Show More
@@ -1,542 +1,546 b'' | |||
|
1 | 1 | # wrapper.py - methods wrapping core mercurial logic |
|
2 | 2 | # |
|
3 | 3 | # Copyright 2017 Facebook, Inc. |
|
4 | 4 | # |
|
5 | 5 | # This software may be used and distributed according to the terms of the |
|
6 | 6 | # GNU General Public License version 2 or any later version. |
|
7 | 7 | |
|
8 | 8 | from __future__ import absolute_import |
|
9 | 9 | |
|
10 | 10 | import hashlib |
|
11 | 11 | |
|
12 | 12 | from mercurial.i18n import _ |
|
13 | 13 | from mercurial.node import bin, hex, nullid, short |
|
14 | 14 | from mercurial.pycompat import ( |
|
15 | 15 | getattr, |
|
16 | 16 | setattr, |
|
17 | 17 | ) |
|
18 | 18 | |
|
19 | 19 | from mercurial import ( |
|
20 | 20 | bundle2, |
|
21 | 21 | changegroup, |
|
22 | 22 | cmdutil, |
|
23 | 23 | context, |
|
24 | 24 | error, |
|
25 | 25 | exchange, |
|
26 | 26 | exthelper, |
|
27 | 27 | localrepo, |
|
28 | 28 | pycompat, |
|
29 | 29 | revlog, |
|
30 | 30 | scmutil, |
|
31 | 31 | upgrade, |
|
32 | 32 | util, |
|
33 | 33 | vfs as vfsmod, |
|
34 | 34 | wireprotov1server, |
|
35 | 35 | ) |
|
36 | 36 | |
|
37 | 37 | from mercurial.interfaces import repository |
|
38 | 38 | |
|
39 | 39 | from mercurial.utils import ( |
|
40 | 40 | storageutil, |
|
41 | 41 | stringutil, |
|
42 | 42 | ) |
|
43 | 43 | |
|
44 | 44 | from ..largefiles import lfutil |
|
45 | 45 | |
|
46 | 46 | from . import ( |
|
47 | 47 | blobstore, |
|
48 | 48 | pointer, |
|
49 | 49 | ) |
|
50 | 50 | |
|
51 | 51 | eh = exthelper.exthelper() |
|
52 | 52 | |
|
53 | 53 | |
|
54 | 54 | @eh.wrapfunction(localrepo, b'makefilestorage') |
|
55 | 55 | def localrepomakefilestorage(orig, requirements, features, **kwargs): |
|
56 | 56 | if b'lfs' in requirements: |
|
57 | 57 | features.add(repository.REPO_FEATURE_LFS) |
|
58 | 58 | |
|
59 | 59 | return orig(requirements=requirements, features=features, **kwargs) |
|
60 | 60 | |
|
61 | 61 | |
|
62 | 62 | @eh.wrapfunction(changegroup, b'allsupportedversions') |
|
63 | 63 | def allsupportedversions(orig, ui): |
|
64 | 64 | versions = orig(ui) |
|
65 | 65 | versions.add(b'03') |
|
66 | 66 | return versions |
|
67 | 67 | |
|
68 | 68 | |
|
69 | 69 | @eh.wrapfunction(wireprotov1server, b'_capabilities') |
|
70 | 70 | def _capabilities(orig, repo, proto): |
|
71 | 71 | '''Wrap server command to announce lfs server capability''' |
|
72 | 72 | caps = orig(repo, proto) |
|
73 | 73 | if util.safehasattr(repo.svfs, b'lfslocalblobstore'): |
|
74 | 74 | # Advertise a slightly different capability when lfs is *required*, so |
|
75 | 75 | # that the client knows it MUST load the extension. If lfs is not |
|
76 | 76 | # required on the server, there's no reason to autoload the extension |
|
77 | 77 | # on the client. |
|
78 | 78 | if b'lfs' in repo.requirements: |
|
79 | 79 | caps.append(b'lfs-serve') |
|
80 | 80 | |
|
81 | 81 | caps.append(b'lfs') |
|
82 | 82 | return caps |
|
83 | 83 | |
|
84 | 84 | |
|
85 | 85 | def bypasscheckhash(self, text): |
|
86 | 86 | return False |
|
87 | 87 | |
|
88 | 88 | |
|
89 | 89 | def readfromstore(self, text): |
|
90 | 90 | """Read filelog content from local blobstore transform for flagprocessor. |
|
91 | 91 | |
|
92 | 92 | Default tranform for flagprocessor, returning contents from blobstore. |
|
93 | 93 | Returns a 2-typle (text, validatehash) where validatehash is True as the |
|
94 | 94 | contents of the blobstore should be checked using checkhash. |
|
95 | 95 | """ |
|
96 | 96 | p = pointer.deserialize(text) |
|
97 | 97 | oid = p.oid() |
|
98 | 98 | store = self.opener.lfslocalblobstore |
|
99 | 99 | if not store.has(oid): |
|
100 | 100 | p.filename = self.filename |
|
101 | 101 | self.opener.lfsremoteblobstore.readbatch([p], store) |
|
102 | 102 | |
|
103 | 103 | # The caller will validate the content |
|
104 | 104 | text = store.read(oid, verify=False) |
|
105 | 105 | |
|
106 | 106 | # pack hg filelog metadata |
|
107 | 107 | hgmeta = {} |
|
108 | 108 | for k in p.keys(): |
|
109 | 109 | if k.startswith(b'x-hg-'): |
|
110 | 110 | name = k[len(b'x-hg-') :] |
|
111 | 111 | hgmeta[name] = p[k] |
|
112 | 112 | if hgmeta or text.startswith(b'\1\n'): |
|
113 | 113 | text = storageutil.packmeta(hgmeta, text) |
|
114 | 114 | |
|
115 | 115 | return (text, True, {}) |
|
116 | 116 | |
|
117 | 117 | |
|
118 | 118 | def writetostore(self, text, sidedata): |
|
119 | 119 | # hg filelog metadata (includes rename, etc) |
|
120 | 120 | hgmeta, offset = storageutil.parsemeta(text) |
|
121 | 121 | if offset and offset > 0: |
|
122 | 122 | # lfs blob does not contain hg filelog metadata |
|
123 | 123 | text = text[offset:] |
|
124 | 124 | |
|
125 | 125 | # git-lfs only supports sha256 |
|
126 | 126 | oid = hex(hashlib.sha256(text).digest()) |
|
127 | 127 | self.opener.lfslocalblobstore.write(oid, text) |
|
128 | 128 | |
|
129 | 129 | # replace contents with metadata |
|
130 | 130 | longoid = b'sha256:%s' % oid |
|
131 | 131 | metadata = pointer.gitlfspointer(oid=longoid, size=b'%d' % len(text)) |
|
132 | 132 | |
|
133 | 133 | # by default, we expect the content to be binary. however, LFS could also |
|
134 | 134 | # be used for non-binary content. add a special entry for non-binary data. |
|
135 | 135 | # this will be used by filectx.isbinary(). |
|
136 | 136 | if not stringutil.binary(text): |
|
137 | 137 | # not hg filelog metadata (affecting commit hash), no "x-hg-" prefix |
|
138 | 138 | metadata[b'x-is-binary'] = b'0' |
|
139 | 139 | |
|
140 | 140 | # translate hg filelog metadata to lfs metadata with "x-hg-" prefix |
|
141 | 141 | if hgmeta is not None: |
|
142 | 142 | for k, v in pycompat.iteritems(hgmeta): |
|
143 | 143 | metadata[b'x-hg-%s' % k] = v |
|
144 | 144 | |
|
145 | 145 | rawtext = metadata.serialize() |
|
146 | 146 | return (rawtext, False) |
|
147 | 147 | |
|
148 | 148 | |
|
149 | 149 | def _islfs(rlog, node=None, rev=None): |
|
150 | 150 | if rev is None: |
|
151 | 151 | if node is None: |
|
152 | 152 | # both None - likely working copy content where node is not ready |
|
153 | 153 | return False |
|
154 | 154 | rev = rlog.rev(node) |
|
155 | 155 | else: |
|
156 | 156 | node = rlog.node(rev) |
|
157 | 157 | if node == nullid: |
|
158 | 158 | return False |
|
159 | 159 | flags = rlog.flags(rev) |
|
160 | 160 | return bool(flags & revlog.REVIDX_EXTSTORED) |
|
161 | 161 | |
|
162 | 162 | |
|
163 | 163 | # Wrapping may also be applied by remotefilelog |
|
164 | 164 | def filelogaddrevision( |
|
165 | 165 | orig, |
|
166 | 166 | self, |
|
167 | 167 | text, |
|
168 | 168 | transaction, |
|
169 | 169 | link, |
|
170 | 170 | p1, |
|
171 | 171 | p2, |
|
172 | 172 | cachedelta=None, |
|
173 | 173 | node=None, |
|
174 | 174 | flags=revlog.REVIDX_DEFAULT_FLAGS, |
|
175 | 175 | **kwds |
|
176 | 176 | ): |
|
177 | 177 | # The matcher isn't available if reposetup() wasn't called. |
|
178 | 178 | lfstrack = self._revlog.opener.options.get(b'lfstrack') |
|
179 | 179 | |
|
180 | 180 | if lfstrack: |
|
181 | 181 | textlen = len(text) |
|
182 | 182 | # exclude hg rename meta from file size |
|
183 | 183 | meta, offset = storageutil.parsemeta(text) |
|
184 | 184 | if offset: |
|
185 | 185 | textlen -= offset |
|
186 | 186 | |
|
187 | 187 | if lfstrack(self._revlog.filename, textlen): |
|
188 | 188 | flags |= revlog.REVIDX_EXTSTORED |
|
189 | 189 | |
|
190 | 190 | return orig( |
|
191 | 191 | self, |
|
192 | 192 | text, |
|
193 | 193 | transaction, |
|
194 | 194 | link, |
|
195 | 195 | p1, |
|
196 | 196 | p2, |
|
197 | 197 | cachedelta=cachedelta, |
|
198 | 198 | node=node, |
|
199 | 199 | flags=flags, |
|
200 | 200 | **kwds |
|
201 | 201 | ) |
|
202 | 202 | |
|
203 | 203 | |
|
204 | 204 | # Wrapping may also be applied by remotefilelog |
|
205 | 205 | def filelogrenamed(orig, self, node): |
|
206 | 206 | if _islfs(self._revlog, node): |
|
207 | 207 | rawtext = self._revlog.rawdata(node) |
|
208 | 208 | if not rawtext: |
|
209 | 209 | return False |
|
210 | 210 | metadata = pointer.deserialize(rawtext) |
|
211 | 211 | if b'x-hg-copy' in metadata and b'x-hg-copyrev' in metadata: |
|
212 | 212 | return metadata[b'x-hg-copy'], bin(metadata[b'x-hg-copyrev']) |
|
213 | 213 | else: |
|
214 | 214 | return False |
|
215 | 215 | return orig(self, node) |
|
216 | 216 | |
|
217 | 217 | |
|
218 | 218 | # Wrapping may also be applied by remotefilelog |
|
219 | 219 | def filelogsize(orig, self, rev): |
|
220 | 220 | if _islfs(self._revlog, rev=rev): |
|
221 | 221 | # fast path: use lfs metadata to answer size |
|
222 | 222 | rawtext = self._revlog.rawdata(rev) |
|
223 | 223 | metadata = pointer.deserialize(rawtext) |
|
224 | 224 | return int(metadata[b'size']) |
|
225 | 225 | return orig(self, rev) |
|
226 | 226 | |
|
227 | 227 | |
|
228 | 228 | @eh.wrapfunction(revlog, b'_verify_revision') |
|
229 | 229 | def _verify_revision(orig, rl, skipflags, state, node): |
|
230 | 230 | if _islfs(rl, node=node): |
|
231 | 231 | rawtext = rl.rawdata(node) |
|
232 | 232 | metadata = pointer.deserialize(rawtext) |
|
233 | 233 | |
|
234 | 234 | # Don't skip blobs that are stored locally, as local verification is |
|
235 | 235 | # relatively cheap and there's no other way to verify the raw data in |
|
236 | 236 | # the revlog. |
|
237 | 237 | if rl.opener.lfslocalblobstore.has(metadata.oid()): |
|
238 | 238 | skipflags &= ~revlog.REVIDX_EXTSTORED |
|
239 | elif skipflags & revlog.REVIDX_EXTSTORED: | |
|
240 | # The wrapped method will set `skipread`, but there's enough local | |
|
241 | # info to check renames. | |
|
242 | state[b'safe_renamed'].add(node) | |
|
239 | 243 | |
|
240 | 244 | orig(rl, skipflags, state, node) |
|
241 | 245 | |
|
242 | 246 | |
|
243 | 247 | @eh.wrapfunction(context.basefilectx, b'cmp') |
|
244 | 248 | def filectxcmp(orig, self, fctx): |
|
245 | 249 | """returns True if text is different than fctx""" |
|
246 | 250 | # some fctx (ex. hg-git) is not based on basefilectx and do not have islfs |
|
247 | 251 | if self.islfs() and getattr(fctx, 'islfs', lambda: False)(): |
|
248 | 252 | # fast path: check LFS oid |
|
249 | 253 | p1 = pointer.deserialize(self.rawdata()) |
|
250 | 254 | p2 = pointer.deserialize(fctx.rawdata()) |
|
251 | 255 | return p1.oid() != p2.oid() |
|
252 | 256 | return orig(self, fctx) |
|
253 | 257 | |
|
254 | 258 | |
|
255 | 259 | @eh.wrapfunction(context.basefilectx, b'isbinary') |
|
256 | 260 | def filectxisbinary(orig, self): |
|
257 | 261 | if self.islfs(): |
|
258 | 262 | # fast path: use lfs metadata to answer isbinary |
|
259 | 263 | metadata = pointer.deserialize(self.rawdata()) |
|
260 | 264 | # if lfs metadata says nothing, assume it's binary by default |
|
261 | 265 | return bool(int(metadata.get(b'x-is-binary', 1))) |
|
262 | 266 | return orig(self) |
|
263 | 267 | |
|
264 | 268 | |
|
265 | 269 | def filectxislfs(self): |
|
266 | 270 | return _islfs(self.filelog()._revlog, self.filenode()) |
|
267 | 271 | |
|
268 | 272 | |
|
269 | 273 | @eh.wrapfunction(cmdutil, b'_updatecatformatter') |
|
270 | 274 | def _updatecatformatter(orig, fm, ctx, matcher, path, decode): |
|
271 | 275 | orig(fm, ctx, matcher, path, decode) |
|
272 | 276 | fm.data(rawdata=ctx[path].rawdata()) |
|
273 | 277 | |
|
274 | 278 | |
|
275 | 279 | @eh.wrapfunction(scmutil, b'wrapconvertsink') |
|
276 | 280 | def convertsink(orig, sink): |
|
277 | 281 | sink = orig(sink) |
|
278 | 282 | if sink.repotype == b'hg': |
|
279 | 283 | |
|
280 | 284 | class lfssink(sink.__class__): |
|
281 | 285 | def putcommit( |
|
282 | 286 | self, |
|
283 | 287 | files, |
|
284 | 288 | copies, |
|
285 | 289 | parents, |
|
286 | 290 | commit, |
|
287 | 291 | source, |
|
288 | 292 | revmap, |
|
289 | 293 | full, |
|
290 | 294 | cleanp2, |
|
291 | 295 | ): |
|
292 | 296 | pc = super(lfssink, self).putcommit |
|
293 | 297 | node = pc( |
|
294 | 298 | files, |
|
295 | 299 | copies, |
|
296 | 300 | parents, |
|
297 | 301 | commit, |
|
298 | 302 | source, |
|
299 | 303 | revmap, |
|
300 | 304 | full, |
|
301 | 305 | cleanp2, |
|
302 | 306 | ) |
|
303 | 307 | |
|
304 | 308 | if b'lfs' not in self.repo.requirements: |
|
305 | 309 | ctx = self.repo[node] |
|
306 | 310 | |
|
307 | 311 | # The file list may contain removed files, so check for |
|
308 | 312 | # membership before assuming it is in the context. |
|
309 | 313 | if any(f in ctx and ctx[f].islfs() for f, n in files): |
|
310 | 314 | self.repo.requirements.add(b'lfs') |
|
311 | 315 | self.repo._writerequirements() |
|
312 | 316 | |
|
313 | 317 | return node |
|
314 | 318 | |
|
315 | 319 | sink.__class__ = lfssink |
|
316 | 320 | |
|
317 | 321 | return sink |
|
318 | 322 | |
|
319 | 323 | |
|
320 | 324 | # bundlerepo uses "vfsmod.readonlyvfs(othervfs)", we need to make sure lfs |
|
321 | 325 | # options and blob stores are passed from othervfs to the new readonlyvfs. |
|
322 | 326 | @eh.wrapfunction(vfsmod.readonlyvfs, b'__init__') |
|
323 | 327 | def vfsinit(orig, self, othervfs): |
|
324 | 328 | orig(self, othervfs) |
|
325 | 329 | # copy lfs related options |
|
326 | 330 | for k, v in othervfs.options.items(): |
|
327 | 331 | if k.startswith(b'lfs'): |
|
328 | 332 | self.options[k] = v |
|
329 | 333 | # also copy lfs blobstores. note: this can run before reposetup, so lfs |
|
330 | 334 | # blobstore attributes are not always ready at this time. |
|
331 | 335 | for name in [b'lfslocalblobstore', b'lfsremoteblobstore']: |
|
332 | 336 | if util.safehasattr(othervfs, name): |
|
333 | 337 | setattr(self, name, getattr(othervfs, name)) |
|
334 | 338 | |
|
335 | 339 | |
|
336 | 340 | def _prefetchfiles(repo, revs, match): |
|
337 | 341 | """Ensure that required LFS blobs are present, fetching them as a group if |
|
338 | 342 | needed.""" |
|
339 | 343 | if not util.safehasattr(repo.svfs, b'lfslocalblobstore'): |
|
340 | 344 | return |
|
341 | 345 | |
|
342 | 346 | pointers = [] |
|
343 | 347 | oids = set() |
|
344 | 348 | localstore = repo.svfs.lfslocalblobstore |
|
345 | 349 | |
|
346 | 350 | for rev in revs: |
|
347 | 351 | ctx = repo[rev] |
|
348 | 352 | for f in ctx.walk(match): |
|
349 | 353 | p = pointerfromctx(ctx, f) |
|
350 | 354 | if p and p.oid() not in oids and not localstore.has(p.oid()): |
|
351 | 355 | p.filename = f |
|
352 | 356 | pointers.append(p) |
|
353 | 357 | oids.add(p.oid()) |
|
354 | 358 | |
|
355 | 359 | if pointers: |
|
356 | 360 | # Recalculating the repo store here allows 'paths.default' that is set |
|
357 | 361 | # on the repo by a clone command to be used for the update. |
|
358 | 362 | blobstore.remote(repo).readbatch(pointers, localstore) |
|
359 | 363 | |
|
360 | 364 | |
|
361 | 365 | def _canskipupload(repo): |
|
362 | 366 | # Skip if this hasn't been passed to reposetup() |
|
363 | 367 | if not util.safehasattr(repo.svfs, b'lfsremoteblobstore'): |
|
364 | 368 | return True |
|
365 | 369 | |
|
366 | 370 | # if remotestore is a null store, upload is a no-op and can be skipped |
|
367 | 371 | return isinstance(repo.svfs.lfsremoteblobstore, blobstore._nullremote) |
|
368 | 372 | |
|
369 | 373 | |
|
370 | 374 | def candownload(repo): |
|
371 | 375 | # Skip if this hasn't been passed to reposetup() |
|
372 | 376 | if not util.safehasattr(repo.svfs, b'lfsremoteblobstore'): |
|
373 | 377 | return False |
|
374 | 378 | |
|
375 | 379 | # if remotestore is a null store, downloads will lead to nothing |
|
376 | 380 | return not isinstance(repo.svfs.lfsremoteblobstore, blobstore._nullremote) |
|
377 | 381 | |
|
378 | 382 | |
|
379 | 383 | def uploadblobsfromrevs(repo, revs): |
|
380 | 384 | '''upload lfs blobs introduced by revs |
|
381 | 385 | |
|
382 | 386 | Note: also used by other extensions e. g. infinitepush. avoid renaming. |
|
383 | 387 | ''' |
|
384 | 388 | if _canskipupload(repo): |
|
385 | 389 | return |
|
386 | 390 | pointers = extractpointers(repo, revs) |
|
387 | 391 | uploadblobs(repo, pointers) |
|
388 | 392 | |
|
389 | 393 | |
|
390 | 394 | def prepush(pushop): |
|
391 | 395 | """Prepush hook. |
|
392 | 396 | |
|
393 | 397 | Read through the revisions to push, looking for filelog entries that can be |
|
394 | 398 | deserialized into metadata so that we can block the push on their upload to |
|
395 | 399 | the remote blobstore. |
|
396 | 400 | """ |
|
397 | 401 | return uploadblobsfromrevs(pushop.repo, pushop.outgoing.missing) |
|
398 | 402 | |
|
399 | 403 | |
|
400 | 404 | @eh.wrapfunction(exchange, b'push') |
|
401 | 405 | def push(orig, repo, remote, *args, **kwargs): |
|
402 | 406 | """bail on push if the extension isn't enabled on remote when needed, and |
|
403 | 407 | update the remote store based on the destination path.""" |
|
404 | 408 | if b'lfs' in repo.requirements: |
|
405 | 409 | # If the remote peer is for a local repo, the requirement tests in the |
|
406 | 410 | # base class method enforce lfs support. Otherwise, some revisions in |
|
407 | 411 | # this repo use lfs, and the remote repo needs the extension loaded. |
|
408 | 412 | if not remote.local() and not remote.capable(b'lfs'): |
|
409 | 413 | # This is a copy of the message in exchange.push() when requirements |
|
410 | 414 | # are missing between local repos. |
|
411 | 415 | m = _(b"required features are not supported in the destination: %s") |
|
412 | 416 | raise error.Abort( |
|
413 | 417 | m % b'lfs', hint=_(b'enable the lfs extension on the server') |
|
414 | 418 | ) |
|
415 | 419 | |
|
416 | 420 | # Repositories where this extension is disabled won't have the field. |
|
417 | 421 | # But if there's a requirement, then the extension must be loaded AND |
|
418 | 422 | # there may be blobs to push. |
|
419 | 423 | remotestore = repo.svfs.lfsremoteblobstore |
|
420 | 424 | try: |
|
421 | 425 | repo.svfs.lfsremoteblobstore = blobstore.remote(repo, remote.url()) |
|
422 | 426 | return orig(repo, remote, *args, **kwargs) |
|
423 | 427 | finally: |
|
424 | 428 | repo.svfs.lfsremoteblobstore = remotestore |
|
425 | 429 | else: |
|
426 | 430 | return orig(repo, remote, *args, **kwargs) |
|
427 | 431 | |
|
428 | 432 | |
|
429 | 433 | # when writing a bundle via "hg bundle" command, upload related LFS blobs |
|
430 | 434 | @eh.wrapfunction(bundle2, b'writenewbundle') |
|
431 | 435 | def writenewbundle( |
|
432 | 436 | orig, ui, repo, source, filename, bundletype, outgoing, *args, **kwargs |
|
433 | 437 | ): |
|
434 | 438 | """upload LFS blobs added by outgoing revisions on 'hg bundle'""" |
|
435 | 439 | uploadblobsfromrevs(repo, outgoing.missing) |
|
436 | 440 | return orig( |
|
437 | 441 | ui, repo, source, filename, bundletype, outgoing, *args, **kwargs |
|
438 | 442 | ) |
|
439 | 443 | |
|
440 | 444 | |
|
441 | 445 | def extractpointers(repo, revs): |
|
442 | 446 | """return a list of lfs pointers added by given revs""" |
|
443 | 447 | repo.ui.debug(b'lfs: computing set of blobs to upload\n') |
|
444 | 448 | pointers = {} |
|
445 | 449 | |
|
446 | 450 | makeprogress = repo.ui.makeprogress |
|
447 | 451 | with makeprogress( |
|
448 | 452 | _(b'lfs search'), _(b'changesets'), len(revs) |
|
449 | 453 | ) as progress: |
|
450 | 454 | for r in revs: |
|
451 | 455 | ctx = repo[r] |
|
452 | 456 | for p in pointersfromctx(ctx).values(): |
|
453 | 457 | pointers[p.oid()] = p |
|
454 | 458 | progress.increment() |
|
455 | 459 | return sorted(pointers.values(), key=lambda p: p.oid()) |
|
456 | 460 | |
|
457 | 461 | |
|
458 | 462 | def pointerfromctx(ctx, f, removed=False): |
|
459 | 463 | """return a pointer for the named file from the given changectx, or None if |
|
460 | 464 | the file isn't LFS. |
|
461 | 465 | |
|
462 | 466 | Optionally, the pointer for a file deleted from the context can be returned. |
|
463 | 467 | Since no such pointer is actually stored, and to distinguish from a non LFS |
|
464 | 468 | file, this pointer is represented by an empty dict. |
|
465 | 469 | """ |
|
466 | 470 | _ctx = ctx |
|
467 | 471 | if f not in ctx: |
|
468 | 472 | if not removed: |
|
469 | 473 | return None |
|
470 | 474 | if f in ctx.p1(): |
|
471 | 475 | _ctx = ctx.p1() |
