |
|
36 | ) | |
37 | from .i18n import _ |
|
37 | from .i18n import _ | |
38 | from .revlogutils.constants import ( |
|
38 | from .revlogutils.constants import ( | |
39 | FLAG_GENERALDELTA, |
|
39 | FLAG_GENERALDELTA, | |
40 | FLAG_INLINE_DATA, |
|
40 | FLAG_INLINE_DATA, | |
41 | REVLOGV0, |
|
41 | REVLOGV0, | |
42 | REVLOGV1, |
|
42 | REVLOGV1, | |
43 | REVLOGV1_FLAGS, |
|
43 | REVLOGV1_FLAGS, | |
44 | REVLOGV2, |
|
44 | REVLOGV2, | |
45 | REVLOGV2_FLAGS, |
|
45 | REVLOGV2_FLAGS, | |
46 | REVLOG_DEFAULT_FLAGS, |
|
46 | REVLOG_DEFAULT_FLAGS, | |
47 | REVLOG_DEFAULT_FORMAT, |
|
47 | REVLOG_DEFAULT_FORMAT, | |
48 | REVLOG_DEFAULT_VERSION, |
|
48 | REVLOG_DEFAULT_VERSION, | |
49 | ) |
|
49 | ) | |
50 | from .revlogutils.flagutil import ( |
|
50 | from .revlogutils.flagutil import ( | |
51 | REVIDX_DEFAULT_FLAGS, |
|
51 | REVIDX_DEFAULT_FLAGS, | |
52 | REVIDX_ELLIPSIS, |
|
52 | REVIDX_ELLIPSIS, | |
53 | REVIDX_EXTSTORED, |
|
53 | REVIDX_EXTSTORED, | |
54 | REVIDX_FLAGS_ORDER, |
|
54 | REVIDX_FLAGS_ORDER, | |
55 | REVIDX_ISCENSORED, |
|
55 | REVIDX_ISCENSORED, | |
56 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
56 | REVIDX_RAWTEXT_CHANGING_FLAGS, | |
57 | ) |
|
57 | ) | |
58 | from .thirdparty import ( |
|
58 | from .thirdparty import ( | |
59 | attr, |
|
59 | attr, | |
60 | ) |
|
60 | ) | |
61 | from . import ( |
|
61 | from . import ( | |
62 | ancestor, |
|
62 | ancestor, | |
63 | dagop, |
|
63 | dagop, | |
64 | error, |
|
64 | error, | |
65 | mdiff, |
|
65 | mdiff, | |
66 | policy, |
|
66 | policy, | |
67 | pycompat, |
|
67 | pycompat, | |
68 | templatefilters, |
|
68 | templatefilters, | |
69 | util, |
|
69 | util, | |
70 | ) |
|
70 | ) | |
71 | from .interfaces import ( |
|
71 | from .interfaces import ( | |
72 | repository, |
|
72 | repository, | |
73 | util as interfaceutil, |
|
73 | util as interfaceutil, | |
74 | ) |
|
74 | ) | |
75 | from .revlogutils import ( |
|
75 | from .revlogutils import ( | |
76 | deltas as deltautil, |
|
76 | deltas as deltautil, | |
77 | flagutil, |
|
77 | flagutil, | |
78 | ) |
|
78 | ) | |
79 | from .utils import ( |
|
79 | from .utils import ( | |
80 | storageutil, |
|
80 | storageutil, | |
81 | stringutil, |
|
81 | stringutil, | |
82 | ) |
|
82 | ) | |
83 |
|
83 | |||
84 | # blanked usage of all the name to prevent pyflakes constraints |
|
84 | # blanked usage of all the name to prevent pyflakes constraints | |
85 | # We need these name available in the module for extensions. |
|
85 | # We need these name available in the module for extensions. | |
86 | REVLOGV0 |
|
86 | REVLOGV0 | |
87 | REVLOGV1 |
|
87 | REVLOGV1 | |
88 | REVLOGV2 |
|
88 | REVLOGV2 | |
89 | FLAG_INLINE_DATA |
|
89 | FLAG_INLINE_DATA | |
90 | FLAG_GENERALDELTA |
|
90 | FLAG_GENERALDELTA | |
91 | REVLOG_DEFAULT_FLAGS |
|
91 | REVLOG_DEFAULT_FLAGS | |
92 | REVLOG_DEFAULT_FORMAT |
|
92 | REVLOG_DEFAULT_FORMAT | |
93 | REVLOG_DEFAULT_VERSION |
|
93 | REVLOG_DEFAULT_VERSION | |
94 | REVLOGV1_FLAGS |
|
94 | REVLOGV1_FLAGS | |
95 | REVLOGV2_FLAGS |
|
95 | REVLOGV2_FLAGS | |
96 | REVIDX_ISCENSORED |
|
96 | REVIDX_ISCENSORED | |
97 | REVIDX_ELLIPSIS |
|
97 | REVIDX_ELLIPSIS | |
98 | REVIDX_EXTSTORED |
|
98 | REVIDX_EXTSTORED | |
99 | REVIDX_DEFAULT_FLAGS |
|
99 | REVIDX_DEFAULT_FLAGS | |
100 | REVIDX_FLAGS_ORDER |
|
100 | REVIDX_FLAGS_ORDER | |
101 | REVIDX_RAWTEXT_CHANGING_FLAGS |
|
101 | REVIDX_RAWTEXT_CHANGING_FLAGS | |
102 |
|
102 | |||
103 | parsers = policy.importmod(r'parsers') |
|
103 | parsers = policy.importmod(r'parsers') | |
104 | rustancestor = policy.importrust(r'ancestor') |
|
104 | rustancestor = policy.importrust(r'ancestor') | |
105 | rustdagop = policy.importrust(r'dagop') |
|
105 | rustdagop = policy.importrust(r'dagop') | |
106 |
|
106 | |||
107 | # Aliased for performance. |
|
107 | # Aliased for performance. | |
108 | _zlibdecompress = zlib.decompress |
|
108 | _zlibdecompress = zlib.decompress | |
109 |
|
109 | |||
110 | # max size of revlog with inline data |
|
110 | # max size of revlog with inline data | |
111 | _maxinline = 131072 |
|
111 | _maxinline = 131072 | |
112 | _chunksize = 1048576 |
|
112 | _chunksize = 1048576 | |
113 |
|
113 | |||
114 | # Flag processors for REVIDX_ELLIPSIS. |
|
114 | # Flag processors for REVIDX_ELLIPSIS. | |
115 | def ellipsisreadprocessor(rl, text): |
|
115 | def ellipsisreadprocessor(rl, text): | |
116 | return text, False |
|
116 | return text, False | |
117 |
|
117 | |||
118 | def ellipsiswriteprocessor(rl, text): |
|
118 | def ellipsiswriteprocessor(rl, text): | |
119 | return text, False |
|
119 | return text, False | |
120 |
|
120 | |||
121 | def ellipsisrawprocessor(rl, text): |
|
121 | def ellipsisrawprocessor(rl, text): | |
122 | return False |
|
122 | return False | |
123 |
|
123 | |||
124 | ellipsisprocessor = ( |
|
124 | ellipsisprocessor = ( | |
125 | ellipsisreadprocessor, |
|
125 | ellipsisreadprocessor, | |
126 | ellipsiswriteprocessor, |
|
126 | ellipsiswriteprocessor, | |
127 | ellipsisrawprocessor, |
|
127 | ellipsisrawprocessor, | |
128 | ) |
|
128 | ) | |
129 |
|
129 | |||
130 | def getoffset(q): |
|
130 | def getoffset(q): | |
131 | return int(q >> 16) |
|
131 | return int(q >> 16) | |
132 |
|
132 | |||
133 | def gettype(q): |
|
133 | def gettype(q): | |
134 | return int(q & 0xFFFF) |
|
134 | return int(q & 0xFFFF) | |
135 |
|
135 | |||
136 | def offset_type(offset, type): |
|
136 | def offset_type(offset, type): | |
137 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: |
|
137 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: | |
138 | raise ValueError('unknown revlog index flags') |
|
138 | raise ValueError('unknown revlog index flags') | |
139 | return int(int(offset) << 16 | type) |
|
139 | return int(int(offset) << 16 | type) | |
140 |
|
140 | |||
141 | @attr.s(slots=True, frozen=True) |
|
141 | @attr.s(slots=True, frozen=True) | |
142 | class _revisioninfo(object): |
|
142 | class _revisioninfo(object): | |
143 | """Information about a revision that allows building its fulltext |
|
143 | """Information about a revision that allows building its fulltext | |
144 | node: expected hash of the revision |
|
144 | node: expected hash of the revision | |
145 | p1, p2: parent revs of the revision |
|
145 | p1, p2: parent revs of the revision | |
146 | btext: built text cache consisting of a one-element list |
|
146 | btext: built text cache consisting of a one-element list | |
147 | cachedelta: (baserev, uncompressed_delta) or None |
|
147 | cachedelta: (baserev, uncompressed_delta) or None | |
148 | flags: flags associated to the revision storage |
|
148 | flags: flags associated to the revision storage | |
149 |
|
149 | |||
150 | One of btext[0] or cachedelta must be set. |
|
150 | One of btext[0] or cachedelta must be set. | |
151 | """ |
|
151 | """ | |
152 | node = attr.ib() |
|
152 | node = attr.ib() | |
153 | p1 = attr.ib() |
|
153 | p1 = attr.ib() | |
154 | p2 = attr.ib() |
|
154 | p2 = attr.ib() | |
155 | btext = attr.ib() |
|
155 | btext = attr.ib() | |
156 | textlen = attr.ib() |
|
156 | textlen = attr.ib() | |
157 | cachedelta = attr.ib() |
|
157 | cachedelta = attr.ib() | |
158 | flags = attr.ib() |
|
158 | flags = attr.ib() | |
159 |
|
159 | |||
160 | @interfaceutil.implementer(repository.irevisiondelta) |
|
160 | @interfaceutil.implementer(repository.irevisiondelta) | |
161 | @attr.s(slots=True) |
|
161 | @attr.s(slots=True) | |
162 | class revlogrevisiondelta(object): |
|
162 | class revlogrevisiondelta(object): | |
163 | node = attr.ib() |
|
163 | node = attr.ib() | |
164 | p1node = attr.ib() |
|
164 | p1node = attr.ib() | |
165 | p2node = attr.ib() |
|
165 | p2node = attr.ib() | |
166 | basenode = attr.ib() |
|
166 | basenode = attr.ib() | |
167 | flags = attr.ib() |
|
167 | flags = attr.ib() | |
168 | baserevisionsize = attr.ib() |
|
168 | baserevisionsize = attr.ib() | |
169 | revision = attr.ib() |
|
169 | revision = attr.ib() | |
170 | delta = attr.ib() |
|
170 | delta = attr.ib() | |
171 | linknode = attr.ib(default=None) |
|
171 | linknode = attr.ib(default=None) | |
172 |
|
172 | |||
173 | @interfaceutil.implementer(repository.iverifyproblem) |
|
173 | @interfaceutil.implementer(repository.iverifyproblem) | |
174 | @attr.s(frozen=True) |
|
174 | @attr.s(frozen=True) | |
175 | class revlogproblem(object): |
|
175 | class revlogproblem(object): | |
176 | warning = attr.ib(default=None) |
|
176 | warning = attr.ib(default=None) | |
177 | error = attr.ib(default=None) |
|
177 | error = attr.ib(default=None) | |
178 | node = attr.ib(default=None) |
|
178 | node = attr.ib(default=None) | |
179 |
|
179 | |||
180 | # index v0: |
|
180 | # index v0: | |
181 | # 4 bytes: offset |
|
181 | # 4 bytes: offset | |
182 | # 4 bytes: compressed length |
|
182 | # 4 bytes: compressed length | |
183 | # 4 bytes: base rev |
|
183 | # 4 bytes: base rev | |
184 | # 4 bytes: link rev |
|
184 | # 4 bytes: link rev | |
185 | # 20 bytes: parent 1 nodeid |
|
185 | # 20 bytes: parent 1 nodeid | |
186 | # 20 bytes: parent 2 nodeid |
|
186 | # 20 bytes: parent 2 nodeid | |
187 | # 20 bytes: nodeid |
|
187 | # 20 bytes: nodeid | |
188 | indexformatv0 = struct.Struct(">4l20s20s20s") |
|
188 | indexformatv0 = struct.Struct(">4l20s20s20s") | |
189 | indexformatv0_pack = indexformatv0.pack |
|
189 | indexformatv0_pack = indexformatv0.pack | |
190 | indexformatv0_unpack = indexformatv0.unpack |
|
190 | indexformatv0_unpack = indexformatv0.unpack | |
191 |
|
191 | |||
192 | class revlogoldindex(list): |
|
192 | class revlogoldindex(list): | |
193 | def __getitem__(self, i): |
|
193 | def __getitem__(self, i): | |
194 | if i == -1: |
|
194 | if i == -1: | |
195 | return (0, 0, 0, -1, -1, -1, -1, nullid) |
|
195 | return (0, 0, 0, -1, -1, -1, -1, nullid) | |
196 | return list.__getitem__(self, i) |
|
196 | return list.__getitem__(self, i) | |
197 |
|
197 | |||
198 | class revlogoldio(object): |
|
198 | class revlogoldio(object): | |
199 | def __init__(self): |
|
199 | def __init__(self): | |
200 | self.size = indexformatv0.size |
|
200 | self.size = indexformatv0.size | |
201 |
|
201 | |||
202 | def parseindex(self, data, inline): |
|
202 | def parseindex(self, data, inline): | |
203 | s = self.size |
|
203 | s = self.size | |
204 | index = [] |
|
204 | index = [] | |
205 | nodemap = {nullid: nullrev} |
|
205 | nodemap = {nullid: nullrev} | |
206 | n = off = 0 |
|
206 | n = off = 0 | |
207 | l = len(data) |
|
207 | l = len(data) | |
208 | while off + s <= l: |
|
208 | while off + s <= l: | |
209 | cur = data[off:off + s] |
|
209 | cur = data[off:off + s] | |
210 | off += s |
|
210 | off += s | |
211 | e = indexformatv0_unpack(cur) |
|
211 | e = indexformatv0_unpack(cur) | |
212 | # transform to revlogv1 format |
|
212 | # transform to revlogv1 format | |
213 | e2 = (offset_type(e[0], 0), e[1], -1, e[2], e[3], |
|
213 | e2 = (offset_type(e[0], 0), e[1], -1, e[2], e[3], | |
214 | nodemap.get(e[4], nullrev), nodemap.get(e[5], nullrev), e[6]) |
|
214 | nodemap.get(e[4], nullrev), nodemap.get(e[5], nullrev), e[6]) | |
215 | index.append(e2) |
|
215 | index.append(e2) | |
216 | nodemap[e[6]] = n |
|
216 | nodemap[e[6]] = n | |
217 | n += 1 |
|
217 | n += 1 | |
218 |
|
218 | |||
219 | return revlogoldindex(index), nodemap, None |
|
219 | return revlogoldindex(index), nodemap, None | |
220 |
|
220 | |||
221 | def packentry(self, entry, node, version, rev): |
|
221 | def packentry(self, entry, node, version, rev): | |
222 | if gettype(entry[0]): |
|
222 | if gettype(entry[0]): | |
223 | raise error.RevlogError(_('index entry flags need revlog ' |
|
223 | raise error.RevlogError(_('index entry flags need revlog ' | |
224 | 'version 1')) |
|
224 | 'version 1')) | |
225 | e2 = (getoffset(entry[0]), entry[1], entry[3], entry[4], |
|
225 | e2 = (getoffset(entry[0]), entry[1], entry[3], entry[4], | |
226 | node(entry[5]), node(entry[6]), entry[7]) |
|
226 | node(entry[5]), node(entry[6]), entry[7]) | |
227 | return indexformatv0_pack(*e2) |
|
227 | return indexformatv0_pack(*e2) | |
228 |
|
228 | |||
229 | # index ng: |
|
229 | # index ng: | |
230 | # 6 bytes: offset |
|
230 | # 6 bytes: offset | |
231 | # 2 bytes: flags |
|
231 | # 2 bytes: flags | |
232 | # 4 bytes: compressed length |
|
232 | # 4 bytes: compressed length | |
233 | # 4 bytes: uncompressed length |
|
233 | # 4 bytes: uncompressed length | |
234 | # 4 bytes: base rev |
|
234 | # 4 bytes: base rev | |
235 | # 4 bytes: link rev |
|
235 | # 4 bytes: link rev | |
236 | # 4 bytes: parent 1 rev |
|
236 | # 4 bytes: parent 1 rev | |
237 | # 4 bytes: parent 2 rev |
|
237 | # 4 bytes: parent 2 rev | |
238 | # 32 bytes: nodeid |
|
238 | # 32 bytes: nodeid | |
239 | indexformatng = struct.Struct(">Qiiiiii20s12x") |
|
239 | indexformatng = struct.Struct(">Qiiiiii20s12x") | |
240 | indexformatng_pack = indexformatng.pack |
|
240 | indexformatng_pack = indexformatng.pack | |
241 | versionformat = struct.Struct(">I") |
|
241 | versionformat = struct.Struct(">I") | |
242 | versionformat_pack = versionformat.pack |
|
242 | versionformat_pack = versionformat.pack | |
243 | versionformat_unpack = versionformat.unpack |
|
243 | versionformat_unpack = versionformat.unpack | |
244 |
|
244 | |||
245 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte |
|
245 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte | |
246 | # signed integer) |
|
246 | # signed integer) | |
247 | _maxentrysize = 0x7fffffff |
|
247 | _maxentrysize = 0x7fffffff | |
248 |
|
248 | |||
249 | class revlogio(object): |
|
249 | class revlogio(object): | |
250 | def __init__(self): |
|
250 | def __init__(self): | |
251 | self.size = indexformatng.size |
|
251 | self.size = indexformatng.size | |
252 |
|
252 | |||
253 | def parseindex(self, data, inline): |
|
253 | def parseindex(self, data, inline): | |
254 | # call the C implementation to parse the index data |
|
254 | # call the C implementation to parse the index data | |
255 | index, cache = parsers.parse_index2(data, inline) |
|
255 | index, cache = parsers.parse_index2(data, inline) | |
256 | return index, getattr(index, 'nodemap', None), cache |
|
256 | return index, getattr(index, 'nodemap', None), cache | |
257 |
|
257 | |||
258 | def packentry(self, entry, node, version, rev): |
|
258 | def packentry(self, entry, node, version, rev): | |
259 | p = indexformatng_pack(*entry) |
|
259 | p = indexformatng_pack(*entry) | |
260 | if rev == 0: |
|
260 | if rev == 0: | |
261 | p = versionformat_pack(version) + p[4:] |
|
261 | p = versionformat_pack(version) + p[4:] | |
262 | return p |
|
262 | return p | |
263 |
|
263 | |||
264 | class revlog(flagutil.flagprocessorsmixin): |
|
264 | class revlog(flagutil.flagprocessorsmixin): | |
265 | """ |
|
265 | """ | |
266 | the underlying revision storage object |
|
266 | the underlying revision storage object | |
267 |
|
267 | |||
268 | A revlog consists of two parts, an index and the revision data. |
|
268 | A revlog consists of two parts, an index and the revision data. | |
269 |
|
269 | |||
270 | The index is a file with a fixed record size containing |
|
270 | The index is a file with a fixed record size containing | |
271 | information on each revision, including its nodeid (hash), the |
|
271 | information on each revision, including its nodeid (hash), the | |
272 | nodeids of its parents, the position and offset of its data within |
|
272 | nodeids of its parents, the position and offset of its data within | |
273 | the data file, and the revision it's based on. Finally, each entry |
|
273 | the data file, and the revision it's based on. Finally, each entry | |
274 | contains a linkrev entry that can serve as a pointer to external |
|
274 | contains a linkrev entry that can serve as a pointer to external | |
275 | data. |
|
275 | data. | |
276 |
|
276 | |||
277 | The revision data itself is a linear collection of data chunks. |
|
277 | The revision data itself is a linear collection of data chunks. | |
278 | Each chunk represents a revision and is usually represented as a |
|
278 | Each chunk represents a revision and is usually represented as a | |
279 | delta against the previous chunk. To bound lookup time, runs of |
|
279 | delta against the previous chunk. To bound lookup time, runs of | |
280 | deltas are limited to about 2 times the length of the original |
|
280 | deltas are limited to about 2 times the length of the original | |
281 | version data. This makes retrieval of a version proportional to |
|
281 | version data. This makes retrieval of a version proportional to | |
282 | its size, or O(1) relative to the number of revisions. |
|
282 | its size, or O(1) relative to the number of revisions. | |
283 |
|
283 | |||
284 | Both pieces of the revlog are written to in an append-only |
|
284 | Both pieces of the revlog are written to in an append-only | |
285 | fashion, which means we never need to rewrite a file to insert or |
|
285 | fashion, which means we never need to rewrite a file to insert or | |
286 | remove data, and can use some simple techniques to avoid the need |
|
286 | remove data, and can use some simple techniques to avoid the need | |
287 | for locking while reading. |
|
287 | for locking while reading. | |
288 |
|
288 | |||
289 | If checkambig, indexfile is opened with checkambig=True at |
|
289 | If checkambig, indexfile is opened with checkambig=True at | |
290 | writing, to avoid file stat ambiguity. |
|
290 | writing, to avoid file stat ambiguity. | |
291 |
|
291 | |||
292 | If mmaplargeindex is True, and an mmapindexthreshold is set, the |
|
292 | If mmaplargeindex is True, and an mmapindexthreshold is set, the | |
293 | index will be mmapped rather than read if it is larger than the |
|
293 | index will be mmapped rather than read if it is larger than the | |
294 | configured threshold. |
|
294 | configured threshold. | |
295 |
|
295 | |||
296 | If censorable is True, the revlog can have censored revisions. |
|
296 | If censorable is True, the revlog can have censored revisions. | |
297 |
|
297 | |||
298 | If `upperboundcomp` is not None, this is the expected maximal gain from |
|
298 | If `upperboundcomp` is not None, this is the expected maximal gain from | |
299 | compression for the data content. |
|
299 | compression for the data content. | |
300 | """ |
|
300 | """ | |
301 | def __init__(self, opener, indexfile, datafile=None, checkambig=False, |
|
301 | def __init__(self, opener, indexfile, datafile=None, checkambig=False, | |
302 | mmaplargeindex=False, censorable=False, |
|
302 | mmaplargeindex=False, censorable=False, | |
303 | upperboundcomp=None): |
|
303 | upperboundcomp=None): | |
304 | """ |
|
304 | """ | |
305 | create a revlog object |
|
305 | create a revlog object | |
306 |
|
306 | |||
307 | opener is a function that abstracts the file opening operation |
|
307 | opener is a function that abstracts the file opening operation | |
308 | and can be used to implement COW semantics or the like. |
|
308 | and can be used to implement COW semantics or the like. | |
309 |
|
309 | |||
310 | """ |
|
310 | """ | |
311 | self.upperboundcomp = upperboundcomp |
|
311 | self.upperboundcomp = upperboundcomp | |
312 | self.indexfile = indexfile |
|
312 | self.indexfile = indexfile | |
313 | self.datafile = datafile or (indexfile[:-2] + ".d") |
|
313 | self.datafile = datafile or (indexfile[:-2] + ".d") | |
314 | self.opener = opener |
|
314 | self.opener = opener | |
315 | # When True, indexfile is opened with checkambig=True at writing, to |
|
315 | # When True, indexfile is opened with checkambig=True at writing, to | |
316 | # avoid file stat ambiguity. |
|
316 | # avoid file stat ambiguity. | |
317 | self._checkambig = checkambig |
|
317 | self._checkambig = checkambig | |
318 | self._mmaplargeindex = mmaplargeindex |
|
318 | self._mmaplargeindex = mmaplargeindex | |
319 | self._censorable = censorable |
|
319 | self._censorable = censorable | |
320 | # 3-tuple of (node, rev, text) for a raw revision. |
|
320 | # 3-tuple of (node, rev, text) for a raw revision. | |
321 | self._revisioncache = None |
|
321 | self._revisioncache = None | |
322 | # Maps rev to chain base rev. |
|
322 | # Maps rev to chain base rev. | |
323 | self._chainbasecache = util.lrucachedict(100) |
|
323 | self._chainbasecache = util.lrucachedict(100) | |
324 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. |
|
324 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. | |
325 | self._chunkcache = (0, '') |
|
325 | self._chunkcache = (0, '') | |
326 | # How much data to read and cache into the raw revlog data cache. |
|
326 | # How much data to read and cache into the raw revlog data cache. | |
327 | self._chunkcachesize = 65536 |
|
327 | self._chunkcachesize = 65536 | |
328 | self._maxchainlen = None |
|
328 | self._maxchainlen = None | |
329 | self._deltabothparents = True |
|
329 | self._deltabothparents = True | |
330 | self.index = [] |
|
330 | self.index = [] | |
331 | # Mapping of partial identifiers to full nodes. |
|
331 | # Mapping of partial identifiers to full nodes. | |
332 | self._pcache = {} |
|
332 | self._pcache = {} | |
333 | # Mapping of revision integer to full node. |
|
333 | # Mapping of revision integer to full node. | |
334 | self._nodecache = {nullid: nullrev} |
|
334 | self._nodecache = {nullid: nullrev} | |
335 | self._nodepos = None |
|
335 | self._nodepos = None | |
336 | self._compengine = 'zlib' |
|
336 | self._compengine = 'zlib' | |
337 | self._compengineopts = {} |
|
337 | self._compengineopts = {} | |
338 | self._maxdeltachainspan = -1 |
|
338 | self._maxdeltachainspan = -1 | |
339 | self._withsparseread = False |
|
339 | self._withsparseread = False | |
340 | self._sparserevlog = False |
|
340 | self._sparserevlog = False | |
341 | self._srdensitythreshold = 0.50 |
|
341 | self._srdensitythreshold = 0.50 | |
342 | self._srmingapsize = 262144 |
|
342 | self._srmingapsize = 262144 | |
343 |
|
343 | |||
344 | # Make copy of flag processors so each revlog instance can support |
|
344 | # Make copy of flag processors so each revlog instance can support | |
345 | # custom flags. |
|
345 | # custom flags. | |
346 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
346 | self._flagprocessors = dict(flagutil.flagprocessors) | |
347 |
|
347 | |||
348 | # 2-tuple of file handles being used for active writing. |
|
348 | # 2-tuple of file handles being used for active writing. | |
349 | self._writinghandles = None |
|
349 | self._writinghandles = None | |
350 |
|
350 | |||
351 | self._loadindex() |
|
351 | self._loadindex() | |
352 |
|
352 | |||
353 | def _loadindex(self): |
|
353 | def _loadindex(self): | |
354 | mmapindexthreshold = None |
|
354 | mmapindexthreshold = None | |
355 | opts = getattr(self.opener, 'options', {}) or {} |
|
355 | opts = getattr(self.opener, 'options', {}) or {} | |
356 |
|
356 | |||
357 | if 'revlogv2' in opts: |
|
357 | if 'revlogv2' in opts: | |
358 | newversionflags = REVLOGV2 | FLAG_INLINE_DATA |
|
358 | newversionflags = REVLOGV2 | FLAG_INLINE_DATA | |
359 | elif 'revlogv1' in opts: |
|
359 | elif 'revlogv1' in opts: | |
360 | newversionflags = REVLOGV1 | FLAG_INLINE_DATA |
|
360 | newversionflags = REVLOGV1 | FLAG_INLINE_DATA | |
361 | if 'generaldelta' in opts: |
|
361 | if 'generaldelta' in opts: | |
362 | newversionflags |= FLAG_GENERALDELTA |
|
362 | newversionflags |= FLAG_GENERALDELTA | |
363 | elif getattr(self.opener, 'options', None) is not None: |
|
363 | elif getattr(self.opener, 'options', None) is not None: | |
364 | # If options provided but no 'revlog*' found, the repository |
|
364 | # If options provided but no 'revlog*' found, the repository | |
365 | # would have no 'requires' file in it, which means we have to |
|
365 | # would have no 'requires' file in it, which means we have to | |
366 | # stick to the old format. |
|
366 | # stick to the old format. | |
367 | newversionflags = REVLOGV0 |
|
367 | newversionflags = REVLOGV0 | |
368 | else: |
|
368 | else: | |
369 | newversionflags = REVLOG_DEFAULT_VERSION |
|
369 | newversionflags = REVLOG_DEFAULT_VERSION | |
370 |
|
370 | |||
371 | if 'chunkcachesize' in opts: |
|
371 | if 'chunkcachesize' in opts: | |
372 | self._chunkcachesize = opts['chunkcachesize'] |
|
372 | self._chunkcachesize = opts['chunkcachesize'] | |
373 | if 'maxchainlen' in opts: |
|
373 | if 'maxchainlen' in opts: | |
374 | self._maxchainlen = opts['maxchainlen'] |
|
374 | self._maxchainlen = opts['maxchainlen'] | |
375 | if 'deltabothparents' in opts: |
|
375 | if 'deltabothparents' in opts: | |
376 | self._deltabothparents = opts['deltabothparents'] |
|
376 | self._deltabothparents = opts['deltabothparents'] | |
377 | self._lazydelta = bool(opts.get('lazydelta', True)) |
|