|
472 | 476 | elif f in ctx.p2(): |
|
473 | 477 | _ctx = ctx.p2() |
|
474 | 478 | else: |
|
475 | 479 | return None |
|
476 | 480 | fctx = _ctx[f] |
|
477 | 481 | if not _islfs(fctx.filelog()._revlog, fctx.filenode()): |
|
478 | 482 | return None |
|
479 | 483 | try: |
|
480 | 484 | p = pointer.deserialize(fctx.rawdata()) |
|
481 | 485 | if ctx == _ctx: |
|
482 | 486 | return p |
|
483 | 487 | return {} |
|
484 | 488 | except pointer.InvalidPointer as ex: |
|
485 | 489 | raise error.Abort( |
|
486 | 490 | _(b'lfs: corrupted pointer (%s@%s): %s\n') |
|
487 | 491 | % (f, short(_ctx.node()), ex) |
|
488 | 492 | ) |
|
489 | 493 | |
|
490 | 494 | |
|
491 | 495 | def pointersfromctx(ctx, removed=False): |
|
492 | 496 | """return a dict {path: pointer} for given single changectx. |
|
493 | 497 | |
|
494 | 498 | If ``removed`` == True and the LFS file was removed from ``ctx``, the value |
|
495 | 499 | stored for the path is an empty dict. |
|
496 | 500 | """ |
|
497 | 501 | result = {} |
|
498 | 502 | m = ctx.repo().narrowmatch() |
|
499 | 503 | |
|
500 | 504 | # TODO: consider manifest.fastread() instead |
|
501 | 505 | for f in ctx.files(): |
|
502 | 506 | if not m(f): |
|
503 | 507 | continue |
|
504 | 508 | p = pointerfromctx(ctx, f, removed=removed) |
|
505 | 509 | if p is not None: |
|
506 | 510 | result[f] = p |
|
507 | 511 | return result |
|
508 | 512 | |
|
509 | 513 | |
|
510 | 514 | def uploadblobs(repo, pointers): |
|
511 | 515 | """upload given pointers from local blobstore""" |
|
512 | 516 | if not pointers: |
|
513 | 517 | return |
|
514 | 518 | |
|
515 | 519 | remoteblob = repo.svfs.lfsremoteblobstore |
|
516 | 520 | remoteblob.writebatch(pointers, repo.svfs.lfslocalblobstore) |
|
517 | 521 | |
|
518 | 522 | |
|
519 | 523 | @eh.wrapfunction(upgrade, b'_finishdatamigration') |
|
520 | 524 | def upgradefinishdatamigration(orig, ui, srcrepo, dstrepo, requirements): |
|
521 | 525 | orig(ui, srcrepo, dstrepo, requirements) |
|
522 | 526 | |
|
523 | 527 | # Skip if this hasn't been passed to reposetup() |
|
524 | 528 | if util.safehasattr( |
|
525 | 529 | srcrepo.svfs, b'lfslocalblobstore' |
|
526 | 530 | ) and util.safehasattr(dstrepo.svfs, b'lfslocalblobstore'): |
|
527 | 531 | srclfsvfs = srcrepo.svfs.lfslocalblobstore.vfs |
|
528 | 532 | dstlfsvfs = dstrepo.svfs.lfslocalblobstore.vfs |
|
529 | 533 | |
|
530 | 534 | for dirpath, dirs, files in srclfsvfs.walk(): |
|
531 | 535 | for oid in files: |
|
532 | 536 | ui.write(_(b'copying lfs blob %s\n') % oid) |
|
533 | 537 | lfutil.link(srclfsvfs.join(oid), dstlfsvfs.join(oid)) |
|
534 | 538 | |
|
535 | 539 | |
|
536 | 540 | @eh.wrapfunction(upgrade, b'preservedrequirements') |
|
537 | 541 | @eh.wrapfunction(upgrade, b'supporteddestrequirements') |
|
538 | 542 | def upgraderequirements(orig, repo): |
|
539 | 543 | reqs = orig(repo) |
|
540 | 544 | if b'lfs' in repo.requirements: |
|
541 | 545 | reqs.add(b'lfs') |
|
542 | 546 | return reqs |
@@ -1,1984 +1,1986 b'' | |||
|
1 | 1 | # repository.py - Interfaces and base classes for repositories and peers. |
|
2 | 2 | # |
|
3 | 3 | # Copyright 2017 Gregory Szorc <gregory.szorc@gmail.com> |
|
4 | 4 | # |
|
5 | 5 | # This software may be used and distributed according to the terms of the |
|
6 | 6 | # GNU General Public License version 2 or any later version. |
|
7 | 7 | |
|
8 | 8 | from __future__ import absolute_import |
|
9 | 9 | |
|
10 | 10 | from ..i18n import _ |
|
11 | 11 | from .. import error |
|
12 | 12 | from . import util as interfaceutil |
|
13 | 13 | |
|
14 | 14 | # When narrowing is finalized and no longer subject to format changes, |
|
15 | 15 | # we should move this to just "narrow" or similar. |
|
16 | 16 | NARROW_REQUIREMENT = b'narrowhg-experimental' |
|
17 | 17 | |
|
18 | 18 | # Local repository feature string. |
|
19 | 19 | |
|
20 | 20 | # Revlogs are being used for file storage. |
|
21 | 21 | REPO_FEATURE_REVLOG_FILE_STORAGE = b'revlogfilestorage' |
|
22 | 22 | # The storage part of the repository is shared from an external source. |
|
23 | 23 | REPO_FEATURE_SHARED_STORAGE = b'sharedstore' |
|
24 | 24 | # LFS supported for backing file storage. |
|
25 | 25 | REPO_FEATURE_LFS = b'lfs' |
|
26 | 26 | # Repository supports being stream cloned. |
|
27 | 27 | REPO_FEATURE_STREAM_CLONE = b'streamclone' |
|
28 | 28 | # Files storage may lack data for all ancestors. |
|
29 | 29 | REPO_FEATURE_SHALLOW_FILE_STORAGE = b'shallowfilestorage' |
|
30 | 30 | |
|
31 | 31 | REVISION_FLAG_CENSORED = 1 << 15 |
|
32 | 32 | REVISION_FLAG_ELLIPSIS = 1 << 14 |
|
33 | 33 | REVISION_FLAG_EXTSTORED = 1 << 13 |
|
34 | 34 | REVISION_FLAG_SIDEDATA = 1 << 12 |
|
35 | 35 | |
|
36 | 36 | REVISION_FLAGS_KNOWN = ( |
|
37 | 37 | REVISION_FLAG_CENSORED |
|
38 | 38 | | REVISION_FLAG_ELLIPSIS |
|
39 | 39 | | REVISION_FLAG_EXTSTORED |
|
40 | 40 | | REVISION_FLAG_SIDEDATA |
|
41 | 41 | ) |
|
42 | 42 | |
|
43 | 43 | CG_DELTAMODE_STD = b'default' |
|
44 | 44 | CG_DELTAMODE_PREV = b'previous' |
|
45 | 45 | CG_DELTAMODE_FULL = b'fulltext' |
|
46 | 46 | CG_DELTAMODE_P1 = b'p1' |
|
47 | 47 | |
|
48 | 48 | |
|
49 | 49 | class ipeerconnection(interfaceutil.Interface): |
|
50 | 50 | """Represents a "connection" to a repository. |
|
51 | 51 | |
|
52 | 52 | This is the base interface for representing a connection to a repository. |
|
53 | 53 | It holds basic properties and methods applicable to all peer types. |
|
54 | 54 | |
|
55 | 55 | This is not a complete interface definition and should not be used |
|
56 | 56 | outside of this module. |
|
57 | 57 | """ |
|
58 | 58 | |
|
59 | 59 | ui = interfaceutil.Attribute("""ui.ui instance""") |
|
60 | 60 | |
|
61 | 61 | def url(): |
|
62 | 62 | """Returns a URL string representing this peer. |
|
63 | 63 | |
|
64 | 64 | Currently, implementations expose the raw URL used to construct the |
|
65 | 65 | instance. It may contain credentials as part of the URL. The |
|
66 | 66 | expectations of the value aren't well-defined and this could lead to |
|
67 | 67 | data leakage. |
|
68 | 68 | |
|
69 | 69 | TODO audit/clean consumers and more clearly define the contents of this |
|
70 | 70 | value. |
|
71 | 71 | """ |
|
72 | 72 | |
|
73 | 73 | def local(): |
|
74 | 74 | """Returns a local repository instance. |
|
75 | 75 | |
|
76 | 76 | If the peer represents a local repository, returns an object that |
|
77 | 77 | can be used to interface with it. Otherwise returns ``None``. |
|
78 | 78 | """ |
|
79 | 79 | |
|
80 | 80 | def peer(): |
|
81 | 81 | """Returns an object conforming to this interface. |
|
82 | 82 | |
|
83 | 83 | Most implementations will ``return self``. |
|
84 | 84 | """ |
|
85 | 85 | |
|
86 | 86 | def canpush(): |
|
87 | 87 | """Returns a boolean indicating if this peer can be pushed to.""" |
|
88 | 88 | |
|
89 | 89 | def close(): |
|
90 | 90 | """Close the connection to this peer. |
|
91 | 91 | |
|
92 | 92 | This is called when the peer will no longer be used. Resources |
|
93 | 93 | associated with the peer should be cleaned up. |
|
94 | 94 | """ |
|
95 | 95 | |
|
96 | 96 | |
|
97 | 97 | class ipeercapabilities(interfaceutil.Interface): |
|
98 | 98 | """Peer sub-interface related to capabilities.""" |
|
99 | 99 | |
|
100 | 100 | def capable(name): |
|
101 | 101 | """Determine support for a named capability. |
|
102 | 102 | |
|
103 | 103 | Returns ``False`` if capability not supported. |
|
104 | 104 | |
|
105 | 105 | Returns ``True`` if boolean capability is supported. Returns a string |
|
106 | 106 | if capability support is non-boolean. |
|
107 | 107 | |
|
108 | 108 | Capability strings may or may not map to wire protocol capabilities. |
|
109 | 109 | """ |
|
110 | 110 | |
|
111 | 111 | def requirecap(name, purpose): |
|
112 | 112 | """Require a capability to be present. |
|
113 | 113 | |
|
114 | 114 | Raises a ``CapabilityError`` if the capability isn't present. |
|
115 | 115 | """ |
|
116 | 116 | |
|
117 | 117 | |
|
118 | 118 | class ipeercommands(interfaceutil.Interface): |
|
119 | 119 | """Client-side interface for communicating over the wire protocol. |
|
120 | 120 | |
|
121 | 121 | This interface is used as a gateway to the Mercurial wire protocol. |
|
122 | 122 | methods commonly call wire protocol commands of the same name. |
|
123 | 123 | """ |
|
124 | 124 | |
|
125 | 125 | def branchmap(): |
|
126 | 126 | """Obtain heads in named branches. |
|
127 | 127 | |
|
128 | 128 | Returns a dict mapping branch name to an iterable of nodes that are |
|
129 | 129 | heads on that branch. |
|
130 | 130 | """ |
|
131 | 131 | |
|
132 | 132 | def capabilities(): |
|
133 | 133 | """Obtain capabilities of the peer. |
|
134 | 134 | |
|
135 | 135 | Returns a set of string capabilities. |
|
136 | 136 | """ |
|
137 | 137 | |
|
138 | 138 | def clonebundles(): |
|
139 | 139 | """Obtains the clone bundles manifest for the repo. |
|
140 | 140 | |
|
141 | 141 | Returns the manifest as unparsed bytes. |
|
142 | 142 | """ |
|
143 | 143 | |
|
144 | 144 | def debugwireargs(one, two, three=None, four=None, five=None): |
|
145 | 145 | """Used to facilitate debugging of arguments passed over the wire.""" |
|
146 | 146 | |
|
147 | 147 | def getbundle(source, **kwargs): |
|
148 | 148 | """Obtain remote repository data as a bundle. |
|
149 | 149 | |
|
150 | 150 | This command is how the bulk of repository data is transferred from |
|
151 | 151 | the peer to the local repository |
|
152 | 152 | |
|
153 | 153 | Returns a generator of bundle data. |
|
154 | 154 | """ |
|
155 | 155 | |
|
156 | 156 | def heads(): |
|
157 | 157 | """Determine all known head revisions in the peer. |
|
158 | 158 | |
|
159 | 159 | Returns an iterable of binary nodes. |
|
160 | 160 | """ |
|
161 | 161 | |
|
162 | 162 | def known(nodes): |
|
163 | 163 | """Determine whether multiple nodes are known. |
|
164 | 164 | |
|
165 | 165 | Accepts an iterable of nodes whose presence to check for. |
|
166 | 166 | |
|
167 | 167 | Returns an iterable of booleans indicating of the corresponding node |
|
168 | 168 | at that index is known to the peer. |
|
169 | 169 | """ |
|
170 | 170 | |
|
171 | 171 | def listkeys(namespace): |
|
172 | 172 | """Obtain all keys in a pushkey namespace. |
|
173 | 173 | |
|
174 | 174 | Returns an iterable of key names. |
|
175 | 175 | """ |
|
176 | 176 | |
|
177 | 177 | def lookup(key): |
|
178 | 178 | """Resolve a value to a known revision. |
|
179 | 179 | |
|
180 | 180 | Returns a binary node of the resolved revision on success. |
|
181 | 181 | """ |
|
182 | 182 | |
|
183 | 183 | def pushkey(namespace, key, old, new): |
|
184 | 184 | """Set a value using the ``pushkey`` protocol. |
|
185 | 185 | |
|
186 | 186 | Arguments correspond to the pushkey namespace and key to operate on and |
|
187 | 187 | the old and new values for that key. |
|
188 | 188 | |
|
189 | 189 | Returns a string with the peer result. The value inside varies by the |
|
190 | 190 | namespace. |
|
191 | 191 | """ |
|
192 | 192 | |
|
193 | 193 | def stream_out(): |
|
194 | 194 | """Obtain streaming clone data. |
|
195 | 195 | |
|
196 | 196 | Successful result should be a generator of data chunks. |
|
197 | 197 | """ |
|
198 | 198 | |
|
199 | 199 | def unbundle(bundle, heads, url): |
|
200 | 200 | """Transfer repository data to the peer. |
|
201 | 201 | |
|
202 | 202 | This is how the bulk of data during a push is transferred. |
|
203 | 203 | |
|
204 | 204 | Returns the integer number of heads added to the peer. |
|
205 | 205 | """ |
|
206 | 206 | |
|
207 | 207 | |
|
208 | 208 | class ipeerlegacycommands(interfaceutil.Interface): |
|
209 | 209 | """Interface for implementing support for legacy wire protocol commands. |
|
210 | 210 | |
|
211 | 211 | Wire protocol commands transition to legacy status when they are no longer |
|
212 | 212 | used by modern clients. To facilitate identifying which commands are |
|
213 | 213 | legacy, the interfaces are split. |
|
214 | 214 | """ |
|
215 | 215 | |
|
216 | 216 | def between(pairs): |
|
217 | 217 | """Obtain nodes between pairs of nodes. |
|
218 | 218 | |
|
219 | 219 | ``pairs`` is an iterable of node pairs. |
|
220 | 220 | |
|
221 | 221 | Returns an iterable of iterables of nodes corresponding to each |
|
222 | 222 | requested pair. |
|
223 | 223 | """ |
|
224 | 224 | |
|
225 | 225 | def branches(nodes): |
|
226 | 226 | """Obtain ancestor changesets of specific nodes back to a branch point. |
|
227 | 227 | |
|
228 | 228 | For each requested node, the peer finds the first ancestor node that is |
|
229 | 229 | a DAG root or is a merge. |
|
230 | 230 | |
|
231 | 231 | Returns an iterable of iterables with the resolved values for each node. |
|
232 | 232 | """ |
|
233 | 233 | |
|
234 | 234 | def changegroup(nodes, source): |
|
235 | 235 | """Obtain a changegroup with data for descendants of specified nodes.""" |
|
236 | 236 | |
|
237 | 237 | def changegroupsubset(bases, heads, source): |
|
238 | 238 | pass |
|
239 | 239 | |
|
240 | 240 | |
|
241 | 241 | class ipeercommandexecutor(interfaceutil.Interface): |
|
242 | 242 | """Represents a mechanism to execute remote commands. |
|
243 | 243 | |
|
244 | 244 | This is the primary interface for requesting that wire protocol commands |
|
245 | 245 | be executed. Instances of this interface are active in a context manager |
|
246 | 246 | and have a well-defined lifetime. When the context manager exits, all |
|
247 | 247 | outstanding requests are waited on. |
|
248 | 248 | """ |
|
249 | 249 | |
|
250 | 250 | def callcommand(name, args): |
|
251 | 251 | """Request that a named command be executed. |
|
252 | 252 | |
|
253 | 253 | Receives the command name and a dictionary of command arguments. |
|
254 | 254 | |
|
255 | 255 | Returns a ``concurrent.futures.Future`` that will resolve to the |
|
256 | 256 | result of that command request. That exact value is left up to |
|
257 | 257 | the implementation and possibly varies by command. |
|
258 | 258 | |
|
259 | 259 | Not all commands can coexist with other commands in an executor |
|
260 | 260 | instance: it depends on the underlying wire protocol transport being |
|
261 | 261 | used and the command itself. |
|
262 | 262 | |
|
263 | 263 | Implementations MAY call ``sendcommands()`` automatically if the |
|
264 | 264 | requested command can not coexist with other commands in this executor. |
|
265 | 265 | |
|
266 | 266 | Implementations MAY call ``sendcommands()`` automatically when the |
|
267 | 267 | future's ``result()`` is called. So, consumers using multiple |
|
268 | 268 | commands with an executor MUST ensure that ``result()`` is not called |
|
269 | 269 | until all command requests have been issued. |
|
270 | 270 | """ |
|
271 | 271 | |
|
272 | 272 | def sendcommands(): |
|
273 | 273 | """Trigger submission of queued command requests. |
|
274 | 274 | |
|
275 | 275 | Not all transports submit commands as soon as they are requested to |
|
276 | 276 | run. When called, this method forces queued command requests to be |
|
277 | 277 | issued. It will no-op if all commands have already been sent. |
|
278 | 278 | |
|
279 | 279 | When called, no more new commands may be issued with this executor. |
|
280 | 280 | """ |
|
281 | 281 | |
|
282 | 282 | def close(): |
|
283 | 283 | """Signal that this command request is finished. |
|
284 | 284 | |
|
285 | 285 | When called, no more new commands may be issued. All outstanding |
|
286 | 286 | commands that have previously been issued are waited on before |
|
287 | 287 | returning. This not only includes waiting for the futures to resolve, |
|
288 | 288 | but also waiting for all response data to arrive. In other words, |
|
289 | 289 | calling this waits for all on-wire state for issued command requests |
|
290 | 290 | to finish. |
|
291 | 291 | |
|
292 | 292 | When used as a context manager, this method is called when exiting the |
|
293 | 293 | context manager. |
|
294 | 294 | |
|
295 | 295 | This method may call ``sendcommands()`` if there are buffered commands. |
|
296 | 296 | """ |
|
297 | 297 | |
|
298 | 298 | |
|
299 | 299 | class ipeerrequests(interfaceutil.Interface): |
|
300 | 300 | """Interface for executing commands on a peer.""" |
|
301 | 301 | |
|
302 | 302 | limitedarguments = interfaceutil.Attribute( |
|
303 | 303 | """True if the peer cannot receive large argument value for commands.""" |
|
304 | 304 | ) |
|
305 | 305 | |
|
306 | 306 | def commandexecutor(): |
|
307 | 307 | """A context manager that resolves to an ipeercommandexecutor. |
|
308 | 308 | |
|
309 | 309 | The object this resolves to can be used to issue command requests |
|
310 | 310 | to the peer. |
|
311 | 311 | |
|
312 | 312 | Callers should call its ``callcommand`` method to issue command |
|
313 | 313 | requests. |
|
314 | 314 | |
|
315 | 315 | A new executor should be obtained for each distinct set of commands |
|
316 | 316 | (possibly just a single command) that the consumer wants to execute |
|
317 | 317 | as part of a single operation or round trip. This is because some |
|
318 | 318 | peers are half-duplex and/or don't support persistent connections. |
|
319 | 319 | e.g. in the case of HTTP peers, commands sent to an executor represent |
|
320 | 320 | a single HTTP request. While some peers may support multiple command |
|
321 | 321 | sends over the wire per executor, consumers need to code to the least |
|
322 | 322 | capable peer. So it should be assumed that command executors buffer |
|
323 | 323 | called commands until they are told to send them and that each |
|
324 | 324 | command executor could result in a new connection or wire-level request |
|
325 | 325 | being issued. |
|
326 | 326 | """ |
|
327 | 327 | |
|
328 | 328 | |
|
329 | 329 | class ipeerbase(ipeerconnection, ipeercapabilities, ipeerrequests): |
|
330 | 330 | """Unified interface for peer repositories. |
|
331 | 331 | |
|
332 | 332 | All peer instances must conform to this interface. |
|
333 | 333 | """ |
|
334 | 334 | |
|
335 | 335 | |
|
336 | 336 | class ipeerv2(ipeerconnection, ipeercapabilities, ipeerrequests): |
|
337 | 337 | """Unified peer interface for wire protocol version 2 peers.""" |
|
338 | 338 | |
|
339 | 339 | apidescriptor = interfaceutil.Attribute( |
|
340 | 340 | """Data structure holding description of server API.""" |
|
341 | 341 | ) |
|
342 | 342 | |
|
343 | 343 | |
|
344 | 344 | @interfaceutil.implementer(ipeerbase) |
|
345 | 345 | class peer(object): |
|
346 | 346 | """Base class for peer repositories.""" |
|
347 | 347 | |
|
348 | 348 | limitedarguments = False |
|
349 | 349 | |
|
350 | 350 | def capable(self, name): |
|
351 | 351 | caps = self.capabilities() |
|
352 | 352 | if name in caps: |
|
353 | 353 | return True |
|
354 | 354 | |
|
355 | 355 | name = b'%s=' % name |
|
356 | 356 | for cap in caps: |
|
357 | 357 | if cap.startswith(name): |
|
358 | 358 | return cap[len(name) :] |
|
359 | 359 | |
|
360 | 360 | return False |
|
361 | 361 | |
|
362 | 362 | def requirecap(self, name, purpose): |
|
363 | 363 | if self.capable(name): |
|
364 | 364 | return |
|
365 | 365 | |
|
366 | 366 | raise error.CapabilityError( |
|
367 | 367 | _( |
|
368 | 368 | b'cannot %s; remote repository does not support the ' |
|
369 | 369 | b'\'%s\' capability' |
|
370 | 370 | ) |
|
371 | 371 | % (purpose, name) |
|
372 | 372 | ) |
|
373 | 373 | |
|
374 | 374 | |
|
375 | 375 | class iverifyproblem(interfaceutil.Interface): |
|
376 | 376 | """Represents a problem with the integrity of the repository. |
|
377 | 377 | |
|
378 | 378 | Instances of this interface are emitted to describe an integrity issue |
|
379 | 379 | with a repository (e.g. corrupt storage, missing data, etc). |
|
380 | 380 | |
|
381 | 381 | Instances are essentially messages associated with severity. |
|
382 | 382 | """ |
|
383 | 383 | |
|
384 | 384 | warning = interfaceutil.Attribute( |
|
385 | 385 | """Message indicating a non-fatal problem.""" |
|
386 | 386 | ) |
|
387 | 387 | |
|
388 | 388 | error = interfaceutil.Attribute("""Message indicating a fatal problem.""") |
|
389 | 389 | |
|
390 | 390 | node = interfaceutil.Attribute( |
|
391 | 391 | """Revision encountering the problem. |
|
392 | 392 | |
|
393 | 393 | ``None`` means the problem doesn't apply to a single revision. |
|
394 | 394 | """ |
|
395 | 395 | ) |
|
396 | 396 | |
|
397 | 397 | |
|
398 | 398 | class irevisiondelta(interfaceutil.Interface): |
|
399 | 399 | """Represents a delta between one revision and another. |
|
400 | 400 | |
|
401 | 401 | Instances convey enough information to allow a revision to be exchanged |
|
402 | 402 | with another repository. |
|
403 | 403 | |
|
404 | 404 | Instances represent the fulltext revision data or a delta against |
|
405 | 405 | another revision. Therefore the ``revision`` and ``delta`` attributes |
|
406 | 406 | are mutually exclusive. |
|
407 | 407 | |
|
408 | 408 | Typically used for changegroup generation. |
|
409 | 409 | """ |
|
410 | 410 | |
|
411 | 411 | node = interfaceutil.Attribute("""20 byte node of this revision.""") |
|
412 | 412 | |
|
413 | 413 | p1node = interfaceutil.Attribute( |
|
414 | 414 | """20 byte node of 1st parent of this revision.""" |
|
415 | 415 | ) |
|
416 | 416 | |
|
417 | 417 | p2node = interfaceutil.Attribute( |
|
418 | 418 | """20 byte node of 2nd parent of this revision.""" |
|
419 | 419 | ) |
|
420 | 420 | |
|
421 | 421 | linknode = interfaceutil.Attribute( |
|
422 | 422 | """20 byte node of the changelog revision this node is linked to.""" |
|
423 | 423 | ) |
|
424 | 424 | |
|
425 | 425 | flags = interfaceutil.Attribute( |
|
426 | 426 | """2 bytes of integer flags that apply to this revision. |
|
427 | 427 | |
|
428 | 428 | This is a bitwise composition of the ``REVISION_FLAG_*`` constants. |
|
429 | 429 | """ |
|
430 | 430 | ) |
|
431 | 431 | |
|
432 | 432 | basenode = interfaceutil.Attribute( |
|
433 | 433 | """20 byte node of the revision this data is a delta against. |
|
434 | 434 | |
|
435 | 435 | ``nullid`` indicates that the revision is a full revision and not |
|
436 | 436 | a delta. |
|
437 | 437 | """ |
|
438 | 438 | ) |
|
439 | 439 | |
|
440 | 440 | baserevisionsize = interfaceutil.Attribute( |
|
441 | 441 | """Size of base revision this delta is against. |
|
442 | 442 | |
|
443 | 443 | May be ``None`` if ``basenode`` is ``nullid``. |
|
444 | 444 | """ |
|
445 | 445 | ) |
|
446 | 446 | |
|
447 | 447 | revision = interfaceutil.Attribute( |
|
448 | 448 | """Raw fulltext of revision data for this node.""" |
|
449 | 449 | ) |
|
450 | 450 | |
|
451 | 451 | delta = interfaceutil.Attribute( |
|
452 | 452 | """Delta between ``basenode`` and ``node``. |
|
453 | 453 | |
|
454 | 454 | Stored in the bdiff delta format. |
|
455 | 455 | """ |
|
456 | 456 | ) |
|
457 | 457 | |
|
458 | 458 | |
|
459 | 459 | class ifilerevisionssequence(interfaceutil.Interface): |
|
460 | 460 | """Contains index data for all revisions of a file. |
|
461 | 461 | |
|
462 | 462 | Types implementing this behave like lists of tuples. The index |
|
463 | 463 | in the list corresponds to the revision number. The values contain |
|
464 | 464 | index metadata. |
|
465 | 465 | |
|
466 | 466 | The *null* revision (revision number -1) is always the last item |
|
467 | 467 | in the index. |
|
468 | 468 | """ |
|
469 | 469 | |
|
470 | 470 | def __len__(): |
|
471 | 471 | """The total number of revisions.""" |
|
472 | 472 | |
|
473 | 473 | def __getitem__(rev): |
|
474 | 474 | """Returns the object having a specific revision number. |
|
475 | 475 | |
|
476 | 476 | Returns an 8-tuple with the following fields: |
|
477 | 477 | |
|
478 | 478 | offset+flags |
|
479 | 479 | Contains the offset and flags for the revision. 64-bit unsigned |
|
480 | 480 | integer where first 6 bytes are the offset and the next 2 bytes |
|
481 | 481 | are flags. The offset can be 0 if it is not used by the store. |
|
482 | 482 | compressed size |
|
483 | 483 | Size of the revision data in the store. It can be 0 if it isn't |
|
484 | 484 | needed by the store. |
|
485 | 485 | uncompressed size |
|
486 | 486 | Fulltext size. It can be 0 if it isn't needed by the store. |
|
487 | 487 | base revision |
|
488 | 488 | Revision number of revision the delta for storage is encoded |
|
489 | 489 | against. -1 indicates not encoded against a base revision. |
|
490 | 490 | link revision |
|
491 | 491 | Revision number of changelog revision this entry is related to. |
|
492 | 492 | p1 revision |
|
493 | 493 | Revision number of 1st parent. -1 if no 1st parent. |
|
494 | 494 | p2 revision |
|
495 | 495 | Revision number of 2nd parent. -1 if no 1st parent. |
|
496 | 496 | node |
|
497 | 497 | Binary node value for this revision number. |
|
498 | 498 | |
|
499 | 499 | Negative values should index off the end of the sequence. ``-1`` |
|
500 | 500 | should return the null revision. ``-2`` should return the most |
|
501 | 501 | recent revision. |
|
502 | 502 | """ |
|
503 | 503 | |
|
504 | 504 | def __contains__(rev): |
|
505 | 505 | """Whether a revision number exists.""" |
|
506 | 506 | |
|
507 | 507 | def insert(self, i, entry): |
|
508 | 508 | """Add an item to the index at specific revision.""" |
|
509 | 509 | |
|
510 | 510 | |
|
511 | 511 | class ifileindex(interfaceutil.Interface): |
|
512 | 512 | """Storage interface for index data of a single file. |
|
513 | 513 | |
|
514 | 514 | File storage data is divided into index metadata and data storage. |
|
515 | 515 | This interface defines the index portion of the interface. |
|
516 | 516 | |
|
517 | 517 | The index logically consists of: |
|
518 | 518 | |
|
519 | 519 | * A mapping between revision numbers and nodes. |
|
520 | 520 | * DAG data (storing and querying the relationship between nodes). |
|
521 | 521 | * Metadata to facilitate storage. |
|
522 | 522 | """ |
|
523 | 523 | |
|
524 | 524 | def __len__(): |
|
525 | 525 | """Obtain the number of revisions stored for this file.""" |
|
526 | 526 | |
|
527 | 527 | def __iter__(): |
|
528 | 528 | """Iterate over revision numbers for this file.""" |
|
529 | 529 | |
|
530 | 530 | def hasnode(node): |
|
531 | 531 | """Returns a bool indicating if a node is known to this store. |
|
532 | 532 | |
|
533 | 533 | Implementations must only return True for full, binary node values: |
|
534 | 534 | hex nodes, revision numbers, and partial node matches must be |
|
535 | 535 | rejected. |
|
536 | 536 | |
|
537 | 537 | The null node is never present. |
|
538 | 538 | """ |
|
539 | 539 | |
|
540 | 540 | def revs(start=0, stop=None): |
|
541 | 541 | """Iterate over revision numbers for this file, with control.""" |
|
542 | 542 | |
|
543 | 543 | def parents(node): |
|
544 | 544 | """Returns a 2-tuple of parent nodes for a revision. |
|
545 | 545 | |
|
546 | 546 | Values will be ``nullid`` if the parent is empty. |
|
547 | 547 | """ |
|
548 | 548 | |
|
549 | 549 | def parentrevs(rev): |
|
550 | 550 | """Like parents() but operates on revision numbers.""" |
|
551 | 551 | |
|
552 | 552 | def rev(node): |
|
553 | 553 | """Obtain the revision number given a node. |
|
554 | 554 | |
|
555 | 555 | Raises ``error.LookupError`` if the node is not known. |
|
556 | 556 | """ |
|
557 | 557 | |
|
558 | 558 | def node(rev): |
|
559 | 559 | """Obtain the node value given a revision number. |
|
560 | 560 | |
|
561 | 561 | Raises ``IndexError`` if the node is not known. |
|
562 | 562 | """ |
|
563 | 563 | |
|
564 | 564 | def lookup(node): |
|
565 | 565 | """Attempt to resolve a value to a node. |
|
566 | 566 | |
|
567 | 567 | Value can be a binary node, hex node, revision number, or a string |
|
568 | 568 | that can be converted to an integer. |
|
569 | 569 | |
|
570 | 570 | Raises ``error.LookupError`` if a node could not be resolved. |
|
571 | 571 | """ |
|
572 | 572 | |
|
573 | 573 | def linkrev(rev): |
|
574 | 574 | """Obtain the changeset revision number a revision is linked to.""" |
|
575 | 575 | |
|
576 | 576 | def iscensored(rev): |
|
577 | 577 | """Return whether a revision's content has been censored.""" |
|
578 | 578 | |
|
579 | 579 | def commonancestorsheads(node1, node2): |
|
580 | 580 | """Obtain an iterable of nodes containing heads of common ancestors. |
|
581 | 581 | |
|
582 | 582 | See ``ancestor.commonancestorsheads()``. |
|
583 | 583 | """ |
|
584 | 584 | |
|
585 | 585 | def descendants(revs): |
|
586 | 586 | """Obtain descendant revision numbers for a set of revision numbers. |
|
587 | 587 | |
|
588 | 588 | If ``nullrev`` is in the set, this is equivalent to ``revs()``. |
|
589 | 589 | """ |
|
590 | 590 | |
|
591 | 591 | def heads(start=None, stop=None): |
|
592 | 592 | """Obtain a list of nodes that are DAG heads, with control. |
|
593 | 593 | |
|
594 | 594 | The set of revisions examined can be limited by specifying |
|
595 | 595 | ``start`` and ``stop``. ``start`` is a node. ``stop`` is an |
|
596 | 596 | iterable of nodes. DAG traversal starts at earlier revision |
|
597 | 597 | ``start`` and iterates forward until any node in ``stop`` is |
|
598 | 598 | encountered. |
|
599 | 599 | """ |
|
600 | 600 | |
|
601 | 601 | def children(node): |
|
602 | 602 | """Obtain nodes that are children of a node. |
|
603 | 603 | |
|
604 | 604 | Returns a list of nodes. |
|
605 | 605 | """ |
|
606 | 606 | |
|
607 | 607 | |
|
608 | 608 | class ifiledata(interfaceutil.Interface): |
|
609 | 609 | """Storage interface for data storage of a specific file. |
|
610 | 610 | |
|
611 | 611 | This complements ``ifileindex`` and provides an interface for accessing |
|
612 | 612 | data for a tracked file. |
|
613 | 613 | """ |
|
614 | 614 | |
|
615 | 615 | def size(rev): |
|
616 | 616 | """Obtain the fulltext size of file data. |
|
617 | 617 | |
|
618 | 618 | Any metadata is excluded from size measurements. |
|
619 | 619 | """ |
|
620 | 620 | |
|
621 | 621 | def revision(node, raw=False): |
|
622 | 622 | """"Obtain fulltext data for a node. |
|
623 | 623 | |
|
624 | 624 | By default, any storage transformations are applied before the data |
|
625 | 625 | is returned. If ``raw`` is True, non-raw storage transformations |
|
626 | 626 | are not applied. |
|
627 | 627 | |
|
628 | 628 | The fulltext data may contain a header containing metadata. Most |
|
629 | 629 | consumers should use ``read()`` to obtain the actual file data. |
|
630 | 630 | """ |
|
631 | 631 | |
|
632 | 632 | def rawdata(node): |
|
633 | 633 | """Obtain raw data for a node. |
|
634 | 634 | """ |
|
635 | 635 | |
|
636 | 636 | def read(node): |
|
637 | 637 | """Resolve file fulltext data. |
|
638 | 638 | |
|
639 | 639 | This is similar to ``revision()`` except any metadata in the data |
|
640 | 640 | headers is stripped. |
|
641 | 641 | """ |
|
642 | 642 | |
|
643 | 643 | def renamed(node): |
|
644 | 644 | """Obtain copy metadata for a node. |
|
645 | 645 | |
|
646 | 646 | Returns ``False`` if no copy metadata is stored or a 2-tuple of |
|
647 | 647 | (path, node) from which this revision was copied. |
|
648 | 648 | """ |
|
649 | 649 | |
|
650 | 650 | def cmp(node, fulltext): |
|
651 | 651 | """Compare fulltext to another revision. |
|
652 | 652 | |
|
653 | 653 | Returns True if the fulltext is different from what is stored. |
|
654 | 654 | |
|
655 | 655 | This takes copy metadata into account. |
|
656 | 656 | |
|
657 | 657 | TODO better document the copy metadata and censoring logic. |
|
658 | 658 | """ |
|
659 | 659 | |
|
660 | 660 | def emitrevisions( |
|
661 | 661 | nodes, |
|
662 | 662 | nodesorder=None, |
|
663 | 663 | revisiondata=False, |
|
664 | 664 | assumehaveparentrevisions=False, |
|
665 | 665 | deltamode=CG_DELTAMODE_STD, |
|
666 | 666 | ): |
|
667 | 667 | """Produce ``irevisiondelta`` for revisions. |
|
668 | 668 | |
|
669 | 669 | Given an iterable of nodes, emits objects conforming to the |
|
670 | 670 | ``irevisiondelta`` interface that describe revisions in storage. |
|
671 | 671 | |
|
672 | 672 | This method is a generator. |
|
673 | 673 | |
|
674 | 674 | The input nodes may be unordered. Implementations must ensure that a |
|
675 | 675 | node's parents are emitted before the node itself. Transitively, this |
|
676 | 676 | means that a node may only be emitted once all its ancestors in |
|
677 | 677 | ``nodes`` have also been emitted. |
|
678 | 678 | |
|
679 | 679 | By default, emits "index" data (the ``node``, ``p1node``, and |
|
680 | 680 | ``p2node`` attributes). If ``revisiondata`` is set, revision data |
|
681 | 681 | will also be present on the emitted objects. |
|
682 | 682 | |
|
683 | 683 | With default argument values, implementations can choose to emit |
|
684 | 684 | either fulltext revision data or a delta. When emitting deltas, |
|
685 | 685 | implementations must consider whether the delta's base revision |
|
686 | 686 | fulltext is available to the receiver. |
|
687 | 687 | |
|
688 | 688 | The base revision fulltext is guaranteed to be available if any of |
|
689 | 689 | the following are met: |
|
690 | 690 | |
|
691 | 691 | * Its fulltext revision was emitted by this method call. |
|
692 | 692 | * A delta for that revision was emitted by this method call. |
|
693 | 693 | * ``assumehaveparentrevisions`` is True and the base revision is a |
|
694 | 694 | parent of the node. |
|
695 | 695 | |
|
696 | 696 | ``nodesorder`` can be used to control the order that revisions are |
|
697 | 697 | emitted. By default, revisions can be reordered as long as they are |
|
698 | 698 | in DAG topological order (see above). If the value is ``nodes``, |
|
699 | 699 | the iteration order from ``nodes`` should be used. If the value is |
|
700 | 700 | ``storage``, then the native order from the backing storage layer |
|
701 | 701 | is used. (Not all storage layers will have strong ordering and behavior |
|
702 | 702 | of this mode is storage-dependent.) ``nodes`` ordering can force |
|
703 | 703 | revisions to be emitted before their ancestors, so consumers should |
|
704 | 704 | use it with care. |
|
705 | 705 | |
|
706 | 706 | The ``linknode`` attribute on the returned ``irevisiondelta`` may not |
|
707 | 707 | be set and it is the caller's responsibility to resolve it, if needed. |
|
708 | 708 | |
|
709 | 709 | If ``deltamode`` is CG_DELTAMODE_PREV and revision data is requested, |
|
710 | 710 | all revision data should be emitted as deltas against the revision |
|
711 | 711 | emitted just prior. The initial revision should be a delta against its |
|
712 | 712 | 1st parent. |
|
713 | 713 | """ |
|
714 | 714 | |
|
715 | 715 | |
|
716 | 716 | class ifilemutation(interfaceutil.Interface): |
|
717 | 717 | """Storage interface for mutation events of a tracked file.""" |
|
718 | 718 | |
|
719 | 719 | def add(filedata, meta, transaction, linkrev, p1, p2): |
|
720 | 720 | """Add a new revision to the store. |
|
721 | 721 | |
|
722 | 722 | Takes file data, dictionary of metadata, a transaction, linkrev, |
|
723 | 723 | and parent nodes. |
|
724 | 724 | |
|
725 | 725 | Returns the node that was added. |
|
726 | 726 | |
|
727 | 727 | May no-op if a revision matching the supplied data is already stored. |
|
728 | 728 | """ |
|
729 | 729 | |
|
730 | 730 | def addrevision( |
|
731 | 731 | revisiondata, |
|
732 | 732 | transaction, |
|
733 | 733 | linkrev, |
|
734 | 734 | p1, |
|
735 | 735 | p2, |
|
736 | 736 | node=None, |
|
737 | 737 | flags=0, |
|
738 | 738 | cachedelta=None, |
|
739 | 739 | ): |
|
740 | 740 | """Add a new revision to the store. |
|
741 | 741 | |
|
742 | 742 | This is similar to ``add()`` except it operates at a lower level. |
|
743 | 743 | |
|
744 | 744 | The data passed in already contains a metadata header, if any. |
|
745 | 745 | |
|
746 | 746 | ``node`` and ``flags`` can be used to define the expected node and |
|
747 | 747 | the flags to use with storage. ``flags`` is a bitwise value composed |
|
748 | 748 | of the various ``REVISION_FLAG_*`` constants. |
|
749 | 749 | |
|
750 | 750 | ``add()`` is usually called when adding files from e.g. the working |
|
751 | 751 | directory. ``addrevision()`` is often called by ``add()`` and for |
|
752 | 752 | scenarios where revision data has already been computed, such as when |
|
753 | 753 | applying raw data from a peer repo. |
|
754 | 754 | """ |
|
755 | 755 | |
|
756 | 756 | def addgroup( |
|
757 | 757 | deltas, |
|
758 | 758 | linkmapper, |
|
759 | 759 | transaction, |
|
760 | 760 | addrevisioncb=None, |
|
761 | 761 | maybemissingparents=False, |
|
762 | 762 | ): |
|
763 | 763 | """Process a series of deltas for storage. |
|
764 | 764 | |
|
765 | 765 | ``deltas`` is an iterable of 7-tuples of |
|
766 | 766 | (node, p1, p2, linknode, deltabase, delta, flags) defining revisions |
|
767 | 767 | to add. |
|
768 | 768 | |
|
769 | 769 | The ``delta`` field contains ``mpatch`` data to apply to a base |
|
770 | 770 | revision, identified by ``deltabase``. The base node can be |
|
771 | 771 | ``nullid``, in which case the header from the delta can be ignored |
|
772 | 772 | and the delta used as the fulltext. |
|
773 | 773 | |
|
774 | 774 | ``addrevisioncb`` should be called for each node as it is committed. |
|
775 | 775 | |
|
776 | 776 | ``maybemissingparents`` is a bool indicating whether the incoming |
|
777 | 777 | data may reference parents/ancestor revisions that aren't present. |
|
778 | 778 | This flag is set when receiving data into a "shallow" store that |
|
779 | 779 | doesn't hold all history. |
|
780 | 780 | |
|
781 | 781 | Returns a list of nodes that were processed. A node will be in the list |
|
782 | 782 | even if it existed in the store previously. |
|
783 | 783 | """ |
|
784 | 784 | |
|
785 | 785 | def censorrevision(tr, node, tombstone=b''): |
|
786 | 786 | """Remove the content of a single revision. |
|
787 | 787 | |
|
788 | 788 | The specified ``node`` will have its content purged from storage. |
|
789 | 789 | Future attempts to access the revision data for this node will |
|
790 | 790 | result in failure. |
|
791 | 791 | |
|
792 | 792 | A ``tombstone`` message can optionally be stored. This message may be |
|
793 | 793 | displayed to users when they attempt to access the missing revision |
|
794 | 794 | data. |
|
795 | 795 | |
|
796 | 796 | Storage backends may have stored deltas against the previous content |
|
797 | 797 | in this revision. As part of censoring a revision, these storage |
|
798 | 798 | backends are expected to rewrite any internally stored deltas such |
|
799 | 799 | that they no longer reference the deleted content. |
|
800 | 800 | """ |
|
801 | 801 | |
|
802 | 802 | def getstrippoint(minlink): |
|
803 | 803 | """Find the minimum revision that must be stripped to strip a linkrev. |
|
804 | 804 | |
|
805 | 805 | Returns a 2-tuple containing the minimum revision number and a set |
|
806 | 806 | of all revisions numbers that would be broken by this strip. |
|
807 | 807 | |
|
808 | 808 | TODO this is highly revlog centric and should be abstracted into |
|
809 | 809 | a higher-level deletion API. ``repair.strip()`` relies on this. |
|
810 | 810 | """ |
|
811 | 811 | |
|
812 | 812 | def strip(minlink, transaction): |
|
813 | 813 | """Remove storage of items starting at a linkrev. |
|
814 | 814 | |
|
815 | 815 | This uses ``getstrippoint()`` to determine the first node to remove. |
|
816 | 816 | Then it effectively truncates storage for all revisions after that. |
|
817 | 817 | |
|
818 | 818 | TODO this is highly revlog centric and should be abstracted into a |
|
819 | 819 | higher-level deletion API. |
|
820 | 820 | """ |
|
821 | 821 | |
|
822 | 822 | |
|
823 | 823 | class ifilestorage(ifileindex, ifiledata, ifilemutation): |
|
824 | 824 | """Complete storage interface for a single tracked file.""" |
|
825 | 825 | |
|
826 | 826 | def files(): |
|
827 | 827 | """Obtain paths that are backing storage for this file. |
|
828 | 828 | |
|
829 | 829 | TODO this is used heavily by verify code and there should probably |
|
830 | 830 | be a better API for that. |
|
831 | 831 | """ |
|
832 | 832 | |
|
833 | 833 | def storageinfo( |