377 | self._lazydelta = bool(opts.get('lazydelta', True)) | |
378 | self._lazydeltabase = False |
|
378 | self._lazydeltabase = False | |
379 | if self._lazydelta: |
|
379 | if self._lazydelta: | |
380 | self._lazydeltabase = bool(opts.get('lazydeltabase', False)) |
|
380 | self._lazydeltabase = bool(opts.get('lazydeltabase', False)) | |
381 | if 'compengine' in opts: |
|
381 | if 'compengine' in opts: | |
382 | self._compengine = opts['compengine'] |
|
382 | self._compengine = opts['compengine'] | |
383 | if 'zlib.level' in opts: |
|
383 | if 'zlib.level' in opts: | |
384 | self._compengineopts['zlib.level'] = opts['zlib.level'] |
|
384 | self._compengineopts['zlib.level'] = opts['zlib.level'] | |
385 | if 'zstd.level' in opts: |
|
385 | if 'zstd.level' in opts: | |
386 | self._compengineopts['zstd.level'] = opts['zstd.level'] |
|
386 | self._compengineopts['zstd.level'] = opts['zstd.level'] | |
387 | if 'maxdeltachainspan' in opts: |
|
387 | if 'maxdeltachainspan' in opts: | |
388 | self._maxdeltachainspan = opts['maxdeltachainspan'] |
|
388 | self._maxdeltachainspan = opts['maxdeltachainspan'] | |
389 | if self._mmaplargeindex and 'mmapindexthreshold' in opts: |
|
389 | if self._mmaplargeindex and 'mmapindexthreshold' in opts: | |
390 | mmapindexthreshold = opts['mmapindexthreshold'] |
|
390 | mmapindexthreshold = opts['mmapindexthreshold'] | |
391 | self._sparserevlog = bool(opts.get('sparse-revlog', False)) |
|
391 | self._sparserevlog = bool(opts.get('sparse-revlog', False)) | |
392 | withsparseread = bool(opts.get('with-sparse-read', False)) |
|
392 | withsparseread = bool(opts.get('with-sparse-read', False)) | |
393 | # sparse-revlog forces sparse-read |
|
393 | # sparse-revlog forces sparse-read | |
394 | self._withsparseread = self._sparserevlog or withsparseread |
|
394 | self._withsparseread = self._sparserevlog or withsparseread | |
395 | if 'sparse-read-density-threshold' in opts: |
|
395 | if 'sparse-read-density-threshold' in opts: | |
396 | self._srdensitythreshold = opts['sparse-read-density-threshold'] |
|
396 | self._srdensitythreshold = opts['sparse-read-density-threshold'] | |
397 | if 'sparse-read-min-gap-size' in opts: |
|
397 | if 'sparse-read-min-gap-size' in opts: | |
398 | self._srmingapsize = opts['sparse-read-min-gap-size'] |
|
398 | self._srmingapsize = opts['sparse-read-min-gap-size'] | |
399 | if opts.get('enableellipsis'): |
|
399 | if opts.get('enableellipsis'): | |
400 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor |
|
400 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor | |
401 |
|
401 | |||
402 | # revlog v0 doesn't have flag processors |
|
402 | # revlog v0 doesn't have flag processors | |
403 | for flag, processor in opts.get(b'flagprocessors', {}).iteritems(): |
|
403 | for flag, processor in opts.get(b'flagprocessors', {}).iteritems(): | |
404 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) |
|
404 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) | |
405 |
|
405 | |||
406 | if self._chunkcachesize <= 0: |
|
406 | if self._chunkcachesize <= 0: | |
407 | raise error.RevlogError(_('revlog chunk cache size %r is not ' |
|
407 | raise error.RevlogError(_('revlog chunk cache size %r is not ' | |
408 | 'greater than 0') % self._chunkcachesize) |
|
408 | 'greater than 0') % self._chunkcachesize) | |
409 | elif self._chunkcachesize & (self._chunkcachesize - 1): |
|
409 | elif self._chunkcachesize & (self._chunkcachesize - 1): | |
410 | raise error.RevlogError(_('revlog chunk cache size %r is not a ' |
|
410 | raise error.RevlogError(_('revlog chunk cache size %r is not a ' | |
411 | 'power of 2') % self._chunkcachesize) |
|
411 | 'power of 2') % self._chunkcachesize) | |
412 |
|
412 | |||
413 | indexdata = '' |
|
413 | indexdata = '' | |
414 | self._initempty = True |
|
414 | self._initempty = True | |
415 | try: |
|
415 | try: | |
416 | with self._indexfp() as f: |
|
416 | with self._indexfp() as f: | |
417 | if (mmapindexthreshold is not None and |
|
417 | if (mmapindexthreshold is not None and | |
418 | self.opener.fstat(f).st_size >= mmapindexthreshold): |
|
418 | self.opener.fstat(f).st_size >= mmapindexthreshold): | |
419 | # TODO: should .close() to release resources without |
|
419 | # TODO: should .close() to release resources without | |
420 | # relying on Python GC |
|
420 | # relying on Python GC | |
421 | indexdata = util.buffer(util.mmapread(f)) |
|
421 | indexdata = util.buffer(util.mmapread(f)) | |
422 | else: |
|
422 | else: | |
423 | indexdata = f.read() |
|
423 | indexdata = f.read() | |
424 | if len(indexdata) > 0: |
|
424 | if len(indexdata) > 0: | |
425 | versionflags = versionformat_unpack(indexdata[:4])[0] |
|
425 | versionflags = versionformat_unpack(indexdata[:4])[0] | |
426 | self._initempty = False |
|
426 | self._initempty = False | |
427 | else: |
|
427 | else: | |
428 | versionflags = newversionflags |
|
428 | versionflags = newversionflags | |
429 | except IOError as inst: |
|
429 | except IOError as inst: | |
430 | if inst.errno != errno.ENOENT: |
|
430 | if inst.errno != errno.ENOENT: | |
431 | raise |
|
431 | raise | |
432 |
|
432 | |||
433 | versionflags = newversionflags |
|
433 | versionflags = newversionflags | |
434 |
|
434 | |||
435 | self.version = versionflags |
|
435 | self.version = versionflags | |
436 |
|
436 | |||
437 | flags = versionflags & ~0xFFFF |
|
437 | flags = versionflags & ~0xFFFF | |
438 | fmt = versionflags & 0xFFFF |
|
438 | fmt = versionflags & 0xFFFF | |
439 |
|
439 | |||
440 | if fmt == REVLOGV0: |
|
440 | if fmt == REVLOGV0: | |
441 | if flags: |
|
441 | if flags: | |
442 | raise error.RevlogError(_('unknown flags (%#04x) in version %d ' |
|
442 | raise error.RevlogError(_('unknown flags (%#04x) in version %d ' | |
443 | 'revlog %s') % |
|
443 | 'revlog %s') % | |
444 | (flags >> 16, fmt, self.indexfile)) |
|
444 | (flags >> 16, fmt, self.indexfile)) | |
445 |
|
445 | |||
446 | self._inline = False |
|
446 | self._inline = False | |
447 | self._generaldelta = False |
|
447 | self._generaldelta = False | |
448 |
|
448 | |||
449 | elif fmt == REVLOGV1: |
|
449 | elif fmt == REVLOGV1: | |
450 | if flags & ~REVLOGV1_FLAGS: |
|
450 | if flags & ~REVLOGV1_FLAGS: | |
451 | raise error.RevlogError(_('unknown flags (%#04x) in version %d ' |
|
451 | raise error.RevlogError(_('unknown flags (%#04x) in version %d ' | |
452 | 'revlog %s') % |
|
452 | 'revlog %s') % | |
453 | (flags >> 16, fmt, self.indexfile)) |
|
453 | (flags >> 16, fmt, self.indexfile)) | |
454 |
|
454 | |||
455 | self._inline = versionflags & FLAG_INLINE_DATA |
|
455 | self._inline = versionflags & FLAG_INLINE_DATA | |
456 | self._generaldelta = versionflags & FLAG_GENERALDELTA |
|
456 | self._generaldelta = versionflags & FLAG_GENERALDELTA | |
457 |
|
457 | |||
458 | elif fmt == REVLOGV2: |
|
458 | elif fmt == REVLOGV2: | |
459 | if flags & ~REVLOGV2_FLAGS: |
|
459 | if flags & ~REVLOGV2_FLAGS: | |
460 | raise error.RevlogError(_('unknown flags (%#04x) in version %d ' |
|
460 | raise error.RevlogError(_('unknown flags (%#04x) in version %d ' | |
461 | 'revlog %s') % |
|
461 | 'revlog %s') % | |
462 | (flags >> 16, fmt, self.indexfile)) |
|
462 | (flags >> 16, fmt, self.indexfile)) | |
463 |
|
463 | |||
464 | self._inline = versionflags & FLAG_INLINE_DATA |
|
464 | self._inline = versionflags & FLAG_INLINE_DATA | |
465 | # generaldelta implied by version 2 revlogs. |
|
465 | # generaldelta implied by version 2 revlogs. | |
466 | self._generaldelta = True |
|
466 | self._generaldelta = True | |
467 |
|
467 | |||
468 | else: |
|
468 | else: | |
469 | raise error.RevlogError(_('unknown version (%d) in revlog %s') % |
|
469 | raise error.RevlogError(_('unknown version (%d) in revlog %s') % | |
470 | (fmt, self.indexfile)) |
|
470 | (fmt, self.indexfile)) | |
471 | # sparse-revlog can't be on without general-delta (issue6056) |
|
471 | # sparse-revlog can't be on without general-delta (issue6056) | |
472 | if not self._generaldelta: |
|
472 | if not self._generaldelta: | |
473 | self._sparserevlog = False |
|
473 | self._sparserevlog = False | |
474 |
|
474 | |||
475 | self._storedeltachains = True |
|
475 | self._storedeltachains = True | |
476 |
|
476 | |||
477 | self._io = revlogio() |
|
477 | self._io = revlogio() | |
478 | if self.version == REVLOGV0: |
|
478 | if self.version == REVLOGV0: | |
479 | self._io = revlogoldio() |
|
479 | self._io = revlogoldio() | |
480 | try: |
|
480 | try: | |
481 | d = self._io.parseindex(indexdata, self._inline) |
|
481 | d = self._io.parseindex(indexdata, self._inline) | |
482 | except (ValueError, IndexError): |
|
482 | except (ValueError, IndexError): | |
483 | raise error.RevlogError(_("index %s is corrupted") % |
|
483 | raise error.RevlogError(_("index %s is corrupted") % | |
484 | self.indexfile) |
|
484 | self.indexfile) | |
485 | self.index, nodemap, self._chunkcache = d |
|
485 | self.index, nodemap, self._chunkcache = d | |
486 | if nodemap is not None: |
|
486 | if nodemap is not None: | |
487 | self.nodemap = self._nodecache = nodemap |
|
487 | self.nodemap = self._nodecache = nodemap | |
488 | if not self._chunkcache: |
|
488 | if not self._chunkcache: | |
489 | self._chunkclear() |
|
489 | self._chunkclear() | |
490 | # revnum -> (chain-length, sum-delta-length) |
|
490 | # revnum -> (chain-length, sum-delta-length) | |
491 | self._chaininfocache = {} |
|
491 | self._chaininfocache = {} | |
492 | # revlog header -> revlog compressor |
|
492 | # revlog header -> revlog compressor | |
493 | self._decompressors = {} |
|
493 | self._decompressors = {} | |
494 |
|
494 | |||
495 | @util.propertycache |
|
495 | @util.propertycache | |
496 | def _compressor(self): |
|
496 | def _compressor(self): | |
497 | engine = util.compengines[self._compengine] |
|
497 | engine = util.compengines[self._compengine] | |
498 | return engine.revlogcompressor(self._compengineopts) |
|
498 | return engine.revlogcompressor(self._compengineopts) | |
499 |
|
499 | |||
500 | def _indexfp(self, mode='r'): |
|
500 | def _indexfp(self, mode='r'): | |
501 | """file object for the revlog's index file""" |
|
501 | """file object for the revlog's index file""" | |
502 | args = {r'mode': mode} |
|
502 | args = {r'mode': mode} | |
503 | if mode != 'r': |
|
503 | if mode != 'r': | |
504 | args[r'checkambig'] = self._checkambig |
|
504 | args[r'checkambig'] = self._checkambig | |
505 | if mode == 'w': |
|
505 | if mode == 'w': | |
506 | args[r'atomictemp'] = True |
|
506 | args[r'atomictemp'] = True | |
507 | return self.opener(self.indexfile, **args) |
|
507 | return self.opener(self.indexfile, **args) | |
508 |
|
508 | |||
509 | def _datafp(self, mode='r'): |
|
509 | def _datafp(self, mode='r'): | |
510 | """file object for the revlog's data file""" |
|
510 | """file object for the revlog's data file""" | |
511 | return self.opener(self.datafile, mode=mode) |
|
511 | return self.opener(self.datafile, mode=mode) | |
512 |
|
512 | |||
513 | @contextlib.contextmanager |
|
513 | @contextlib.contextmanager | |
514 | def _datareadfp(self, existingfp=None): |
|
514 | def _datareadfp(self, existingfp=None): | |
515 | """file object suitable to read data""" |
|
515 | """file object suitable to read data""" | |
516 | # Use explicit file handle, if given. |
|
516 | # Use explicit file handle, if given. | |
517 | if existingfp is not None: |
|
517 | if existingfp is not None: | |
518 | yield existingfp |
|
518 | yield existingfp | |
519 |
|
519 | |||
520 | # Use a file handle being actively used for writes, if available. |
|
520 | # Use a file handle being actively used for writes, if available. | |
521 | # There is some danger to doing this because reads will seek the |
|
521 | # There is some danger to doing this because reads will seek the | |
522 | # file. However, _writeentry() performs a SEEK_END before all writes, |
|
522 | # file. However, _writeentry() performs a SEEK_END before all writes, | |
523 | # so we should be safe. |
|
523 | # so we should be safe. | |
524 | elif self._writinghandles: |
|
524 | elif self._writinghandles: | |
525 | if self._inline: |
|
525 | if self._inline: | |
526 | yield self._writinghandles[0] |
|
526 | yield self._writinghandles[0] | |
527 | else: |
|
527 | else: | |
528 | yield self._writinghandles[1] |
|
528 | yield self._writinghandles[1] | |
529 |
|
529 | |||
530 | # Otherwise open a new file handle. |
|
530 | # Otherwise open a new file handle. | |
531 | else: |
|
531 | else: | |
532 | if self._inline: |
|
532 | if self._inline: | |
533 | func = self._indexfp |
|
533 | func = self._indexfp | |
534 | else: |
|
534 | else: | |
535 | func = self._datafp |
|
535 | func = self._datafp | |
536 | with func() as fp: |
|
536 | with func() as fp: | |
537 | yield fp |
|
537 | yield fp | |
538 |
|
538 | |||
539 | def tip(self): |
|
539 | def tip(self): | |
540 | return self.node(len(self.index) - 1) |
|
540 | return self.node(len(self.index) - 1) | |
541 | def __contains__(self, rev): |
|
541 | def __contains__(self, rev): | |
542 | return 0 <= rev < len(self) |
|
542 | return 0 <= rev < len(self) | |
543 | def __len__(self): |
|
543 | def __len__(self): | |
544 | return len(self.index) |
|
544 | return len(self.index) | |
545 | def __iter__(self): |
|
545 | def __iter__(self): | |
546 | return iter(pycompat.xrange(len(self))) |
|
546 | return iter(pycompat.xrange(len(self))) | |
547 | def revs(self, start=0, stop=None): |
|
547 | def revs(self, start=0, stop=None): | |
548 | """iterate over all rev in this revlog (from start to stop)""" |
|
548 | """iterate over all rev in this revlog (from start to stop)""" | |
549 | return storageutil.iterrevs(len(self), start=start, stop=stop) |
|
549 | return storageutil.iterrevs(len(self), start=start, stop=stop) | |
550 |
|
550 | |||
551 | @util.propertycache |
|
551 | @util.propertycache | |
552 | def nodemap(self): |
|
552 | def nodemap(self): | |
553 | if self.index: |
|
553 | if self.index: | |
554 | # populate mapping down to the initial node |
|
554 | # populate mapping down to the initial node | |
555 | node0 = self.index[0][7] # get around changelog filtering |
|
555 | node0 = self.index[0][7] # get around changelog filtering | |
556 | self.rev(node0) |
|
556 | self.rev(node0) | |
557 | return self._nodecache |
|
557 | return self._nodecache | |
558 |
|
558 | |||
559 | def hasnode(self, node): |
|
559 | def hasnode(self, node): | |
560 | try: |
|
560 | try: | |
561 | self.rev(node) |
|
561 | self.rev(node) | |
562 | return True |
|
562 | return True | |
563 | except KeyError: |
|
563 | except KeyError: | |
564 | return False |
|
564 | return False | |
565 |
|
565 | |||
566 | def candelta(self, baserev, rev): |
|
566 | def candelta(self, baserev, rev): | |
567 | """whether two revisions (baserev, rev) can be delta-ed or not""" |
|
567 | """whether two revisions (baserev, rev) can be delta-ed or not""" | |
568 | # Disable delta if either rev requires a content-changing flag |
|
568 | # Disable delta if either rev requires a content-changing flag | |
569 | # processor (ex. LFS). This is because such flag processor can alter |
|
569 | # processor (ex. LFS). This is because such flag processor can alter | |
570 | # the rawtext content that the delta will be based on, and two clients |
|
570 | # the rawtext content that the delta will be based on, and two clients | |
571 | # could have a same revlog node with different flags (i.e. different |
|
571 | # could have a same revlog node with different flags (i.e. different | |
572 | # rawtext contents) and the delta could be incompatible. |
|
572 | # rawtext contents) and the delta could be incompatible. | |
573 | if ((self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) |
|
573 | if ((self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) | |
574 | or (self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS)): |
|
574 | or (self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS)): | |
575 | return False |
|
575 | return False | |
576 | return True |
|
576 | return True | |
577 |
|
577 | |||
578 | def clearcaches(self): |
|
578 | def clearcaches(self): | |
579 | self._revisioncache = None |
|
579 | self._revisioncache = None | |
580 | self._chainbasecache.clear() |
|
580 | self._chainbasecache.clear() | |
581 | self._chunkcache = (0, '') |
|
581 | self._chunkcache = (0, '') | |
582 | self._pcache = {} |
|
582 | self._pcache = {} | |
583 |
|
583 | |||
584 | try: |
|
584 | try: | |
585 | # If we are using the native C version, you are in a fun case |
|
585 | # If we are using the native C version, you are in a fun case | |
586 | # where self.index, self.nodemap and self._nodecaches is the same |
|
586 | # where self.index, self.nodemap and self._nodecaches is the same | |
587 | # object. |
|
587 | # object. | |
588 | self._nodecache.clearcaches() |
|
588 | self._nodecache.clearcaches() | |
589 | except AttributeError: |
|
589 | except AttributeError: | |
590 | self._nodecache = {nullid: nullrev} |
|
590 | self._nodecache = {nullid: nullrev} | |
591 | self._nodepos = None |
|
591 | self._nodepos = None | |
592 |
|
592 | |||
593 | def rev(self, node): |
|
593 | def rev(self, node): | |
594 | try: |
|
594 | try: | |
595 | return self._nodecache[node] |
|
595 | return self._nodecache[node] | |
596 | except TypeError: |
|
596 | except TypeError: | |
597 | raise |
|
597 | raise | |
598 | except error.RevlogError: |
|
598 | except error.RevlogError: | |
599 | # parsers.c radix tree lookup failed |
|
599 | # parsers.c radix tree lookup failed | |
600 | if node == wdirid or node in wdirfilenodeids: |
|
600 | if node == wdirid or node in wdirfilenodeids: | |
601 | raise error.WdirUnsupported |
|
601 | raise error.WdirUnsupported | |
602 | raise error.LookupError(node, self.indexfile, _('no node')) |
|
602 | raise error.LookupError(node, self.indexfile, _('no node')) | |
603 | except KeyError: |
|
603 | except KeyError: | |
604 | # pure python cache lookup failed |
|
604 | # pure python cache lookup failed | |
605 | n = self._nodecache |
|
605 | n = self._nodecache | |
606 | i = self.index |
|
606 | i = self.index | |
607 | p = self._nodepos |
|
607 | p = self._nodepos | |
608 | if p is None: |
|
608 | if p is None: | |
609 | p = len(i) - 1 |
|
609 | p = len(i) - 1 | |
610 | else: |
|
610 | else: | |
611 | assert p < len(i) |
|
611 | assert p < len(i) | |
612 | for r in pycompat.xrange(p, -1, -1): |
|
612 | for r in pycompat.xrange(p, -1, -1): | |
613 | v = i[r][7] |
|
613 | v = i[r][7] | |
614 | n[v] = r |
|
614 | n[v] = r | |
615 | if v == node: |
|
615 | if v == node: | |
616 | self._nodepos = r - 1 |
|
616 | self._nodepos = r - 1 | |
617 | return r |
|
617 | return r | |
618 | if node == wdirid or node in wdirfilenodeids: |
|
618 | if node == wdirid or node in wdirfilenodeids: | |
619 | raise error.WdirUnsupported |
|
619 | raise error.WdirUnsupported | |
620 | raise error.LookupError(node, self.indexfile, _('no node')) |
|
620 | raise error.LookupError(node, self.indexfile, _('no node')) | |
621 |
|
621 | |||
622 | # Accessors for index entries. |
|
622 | # Accessors for index entries. | |
623 |
|
623 | |||
624 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes |
|
624 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes | |
625 | # are flags. |
|
625 | # are flags. | |
626 | def start(self, rev): |
|
626 | def start(self, rev): | |
627 | return int(self.index[rev][0] >> 16) |
|
627 | return int(self.index[rev][0] >> 16) | |
628 |
|
628 | |||
629 | def flags(self, rev): |
|
629 | def flags(self, rev): | |
630 | return self.index[rev][0] & 0xFFFF |
|
630 | return self.index[rev][0] & 0xFFFF | |
631 |
|
631 | |||
632 | def length(self, rev): |
|
632 | def length(self, rev): | |
633 | return self.index[rev][1] |
|
633 | return self.index[rev][1] | |
634 |
|
634 | |||
635 | def rawsize(self, rev): |
|
635 | def rawsize(self, rev): | |
636 | """return the length of the uncompressed text for a given revision""" |
|
636 | """return the length of the uncompressed text for a given revision""" | |
637 | l = self.index[rev][2] |
|
637 | l = self.index[rev][2] | |
638 | if l >= 0: |
|
638 | if l >= 0: | |
639 | return l |
|
639 | return l | |
640 |
|
640 | |||
641 | t = self.rawdata(rev) |
|
641 | t = self.rawdata(rev) | |
642 | return len(t) |
|
642 | return len(t) | |
643 |
|
643 | |||
644 | def size(self, rev): |
|
644 | def size(self, rev): | |
645 | """length of non-raw text (processed by a "read" flag processor)""" |
|
645 | """length of non-raw text (processed by a "read" flag processor)""" | |
646 | # fast path: if no "read" flag processor could change the content, |
|
646 | # fast path: if no "read" flag processor could change the content, | |
647 | # size is rawsize. note: ELLIPSIS is known to not change the content. |
|
647 | # size is rawsize. note: ELLIPSIS is known to not change the content. | |
648 | flags = self.flags(rev) |
|
648 | flags = self.flags(rev) | |
649 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: |
|
649 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: | |
650 | return self.rawsize(rev) |
|
650 | return self.rawsize(rev) | |
651 |
|
651 | |||
652 | return len(self.revision(rev, raw=False)) |
|
652 | return len(self.revision(rev, raw=False)) | |
653 |
|
653 | |||
654 | def chainbase(self, rev): |
|
654 | def chainbase(self, rev): | |
655 | base = self._chainbasecache.get(rev) |
|
655 | base = self._chainbasecache.get(rev) | |
656 | if base is not None: |
|
656 | if base is not None: | |
657 | return base |
|
657 | return base | |
658 |
|
658 | |||
659 | index = self.index |
|
659 | index = self.index | |
660 | iterrev = rev |
|
660 | iterrev = rev | |
661 | base = index[iterrev][3] |
|
661 | base = index[iterrev][3] | |
662 | while base != iterrev: |
|
662 | while base != iterrev: | |
663 | iterrev = base |
|
663 | iterrev = base | |
664 | base = index[iterrev][3] |
|
664 | base = index[iterrev][3] | |
665 |
|
665 | |||
666 | self._chainbasecache[rev] = base |
|
666 | self._chainbasecache[rev] = base | |
667 | return base |
|
667 | return base | |
668 |
|
668 | |||
669 | def linkrev(self, rev): |
|
669 | def linkrev(self, rev): | |
670 | return self.index[rev][4] |
|
670 | return self.index[rev][4] | |
671 |
|
671 | |||
672 | def parentrevs(self, rev): |
|
672 | def parentrevs(self, rev): | |
673 | try: |
|
673 | try: | |
674 | entry = self.index[rev] |
|
674 | entry = self.index[rev] | |
675 | except IndexError: |
|
675 | except IndexError: | |
676 | if rev == wdirrev: |
|
676 | if rev == wdirrev: | |
677 | raise error.WdirUnsupported |
|
677 | raise error.WdirUnsupported | |
678 | raise |
|
678 | raise | |
679 |
|
679 | |||
680 | return entry[5], entry[6] |
|
680 | return entry[5], entry[6] | |
681 |
|
681 | |||
682 | # fast parentrevs(rev) where rev isn't filtered |
|
682 | # fast parentrevs(rev) where rev isn't filtered | |
683 | _uncheckedparentrevs = parentrevs |
|
683 | _uncheckedparentrevs = parentrevs | |
684 |
|
684 | |||
685 | def node(self, rev): |
|
685 | def node(self, rev): | |
686 | try: |
|
686 | try: | |
687 | return self.index[rev][7] |
|
687 | return self.index[rev][7] | |
688 | except IndexError: |
|
688 | except IndexError: | |
689 | if rev == wdirrev: |
|
689 | if rev == wdirrev: | |
690 | raise error.WdirUnsupported |
|
690 | raise error.WdirUnsupported | |
691 | raise |
|
691 | raise | |
692 |
|
692 | |||
693 | # Derived from index values. |
|
693 | # Derived from index values. | |
694 |
|
694 | |||
695 | def end(self, rev): |
|
695 | def end(self, rev): | |
696 | return self.start(rev) + self.length(rev) |
|
696 | return self.start(rev) + self.length(rev) | |
697 |
|
697 | |||
698 | def parents(self, node): |
|
698 | def parents(self, node): | |
699 | i = self.index |
|
699 | i = self.index | |
700 | d = i[self.rev(node)] |
|
700 | d = i[self.rev(node)] | |
701 | return i[d[5]][7], i[d[6]][7] # map revisions to nodes inline |
|
701 | return i[d[5]][7], i[d[6]][7] # map revisions to nodes inline | |
702 |
|
702 | |||
703 | def chainlen(self, rev): |
|
703 | def chainlen(self, rev): | |
704 | return self._chaininfo(rev)[0] |
|
704 | return self._chaininfo(rev)[0] | |
705 |
|
705 | |||
706 | def _chaininfo(self, rev): |
|
706 | def _chaininfo(self, rev): | |
707 | chaininfocache = self._chaininfocache |
|
707 | chaininfocache = self._chaininfocache | |
708 | if rev in chaininfocache: |
|
708 | if rev in chaininfocache: | |
709 | return chaininfocache[rev] |
|
709 | return chaininfocache[rev] | |
710 | index = self.index |
|
710 | index = self.index | |
711 | generaldelta = self._generaldelta |
|
711 | generaldelta = self._generaldelta | |
712 | iterrev = rev |
|
712 | iterrev = rev | |
713 | e = index[iterrev] |
|
713 | e = index[iterrev] | |