|
834 | 834 | exclusivefiles=False, |
|
835 | 835 | sharedfiles=False, |
|
836 | 836 | revisionscount=False, |
|
837 | 837 | trackedsize=False, |
|
838 | 838 | storedsize=False, |
|
839 | 839 | ): |
|
840 | 840 | """Obtain information about storage for this file's data. |
|
841 | 841 | |
|
842 | 842 | Returns a dict describing storage for this tracked path. The keys |
|
843 | 843 | in the dict map to arguments of the same. The arguments are bools |
|
844 | 844 | indicating whether to calculate and obtain that data. |
|
845 | 845 | |
|
846 | 846 | exclusivefiles |
|
847 | 847 | Iterable of (vfs, path) describing files that are exclusively |
|
848 | 848 | used to back storage for this tracked path. |
|
849 | 849 | |
|
850 | 850 | sharedfiles |
|
851 | 851 | Iterable of (vfs, path) describing files that are used to back |
|
852 | 852 | storage for this tracked path. Those files may also provide storage |
|
853 | 853 | for other stored entities. |
|
854 | 854 | |
|
855 | 855 | revisionscount |
|
856 | 856 | Number of revisions available for retrieval. |
|
857 | 857 | |
|
858 | 858 | trackedsize |
|
859 | 859 | Total size in bytes of all tracked revisions. This is a sum of the |
|
860 | 860 | length of the fulltext of all revisions. |
|
861 | 861 | |
|
862 | 862 | storedsize |
|
863 | 863 | Total size in bytes used to store data for all tracked revisions. |
|
864 | 864 | This is commonly less than ``trackedsize`` due to internal usage |
|
865 | 865 | of deltas rather than fulltext revisions. |
|
866 | 866 | |
|
867 | 867 | Not all storage backends may support all queries are have a reasonable |
|
868 | 868 | value to use. In that case, the value should be set to ``None`` and |
|
869 | 869 | callers are expected to handle this special value. |
|
870 | 870 | """ |
|
871 | 871 | |
|
872 | 872 | def verifyintegrity(state): |
|
873 | 873 | """Verifies the integrity of file storage. |
|
874 | 874 | |
|
875 | 875 | ``state`` is a dict holding state of the verifier process. It can be |
|
876 | 876 | used to communicate data between invocations of multiple storage |
|
877 | 877 | primitives. |
|
878 | 878 | |
|
879 | 879 | If individual revisions cannot have their revision content resolved, |
|
880 | 880 | the method is expected to set the ``skipread`` key to a set of nodes |
|
881 | that encountered problems. | |
|
881 | that encountered problems. If set, the method can also add the node(s) | |
|
882 | to ``safe_renamed`` in order to indicate nodes that may perform the | |
|
883 | rename checks with currently accessible data. | |
|
882 | 884 | |
|
883 | 885 | The method yields objects conforming to the ``iverifyproblem`` |
|
884 | 886 | interface. |
|
885 | 887 | """ |
|
886 | 888 | |
|
887 | 889 | |
|
888 | 890 | class idirs(interfaceutil.Interface): |
|
889 | 891 | """Interface representing a collection of directories from paths. |
|
890 | 892 | |
|
891 | 893 | This interface is essentially a derived data structure representing |
|
892 | 894 | directories from a collection of paths. |
|
893 | 895 | """ |
|
894 | 896 | |
|
895 | 897 | def addpath(path): |
|
896 | 898 | """Add a path to the collection. |
|
897 | 899 | |
|
898 | 900 | All directories in the path will be added to the collection. |
|
899 | 901 | """ |
|
900 | 902 | |
|
901 | 903 | def delpath(path): |
|
902 | 904 | """Remove a path from the collection. |
|
903 | 905 | |
|
904 | 906 | If the removal was the last path in a particular directory, the |
|
905 | 907 | directory is removed from the collection. |
|
906 | 908 | """ |
|
907 | 909 | |
|
908 | 910 | def __iter__(): |
|
909 | 911 | """Iterate over the directories in this collection of paths.""" |
|
910 | 912 | |
|
911 | 913 | def __contains__(path): |
|
912 | 914 | """Whether a specific directory is in this collection.""" |
|
913 | 915 | |
|
914 | 916 | |
|
915 | 917 | class imanifestdict(interfaceutil.Interface): |
|
916 | 918 | """Interface representing a manifest data structure. |
|
917 | 919 | |
|
918 | 920 | A manifest is effectively a dict mapping paths to entries. Each entry |
|
919 | 921 | consists of a binary node and extra flags affecting that entry. |
|
920 | 922 | """ |
|
921 | 923 | |
|
922 | 924 | def __getitem__(path): |
|
923 | 925 | """Returns the binary node value for a path in the manifest. |
|
924 | 926 | |
|
925 | 927 | Raises ``KeyError`` if the path does not exist in the manifest. |
|
926 | 928 | |
|
927 | 929 | Equivalent to ``self.find(path)[0]``. |
|
928 | 930 | """ |
|
929 | 931 | |
|
930 | 932 | def find(path): |
|
931 | 933 | """Returns the entry for a path in the manifest. |
|
932 | 934 | |
|
933 | 935 | Returns a 2-tuple of (node, flags). |
|
934 | 936 | |
|
935 | 937 | Raises ``KeyError`` if the path does not exist in the manifest. |
|
936 | 938 | """ |
|
937 | 939 | |
|
938 | 940 | def __len__(): |
|
939 | 941 | """Return the number of entries in the manifest.""" |
|
940 | 942 | |
|
941 | 943 | def __nonzero__(): |
|
942 | 944 | """Returns True if the manifest has entries, False otherwise.""" |
|
943 | 945 | |
|
944 | 946 | __bool__ = __nonzero__ |
|
945 | 947 | |
|
946 | 948 | def __setitem__(path, node): |
|
947 | 949 | """Define the node value for a path in the manifest. |
|
948 | 950 | |
|
949 | 951 | If the path is already in the manifest, its flags will be copied to |
|
950 | 952 | the new entry. |
|
951 | 953 | """ |
|
952 | 954 | |
|
953 | 955 | def __contains__(path): |
|
954 | 956 | """Whether a path exists in the manifest.""" |
|
955 | 957 | |
|
956 | 958 | def __delitem__(path): |
|
957 | 959 | """Remove a path from the manifest. |
|
958 | 960 | |
|
959 | 961 | Raises ``KeyError`` if the path is not in the manifest. |
|
960 | 962 | """ |
|
961 | 963 | |
|
962 | 964 | def __iter__(): |
|
963 | 965 | """Iterate over paths in the manifest.""" |
|
964 | 966 | |
|
965 | 967 | def iterkeys(): |
|
966 | 968 | """Iterate over paths in the manifest.""" |
|
967 | 969 | |
|
968 | 970 | def keys(): |
|
969 | 971 | """Obtain a list of paths in the manifest.""" |
|
970 | 972 | |
|
971 | 973 | def filesnotin(other, match=None): |
|
972 | 974 | """Obtain the set of paths in this manifest but not in another. |
|
973 | 975 | |
|
974 | 976 | ``match`` is an optional matcher function to be applied to both |
|
975 | 977 | manifests. |
|
976 | 978 | |
|
977 | 979 | Returns a set of paths. |
|
978 | 980 | """ |
|
979 | 981 | |
|
980 | 982 | def dirs(): |
|
981 | 983 | """Returns an object implementing the ``idirs`` interface.""" |
|
982 | 984 | |
|
983 | 985 | def hasdir(dir): |
|
984 | 986 | """Returns a bool indicating if a directory is in this manifest.""" |
|
985 | 987 | |
|
986 | 988 | def matches(match): |
|
987 | 989 | """Generate a new manifest filtered through a matcher. |
|
988 | 990 | |
|
989 | 991 | Returns an object conforming to the ``imanifestdict`` interface. |
|
990 | 992 | """ |
|
991 | 993 | |
|
992 | 994 | def walk(match): |
|
993 | 995 | """Generator of paths in manifest satisfying a matcher. |
|
994 | 996 | |
|
995 | 997 | This is equivalent to ``self.matches(match).iterkeys()`` except a new |
|
996 | 998 | manifest object is not created. |
|
997 | 999 | |
|
998 | 1000 | If the matcher has explicit files listed and they don't exist in |
|
999 | 1001 | the manifest, ``match.bad()`` is called for each missing file. |
|
1000 | 1002 | """ |
|
1001 | 1003 | |
|
1002 | 1004 | def diff(other, match=None, clean=False): |
|
1003 | 1005 | """Find differences between this manifest and another. |
|
1004 | 1006 | |
|
1005 | 1007 | This manifest is compared to ``other``. |
|
1006 | 1008 | |
|
1007 | 1009 | If ``match`` is provided, the two manifests are filtered against this |
|
1008 | 1010 | matcher and only entries satisfying the matcher are compared. |
|
1009 | 1011 | |
|
1010 | 1012 | If ``clean`` is True, unchanged files are included in the returned |
|
1011 | 1013 | object. |
|
1012 | 1014 | |
|
1013 | 1015 | Returns a dict with paths as keys and values of 2-tuples of 2-tuples of |
|
1014 | 1016 | the form ``((node1, flag1), (node2, flag2))`` where ``(node1, flag1)`` |
|
1015 | 1017 | represents the node and flags for this manifest and ``(node2, flag2)`` |
|
1016 | 1018 | are the same for the other manifest. |
|
1017 | 1019 | """ |
|
1018 | 1020 | |
|
1019 | 1021 | def setflag(path, flag): |
|
1020 | 1022 | """Set the flag value for a given path. |
|
1021 | 1023 | |
|
1022 | 1024 | Raises ``KeyError`` if the path is not already in the manifest. |
|
1023 | 1025 | """ |
|
1024 | 1026 | |
|
1025 | 1027 | def get(path, default=None): |
|
1026 | 1028 | """Obtain the node value for a path or a default value if missing.""" |
|
1027 | 1029 | |
|
1028 | 1030 | def flags(path, default=b''): |
|
1029 | 1031 | """Return the flags value for a path or a default value if missing.""" |
|
1030 | 1032 | |
|
1031 | 1033 | def copy(): |
|
1032 | 1034 | """Return a copy of this manifest.""" |
|
1033 | 1035 | |
|
1034 | 1036 | def items(): |
|
1035 | 1037 | """Returns an iterable of (path, node) for items in this manifest.""" |
|
1036 | 1038 | |
|
1037 | 1039 | def iteritems(): |
|
1038 | 1040 | """Identical to items().""" |
|
1039 | 1041 | |
|
1040 | 1042 | def iterentries(): |
|
1041 | 1043 | """Returns an iterable of (path, node, flags) for this manifest. |
|
1042 | 1044 | |
|
1043 | 1045 | Similar to ``iteritems()`` except items are a 3-tuple and include |
|
1044 | 1046 | flags. |
|
1045 | 1047 | """ |
|
1046 | 1048 | |
|
1047 | 1049 | def text(): |
|
1048 | 1050 | """Obtain the raw data representation for this manifest. |
|
1049 | 1051 | |
|
1050 | 1052 | Result is used to create a manifest revision. |
|
1051 | 1053 | """ |
|
1052 | 1054 | |
|
1053 | 1055 | def fastdelta(base, changes): |
|
1054 | 1056 | """Obtain a delta between this manifest and another given changes. |
|
1055 | 1057 | |
|
1056 | 1058 | ``base`` in the raw data representation for another manifest. |
|
1057 | 1059 | |
|
1058 | 1060 | ``changes`` is an iterable of ``(path, to_delete)``. |
|
1059 | 1061 | |
|
1060 | 1062 | Returns a 2-tuple containing ``bytearray(self.text())`` and the |
|
1061 | 1063 | delta between ``base`` and this manifest. |
|
1062 | 1064 | """ |
|
1063 | 1065 | |
|
1064 | 1066 | |
|
1065 | 1067 | class imanifestrevisionbase(interfaceutil.Interface): |
|
1066 | 1068 | """Base interface representing a single revision of a manifest. |
|
1067 | 1069 | |
|
1068 | 1070 | Should not be used as a primary interface: should always be inherited |
|
1069 | 1071 | as part of a larger interface. |
|
1070 | 1072 | """ |
|
1071 | 1073 | |
|
1072 | 1074 | def new(): |
|
1073 | 1075 | """Obtain a new manifest instance. |
|
1074 | 1076 | |
|
1075 | 1077 | Returns an object conforming to the ``imanifestrevisionwritable`` |
|
1076 | 1078 | interface. The instance will be associated with the same |
|
1077 | 1079 | ``imanifestlog`` collection as this instance. |
|
1078 | 1080 | """ |
|
1079 | 1081 | |
|
1080 | 1082 | def copy(): |
|
1081 | 1083 | """Obtain a copy of this manifest instance. |
|
1082 | 1084 | |
|
1083 | 1085 | Returns an object conforming to the ``imanifestrevisionwritable`` |
|
1084 | 1086 | interface. The instance will be associated with the same |
|
1085 | 1087 | ``imanifestlog`` collection as this instance. |
|
1086 | 1088 | """ |
|
1087 | 1089 | |
|
1088 | 1090 | def read(): |
|
1089 | 1091 | """Obtain the parsed manifest data structure. |
|
1090 | 1092 | |
|
1091 | 1093 | The returned object conforms to the ``imanifestdict`` interface. |
|
1092 | 1094 | """ |
|
1093 | 1095 | |
|
1094 | 1096 | |
|
1095 | 1097 | class imanifestrevisionstored(imanifestrevisionbase): |
|
1096 | 1098 | """Interface representing a manifest revision committed to storage.""" |
|
1097 | 1099 | |
|
1098 | 1100 | def node(): |
|
1099 | 1101 | """The binary node for this manifest.""" |
|
1100 | 1102 | |
|
1101 | 1103 | parents = interfaceutil.Attribute( |
|
1102 | 1104 | """List of binary nodes that are parents for this manifest revision.""" |
|
1103 | 1105 | ) |
|
1104 | 1106 | |
|
1105 | 1107 | def readdelta(shallow=False): |
|
1106 | 1108 | """Obtain the manifest data structure representing changes from parent. |
|
1107 | 1109 | |
|
1108 | 1110 | This manifest is compared to its 1st parent. A new manifest representing |
|
1109 | 1111 | those differences is constructed. |
|
1110 | 1112 | |
|
1111 | 1113 | The returned object conforms to the ``imanifestdict`` interface. |
|
1112 | 1114 | """ |
|
1113 | 1115 | |
|
1114 | 1116 | def readfast(shallow=False): |
|
1115 | 1117 | """Calls either ``read()`` or ``readdelta()``. |
|
1116 | 1118 | |
|
1117 | 1119 | The faster of the two options is called. |
|
1118 | 1120 | """ |
|
1119 | 1121 | |
|
1120 | 1122 | def find(key): |
|
1121 | 1123 | """Calls self.read().find(key)``. |
|
1122 | 1124 | |
|
1123 | 1125 | Returns a 2-tuple of ``(node, flags)`` or raises ``KeyError``. |
|
1124 | 1126 | """ |
|
1125 | 1127 | |
|
1126 | 1128 | |
|
1127 | 1129 | class imanifestrevisionwritable(imanifestrevisionbase): |
|
1128 | 1130 | """Interface representing a manifest revision that can be committed.""" |
|
1129 | 1131 | |
|
1130 | 1132 | def write(transaction, linkrev, p1node, p2node, added, removed, match=None): |
|
1131 | 1133 | """Add this revision to storage. |
|
1132 | 1134 | |
|
1133 | 1135 | Takes a transaction object, the changeset revision number it will |
|
1134 | 1136 | be associated with, its parent nodes, and lists of added and |
|
1135 | 1137 | removed paths. |
|
1136 | 1138 | |
|
1137 | 1139 | If match is provided, storage can choose not to inspect or write out |
|
1138 | 1140 | items that do not match. Storage is still required to be able to provide |
|
1139 | 1141 | the full manifest in the future for any directories written (these |
|
1140 | 1142 | manifests should not be "narrowed on disk"). |
|
1141 | 1143 | |
|
1142 | 1144 | Returns the binary node of the created revision. |
|
1143 | 1145 | """ |
|
1144 | 1146 | |
|
1145 | 1147 | |
|
1146 | 1148 | class imanifeststorage(interfaceutil.Interface): |
|
1147 | 1149 | """Storage interface for manifest data.""" |
|
1148 | 1150 | |
|
1149 | 1151 | tree = interfaceutil.Attribute( |
|
1150 | 1152 | """The path to the directory this manifest tracks. |
|
1151 | 1153 | |
|
1152 | 1154 | The empty bytestring represents the root manifest. |
|
1153 | 1155 | """ |
|
1154 | 1156 | ) |
|
1155 | 1157 | |
|
1156 | 1158 | index = interfaceutil.Attribute( |
|
1157 | 1159 | """An ``ifilerevisionssequence`` instance.""" |
|
1158 | 1160 | ) |
|
1159 | 1161 | |
|
1160 | 1162 | indexfile = interfaceutil.Attribute( |
|
1161 | 1163 | """Path of revlog index file. |
|
1162 | 1164 | |
|
1163 | 1165 | TODO this is revlog specific and should not be exposed. |
|
1164 | 1166 | """ |
|
1165 | 1167 | ) |
|
1166 | 1168 | |
|
1167 | 1169 | opener = interfaceutil.Attribute( |
|
1168 | 1170 | """VFS opener to use to access underlying files used for storage. |
|
1169 | 1171 | |
|
1170 | 1172 | TODO this is revlog specific and should not be exposed. |
|
1171 | 1173 | """ |
|
1172 | 1174 | ) |
|
1173 | 1175 | |
|
1174 | 1176 | version = interfaceutil.Attribute( |
|
1175 | 1177 | """Revlog version number. |
|
1176 | 1178 | |
|
1177 | 1179 | TODO this is revlog specific and should not be exposed. |
|
1178 | 1180 | """ |
|
1179 | 1181 | ) |
|
1180 | 1182 | |
|
1181 | 1183 | _generaldelta = interfaceutil.Attribute( |
|
1182 | 1184 | """Whether generaldelta storage is being used. |
|
1183 | 1185 | |
|
1184 | 1186 | TODO this is revlog specific and should not be exposed. |
|
1185 | 1187 | """ |
|
1186 | 1188 | ) |
|
1187 | 1189 | |
|
1188 | 1190 | fulltextcache = interfaceutil.Attribute( |
|
1189 | 1191 | """Dict with cache of fulltexts. |
|
1190 | 1192 | |
|
1191 | 1193 | TODO this doesn't feel appropriate for the storage interface. |
|
1192 | 1194 | """ |
|
1193 | 1195 | ) |
|
1194 | 1196 | |
|
1195 | 1197 | def __len__(): |
|
1196 | 1198 | """Obtain the number of revisions stored for this manifest.""" |
|
1197 | 1199 | |
|
1198 | 1200 | def __iter__(): |
|
1199 | 1201 | """Iterate over revision numbers for this manifest.""" |
|
1200 | 1202 | |
|
1201 | 1203 | def rev(node): |
|
1202 | 1204 | """Obtain the revision number given a binary node. |
|
1203 | 1205 | |
|
1204 | 1206 | Raises ``error.LookupError`` if the node is not known. |
|
1205 | 1207 | """ |
|
1206 | 1208 | |
|
1207 | 1209 | def node(rev): |
|
1208 | 1210 | """Obtain the node value given a revision number. |
|
1209 | 1211 | |
|
1210 | 1212 | Raises ``error.LookupError`` if the revision is not known. |
|
1211 | 1213 | """ |
|
1212 | 1214 | |
|
1213 | 1215 | def lookup(value): |
|
1214 | 1216 | """Attempt to resolve a value to a node. |
|
1215 | 1217 | |
|
1216 | 1218 | Value can be a binary node, hex node, revision number, or a bytes |
|
1217 | 1219 | that can be converted to an integer. |
|
1218 | 1220 | |
|
1219 | 1221 | Raises ``error.LookupError`` if a ndoe could not be resolved. |
|
1220 | 1222 | """ |
|
1221 | 1223 | |
|
1222 | 1224 | def parents(node): |
|
1223 | 1225 | """Returns a 2-tuple of parent nodes for a node. |
|
1224 | 1226 | |
|
1225 | 1227 | Values will be ``nullid`` if the parent is empty. |
|
1226 | 1228 | """ |
|
1227 | 1229 | |
|
1228 | 1230 | def parentrevs(rev): |
|
1229 | 1231 | """Like parents() but operates on revision numbers.""" |
|
1230 | 1232 | |
|
1231 | 1233 | def linkrev(rev): |
|
1232 | 1234 | """Obtain the changeset revision number a revision is linked to.""" |
|
1233 | 1235 | |
|
1234 | 1236 | def revision(node, _df=None, raw=False): |
|
1235 | 1237 | """Obtain fulltext data for a node.""" |
|
1236 | 1238 | |
|
1237 | 1239 | def rawdata(node, _df=None): |
|
1238 | 1240 | """Obtain raw data for a node.""" |
|
1239 | 1241 | |
|
1240 | 1242 | def revdiff(rev1, rev2): |
|
1241 | 1243 | """Obtain a delta between two revision numbers. |
|
1242 | 1244 | |
|
1243 | 1245 | The returned data is the result of ``bdiff.bdiff()`` on the raw |
|
1244 | 1246 | revision data. |
|
1245 | 1247 | """ |
|
1246 | 1248 | |
|
1247 | 1249 | def cmp(node, fulltext): |
|
1248 | 1250 | """Compare fulltext to another revision. |
|
1249 | 1251 | |
|
1250 | 1252 | Returns True if the fulltext is different from what is stored. |
|
1251 | 1253 | """ |
|
1252 | 1254 | |
|
1253 | 1255 | def emitrevisions( |
|
1254 | 1256 | nodes, |
|
1255 | 1257 | nodesorder=None, |
|
1256 | 1258 | revisiondata=False, |
|
1257 | 1259 | assumehaveparentrevisions=False, |
|
1258 | 1260 | ): |
|
1259 | 1261 | """Produce ``irevisiondelta`` describing revisions. |
|
1260 | 1262 | |
|
1261 | 1263 | See the documentation for ``ifiledata`` for more. |
|
1262 | 1264 | """ |
|
1263 | 1265 | |
|
1264 | 1266 | def addgroup(deltas, linkmapper, transaction, addrevisioncb=None): |
|
1265 | 1267 | """Process a series of deltas for storage. |
|
1266 | 1268 | |
|
1267 | 1269 | See the documentation in ``ifilemutation`` for more. |
|
1268 | 1270 | """ |
|
1269 | 1271 | |
|
1270 | 1272 | def rawsize(rev): |
|
1271 | 1273 | """Obtain the size of tracked data. |
|
1272 | 1274 | |
|
1273 | 1275 | Is equivalent to ``len(m.rawdata(node))``. |
|
1274 | 1276 | |
|
1275 | 1277 | TODO this method is only used by upgrade code and may be removed. |
|
1276 | 1278 | """ |
|
1277 | 1279 | |
|
1278 | 1280 | def getstrippoint(minlink): |
|
1279 | 1281 | """Find minimum revision that must be stripped to strip a linkrev. |
|
1280 | 1282 | |
|
1281 | 1283 | See the documentation in ``ifilemutation`` for more. |
|
1282 | 1284 | """ |
|
1283 | 1285 | |
|
1284 | 1286 | def strip(minlink, transaction): |
|
1285 | 1287 | """Remove storage of items starting at a linkrev. |
|
1286 | 1288 | |
|
1287 | 1289 | See the documentation in ``ifilemutation`` for more. |
|
1288 | 1290 | """ |
|
1289 | 1291 | |
|
1290 | 1292 | def checksize(): |
|
1291 | 1293 | """Obtain the expected sizes of backing files. |
|
1292 | 1294 | |
|