714 | clen = 0 |
|
714 | clen = 0 | |
715 | compresseddeltalen = 0 |
|
715 | compresseddeltalen = 0 | |
716 | while iterrev != e[3]: |
|
716 | while iterrev != e[3]: | |
717 | clen += 1 |
|
717 | clen += 1 | |
718 | compresseddeltalen += e[1] |
|
718 | compresseddeltalen += e[1] | |
719 | if generaldelta: |
|
719 | if generaldelta: | |
720 | iterrev = e[3] |
|
720 | iterrev = e[3] | |
721 | else: |
|
721 | else: | |
722 | iterrev -= 1 |
|
722 | iterrev -= 1 | |
723 | if iterrev in chaininfocache: |
|
723 | if iterrev in chaininfocache: | |
724 | t = chaininfocache[iterrev] |
|
724 | t = chaininfocache[iterrev] | |
725 | clen += t[0] |
|
725 | clen += t[0] | |
726 | compresseddeltalen += t[1] |
|
726 | compresseddeltalen += t[1] | |
727 | break |
|
727 | break | |
728 | e = index[iterrev] |
|
728 | e = index[iterrev] | |
729 | else: |
|
729 | else: | |
730 | # Add text length of base since decompressing that also takes |
|
730 | # Add text length of base since decompressing that also takes | |
731 | # work. For cache hits the length is already included. |
|
731 | # work. For cache hits the length is already included. | |
732 | compresseddeltalen += e[1] |
|
732 | compresseddeltalen += e[1] | |
733 | r = (clen, compresseddeltalen) |
|
733 | r = (clen, compresseddeltalen) | |
734 | chaininfocache[rev] = r |
|
734 | chaininfocache[rev] = r | |
735 | return r |
|
735 | return r | |
736 |
|
736 | |||
737 | def _deltachain(self, rev, stoprev=None): |
|
737 | def _deltachain(self, rev, stoprev=None): | |
738 | """Obtain the delta chain for a revision. |
|
738 | """Obtain the delta chain for a revision. | |
739 |
|
739 | |||
740 | ``stoprev`` specifies a revision to stop at. If not specified, we |
|
740 | ``stoprev`` specifies a revision to stop at. If not specified, we | |
741 | stop at the base of the chain. |
|
741 | stop at the base of the chain. | |
742 |
|
742 | |||
743 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of |
|
743 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of | |
744 | revs in ascending order and ``stopped`` is a bool indicating whether |
|
744 | revs in ascending order and ``stopped`` is a bool indicating whether | |
745 | ``stoprev`` was hit. |
|
745 | ``stoprev`` was hit. | |
746 | """ |
|
746 | """ | |
747 | # Try C implementation. |
|
747 | # Try C implementation. | |
748 | try: |
|
748 | try: | |
749 | return self.index.deltachain(rev, stoprev, self._generaldelta) |
|
749 | return self.index.deltachain(rev, stoprev, self._generaldelta) | |
750 | except AttributeError: |
|
750 | except AttributeError: | |
751 | pass |
|
751 | pass | |
752 |
|
752 | |||
753 | chain = [] |
|
753 | chain = [] | |
754 |
|
754 | |||
755 | # Alias to prevent attribute lookup in tight loop. |
|
755 | # Alias to prevent attribute lookup in tight loop. | |
756 | index = self.index |
|
756 | index = self.index | |
757 | generaldelta = self._generaldelta |
|
757 | generaldelta = self._generaldelta | |
758 |
|
758 | |||
759 | iterrev = rev |
|
759 | iterrev = rev | |
760 | e = index[iterrev] |
|
760 | e = index[iterrev] | |
761 | while iterrev != e[3] and iterrev != stoprev: |
|
761 | while iterrev != e[3] and iterrev != stoprev: | |
762 | chain.append(iterrev) |
|
762 | chain.append(iterrev) | |
763 | if generaldelta: |
|
763 | if generaldelta: | |
764 | iterrev = e[3] |
|
764 | iterrev = e[3] | |
765 | else: |
|
765 | else: | |
766 | iterrev -= 1 |
|
766 | iterrev -= 1 | |
767 | e = index[iterrev] |
|
767 | e = index[iterrev] | |
768 |
|
768 | |||
769 | if iterrev == stoprev: |
|
769 | if iterrev == stoprev: | |
770 | stopped = True |
|
770 | stopped = True | |
771 | else: |
|
771 | else: | |
772 | chain.append(iterrev) |
|
772 | chain.append(iterrev) | |
773 | stopped = False |
|
773 | stopped = False | |
774 |
|
774 | |||
775 | chain.reverse() |
|
775 | chain.reverse() | |
776 | return chain, stopped |
|
776 | return chain, stopped | |
777 |
|
777 | |||
778 | def ancestors(self, revs, stoprev=0, inclusive=False): |
|
778 | def ancestors(self, revs, stoprev=0, inclusive=False): | |
779 | """Generate the ancestors of 'revs' in reverse revision order. |
|
779 | """Generate the ancestors of 'revs' in reverse revision order. | |
780 | Does not generate revs lower than stoprev. |
|
780 | Does not generate revs lower than stoprev. | |
781 |
|
781 | |||
782 | See the documentation for ancestor.lazyancestors for more details.""" |
|
782 | See the documentation for ancestor.lazyancestors for more details.""" | |
783 |
|
783 | |||
784 | # first, make sure start revisions aren't filtered |
|
784 | # first, make sure start revisions aren't filtered | |
785 | revs = list(revs) |
|
785 | revs = list(revs) | |
786 | checkrev = self.node |
|
786 | checkrev = self.node | |
787 | for r in revs: |
|
787 | for r in revs: | |
788 | checkrev(r) |
|
788 | checkrev(r) | |
789 | # and we're sure ancestors aren't filtered as well |
|
789 | # and we're sure ancestors aren't filtered as well | |
790 |
|
790 | |||
791 | if rustancestor is not None: |
|
791 | if rustancestor is not None: | |
792 | lazyancestors = rustancestor.LazyAncestors |
|
792 | lazyancestors = rustancestor.LazyAncestors | |
793 | arg = self.index |
|
793 | arg = self.index | |
794 | elif util.safehasattr(parsers, 'rustlazyancestors'): |
|
794 | elif util.safehasattr(parsers, 'rustlazyancestors'): | |
795 | lazyancestors = ancestor.rustlazyancestors |
|
795 | lazyancestors = ancestor.rustlazyancestors | |
796 | arg = self.index |
|
796 | arg = self.index | |
797 | else: |
|
797 | else: | |
798 | lazyancestors = ancestor.lazyancestors |
|
798 | lazyancestors = ancestor.lazyancestors | |
799 | arg = self._uncheckedparentrevs |
|
799 | arg = self._uncheckedparentrevs | |
800 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) |
|
800 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) | |
801 |
|
801 | |||
802 | def descendants(self, revs): |
|
802 | def descendants(self, revs): | |
803 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
803 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) | |
804 |
|
804 | |||
805 | def findcommonmissing(self, common=None, heads=None): |
|
805 | def findcommonmissing(self, common=None, heads=None): | |
806 | """Return a tuple of the ancestors of common and the ancestors of heads |
|
806 | """Return a tuple of the ancestors of common and the ancestors of heads | |
807 | that are not ancestors of common. In revset terminology, we return the |
|
807 | that are not ancestors of common. In revset terminology, we return the | |
808 | tuple: |
|
808 | tuple: | |
809 |
|
809 | |||
810 | ::common, (::heads) - (::common) |
|
810 | ::common, (::heads) - (::common) | |
811 |
|
811 | |||
812 | The list is sorted by revision number, meaning it is |
|
812 | The list is sorted by revision number, meaning it is | |
813 | topologically sorted. |
|
813 | topologically sorted. | |
814 |
|
814 | |||
815 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
815 | 'heads' and 'common' are both lists of node IDs. If heads is | |
816 | not supplied, uses all of the revlog's heads. If common is not |
|
816 | not supplied, uses all of the revlog's heads. If common is not | |
817 | supplied, uses nullid.""" |
|
817 | supplied, uses nullid.""" | |
818 | if common is None: |
|
818 | if common is None: | |
819 | common = [nullid] |
|
819 | common = [nullid] | |
820 | if heads is None: |
|
820 | if heads is None: | |
821 | heads = self.heads() |
|
821 | heads = self.heads() | |
822 |
|
822 | |||
823 | common = [self.rev(n) for n in common] |
|
823 | common = [self.rev(n) for n in common] | |
824 | heads = [self.rev(n) for n in heads] |
|
824 | heads = [self.rev(n) for n in heads] | |
825 |
|
825 | |||
826 | # we want the ancestors, but inclusive |
|
826 | # we want the ancestors, but inclusive | |
827 | class lazyset(object): |
|
827 | class lazyset(object): | |
828 | def __init__(self, lazyvalues): |
|
828 | def __init__(self, lazyvalues): | |
829 | self.addedvalues = set() |
|
829 | self.addedvalues = set() | |
830 | self.lazyvalues = lazyvalues |
|
830 | self.lazyvalues = lazyvalues | |
831 |
|
831 | |||
832 | def __contains__(self, value): |
|
832 | def __contains__(self, value): | |
833 | return value in self.addedvalues or value in self.lazyvalues |
|
833 | return value in self.addedvalues or value in self.lazyvalues | |
834 |
|
834 | |||
835 | def __iter__(self): |
|
835 | def __iter__(self): | |
836 | added = self.addedvalues |
|
836 | added = self.addedvalues | |
837 | for r in added: |
|
837 | for r in added: | |
838 | yield r |
|
838 | yield r | |
839 | for r in self.lazyvalues: |
|
839 | for r in self.lazyvalues: | |
840 | if not r in added: |
|
840 | if not r in added: | |
841 | yield r |
|
841 | yield r | |
842 |
|
842 | |||
843 | def add(self, value): |
|
843 | def add(self, value): | |
844 | self.addedvalues.add(value) |
|
844 | self.addedvalues.add(value) | |
845 |
|
845 | |||
846 | def update(self, values): |
|
846 | def update(self, values): | |
847 | self.addedvalues.update(values) |
|
847 | self.addedvalues.update(values) | |
848 |
|
848 | |||
849 | has = lazyset(self.ancestors(common)) |
|
849 | has = lazyset(self.ancestors(common)) | |
850 | has.add(nullrev) |
|
850 | has.add(nullrev) | |
851 | has.update(common) |
|
851 | has.update(common) | |
852 |
|
852 | |||
853 | # take all ancestors from heads that aren't in has |
|
853 | # take all ancestors from heads that aren't in has | |
854 | missing = set() |
|
854 | missing = set() | |
855 | visit = collections.deque(r for r in heads if r not in has) |
|
855 | visit = collections.deque(r for r in heads if r not in has) | |
856 | while visit: |
|
856 | while visit: | |
857 | r = visit.popleft() |
|
857 | r = visit.popleft() | |
858 | if r in missing: |
|
858 | if r in missing: | |
859 | continue |
|
859 | continue | |
860 | else: |
|
860 | else: | |
861 | missing.add(r) |
|
861 | missing.add(r) | |
862 | for p in self.parentrevs(r): |
|
862 | for p in self.parentrevs(r): | |
863 | if p not in has: |
|
863 | if p not in has: | |
864 | visit.append(p) |
|
864 | visit.append(p) | |
865 | missing = list(missing) |
|
865 | missing = list(missing) | |
866 | missing.sort() |
|
866 | missing.sort() | |
867 | return has, [self.node(miss) for miss in missing] |
|
867 | return has, [self.node(miss) for miss in missing] | |
868 |
|
868 | |||
869 | def incrementalmissingrevs(self, common=None): |
|
869 | def incrementalmissingrevs(self, common=None): | |
870 | """Return an object that can be used to incrementally compute the |
|
870 | """Return an object that can be used to incrementally compute the | |
871 | revision numbers of the ancestors of arbitrary sets that are not |
|
871 | revision numbers of the ancestors of arbitrary sets that are not | |
872 | ancestors of common. This is an ancestor.incrementalmissingancestors |
|
872 | ancestors of common. This is an ancestor.incrementalmissingancestors | |
873 | object. |
|
873 | object. | |
874 |
|
874 | |||
875 | 'common' is a list of revision numbers. If common is not supplied, uses |
|
875 | 'common' is a list of revision numbers. If common is not supplied, uses | |
876 | nullrev. |
|
876 | nullrev. | |
877 | """ |
|
877 | """ | |
878 | if common is None: |
|
878 | if common is None: | |
879 | common = [nullrev] |
|
879 | common = [nullrev] | |
880 |
|
880 | |||
881 | if rustancestor is not None: |
|
881 | if rustancestor is not None: | |
882 | return rustancestor.MissingAncestors(self.index, common) |
|
882 | return rustancestor.MissingAncestors(self.index, common) | |
883 | return ancestor.incrementalmissingancestors(self.parentrevs, common) |
|
883 | return ancestor.incrementalmissingancestors(self.parentrevs, common) | |
884 |
|
884 | |||
885 | def findmissingrevs(self, common=None, heads=None): |
|
885 | def findmissingrevs(self, common=None, heads=None): | |
886 | """Return the revision numbers of the ancestors of heads that |
|
886 | """Return the revision numbers of the ancestors of heads that | |
887 | are not ancestors of common. |
|
887 | are not ancestors of common. | |
888 |
|
888 | |||
889 | More specifically, return a list of revision numbers corresponding to |
|
889 | More specifically, return a list of revision numbers corresponding to | |
890 | nodes N such that every N satisfies the following constraints: |
|
890 | nodes N such that every N satisfies the following constraints: | |
891 |
|
891 | |||
892 | 1. N is an ancestor of some node in 'heads' |
|
892 | 1. N is an ancestor of some node in 'heads' | |
893 | 2. N is not an ancestor of any node in 'common' |
|
893 | 2. N is not an ancestor of any node in 'common' | |
894 |
|
894 | |||
895 | The list is sorted by revision number, meaning it is |
|
895 | The list is sorted by revision number, meaning it is | |
896 | topologically sorted. |
|
896 | topologically sorted. | |
897 |
|
897 | |||
898 | 'heads' and 'common' are both lists of revision numbers. If heads is |
|
898 | 'heads' and 'common' are both lists of revision numbers. If heads is | |
899 | not supplied, uses all of the revlog's heads. If common is not |
|
899 | not supplied, uses all of the revlog's heads. If common is not | |
900 | supplied, uses nullid.""" |
|
900 | supplied, uses nullid.""" | |
901 | if common is None: |
|
901 | if common is None: | |
902 | common = [nullrev] |
|
902 | common = [nullrev] | |
903 | if heads is None: |
|
903 | if heads is None: | |
904 | heads = self.headrevs() |
|
904 | heads = self.headrevs() | |
905 |
|
905 | |||
906 | inc = self.incrementalmissingrevs(common=common) |
|
906 | inc = self.incrementalmissingrevs(common=common) | |
907 | return inc.missingancestors(heads) |
|
907 | return inc.missingancestors(heads) | |
908 |
|
908 | |||
909 | def findmissing(self, common=None, heads=None): |
|
909 | def findmissing(self, common=None, heads=None): | |
910 | """Return the ancestors of heads that are not ancestors of common. |
|
910 | """Return the ancestors of heads that are not ancestors of common. | |
911 |
|
911 | |||
912 | More specifically, return a list of nodes N such that every N |
|
912 | More specifically, return a list of nodes N such that every N | |
913 | satisfies the following constraints: |
|
913 | satisfies the following constraints: | |
914 |
|
914 | |||
915 | 1. N is an ancestor of some node in 'heads' |
|
915 | 1. N is an ancestor of some node in 'heads' | |
916 | 2. N is not an ancestor of any node in 'common' |
|
916 | 2. N is not an ancestor of any node in 'common' | |
917 |
|
917 | |||
918 | The list is sorted by revision number, meaning it is |
|
918 | The list is sorted by revision number, meaning it is | |
919 | topologically sorted. |
|
919 | topologically sorted. | |
920 |
|
920 | |||
921 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
921 | 'heads' and 'common' are both lists of node IDs. If heads is | |
922 | not supplied, uses all of the revlog's heads. If common is not |
|
922 | not supplied, uses all of the revlog's heads. If common is not | |
923 | supplied, uses nullid.""" |
|
923 | supplied, uses nullid.""" | |
924 | if common is None: |
|
924 | if common is None: | |
925 | common = [nullid] |
|
925 | common = [nullid] | |
926 | if heads is None: |
|
926 | if heads is None: | |
927 | heads = self.heads() |
|
927 | heads = self.heads() | |
928 |
|
928 | |||
929 | common = [self.rev(n) for n in common] |
|
929 | common = [self.rev(n) for n in common] | |
930 | heads = [self.rev(n) for n in heads] |
|
930 | heads = [self.rev(n) for n in heads] | |
931 |
|
931 | |||
932 | inc = self.incrementalmissingrevs(common=common) |
|
932 | inc = self.incrementalmissingrevs(common=common) | |
933 | return [self.node(r) for r in inc.missingancestors(heads)] |
|
933 | return [self.node(r) for r in inc.missingancestors(heads)] | |
934 |
|
934 | |||
935 | def nodesbetween(self, roots=None, heads=None): |
|
935 | def nodesbetween(self, roots=None, heads=None): | |
936 | """Return a topological path from 'roots' to 'heads'. |
|
936 | """Return a topological path from 'roots' to 'heads'. | |
937 |
|
937 | |||
938 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a |
|
938 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a | |
939 | topologically sorted list of all nodes N that satisfy both of |
|
939 | topologically sorted list of all nodes N that satisfy both of | |
940 | these constraints: |
|
940 | these constraints: | |
941 |
|
941 | |||
942 | 1. N is a descendant of some node in 'roots' |
|
942 | 1. N is a descendant of some node in 'roots' | |
943 | 2. N is an ancestor of some node in 'heads' |
|
943 | 2. N is an ancestor of some node in 'heads' | |
944 |
|
944 | |||
945 | Every node is considered to be both a descendant and an ancestor |
|
945 | Every node is considered to be both a descendant and an ancestor | |
946 | of itself, so every reachable node in 'roots' and 'heads' will be |
|
946 | of itself, so every reachable node in 'roots' and 'heads' will be | |
947 | included in 'nodes'. |
|
947 | included in 'nodes'. | |
948 |
|
948 | |||
949 | 'outroots' is the list of reachable nodes in 'roots', i.e., the |
|
949 | 'outroots' is the list of reachable nodes in 'roots', i.e., the | |
950 | subset of 'roots' that is returned in 'nodes'. Likewise, |
|
950 | subset of 'roots' that is returned in 'nodes'. Likewise, | |
951 | 'outheads' is the subset of 'heads' that is also in 'nodes'. |
|
951 | 'outheads' is the subset of 'heads' that is also in 'nodes'. | |
952 |
|
952 | |||
953 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is |
|
953 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is | |
954 | unspecified, uses nullid as the only root. If 'heads' is |
|
954 | unspecified, uses nullid as the only root. If 'heads' is | |
955 | unspecified, uses list of all of the revlog's heads.""" |
|
955 | unspecified, uses list of all of the revlog's heads.""" | |
956 | nonodes = ([], [], []) |
|
956 | nonodes = ([], [], []) | |
957 | if roots is not None: |
|
957 | if roots is not None: | |
958 | roots = list(roots) |
|
958 | roots = list(roots) | |
959 | if not roots: |
|
959 | if not roots: | |
960 | return nonodes |
|
960 | return nonodes | |
961 | lowestrev = min([self.rev(n) for n in roots]) |
|
961 | lowestrev = min([self.rev(n) for n in roots]) | |
962 | else: |
|
962 | else: | |
963 | roots = [nullid] # Everybody's a descendant of nullid |
|
963 | roots = [nullid] # Everybody's a descendant of nullid | |
964 | lowestrev = nullrev |
|
964 | lowestrev = nullrev | |
965 | if (lowestrev == nullrev) and (heads is None): |
|
965 | if (lowestrev == nullrev) and (heads is None): | |
966 | # We want _all_ the nodes! |
|
966 | # We want _all_ the nodes! | |
967 | return ([self.node(r) for r in self], [nullid], list(self.heads())) |
|
967 | return ([self.node(r) for r in self], [nullid], list(self.heads())) | |
968 | if heads is None: |
|
968 | if heads is None: | |
969 | # All nodes are ancestors, so the latest ancestor is the last |
|
969 | # All nodes are ancestors, so the latest ancestor is the last | |
970 | # node. |
|
970 | # node. | |
971 | highestrev = len(self) - 1 |
|
971 | highestrev = len(self) - 1 | |
972 | # Set ancestors to None to signal that every node is an ancestor. |
|
972 | # Set ancestors to None to signal that every node is an ancestor. | |
973 | ancestors = None |
|
973 | ancestors = None | |
974 | # Set heads to an empty dictionary for later discovery of heads |
|
974 | # Set heads to an empty dictionary for later discovery of heads | |
975 | heads = {} |
|
975 | heads = {} | |
976 | else: |
|
976 | else: | |
977 | heads = list(heads) |
|
977 | heads = list(heads) | |
978 | if not heads: |
|
978 | if not heads: | |
979 | return nonodes |
|
979 | return nonodes | |
980 | ancestors = set() |
|
980 | ancestors = set() | |
981 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
981 | # Turn heads into a dictionary so we can remove 'fake' heads. | |
982 | # Also, later we will be using it to filter out the heads we can't |
|
982 | # Also, later we will be using it to filter out the heads we can't | |
983 | # find from roots. |
|
983 | # find from roots. | |
984 | heads = dict.fromkeys(heads, False) |
|
984 | heads = dict.fromkeys(heads, False) | |
985 | # Start at the top and keep marking parents until we're done. |
|
985 | # Start at the top and keep marking parents until we're done. | |
986 | nodestotag = set(heads) |
|
986 | nodestotag = set(heads) | |
987 | # Remember where the top was so we can use it as a limit later. |
|
987 | # Remember where the top was so we can use it as a limit later. | |
988 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
988 | highestrev = max([self.rev(n) for n in nodestotag]) | |
989 | while nodestotag: |
|
989 | while nodestotag: | |
990 | # grab a node to tag |
|
990 | # grab a node to tag | |
991 | n = nodestotag.pop() |
|
991 | n = nodestotag.pop() | |
992 | # Never tag nullid |
|
992 | # Never tag nullid | |
993 | if n == nullid: |
|
993 | if n == nullid: | |
994 | continue |
|
994 | continue | |
995 | # A node's revision number represents its place in a |
|
995 | # A node's revision number represents its place in a | |
996 | # topologically sorted list of nodes. |
|
996 | # topologically sorted list of nodes. | |
997 | r = self.rev(n) |
|
997 | r = self.rev(n) | |
998 | if r >= lowestrev: |
|
998 | if r >= lowestrev: | |
999 | if n not in ancestors: |
|
999 | if n not in ancestors: | |
1000 | # If we are possibly a descendant of one of the roots |
|
1000 | # If we are possibly a descendant of one of the roots | |
1001 | # and we haven't already been marked as an ancestor |
|
1001 | # and we haven't already been marked as an ancestor | |
1002 | ancestors.add(n) # Mark as ancestor |
|
1002 | ancestors.add(n) # Mark as ancestor | |
1003 | # Add non-nullid parents to list of nodes to tag. |
|
1003 | # Add non-nullid parents to list of nodes to tag. | |
1004 | nodestotag.update([p for p in self.parents(n) if |
|
1004 | nodestotag.update([p for p in self.parents(n) if | |
1005 | p != nullid]) |
|
1005 | p != nullid]) | |
1006 | elif n in heads: # We've seen it before, is it a fake head? |
|
1006 | elif n in heads: # We've seen it before, is it a fake head? | |
1007 | # So it is, real heads should not be the ancestors of |
|
1007 | # So it is, real heads should not be the ancestors of | |
1008 | # any other heads. |
|
1008 | # any other heads. | |
1009 | heads.pop(n) |
|
1009 | heads.pop(n) | |
1010 | if not ancestors: |
|
1010 | if not ancestors: | |
1011 | return nonodes |
|
1011 | return nonodes | |
1012 | # Now that we have our set of ancestors, we want to remove any |
|
1012 | # Now that we have our set of ancestors, we want to remove any | |
1013 | # roots that are not ancestors. |
|
1013 | # roots that are not ancestors. | |
1014 |
|
1014 | |||
1015 | # If one of the roots was nullid, everything is included anyway. |
|
1015 | # If one of the roots was nullid, everything is included anyway. | |
1016 | if lowestrev > nullrev: |
|
1016 | if lowestrev > nullrev: | |
1017 | # But, since we weren't, let's recompute the lowest rev to not |
|
1017 | # But, since we weren't, let's recompute the lowest rev to not | |
1018 | # include roots that aren't ancestors. |
|
1018 | # include roots that aren't ancestors. | |
1019 |
|
1019 | |||
1020 | # Filter out roots that aren't ancestors of heads |
|
1020 | # Filter out roots that aren't ancestors of heads | |
1021 | roots = [root for root in roots if root in ancestors] |
|
1021 | roots = [root for root in roots if root in ancestors] | |
1022 | # Recompute the lowest revision |
|
1022 | # Recompute the lowest revision | |
1023 | if roots: |
|
1023 | if roots: | |
1024 | lowestrev = min([self.rev(root) for root in roots]) |
|
1024 | lowestrev = min([self.rev(root) for root in roots]) | |
1025 | else: |
|
1025 | else: | |
1026 | # No more roots? Return empty list |
|
1026 | # No more roots? Return empty list | |
1027 | return nonodes |
|
1027 | return nonodes | |
1028 | else: |
|
1028 | else: | |
1029 | # We are descending from nullid, and don't need to care about |
|
1029 | # We are descending from nullid, and don't need to care about | |
1030 | # any other roots. |
|
1030 | # any other roots. | |
1031 | lowestrev = nullrev |
|
1031 | lowestrev = nullrev | |
1032 | roots = [nullid] |
|
1032 | roots = [nullid] | |
1033 | # Transform our roots list into a set. |
|
1033 | # Transform our roots list into a set. | |
1034 | descendants = set(roots) |
|
1034 | descendants = set(roots) | |
1035 | # Also, keep the original roots so we can filter out roots that aren't |
|
1035 | # Also, keep the original roots so we can filter out roots that aren't | |