1293 | 1295 | TODO this is used by verify and it should not be part of the interface. |
|
1294 | 1296 | """ |
|
1295 | 1297 | |
|
1296 | 1298 | def files(): |
|
1297 | 1299 | """Obtain paths that are backing storage for this manifest. |
|
1298 | 1300 | |
|
1299 | 1301 | TODO this is used by verify and there should probably be a better API |
|
1300 | 1302 | for this functionality. |
|
1301 | 1303 | """ |
|
1302 | 1304 | |
|
1303 | 1305 | def deltaparent(rev): |
|
1304 | 1306 | """Obtain the revision that a revision is delta'd against. |
|
1305 | 1307 | |
|
1306 | 1308 | TODO delta encoding is an implementation detail of storage and should |
|
1307 | 1309 | not be exposed to the storage interface. |
|
1308 | 1310 | """ |
|
1309 | 1311 | |
|
1310 | 1312 | def clone(tr, dest, **kwargs): |
|
1311 | 1313 | """Clone this instance to another.""" |
|
1312 | 1314 | |
|
1313 | 1315 | def clearcaches(clear_persisted_data=False): |
|
1314 | 1316 | """Clear any caches associated with this instance.""" |
|
1315 | 1317 | |
|
1316 | 1318 | def dirlog(d): |
|
1317 | 1319 | """Obtain a manifest storage instance for a tree.""" |
|
1318 | 1320 | |
|
1319 | 1321 | def add( |
|
1320 | 1322 | m, transaction, link, p1, p2, added, removed, readtree=None, match=None |
|
1321 | 1323 | ): |
|
1322 | 1324 | """Add a revision to storage. |
|
1323 | 1325 | |
|
1324 | 1326 | ``m`` is an object conforming to ``imanifestdict``. |
|
1325 | 1327 | |
|
1326 | 1328 | ``link`` is the linkrev revision number. |
|
1327 | 1329 | |
|
1328 | 1330 | ``p1`` and ``p2`` are the parent revision numbers. |
|
1329 | 1331 | |
|
1330 | 1332 | ``added`` and ``removed`` are iterables of added and removed paths, |
|
1331 | 1333 | respectively. |
|
1332 | 1334 | |
|
1333 | 1335 | ``readtree`` is a function that can be used to read the child tree(s) |
|
1334 | 1336 | when recursively writing the full tree structure when using |
|
1335 | 1337 | treemanifets. |
|
1336 | 1338 | |
|
1337 | 1339 | ``match`` is a matcher that can be used to hint to storage that not all |
|
1338 | 1340 | paths must be inspected; this is an optimization and can be safely |
|
1339 | 1341 | ignored. Note that the storage must still be able to reproduce a full |
|
1340 | 1342 | manifest including files that did not match. |
|
1341 | 1343 | """ |
|
1342 | 1344 | |
|
1343 | 1345 | def storageinfo( |
|
1344 | 1346 | exclusivefiles=False, |
|
1345 | 1347 | sharedfiles=False, |
|
1346 | 1348 | revisionscount=False, |
|
1347 | 1349 | trackedsize=False, |
|
1348 | 1350 | storedsize=False, |
|
1349 | 1351 | ): |
|
1350 | 1352 | """Obtain information about storage for this manifest's data. |
|
1351 | 1353 | |
|
1352 | 1354 | See ``ifilestorage.storageinfo()`` for a description of this method. |
|
1353 | 1355 | This one behaves the same way, except for manifest data. |
|
1354 | 1356 | """ |
|
1355 | 1357 | |
|
1356 | 1358 | |
|
1357 | 1359 | class imanifestlog(interfaceutil.Interface): |
|
1358 | 1360 | """Interface representing a collection of manifest snapshots. |
|
1359 | 1361 | |
|
1360 | 1362 | Represents the root manifest in a repository. |
|
1361 | 1363 | |
|
1362 | 1364 | Also serves as a means to access nested tree manifests and to cache |
|
1363 | 1365 | tree manifests. |
|
1364 | 1366 | """ |
|
1365 | 1367 | |
|
1366 | 1368 | def __getitem__(node): |
|
1367 | 1369 | """Obtain a manifest instance for a given binary node. |
|
1368 | 1370 | |
|
1369 | 1371 | Equivalent to calling ``self.get('', node)``. |
|
1370 | 1372 | |
|
1371 | 1373 | The returned object conforms to the ``imanifestrevisionstored`` |
|
1372 | 1374 | interface. |
|
1373 | 1375 | """ |
|
1374 | 1376 | |
|
1375 | 1377 | def get(tree, node, verify=True): |
|
1376 | 1378 | """Retrieve the manifest instance for a given directory and binary node. |
|
1377 | 1379 | |
|
1378 | 1380 | ``node`` always refers to the node of the root manifest (which will be |
|
1379 | 1381 | the only manifest if flat manifests are being used). |
|
1380 | 1382 | |
|
1381 | 1383 | If ``tree`` is the empty string, the root manifest is returned. |
|
1382 | 1384 | Otherwise the manifest for the specified directory will be returned |
|
1383 | 1385 | (requires tree manifests). |
|
1384 | 1386 | |
|
1385 | 1387 | If ``verify`` is True, ``LookupError`` is raised if the node is not |
|
1386 | 1388 | known. |
|
1387 | 1389 | |
|
1388 | 1390 | The returned object conforms to the ``imanifestrevisionstored`` |
|
1389 | 1391 | interface. |
|
1390 | 1392 | """ |
|
1391 | 1393 | |
|
1392 | 1394 | def getstorage(tree): |
|
1393 | 1395 | """Retrieve an interface to storage for a particular tree. |
|
1394 | 1396 | |
|
1395 | 1397 | If ``tree`` is the empty bytestring, storage for the root manifest will |
|
1396 | 1398 | be returned. Otherwise storage for a tree manifest is returned. |
|
1397 | 1399 | |
|
1398 | 1400 | TODO formalize interface for returned object. |
|
1399 | 1401 | """ |
|
1400 | 1402 | |
|
1401 | 1403 | def clearcaches(): |
|
1402 | 1404 | """Clear caches associated with this collection.""" |
|
1403 | 1405 | |
|
1404 | 1406 | def rev(node): |
|
1405 | 1407 | """Obtain the revision number for a binary node. |
|
1406 | 1408 | |
|
1407 | 1409 | Raises ``error.LookupError`` if the node is not known. |
|
1408 | 1410 | """ |
|
1409 | 1411 | |
|
1410 | 1412 | |
|
1411 | 1413 | class ilocalrepositoryfilestorage(interfaceutil.Interface): |
|
1412 | 1414 | """Local repository sub-interface providing access to tracked file storage. |
|
1413 | 1415 | |
|
1414 | 1416 | This interface defines how a repository accesses storage for a single |
|
1415 | 1417 | tracked file path. |
|
1416 | 1418 | """ |
|
1417 | 1419 | |
|
1418 | 1420 | def file(f): |
|
1419 | 1421 | """Obtain a filelog for a tracked path. |
|
1420 | 1422 | |
|
1421 | 1423 | The returned type conforms to the ``ifilestorage`` interface. |
|
1422 | 1424 | """ |
|
1423 | 1425 | |
|
1424 | 1426 | |
|
1425 | 1427 | class ilocalrepositorymain(interfaceutil.Interface): |
|
1426 | 1428 | """Main interface for local repositories. |
|
1427 | 1429 | |
|
1428 | 1430 | This currently captures the reality of things - not how things should be. |
|
1429 | 1431 | """ |
|
1430 | 1432 | |
|
1431 | 1433 | supportedformats = interfaceutil.Attribute( |
|
1432 | 1434 | """Set of requirements that apply to stream clone. |
|
1433 | 1435 | |
|
1434 | 1436 | This is actually a class attribute and is shared among all instances. |
|
1435 | 1437 | """ |
|
1436 | 1438 | ) |
|
1437 | 1439 | |
|
1438 | 1440 | supported = interfaceutil.Attribute( |
|
1439 | 1441 | """Set of requirements that this repo is capable of opening.""" |
|
1440 | 1442 | ) |
|
1441 | 1443 | |
|
1442 | 1444 | requirements = interfaceutil.Attribute( |
|
1443 | 1445 | """Set of requirements this repo uses.""" |
|
1444 | 1446 | ) |
|
1445 | 1447 | |
|
1446 | 1448 | features = interfaceutil.Attribute( |
|
1447 | 1449 | """Set of "features" this repository supports. |
|
1448 | 1450 | |
|
1449 | 1451 | A "feature" is a loosely-defined term. It can refer to a feature |
|
1450 | 1452 | in the classical sense or can describe an implementation detail |
|
1451 | 1453 | of the repository. For example, a ``readonly`` feature may denote |
|
1452 | 1454 | the repository as read-only. Or a ``revlogfilestore`` feature may |
|
1453 | 1455 | denote that the repository is using revlogs for file storage. |
|
1454 | 1456 | |
|
1455 | 1457 | The intent of features is to provide a machine-queryable mechanism |
|
1456 | 1458 | for repo consumers to test for various repository characteristics. |
|
1457 | 1459 | |
|
1458 | 1460 | Features are similar to ``requirements``. The main difference is that |
|
1459 | 1461 | requirements are stored on-disk and represent requirements to open the |
|
1460 | 1462 | repository. Features are more run-time capabilities of the repository |
|
1461 | 1463 | and more granular capabilities (which may be derived from requirements). |
|
1462 | 1464 | """ |
|
1463 | 1465 | ) |
|
1464 | 1466 | |
|
1465 | 1467 | filtername = interfaceutil.Attribute( |
|
1466 | 1468 | """Name of the repoview that is active on this repo.""" |
|
1467 | 1469 | ) |
|
1468 | 1470 | |
|
1469 | 1471 | wvfs = interfaceutil.Attribute( |
|
1470 | 1472 | """VFS used to access the working directory.""" |
|
1471 | 1473 | ) |
|
1472 | 1474 | |
|
1473 | 1475 | vfs = interfaceutil.Attribute( |
|
1474 | 1476 | """VFS rooted at the .hg directory. |
|
1475 | 1477 | |
|
1476 | 1478 | Used to access repository data not in the store. |
|
1477 | 1479 | """ |
|
1478 | 1480 | ) |
|
1479 | 1481 | |
|
1480 | 1482 | svfs = interfaceutil.Attribute( |
|
1481 | 1483 | """VFS rooted at the store. |
|
1482 | 1484 | |
|
1483 | 1485 | Used to access repository data in the store. Typically .hg/store. |
|
1484 | 1486 | But can point elsewhere if the store is shared. |
|
1485 | 1487 | """ |
|
1486 | 1488 | ) |
|
1487 | 1489 | |
|
1488 | 1490 | root = interfaceutil.Attribute( |
|
1489 | 1491 | """Path to the root of the working directory.""" |
|
1490 | 1492 | ) |
|
1491 | 1493 | |
|
1492 | 1494 | path = interfaceutil.Attribute("""Path to the .hg directory.""") |
|
1493 | 1495 | |
|
1494 | 1496 | origroot = interfaceutil.Attribute( |
|
1495 | 1497 | """The filesystem path that was used to construct the repo.""" |
|
1496 | 1498 | ) |
|
1497 | 1499 | |
|
1498 | 1500 | auditor = interfaceutil.Attribute( |
|
1499 | 1501 | """A pathauditor for the working directory. |
|
1500 | 1502 | |
|
1501 | 1503 | This checks if a path refers to a nested repository. |
|
1502 | 1504 | |
|
1503 | 1505 | Operates on the filesystem. |
|
1504 | 1506 | """ |
|
1505 | 1507 | ) |
|
1506 | 1508 | |
|
1507 | 1509 | nofsauditor = interfaceutil.Attribute( |
|
1508 | 1510 | """A pathauditor for the working directory. |
|
1509 | 1511 | |
|
1510 | 1512 | This is like ``auditor`` except it doesn't do filesystem checks. |
|
1511 | 1513 | """ |
|
1512 | 1514 | ) |
|
1513 | 1515 | |
|
1514 | 1516 | baseui = interfaceutil.Attribute( |
|
1515 | 1517 | """Original ui instance passed into constructor.""" |
|
1516 | 1518 | ) |
|
1517 | 1519 | |
|
1518 | 1520 | ui = interfaceutil.Attribute("""Main ui instance for this instance.""") |
|
1519 | 1521 | |
|
1520 | 1522 | sharedpath = interfaceutil.Attribute( |
|
1521 | 1523 | """Path to the .hg directory of the repo this repo was shared from.""" |
|
1522 | 1524 | ) |
|
1523 | 1525 | |
|
1524 | 1526 | store = interfaceutil.Attribute("""A store instance.""") |
|
1525 | 1527 | |
|
1526 | 1528 | spath = interfaceutil.Attribute("""Path to the store.""") |
|
1527 | 1529 | |
|
1528 | 1530 | sjoin = interfaceutil.Attribute("""Alias to self.store.join.""") |
|
1529 | 1531 | |
|
1530 | 1532 | cachevfs = interfaceutil.Attribute( |
|
1531 | 1533 | """A VFS used to access the cache directory. |
|
1532 | 1534 | |
|
1533 | 1535 | Typically .hg/cache. |
|
1534 | 1536 | """ |
|
1535 | 1537 | ) |
|
1536 | 1538 | |
|
1537 | 1539 | wcachevfs = interfaceutil.Attribute( |
|
1538 | 1540 | """A VFS used to access the cache directory dedicated to working copy |
|
1539 | 1541 | |
|
1540 | 1542 | Typically .hg/wcache. |
|
1541 | 1543 | """ |
|
1542 | 1544 | ) |
|
1543 | 1545 | |
|
1544 | 1546 | filteredrevcache = interfaceutil.Attribute( |
|
1545 | 1547 | """Holds sets of revisions to be filtered.""" |
|
1546 | 1548 | ) |
|
1547 | 1549 | |
|
1548 | 1550 | names = interfaceutil.Attribute("""A ``namespaces`` instance.""") |
|
1549 | 1551 | |
|
1550 | 1552 | filecopiesmode = interfaceutil.Attribute( |
|
1551 | 1553 | """The way files copies should be dealt with in this repo.""" |
|
1552 | 1554 | ) |
|
1553 | 1555 | |
|
1554 | 1556 | def close(): |
|
1555 | 1557 | """Close the handle on this repository.""" |
|
1556 | 1558 | |
|
1557 | 1559 | def peer(): |
|
1558 | 1560 | """Obtain an object conforming to the ``peer`` interface.""" |
|
1559 | 1561 | |
|
1560 | 1562 | def unfiltered(): |
|
1561 | 1563 | """Obtain an unfiltered/raw view of this repo.""" |
|
1562 | 1564 | |
|
1563 | 1565 | def filtered(name, visibilityexceptions=None): |
|
1564 | 1566 | """Obtain a named view of this repository.""" |
|
1565 | 1567 | |
|
1566 | 1568 | obsstore = interfaceutil.Attribute("""A store of obsolescence data.""") |
|
1567 | 1569 | |
|
1568 | 1570 | changelog = interfaceutil.Attribute("""A handle on the changelog revlog.""") |
|
1569 | 1571 | |
|
1570 | 1572 | manifestlog = interfaceutil.Attribute( |
|
1571 | 1573 | """An instance conforming to the ``imanifestlog`` interface. |
|
1572 | 1574 | |
|
1573 | 1575 | Provides access to manifests for the repository. |
|
1574 | 1576 | """ |
|
1575 | 1577 | ) |
|
1576 | 1578 | |
|
1577 | 1579 | dirstate = interfaceutil.Attribute("""Working directory state.""") |
|
1578 | 1580 | |
|
1579 | 1581 | narrowpats = interfaceutil.Attribute( |
|
1580 | 1582 | """Matcher patterns for this repository's narrowspec.""" |
|
1581 | 1583 | ) |
|
1582 | 1584 | |
|
1583 | 1585 | def narrowmatch(match=None, includeexact=False): |
|
1584 | 1586 | """Obtain a matcher for the narrowspec.""" |
|
1585 | 1587 | |
|
1586 | 1588 | def setnarrowpats(newincludes, newexcludes): |
|
1587 | 1589 | """Define the narrowspec for this repository.""" |
|
1588 | 1590 | |
|
1589 | 1591 | def __getitem__(changeid): |
|
1590 | 1592 | """Try to resolve a changectx.""" |
|
1591 | 1593 | |
|
1592 | 1594 | def __contains__(changeid): |
|
1593 | 1595 | """Whether a changeset exists.""" |
|
1594 | 1596 | |
|
1595 | 1597 | def __nonzero__(): |
|
1596 | 1598 | """Always returns True.""" |
|
1597 | 1599 | return True |
|
1598 | 1600 | |
|
1599 | 1601 | __bool__ = __nonzero__ |
|
1600 | 1602 | |
|
1601 | 1603 | def __len__(): |
|
1602 | 1604 | """Returns the number of changesets in the repo.""" |
|
1603 | 1605 | |
|
1604 | 1606 | def __iter__(): |
|
1605 | 1607 | """Iterate over revisions in the changelog.""" |
|
1606 | 1608 | |
|
1607 | 1609 | def revs(expr, *args): |
|
1608 | 1610 | """Evaluate a revset. |
|
1609 | 1611 | |
|
1610 | 1612 | Emits revisions. |
|
1611 | 1613 | """ |
|
1612 | 1614 | |
|
1613 | 1615 | def set(expr, *args): |
|
1614 | 1616 | """Evaluate a revset. |
|
1615 | 1617 | |
|
1616 | 1618 | Emits changectx instances. |
|
1617 | 1619 | """ |
|
1618 | 1620 | |
|
1619 | 1621 | def anyrevs(specs, user=False, localalias=None): |
|
1620 | 1622 | """Find revisions matching one of the given revsets.""" |
|
1621 | 1623 | |
|
1622 | 1624 | def url(): |
|
1623 | 1625 | """Returns a string representing the location of this repo.""" |
|
1624 | 1626 | |
|
1625 | 1627 | def hook(name, throw=False, **args): |
|
1626 | 1628 | """Call a hook.""" |
|
1627 | 1629 | |
|
1628 | 1630 | def tags(): |
|
1629 | 1631 | """Return a mapping of tag to node.""" |
|
1630 | 1632 | |
|
1631 | 1633 | def tagtype(tagname): |
|
1632 | 1634 | """Return the type of a given tag.""" |
|
1633 | 1635 | |
|
1634 | 1636 | def tagslist(): |
|
1635 | 1637 | """Return a list of tags ordered by revision.""" |
|
1636 | 1638 | |
|
1637 | 1639 | def nodetags(node): |
|
1638 | 1640 | """Return the tags associated with a node.""" |
|
1639 | 1641 | |
|
1640 | 1642 | def nodebookmarks(node): |
|
1641 | 1643 | """Return the list of bookmarks pointing to the specified node.""" |
|
1642 | 1644 | |
|
1643 | 1645 | def branchmap(): |
|
1644 | 1646 | """Return a mapping of branch to heads in that branch.""" |
|
1645 | 1647 | |
|
1646 | 1648 | def revbranchcache(): |
|
1647 | 1649 | pass |
|
1648 | 1650 | |
|
1649 | 1651 | def branchtip(branchtip, ignoremissing=False): |
|
1650 | 1652 | """Return the tip node for a given branch.""" |
|
1651 | 1653 | |
|
1652 | 1654 | def lookup(key): |
|
1653 | 1655 | """Resolve the node for a revision.""" |
|
1654 | 1656 | |
|
1655 | 1657 | def lookupbranch(key): |
|
1656 | 1658 | """Look up the branch name of the given revision or branch name.""" |
|
1657 | 1659 | |
|
1658 | 1660 | def known(nodes): |
|
1659 | 1661 | """Determine whether a series of nodes is known. |
|
1660 | 1662 | |
|
1661 | 1663 | Returns a list of bools. |
|
1662 | 1664 | """ |
|
1663 | 1665 | |
|
1664 | 1666 | def local(): |
|
1665 | 1667 | """Whether the repository is local.""" |
|
1666 | 1668 | return True |
|
1667 | 1669 | |
|
1668 | 1670 | def publishing(): |
|
1669 | 1671 | """Whether the repository is a publishing repository.""" |
|
1670 | 1672 | |
|
1671 | 1673 | def cancopy(): |
|
1672 | 1674 | pass |
|
1673 | 1675 | |
|
1674 | 1676 | def shared(): |
|
1675 | 1677 | """The type of shared repository or None.""" |
|
1676 | 1678 | |
|
1677 | 1679 | def wjoin(f, *insidef): |
|
1678 | 1680 | """Calls self.vfs.reljoin(self.root, f, *insidef)""" |
|
1679 | 1681 | |
|
1680 | 1682 | def setparents(p1, p2): |
|
1681 | 1683 | """Set the parent nodes of the working directory.""" |
|
1682 | 1684 | |
|
1683 | 1685 | def filectx(path, changeid=None, fileid=None): |
|
1684 | 1686 | """Obtain a filectx for the given file revision.""" |
|
1685 | 1687 | |
|
1686 | 1688 | def getcwd(): |
|
1687 | 1689 | """Obtain the current working directory from the dirstate.""" |
|
1688 | 1690 | |
|
1689 | 1691 | def pathto(f, cwd=None): |
|
1690 | 1692 | """Obtain the relative path to a file.""" |
|
1691 | 1693 | |
|
1692 | 1694 | def adddatafilter(name, fltr): |
|
1693 | 1695 | pass |
|
1694 | 1696 | |
|
1695 | 1697 | def wread(filename): |
|
1696 | 1698 | """Read a file from wvfs, using data filters.""" |
|
1697 | 1699 | |
|
1698 | 1700 | def wwrite(filename, data, flags, backgroundclose=False, **kwargs): |
|
1699 | 1701 | """Write data to a file in the wvfs, using data filters.""" |
|
1700 | 1702 | |
|
1701 | 1703 | def wwritedata(filename, data): |
|
1702 | 1704 | """Resolve data for writing to the wvfs, using data filters.""" |
|
1703 | 1705 | |
|
1704 | 1706 | def currenttransaction(): |
|
1705 | 1707 | """Obtain the current transaction instance or None.""" |
|
1706 | 1708 | |
|
1707 | 1709 | def transaction(desc, report=None): |
|
1708 | 1710 | """Open a new transaction to write to the repository.""" |
|
1709 | 1711 | |
|
1710 | 1712 | def undofiles(): |
|
1711 | 1713 | """Returns a list of (vfs, path) for files to undo transactions.""" |
|
1712 | 1714 | |
|
1713 | 1715 | def recover(): |
|
1714 | 1716 | """Roll back an interrupted transaction.""" |
|
1715 | 1717 | |
|
1716 | 1718 | def rollback(dryrun=False, force=False): |
|
1717 | 1719 | """Undo the last transaction. |
|
1718 | 1720 | |
|
1719 | 1721 | DANGEROUS. |
|
1720 | 1722 | """ |
|
1721 | 1723 | |
|
1722 | 1724 | def updatecaches(tr=None, full=False): |
|
1723 | 1725 | """Warm repo caches.""" |
|
1724 | 1726 | |
|
1725 | 1727 | def invalidatecaches(): |
|
1726 | 1728 | """Invalidate cached data due to the repository mutating.""" |
|
1727 | 1729 | |
|
1728 | 1730 | def invalidatevolatilesets(): |
|
1729 | 1731 | pass |
|
1730 | 1732 | |
|
1731 | 1733 | def invalidatedirstate(): |
|
1732 | 1734 | """Invalidate the dirstate.""" |
|
1733 | 1735 | |
|
1734 | 1736 | def invalidate(clearfilecache=False): |
|
1735 | 1737 | pass |
|
1736 | 1738 | |
|
1737 | 1739 | def invalidateall(): |
|
1738 | 1740 | pass |
|
1739 | 1741 | |
|
1740 | 1742 | def lock(wait=True): |
|
1741 | 1743 | """Lock the repository store and return a lock instance.""" |
|
1742 | 1744 | |
|
1743 | 1745 | def wlock(wait=True): |
|
1744 | 1746 | """Lock the non-store parts of the repository.""" |
|
1745 | 1747 | |
|
1746 | 1748 | def currentwlock(): |
|
1747 | 1749 | """Return the wlock if it's held or None.""" |
|
1748 | 1750 | |
|
1749 | 1751 | def checkcommitpatterns(wctx, match, status, fail): |
|
1750 | 1752 | pass |
|
1751 | 1753 | |
|
1752 | 1754 | def commit( |
|
1753 | 1755 | text=b'', |
|
1754 | 1756 | user=None, |
|
1755 | 1757 | date=None, |
|
1756 | 1758 | match=None, |
|
1757 | 1759 | force=False, |
|
1758 | 1760 | editor=False, |
|
1759 | 1761 | extra=None, |
|
1760 | 1762 | ): |
|
1761 | 1763 | """Add a new revision to the repository.""" |
|
1762 | 1764 | |