1036 | # 'real' roots (i.e. are descended from other roots). |
|
1036 | # 'real' roots (i.e. are descended from other roots). | |
1037 | roots = descendants.copy() |
|
1037 | roots = descendants.copy() | |
1038 | # Our topologically sorted list of output nodes. |
|
1038 | # Our topologically sorted list of output nodes. | |
1039 | orderedout = [] |
|
1039 | orderedout = [] | |
1040 | # Don't start at nullid since we don't want nullid in our output list, |
|
1040 | # Don't start at nullid since we don't want nullid in our output list, | |
1041 | # and if nullid shows up in descendants, empty parents will look like |
|
1041 | # and if nullid shows up in descendants, empty parents will look like | |
1042 | # they're descendants. |
|
1042 | # they're descendants. | |
1043 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): |
|
1043 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): | |
1044 | n = self.node(r) |
|
1044 | n = self.node(r) | |
1045 | isdescendant = False |
|
1045 | isdescendant = False | |
1046 | if lowestrev == nullrev: # Everybody is a descendant of nullid |
|
1046 | if lowestrev == nullrev: # Everybody is a descendant of nullid | |
1047 | isdescendant = True |
|
1047 | isdescendant = True | |
1048 | elif n in descendants: |
|
1048 | elif n in descendants: | |
1049 | # n is already a descendant |
|
1049 | # n is already a descendant | |
1050 | isdescendant = True |
|
1050 | isdescendant = True | |
1051 | # This check only needs to be done here because all the roots |
|
1051 | # This check only needs to be done here because all the roots | |
1052 | # will start being marked is descendants before the loop. |
|
1052 | # will start being marked is descendants before the loop. | |
1053 | if n in roots: |
|
1053 | if n in roots: | |
1054 | # If n was a root, check if it's a 'real' root. |
|
1054 | # If n was a root, check if it's a 'real' root. | |
1055 | p = tuple(self.parents(n)) |
|
1055 | p = tuple(self.parents(n)) | |
1056 | # If any of its parents are descendants, it's not a root. |
|
1056 | # If any of its parents are descendants, it's not a root. | |
1057 | if (p[0] in descendants) or (p[1] in descendants): |
|
1057 | if (p[0] in descendants) or (p[1] in descendants): | |
1058 | roots.remove(n) |
|
1058 | roots.remove(n) | |
1059 | else: |
|
1059 | else: | |
1060 | p = tuple(self.parents(n)) |
|
1060 | p = tuple(self.parents(n)) | |
1061 | # A node is a descendant if either of its parents are |
|
1061 | # A node is a descendant if either of its parents are | |
1062 | # descendants. (We seeded the dependents list with the roots |
|
1062 | # descendants. (We seeded the dependents list with the roots | |
1063 | # up there, remember?) |
|
1063 | # up there, remember?) | |
1064 | if (p[0] in descendants) or (p[1] in descendants): |
|
1064 | if (p[0] in descendants) or (p[1] in descendants): | |
1065 | descendants.add(n) |
|
1065 | descendants.add(n) | |
1066 | isdescendant = True |
|
1066 | isdescendant = True | |
1067 | if isdescendant and ((ancestors is None) or (n in ancestors)): |
|
1067 | if isdescendant and ((ancestors is None) or (n in ancestors)): | |
1068 | # Only include nodes that are both descendants and ancestors. |
|
1068 | # Only include nodes that are both descendants and ancestors. | |
1069 | orderedout.append(n) |
|
1069 | orderedout.append(n) | |
1070 | if (ancestors is not None) and (n in heads): |
|
1070 | if (ancestors is not None) and (n in heads): | |
1071 | # We're trying to figure out which heads are reachable |
|
1071 | # We're trying to figure out which heads are reachable | |
1072 | # from roots. |
|
1072 | # from roots. | |
1073 | # Mark this head as having been reached |
|
1073 | # Mark this head as having been reached | |
1074 | heads[n] = True |
|
1074 | heads[n] = True | |
1075 | elif ancestors is None: |
|
1075 | elif ancestors is None: | |
1076 | # Otherwise, we're trying to discover the heads. |
|
1076 | # Otherwise, we're trying to discover the heads. | |
1077 | # Assume this is a head because if it isn't, the next step |
|
1077 | # Assume this is a head because if it isn't, the next step | |
1078 | # will eventually remove it. |
|
1078 | # will eventually remove it. | |
1079 | heads[n] = True |
|
1079 | heads[n] = True | |
1080 | # But, obviously its parents aren't. |
|
1080 | # But, obviously its parents aren't. | |
1081 | for p in self.parents(n): |
|
1081 | for p in self.parents(n): | |
1082 | heads.pop(p, None) |
|
1082 | heads.pop(p, None) | |
1083 | heads = [head for head, flag in heads.iteritems() if flag] |
|
1083 | heads = [head for head, flag in heads.iteritems() if flag] | |
1084 | roots = list(roots) |
|
1084 | roots = list(roots) | |
1085 | assert orderedout |
|
1085 | assert orderedout | |
1086 | assert roots |
|
1086 | assert roots | |
1087 | assert heads |
|
1087 | assert heads | |
1088 | return (orderedout, roots, heads) |
|
1088 | return (orderedout, roots, heads) | |
1089 |
|
1089 | |||
1090 | def headrevs(self, revs=None): |
|
1090 | def headrevs(self, revs=None): | |
1091 | if revs is None: |
|
1091 | if revs is None: | |
1092 | try: |
|
1092 | try: | |
1093 | return self.index.headrevs() |
|
1093 | return self.index.headrevs() | |
1094 | except AttributeError: |
|
1094 | except AttributeError: | |
1095 | return self._headrevs() |
|
1095 | return self._headrevs() | |
1096 | if rustdagop is not None: |
|
1096 | if rustdagop is not None: | |
1097 | return rustdagop.headrevs(self.index, revs) |
|
1097 | return rustdagop.headrevs(self.index, revs) | |
1098 | return dagop.headrevs(revs, self._uncheckedparentrevs) |
|
1098 | return dagop.headrevs(revs, self._uncheckedparentrevs) | |
1099 |
|
1099 | |||
1100 | def computephases(self, roots): |
|
1100 | def computephases(self, roots): | |
1101 | return self.index.computephasesmapsets(roots) |
|
1101 | return self.index.computephasesmapsets(roots) | |
1102 |
|
1102 | |||
1103 | def _headrevs(self): |
|
1103 | def _headrevs(self): | |
1104 | count = len(self) |
|
1104 | count = len(self) | |
1105 | if not count: |
|
1105 | if not count: | |
1106 | return [nullrev] |
|
1106 | return [nullrev] | |
1107 | # we won't iter over filtered rev so nobody is a head at start |
|
1107 | # we won't iter over filtered rev so nobody is a head at start | |
1108 | ishead = [0] * (count + 1) |
|
1108 | ishead = [0] * (count + 1) | |
1109 | index = self.index |
|
1109 | index = self.index | |
1110 | for r in self: |
|
1110 | for r in self: | |
1111 | ishead[r] = 1 # I may be an head |
|
1111 | ishead[r] = 1 # I may be an head | |
1112 | e = index[r] |
|
1112 | e = index[r] | |
1113 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not |
|
1113 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not | |
1114 | return [r for r, val in enumerate(ishead) if val] |
|
1114 | return [r for r, val in enumerate(ishead) if val] | |
1115 |
|
1115 | |||
1116 | def heads(self, start=None, stop=None): |
|
1116 | def heads(self, start=None, stop=None): | |
1117 | """return the list of all nodes that have no children |
|
1117 | """return the list of all nodes that have no children | |
1118 |
|
1118 | |||
1119 | if start is specified, only heads that are descendants of |
|
1119 | if start is specified, only heads that are descendants of | |
1120 | start will be returned |
|
1120 | start will be returned | |
1121 | if stop is specified, it will consider all the revs from stop |
|
1121 | if stop is specified, it will consider all the revs from stop | |
1122 | as if they had no children |
|
1122 | as if they had no children | |
1123 | """ |
|
1123 | """ | |
1124 | if start is None and stop is None: |
|
1124 | if start is None and stop is None: | |
1125 | if not len(self): |
|
1125 | if not len(self): | |
1126 | return [nullid] |
|
1126 | return [nullid] | |
1127 | return [self.node(r) for r in self.headrevs()] |
|
1127 | return [self.node(r) for r in self.headrevs()] | |
1128 |
|
1128 | |||
1129 | if start is None: |
|
1129 | if start is None: | |
1130 | start = nullrev |
|
1130 | start = nullrev | |
1131 | else: |
|
1131 | else: | |
1132 | start = self.rev(start) |
|
1132 | start = self.rev(start) | |
1133 |
|
1133 | |||
1134 | stoprevs = set(self.rev(n) for n in stop or []) |
|
1134 | stoprevs = set(self.rev(n) for n in stop or []) | |
1135 |
|
1135 | |||
1136 | revs = dagop.headrevssubset(self.revs, self.parentrevs, startrev=start, |
|
1136 | revs = dagop.headrevssubset(self.revs, self.parentrevs, startrev=start, | |
1137 | stoprevs=stoprevs) |
|
1137 | stoprevs=stoprevs) | |
1138 |
|
1138 | |||
1139 | return [self.node(rev) for rev in revs] |
|
1139 | return [self.node(rev) for rev in revs] | |
1140 |
|
1140 | |||
1141 | def children(self, node): |
|
1141 | def children(self, node): | |
1142 | """find the children of a given node""" |
|
1142 | """find the children of a given node""" | |
1143 | c = [] |
|
1143 | c = [] | |
1144 | p = self.rev(node) |
|
1144 | p = self.rev(node) | |
1145 | for r in self.revs(start=p + 1): |
|
1145 | for r in self.revs(start=p + 1): | |
1146 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
1146 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
1147 | if prevs: |
|
1147 | if prevs: | |
1148 | for pr in prevs: |
|
1148 | for pr in prevs: | |
1149 | if pr == p: |
|
1149 | if pr == p: | |
1150 | c.append(self.node(r)) |
|
1150 | c.append(self.node(r)) | |
1151 | elif p == nullrev: |
|
1151 | elif p == nullrev: | |
1152 | c.append(self.node(r)) |
|
1152 | c.append(self.node(r)) | |
1153 | return c |
|
1153 | return c | |
1154 |
|
1154 | |||
1155 | def commonancestorsheads(self, a, b): |
|
1155 | def commonancestorsheads(self, a, b): | |
1156 | """calculate all the heads of the common ancestors of nodes a and b""" |
|
1156 | """calculate all the heads of the common ancestors of nodes a and b""" | |
1157 | a, b = self.rev(a), self.rev(b) |
|
1157 | a, b = self.rev(a), self.rev(b) | |
1158 | ancs = self._commonancestorsheads(a, b) |
|
1158 | ancs = self._commonancestorsheads(a, b) | |
1159 | return pycompat.maplist(self.node, ancs) |
|
1159 | return pycompat.maplist(self.node, ancs) | |
1160 |
|
1160 | |||
1161 | def _commonancestorsheads(self, *revs): |
|
1161 | def _commonancestorsheads(self, *revs): | |
1162 | """calculate all the heads of the common ancestors of revs""" |
|
1162 | """calculate all the heads of the common ancestors of revs""" | |
1163 | try: |
|
1163 | try: | |
1164 | ancs = self.index.commonancestorsheads(*revs) |
|
1164 | ancs = self.index.commonancestorsheads(*revs) | |
1165 | except (AttributeError, OverflowError): # C implementation failed |
|
1165 | except (AttributeError, OverflowError): # C implementation failed | |
1166 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) |
|
1166 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) | |
1167 | return ancs |
|
1167 | return ancs | |
1168 |
|
1168 | |||
1169 | def isancestor(self, a, b): |
|
1169 | def isancestor(self, a, b): | |
1170 | """return True if node a is an ancestor of node b |
|
1170 | """return True if node a is an ancestor of node b | |
1171 |
|
1171 | |||
1172 | A revision is considered an ancestor of itself.""" |
|
1172 | A revision is considered an ancestor of itself.""" | |
1173 | a, b = self.rev(a), self.rev(b) |
|
1173 | a, b = self.rev(a), self.rev(b) | |
1174 | return self.isancestorrev(a, b) |
|
1174 | return self.isancestorrev(a, b) | |
1175 |
|
1175 | |||
1176 | def isancestorrev(self, a, b): |
|
1176 | def isancestorrev(self, a, b): | |
1177 | """return True if revision a is an ancestor of revision b |
|
1177 | """return True if revision a is an ancestor of revision b | |
1178 |
|
1178 | |||
1179 | A revision is considered an ancestor of itself. |
|
1179 | A revision is considered an ancestor of itself. | |
1180 |
|
1180 | |||
1181 | The implementation of this is trivial but the use of |
|
1181 | The implementation of this is trivial but the use of | |
1182 | reachableroots is not.""" |
|
1182 | reachableroots is not.""" | |
1183 | if a == nullrev: |
|
1183 | if a == nullrev: | |
1184 | return True |
|
1184 | return True | |
1185 | elif a == b: |
|
1185 | elif a == b: | |
1186 | return True |
|
1186 | return True | |
1187 | elif a > b: |
|
1187 | elif a > b: | |
1188 | return False |
|
1188 | return False | |
1189 | return bool(self.reachableroots(a, [b], [a], includepath=False)) |
|
1189 | return bool(self.reachableroots(a, [b], [a], includepath=False)) | |
1190 |
|
1190 | |||
1191 | def reachableroots(self, minroot, heads, roots, includepath=False): |
|
1191 | def reachableroots(self, minroot, heads, roots, includepath=False): | |
1192 | """return (heads(::<roots> and <roots>::<heads>)) |
|
1192 | """return (heads(::<roots> and <roots>::<heads>)) | |
1193 |
|
1193 | |||
1194 | If includepath is True, return (<roots>::<heads>).""" |
|
1194 | If includepath is True, return (<roots>::<heads>).""" | |
1195 | try: |
|
1195 | try: | |
1196 | return self.index.reachableroots2(minroot, heads, roots, |
|
1196 | return self.index.reachableroots2(minroot, heads, roots, | |
1197 | includepath) |
|
1197 | includepath) | |
1198 | except AttributeError: |
|
1198 | except AttributeError: | |
1199 | return dagop._reachablerootspure(self.parentrevs, |
|
1199 | return dagop._reachablerootspure(self.parentrevs, | |
1200 | minroot, roots, heads, includepath) |
|
1200 | minroot, roots, heads, includepath) | |
1201 |
|
1201 | |||
1202 | def ancestor(self, a, b): |
|
1202 | def ancestor(self, a, b): | |
1203 | """calculate the "best" common ancestor of nodes a and b""" |
|
1203 | """calculate the "best" common ancestor of nodes a and b""" | |
1204 |
|
1204 | |||
1205 | a, b = self.rev(a), self.rev(b) |
|
1205 | a, b = self.rev(a), self.rev(b) | |
1206 | try: |
|
1206 | try: | |
1207 | ancs = self.index.ancestors(a, b) |
|
1207 | ancs = self.index.ancestors(a, b) | |
1208 | except (AttributeError, OverflowError): |
|
1208 | except (AttributeError, OverflowError): | |
1209 | ancs = ancestor.ancestors(self.parentrevs, a, b) |
|
1209 | ancs = ancestor.ancestors(self.parentrevs, a, b) | |
1210 | if ancs: |
|
1210 | if ancs: | |
1211 | # choose a consistent winner when there's a tie |
|
1211 | # choose a consistent winner when there's a tie | |
1212 | return min(map(self.node, ancs)) |
|
1212 | return min(map(self.node, ancs)) | |
1213 | return nullid |
|
1213 | return nullid | |
1214 |
|
1214 | |||
1215 | def _match(self, id): |
|
1215 | def _match(self, id): | |
1216 | if isinstance(id, int): |
|
1216 | if isinstance(id, int): | |
1217 | # rev |
|
1217 | # rev | |
1218 | return self.node(id) |
|
1218 | return self.node(id) | |
1219 | if len(id) == 20: |
|
1219 | if len(id) == 20: | |
1220 | # possibly a binary node |
|
1220 | # possibly a binary node | |
1221 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
1221 | # odds of a binary node being all hex in ASCII are 1 in 10**25 | |
1222 | try: |
|
1222 | try: | |
1223 | node = id |
|
1223 | node = id | |
1224 | self.rev(node) # quick search the index |
|
1224 | self.rev(node) # quick search the index | |
1225 | return node |
|
1225 | return node | |
1226 | except error.LookupError: |
|
1226 | except error.LookupError: | |
1227 | pass # may be partial hex id |
|
1227 | pass # may be partial hex id | |
1228 | try: |
|
1228 | try: | |
1229 | # str(rev) |
|
1229 | # str(rev) | |
1230 | rev = int(id) |
|
1230 | rev = int(id) | |
1231 | if "%d" % rev != id: |
|
1231 | if "%d" % rev != id: | |
1232 | raise ValueError |
|
1232 | raise ValueError | |
1233 | if rev < 0: |
|
1233 | if rev < 0: | |
1234 | rev = len(self) + rev |
|
1234 | rev = len(self) + rev | |
1235 | if rev < 0 or rev >= len(self): |
|
1235 | if rev < 0 or rev >= len(self): | |
1236 | raise ValueError |
|
1236 | raise ValueError | |
1237 | return self.node(rev) |
|
1237 | return self.node(rev) | |
1238 | except (ValueError, OverflowError): |
|
1238 | except (ValueError, OverflowError): | |
1239 | pass |
|
1239 | pass | |
1240 | if len(id) == 40: |
|
1240 | if len(id) == 40: | |
1241 | try: |
|
1241 | try: | |
1242 | # a full hex nodeid? |
|
1242 | # a full hex nodeid? | |
1243 | node = bin(id) |
|
1243 | node = bin(id) | |
1244 | self.rev(node) |
|
1244 | self.rev(node) | |
1245 | return node |
|
1245 | return node | |
1246 | except (TypeError, error.LookupError): |
|
1246 | except (TypeError, error.LookupError): | |
1247 | pass |
|
1247 | pass | |
1248 |
|
1248 | |||
1249 | def _partialmatch(self, id): |
|
1249 | def _partialmatch(self, id): | |
1250 | # we don't care wdirfilenodeids as they should be always full hash |
|
1250 | # we don't care wdirfilenodeids as they should be always full hash | |
1251 | maybewdir = wdirhex.startswith(id) |
|
1251 | maybewdir = wdirhex.startswith(id) | |
1252 | try: |
|
1252 | try: | |
1253 | partial = self.index.partialmatch(id) |
|
1253 | partial = self.index.partialmatch(id) | |
1254 | if partial and self.hasnode(partial): |
|
1254 | if partial and self.hasnode(partial): | |
1255 | if maybewdir: |
|
1255 | if maybewdir: | |
1256 | # single 'ff...' match in radix tree, ambiguous with wdir |
|
1256 | # single 'ff...' match in radix tree, ambiguous with wdir | |
1257 | raise error.RevlogError |
|
1257 | raise error.RevlogError | |
1258 | return partial |
|
1258 | return partial | |
1259 | if maybewdir: |
|
1259 | if maybewdir: | |
1260 | # no 'ff...' match in radix tree, wdir identified |
|
1260 | # no 'ff...' match in radix tree, wdir identified | |
1261 | raise error.WdirUnsupported |
|
1261 | raise error.WdirUnsupported | |
1262 | return None |
|
1262 | return None | |
1263 | except error.RevlogError: |
|
1263 | except error.RevlogError: | |
1264 | # parsers.c radix tree lookup gave multiple matches |
|
1264 | # parsers.c radix tree lookup gave multiple matches | |
1265 | # fast path: for unfiltered changelog, radix tree is accurate |
|
1265 | # fast path: for unfiltered changelog, radix tree is accurate | |
1266 | if not getattr(self, 'filteredrevs', None): |
|
1266 | if not getattr(self, 'filteredrevs', None): | |
1267 | raise error.AmbiguousPrefixLookupError( |
|
1267 | raise error.AmbiguousPrefixLookupError( | |
1268 | id, self.indexfile, _('ambiguous identifier')) |
|
1268 | id, self.indexfile, _('ambiguous identifier')) | |
1269 | # fall through to slow path that filters hidden revisions |
|
1269 | # fall through to slow path that filters hidden revisions | |
1270 | except (AttributeError, ValueError): |
|
1270 | except (AttributeError, ValueError): | |
1271 | # we are pure python, or key was too short to search radix tree |
|
1271 | # we are pure python, or key was too short to search radix tree | |
1272 | pass |
|
1272 | pass | |
1273 |
|
1273 | |||
1274 | if id in self._pcache: |
|
1274 | if id in self._pcache: | |
1275 | return self._pcache[id] |
|
1275 | return self._pcache[id] | |
1276 |
|
1276 | |||
1277 | if len(id) <= 40: |
|
1277 | if len(id) <= 40: | |
1278 | try: |
|
1278 | try: | |
1279 | # hex(node)[:...] |
|
1279 | # hex(node)[:...] | |
1280 | l = len(id) // 2 # grab an even number of digits |
|
1280 | l = len(id) // 2 # grab an even number of digits | |
1281 | prefix = bin(id[:l * 2]) |
|
1281 | prefix = bin(id[:l * 2]) | |
1282 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] |
|
1282 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] | |
1283 | nl = [n for n in nl if hex(n).startswith(id) and |
|
1283 | nl = [n for n in nl if hex(n).startswith(id) and | |
1284 | self.hasnode(n)] |
|
1284 | self.hasnode(n)] | |
1285 | if nullhex.startswith(id): |
|
1285 | if nullhex.startswith(id): | |
1286 | nl.append(nullid) |
|
1286 | nl.append(nullid) | |
1287 | if len(nl) > 0: |
|
1287 | if len(nl) > 0: | |
1288 | if len(nl) == 1 and not maybewdir: |
|
1288 | if len(nl) == 1 and not maybewdir: | |
1289 | self._pcache[id] = nl[0] |
|
1289 | self._pcache[id] = nl[0] | |
1290 | return nl[0] |
|
1290 | return nl[0] | |
1291 | raise error.AmbiguousPrefixLookupError( |
|
1291 | raise error.AmbiguousPrefixLookupError( | |
1292 | id, self.indexfile, _('ambiguous identifier')) |
|
1292 | id, self.indexfile, _('ambiguous identifier')) | |
1293 | if maybewdir: |
|
1293 | if maybewdir: | |
1294 | raise error.WdirUnsupported |
|
1294 | raise error.WdirUnsupported | |
1295 | return None |
|
1295 | return None | |
1296 | except TypeError: |
|
1296 | except TypeError: | |
1297 | pass |
|
1297 | pass | |
1298 |
|
1298 | |||
1299 | def lookup(self, id): |
|
1299 | def lookup(self, id): | |
1300 | """locate a node based on: |
|
1300 | """locate a node based on: | |
1301 | - revision number or str(revision number) |
|
1301 | - revision number or str(revision number) | |
1302 | - nodeid or subset of hex nodeid |
|
1302 | - nodeid or subset of hex nodeid | |
1303 | """ |
|
1303 | """ | |
1304 | n = self._match(id) |
|
1304 | n = self._match(id) | |
1305 | if n is not None: |
|
1305 | if n is not None: | |
1306 | return n |
|
1306 | return n | |
1307 | n = self._partialmatch(id) |
|
1307 | n = self._partialmatch(id) | |
1308 | if n: |
|
1308 | if n: | |
1309 | return n |
|
1309 | return n | |
1310 |
|
1310 | |||
1311 | raise error.LookupError(id, self.indexfile, _('no match found')) |
|
1311 | raise error.LookupError(id, self.indexfile, _('no match found')) | |
1312 |
|
1312 | |||
1313 | def shortest(self, node, minlength=1): |
|
1313 | def shortest(self, node, minlength=1): | |
1314 | """Find the shortest unambiguous prefix that matches node.""" |
|
1314 | """Find the shortest unambiguous prefix that matches node.""" | |
1315 | def isvalid(prefix): |
|
1315 | def isvalid(prefix): | |
1316 | try: |
|
1316 | try: | |
1317 | matchednode = self._partialmatch(prefix) |
|
1317 | matchednode = self._partialmatch(prefix) | |
1318 | except error.AmbiguousPrefixLookupError: |
|
1318 | except error.AmbiguousPrefixLookupError: | |
1319 | return False |
|
1319 | return False | |
1320 | except error.WdirUnsupported: |
|
1320 | except error.WdirUnsupported: | |
1321 | # single 'ff...' match |
|
1321 | # single 'ff...' match | |
1322 | return True |
|
1322 | return True | |
1323 | if matchednode is None: |
|
1323 | if matchednode is None: | |
1324 | raise error.LookupError(node, self.indexfile, _('no node')) |
|
1324 | raise error.LookupError(node, self.indexfile, _('no node')) | |
1325 | return True |
|
1325 | return True | |
1326 |
|
1326 | |||
1327 | def maybewdir(prefix): |
|
1327 | def maybewdir(prefix): | |
1328 | return all(c == 'f' for c in pycompat.iterbytestr(prefix)) |
|
1328 | return all(c == 'f' for c in pycompat.iterbytestr(prefix)) | |
1329 |
|
1329 | |||
1330 | hexnode = hex(node) |
|
1330 | hexnode = hex(node) | |
1331 |
|
1331 | |||
1332 | def disambiguate(hexnode, minlength): |
|
1332 | def disambiguate(hexnode, minlength): | |
1333 | """Disambiguate against wdirid.""" |
|
1333 | """Disambiguate against wdirid.""" | |
1334 | for length in range(minlength, 41): |
|
1334 | for length in range(minlength, 41): | |
1335 | prefix = hexnode[:length] |
|
1335 | prefix = hexnode[:length] | |
1336 | if not maybewdir(prefix): |
|
1336 | if not maybewdir(prefix): | |
1337 | return prefix |
|
1337 | return prefix | |
1338 |
|
1338 | |||
1339 | if not getattr(self, 'filteredrevs', None): |
|
1339 | if not getattr(self, 'filteredrevs', None): | |
1340 | try: |
|
1340 | try: | |
1341 | length = max(self.index.shortest(node), minlength) |
|
1341 | length = max(self.index.shortest(node), minlength) | |
1342 | return disambiguate(hexnode, length) |
|
1342 | return disambiguate(hexnode, length) | |
1343 | except error.RevlogError: |
|
1343 | except error.RevlogError: | |
1344 | if node != wdirid: |
|
1344 | if node != wdirid: | |
1345 | raise error.LookupError(node, self.indexfile, _('no node')) |
|
1345 | raise error.LookupError(node, self.indexfile, _('no node')) | |
1346 | except AttributeError: |
|
1346 | except AttributeError: | |
1347 | # Fall through to pure code |
|
1347 | # Fall through to pure code | |
1348 | pass |
|
1348 | pass | |
1349 |
|
1349 | |||
1350 | if node == wdirid: |
|
1350 | if node == wdirid: | |
1351 | for length in range(minlength, 41): |
|
1351 | for length in range(minlength, 41): | |
1352 | prefix = hexnode[:length] |
|
1352 | prefix = hexnode[:length] | |
1353 | if isvalid(prefix): |
|
1353 | if isvalid(prefix): | |
1354 | return prefix |
|
1354 | return prefix | |
1355 |
|
1355 | |||
1356 | for length in range(minlength, 41): |
|
1356 | for length in range(minlength, 41): | |
1357 | prefix = hexnode[:length] |
|
1357 | prefix = hexnode[:length] | |
1358 | if isvalid(prefix): |
|
1358 | if isvalid(prefix): | |
1359 | return disambiguate(hexnode, length) |
|
1359 | return disambiguate(hexnode, length) | |