|
1763 | 1765 | def commitctx(ctx, error=False, origctx=None): |
|
1764 | 1766 | """Commit a commitctx instance to the repository.""" |
|
1765 | 1767 | |
|
1766 | 1768 | def destroying(): |
|
1767 | 1769 | """Inform the repository that nodes are about to be destroyed.""" |
|
1768 | 1770 | |
|
1769 | 1771 | def destroyed(): |
|
1770 | 1772 | """Inform the repository that nodes have been destroyed.""" |
|
1771 | 1773 | |
|
1772 | 1774 | def status( |
|
1773 | 1775 | node1=b'.', |
|
1774 | 1776 | node2=None, |
|
1775 | 1777 | match=None, |
|
1776 | 1778 | ignored=False, |
|
1777 | 1779 | clean=False, |
|
1778 | 1780 | unknown=False, |
|
1779 | 1781 | listsubrepos=False, |
|
1780 | 1782 | ): |
|
1781 | 1783 | """Convenience method to call repo[x].status().""" |
|
1782 | 1784 | |
|
1783 | 1785 | def addpostdsstatus(ps): |
|
1784 | 1786 | pass |
|
1785 | 1787 | |
|
1786 | 1788 | def postdsstatus(): |
|
1787 | 1789 | pass |
|
1788 | 1790 | |
|
1789 | 1791 | def clearpostdsstatus(): |
|
1790 | 1792 | pass |
|
1791 | 1793 | |
|
1792 | 1794 | def heads(start=None): |
|
1793 | 1795 | """Obtain list of nodes that are DAG heads.""" |
|
1794 | 1796 | |
|
1795 | 1797 | def branchheads(branch=None, start=None, closed=False): |
|
1796 | 1798 | pass |
|
1797 | 1799 | |
|
1798 | 1800 | def branches(nodes): |
|
1799 | 1801 | pass |
|
1800 | 1802 | |
|
1801 | 1803 | def between(pairs): |
|
1802 | 1804 | pass |
|
1803 | 1805 | |
|
1804 | 1806 | def checkpush(pushop): |
|
1805 | 1807 | pass |
|
1806 | 1808 | |
|
1807 | 1809 | prepushoutgoinghooks = interfaceutil.Attribute("""util.hooks instance.""") |
|
1808 | 1810 | |
|
1809 | 1811 | def pushkey(namespace, key, old, new): |
|
1810 | 1812 | pass |
|
1811 | 1813 | |
|
1812 | 1814 | def listkeys(namespace): |
|
1813 | 1815 | pass |
|
1814 | 1816 | |
|
1815 | 1817 | def debugwireargs(one, two, three=None, four=None, five=None): |
|
1816 | 1818 | pass |
|
1817 | 1819 | |
|
1818 | 1820 | def savecommitmessage(text): |
|
1819 | 1821 | pass |
|
1820 | 1822 | |
|
1821 | 1823 | |
|
1822 | 1824 | class completelocalrepository( |
|
1823 | 1825 | ilocalrepositorymain, ilocalrepositoryfilestorage |
|
1824 | 1826 | ): |
|
1825 | 1827 | """Complete interface for a local repository.""" |
|
1826 | 1828 | |
|
1827 | 1829 | |
|
1828 | 1830 | class iwireprotocolcommandcacher(interfaceutil.Interface): |
|
1829 | 1831 | """Represents a caching backend for wire protocol commands. |
|
1830 | 1832 | |
|
1831 | 1833 | Wire protocol version 2 supports transparent caching of many commands. |
|
1832 | 1834 | To leverage this caching, servers can activate objects that cache |
|
1833 | 1835 | command responses. Objects handle both cache writing and reading. |
|
1834 | 1836 | This interface defines how that response caching mechanism works. |
|
1835 | 1837 | |
|
1836 | 1838 | Wire protocol version 2 commands emit a series of objects that are |
|
1837 | 1839 | serialized and sent to the client. The caching layer exists between |
|
1838 | 1840 | the invocation of the command function and the sending of its output |
|
1839 | 1841 | objects to an output layer. |
|
1840 | 1842 | |
|
1841 | 1843 | Instances of this interface represent a binding to a cache that |
|
1842 | 1844 | can serve a response (in place of calling a command function) and/or |
|
1843 | 1845 | write responses to a cache for subsequent use. |
|
1844 | 1846 | |
|
1845 | 1847 | When a command request arrives, the following happens with regards |
|
1846 | 1848 | to this interface: |
|
1847 | 1849 | |
|
1848 | 1850 | 1. The server determines whether the command request is cacheable. |
|
1849 | 1851 | 2. If it is, an instance of this interface is spawned. |
|
1850 | 1852 | 3. The cacher is activated in a context manager (``__enter__`` is called). |
|
1851 | 1853 | 4. A cache *key* for that request is derived. This will call the |
|
1852 | 1854 | instance's ``adjustcachekeystate()`` method so the derivation |
|
1853 | 1855 | can be influenced. |
|
1854 | 1856 | 5. The cacher is informed of the derived cache key via a call to |
|
1855 | 1857 | ``setcachekey()``. |
|
1856 | 1858 | 6. The cacher's ``lookup()`` method is called to test for presence of |
|
1857 | 1859 | the derived key in the cache. |
|
1858 | 1860 | 7. If ``lookup()`` returns a hit, that cached result is used in place |
|
1859 | 1861 | of invoking the command function. ``__exit__`` is called and the instance |
|
1860 | 1862 | is discarded. |
|
1861 | 1863 | 8. The command function is invoked. |
|
1862 | 1864 | 9. ``onobject()`` is called for each object emitted by the command |
|
1863 | 1865 | function. |
|
1864 | 1866 | 10. After the final object is seen, ``onfinished()`` is called. |
|
1865 | 1867 | 11. ``__exit__`` is called to signal the end of use of the instance. |
|
1866 | 1868 | |
|
1867 | 1869 | Cache *key* derivation can be influenced by the instance. |
|
1868 | 1870 | |
|
1869 | 1871 | Cache keys are initially derived by a deterministic representation of |
|
1870 | 1872 | the command request. This includes the command name, arguments, protocol |
|
1871 | 1873 | version, etc. This initial key derivation is performed by CBOR-encoding a |
|
1872 | 1874 | data structure and feeding that output into a hasher. |
|
1873 | 1875 | |
|
1874 | 1876 | Instances of this interface can influence this initial key derivation |
|
1875 | 1877 | via ``adjustcachekeystate()``. |
|
1876 | 1878 | |
|
1877 | 1879 | The instance is informed of the derived cache key via a call to |
|
1878 | 1880 | ``setcachekey()``. The instance must store the key locally so it can |
|
1879 | 1881 | be consulted on subsequent operations that may require it. |
|
1880 | 1882 | |
|
1881 | 1883 | When constructed, the instance has access to a callable that can be used |
|
1882 | 1884 | for encoding response objects. This callable receives as its single |
|
1883 | 1885 | argument an object emitted by a command function. It returns an iterable |
|
1884 | 1886 | of bytes chunks representing the encoded object. Unless the cacher is |
|
1885 | 1887 | caching native Python objects in memory or has a way of reconstructing |
|
1886 | 1888 | the original Python objects, implementations typically call this function |
|
1887 | 1889 | to produce bytes from the output objects and then store those bytes in |
|
1888 | 1890 | the cache. When it comes time to re-emit those bytes, they are wrapped |
|
1889 | 1891 | in a ``wireprototypes.encodedresponse`` instance to tell the output |
|
1890 | 1892 | layer that they are pre-encoded. |
|
1891 | 1893 | |
|
1892 | 1894 | When receiving the objects emitted by the command function, instances |
|
1893 | 1895 | can choose what to do with those objects. The simplest thing to do is |
|
1894 | 1896 | re-emit the original objects. They will be forwarded to the output |
|
1895 | 1897 | layer and will be processed as if the cacher did not exist. |
|
1896 | 1898 | |
|
1897 | 1899 | Implementations could also choose to not emit objects - instead locally |
|
1898 | 1900 | buffering objects or their encoded representation. They could then emit |
|
1899 | 1901 | a single "coalesced" object when ``onfinished()`` is called. In |
|
1900 | 1902 | this way, the implementation would function as a filtering layer of |
|
1901 | 1903 | sorts. |
|
1902 | 1904 | |
|
1903 | 1905 | When caching objects, typically the encoded form of the object will |
|
1904 | 1906 | be stored. Keep in mind that if the original object is forwarded to |
|
1905 | 1907 | the output layer, it will need to be encoded there as well. For large |
|
1906 | 1908 | output, this redundant encoding could add overhead. Implementations |
|
1907 | 1909 | could wrap the encoded object data in ``wireprototypes.encodedresponse`` |
|
1908 | 1910 | instances to avoid this overhead. |
|
1909 | 1911 | """ |
|
1910 | 1912 | |
|
1911 | 1913 | def __enter__(): |
|
1912 | 1914 | """Marks the instance as active. |
|
1913 | 1915 | |
|
1914 | 1916 | Should return self. |
|
1915 | 1917 | """ |
|
1916 | 1918 | |
|
1917 | 1919 | def __exit__(exctype, excvalue, exctb): |
|
1918 | 1920 | """Called when cacher is no longer used. |
|
1919 | 1921 | |
|
1920 | 1922 | This can be used by implementations to perform cleanup actions (e.g. |
|
1921 | 1923 | disconnecting network sockets, aborting a partially cached response. |
|
1922 | 1924 | """ |
|
1923 | 1925 | |
|
1924 | 1926 | def adjustcachekeystate(state): |
|
1925 | 1927 | """Influences cache key derivation by adjusting state to derive key. |
|
1926 | 1928 | |
|
1927 | 1929 | A dict defining the state used to derive the cache key is passed. |
|
1928 | 1930 | |
|
1929 | 1931 | Implementations can modify this dict to record additional state that |
|
1930 | 1932 | is wanted to influence key derivation. |
|
1931 | 1933 | |
|
1932 | 1934 | Implementations are *highly* encouraged to not modify or delete |
|
1933 | 1935 | existing keys. |
|
1934 | 1936 | """ |
|
1935 | 1937 | |
|
1936 | 1938 | def setcachekey(key): |
|
1937 | 1939 | """Record the derived cache key for this request. |
|
1938 | 1940 | |
|
1939 | 1941 | Instances may mutate the key for internal usage, as desired. e.g. |
|
1940 | 1942 | instances may wish to prepend the repo name, introduce path |
|
1941 | 1943 | components for filesystem or URL addressing, etc. Behavior is up to |
|
1942 | 1944 | the cache. |
|
1943 | 1945 | |
|
1944 | 1946 | Returns a bool indicating if the request is cacheable by this |
|
1945 | 1947 | instance. |
|
1946 | 1948 | """ |
|
1947 | 1949 | |
|
1948 | 1950 | def lookup(): |
|
1949 | 1951 | """Attempt to resolve an entry in the cache. |
|
1950 | 1952 | |
|
1951 | 1953 | The instance is instructed to look for the cache key that it was |
|
1952 | 1954 | informed about via the call to ``setcachekey()``. |
|
1953 | 1955 | |
|
1954 | 1956 | If there's no cache hit or the cacher doesn't wish to use the cached |
|
1955 | 1957 | entry, ``None`` should be returned. |
|
1956 | 1958 | |
|
1957 | 1959 | Else, a dict defining the cached result should be returned. The |
|
1958 | 1960 | dict may have the following keys: |
|
1959 | 1961 | |
|
1960 | 1962 | objs |
|
1961 | 1963 | An iterable of objects that should be sent to the client. That |
|
1962 | 1964 | iterable of objects is expected to be what the command function |
|
1963 | 1965 | would return if invoked or an equivalent representation thereof. |
|
1964 | 1966 | """ |
|
1965 | 1967 | |
|
1966 | 1968 | def onobject(obj): |
|
1967 | 1969 | """Called when a new object is emitted from the command function. |
|
1968 | 1970 | |
|
1969 | 1971 | Receives as its argument the object that was emitted from the |
|
1970 | 1972 | command function. |
|
1971 | 1973 | |
|
1972 | 1974 | This method returns an iterator of objects to forward to the output |
|
1973 | 1975 | layer. The easiest implementation is a generator that just |
|
1974 | 1976 | ``yield obj``. |
|
1975 | 1977 | """ |
|
1976 | 1978 | |
|
1977 | 1979 | def onfinished(): |
|
1978 | 1980 | """Called after all objects have been emitted from the command function. |
|
1979 | 1981 | |
|
1980 | 1982 | Implementations should return an iterator of objects to forward to |
|
1981 | 1983 | the output layer. |
|
1982 | 1984 | |
|
1983 | 1985 | This method can be a generator. |
|
1984 | 1986 | """ |
@@ -1,2988 +1,2989 b'' | |||
|
1 | 1 | # revlog.py - storage back-end for mercurial |
|
2 | 2 | # |
|
3 | 3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
4 | 4 | # |
|
5 | 5 | # This software may be used and distributed according to the terms of the |
|
6 | 6 | # GNU General Public License version 2 or any later version. |
|
7 | 7 | |
|
8 | 8 | """Storage back-end for Mercurial. |
|
9 | 9 | |
|
10 | 10 | This provides efficient delta storage with O(1) retrieve and append |
|
11 | 11 | and O(changes) merge between branches. |
|
12 | 12 | """ |
|
13 | 13 | |
|
14 | 14 | from __future__ import absolute_import |
|
15 | 15 | |
|
16 | 16 | import collections |
|
17 | 17 | import contextlib |
|
18 | 18 | import errno |
|
19 | 19 | import io |
|
20 | 20 | import os |
|
21 | 21 | import struct |
|
22 | 22 | import zlib |
|
23 | 23 | |
|
24 | 24 | # import stuff from node for others to import from revlog |
|
25 | 25 | from .node import ( |
|
26 | 26 | bin, |
|
27 | 27 | hex, |
|
28 | 28 | nullhex, |
|
29 | 29 | nullid, |
|
30 | 30 | nullrev, |
|
31 | 31 | short, |
|
32 | 32 | wdirfilenodeids, |
|
33 | 33 | wdirhex, |
|
34 | 34 | wdirid, |
|
35 | 35 | wdirrev, |
|
36 | 36 | ) |
|
37 | 37 | from .i18n import _ |
|
38 | 38 | from .pycompat import getattr |
|
39 | 39 | from .revlogutils.constants import ( |
|
40 | 40 | FLAG_GENERALDELTA, |
|
41 | 41 | FLAG_INLINE_DATA, |
|
42 | 42 | REVLOGV0, |
|
43 | 43 | REVLOGV1, |
|
44 | 44 | REVLOGV1_FLAGS, |
|
45 | 45 | REVLOGV2, |
|
46 | 46 | REVLOGV2_FLAGS, |
|
47 | 47 | REVLOG_DEFAULT_FLAGS, |
|
48 | 48 | REVLOG_DEFAULT_FORMAT, |
|
49 | 49 | REVLOG_DEFAULT_VERSION, |
|
50 | 50 | ) |
|
51 | 51 | from .revlogutils.flagutil import ( |
|
52 | 52 | REVIDX_DEFAULT_FLAGS, |
|
53 | 53 | REVIDX_ELLIPSIS, |
|
54 | 54 | REVIDX_EXTSTORED, |
|
55 | 55 | REVIDX_FLAGS_ORDER, |
|
56 | 56 | REVIDX_ISCENSORED, |
|
57 | 57 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
58 | 58 | REVIDX_SIDEDATA, |
|
59 | 59 | ) |
|
60 | 60 | from .thirdparty import attr |
|
61 | 61 | from . import ( |
|
62 | 62 | ancestor, |
|
63 | 63 | dagop, |
|
64 | 64 | error, |
|
65 | 65 | mdiff, |
|
66 | 66 | policy, |
|
67 | 67 | pycompat, |
|
68 | 68 | templatefilters, |
|
69 | 69 | util, |
|
70 | 70 | ) |
|
71 | 71 | from .interfaces import ( |
|
72 | 72 | repository, |
|
73 | 73 | util as interfaceutil, |
|
74 | 74 | ) |
|
75 | 75 | from .revlogutils import ( |
|
76 | 76 | deltas as deltautil, |
|
77 | 77 | flagutil, |
|
78 | 78 | nodemap as nodemaputil, |
|
79 | 79 | sidedata as sidedatautil, |
|
80 | 80 | ) |
|
81 | 81 | from .utils import ( |
|
82 | 82 | storageutil, |
|
83 | 83 | stringutil, |
|
84 | 84 | ) |
|
85 | 85 | |
|
86 | 86 | # blanked usage of all the name to prevent pyflakes constraints |
|
87 | 87 | # We need these name available in the module for extensions. |
|
88 | 88 | REVLOGV0 |
|
89 | 89 | REVLOGV1 |
|
90 | 90 | REVLOGV2 |
|
91 | 91 | FLAG_INLINE_DATA |
|
92 | 92 | FLAG_GENERALDELTA |
|
93 | 93 | REVLOG_DEFAULT_FLAGS |
|
94 | 94 | REVLOG_DEFAULT_FORMAT |
|
95 | 95 | REVLOG_DEFAULT_VERSION |
|
96 | 96 | REVLOGV1_FLAGS |
|
97 | 97 | REVLOGV2_FLAGS |
|
98 | 98 | REVIDX_ISCENSORED |
|
99 | 99 | REVIDX_ELLIPSIS |
|
100 | 100 | REVIDX_SIDEDATA |
|
101 | 101 | REVIDX_EXTSTORED |
|
102 | 102 | REVIDX_DEFAULT_FLAGS |
|
103 | 103 | REVIDX_FLAGS_ORDER |
|
104 | 104 | REVIDX_RAWTEXT_CHANGING_FLAGS |
|
105 | 105 | |
|
106 | 106 | parsers = policy.importmod('parsers') |
|
107 | 107 | rustancestor = policy.importrust('ancestor') |
|
108 | 108 | rustdagop = policy.importrust('dagop') |
|
109 | 109 | rustrevlog = policy.importrust('revlog') |
|
110 | 110 | |
|
111 | 111 | # Aliased for performance. |
|
112 | 112 | _zlibdecompress = zlib.decompress |
|
113 | 113 | |
|
114 | 114 | # max size of revlog with inline data |
|
115 | 115 | _maxinline = 131072 |
|
116 | 116 | _chunksize = 1048576 |
|
117 | 117 | |
|
118 | 118 | # Flag processors for REVIDX_ELLIPSIS. |
|
119 | 119 | def ellipsisreadprocessor(rl, text): |
|
120 | 120 | return text, False, {} |
|
121 | 121 | |
|
122 | 122 | |
|
123 | 123 | def ellipsiswriteprocessor(rl, text, sidedata): |
|
124 | 124 | return text, False |
|
125 | 125 | |
|
126 | 126 | |
|
127 | 127 | def ellipsisrawprocessor(rl, text): |
|
128 | 128 | return False |
|
129 | 129 | |
|
130 | 130 | |
|
131 | 131 | ellipsisprocessor = ( |
|
132 | 132 | ellipsisreadprocessor, |
|
133 | 133 | ellipsiswriteprocessor, |
|
134 | 134 | ellipsisrawprocessor, |
|
135 | 135 | ) |
|
136 | 136 | |
|
137 | 137 | |
|
138 | 138 | def getoffset(q): |
|
139 | 139 | return int(q >> 16) |
|
140 | 140 | |
|
141 | 141 | |
|
142 | 142 | def gettype(q): |
|
143 | 143 | return int(q & 0xFFFF) |
|
144 | 144 | |
|
145 | 145 | |
|
146 | 146 | def offset_type(offset, type): |
|
147 | 147 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: |
|
148 | 148 | raise ValueError(b'unknown revlog index flags') |
|
149 | 149 | return int(int(offset) << 16 | type) |
|
150 | 150 | |
|
151 | 151 | |
|
152 | 152 | def _verify_revision(rl, skipflags, state, node): |
|
153 | 153 | """Verify the integrity of the given revlog ``node`` while providing a hook |
|
154 | 154 | point for extensions to influence the operation.""" |
|
155 | 155 | if skipflags: |
|
156 | 156 | state[b'skipread'].add(node) |
|
157 | 157 | else: |
|
158 | 158 | # Side-effect: read content and verify hash. |
|
159 | 159 | rl.revision(node) |
|
160 | 160 | |
|
161 | 161 | |
|
162 | 162 | @attr.s(slots=True, frozen=True) |
|
163 | 163 | class _revisioninfo(object): |
|
164 | 164 | """Information about a revision that allows building its fulltext |
|
165 | 165 | node: expected hash of the revision |
|
166 | 166 | p1, p2: parent revs of the revision |
|
167 | 167 | btext: built text cache consisting of a one-element list |
|
168 | 168 | cachedelta: (baserev, uncompressed_delta) or None |
|
169 | 169 | flags: flags associated to the revision storage |
|
170 | 170 | |
|
171 | 171 | One of btext[0] or cachedelta must be set. |
|
172 | 172 | """ |
|
173 | 173 | |
|
174 | 174 | node = attr.ib() |
|
175 | 175 | p1 = attr.ib() |
|
176 | 176 | p2 = attr.ib() |
|
177 | 177 | btext = attr.ib() |
|
178 | 178 | textlen = attr.ib() |
|
179 | 179 | cachedelta = attr.ib() |
|
180 | 180 | flags = attr.ib() |
|
181 | 181 | |
|
182 | 182 | |
|
183 | 183 | @interfaceutil.implementer(repository.irevisiondelta) |
|
184 | 184 | @attr.s(slots=True) |
|
185 | 185 | class revlogrevisiondelta(object): |
|
186 | 186 | node = attr.ib() |
|
187 | 187 | p1node = attr.ib() |
|
188 | 188 | p2node = attr.ib() |
|
189 | 189 | basenode = attr.ib() |
|
190 | 190 | flags = attr.ib() |
|
191 | 191 | baserevisionsize = attr.ib() |
|
192 | 192 | revision = attr.ib() |
|
193 | 193 | delta = attr.ib() |
|
194 | 194 | linknode = attr.ib(default=None) |
|
195 | 195 | |
|
196 | 196 | |
|
197 | 197 | @interfaceutil.implementer(repository.iverifyproblem) |
|
198 | 198 | @attr.s(frozen=True) |
|
199 | 199 | class revlogproblem(object): |
|
200 | 200 | warning = attr.ib(default=None) |
|
201 | 201 | error = attr.ib(default=None) |
|
202 | 202 | node = attr.ib(default=None) |
|
203 | 203 | |
|
204 | 204 | |
|
205 | 205 | # index v0: |
|
206 | 206 | # 4 bytes: offset |
|
207 | 207 | # 4 bytes: compressed length |
|
208 | 208 | # 4 bytes: base rev |
|
209 | 209 | # 4 bytes: link rev |
|
210 | 210 | # 20 bytes: parent 1 nodeid |
|
211 | 211 | # 20 bytes: parent 2 nodeid |
|
212 | 212 | # 20 bytes: nodeid |
|
213 | 213 | indexformatv0 = struct.Struct(b">4l20s20s20s") |
|
214 | 214 | indexformatv0_pack = indexformatv0.pack |
|
215 | 215 | indexformatv0_unpack = indexformatv0.unpack |
|
216 | 216 | |
|
217 | 217 | |
|
218 | 218 | class revlogoldindex(list): |
|
219 | 219 | @property |
|
220 | 220 | def nodemap(self): |
|
221 | 221 | msg = b"index.nodemap is deprecated, use index.[has_node|rev|get_rev]" |
|
222 | 222 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
223 | 223 | return self._nodemap |
|
224 | 224 | |
|
225 | 225 | @util.propertycache |
|
226 | 226 | def _nodemap(self): |
|
227 | 227 | nodemap = nodemaputil.NodeMap({nullid: nullrev}) |
|
228 | 228 | for r in range(0, len(self)): |
|
229 | 229 | n = self[r][7] |
|
230 | 230 | nodemap[n] = r |
|
231 | 231 | return nodemap |
|
232 | 232 | |
|
233 | 233 | def has_node(self, node): |
|
234 | 234 | """return True if the node exist in the index""" |
|
235 | 235 | return node in self._nodemap |
|
236 | 236 | |
|
237 | 237 | def rev(self, node): |
|
238 | 238 | """return a revision for a node |
|
239 | 239 | |
|
240 | 240 | If the node is unknown, raise a RevlogError""" |
|