1360 |
|
1360 | |||
1361 | def cmp(self, node, text): |
|
1361 | def cmp(self, node, text): | |
1362 | """compare text with a given file revision |
|
1362 | """compare text with a given file revision | |
1363 |
|
1363 | |||
1364 | returns True if text is different than what is stored. |
|
1364 | returns True if text is different than what is stored. | |
1365 | """ |
|
1365 | """ | |
1366 | p1, p2 = self.parents(node) |
|
1366 | p1, p2 = self.parents(node) | |
1367 | return storageutil.hashrevisionsha1(text, p1, p2) != node |
|
1367 | return storageutil.hashrevisionsha1(text, p1, p2) != node | |
1368 |
|
1368 | |||
1369 | def _cachesegment(self, offset, data): |
|
1369 | def _cachesegment(self, offset, data): | |
1370 | """Add a segment to the revlog cache. |
|
1370 | """Add a segment to the revlog cache. | |
1371 |
|
1371 | |||
1372 | Accepts an absolute offset and the data that is at that location. |
|
1372 | Accepts an absolute offset and the data that is at that location. | |
1373 | """ |
|
1373 | """ | |
1374 | o, d = self._chunkcache |
|
1374 | o, d = self._chunkcache | |
1375 | # try to add to existing cache |
|
1375 | # try to add to existing cache | |
1376 | if o + len(d) == offset and len(d) + len(data) < _chunksize: |
|
1376 | if o + len(d) == offset and len(d) + len(data) < _chunksize: | |
1377 | self._chunkcache = o, d + data |
|
1377 | self._chunkcache = o, d + data | |
1378 | else: |
|
1378 | else: | |
1379 | self._chunkcache = offset, data |
|
1379 | self._chunkcache = offset, data | |
1380 |
|
1380 | |||
1381 | def _readsegment(self, offset, length, df=None): |
|
1381 | def _readsegment(self, offset, length, df=None): | |
1382 | """Load a segment of raw data from the revlog. |
|
1382 | """Load a segment of raw data from the revlog. | |
1383 |
|
1383 | |||
1384 | Accepts an absolute offset, length to read, and an optional existing |
|
1384 | Accepts an absolute offset, length to read, and an optional existing | |
1385 | file handle to read from. |
|
1385 | file handle to read from. | |
1386 |
|
1386 | |||
1387 | If an existing file handle is passed, it will be seeked and the |
|
1387 | If an existing file handle is passed, it will be seeked and the | |
1388 | original seek position will NOT be restored. |
|
1388 | original seek position will NOT be restored. | |
1389 |
|
1389 | |||
1390 | Returns a str or buffer of raw byte data. |
|
1390 | Returns a str or buffer of raw byte data. | |
1391 |
|
1391 | |||
1392 | Raises if the requested number of bytes could not be read. |
|
1392 | Raises if the requested number of bytes could not be read. | |
1393 | """ |
|
1393 | """ | |
1394 | # Cache data both forward and backward around the requested |
|
1394 | # Cache data both forward and backward around the requested | |
1395 | # data, in a fixed size window. This helps speed up operations |
|
1395 | # data, in a fixed size window. This helps speed up operations | |
1396 | # involving reading the revlog backwards. |
|
1396 | # involving reading the revlog backwards. | |
1397 | cachesize = self._chunkcachesize |
|
1397 | cachesize = self._chunkcachesize | |
1398 | realoffset = offset & ~(cachesize - 1) |
|
1398 | realoffset = offset & ~(cachesize - 1) | |
1399 | reallength = (((offset + length + cachesize) & ~(cachesize - 1)) |
|
1399 | reallength = (((offset + length + cachesize) & ~(cachesize - 1)) | |
1400 | - realoffset) |
|
1400 | - realoffset) | |
1401 | with self._datareadfp(df) as df: |
|
1401 | with self._datareadfp(df) as df: | |
1402 | df.seek(realoffset) |
|
1402 | df.seek(realoffset) | |
1403 | d = df.read(reallength) |
|
1403 | d = df.read(reallength) | |
1404 |
|
1404 | |||
1405 | self._cachesegment(realoffset, d) |
|
1405 | self._cachesegment(realoffset, d) | |
1406 | if offset != realoffset or reallength != length: |
|
1406 | if offset != realoffset or reallength != length: | |
1407 | startoffset = offset - realoffset |
|
1407 | startoffset = offset - realoffset | |
1408 | if len(d) - startoffset < length: |
|
1408 | if len(d) - startoffset < length: | |
1409 | raise error.RevlogError( |
|
1409 | raise error.RevlogError( | |
1410 | _('partial read of revlog %s; expected %d bytes from ' |
|
1410 | _('partial read of revlog %s; expected %d bytes from ' | |
1411 | 'offset %d, got %d') % |
|
1411 | 'offset %d, got %d') % | |
1412 | (self.indexfile if self._inline else self.datafile, |
|
1412 | (self.indexfile if self._inline else self.datafile, | |
1413 | length, realoffset, len(d) - startoffset)) |
|
1413 | length, realoffset, len(d) - startoffset)) | |
1414 |
|
1414 | |||
1415 | return util.buffer(d, startoffset, length) |
|
1415 | return util.buffer(d, startoffset, length) | |
1416 |
|
1416 | |||
1417 | if len(d) < length: |
|
1417 | if len(d) < length: | |
1418 | raise error.RevlogError( |
|
1418 | raise error.RevlogError( | |
1419 | _('partial read of revlog %s; expected %d bytes from offset ' |
|
1419 | _('partial read of revlog %s; expected %d bytes from offset ' | |
1420 | '%d, got %d') % |
|
1420 | '%d, got %d') % | |
1421 | (self.indexfile if self._inline else self.datafile, |
|
1421 | (self.indexfile if self._inline else self.datafile, | |
1422 | length, offset, len(d))) |
|
1422 | length, offset, len(d))) | |
1423 |
|
1423 | |||
1424 | return d |
|
1424 | return d | |
1425 |
|
1425 | |||
1426 | def _getsegment(self, offset, length, df=None): |
|
1426 | def _getsegment(self, offset, length, df=None): | |
1427 | """Obtain a segment of raw data from the revlog. |
|
1427 | """Obtain a segment of raw data from the revlog. | |
1428 |
|
1428 | |||
1429 | Accepts an absolute offset, length of bytes to obtain, and an |
|
1429 | Accepts an absolute offset, length of bytes to obtain, and an | |
1430 | optional file handle to the already-opened revlog. If the file |
|
1430 | optional file handle to the already-opened revlog. If the file | |
1431 | handle is used, it's original seek position will not be preserved. |
|
1431 | handle is used, it's original seek position will not be preserved. | |
1432 |
|
1432 | |||
1433 | Requests for data may be returned from a cache. |
|
1433 | Requests for data may be returned from a cache. | |
1434 |
|
1434 | |||
1435 | Returns a str or a buffer instance of raw byte data. |
|
1435 | Returns a str or a buffer instance of raw byte data. | |
1436 | """ |
|
1436 | """ | |
1437 | o, d = self._chunkcache |
|
1437 | o, d = self._chunkcache | |
1438 | l = len(d) |
|
1438 | l = len(d) | |
1439 |
|
1439 | |||
1440 | # is it in the cache? |
|
1440 | # is it in the cache? | |
1441 | cachestart = offset - o |
|
1441 | cachestart = offset - o | |
1442 | cacheend = cachestart + length |
|
1442 | cacheend = cachestart + length | |
1443 | if cachestart >= 0 and cacheend <= l: |
|
1443 | if cachestart >= 0 and cacheend <= l: | |
1444 | if cachestart == 0 and cacheend == l: |
|
1444 | if cachestart == 0 and cacheend == l: | |
1445 | return d # avoid a copy |
|
1445 | return d # avoid a copy | |
1446 | return util.buffer(d, cachestart, cacheend - cachestart) |
|
1446 | return util.buffer(d, cachestart, cacheend - cachestart) | |
1447 |
|
1447 | |||
1448 | return self._readsegment(offset, length, df=df) |
|
1448 | return self._readsegment(offset, length, df=df) | |
1449 |
|
1449 | |||
1450 | def _getsegmentforrevs(self, startrev, endrev, df=None): |
|
1450 | def _getsegmentforrevs(self, startrev, endrev, df=None): | |
1451 | """Obtain a segment of raw data corresponding to a range of revisions. |
|
1451 | """Obtain a segment of raw data corresponding to a range of revisions. | |
1452 |
|
1452 | |||
1453 | Accepts the start and end revisions and an optional already-open |
|
1453 | Accepts the start and end revisions and an optional already-open | |
1454 | file handle to be used for reading. If the file handle is read, its |
|
1454 | file handle to be used for reading. If the file handle is read, its | |
1455 | seek position will not be preserved. |
|
1455 | seek position will not be preserved. | |
1456 |
|
1456 | |||
1457 | Requests for data may be satisfied by a cache. |
|
1457 | Requests for data may be satisfied by a cache. | |
1458 |
|
1458 | |||
1459 | Returns a 2-tuple of (offset, data) for the requested range of |
|
1459 | Returns a 2-tuple of (offset, data) for the requested range of | |
1460 | revisions. Offset is the integer offset from the beginning of the |
|
1460 | revisions. Offset is the integer offset from the beginning of the | |
1461 | revlog and data is a str or buffer of the raw byte data. |
|
1461 | revlog and data is a str or buffer of the raw byte data. | |
1462 |
|
1462 | |||
1463 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` |
|
1463 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` | |
1464 | to determine where each revision's data begins and ends. |
|
1464 | to determine where each revision's data begins and ends. | |
1465 | """ |
|
1465 | """ | |
1466 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons |
|
1466 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons | |
1467 | # (functions are expensive). |
|
1467 | # (functions are expensive). | |
1468 | index = self.index |
|
1468 | index = self.index | |
1469 | istart = index[startrev] |
|
1469 | istart = index[startrev] | |
1470 | start = int(istart[0] >> 16) |
|
1470 | start = int(istart[0] >> 16) | |
1471 | if startrev == endrev: |
|
1471 | if startrev == endrev: | |
1472 | end = start + istart[1] |
|
1472 | end = start + istart[1] | |
1473 | else: |
|
1473 | else: | |
1474 | iend = index[endrev] |
|
1474 | iend = index[endrev] | |
1475 | end = int(iend[0] >> 16) + iend[1] |
|
1475 | end = int(iend[0] >> 16) + iend[1] | |
1476 |
|
1476 | |||
1477 | if self._inline: |
|
1477 | if self._inline: | |
1478 | start += (startrev + 1) * self._io.size |
|
1478 | start += (startrev + 1) * self._io.size | |
1479 | end += (endrev + 1) * self._io.size |
|
1479 | end += (endrev + 1) * self._io.size | |
1480 | length = end - start |
|
1480 | length = end - start | |
1481 |
|
1481 | |||
1482 | return start, self._getsegment(start, length, df=df) |
|
1482 | return start, self._getsegment(start, length, df=df) | |
1483 |
|
1483 | |||
1484 | def _chunk(self, rev, df=None): |
|
1484 | def _chunk(self, rev, df=None): | |
1485 | """Obtain a single decompressed chunk for a revision. |
|
1485 | """Obtain a single decompressed chunk for a revision. | |
1486 |
|
1486 | |||
1487 | Accepts an integer revision and an optional already-open file handle |
|
1487 | Accepts an integer revision and an optional already-open file handle | |
1488 | to be used for reading. If used, the seek position of the file will not |
|
1488 | to be used for reading. If used, the seek position of the file will not | |
1489 | be preserved. |
|
1489 | be preserved. | |
1490 |
|
1490 | |||
1491 | Returns a str holding uncompressed data for the requested revision. |
|
1491 | Returns a str holding uncompressed data for the requested revision. | |
1492 | """ |
|
1492 | """ | |
1493 | return self.decompress(self._getsegmentforrevs(rev, rev, df=df)[1]) |
|
1493 | return self.decompress(self._getsegmentforrevs(rev, rev, df=df)[1]) | |
1494 |
|
1494 | |||
1495 | def _chunks(self, revs, df=None, targetsize=None): |
|
1495 | def _chunks(self, revs, df=None, targetsize=None): | |
1496 | """Obtain decompressed chunks for the specified revisions. |
|
1496 | """Obtain decompressed chunks for the specified revisions. | |
1497 |
|
1497 | |||
1498 | Accepts an iterable of numeric revisions that are assumed to be in |
|
1498 | Accepts an iterable of numeric revisions that are assumed to be in | |
1499 | ascending order. Also accepts an optional already-open file handle |
|
1499 | ascending order. Also accepts an optional already-open file handle | |
1500 | to be used for reading. If used, the seek position of the file will |
|
1500 | to be used for reading. If used, the seek position of the file will | |
1501 | not be preserved. |
|
1501 | not be preserved. | |
1502 |
|
1502 | |||
1503 | This function is similar to calling ``self._chunk()`` multiple times, |
|
1503 | This function is similar to calling ``self._chunk()`` multiple times, | |
1504 | but is faster. |
|
1504 | but is faster. | |
1505 |
|
1505 | |||
1506 | Returns a list with decompressed data for each requested revision. |
|
1506 | Returns a list with decompressed data for each requested revision. | |
1507 | """ |
|
1507 | """ | |
1508 | if not revs: |
|
1508 | if not revs: | |
1509 | return [] |
|
1509 | return [] | |
1510 | start = self.start |
|
1510 | start = self.start | |
1511 | length = self.length |
|
1511 | length = self.length | |
1512 | inline = self._inline |
|
1512 | inline = self._inline | |
1513 | iosize = self._io.size |
|
1513 | iosize = self._io.size | |
1514 | buffer = util.buffer |
|
1514 | buffer = util.buffer | |
1515 |
|
1515 | |||
1516 | l = [] |
|
1516 | l = [] | |
1517 | ladd = l.append |
|
1517 | ladd = l.append | |
1518 |
|
1518 | |||
1519 | if not self._withsparseread: |
|
1519 | if not self._withsparseread: | |
1520 | slicedchunks = (revs,) |
|
1520 | slicedchunks = (revs,) | |
1521 | else: |
|
1521 | else: | |
1522 | slicedchunks = deltautil.slicechunk(self, revs, |
|
1522 | slicedchunks = deltautil.slicechunk(self, revs, | |
1523 | targetsize=targetsize) |
|
1523 | targetsize=targetsize) | |
1524 |
|
1524 | |||
1525 | for revschunk in slicedchunks: |
|
1525 | for revschunk in slicedchunks: | |
1526 | firstrev = revschunk[0] |
|
1526 | firstrev = revschunk[0] | |
1527 | # Skip trailing revisions with empty diff |
|
1527 | # Skip trailing revisions with empty diff | |
1528 | for lastrev in revschunk[::-1]: |
|
1528 | for lastrev in revschunk[::-1]: | |
1529 | if length(lastrev) != 0: |
|
1529 | if length(lastrev) != 0: | |
1530 | break |
|
1530 | break | |
1531 |
|
1531 | |||
1532 | try: |
|
1532 | try: | |
1533 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) |
|
1533 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) | |
1534 | except OverflowError: |
|
1534 | except OverflowError: | |
1535 | # issue4215 - we can't cache a run of chunks greater than |
|
1535 | # issue4215 - we can't cache a run of chunks greater than | |
1536 | # 2G on Windows |
|
1536 | # 2G on Windows | |
1537 | return [self._chunk(rev, df=df) for rev in revschunk] |
|
1537 | return [self._chunk(rev, df=df) for rev in revschunk] | |
1538 |
|
1538 | |||
1539 | decomp = self.decompress |
|
1539 | decomp = self.decompress | |
1540 | for rev in revschunk: |
|
1540 | for rev in revschunk: | |
1541 | chunkstart = start(rev) |
|
1541 | chunkstart = start(rev) | |
1542 | if inline: |
|
1542 | if inline: | |
1543 | chunkstart += (rev + 1) * iosize |
|
1543 | chunkstart += (rev + 1) * iosize | |
1544 | chunklength = length(rev) |
|
1544 | chunklength = length(rev) | |
1545 | ladd(decomp(buffer(data, chunkstart - offset, chunklength))) |
|
1545 | ladd(decomp(buffer(data, chunkstart - offset, chunklength))) | |
1546 |
|
1546 | |||
1547 | return l |
|
1547 | return l | |
1548 |
|
1548 | |||
1549 | def _chunkclear(self): |
|
1549 | def _chunkclear(self): | |
1550 | """Clear the raw chunk cache.""" |
|
1550 | """Clear the raw chunk cache.""" | |
1551 | self._chunkcache = (0, '') |
|
1551 | self._chunkcache = (0, '') | |
1552 |
|
1552 | |||
1553 | def deltaparent(self, rev): |
|
1553 | def deltaparent(self, rev): | |
1554 | """return deltaparent of the given revision""" |
|
1554 | """return deltaparent of the given revision""" | |
1555 | base = self.index[rev][3] |
|
1555 | base = self.index[rev][3] | |
1556 | if base == rev: |
|
1556 | if base == rev: | |
1557 | return nullrev |
|
1557 | return nullrev | |
1558 | elif self._generaldelta: |
|
1558 | elif self._generaldelta: | |
1559 | return base |
|
1559 | return base | |
1560 | else: |
|
1560 | else: | |
1561 | return rev - 1 |
|
1561 | return rev - 1 | |
1562 |
|
1562 | |||
1563 | def issnapshot(self, rev): |
|
1563 | def issnapshot(self, rev): | |
1564 | """tells whether rev is a snapshot |
|
1564 | """tells whether rev is a snapshot | |
1565 | """ |
|
1565 | """ | |
1566 | if not self._sparserevlog: |
|
1566 | if not self._sparserevlog: | |
1567 | return self.deltaparent(rev) == nullrev |
|
1567 | return self.deltaparent(rev) == nullrev | |
1568 | elif util.safehasattr(self.index, 'issnapshot'): |
|
1568 | elif util.safehasattr(self.index, 'issnapshot'): | |
1569 | # directly assign the method to cache the testing and access |
|
1569 | # directly assign the method to cache the testing and access | |
1570 | self.issnapshot = self.index.issnapshot |
|
1570 | self.issnapshot = self.index.issnapshot | |
1571 | return self.issnapshot(rev) |
|
1571 | return self.issnapshot(rev) | |
1572 | if rev == nullrev: |
|
1572 | if rev == nullrev: | |
1573 | return True |
|
1573 | return True | |
1574 | entry = self.index[rev] |
|
1574 | entry = self.index[rev] | |
1575 | base = entry[3] |
|
1575 | base = entry[3] | |
1576 | if base == rev: |
|
1576 | if base == rev: | |
1577 | return True |
|
1577 | return True | |
1578 | if base == nullrev: |
|
1578 | if base == nullrev: | |
1579 | return True |
|
1579 | return True | |
1580 | p1 = entry[5] |
|
1580 | p1 = entry[5] | |
1581 | p2 = entry[6] |
|
1581 | p2 = entry[6] | |
1582 | if base == p1 or base == p2: |
|
1582 | if base == p1 or base == p2: | |
1583 | return False |
|
1583 | return False | |
1584 | return self.issnapshot(base) |
|
1584 | return self.issnapshot(base) | |
1585 |
|
1585 | |||
1586 | def snapshotdepth(self, rev): |
|
1586 | def snapshotdepth(self, rev): | |
1587 | """number of snapshot in the chain before this one""" |
|
1587 | """number of snapshot in the chain before this one""" | |
1588 | if not self.issnapshot(rev): |
|
1588 | if not self.issnapshot(rev): | |
1589 | raise error.ProgrammingError('revision %d not a snapshot') |
|
1589 | raise error.ProgrammingError('revision %d not a snapshot') | |
1590 | return len(self._deltachain(rev)[0]) - 1 |
|
1590 | return len(self._deltachain(rev)[0]) - 1 | |
1591 |
|
1591 | |||
1592 | def revdiff(self, rev1, rev2): |
|
1592 | def revdiff(self, rev1, rev2): | |
1593 | """return or calculate a delta between two revisions |
|
1593 | """return or calculate a delta between two revisions | |
1594 |
|
1594 | |||
1595 | The delta calculated is in binary form and is intended to be written to |
|
1595 | The delta calculated is in binary form and is intended to be written to | |
1596 | revlog data directly. So this function needs raw revision data. |
|
1596 | revlog data directly. So this function needs raw revision data. | |
1597 | """ |
|
1597 | """ | |
1598 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: |
|
1598 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: | |
1599 | return bytes(self._chunk(rev2)) |
|
1599 | return bytes(self._chunk(rev2)) | |
1600 |
|
1600 | |||
1601 | return mdiff.textdiff(self.rawdata(rev1), |
|
1601 | return mdiff.textdiff(self.rawdata(rev1), | |
1602 | self.rawdata(rev2)) |
|
1602 | self.rawdata(rev2)) | |
1603 |
|
1603 | |||
1604 | def revision(self, nodeorrev, _df=None, raw=False): |
|
1604 | def revision(self, nodeorrev, _df=None, raw=False): | |
1605 | """return an uncompressed revision of a given node or revision |
|
1605 | """return an uncompressed revision of a given node or revision | |
1606 | number. |
|
1606 | number. | |
1607 |
|
1607 | |||
1608 | _df - an existing file handle to read from. (internal-only) |
|
1608 | _df - an existing file handle to read from. (internal-only) | |
1609 | raw - an optional argument specifying if the revision data is to be |
|
1609 | raw - an optional argument specifying if the revision data is to be | |
1610 | treated as raw data when applying flag transforms. 'raw' should be set |
|
1610 | treated as raw data when applying flag transforms. 'raw' should be set | |
1611 | to True when generating changegroups or in debug commands. |
|
1611 | to True when generating changegroups or in debug commands. | |
1612 | """ |
|
1612 | """ | |
1613 | if raw: |
|
1613 | if raw: | |
1614 | msg = ('revlog.revision(..., raw=True) is deprecated, ' |
|
1614 | msg = ('revlog.revision(..., raw=True) is deprecated, ' | |
1615 | 'use revlog.rawdata(...)') |
|
1615 | 'use revlog.rawdata(...)') | |
1616 | util.nouideprecwarn(msg, '5.2', stacklevel=2) |
|
1616 | util.nouideprecwarn(msg, '5.2', stacklevel=2) | |
1617 | return self._revisiondata(nodeorrev, _df, raw=raw) |
|
1617 | return self._revisiondata(nodeorrev, _df, raw=raw) | |
1618 |
|
1618 | |||
1619 | def _revisiondata(self, nodeorrev, _df=None, raw=False): |
|
1619 | def _revisiondata(self, nodeorrev, _df=None, raw=False): | |
1620 | # deal with <nodeorrev> argument type |
|
1620 | # deal with <nodeorrev> argument type | |
1621 | if isinstance(nodeorrev, int): |
|
1621 | if isinstance(nodeorrev, int): | |
1622 | rev = nodeorrev |
|
1622 | rev = nodeorrev | |
1623 | node = self.node(rev) |
|
1623 | node = self.node(rev) | |
1624 | else: |
|
1624 | else: | |
1625 | node = nodeorrev |
|
1625 | node = nodeorrev | |
1626 | rev = None |
|
1626 | rev = None | |
1627 |
|
1627 | |||
1628 | # fast path the special `nullid` rev |
|
1628 | # fast path the special `nullid` rev | |
1629 | if node == nullid: |
|
1629 | if node == nullid: | |
1630 | return "" |
|
1630 | return "" | |
1631 |
|
1631 | |||
1632 | # The text as stored inside the revlog. Might be the revision or might |
|
1632 | # The text as stored inside the revlog. Might be the revision or might | |
1633 | # need to be processed to retrieve the revision. |
|
1633 | # need to be processed to retrieve the revision. | |
1634 | rawtext = None |
|
1634 | rawtext = None | |
1635 |
|
1635 | |||
1636 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) |
|
1636 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) | |
1637 |
|
1637 | |||
1638 | if raw and validated: |
|
1638 | if raw and validated: | |
1639 | # if we don't want to process the raw text and that raw |
|
1639 | # if we don't want to process the raw text and that raw | |
1640 | # text is cached, we can exit early. |
|
1640 | # text is cached, we can exit early. | |
1641 | return rawtext |
|
1641 | return rawtext | |
1642 | if rev is None: |
|
1642 | if rev is None: | |
1643 | rev = self.rev(node) |
|
1643 | rev = self.rev(node) | |
1644 | # the revlog's flag for this revision |
|
1644 | # the revlog's flag for this revision | |
1645 | # (usually alter its state or content) |
|
1645 | # (usually alter its state or content) | |
1646 | flags = self.flags(rev) |
|
1646 | flags = self.flags(rev) | |
1647 |
|
1647 | |||
1648 | if validated and flags == REVIDX_DEFAULT_FLAGS: |
|
1648 | if validated and flags == REVIDX_DEFAULT_FLAGS: | |
1649 | # no extra flags set, no flag processor runs, text = rawtext |
|
1649 | # no extra flags set, no flag processor runs, text = rawtext | |
1650 | return rawtext |
|
1650 | return rawtext | |
1651 |
|
1651 | |||
1652 | text, validatehash = self._processflags(rawtext, flags, 'read', raw=raw) |
|
1652 | if raw: | |
|
1653 | validatehash = self._processflagsraw(rawtext, flags) | |||
|
1654 | text = rawtext | |||
|
1655 | else: | |||
|
1656 | text, validatehash = self._processflagsread(rawtext, flags) | |||
1653 | if validatehash: |
|
1657 | if validatehash: | |
1654 | self.checkhash(text, node, rev=rev) |
|
1658 | self.checkhash(text, node, rev=rev) | |
1655 | if not validated: |
|
1659 | if not validated: | |
1656 | self._revisioncache = (node, rev, rawtext) |
|
1660 | self._revisioncache = (node, rev, rawtext) | |
1657 |
|
1661 | |||
1658 | return text |
|
1662 | return text | |
1659 |
|
1663 | |||
1660 | def _rawtext(self, node, rev, _df=None): |
|
1664 | def _rawtext(self, node, rev, _df=None): | |
1661 | """return the possibly unvalidated rawtext for a revision |
|
1665 | """return the possibly unvalidated rawtext for a revision | |
1662 |
|
1666 | |||
1663 | returns (rev, rawtext, validated) |
|
1667 | returns (rev, rawtext, validated) | |
1664 | """ |
|
1668 | """ | |
1665 |
|
1669 | |||
1666 | # revision in the cache (could be useful to apply delta) |
|
1670 | # revision in the cache (could be useful to apply delta) | |
1667 | cachedrev = None |
|
1671 | cachedrev = None | |
1668 | # An intermediate text to apply deltas to |
|
1672 | # An intermediate text to apply deltas to | |
1669 | basetext = None |
|
1673 | basetext = None | |
1670 |
|
1674 | |||
1671 | # Check if we have the entry in cache |
|
1675 | # Check if we have the entry in cache | |
1672 | # The cache entry looks like (node, rev, rawtext) |
|
1676 | # The cache entry looks like (node, rev, rawtext) | |
1673 | if self._revisioncache: |
|
1677 | if self._revisioncache: | |
1674 | if self._revisioncache[0] == node: |