241 | 241 | return self._nodemap[node] |
|
242 | 242 | |
|
243 | 243 | def get_rev(self, node): |
|
244 | 244 | """return a revision for a node |
|
245 | 245 | |
|
246 | 246 | If the node is unknown, return None""" |
|
247 | 247 | return self._nodemap.get(node) |
|
248 | 248 | |
|
249 | 249 | def append(self, tup): |
|
250 | 250 | self._nodemap[tup[7]] = len(self) |
|
251 | 251 | super(revlogoldindex, self).append(tup) |
|
252 | 252 | |
|
253 | 253 | def __delitem__(self, i): |
|
254 | 254 | if not isinstance(i, slice) or not i.stop == -1 or i.step is not None: |
|
255 | 255 | raise ValueError(b"deleting slices only supports a:-1 with step 1") |
|
256 | 256 | for r in pycompat.xrange(i.start, len(self)): |
|
257 | 257 | del self._nodemap[self[r][7]] |
|
258 | 258 | super(revlogoldindex, self).__delitem__(i) |
|
259 | 259 | |
|
260 | 260 | def clearcaches(self): |
|
261 | 261 | self.__dict__.pop('_nodemap', None) |
|
262 | 262 | |
|
263 | 263 | def __getitem__(self, i): |
|
264 | 264 | if i == -1: |
|
265 | 265 | return (0, 0, 0, -1, -1, -1, -1, nullid) |
|
266 | 266 | return list.__getitem__(self, i) |
|
267 | 267 | |
|
268 | 268 | |
|
269 | 269 | class revlogoldio(object): |
|
270 | 270 | def __init__(self): |
|
271 | 271 | self.size = indexformatv0.size |
|
272 | 272 | |
|
273 | 273 | def parseindex(self, data, inline): |
|
274 | 274 | s = self.size |
|
275 | 275 | index = [] |
|
276 | 276 | nodemap = nodemaputil.NodeMap({nullid: nullrev}) |
|
277 | 277 | n = off = 0 |
|
278 | 278 | l = len(data) |
|
279 | 279 | while off + s <= l: |
|
280 | 280 | cur = data[off : off + s] |
|
281 | 281 | off += s |
|
282 | 282 | e = indexformatv0_unpack(cur) |
|
283 | 283 | # transform to revlogv1 format |
|
284 | 284 | e2 = ( |
|
285 | 285 | offset_type(e[0], 0), |
|
286 | 286 | e[1], |
|
287 | 287 | -1, |
|
288 | 288 | e[2], |
|
289 | 289 | e[3], |
|
290 | 290 | nodemap.get(e[4], nullrev), |
|
291 | 291 | nodemap.get(e[5], nullrev), |
|
292 | 292 | e[6], |
|
293 | 293 | ) |
|
294 | 294 | index.append(e2) |
|
295 | 295 | nodemap[e[6]] = n |
|
296 | 296 | n += 1 |
|
297 | 297 | |
|
298 | 298 | index = revlogoldindex(index) |
|
299 | 299 | return index, None |
|
300 | 300 | |
|
301 | 301 | def packentry(self, entry, node, version, rev): |
|
302 | 302 | if gettype(entry[0]): |
|
303 | 303 | raise error.RevlogError( |
|
304 | 304 | _(b'index entry flags need revlog version 1') |
|
305 | 305 | ) |
|
306 | 306 | e2 = ( |
|
307 | 307 | getoffset(entry[0]), |
|
308 | 308 | entry[1], |
|
309 | 309 | entry[3], |
|
310 | 310 | entry[4], |
|
311 | 311 | node(entry[5]), |
|
312 | 312 | node(entry[6]), |
|
313 | 313 | entry[7], |
|
314 | 314 | ) |
|
315 | 315 | return indexformatv0_pack(*e2) |
|
316 | 316 | |
|
317 | 317 | |
|
318 | 318 | # index ng: |
|
319 | 319 | # 6 bytes: offset |
|
320 | 320 | # 2 bytes: flags |
|
321 | 321 | # 4 bytes: compressed length |
|
322 | 322 | # 4 bytes: uncompressed length |
|
323 | 323 | # 4 bytes: base rev |
|
324 | 324 | # 4 bytes: link rev |
|
325 | 325 | # 4 bytes: parent 1 rev |
|
326 | 326 | # 4 bytes: parent 2 rev |
|
327 | 327 | # 32 bytes: nodeid |
|
328 | 328 | indexformatng = struct.Struct(b">Qiiiiii20s12x") |
|
329 | 329 | indexformatng_pack = indexformatng.pack |
|
330 | 330 | versionformat = struct.Struct(b">I") |
|
331 | 331 | versionformat_pack = versionformat.pack |
|
332 | 332 | versionformat_unpack = versionformat.unpack |
|
333 | 333 | |
|
334 | 334 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte |
|
335 | 335 | # signed integer) |
|
336 | 336 | _maxentrysize = 0x7FFFFFFF |
|
337 | 337 | |
|
338 | 338 | |
|
339 | 339 | class revlogio(object): |
|
340 | 340 | def __init__(self): |
|
341 | 341 | self.size = indexformatng.size |
|
342 | 342 | |
|
343 | 343 | def parseindex(self, data, inline): |
|
344 | 344 | # call the C implementation to parse the index data |
|
345 | 345 | index, cache = parsers.parse_index2(data, inline) |
|
346 | 346 | return index, cache |
|
347 | 347 | |
|
348 | 348 | def packentry(self, entry, node, version, rev): |
|
349 | 349 | p = indexformatng_pack(*entry) |
|
350 | 350 | if rev == 0: |
|
351 | 351 | p = versionformat_pack(version) + p[4:] |
|
352 | 352 | return p |
|
353 | 353 | |
|
354 | 354 | |
|
355 | 355 | class rustrevlogio(revlogio): |
|
356 | 356 | def parseindex(self, data, inline): |
|
357 | 357 | index, cache = super(rustrevlogio, self).parseindex(data, inline) |
|
358 | 358 | return rustrevlog.MixedIndex(index), cache |
|
359 | 359 | |
|
360 | 360 | |
|
361 | 361 | class revlog(object): |
|
362 | 362 | """ |
|
363 | 363 | the underlying revision storage object |
|
364 | 364 | |
|
365 | 365 | A revlog consists of two parts, an index and the revision data. |
|
366 | 366 | |
|
367 | 367 | The index is a file with a fixed record size containing |
|
368 | 368 | information on each revision, including its nodeid (hash), the |
|
369 | 369 | nodeids of its parents, the position and offset of its data within |
|
370 | 370 | the data file, and the revision it's based on. Finally, each entry |
|
371 | 371 | contains a linkrev entry that can serve as a pointer to external |
|
372 | 372 | data. |
|
373 | 373 | |
|
374 | 374 | The revision data itself is a linear collection of data chunks. |
|
375 | 375 | Each chunk represents a revision and is usually represented as a |
|
376 | 376 | delta against the previous chunk. To bound lookup time, runs of |
|
377 | 377 | deltas are limited to about 2 times the length of the original |
|
378 | 378 | version data. This makes retrieval of a version proportional to |
|
379 | 379 | its size, or O(1) relative to the number of revisions. |
|
380 | 380 | |
|
381 | 381 | Both pieces of the revlog are written to in an append-only |
|
382 | 382 | fashion, which means we never need to rewrite a file to insert or |
|
383 | 383 | remove data, and can use some simple techniques to avoid the need |
|
384 | 384 | for locking while reading. |
|
385 | 385 | |
|
386 | 386 | If checkambig, indexfile is opened with checkambig=True at |
|
387 | 387 | writing, to avoid file stat ambiguity. |
|
388 | 388 | |
|
389 | 389 | If mmaplargeindex is True, and an mmapindexthreshold is set, the |
|
390 | 390 | index will be mmapped rather than read if it is larger than the |
|
391 | 391 | configured threshold. |
|
392 | 392 | |
|
393 | 393 | If censorable is True, the revlog can have censored revisions. |
|
394 | 394 | |
|
395 | 395 | If `upperboundcomp` is not None, this is the expected maximal gain from |
|
396 | 396 | compression for the data content. |
|
397 | 397 | """ |
|
398 | 398 | |
|
399 | 399 | _flagserrorclass = error.RevlogError |
|
400 | 400 | |
|
401 | 401 | def __init__( |
|
402 | 402 | self, |
|
403 | 403 | opener, |
|
404 | 404 | indexfile, |
|
405 | 405 | datafile=None, |
|
406 | 406 | checkambig=False, |
|
407 | 407 | mmaplargeindex=False, |
|
408 | 408 | censorable=False, |
|
409 | 409 | upperboundcomp=None, |
|
410 | 410 | ): |
|
411 | 411 | """ |
|
412 | 412 | create a revlog object |
|
413 | 413 | |
|
414 | 414 | opener is a function that abstracts the file opening operation |
|
415 | 415 | and can be used to implement COW semantics or the like. |
|
416 | 416 | |
|
417 | 417 | """ |
|
418 | 418 | self.upperboundcomp = upperboundcomp |
|
419 | 419 | self.indexfile = indexfile |
|
420 | 420 | self.datafile = datafile or (indexfile[:-2] + b".d") |
|
421 | 421 | self.opener = opener |
|
422 | 422 | # When True, indexfile is opened with checkambig=True at writing, to |
|
423 | 423 | # avoid file stat ambiguity. |
|
424 | 424 | self._checkambig = checkambig |
|
425 | 425 | self._mmaplargeindex = mmaplargeindex |
|
426 | 426 | self._censorable = censorable |
|
427 | 427 | # 3-tuple of (node, rev, text) for a raw revision. |
|
428 | 428 | self._revisioncache = None |
|
429 | 429 | # Maps rev to chain base rev. |
|
430 | 430 | self._chainbasecache = util.lrucachedict(100) |
|
431 | 431 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. |
|
432 | 432 | self._chunkcache = (0, b'') |
|
433 | 433 | # How much data to read and cache into the raw revlog data cache. |
|
434 | 434 | self._chunkcachesize = 65536 |
|
435 | 435 | self._maxchainlen = None |
|
436 | 436 | self._deltabothparents = True |
|
437 | 437 | self.index = None |
|
438 | 438 | # Mapping of partial identifiers to full nodes. |
|
439 | 439 | self._pcache = {} |
|
440 | 440 | # Mapping of revision integer to full node. |
|
441 | 441 | self._compengine = b'zlib' |
|
442 | 442 | self._compengineopts = {} |
|
443 | 443 | self._maxdeltachainspan = -1 |
|
444 | 444 | self._withsparseread = False |
|
445 | 445 | self._sparserevlog = False |
|
446 | 446 | self._srdensitythreshold = 0.50 |
|
447 | 447 | self._srmingapsize = 262144 |
|
448 | 448 | |
|
449 | 449 | # Make copy of flag processors so each revlog instance can support |
|
450 | 450 | # custom flags. |
|
451 | 451 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
452 | 452 | |
|
453 | 453 | # 2-tuple of file handles being used for active writing. |
|
454 | 454 | self._writinghandles = None |
|
455 | 455 | |
|
456 | 456 | self._loadindex() |
|
457 | 457 | |
|
458 | 458 | def _loadindex(self): |
|
459 | 459 | mmapindexthreshold = None |
|
460 | 460 | opts = self.opener.options |
|
461 | 461 | |
|
462 | 462 | if b'revlogv2' in opts: |
|
463 | 463 | newversionflags = REVLOGV2 | FLAG_INLINE_DATA |
|
464 | 464 | elif b'revlogv1' in opts: |
|
465 | 465 | newversionflags = REVLOGV1 | FLAG_INLINE_DATA |
|
466 | 466 | if b'generaldelta' in opts: |
|
467 | 467 | newversionflags |= FLAG_GENERALDELTA |
|
468 | 468 | elif b'revlogv0' in self.opener.options: |
|
469 | 469 | newversionflags = REVLOGV0 |
|
470 | 470 | else: |
|
471 | 471 | newversionflags = REVLOG_DEFAULT_VERSION |
|
472 | 472 | |
|
473 | 473 | if b'chunkcachesize' in opts: |
|
474 | 474 | self._chunkcachesize = opts[b'chunkcachesize'] |
|
475 | 475 | if b'maxchainlen' in opts: |
|
476 | 476 | self._maxchainlen = opts[b'maxchainlen'] |
|
477 | 477 | if b'deltabothparents' in opts: |
|
478 | 478 | self._deltabothparents = opts[b'deltabothparents'] |
|
479 | 479 | self._lazydelta = bool(opts.get(b'lazydelta', True)) |
|
480 | 480 | self._lazydeltabase = False |
|
481 | 481 | if self._lazydelta: |
|
482 | 482 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) |
|
483 | 483 | if b'compengine' in opts: |
|
484 | 484 | self._compengine = opts[b'compengine'] |
|
485 | 485 | if b'zlib.level' in opts: |
|
486 | 486 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] |
|
487 | 487 | if b'zstd.level' in opts: |
|
488 | 488 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] |
|
489 | 489 | if b'maxdeltachainspan' in opts: |
|
490 | 490 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] |
|
491 | 491 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: |
|
492 | 492 | mmapindexthreshold = opts[b'mmapindexthreshold'] |
|
493 | 493 | self.hassidedata = bool(opts.get(b'side-data', False)) |
|
494 | 494 | if self.hassidedata: |
|
495 | 495 | self._flagprocessors[REVIDX_SIDEDATA] = sidedatautil.processors |
|
496 | 496 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) |
|
497 | 497 | withsparseread = bool(opts.get(b'with-sparse-read', False)) |
|
498 | 498 | # sparse-revlog forces sparse-read |
|
499 | 499 | self._withsparseread = self._sparserevlog or withsparseread |
|
500 | 500 | if b'sparse-read-density-threshold' in opts: |
|
501 | 501 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] |
|
502 | 502 | if b'sparse-read-min-gap-size' in opts: |
|
503 | 503 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] |
|
504 | 504 | if opts.get(b'enableellipsis'): |
|
505 | 505 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor |
|
506 | 506 | |
|
507 | 507 | # revlog v0 doesn't have flag processors |
|
508 | 508 | for flag, processor in pycompat.iteritems( |
|
509 | 509 | opts.get(b'flagprocessors', {}) |
|
510 | 510 | ): |
|
511 | 511 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) |
|
512 | 512 | |
|
513 | 513 | if self._chunkcachesize <= 0: |
|
514 | 514 | raise error.RevlogError( |
|
515 | 515 | _(b'revlog chunk cache size %r is not greater than 0') |
|
516 | 516 | % self._chunkcachesize |
|
517 | 517 | ) |
|
518 | 518 | elif self._chunkcachesize & (self._chunkcachesize - 1): |
|
519 | 519 | raise error.RevlogError( |
|
520 | 520 | _(b'revlog chunk cache size %r is not a power of 2') |
|
521 | 521 | % self._chunkcachesize |
|
522 | 522 | ) |
|
523 | 523 | |
|
524 | 524 | indexdata = b'' |
|
525 | 525 | self._initempty = True |
|
526 | 526 | try: |
|
527 | 527 | with self._indexfp() as f: |
|
528 | 528 | if ( |
|
529 | 529 | mmapindexthreshold is not None |
|
530 | 530 | and self.opener.fstat(f).st_size >= mmapindexthreshold |
|
531 | 531 | ): |
|
532 | 532 | # TODO: should .close() to release resources without |
|
533 | 533 | # relying on Python GC |
|
534 | 534 | indexdata = util.buffer(util.mmapread(f)) |
|
535 | 535 | else: |
|
536 | 536 | indexdata = f.read() |
|
537 | 537 | if len(indexdata) > 0: |
|
538 | 538 | versionflags = versionformat_unpack(indexdata[:4])[0] |
|
539 | 539 | self._initempty = False |
|
540 | 540 | else: |
|
541 | 541 | versionflags = newversionflags |
|
542 | 542 | except IOError as inst: |
|
543 | 543 | if inst.errno != errno.ENOENT: |
|
544 | 544 | raise |
|
545 | 545 | |
|
546 | 546 | versionflags = newversionflags |
|
547 | 547 | |
|
548 | 548 | self.version = versionflags |
|
549 | 549 | |
|
550 | 550 | flags = versionflags & ~0xFFFF |
|
551 | 551 | fmt = versionflags & 0xFFFF |
|
552 | 552 | |
|
553 | 553 | if fmt == REVLOGV0: |
|
554 | 554 | if flags: |
|
555 | 555 | raise error.RevlogError( |
|
556 | 556 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
557 | 557 | % (flags >> 16, fmt, self.indexfile) |
|
558 | 558 | ) |
|
559 | 559 | |
|
560 | 560 | self._inline = False |
|
561 | 561 | self._generaldelta = False |
|
562 | 562 | |
|
563 | 563 | elif fmt == REVLOGV1: |
|
564 | 564 | if flags & ~REVLOGV1_FLAGS: |
|
565 | 565 | raise error.RevlogError( |
|
566 | 566 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
567 | 567 | % (flags >> 16, fmt, self.indexfile) |
|
568 | 568 | ) |
|
569 | 569 | |
|
570 | 570 | self._inline = versionflags & FLAG_INLINE_DATA |
|
571 | 571 | self._generaldelta = versionflags & FLAG_GENERALDELTA |
|
572 | 572 | |
|
573 | 573 | elif fmt == REVLOGV2: |
|
574 | 574 | if flags & ~REVLOGV2_FLAGS: |
|
575 | 575 | raise error.RevlogError( |
|
576 | 576 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
577 | 577 | % (flags >> 16, fmt, self.indexfile) |
|
578 | 578 | ) |
|
579 | 579 | |
|
580 | 580 | self._inline = versionflags & FLAG_INLINE_DATA |
|
581 | 581 | # generaldelta implied by version 2 revlogs. |
|
582 | 582 | self._generaldelta = True |
|
583 | 583 | |
|
584 | 584 | else: |
|
585 | 585 | raise error.RevlogError( |
|
586 | 586 | _(b'unknown version (%d) in revlog %s') % (fmt, self.indexfile) |
|
587 | 587 | ) |
|
588 | 588 | # sparse-revlog can't be on without general-delta (issue6056) |
|
589 | 589 | if not self._generaldelta: |
|
590 | 590 | self._sparserevlog = False |
|
591 | 591 | |
|
592 | 592 | self._storedeltachains = True |
|
593 | 593 | |
|
594 | 594 | self._io = revlogio() |
|
595 | 595 | if self.version == REVLOGV0: |
|
596 | 596 | self._io = revlogoldio() |
|
597 | 597 | elif rustrevlog is not None and self.opener.options.get(b'rust.index'): |
|
598 | 598 | self._io = rustrevlogio() |
|
599 | 599 | try: |
|
600 | 600 | d = self._io.parseindex(indexdata, self._inline) |
|
601 | 601 | except (ValueError, IndexError): |
|
602 | 602 | raise error.RevlogError( |
|
603 | 603 | _(b"index %s is corrupted") % self.indexfile |
|
604 | 604 | ) |
|
605 | 605 | self.index, self._chunkcache = d |
|
606 | 606 | if not self._chunkcache: |
|
607 | 607 | self._chunkclear() |
|
608 | 608 | # revnum -> (chain-length, sum-delta-length) |
|
609 | 609 | self._chaininfocache = {} |
|
610 | 610 | # revlog header -> revlog compressor |
|
611 | 611 | self._decompressors = {} |
|
612 | 612 | |
|
613 | 613 | @util.propertycache |
|
614 | 614 | def _compressor(self): |
|
615 | 615 | engine = util.compengines[self._compengine] |
|
616 | 616 | return engine.revlogcompressor(self._compengineopts) |
|
617 | 617 | |
|
618 | 618 | def _indexfp(self, mode=b'r'): |
|
619 | 619 | """file object for the revlog's index file""" |
|
620 | 620 | args = {'mode': mode} |
|
621 | 621 | if mode != b'r': |
|
622 | 622 | args['checkambig'] = self._checkambig |
|
623 | 623 | if mode == b'w': |
|
624 | 624 | args['atomictemp'] = True |
|
625 | 625 | return self.opener(self.indexfile, **args) |
|
626 | 626 | |
|
627 | 627 | def _datafp(self, mode=b'r'): |
|
628 | 628 | """file object for the revlog's data file""" |
|
629 | 629 | return self.opener(self.datafile, mode=mode) |
|
630 | 630 | |
|
631 | 631 | @contextlib.contextmanager |
|
632 | 632 | def _datareadfp(self, existingfp=None): |
|
633 | 633 | """file object suitable to read data""" |
|
634 | 634 | # Use explicit file handle, if given. |
|
635 | 635 | if existingfp is not None: |
|
636 | 636 | yield existingfp |
|
637 | 637 | |
|
638 | 638 | # Use a file handle being actively used for writes, if available. |
|
639 | 639 | # There is some danger to doing this because reads will seek the |
|
640 | 640 | # file. However, _writeentry() performs a SEEK_END before all writes, |
|
641 | 641 | # so we should be safe. |
|
642 | 642 | elif self._writinghandles: |
|
643 | 643 | if self._inline: |
|
644 | 644 | yield self._writinghandles[0] |
|
645 | 645 | else: |
|
646 | 646 | yield self._writinghandles[1] |
|
647 | 647 | |
|
648 | 648 | # Otherwise open a new file handle. |
|
649 | 649 | else: |
|
650 | 650 | if self._inline: |
|
651 | 651 | func = self._indexfp |
|
652 | 652 | else: |
|
653 | 653 | func = self._datafp |
|
654 | 654 | with func() as fp: |
|
655 | 655 | yield fp |
|
656 | 656 | |
|
657 | 657 | def tiprev(self): |
|
658 | 658 | return len(self.index) - 1 |
|
659 | 659 | |
|
660 | 660 | def tip(self): |
|
661 | 661 | return self.node(self.tiprev()) |
|
662 | 662 | |
|
663 | 663 | def __contains__(self, rev): |
|
664 | 664 | return 0 <= rev < len(self) |
|
665 | 665 | |
|
666 | 666 | def __len__(self): |
|
667 | 667 | return len(self.index) |
|
668 | 668 | |
|
669 | 669 | def __iter__(self): |
|
670 | 670 | return iter(pycompat.xrange(len(self))) |
|
671 | 671 | |
|
672 | 672 | def revs(self, start=0, stop=None): |
|
673 | 673 | """iterate over all rev in this revlog (from start to stop)""" |
|
674 | 674 | return storageutil.iterrevs(len(self), start=start, stop=stop) |
|
675 | 675 | |
|
676 | 676 | @property |
|
677 | 677 | def nodemap(self): |
|
678 | 678 | msg = ( |
|
679 | 679 | b"revlog.nodemap is deprecated, " |
|
680 | 680 | b"use revlog.index.[has_node|rev|get_rev]" |
|
681 | 681 | ) |
|
682 | 682 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
683 | 683 | return self.index.nodemap |
|
684 | 684 | |
|
685 | 685 | @property |
|
686 | 686 | def _nodecache(self): |
|
687 | 687 | msg = b"revlog._nodecache is deprecated, use revlog.index.nodemap" |
|
688 | 688 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
689 | 689 | return self.index.nodemap |
|
690 | 690 | |
|
691 | 691 | def hasnode(self, node): |
|
692 | 692 | try: |
|
693 | 693 | self.rev(node) |
|
694 | 694 | return True |
|
695 | 695 | except KeyError: |
|
696 | 696 | return False |
|
697 | 697 | |
|
698 | 698 | def candelta(self, baserev, rev): |
|
699 | 699 | """whether two revisions (baserev, rev) can be delta-ed or not""" |
|
700 | 700 | # Disable delta if either rev requires a content-changing flag |
|
701 | 701 | # processor (ex. LFS). This is because such flag processor can alter |
|
702 | 702 | # the rawtext content that the delta will be based on, and two clients |
|
703 | 703 | # could have a same revlog node with different flags (i.e. different |