|
1678 | if self._revisioncache[0] == node: | |
1675 | return (rev, self._revisioncache[2], True) |
|
1679 | return (rev, self._revisioncache[2], True) | |
1676 | cachedrev = self._revisioncache[1] |
|
1680 | cachedrev = self._revisioncache[1] | |
1677 |
|
1681 | |||
1678 | if rev is None: |
|
1682 | if rev is None: | |
1679 | rev = self.rev(node) |
|
1683 | rev = self.rev(node) | |
1680 |
|
1684 | |||
1681 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) |
|
1685 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) | |
1682 | if stopped: |
|
1686 | if stopped: | |
1683 | basetext = self._revisioncache[2] |
|
1687 | basetext = self._revisioncache[2] | |
1684 |
|
1688 | |||
1685 | # drop cache to save memory, the caller is expected to |
|
1689 | # drop cache to save memory, the caller is expected to | |
1686 | # update self._revisioncache after validating the text |
|
1690 | # update self._revisioncache after validating the text | |
1687 | self._revisioncache = None |
|
1691 | self._revisioncache = None | |
1688 |
|
1692 | |||
1689 | targetsize = None |
|
1693 | targetsize = None | |
1690 | rawsize = self.index[rev][2] |
|
1694 | rawsize = self.index[rev][2] | |
1691 | if 0 <= rawsize: |
|
1695 | if 0 <= rawsize: | |
1692 | targetsize = 4 * rawsize |
|
1696 | targetsize = 4 * rawsize | |
1693 |
|
1697 | |||
1694 | bins = self._chunks(chain, df=_df, targetsize=targetsize) |
|
1698 | bins = self._chunks(chain, df=_df, targetsize=targetsize) | |
1695 | if basetext is None: |
|
1699 | if basetext is None: | |
1696 | basetext = bytes(bins[0]) |
|
1700 | basetext = bytes(bins[0]) | |
1697 | bins = bins[1:] |
|
1701 | bins = bins[1:] | |
1698 |
|
1702 | |||
1699 | rawtext = mdiff.patches(basetext, bins) |
|
1703 | rawtext = mdiff.patches(basetext, bins) | |
1700 | del basetext # let us have a chance to free memory early |
|
1704 | del basetext # let us have a chance to free memory early | |
1701 | return (rev, rawtext, False) |
|
1705 | return (rev, rawtext, False) | |
1702 |
|
1706 | |||
1703 | def rawdata(self, nodeorrev, _df=None): |
|
1707 | def rawdata(self, nodeorrev, _df=None): | |
1704 | """return an uncompressed raw data of a given node or revision number. |
|
1708 | """return an uncompressed raw data of a given node or revision number. | |
1705 |
|
1709 | |||
1706 | _df - an existing file handle to read from. (internal-only) |
|
1710 | _df - an existing file handle to read from. (internal-only) | |
1707 | """ |
|
1711 | """ | |
1708 | return self._revisiondata(nodeorrev, _df, raw=True) |
|
1712 | return self._revisiondata(nodeorrev, _df, raw=True) | |
1709 |
|
1713 | |||
1710 | def hash(self, text, p1, p2): |
|
1714 | def hash(self, text, p1, p2): | |
1711 | """Compute a node hash. |
|
1715 | """Compute a node hash. | |
1712 |
|
1716 | |||
1713 | Available as a function so that subclasses can replace the hash |
|
1717 | Available as a function so that subclasses can replace the hash | |
1714 | as needed. |
|
1718 | as needed. | |
1715 | """ |
|
1719 | """ | |
1716 | return storageutil.hashrevisionsha1(text, p1, p2) |
|
1720 | return storageutil.hashrevisionsha1(text, p1, p2) | |
1717 |
|
1721 | |||
1718 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
1722 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
1719 | """Check node hash integrity. |
|
1723 | """Check node hash integrity. | |
1720 |
|
1724 | |||
1721 | Available as a function so that subclasses can extend hash mismatch |
|
1725 | Available as a function so that subclasses can extend hash mismatch | |
1722 | behaviors as needed. |
|
1726 | behaviors as needed. | |
1723 | """ |
|
1727 | """ | |
1724 | try: |
|
1728 | try: | |
1725 | if p1 is None and p2 is None: |
|
1729 | if p1 is None and p2 is None: | |
1726 | p1, p2 = self.parents(node) |
|
1730 | p1, p2 = self.parents(node) | |
1727 | if node != self.hash(text, p1, p2): |
|
1731 | if node != self.hash(text, p1, p2): | |
1728 | # Clear the revision cache on hash failure. The revision cache |
|
1732 | # Clear the revision cache on hash failure. The revision cache | |
1729 | # only stores the raw revision and clearing the cache does have |
|
1733 | # only stores the raw revision and clearing the cache does have | |
1730 | # the side-effect that we won't have a cache hit when the raw |
|
1734 | # the side-effect that we won't have a cache hit when the raw | |
1731 | # revision data is accessed. But this case should be rare and |
|
1735 | # revision data is accessed. But this case should be rare and | |
1732 | # it is extra work to teach the cache about the hash |
|
1736 | # it is extra work to teach the cache about the hash | |
1733 | # verification state. |
|
1737 | # verification state. | |
1734 | if self._revisioncache and self._revisioncache[0] == node: |
|
1738 | if self._revisioncache and self._revisioncache[0] == node: | |
1735 | self._revisioncache = None |
|
1739 | self._revisioncache = None | |
1736 |
|
1740 | |||
1737 | revornode = rev |
|
1741 | revornode = rev | |
1738 | if revornode is None: |
|
1742 | if revornode is None: | |
1739 | revornode = templatefilters.short(hex(node)) |
|
1743 | revornode = templatefilters.short(hex(node)) | |
1740 | raise error.RevlogError(_("integrity check failed on %s:%s") |
|
1744 | raise error.RevlogError(_("integrity check failed on %s:%s") | |
1741 | % (self.indexfile, pycompat.bytestr(revornode))) |
|
1745 | % (self.indexfile, pycompat.bytestr(revornode))) | |
1742 | except error.RevlogError: |
|
1746 | except error.RevlogError: | |
1743 | if self._censorable and storageutil.iscensoredtext(text): |
|
1747 | if self._censorable and storageutil.iscensoredtext(text): | |
1744 | raise error.CensoredNodeError(self.indexfile, node, text) |
|
1748 | raise error.CensoredNodeError(self.indexfile, node, text) | |
1745 | raise |
|
1749 | raise | |
1746 |
|
1750 | |||
1747 | def _enforceinlinesize(self, tr, fp=None): |
|
1751 | def _enforceinlinesize(self, tr, fp=None): | |
1748 | """Check if the revlog is too big for inline and convert if so. |
|
1752 | """Check if the revlog is too big for inline and convert if so. | |
1749 |
|
1753 | |||
1750 | This should be called after revisions are added to the revlog. If the |
|
1754 | This should be called after revisions are added to the revlog. If the | |
1751 | revlog has grown too large to be an inline revlog, it will convert it |
|
1755 | revlog has grown too large to be an inline revlog, it will convert it | |
1752 | to use multiple index and data files. |
|
1756 | to use multiple index and data files. | |
1753 | """ |
|
1757 | """ | |
1754 | tiprev = len(self) - 1 |
|
1758 | tiprev = len(self) - 1 | |
1755 | if (not self._inline or |
|
1759 | if (not self._inline or | |
1756 | (self.start(tiprev) + self.length(tiprev)) < _maxinline): |
|
1760 | (self.start(tiprev) + self.length(tiprev)) < _maxinline): | |
1757 | return |
|
1761 | return | |
1758 |
|
1762 | |||
1759 | trinfo = tr.find(self.indexfile) |
|
1763 | trinfo = tr.find(self.indexfile) | |
1760 | if trinfo is None: |
|
1764 | if trinfo is None: | |
1761 | raise error.RevlogError(_("%s not found in the transaction") |
|
1765 | raise error.RevlogError(_("%s not found in the transaction") | |
1762 | % self.indexfile) |
|
1766 | % self.indexfile) | |
1763 |
|
1767 | |||
1764 | trindex = trinfo[2] |
|
1768 | trindex = trinfo[2] | |
1765 | if trindex is not None: |
|
1769 | if trindex is not None: | |
1766 | dataoff = self.start(trindex) |
|
1770 | dataoff = self.start(trindex) | |
1767 | else: |
|
1771 | else: | |
1768 | # revlog was stripped at start of transaction, use all leftover data |
|
1772 | # revlog was stripped at start of transaction, use all leftover data | |
1769 | trindex = len(self) - 1 |
|
1773 | trindex = len(self) - 1 | |
1770 | dataoff = self.end(tiprev) |
|
1774 | dataoff = self.end(tiprev) | |
1771 |
|
1775 | |||
1772 | tr.add(self.datafile, dataoff) |
|
1776 | tr.add(self.datafile, dataoff) | |
1773 |
|
1777 | |||
1774 | if fp: |
|
1778 | if fp: | |
1775 | fp.flush() |
|
1779 | fp.flush() | |
1776 | fp.close() |
|
1780 | fp.close() | |
1777 | # We can't use the cached file handle after close(). So prevent |
|
1781 | # We can't use the cached file handle after close(). So prevent | |
1778 | # its usage. |
|
1782 | # its usage. | |
1779 | self._writinghandles = None |
|
1783 | self._writinghandles = None | |
1780 |
|
1784 | |||
1781 | with self._indexfp('r') as ifh, self._datafp('w') as dfh: |
|
1785 | with self._indexfp('r') as ifh, self._datafp('w') as dfh: | |
1782 | for r in self: |
|
1786 | for r in self: | |
1783 | dfh.write(self._getsegmentforrevs(r, r, df=ifh)[1]) |
|
1787 | dfh.write(self._getsegmentforrevs(r, r, df=ifh)[1]) | |
1784 |
|
1788 | |||
1785 | with self._indexfp('w') as fp: |
|
1789 | with self._indexfp('w') as fp: | |
1786 | self.version &= ~FLAG_INLINE_DATA |
|
1790 | self.version &= ~FLAG_INLINE_DATA | |
1787 | self._inline = False |
|
1791 | self._inline = False | |
1788 | io = self._io |
|
1792 | io = self._io | |
1789 | for i in self: |
|
1793 | for i in self: | |
1790 | e = io.packentry(self.index[i], self.node, self.version, i) |
|
1794 | e = io.packentry(self.index[i], self.node, self.version, i) | |
1791 | fp.write(e) |
|
1795 | fp.write(e) | |
1792 |
|
1796 | |||
1793 | # the temp file replace the real index when we exit the context |
|
1797 | # the temp file replace the real index when we exit the context | |
1794 | # manager |
|
1798 | # manager | |
1795 |
|
1799 | |||
1796 | tr.replace(self.indexfile, trindex * self._io.size) |
|
1800 | tr.replace(self.indexfile, trindex * self._io.size) | |
1797 | self._chunkclear() |
|
1801 | self._chunkclear() | |
1798 |
|
1802 | |||
1799 | def _nodeduplicatecallback(self, transaction, node): |
|
1803 | def _nodeduplicatecallback(self, transaction, node): | |
1800 | """called when trying to add a node already stored. |
|
1804 | """called when trying to add a node already stored. | |
1801 | """ |
|
1805 | """ | |
1802 |
|
1806 | |||
1803 | def addrevision(self, text, transaction, link, p1, p2, cachedelta=None, |
|
1807 | def addrevision(self, text, transaction, link, p1, p2, cachedelta=None, | |
1804 | node=None, flags=REVIDX_DEFAULT_FLAGS, deltacomputer=None): |
|
1808 | node=None, flags=REVIDX_DEFAULT_FLAGS, deltacomputer=None): | |
1805 | """add a revision to the log |
|
1809 | """add a revision to the log | |
1806 |
|
1810 | |||
1807 | text - the revision data to add |
|
1811 | text - the revision data to add | |
1808 | transaction - the transaction object used for rollback |
|
1812 | transaction - the transaction object used for rollback | |
1809 | link - the linkrev data to add |
|
1813 | link - the linkrev data to add | |
1810 | p1, p2 - the parent nodeids of the revision |
|
1814 | p1, p2 - the parent nodeids of the revision | |
1811 | cachedelta - an optional precomputed delta |
|
1815 | cachedelta - an optional precomputed delta | |
1812 | node - nodeid of revision; typically node is not specified, and it is |
|
1816 | node - nodeid of revision; typically node is not specified, and it is | |
1813 | computed by default as hash(text, p1, p2), however subclasses might |
|
1817 | computed by default as hash(text, p1, p2), however subclasses might | |
1814 | use different hashing method (and override checkhash() in such case) |
|
1818 | use different hashing method (and override checkhash() in such case) | |
1815 | flags - the known flags to set on the revision |
|
1819 | flags - the known flags to set on the revision | |
1816 | deltacomputer - an optional deltacomputer instance shared between |
|
1820 | deltacomputer - an optional deltacomputer instance shared between | |
1817 | multiple calls |
|
1821 | multiple calls | |
1818 | """ |
|
1822 | """ | |
1819 | if link == nullrev: |
|
1823 | if link == nullrev: | |
1820 | raise error.RevlogError(_("attempted to add linkrev -1 to %s") |
|
1824 | raise error.RevlogError(_("attempted to add linkrev -1 to %s") | |
1821 | % self.indexfile) |
|
1825 | % self.indexfile) | |
1822 |
|
1826 | |||
1823 | if flags: |
|
1827 | if flags: | |
1824 | node = node or self.hash(text, p1, p2) |
|
1828 | node = node or self.hash(text, p1, p2) | |
1825 |
|
1829 | |||
1826 | rawtext, validatehash = self._processflagswrite(text, flags) |
|
1830 | rawtext, validatehash = self._processflagswrite(text, flags) | |
1827 |
|
1831 | |||
1828 | # If the flag processor modifies the revision data, ignore any provided |
|
1832 | # If the flag processor modifies the revision data, ignore any provided | |
1829 | # cachedelta. |
|
1833 | # cachedelta. | |
1830 | if rawtext != text: |
|
1834 | if rawtext != text: | |
1831 | cachedelta = None |
|
1835 | cachedelta = None | |
1832 |
|
1836 | |||
1833 | if len(rawtext) > _maxentrysize: |
|
1837 | if len(rawtext) > _maxentrysize: | |
1834 | raise error.RevlogError( |
|
1838 | raise error.RevlogError( | |
1835 | _("%s: size of %d bytes exceeds maximum revlog storage of 2GiB") |
|
1839 | _("%s: size of %d bytes exceeds maximum revlog storage of 2GiB") | |
1836 | % (self.indexfile, len(rawtext))) |
|
1840 | % (self.indexfile, len(rawtext))) | |
1837 |
|
1841 | |||
1838 | node = node or self.hash(rawtext, p1, p2) |
|
1842 | node = node or self.hash(rawtext, p1, p2) | |
1839 | if node in self.nodemap: |
|
1843 | if node in self.nodemap: | |
1840 | return node |
|
1844 | return node | |
1841 |
|
1845 | |||
1842 | if validatehash: |
|
1846 | if validatehash: | |
1843 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
1847 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
1844 |
|
1848 | |||
1845 | return self.addrawrevision(rawtext, transaction, link, p1, p2, node, |
|
1849 | return self.addrawrevision(rawtext, transaction, link, p1, p2, node, | |
1846 | flags, cachedelta=cachedelta, |
|
1850 | flags, cachedelta=cachedelta, | |
1847 | deltacomputer=deltacomputer) |
|
1851 | deltacomputer=deltacomputer) | |
1848 |
|
1852 | |||
1849 | def addrawrevision(self, rawtext, transaction, link, p1, p2, node, flags, |
|
1853 | def addrawrevision(self, rawtext, transaction, link, p1, p2, node, flags, | |
1850 | cachedelta=None, deltacomputer=None): |
|
1854 | cachedelta=None, deltacomputer=None): | |
1851 | """add a raw revision with known flags, node and parents |
|
1855 | """add a raw revision with known flags, node and parents | |
1852 | useful when reusing a revision not stored in this revlog (ex: received |
|
1856 | useful when reusing a revision not stored in this revlog (ex: received | |
1853 | over wire, or read from an external bundle). |
|
1857 | over wire, or read from an external bundle). | |
1854 | """ |
|
1858 | """ | |
1855 | dfh = None |
|
1859 | dfh = None | |
1856 | if not self._inline: |
|
1860 | if not self._inline: | |
1857 | dfh = self._datafp("a+") |
|
1861 | dfh = self._datafp("a+") | |
1858 | ifh = self._indexfp("a+") |
|
1862 | ifh = self._indexfp("a+") | |
1859 | try: |
|
1863 | try: | |
1860 | return self._addrevision(node, rawtext, transaction, link, p1, p2, |
|
1864 | return self._addrevision(node, rawtext, transaction, link, p1, p2, | |
1861 | flags, cachedelta, ifh, dfh, |
|
1865 | flags, cachedelta, ifh, dfh, | |
1862 | deltacomputer=deltacomputer) |
|
1866 | deltacomputer=deltacomputer) | |
1863 | finally: |
|
1867 | finally: | |
1864 | if dfh: |
|
1868 | if dfh: | |
1865 | dfh.close() |
|
1869 | dfh.close() | |
1866 | ifh.close() |
|
1870 | ifh.close() | |
1867 |
|
1871 | |||
1868 | def compress(self, data): |
|
1872 | def compress(self, data): | |
1869 | """Generate a possibly-compressed representation of data.""" |
|
1873 | """Generate a possibly-compressed representation of data.""" | |
1870 | if not data: |
|
1874 | if not data: | |
1871 | return '', data |
|
1875 | return '', data | |
1872 |
|
1876 | |||
1873 | compressed = self._compressor.compress(data) |
|
1877 | compressed = self._compressor.compress(data) | |
1874 |
|
1878 | |||
1875 | if compressed: |
|
1879 | if compressed: | |
1876 | # The revlog compressor added the header in the returned data. |
|
1880 | # The revlog compressor added the header in the returned data. | |
1877 | return '', compressed |
|
1881 | return '', compressed | |
1878 |
|
1882 | |||
1879 | if data[0:1] == '\0': |
|
1883 | if data[0:1] == '\0': | |
1880 | return '', data |
|
1884 | return '', data | |
1881 | return 'u', data |
|
1885 | return 'u', data | |
1882 |
|
1886 | |||
1883 | def decompress(self, data): |
|
1887 | def decompress(self, data): | |
1884 | """Decompress a revlog chunk. |
|
1888 | """Decompress a revlog chunk. | |
1885 |
|
1889 | |||
1886 | The chunk is expected to begin with a header identifying the |
|
1890 | The chunk is expected to begin with a header identifying the | |
1887 | format type so it can be routed to an appropriate decompressor. |
|
1891 | format type so it can be routed to an appropriate decompressor. | |
1888 | """ |
|
1892 | """ | |
1889 | if not data: |
|
1893 | if not data: | |
1890 | return data |
|
1894 | return data | |
1891 |
|
1895 | |||
1892 | # Revlogs are read much more frequently than they are written and many |
|
1896 | # Revlogs are read much more frequently than they are written and many | |
1893 | # chunks only take microseconds to decompress, so performance is |
|
1897 | # chunks only take microseconds to decompress, so performance is | |
1894 | # important here. |
|
1898 | # important here. | |
1895 | # |
|
1899 | # | |
1896 | # We can make a few assumptions about revlogs: |
|
1900 | # We can make a few assumptions about revlogs: | |
1897 | # |
|
1901 | # | |
1898 | # 1) the majority of chunks will be compressed (as opposed to inline |
|
1902 | # 1) the majority of chunks will be compressed (as opposed to inline | |
1899 | # raw data). |
|
1903 | # raw data). | |
1900 | # 2) decompressing *any* data will likely by at least 10x slower than |
|
1904 | # 2) decompressing *any* data will likely by at least 10x slower than | |
1901 | # returning raw inline data. |
|
1905 | # returning raw inline data. | |
1902 | # 3) we want to prioritize common and officially supported compression |
|
1906 | # 3) we want to prioritize common and officially supported compression | |
1903 | # engines |
|
1907 | # engines | |
1904 | # |
|
1908 | # | |
1905 | # It follows that we want to optimize for "decompress compressed data |
|
1909 | # It follows that we want to optimize for "decompress compressed data | |
1906 | # when encoded with common and officially supported compression engines" |
|
1910 | # when encoded with common and officially supported compression engines" | |
1907 | # case over "raw data" and "data encoded by less common or non-official |
|
1911 | # case over "raw data" and "data encoded by less common or non-official | |
1908 | # compression engines." That is why we have the inline lookup first |
|
1912 | # compression engines." That is why we have the inline lookup first | |
1909 | # followed by the compengines lookup. |
|
1913 | # followed by the compengines lookup. | |
1910 | # |
|
1914 | # | |
1911 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib |
|
1915 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib | |
1912 | # compressed chunks. And this matters for changelog and manifest reads. |
|
1916 | # compressed chunks. And this matters for changelog and manifest reads. | |
1913 | t = data[0:1] |
|
1917 | t = data[0:1] | |
1914 |
|
1918 | |||
1915 | if t == 'x': |
|
1919 | if t == 'x': | |
1916 | try: |
|
1920 | try: | |
1917 | return _zlibdecompress(data) |
|
1921 | return _zlibdecompress(data) | |
1918 | except zlib.error as e: |
|
1922 | except zlib.error as e: | |
1919 | raise error.RevlogError(_('revlog decompress error: %s') % |
|
1923 | raise error.RevlogError(_('revlog decompress error: %s') % | |
1920 | stringutil.forcebytestr(e)) |
|
1924 | stringutil.forcebytestr(e)) | |
1921 | # '\0' is more common than 'u' so it goes first. |
|
1925 | # '\0' is more common than 'u' so it goes first. | |
1922 | elif t == '\0': |
|
1926 | elif t == '\0': | |
1923 | return data |
|
1927 | return data | |
1924 | elif t == 'u': |
|
1928 | elif t == 'u': | |
1925 | return util.buffer(data, 1) |
|
1929 | return util.buffer(data, 1) | |
1926 |
|
1930 | |||
1927 | try: |
|
1931 | try: | |
1928 | compressor = self._decompressors[t] |
|
1932 | compressor = self._decompressors[t] | |
1929 | except KeyError: |
|
1933 | except KeyError: | |
1930 | try: |
|
1934 | try: | |
1931 | engine = util.compengines.forrevlogheader(t) |
|
1935 | engine = util.compengines.forrevlogheader(t) | |
1932 | compressor = engine.revlogcompressor(self._compengineopts) |
|
1936 | compressor = engine.revlogcompressor(self._compengineopts) | |
1933 | self._decompressors[t] = compressor |
|
1937 | self._decompressors[t] = compressor | |
1934 | except KeyError: |
|
1938 | except KeyError: | |
1935 | raise error.RevlogError(_('unknown compression type %r') % t) |
|
1939 | raise error.RevlogError(_('unknown compression type %r') % t) | |
1936 |
|
1940 | |||
1937 | return compressor.decompress(data) |
|
1941 | return compressor.decompress(data) | |
1938 |
|
1942 | |||
1939 | def _addrevision(self, node, rawtext, transaction, link, p1, p2, flags, |
|
1943 | def _addrevision(self, node, rawtext, transaction, link, p1, p2, flags, | |
1940 | cachedelta, ifh, dfh, alwayscache=False, |
|
1944 | cachedelta, ifh, dfh, alwayscache=False, | |
1941 | deltacomputer=None): |
|
1945 | deltacomputer=None): | |
1942 | """internal function to add revisions to the log |
|
1946 | """internal function to add revisions to the log | |
1943 |
|
1947 | |||
1944 | see addrevision for argument descriptions. |
|
1948 | see addrevision for argument descriptions. | |
1945 |
|
1949 | |||
1946 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. |
|
1950 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. | |
1947 |
|
1951 | |||
1948 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will |
|
1952 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will | |
1949 | be used. |
|
1953 | be used. | |
1950 |
|
1954 | |||
1951 | invariants: |
|
1955 | invariants: | |
1952 | - rawtext is optional (can be None); if not set, cachedelta must be set. |
|
1956 | - rawtext is optional (can be None); if not set, cachedelta must be set. | |
1953 | if both are set, they must correspond to each other. |
|
1957 | if both are set, they must correspond to each other. | |
1954 | """ |
|
1958 | """ | |
1955 | if node == nullid: |
|
1959 | if node == nullid: | |
1956 | raise error.RevlogError(_("%s: attempt to add null revision") % |
|
1960 | raise error.RevlogError(_("%s: attempt to add null revision") % | |
1957 | self.indexfile) |
|
1961 | self.indexfile) | |
1958 | if node == wdirid or node in wdirfilenodeids: |
|
1962 | if node == wdirid or node in wdirfilenodeids: | |
1959 | raise error.RevlogError(_("%s: attempt to add wdir revision") % |
|
1963 | raise error.RevlogError(_("%s: attempt to add wdir revision") % | |
1960 | self.indexfile) |
|
1964 | self.indexfile) | |
1961 |
|
1965 | |||
1962 | if self._inline: |
|
1966 | if self._inline: | |
1963 | fh = ifh |
|
1967 | fh = ifh | |
1964 | else: |
|
1968 | else: | |
1965 | fh = dfh |
|
1969 | fh = dfh | |
1966 |
|
1970 | |||
1967 | btext = [rawtext] |
|
1971 | btext = [rawtext] | |
1968 |
|
1972 | |||
1969 | curr = len(self) |
|
1973 | curr = len(self) | |
1970 | prev = curr - 1 |
|
1974 | prev = curr - 1 | |
1971 | offset = self.end(prev) |
|
1975 | offset = self.end(prev) | |
1972 | p1r, p2r = self.rev(p1), self.rev(p2) |
|
1976 | p1r, p2r = self.rev(p1), self.rev(p2) | |
1973 |
|
1977 | |||
1974 | # full versions are inserted when the needed deltas |
|
1978 | # full versions are inserted when the needed deltas | |
1975 | # become comparable to the uncompressed text |
|
1979 | # become comparable to the uncompressed text | |
1976 | if rawtext is None: |
|
1980 | if rawtext is None: | |
1977 | # need rawtext size, before changed by flag processors, which is |
|
1981 | # need rawtext size, before changed by flag processors, which is | |
1978 | # the non-raw size. use revlog explicitly to avoid filelog's extra |
|
1982 | # the non-raw size. use revlog explicitly to avoid filelog's extra | |
1979 | # logic that might remove metadata size. |
|
1983 | # logic that might remove metadata size. | |
1980 | textlen = mdiff.patchedsize(revlog.size(self, cachedelta[0]), |
|
1984 | textlen = mdiff.patchedsize(revlog.size(self, cachedelta[0]), | |
1981 | cachedelta[1]) |
|
1985 | cachedelta[1]) | |
1982 | else: |
|
1986 | else: | |
1983 | textlen = len(rawtext) |
|
1987 | textlen = len(rawtext) | |
1984 |
|
1988 | |||
1985 | if deltacomputer is None: |
|
1989 | if deltacomputer is None: | |
1986 | deltacomputer = deltautil.deltacomputer(self) |
|
1990 | deltacomputer = deltautil.deltacomputer(self) | |
1987 |
|
1991 | |||
1988 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) |
|
1992 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) | |
1989 |
|
1993 | |||
1990 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) |
|
1994 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) | |
1991 |
|
1995 | |||
1992 | e = (offset_type(offset, flags), deltainfo.deltalen, textlen, |
|
1996 | e = (offset_type(offset, flags), deltainfo.deltalen, textlen, | |
1993 | deltainfo.base, link, p1r, p2r, node) |
|
1997 | deltainfo.base, link, p1r, p2r, node) | |
1994 | self.index.append(e) |
|
1998 | self.index.append(e) | |
1995 | self.nodemap[node] = curr |
|
1999 | self.nodemap[node] = curr | |
1996 |
|
2000 | |||
1997 | # Reset the pure node cache start lookup offset to account for new |
|
2001 | # Reset the pure node cache start lookup offset to account for new | |
1998 | # revision. |
|
2002 | # revision. | |
1999 | if self._nodepos is not None: |
|
2003 | if self._nodepos is not None: | |
2000 | self._nodepos = curr |
|
2004 | self._nodepos = curr | |
2001 |
|
2005 | |||
2002 | entry = self._io.packentry(e, self.node, self.version, curr) |
|
2006 | entry = self._io.packentry(e, self.node, self.version, curr) | |
2003 | self._writeentry(transaction, ifh, dfh, entry, deltainfo.data, |
|
2007 | self._writeentry(transaction, ifh, dfh, entry, deltainfo.data, | |
2004 | link, offset) |
|
2008 | link, offset) | |
2005 |
|
2009 | |||
2006 | rawtext = btext[0] |
|
2010 | rawtext = btext[0] | |
2007 |
|
2011 | |||
2008 | if alwayscache and rawtext is None: |
|
2012 | if alwayscache and rawtext is None: | |
2009 | rawtext = deltacomputer.buildtext(revinfo, fh) |
|
2013 | rawtext = deltacomputer.buildtext(revinfo, fh) | |
2010 |
|
2014 | |||
2011 | if type(rawtext) == bytes: # only accept immutable objects |
|
2015 | if type(rawtext) == bytes: # only accept immutable objects | |
2012 | self._revisioncache = (node, curr, rawtext) |
|
2016 | self._revisioncache = (node, curr, rawtext) | |
2013 | self._chainbasecache[curr] = deltainfo.chainbase |
|
2017 | self._chainbasecache[curr] = deltainfo.chainbase | |
2014 | return node |
|
2018 | return node | |
2015 |
|
2019 | |||
2016 | def _writeentry(self, transaction, ifh, dfh, entry, data, link, offset): |
|
2020 | def _writeentry(self, transaction, ifh, dfh, entry, data, link, offset): | |
2017 | # Files opened in a+ mode have inconsistent behavior on various |
|
2021 | # Files opened in a+ mode have inconsistent behavior on various | |
2018 | # platforms. Windows requires that a file positioning call be made |
|
2022 | # platforms. Windows requires that a file positioning call be made | |
2019 | # when the file handle transitions between reads and writes. See |
|
2023 | # when the file handle transitions between reads and writes. See | |
2020 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
2024 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other | |
2021 | # platforms, Python or the platform itself can be buggy. Some versions |
|
2025 | # platforms, Python or the platform itself can be buggy. Some versions | |
2022 | # of Solaris have been observed to not append at the end of the file |
|
2026 | # of Solaris have been observed to not append at the end of the file | |
2023 | # if the file was seeked to before the end. See issue4943 for more. |
|
2027 | # if the file was seeked to before the end. See issue4943 for more. | |
2024 | # |
|
2028 | # | |
2025 | # We work around this issue by inserting a seek() before writing. |
|
2029 | # We work around this issue by inserting a seek() before writing. | |
2026 | # Note: This is likely not necessary on Python 3. However, because |
|
2030 | # Note: This is likely not necessary on Python 3. However, because | |
2027 | # the file handle is reused for reads and may be seeked there, we need |
|
2031 | # the file handle is reused for reads and may be seeked there, we need | |
2028 | # to be careful before changing this. |
|
2032 | # to be careful before changing this. | |
2029 | ifh.seek(0, os.SEEK_END) |
|
2033 | ifh.seek(0, os.SEEK_END) | |
2030 | if dfh: |
|
2034 | if dfh: | |
2031 | dfh.seek(0, os.SEEK_END) |
|
2035 | dfh.seek(0, os.SEEK_END) | |
2032 |
|
2036 | |||
2033 | curr = len(self) - 1 |
|
2037 | curr = len(self) - 1 | |
2034 | if not self._inline: |
|
2038 | if not self._inline: | |
2035 | transaction.add(self.datafile, offset) |
|
2039 | transaction.add(self.datafile, offset) | |
2036 | transaction.add(self.indexfile, curr * len(entry)) |
|
2040 | transaction.add(self.indexfile, curr * len(entry)) | |
2037 | if data[0]: |
|
2041 | if data[0]: | |
2038 | dfh.write(data[0]) |
|
2042 | dfh.write(data[0]) | |
2039 | dfh.write(data[1]) |
|
2043 | dfh.write(data[1]) | |
2040 | ifh.write(entry) |
|
2044 | ifh.write(entry) | |
2041 | else: |
|
2045 | else: | |
2042 | offset += curr * self._io.size |
|
2046 | offset += curr * self._io.size | |
2043 | transaction.add(self.indexfile, offset, curr) |
|
2047 | transaction.add(self.indexfile, offset, curr) | |
2044 | ifh.write(entry) |
|
2048 | ifh.write(entry) | |
2045 | ifh.write(data[0]) |
|
2049 | ifh.write(data[0]) | |
2046 | ifh.write(data[1]) |
|
2050 | ifh.write(data[1]) | |
2047 | self._enforceinlinesize(transaction, ifh) |
|
2051 | self._enforceinlinesize(transaction, ifh) | |
2048 |
|
2052 | |||
2049 | def addgroup(self, deltas, linkmapper, transaction, addrevisioncb=None): |
|
2053 | def addgroup(self, deltas, linkmapper, transaction, addrevisioncb=None): | |
2050 | """ |
|
2054 | """ | |
2051 | add a delta group |
|
2055 | add a delta group | |
2052 |
|
2056 | |||
2053 | given a set of deltas, add them to the revision log. the |
|
2057 | given a set of deltas, add them to the revision log. the | |
2054 | first delta is against its parent, which should be in our |
|
2058 | first delta is against its parent, which should be in our | |
2055 | log, the rest are against the previous delta. |
|
2059 | log, the rest are against the previous delta. | |
2056 |
|
2060 | |||
2057 | If ``addrevisioncb`` is defined, it will be called with arguments of |
|
2061 | If ``addrevisioncb`` is defined, it will be called with arguments of | |
2058 | this revlog and the node that was added. |
|
2062 | this revlog and the node that was added. | |
2059 | """ |
|
2063 | """ | |
2060 |
|
2064 | |||
2061 | if self._writinghandles: |
|
2065 | if self._writinghandles: | |
2062 | raise error.ProgrammingError('cannot nest addgroup() calls') |
|
2066 | raise error.ProgrammingError('cannot nest addgroup() calls') | |
2063 |
|
2067 | |||
2064 | nodes = [] |
|
2068 | nodes = [] | |
2065 |
|
2069 | |||
2066 | r = len(self) |
|
2070 | r = len(self) | |
2067 | end = 0 |
|
2071 | end = 0 | |
2068 | if r: |
|
2072 | if r: | |
2069 | end = self.end(r - 1) |
|
2073 | end = self.end(r - 1) | |
2070 | ifh = self._indexfp("a+") |
|
2074 | ifh = self._indexfp("a+") | |
2071 | isize = r * self._io.size |
|
2075 | isize = r * self._io.size | |
2072 | if self._inline: |
|
2076 | if self._inline: | |
2073 | transaction.add(self.indexfile, end + isize, r) |
|
2077 | transaction.add(self.indexfile, end + isize, r) | |
2074 | dfh = None |
|
2078 | dfh = None | |
2075 | else: |
|
2079 | else: | |
2076 | transaction.add(self.indexfile, isize, r) |
|
2080 | transaction.add(self.indexfile, isize, r) | |
2077 | transaction.add(self.datafile, end) |
|
2081 | transaction.add(self.datafile, end) | |
2078 | dfh = self._datafp("a+") |
|
2082 | dfh = self._datafp("a+") | |
2079 | def flush(): |
|
2083 | def flush(): | |
2080 | if dfh: |
|
2084 | if dfh: | |
2081 | dfh.flush() |
|
2085 | dfh.flush() | |
2082 | ifh.flush() |
|
2086 | ifh.flush() | |
2083 |
|
2087 | |||
2084 | self._writinghandles = (ifh, dfh) |
|
2088 | self._writinghandles = (ifh, dfh) | |
2085 |
|
2089 | |||
2086 | try: |
|
2090 | try: | |
2087 | deltacomputer = deltautil.deltacomputer(self) |
|
2091 | deltacomputer = deltautil.deltacomputer(self) | |
2088 | # loop through our set of deltas |
|
2092 | # loop through our set of deltas | |
2089 | for data in deltas: |
|
2093 | for data in deltas: | |
2090 | node, p1, p2, linknode, deltabase, delta, flags = data |
|
2094 | node, p1, p2, linknode, deltabase, delta, flags = data | |
2091 | link = linkmapper(linknode) |
|
2095 | link = linkmapper(linknode) | |
2092 | flags = flags or REVIDX_DEFAULT_FLAGS |
|
2096 | flags = flags or REVIDX_DEFAULT_FLAGS | |
2093 |
|
2097 | |||
2094 | nodes.append(node) |
|
2098 | nodes.append(node) | |
2095 |
|
2099 | |||
2096 | if node in self.nodemap: |
|
2100 | if node in self.nodemap: | |
2097 | self._nodeduplicatecallback(transaction, node) |
|
2101 | self._nodeduplicatecallback(transaction, node) | |
2098 | # this can happen if two branches make the same change |
|
2102 | # this can happen if two branches make the same change | |
2099 | continue |
|
2103 | continue | |
2100 |
|
2104 | |||
2101 | for p in (p1, p2): |
|
2105 | for p in (p1, p2): | |
2102 | if p not in self.nodemap: |
|
2106 | if p not in self.nodemap: | |
2103 | raise error.LookupError(p, self.indexfile, |
|
2107 | raise error.LookupError(p, self.indexfile, | |
2104 | _('unknown parent')) |
|
2108 | _('unknown parent')) | |
2105 |
|
2109 | |||
2106 | if deltabase not in self.nodemap: |
|
2110 | if deltabase not in self.nodemap: | |
2107 | raise error.LookupError(deltabase, self.indexfile, |
|
2111 | raise error.LookupError(deltabase, self.indexfile, | |
2108 | _('unknown delta base')) |
|
2112 | _('unknown delta base')) | |
2109 |
|
2113 | |||
2110 | baserev = self.rev(deltabase) |
|
2114 | baserev = self.rev(deltabase) | |
2111 |
|
2115 | |||
2112 | if baserev != nullrev and self.iscensored(baserev): |
|
2116 | if baserev != nullrev and self.iscensored(baserev): | |
2113 | # if base is censored, delta must be full replacement in a |
|
2117 | # if base is censored, delta must be full replacement in a | |
2114 | # single patch operation |
|
2118 | # single patch operation | |
2115 | hlen = struct.calcsize(">lll") |
|
2119 | hlen = struct.calcsize(">lll") | |
2116 | oldlen = self.rawsize(baserev) |
|
2120 | oldlen = self.rawsize(baserev) | |
2117 | newlen = len(delta) - hlen |
|
2121 | newlen = len(delta) - hlen | |
2118 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): |
|
2122 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): | |
2119 | raise error.CensoredBaseError(self.indexfile, |
|
2123 | raise error.CensoredBaseError(self.indexfile, | |
2120 | self.node(baserev)) |
|
2124 | self.node(baserev)) | |
2121 |
|
2125 | |||
2122 | if not flags and self._peek_iscensored(baserev, delta, flush): |
|
2126 | if not flags and self._peek_iscensored(baserev, delta, flush): | |
2123 | flags |= REVIDX_ISCENSORED |
|
2127 | flags |= REVIDX_ISCENSORED | |
2124 |
|
2128 | |||
2125 | # We assume consumers of addrevisioncb will want to retrieve |
|
2129 | # We assume consumers of addrevisioncb will want to retrieve | |
2126 | # the added revision, which will require a call to |
|
2130 | # the added revision, which will require a call to | |
2127 | # revision(). revision() will fast path if there is a cache |
|
2131 | # revision(). revision() will fast path if there is a cache | |
2128 | # hit. So, we tell _addrevision() to always cache in this case. |
|
2132 | # hit. So, we tell _addrevision() to always cache in this case. | |
2129 | # We're only using addgroup() in the context of changegroup |
|
2133 | # We're only using addgroup() in the context of changegroup | |
2130 | # generation so the revision data can always be handled as raw |
|
2134 | # generation so the revision data can always be handled as raw | |
2131 | # by the flagprocessor. |
|
2135 | # by the flagprocessor. | |
2132 | self._addrevision(node, None, transaction, link, |
|
2136 | self._addrevision(node, None, transaction, link, | |
2133 | p1, p2, flags, (baserev, delta), |
|
2137 | p1, p2, flags, (baserev, delta), | |
2134 | ifh, dfh, |
|
2138 | ifh, dfh, | |
2135 | alwayscache=bool(addrevisioncb), |
|
2139 | alwayscache=bool(addrevisioncb), | |
2136 | deltacomputer=deltacomputer) |
|
2140 | deltacomputer=deltacomputer) | |
2137 |
|
2141 | |||
2138 | if addrevisioncb: |
|
2142 | if addrevisioncb: | |
2139 | addrevisioncb(self, node) |
|
2143 | addrevisioncb(self, node) | |
2140 |
|
2144 | |||
2141 | if not dfh and not self._inline: |
|
2145 | if not dfh and not self._inline: | |
2142 | # addrevision switched from inline to conventional |
|
2146 | # addrevision switched from inline to conventional | |
2143 | # reopen the index |
|
2147 | # reopen the index | |
2144 | ifh.close() |
|
2148 | ifh.close() | |
2145 | dfh = self._datafp("a+") |
|
2149 | dfh = self._datafp("a+") | |
2146 | ifh = self._indexfp("a+") |
|
2150 | ifh = self._indexfp("a+") | |
2147 | self._writinghandles = (ifh, dfh) |
|
2151 | self._writinghandles = (ifh, dfh) | |
2148 | finally: |
|
2152 | finally: | |
2149 | self._writinghandles = None |
|
2153 | self._writinghandles = None | |
2150 |
|
2154 | |||
2151 | if dfh: |
|
2155 | if dfh: | |
2152 | dfh.close() |
|
2156 | dfh.close() | |
2153 | ifh.close() |
|
2157 | ifh.close() | |
2154 |
|
2158 | |||
2155 | return nodes |
|
2159 | return nodes | |
2156 |
|
2160 | |||
2157 | def iscensored(self, rev): |
|
2161 | def iscensored(self, rev): | |
2158 | """Check if a file revision is censored.""" |
|
2162 | """Check if a file revision is censored.""" | |
2159 | if not self._censorable: |
|
2163 | if not self._censorable: | |
2160 | return False |
|
2164 | return False | |
2161 |
|
2165 | |||
2162 | return self.flags(rev) & REVIDX_ISCENSORED |
|
2166 | return self.flags(rev) & REVIDX_ISCENSORED | |
2163 |
|
2167 | |||
2164 | def _peek_iscensored(self, baserev, delta, flush): |
|
2168 | def _peek_iscensored(self, baserev, delta, flush): | |
2165 | """Quickly check if a delta produces a censored revision.""" |
|
2169 | """Quickly check if a delta produces a censored revision.""" | |
2166 | if not self._censorable: |
|
2170 | if not self._censorable: | |
2167 | return False |
|
2171 | return False | |
2168 |
|
2172 | |||
2169 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) |
|
2173 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) | |
2170 |
|
2174 | |||
2171 | def getstrippoint(self, minlink): |
|
2175 | def getstrippoint(self, minlink): | |
2172 | """find the minimum rev that must be stripped to strip the linkrev |
|
2176 | """find the minimum rev that must be stripped to strip the linkrev | |
2173 |
|
2177 | |||
2174 | Returns a tuple containing the minimum rev and a set of all revs that |
|
2178 | Returns a tuple containing the minimum rev and a set of all revs that | |
2175 | have linkrevs that will be broken by this strip. |
|
2179 | have linkrevs that will be broken by this strip. | |
2176 | """ |
|
2180 | """ | |
2177 | return storageutil.resolvestripinfo(minlink, len(self) - 1, |
|
2181 | return storageutil.resolvestripinfo(minlink, len(self) - 1, | |
2178 | self.headrevs(), |
|
2182 | self.headrevs(), | |
2179 | self.linkrev, self.parentrevs) |
|
2183 | self.linkrev, self.parentrevs) | |
2180 |
|
2184 | |||
2181 | def strip(self, minlink, transaction): |
|
2185 | def strip(self, minlink, transaction): | |
2182 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
2186 | """truncate the revlog on the first revision with a linkrev >= minlink | |
2183 |
|
2187 | |||
2184 | This function is called when we're stripping revision minlink and |
|
2188 | This function is called when we're stripping revision minlink and | |
2185 | its descendants from the repository. |
|
2189 | its descendants from the repository. | |
2186 |
|
2190 | |||
2187 | We have to remove all revisions with linkrev >= minlink, because |
|
2191 | We have to remove all revisions with linkrev >= minlink, because | |
2188 | the equivalent changelog revisions will be renumbered after the |
|
2192 | the equivalent changelog revisions will be renumbered after the | |
2189 | strip. |
|
2193 | strip. | |
2190 |
|
2194 | |||
2191 | So we truncate the revlog on the first of these revisions, and |
|
2195 | So we truncate the revlog on the first of these revisions, and | |
2192 | trust that the caller has saved the revisions that shouldn't be |
|
2196 | trust that the caller has saved the revisions that shouldn't be | |
2193 | removed and that it'll re-add them after this truncation. |
|
2197 | removed and that it'll re-add them after this truncation. | |
2194 | """ |
|
2198 | """ | |
2195 | if len(self) == 0: |
|
2199 | if len(self) == 0: | |
2196 | return |
|
2200 | return | |
2197 |
|
2201 | |||
2198 | rev, _ = self.getstrippoint(minlink) |
|
2202 | rev, _ = self.getstrippoint(minlink) | |
2199 | if rev == len(self): |
|
2203 | if rev == len(self): | |
2200 | return |
|
2204 | return | |
2201 |
|
2205 | |||
2202 | # first truncate the files on disk |
|
2206 | # first truncate the files on disk | |
2203 | end = self.start(rev) |
|
2207 | end = self.start(rev) | |
2204 | if not self._inline: |
|
2208 | if not self._inline: | |
2205 | transaction.add(self.datafile, end) |
|
2209 | transaction.add(self.datafile, end) | |
2206 | end = rev * self._io.size |
|
2210 | end = rev * self._io.size | |
2207 | else: |
|
2211 | else: | |
2208 | end += rev * self._io.size |
|
2212 | end += rev * self._io.size | |
2209 |
|
2213 | |||
2210 | transaction.add(self.indexfile, end) |
|
2214 | transaction.add(self.indexfile, end) | |
2211 |
|
2215 | |||
2212 | # then reset internal state in memory to forget those revisions |
|
2216 | # then reset internal state in memory to forget those revisions | |
2213 | self._revisioncache = None |
|
2217 | self._revisioncache = None | |
2214 | self._chaininfocache = {} |
|
2218 | self._chaininfocache = {} | |
2215 | self._chunkclear() |
|
2219 | self._chunkclear() | |
2216 | for x in pycompat.xrange(rev, len(self)): |
|
2220 | for x in pycompat.xrange(rev, len(self)): | |
2217 | del self.nodemap[self.node(x)] |
|
2221 | del self.nodemap[self.node(x)] | |
2218 |
|
2222 | |||
2219 | del self.index[rev:-1] |
|
2223 | del self.index[rev:-1] | |
2220 | self._nodepos = None |
|
2224 | self._nodepos = None | |
2221 |
|
2225 | |||
2222 | def checksize(self): |
|
2226 | def checksize(self): | |
2223 | """Check size of index and data files |
|
2227 | """Check size of index and data files | |
2224 |
|
2228 | |||
2225 | return a (dd, di) tuple. |
|
2229 | return a (dd, di) tuple. | |
2226 | - dd: extra bytes for the "data" file |
|
2230 | - dd: extra bytes for the "data" file | |
2227 | - di: extra bytes for the "index" file |
|
2231 | - di: extra bytes for the "index" file | |
2228 |
|
2232 | |||
2229 | A healthy revlog will return (0, 0). |
|
2233 | A healthy revlog will return (0, 0). | |
2230 | """ |
|
2234 | """ | |
2231 | expected = 0 |
|
2235 | expected = 0 | |
2232 | if len(self): |
|
2236 | if len(self): | |
2233 | expected = max(0, self.end(len(self) - 1)) |
|
2237 | expected = max(0, self.end(len(self) - 1)) | |
2234 |
|
2238 | |||
2235 | try: |
|
2239 | try: | |
2236 | with self._datafp() as f: |
|
2240 | with self._datafp() as f: | |
2237 | f.seek(0, io.SEEK_END) |
|
2241 | f.seek(0, io.SEEK_END) | |
2238 | actual = f.tell() |
|
2242 | actual = f.tell() | |
2239 | dd = actual - expected |
|
2243 | dd = actual - expected | |
2240 | except IOError as inst: |
|
2244 | except IOError as inst: | |
2241 | if inst.errno != errno.ENOENT: |
|
2245 | if inst.errno != errno.ENOENT: | |
2242 | raise |
|
2246 | raise | |
2243 | dd = 0 |
|
2247 | dd = 0 | |
2244 |
|
2248 | |||
2245 | try: |
|
2249 | try: | |
2246 | f = self.opener(self.indexfile) |
|
2250 | f = self.opener(self.indexfile) | |
2247 | f.seek(0, io.SEEK_END) |
|
2251 | f.seek(0, io.SEEK_END) | |
2248 | actual = f.tell() |
|
2252 | actual = f.tell() | |
2249 | f.close() |
|
2253 | f.close() | |
2250 | s = self._io.size |
|
2254 | s = self._io.size | |
2251 | i = max(0, actual // s) |
|
2255 | i = max(0, actual // s) | |
2252 | di = actual - (i * s) |
|
2256 | di = actual - (i * s) | |
2253 | if self._inline: |
|
2257 | if self._inline: | |
2254 | databytes = 0 |
|
2258 | databytes = 0 | |
2255 | for r in self: |
|
2259 | for r in self: | |
2256 | databytes += max(0, self.length(r)) |
|
2260 | databytes += max(0, self.length(r)) | |
2257 | dd = 0 |
|
2261 | dd = 0 | |
2258 | di = actual - len(self) * s - databytes |
|
2262 | di = actual - len(self) * s - databytes | |
2259 | except IOError as inst: |
|
2263 | except IOError as inst: | |
2260 | if inst.errno != errno.ENOENT: |
|
2264 | if inst.errno != errno.ENOENT: | |
2261 | raise |
|
2265 | raise | |
2262 | di = 0 |
|
2266 | di = 0 | |
2263 |
|
2267 | |||
2264 | return (dd, di) |
|
2268 | return (dd, di) | |
2265 |
|
2269 | |||
2266 | def files(self): |
|
2270 | def files(self): | |
2267 | res = [self.indexfile] |
|
2271 | res = [self.indexfile] | |
2268 | if not self._inline: |
|
2272 | if not self._inline: | |
2269 | res.append(self.datafile) |
|
2273 | res.append(self.datafile) | |
2270 | return res |
|
2274 | return res | |
2271 |
|
2275 | |||
2272 | def emitrevisions(self, nodes, nodesorder=None, revisiondata=False, |
|
2276 | def emitrevisions(self, nodes, nodesorder=None, revisiondata=False, | |
2273 | assumehaveparentrevisions=False, |
|
2277 | assumehaveparentrevisions=False, | |
2274 | deltamode=repository.CG_DELTAMODE_STD): |
|
2278 | deltamode=repository.CG_DELTAMODE_STD): | |
2275 | if nodesorder not in ('nodes', 'storage', 'linear', None): |
|
2279 | if nodesorder not in ('nodes', 'storage', 'linear', None): | |
2276 | raise error.ProgrammingError('unhandled value for nodesorder: %s' % |
|
2280 | raise error.ProgrammingError('unhandled value for nodesorder: %s' % | |
2277 | nodesorder) |
|
2281 | nodesorder) | |
2278 |
|
2282 | |||
2279 | if nodesorder is None and not self._generaldelta: |
|
2283 | if nodesorder is None and not self._generaldelta: | |
2280 | nodesorder = 'storage' |
|
2284 | nodesorder = 'storage' | |
2281 |
|
2285 | |||
2282 | if (not self._storedeltachains and |
|
2286 | if (not self._storedeltachains and | |
2283 | deltamode != repository.CG_DELTAMODE_PREV): |
|
2287 | deltamode != repository.CG_DELTAMODE_PREV): | |
2284 | deltamode = repository.CG_DELTAMODE_FULL |
|
2288 | deltamode = repository.CG_DELTAMODE_FULL | |
2285 |
|
2289 | |||
2286 | return storageutil.emitrevisions( |
|
2290 | return storageutil.emitrevisions( | |
2287 | self, nodes, nodesorder, revlogrevisiondelta, |
|
2291 | self, nodes, nodesorder, revlogrevisiondelta, | |
2288 | deltaparentfn=self.deltaparent, |
|
2292 | deltaparentfn=self.deltaparent, | |
2289 | candeltafn=self.candelta, |
|
2293 | candeltafn=self.candelta, | |
2290 | rawsizefn=self.rawsize, |
|
2294 | rawsizefn=self.rawsize, | |
2291 | revdifffn=self.revdiff, |
|
2295 | revdifffn=self.revdiff, | |
2292 | flagsfn=self.flags, |
|
2296 | flagsfn=self.flags, | |
2293 | deltamode=deltamode, |
|
2297 | deltamode=deltamode, | |
2294 | revisiondata=revisiondata, |
|
2298 | revisiondata=revisiondata, | |
2295 | assumehaveparentrevisions=assumehaveparentrevisions) |
|
2299 | assumehaveparentrevisions=assumehaveparentrevisions) | |
2296 |
|
2300 | |||
2297 | DELTAREUSEALWAYS = 'always' |
|
2301 | DELTAREUSEALWAYS = 'always' | |
2298 | DELTAREUSESAMEREVS = 'samerevs' |
|
2302 | DELTAREUSESAMEREVS = 'samerevs' | |
2299 | DELTAREUSENEVER = 'never' |
|
2303 | DELTAREUSENEVER = 'never' | |
2300 |
|
2304 | |||
2301 | DELTAREUSEFULLADD = 'fulladd' |
|
2305 | DELTAREUSEFULLADD = 'fulladd' | |
2302 |
|
2306 | |||
2303 | DELTAREUSEALL = {'always', 'samerevs', 'never', 'fulladd'} |
|
2307 | DELTAREUSEALL = {'always', 'samerevs', 'never', 'fulladd'} | |
2304 |
|
2308 | |||
2305 | def clone(self, tr, destrevlog, addrevisioncb=None, |
|
2309 | def clone(self, tr, destrevlog, addrevisioncb=None, | |
2306 | deltareuse=DELTAREUSESAMEREVS, forcedeltabothparents=None): |
|
2310 | deltareuse=DELTAREUSESAMEREVS, forcedeltabothparents=None): | |
2307 | """Copy this revlog to another, possibly with format changes. |
|
2311 | """Copy this revlog to another, possibly with format changes. | |
2308 |
|
2312 | |||
2309 | The destination revlog will contain the same revisions and nodes. |
|
2313 | The destination revlog will contain the same revisions and nodes. | |
2310 | However, it may not be bit-for-bit identical due to e.g. delta encoding |
|
2314 | However, it may not be bit-for-bit identical due to e.g. delta encoding | |
2311 | differences. |
|
2315 | differences. | |
2312 |
|
2316 | |||
2313 | The ``deltareuse`` argument control how deltas from the existing revlog |
|
2317 | The ``deltareuse`` argument control how deltas from the existing revlog | |
2314 | are preserved in the destination revlog. The argument can have the |
|
2318 | are preserved in the destination revlog. The argument can have the | |
2315 | following values: |
|
2319 | following values: | |
2316 |
|
2320 | |||
2317 | DELTAREUSEALWAYS |
|
2321 | DELTAREUSEALWAYS | |
2318 | Deltas will always be reused (if possible), even if the destination |
|
2322 | Deltas will always be reused (if possible), even if the destination | |
2319 | revlog would not select the same revisions for the delta. This is the |
|
2323 | revlog would not select the same revisions for the delta. This is the | |
2320 | fastest mode of operation. |
|
2324 | fastest mode of operation. | |
2321 | DELTAREUSESAMEREVS |
|
2325 | DELTAREUSESAMEREVS | |
2322 | Deltas will be reused if the destination revlog would pick the same |
|
2326 | Deltas will be reused if the destination revlog would pick the same | |
2323 | revisions for the delta. This mode strikes a balance between speed |
|
2327 | revisions for the delta. This mode strikes a balance between speed | |
2324 | and optimization. |
|
2328 | and optimization. | |
2325 | DELTAREUSENEVER |
|
2329 | DELTAREUSENEVER | |
2326 | Deltas will never be reused. This is the slowest mode of execution. |
|
2330 | Deltas will never be reused. This is the slowest mode of execution. | |
2327 | This mode can be used to recompute deltas (e.g. if the diff/delta |
|
2331 | This mode can be used to recompute deltas (e.g. if the diff/delta | |
2328 | algorithm changes). |
|
2332 | algorithm changes). | |
2329 |
|
2333 | |||
2330 | Delta computation can be slow, so the choice of delta reuse policy can |
|
2334 | Delta computation can be slow, so the choice of delta reuse policy can | |
2331 | significantly affect run time. |
|
2335 | significantly affect run time. | |
2332 |
|
2336 | |||
2333 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between |
|
2337 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between | |
2334 | two extremes. Deltas will be reused if they are appropriate. But if the |
|
2338 | two extremes. Deltas will be reused if they are appropriate. But if the | |
2335 | delta could choose a better revision, it will do so. This means if you |
|
2339 | delta could choose a better revision, it will do so. This means if you | |
2336 | are converting a non-generaldelta revlog to a generaldelta revlog, |
|
2340 | are converting a non-generaldelta revlog to a generaldelta revlog, | |
2337 | deltas will be recomputed if the delta's parent isn't a parent of the |
|
2341 | deltas will be recomputed if the delta's parent isn't a parent of the | |
2338 | revision. |
|
2342 | revision. | |
2339 |
|
2343 | |||
2340 | In addition to the delta policy, the ``forcedeltabothparents`` |
|
2344 | In addition to the delta policy, the ``forcedeltabothparents`` | |
2341 | argument controls whether to force compute deltas against both parents |
|
2345 | argument controls whether to force compute deltas against both parents | |
2342 | for merges. By default, the current default is used. |
|
2346 | for merges. By default, the current default is used. | |
2343 | """ |
|
2347 | """ | |
2344 | if deltareuse not in self.DELTAREUSEALL: |
|
2348 | if deltareuse not in self.DELTAREUSEALL: | |
2345 | raise ValueError(_('value for deltareuse invalid: %s') % deltareuse) |
|
2349 | raise ValueError(_('value for deltareuse invalid: %s') % deltareuse) | |
2346 |
|
2350 | |||
2347 | if len(destrevlog): |
|
2351 | if len(destrevlog): | |
2348 | raise ValueError(_('destination revlog is not empty')) |
|
2352 | raise ValueError(_('destination revlog is not empty')) | |
2349 |
|
2353 | |||
2350 | if getattr(self, 'filteredrevs', None): |
|
2354 | if getattr(self, 'filteredrevs', None): | |
2351 | raise ValueError(_('source revlog has filtered revisions')) |
|
2355 | raise ValueError(_('source revlog has filtered revisions')) | |
2352 | if getattr(destrevlog, 'filteredrevs', None): |
|
2356 | if getattr(destrevlog, 'filteredrevs', None): | |
2353 | raise ValueError(_('destination revlog has filtered revisions')) |
|
2357 | raise ValueError(_('destination revlog has filtered revisions')) | |
2354 |
|
2358 | |||
2355 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, |
|
2359 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, | |
2356 | # if possible. |
|
2360 | # if possible. | |
2357 | oldlazydelta = destrevlog._lazydelta |
|
2361 | oldlazydelta = destrevlog._lazydelta | |
2358 | oldlazydeltabase = destrevlog._lazydeltabase |
|
2362 | oldlazydeltabase = destrevlog._lazydeltabase | |
2359 | oldamd = destrevlog._deltabothparents |
|
2363 | oldamd = destrevlog._deltabothparents | |
2360 |
|
2364 | |||
2361 | try: |
|
2365 | try: | |
2362 | if deltareuse == self.DELTAREUSEALWAYS: |
|
2366 | if deltareuse == self.DELTAREUSEALWAYS: | |
2363 | destrevlog._lazydeltabase = True |
|
2367 | destrevlog._lazydeltabase = True | |
2364 | destrevlog._lazydelta = True |
|
2368 | destrevlog._lazydelta = True | |
2365 | elif deltareuse == self.DELTAREUSESAMEREVS: |
|
2369 | elif deltareuse == self.DELTAREUSESAMEREVS: | |
2366 | destrevlog._lazydeltabase = False |
|
2370 | destrevlog._lazydeltabase = False | |
2367 | destrevlog._lazydelta = True |
|
2371 | destrevlog._lazydelta = True | |
2368 | elif deltareuse == self.DELTAREUSENEVER: |
|
2372 | elif deltareuse == self.DELTAREUSENEVER: | |
2369 | destrevlog._lazydeltabase = False |
|
2373 | destrevlog._lazydeltabase = False | |
2370 | destrevlog._lazydelta = False |
|
2374 | destrevlog._lazydelta = False | |
2371 |
|
2375 | |||
2372 | destrevlog._deltabothparents = forcedeltabothparents or oldamd |
|
2376 | destrevlog._deltabothparents = forcedeltabothparents or oldamd | |
2373 |
|
2377 | |||
2374 | deltacomputer = deltautil.deltacomputer(destrevlog) |
|
2378 | deltacomputer = deltautil.deltacomputer(destrevlog) | |
2375 | index = self.index |
|
2379 | index = self.index | |
2376 | for rev in self: |
|
2380 | for rev in self: | |
2377 | entry = index[rev] |
|
2381 | entry = index[rev] | |
2378 |
|
2382 | |||
2379 | # Some classes override linkrev to take filtered revs into |
|
2383 | # Some classes override linkrev to take filtered revs into | |
2380 | # account. Use raw entry from index. |
|
2384 | # account. Use raw entry from index. | |
2381 | flags = entry[0] & 0xffff |
|
2385 | flags = entry[0] & 0xffff | |
2382 | linkrev = entry[4] |
|
2386 | linkrev = entry[4] | |
2383 | p1 = index[entry[5]][7] |
|
2387 | p1 = index[entry[5]][7] | |
2384 | p2 = index[entry[6]][7] |
|
2388 | p2 = index[entry[6]][7] | |
2385 | node = entry[7] |
|
2389 | node = entry[7] | |
2386 |
|
2390 | |||
2387 | # (Possibly) reuse the delta from the revlog if allowed and |
|
2391 | # (Possibly) reuse the delta from the revlog if allowed and | |
2388 | # the revlog chunk is a delta. |
|
2392 | # the revlog chunk is a delta. | |
2389 | cachedelta = None |
|
2393 | cachedelta = None | |
2390 | rawtext = None |
|
2394 | rawtext = None | |
2391 | if (deltareuse != self.DELTAREUSEFULLADD |
|
2395 | if (deltareuse != self.DELTAREUSEFULLADD | |
2392 | and destrevlog._lazydelta): |
|
2396 | and destrevlog._lazydelta): | |
2393 | dp = self.deltaparent(rev) |
|
2397 | dp = self.deltaparent(rev) | |
2394 | if dp != nullrev: |
|
2398 | if dp != nullrev: | |
2395 | cachedelta = (dp, bytes(self._chunk(rev))) |
|
2399 | cachedelta = (dp, bytes(self._chunk(rev))) | |
2396 |
|
2400 | |||
2397 | if not cachedelta: |
|
2401 | if not cachedelta: | |
2398 | rawtext = self.rawdata(rev) |
|
2402 | rawtext = self.rawdata(rev) | |
2399 |
|
2403 | |||
2400 |
|
2404 | |||
2401 | if deltareuse == self.DELTAREUSEFULLADD: |
|
2405 | if deltareuse == self.DELTAREUSEFULLADD: | |
2402 | destrevlog.addrevision(rawtext, tr, linkrev, p1, p2, |
|
2406 | destrevlog.addrevision(rawtext, tr, linkrev, p1, p2, | |
2403 | cachedelta=cachedelta, |
|
2407 | cachedelta=cachedelta, | |
2404 | node=node, flags=flags, |
|
2408 | node=node, flags=flags, | |
2405 | deltacomputer=deltacomputer) |
|
2409 | deltacomputer=deltacomputer) | |
2406 | else: |
|
2410 | else: | |
2407 | ifh = destrevlog.opener(destrevlog.indexfile, 'a+', |
|
2411 | ifh = destrevlog.opener(destrevlog.indexfile, 'a+', | |
2408 | checkambig=False) |
|
2412 | checkambig=False) | |
2409 | dfh = None |
|
2413 | dfh = None | |
2410 | if not destrevlog._inline: |
|
2414 | if not destrevlog._inline: | |
2411 | dfh = destrevlog.opener(destrevlog.datafile, 'a+') |
|
2415 | dfh = destrevlog.opener(destrevlog.datafile, 'a+') | |
2412 | try: |
|
2416 | try: | |
2413 | destrevlog._addrevision(node, rawtext, tr, linkrev, p1, |
|
2417 | destrevlog._addrevision(node, rawtext, tr, linkrev, p1, | |
2414 | p2, flags, cachedelta, ifh, dfh, |
|
2418 | p2, flags, cachedelta, ifh, dfh, | |
2415 | deltacomputer=deltacomputer) |
|
2419 | deltacomputer=deltacomputer) | |
2416 | finally: |
|
2420 | finally: | |
2417 | if dfh: |
|
2421 | if dfh: | |
2418 | dfh.close() |
|
2422 | dfh.close() | |
2419 | ifh.close() |
|
2423 | ifh.close() | |
2420 |
|
2424 | |||
2421 | if addrevisioncb: |
|
2425 | if addrevisioncb: | |
2422 | addrevisioncb(self, rev, node) |
|
2426 | addrevisioncb(self, rev, node) | |
2423 | finally: |
|
2427 | finally: | |
2424 | destrevlog._lazydelta = oldlazydelta |
|
2428 | destrevlog._lazydelta = oldlazydelta | |
2425 | destrevlog._lazydeltabase = oldlazydeltabase |
|
2429 | destrevlog._lazydeltabase = oldlazydeltabase | |
2426 | destrevlog._deltabothparents = oldamd |
|
2430 | destrevlog._deltabothparents = oldamd | |
2427 |
|
2431 | |||
2428 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
2432 | def censorrevision(self, tr, censornode, tombstone=b''): | |
2429 | if (self.version & 0xFFFF) == REVLOGV0: |
|
2433 | if (self.version & 0xFFFF) == REVLOGV0: | |
2430 | raise error.RevlogError(_('cannot censor with version %d revlogs') % |
|
2434 | raise error.RevlogError(_('cannot censor with version %d revlogs') % | |
2431 | self.version) |
|
2435 | self.version) | |
2432 |
|
2436 | |||
2433 | censorrev = self.rev(censornode) |
|
2437 | censorrev = self.rev(censornode) | |
2434 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') |
|
2438 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') | |
2435 |
|
2439 | |||
2436 | if len(tombstone) > self.rawsize(censorrev): |
|
2440 | if len(tombstone) > self.rawsize(censorrev): | |
2437 | raise error.Abort(_('censor tombstone must be no longer than ' |
|
2441 | raise error.Abort(_('censor tombstone must be no longer than ' | |
2438 | 'censored data')) |
|
2442 | 'censored data')) | |
2439 |
|
2443 | |||
2440 | # Rewriting the revlog in place is hard. Our strategy for censoring is |
|
2444 | # Rewriting the revlog in place is hard. Our strategy for censoring is | |
2441 | # to create a new revlog, copy all revisions to it, then replace the |
|
2445 | # to create a new revlog, copy all revisions to it, then replace the | |
2442 | # revlogs on transaction close. |
|
2446 | # revlogs on transaction close. | |
2443 |
|
2447 | |||
2444 | newindexfile = self.indexfile + b'.tmpcensored' |
|
2448 | newindexfile = self.indexfile + b'.tmpcensored' | |
2445 | newdatafile = self.datafile + b'.tmpcensored' |
|
2449 | newdatafile = self.datafile + b'.tmpcensored' | |
2446 |
|
2450 | |||
2447 | # This is a bit dangerous. We could easily have a mismatch of state. |
|
2451 | # This is a bit dangerous. We could easily have a mismatch of state. | |
2448 | newrl = revlog(self.opener, newindexfile, newdatafile, |
|
2452 | newrl = revlog(self.opener, newindexfile, newdatafile, | |
2449 | censorable=True) |
|
2453 | censorable=True) | |
2450 | newrl.version = self.version |
|
2454 | newrl.version = self.version | |
2451 | newrl._generaldelta = self._generaldelta |
|
2455 | newrl._generaldelta = self._generaldelta | |
2452 | newrl._io = self._io |
|
2456 | newrl._io = self._io | |
2453 |
|
2457 | |||
2454 | for rev in self.revs(): |
|
2458 | for rev in self.revs(): | |
2455 | node = self.node(rev) |
|
2459 | node = self.node(rev) | |
2456 | p1, p2 = self.parents(node) |
|
2460 | p1, p2 = self.parents(node) | |
2457 |
|
2461 | |||
2458 | if rev == censorrev: |
|
2462 | if rev == censorrev: | |
2459 | newrl.addrawrevision(tombstone, tr, self.linkrev(censorrev), |
|
2463 | newrl.addrawrevision(tombstone, tr, self.linkrev(censorrev), | |
2460 | p1, p2, censornode, REVIDX_ISCENSORED) |
|
2464 | p1, p2, censornode, REVIDX_ISCENSORED) | |
2461 |
|
2465 | |||
2462 | if newrl.deltaparent(rev) != nullrev: |
|
2466 | if newrl.deltaparent(rev) != nullrev: | |
2463 | raise error.Abort(_('censored revision stored as delta; ' |
|
2467 | raise error.Abort(_('censored revision stored as delta; ' | |
2464 | 'cannot censor'), |
|
2468 | 'cannot censor'), | |
2465 | hint=_('censoring of revlogs is not ' |
|
2469 | hint=_('censoring of revlogs is not ' | |
2466 | 'fully implemented; please report ' |
|
2470 | 'fully implemented; please report ' | |
2467 | 'this bug')) |
|
2471 | 'this bug')) | |
2468 | continue |
|
2472 | continue | |
2469 |
|
2473 | |||
2470 | if self.iscensored(rev): |
|
2474 | if self.iscensored(rev): | |
2471 | if self.deltaparent(rev) != nullrev: |
|
2475 | if self.deltaparent(rev) != nullrev: | |
2472 | raise error.Abort(_('cannot censor due to censored ' |
|
2476 | raise error.Abort(_('cannot censor due to censored ' | |
2473 | 'revision having delta stored')) |
|
2477 | 'revision having delta stored')) | |
2474 | rawtext = self._chunk(rev) |
|
2478 | rawtext = self._chunk(rev) | |
2475 | else: |
|
2479 | else: | |
2476 | rawtext = self.rawdata(rev) |
|
2480 | rawtext = self.rawdata(rev) | |
2477 |
|
2481 | |||
2478 | newrl.addrawrevision(rawtext, tr, self.linkrev(rev), p1, p2, node, |
|
2482 | newrl.addrawrevision(rawtext, tr, self.linkrev(rev), p1, p2, node, | |
2479 | self.flags(rev)) |
|
2483 | self.flags(rev)) | |
2480 |
|
2484 | |||
2481 | tr.addbackup(self.indexfile, location='store') |
|
2485 | tr.addbackup(self.indexfile, location='store') | |
2482 | if not self._inline: |
|
2486 | if not self._inline: | |
2483 | tr.addbackup(self.datafile, location='store') |
|
2487 | tr.addbackup(self.datafile, location='store') | |
2484 |
|
2488 | |||
2485 | self.opener.rename(newrl.indexfile, self.indexfile) |
|
2489 | self.opener.rename(newrl.indexfile, self.indexfile) | |
2486 | if not self._inline: |
|
2490 | if not self._inline: | |
2487 | self.opener.rename(newrl.datafile, self.datafile) |
|
2491 | self.opener.rename(newrl.datafile, self.datafile) | |
2488 |
|
2492 | |||
2489 | self.clearcaches() |
|
2493 | self.clearcaches() | |
2490 | self._loadindex() |
|
2494 | self._loadindex() | |
2491 |
|
2495 | |||
2492 | def verifyintegrity(self, state): |
|
2496 | def verifyintegrity(self, state): | |
2493 | """Verifies the integrity of the revlog. |
|
2497 | """Verifies the integrity of the revlog. | |
2494 |
|
2498 | |||
2495 | Yields ``revlogproblem`` instances describing problems that are |
|
2499 | Yields ``revlogproblem`` instances describing problems that are | |
2496 | found. |
|
2500 | found. | |
2497 | """ |
|
2501 | """ | |
2498 | dd, di = self.checksize() |
|
2502 | dd, di = self.checksize() | |
2499 | if dd: |
|
2503 | if dd: | |
2500 | yield revlogproblem(error=_('data length off by %d bytes') % dd) |
|
2504 | yield revlogproblem(error=_('data length off by %d bytes') % dd) | |
2501 | if di: |
|
2505 | if di: | |
2502 | yield revlogproblem(error=_('index contains %d extra bytes') % di) |
|
2506 | yield revlogproblem(error=_('index contains %d extra bytes') % di) | |
2503 |
|
2507 | |||
2504 | version = self.version & 0xFFFF |
|
2508 | version = self.version & 0xFFFF | |
2505 |
|
2509 | |||
2506 | # The verifier tells us what version revlog we should be. |
|
2510 | # The verifier tells us what version revlog we should be. | |
2507 | if version != state['expectedversion']: |
|
2511 | if version != state['expectedversion']: | |
2508 | yield revlogproblem( |
|
2512 | yield revlogproblem( | |
2509 | warning=_("warning: '%s' uses revlog format %d; expected %d") % |
|
2513 | warning=_("warning: '%s' uses revlog format %d; expected %d") % | |
2510 | (self.indexfile, version, state['expectedversion'])) |
|
2514 | (self.indexfile, version, state['expectedversion'])) | |
2511 |
|
2515 | |||
2512 | state['skipread'] = set() |
|
2516 | state['skipread'] = set() | |
2513 |
|
2517 | |||
2514 | for rev in self: |
|
2518 | for rev in self: | |
2515 | node = self.node(rev) |
|
2519 | node = self.node(rev) | |
2516 |
|
2520 | |||
2517 | # Verify contents. 4 cases to care about: |
|
2521 | # Verify contents. 4 cases to care about: | |
2518 | # |
|
2522 | # | |
2519 | # common: the most common case |
|
2523 | # common: the most common case | |
2520 | # rename: with a rename |
|
2524 | # rename: with a rename | |
2521 | # meta: file content starts with b'\1\n', the metadata |
|
2525 | # meta: file content starts with b'\1\n', the metadata | |
2522 | # header defined in filelog.py, but without a rename |
|
2526 | # header defined in filelog.py, but without a rename | |
2523 | # ext: content stored externally |
|
2527 | # ext: content stored externally | |
2524 | # |
|
2528 | # | |
2525 | # More formally, their differences are shown below: |
|
2529 | # More formally, their differences are shown below: | |
2526 | # |
|
2530 | # | |
2527 | # | common | rename | meta | ext |
|
2531 | # | common | rename | meta | ext | |
2528 | # ------------------------------------------------------- |
|
2532 | # ------------------------------------------------------- | |
2529 | # flags() | 0 | 0 | 0 | not 0 |
|
2533 | # flags() | 0 | 0 | 0 | not 0 | |
2530 | # renamed() | False | True | False | ? |
|
2534 | # renamed() | False | True | False | ? | |
2531 | # rawtext[0:2]=='\1\n'| False | True | True | ? |
|
2535 | # rawtext[0:2]=='\1\n'| False | True | True | ? | |
2532 | # |
|
2536 | # | |
2533 | # "rawtext" means the raw text stored in revlog data, which |
|
2537 | # "rawtext" means the raw text stored in revlog data, which | |
2534 | # could be retrieved by "rawdata(rev)". "text" |
|
2538 | # could be retrieved by "rawdata(rev)". "text" | |
2535 | # mentioned below is "revision(rev)". |
|
2539 | # mentioned below is "revision(rev)". | |
2536 | # |
|
2540 | # | |
2537 | # There are 3 different lengths stored physically: |
|
2541 | # There are 3 different lengths stored physically: | |
2538 | # 1. L1: rawsize, stored in revlog index |
|
2542 | # 1. L1: rawsize, stored in revlog index | |
2539 | # 2. L2: len(rawtext), stored in revlog data |
|
2543 | # 2. L2: len(rawtext), stored in revlog data | |
2540 | # 3. L3: len(text), stored in revlog data if flags==0, or |
|
2544 | # 3. L3: len(text), stored in revlog data if flags==0, or | |
2541 | # possibly somewhere else if flags!=0 |
|
2545 | # possibly somewhere else if flags!=0 | |
2542 | # |
|
2546 | # | |
2543 | # L1 should be equal to L2. L3 could be different from them. |
|
2547 | # L1 should be equal to L2. L3 could be different from them. | |
2544 | # "text" may or may not affect commit hash depending on flag |
|
2548 | # "text" may or may not affect commit hash depending on flag | |
2545 | # processors (see flagutil.addflagprocessor). |
|
2549 | # processors (see flagutil.addflagprocessor). | |
2546 | # |
|
2550 | # | |
2547 | # | common | rename | meta | ext |
|
2551 | # | common | rename | meta | ext | |
2548 | # ------------------------------------------------- |
|
2552 | # ------------------------------------------------- | |
2549 | # rawsize() | L1 | L1 | L1 | L1 |
|
2553 | # rawsize() | L1 | L1 | L1 | L1 | |
2550 | # size() | L1 | L2-LM | L1(*) | L1 (?) |
|
2554 | # size() | L1 | L2-LM | L1(*) | L1 (?) | |
2551 | # len(rawtext) | L2 | L2 | L2 | L2 |
|
2555 | # len(rawtext) | L2 | L2 | L2 | L2 | |
2552 | # len(text) | L2 | L2 | L2 | L3 |
|
2556 | # len(text) | L2 | L2 | L2 | L3 | |
2553 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) |
|
2557 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) | |
2554 | # |
|
2558 | # | |
2555 | # LM: length of metadata, depending on rawtext |
|
2559 | # LM: length of metadata, depending on rawtext | |
2556 | # (*): not ideal, see comment in filelog.size |
|
2560 | # (*): not ideal, see comment in filelog.size | |
2557 | # (?): could be "- len(meta)" if the resolved content has |
|
2561 | # (?): could be "- len(meta)" if the resolved content has | |
2558 | # rename metadata |
|
2562 | # rename metadata | |
2559 | # |
|
2563 | # | |
2560 | # Checks needed to be done: |
|
2564 | # Checks needed to be done: | |
2561 | # 1. length check: L1 == L2, in all cases. |
|
2565 | # 1. length check: L1 == L2, in all cases. | |
2562 | # 2. hash check: depending on flag processor, we may need to |
|
2566 | # 2. hash check: depending on flag processor, we may need to | |
2563 | # use either "text" (external), or "rawtext" (in revlog). |
|
2567 | # use either "text" (external), or "rawtext" (in revlog). | |
2564 |
|
2568 | |||
2565 | try: |
|
2569 | try: | |
2566 | skipflags = state.get('skipflags', 0) |
|
2570 | skipflags = state.get('skipflags', 0) | |
2567 | if skipflags: |
|
2571 | if skipflags: | |
2568 | skipflags &= self.flags(rev) |
|
2572 | skipflags &= self.flags(rev) | |
2569 |
|
2573 | |||
2570 | if skipflags: |
|
2574 | if skipflags: | |
2571 | state['skipread'].add(node) |
|
2575 | state['skipread'].add(node) | |
2572 | else: |
|
2576 | else: | |
2573 | # Side-effect: read content and verify hash. |
|
2577 | # Side-effect: read content and verify hash. | |
2574 | self.revision(node) |
|
2578 | self.revision(node) | |
2575 |
|
2579 | |||
2576 | l1 = self.rawsize(rev) |
|
2580 | l1 = self.rawsize(rev) | |
2577 | l2 = len(self.rawdata(node)) |
|
2581 | l2 = len(self.rawdata(node)) | |
2578 |
|
2582 | |||
2579 | if l1 != l2: |
|
2583 | if l1 != l2: | |
2580 | yield revlogproblem( |
|
2584 | yield revlogproblem( | |
2581 | error=_('unpacked size is %d, %d expected') % (l2, l1), |
|
2585 | error=_('unpacked size is %d, %d expected') % (l2, l1), | |
2582 | node=node) |
|
2586 | node=node) | |
2583 |
|
2587 | |||
2584 | except error.CensoredNodeError: |
|
2588 | except error.CensoredNodeError: | |
2585 | if state['erroroncensored']: |
|
2589 | if state['erroroncensored']: | |
2586 | yield revlogproblem(error=_('censored file data'), |
|
2590 | yield revlogproblem(error=_('censored file data'), | |
2587 | node=node) |
|
2591 | node=node) | |
2588 | state['skipread'].add(node) |
|
2592 | state['skipread'].add(node) | |
2589 | except Exception as e: |
|
2593 | except Exception as e: | |
2590 | yield revlogproblem( |
|
2594 | yield revlogproblem( | |
2591 | error=_('unpacking %s: %s') % (short(node), |
|
2595 | error=_('unpacking %s: %s') % (short(node), | |
2592 | stringutil.forcebytestr(e)), |
|
2596 | stringutil.forcebytestr(e)), | |
2593 | node=node) |
|
2597 | node=node) | |
2594 | state['skipread'].add(node) |
|
2598 | state['skipread'].add(node) | |
2595 |
|
2599 | |||
2596 | def storageinfo(self, exclusivefiles=False, sharedfiles=False, |
|
2600 | def storageinfo(self, exclusivefiles=False, sharedfiles=False, | |
2597 | revisionscount=False, trackedsize=False, |
|
2601 | revisionscount=False, trackedsize=False, | |
2598 | storedsize=False): |
|
2602 | storedsize=False): | |
2599 | d = {} |
|
2603 | d = {} | |
2600 |
|
2604 | |||
2601 | if exclusivefiles: |
|
2605 | if exclusivefiles: | |
2602 | d['exclusivefiles'] = [(self.opener, self.indexfile)] |
|
2606 | d['exclusivefiles'] = [(self.opener, self.indexfile)] | |
2603 | if not self._inline: |
|
2607 | if not self._inline: | |
2604 | d['exclusivefiles'].append((self.opener, self.datafile)) |
|
2608 | d['exclusivefiles'].append((self.opener, self.datafile)) | |
2605 |
|
2609 | |||
2606 | if sharedfiles: |
|
2610 | if sharedfiles: | |
2607 | d['sharedfiles'] = [] |
|
2611 | d['sharedfiles'] = [] | |
2608 |
|
2612 | |||
2609 | if revisionscount: |
|
2613 | if revisionscount: | |
2610 | d['revisionscount'] = len(self) |
|
2614 | d['revisionscount'] = len(self) | |
2611 |
|
2615 | |||
2612 | if trackedsize: |
|
2616 | if trackedsize: | |
2613 | d['trackedsize'] = sum(map(self.rawsize, iter(self))) |
|
2617 | d['trackedsize'] = sum(map(self.rawsize, iter(self))) | |
2614 |
|
2618 | |||
2615 | if storedsize: |
|
2619 | if storedsize: | |
2616 | d['storedsize'] = sum(self.opener.stat(path).st_size |
|
2620 | d['storedsize'] = sum(self.opener.stat(path).st_size | |
2617 | for path in self.files()) |
|
2621 | for path in self.files()) | |
2618 |
|
2622 | |||
2619 | return d |
|
2623 | return d |
General Comments 0
You need to be logged in to leave comments.
Login now