|
704 | 704 | # rawtext contents) and the delta could be incompatible. |
|
705 | 705 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
706 | 706 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS |
|
707 | 707 | ): |
|
708 | 708 | return False |
|
709 | 709 | return True |
|
710 | 710 | |
|
711 | 711 | def clearcaches(self): |
|
712 | 712 | self._revisioncache = None |
|
713 | 713 | self._chainbasecache.clear() |
|
714 | 714 | self._chunkcache = (0, b'') |
|
715 | 715 | self._pcache = {} |
|
716 | 716 | self.index.clearcaches() |
|
717 | 717 | |
|
718 | 718 | def rev(self, node): |
|
719 | 719 | try: |
|
720 | 720 | return self.index.rev(node) |
|
721 | 721 | except TypeError: |
|
722 | 722 | raise |
|
723 | 723 | except error.RevlogError: |
|
724 | 724 | # parsers.c radix tree lookup failed |
|
725 | 725 | if node == wdirid or node in wdirfilenodeids: |
|
726 | 726 | raise error.WdirUnsupported |
|
727 | 727 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
728 | 728 | |
|
729 | 729 | # Accessors for index entries. |
|
730 | 730 | |
|
731 | 731 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes |
|
732 | 732 | # are flags. |
|
733 | 733 | def start(self, rev): |
|
734 | 734 | return int(self.index[rev][0] >> 16) |
|
735 | 735 | |
|
736 | 736 | def flags(self, rev): |
|
737 | 737 | return self.index[rev][0] & 0xFFFF |
|
738 | 738 | |
|
739 | 739 | def length(self, rev): |
|
740 | 740 | return self.index[rev][1] |
|
741 | 741 | |
|
742 | 742 | def rawsize(self, rev): |
|
743 | 743 | """return the length of the uncompressed text for a given revision""" |
|
744 | 744 | l = self.index[rev][2] |
|
745 | 745 | if l >= 0: |
|
746 | 746 | return l |
|
747 | 747 | |
|
748 | 748 | t = self.rawdata(rev) |
|
749 | 749 | return len(t) |
|
750 | 750 | |
|
751 | 751 | def size(self, rev): |
|
752 | 752 | """length of non-raw text (processed by a "read" flag processor)""" |
|
753 | 753 | # fast path: if no "read" flag processor could change the content, |
|
754 | 754 | # size is rawsize. note: ELLIPSIS is known to not change the content. |
|
755 | 755 | flags = self.flags(rev) |
|
756 | 756 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: |
|
757 | 757 | return self.rawsize(rev) |
|
758 | 758 | |
|
759 | 759 | return len(self.revision(rev, raw=False)) |
|
760 | 760 | |
|
761 | 761 | def chainbase(self, rev): |
|
762 | 762 | base = self._chainbasecache.get(rev) |
|
763 | 763 | if base is not None: |
|
764 | 764 | return base |
|
765 | 765 | |
|
766 | 766 | index = self.index |
|
767 | 767 | iterrev = rev |
|
768 | 768 | base = index[iterrev][3] |
|
769 | 769 | while base != iterrev: |
|
770 | 770 | iterrev = base |
|
771 | 771 | base = index[iterrev][3] |
|
772 | 772 | |
|
773 | 773 | self._chainbasecache[rev] = base |
|
774 | 774 | return base |
|
775 | 775 | |
|
776 | 776 | def linkrev(self, rev): |
|
777 | 777 | return self.index[rev][4] |
|
778 | 778 | |
|
779 | 779 | def parentrevs(self, rev): |
|
780 | 780 | try: |
|
781 | 781 | entry = self.index[rev] |
|
782 | 782 | except IndexError: |
|
783 | 783 | if rev == wdirrev: |
|
784 | 784 | raise error.WdirUnsupported |
|
785 | 785 | raise |
|
786 | 786 | |
|
787 | 787 | return entry[5], entry[6] |
|
788 | 788 | |
|
789 | 789 | # fast parentrevs(rev) where rev isn't filtered |
|
790 | 790 | _uncheckedparentrevs = parentrevs |
|
791 | 791 | |
|
792 | 792 | def node(self, rev): |
|
793 | 793 | try: |
|
794 | 794 | return self.index[rev][7] |
|
795 | 795 | except IndexError: |
|
796 | 796 | if rev == wdirrev: |
|
797 | 797 | raise error.WdirUnsupported |
|
798 | 798 | raise |
|
799 | 799 | |
|
800 | 800 | # Derived from index values. |
|
801 | 801 | |
|
802 | 802 | def end(self, rev): |
|
803 | 803 | return self.start(rev) + self.length(rev) |
|
804 | 804 | |
|
805 | 805 | def parents(self, node): |
|
806 | 806 | i = self.index |
|
807 | 807 | d = i[self.rev(node)] |
|
808 | 808 | return i[d[5]][7], i[d[6]][7] # map revisions to nodes inline |
|
809 | 809 | |
|
810 | 810 | def chainlen(self, rev): |
|
811 | 811 | return self._chaininfo(rev)[0] |
|
812 | 812 | |
|
813 | 813 | def _chaininfo(self, rev): |
|
814 | 814 | chaininfocache = self._chaininfocache |
|
815 | 815 | if rev in chaininfocache: |
|
816 | 816 | return chaininfocache[rev] |
|
817 | 817 | index = self.index |
|
818 | 818 | generaldelta = self._generaldelta |
|
819 | 819 | iterrev = rev |
|
820 | 820 | e = index[iterrev] |
|
821 | 821 | clen = 0 |
|
822 | 822 | compresseddeltalen = 0 |
|
823 | 823 | while iterrev != e[3]: |
|
824 | 824 | clen += 1 |
|
825 | 825 | compresseddeltalen += e[1] |
|
826 | 826 | if generaldelta: |
|
827 | 827 | iterrev = e[3] |
|
828 | 828 | else: |
|
829 | 829 | iterrev -= 1 |
|
830 | 830 | if iterrev in chaininfocache: |
|
831 | 831 | t = chaininfocache[iterrev] |
|
832 | 832 | clen += t[0] |
|
833 | 833 | compresseddeltalen += t[1] |
|
834 | 834 | break |
|
835 | 835 | e = index[iterrev] |
|
836 | 836 | else: |
|
837 | 837 | # Add text length of base since decompressing that also takes |
|
838 | 838 | # work. For cache hits the length is already included. |
|
839 | 839 | compresseddeltalen += e[1] |
|
840 | 840 | r = (clen, compresseddeltalen) |
|
841 | 841 | chaininfocache[rev] = r |
|
842 | 842 | return r |
|
843 | 843 | |
|
844 | 844 | def _deltachain(self, rev, stoprev=None): |
|
845 | 845 | """Obtain the delta chain for a revision. |
|
846 | 846 | |
|
847 | 847 | ``stoprev`` specifies a revision to stop at. If not specified, we |
|
848 | 848 | stop at the base of the chain. |
|
849 | 849 | |
|
850 | 850 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of |
|
851 | 851 | revs in ascending order and ``stopped`` is a bool indicating whether |
|
852 | 852 | ``stoprev`` was hit. |
|
853 | 853 | """ |
|
854 | 854 | # Try C implementation. |
|
855 | 855 | try: |
|
856 | 856 | return self.index.deltachain(rev, stoprev, self._generaldelta) |
|
857 | 857 | except AttributeError: |
|
858 | 858 | pass |
|
859 | 859 | |
|
860 | 860 | chain = [] |
|
861 | 861 | |
|
862 | 862 | # Alias to prevent attribute lookup in tight loop. |
|
863 | 863 | index = self.index |
|
864 | 864 | generaldelta = self._generaldelta |
|
865 | 865 | |
|
866 | 866 | iterrev = rev |
|
867 | 867 | e = index[iterrev] |
|
868 | 868 | while iterrev != e[3] and iterrev != stoprev: |
|
869 | 869 | chain.append(iterrev) |
|
870 | 870 | if generaldelta: |
|
871 | 871 | iterrev = e[3] |
|
872 | 872 | else: |
|
873 | 873 | iterrev -= 1 |
|
874 | 874 | e = index[iterrev] |
|
875 | 875 | |
|
876 | 876 | if iterrev == stoprev: |
|
877 | 877 | stopped = True |
|
878 | 878 | else: |
|
879 | 879 | chain.append(iterrev) |
|
880 | 880 | stopped = False |
|
881 | 881 | |
|
882 | 882 | chain.reverse() |
|
883 | 883 | return chain, stopped |
|
884 | 884 | |
|
885 | 885 | def ancestors(self, revs, stoprev=0, inclusive=False): |
|
886 | 886 | """Generate the ancestors of 'revs' in reverse revision order. |
|
887 | 887 | Does not generate revs lower than stoprev. |
|
888 | 888 | |
|
889 | 889 | See the documentation for ancestor.lazyancestors for more details.""" |
|
890 | 890 | |
|
891 | 891 | # first, make sure start revisions aren't filtered |
|
892 | 892 | revs = list(revs) |
|
893 | 893 | checkrev = self.node |
|
894 | 894 | for r in revs: |
|
895 | 895 | checkrev(r) |
|
896 | 896 | # and we're sure ancestors aren't filtered as well |
|
897 | 897 | |
|
898 | 898 | if rustancestor is not None: |
|
899 | 899 | lazyancestors = rustancestor.LazyAncestors |
|
900 | 900 | arg = self.index |
|
901 | 901 | elif util.safehasattr(parsers, b'rustlazyancestors'): |
|
902 | 902 | lazyancestors = ancestor.rustlazyancestors |
|
903 | 903 | arg = self.index |
|
904 | 904 | else: |
|
905 | 905 | lazyancestors = ancestor.lazyancestors |
|
906 | 906 | arg = self._uncheckedparentrevs |
|
907 | 907 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) |
|
908 | 908 | |
|
909 | 909 | def descendants(self, revs): |
|
910 | 910 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
911 | 911 | |
|
912 | 912 | def findcommonmissing(self, common=None, heads=None): |
|
913 | 913 | """Return a tuple of the ancestors of common and the ancestors of heads |
|
914 | 914 | that are not ancestors of common. In revset terminology, we return the |
|
915 | 915 | tuple: |
|
916 | 916 | |
|
917 | 917 | ::common, (::heads) - (::common) |
|
918 | 918 | |
|
919 | 919 | The list is sorted by revision number, meaning it is |
|
920 | 920 | topologically sorted. |
|
921 | 921 | |
|
922 | 922 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
923 | 923 | not supplied, uses all of the revlog's heads. If common is not |
|
924 | 924 | supplied, uses nullid.""" |
|
925 | 925 | if common is None: |
|
926 | 926 | common = [nullid] |
|
927 | 927 | if heads is None: |
|
928 | 928 | heads = self.heads() |
|
929 | 929 | |
|
930 | 930 | common = [self.rev(n) for n in common] |
|
931 | 931 | heads = [self.rev(n) for n in heads] |
|
932 | 932 | |
|
933 | 933 | # we want the ancestors, but inclusive |
|
934 | 934 | class lazyset(object): |
|
935 | 935 | def __init__(self, lazyvalues): |
|
936 | 936 | self.addedvalues = set() |
|
937 | 937 | self.lazyvalues = lazyvalues |
|
938 | 938 | |
|
939 | 939 | def __contains__(self, value): |
|
940 | 940 | return value in self.addedvalues or value in self.lazyvalues |
|
941 | 941 | |
|
942 | 942 | def __iter__(self): |
|
943 | 943 | added = self.addedvalues |
|
944 | 944 | for r in added: |
|
945 | 945 | yield r |
|
946 | 946 | for r in self.lazyvalues: |
|
947 | 947 | if not r in added: |
|
948 | 948 | yield r |
|
949 | 949 | |
|
950 | 950 | def add(self, value): |
|
951 | 951 | self.addedvalues.add(value) |
|
952 | 952 | |
|
953 | 953 | def update(self, values): |
|
954 | 954 | self.addedvalues.update(values) |
|
955 | 955 | |
|
956 | 956 | has = lazyset(self.ancestors(common)) |
|
957 | 957 | has.add(nullrev) |
|
958 | 958 | has.update(common) |
|
959 | 959 | |
|
960 | 960 | # take all ancestors from heads that aren't in has |
|
961 | 961 | missing = set() |
|
962 | 962 | visit = collections.deque(r for r in heads if r not in has) |
|
963 | 963 | while visit: |
|
964 | 964 | r = visit.popleft() |
|
965 | 965 | if r in missing: |
|
966 | 966 | continue |
|
967 | 967 | else: |
|
968 | 968 | missing.add(r) |
|
969 | 969 | for p in self.parentrevs(r): |
|
970 | 970 | if p not in has: |
|
971 | 971 | visit.append(p) |
|
972 | 972 | missing = list(missing) |
|
973 | 973 | missing.sort() |
|
974 | 974 | return has, [self.node(miss) for miss in missing] |
|
975 | 975 | |
|
976 | 976 | def incrementalmissingrevs(self, common=None): |
|
977 | 977 | """Return an object that can be used to incrementally compute the |
|
978 | 978 | revision numbers of the ancestors of arbitrary sets that are not |
|
979 | 979 | ancestors of common. This is an ancestor.incrementalmissingancestors |
|
980 | 980 | object. |
|
981 | 981 | |
|
982 | 982 | 'common' is a list of revision numbers. If common is not supplied, uses |
|
983 | 983 | nullrev. |
|
984 | 984 | """ |
|
985 | 985 | if common is None: |
|
986 | 986 | common = [nullrev] |
|
987 | 987 | |
|
988 | 988 | if rustancestor is not None: |
|
989 | 989 | return rustancestor.MissingAncestors(self.index, common) |
|
990 | 990 | return ancestor.incrementalmissingancestors(self.parentrevs, common) |
|
991 | 991 | |
|
992 | 992 | def findmissingrevs(self, common=None, heads=None): |
|
993 | 993 | """Return the revision numbers of the ancestors of heads that |
|
994 | 994 | are not ancestors of common. |
|
995 | 995 | |
|
996 | 996 | More specifically, return a list of revision numbers corresponding to |
|
997 | 997 | nodes N such that every N satisfies the following constraints: |
|
998 | 998 | |
|
999 | 999 | 1. N is an ancestor of some node in 'heads' |
|
1000 | 1000 | 2. N is not an ancestor of any node in 'common' |
|
1001 | 1001 | |
|
1002 | 1002 | The list is sorted by revision number, meaning it is |
|
1003 | 1003 | topologically sorted. |
|
1004 | 1004 | |
|
1005 | 1005 | 'heads' and 'common' are both lists of revision numbers. If heads is |
|
1006 | 1006 | not supplied, uses all of the revlog's heads. If common is not |
|
1007 | 1007 | supplied, uses nullid.""" |
|
1008 | 1008 | if common is None: |
|
1009 | 1009 | common = [nullrev] |
|
1010 | 1010 | if heads is None: |
|
1011 | 1011 | heads = self.headrevs() |
|
1012 | 1012 | |
|
1013 | 1013 | inc = self.incrementalmissingrevs(common=common) |
|
1014 | 1014 | return inc.missingancestors(heads) |
|
1015 | 1015 | |
|
1016 | 1016 | def findmissing(self, common=None, heads=None): |
|
1017 | 1017 | """Return the ancestors of heads that are not ancestors of common. |
|
1018 | 1018 | |
|
1019 | 1019 | More specifically, return a list of nodes N such that every N |
|
1020 | 1020 | satisfies the following constraints: |
|
1021 | 1021 | |
|
1022 | 1022 | 1. N is an ancestor of some node in 'heads' |
|
1023 | 1023 | 2. N is not an ancestor of any node in 'common' |
|
1024 | 1024 | |
|
1025 | 1025 | The list is sorted by revision number, meaning it is |
|
1026 | 1026 | topologically sorted. |
|
1027 | 1027 | |
|
1028 | 1028 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1029 | 1029 | not supplied, uses all of the revlog's heads. If common is not |
|
1030 | 1030 | supplied, uses nullid.""" |
|
1031 | 1031 | if common is None: |
|
1032 | 1032 | common = [nullid] |
|
1033 | 1033 | if heads is None: |
|
1034 | 1034 | heads = self.heads() |
|
1035 | 1035 | |
|
1036 | 1036 | common = [self.rev(n) for n in common] |
|
1037 | 1037 | heads = [self.rev(n) for n in heads] |
|
1038 | 1038 | |
|
1039 | 1039 | inc = self.incrementalmissingrevs(common=common) |
|
1040 | 1040 | return [self.node(r) for r in inc.missingancestors(heads)] |
|
1041 | 1041 | |
|
1042 | 1042 | def nodesbetween(self, roots=None, heads=None): |
|
1043 | 1043 | """Return a topological path from 'roots' to 'heads'. |
|
1044 | 1044 | |
|
1045 | 1045 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a |
|
1046 | 1046 | topologically sorted list of all nodes N that satisfy both of |
|
1047 | 1047 | these constraints: |
|
1048 | 1048 | |
|
1049 | 1049 | 1. N is a descendant of some node in 'roots' |
|
1050 | 1050 | 2. N is an ancestor of some node in 'heads' |
|
1051 | 1051 | |
|
1052 | 1052 | Every node is considered to be both a descendant and an ancestor |
|
1053 | 1053 | of itself, so every reachable node in 'roots' and 'heads' will be |
|
1054 | 1054 | included in 'nodes'. |
|
1055 | 1055 | |
|
1056 | 1056 | 'outroots' is the list of reachable nodes in 'roots', i.e., the |
|
1057 | 1057 | subset of 'roots' that is returned in 'nodes'. Likewise, |
|
1058 | 1058 | 'outheads' is the subset of 'heads' that is also in 'nodes'. |
|
1059 | 1059 | |
|
1060 | 1060 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is |
|
1061 | 1061 | unspecified, uses nullid as the only root. If 'heads' is |
|
1062 | 1062 | unspecified, uses list of all of the revlog's heads.""" |
|
1063 | 1063 | nonodes = ([], [], []) |
|
1064 | 1064 | if roots is not None: |
|
1065 | 1065 | roots = list(roots) |
|
1066 | 1066 | if not roots: |
|
1067 | 1067 | return nonodes |
|
1068 | 1068 | lowestrev = min([self.rev(n) for n in roots]) |
|
1069 | 1069 | else: |
|
1070 | 1070 | roots = [nullid] # Everybody's a descendant of nullid |
|
1071 | 1071 | lowestrev = nullrev |
|
1072 | 1072 | if (lowestrev == nullrev) and (heads is None): |
|
1073 | 1073 | # We want _all_ the nodes! |
|
1074 | 1074 | return ([self.node(r) for r in self], [nullid], list(self.heads())) |
|
1075 | 1075 | if heads is None: |
|
1076 | 1076 | # All nodes are ancestors, so the latest ancestor is the last |
|
1077 | 1077 | # node. |
|
1078 | 1078 | highestrev = len(self) - 1 |
|
1079 | 1079 | # Set ancestors to None to signal that every node is an ancestor. |
|
1080 | 1080 | ancestors = None |
|
1081 | 1081 | # Set heads to an empty dictionary for later discovery of heads |
|
1082 | 1082 | heads = {} |
|
1083 | 1083 | else: |
|
1084 | 1084 | heads = list(heads) |
|
1085 | 1085 | if not heads: |
|
1086 | 1086 | return nonodes |
|
1087 | 1087 | ancestors = set() |
|
1088 | 1088 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
1089 | 1089 | # Also, later we will be using it to filter out the heads we can't |
|
1090 | 1090 | # find from roots. |
|
1091 | 1091 | heads = dict.fromkeys(heads, False) |
|
1092 | 1092 | # Start at the top and keep marking parents until we're done. |
|
1093 | 1093 | nodestotag = set(heads) |
|
1094 | 1094 | # Remember where the top was so we can use it as a limit later. |
|
1095 | 1095 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
1096 | 1096 | while nodestotag: |
|
1097 | 1097 | # grab a node to tag |
|
1098 | 1098 | n = nodestotag.pop() |
|
1099 | 1099 | # Never tag nullid |
|
1100 | 1100 | if n == nullid: |
|
1101 | 1101 | continue |
|
1102 | 1102 | # A node's revision number represents its place in a |
|
1103 | 1103 | # topologically sorted list of nodes. |
|
1104 | 1104 | r = self.rev(n) |
|
1105 | 1105 | if r >= lowestrev: |
|
1106 | 1106 | if n not in ancestors: |
|
1107 | 1107 | # If we are possibly a descendant of one of the roots |
|
1108 | 1108 | # and we haven't already been marked as an ancestor |
|
1109 | 1109 | ancestors.add(n) # Mark as ancestor |
|
1110 | 1110 | # Add non-nullid parents to list of nodes to tag. |
|
1111 | 1111 | nodestotag.update( |
|
1112 | 1112 | [p for p in self.parents(n) if p != nullid] |
|
1113 | 1113 | ) |
|
1114 | 1114 | elif n in heads: # We've seen it before, is it a fake head? |
|
1115 | 1115 | # So it is, real heads should not be the ancestors of |
|
1116 | 1116 | # any other heads. |
|
1117 | 1117 | heads.pop(n) |
|
1118 | 1118 | if not ancestors: |
|
1119 | 1119 | return nonodes |
|
1120 | 1120 | # Now that we have our set of ancestors, we want to remove any |
|
1121 | 1121 | # roots that are not ancestors. |
|
1122 | 1122 | |
|
1123 | 1123 | # If one of the roots was nullid, everything is included anyway. |
|
1124 | 1124 | if lowestrev > nullrev: |
|
1125 | 1125 | # But, since we weren't, let's recompute the lowest rev to not |
|
1126 | 1126 | # include roots that aren't ancestors. |
|
1127 | 1127 | |
|
1128 | 1128 | # Filter out roots that aren't ancestors of heads |
|
1129 | 1129 | roots = [root for root in roots if root in ancestors] |
|
1130 | 1130 | # Recompute the lowest revision |
|
1131 | 1131 | if roots: |
|
1132 | 1132 | lowestrev = min([self.rev(root) for root in roots]) |
|
1133 | 1133 | else: |
|
1134 | 1134 | # No more roots? Return empty list |
|
1135 | 1135 | return nonodes |
|
1136 | 1136 | else: |
|
1137 | 1137 | # We are descending from nullid, and don't need to care about |
|
1138 | 1138 | # any other roots. |
|
1139 | 1139 | lowestrev = nullrev |
|
1140 | 1140 | roots = [nullid] |
|
1141 | 1141 | # Transform our roots list into a set. |
|
1142 | 1142 | descendants = set(roots) |
|
1143 | 1143 | # Also, keep the original roots so we can filter out roots that aren't |
|
1144 | 1144 | # 'real' roots (i.e. are descended from other roots). |
|
1145 | 1145 | roots = descendants.copy() |
|
1146 | 1146 | # Our topologically sorted list of output nodes. |
|
1147 | 1147 | orderedout = [] |
|
1148 | 1148 | # Don't start at nullid since we don't want nullid in our output list, |
|
1149 | 1149 | # and if nullid shows up in descendants, empty parents will look like |
|
1150 | 1150 | # they're descendants. |
|
1151 | 1151 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): |
|
1152 | 1152 | n = self.node(r) |
|
1153 | 1153 | isdescendant = False |
|
1154 | 1154 | if lowestrev == nullrev: # Everybody is a descendant of nullid |
|
1155 | 1155 | isdescendant = True |
|
1156 | 1156 | elif n in descendants: |
|
1157 | 1157 | # n is already a descendant |
|
1158 | 1158 | isdescendant = True |
|
1159 | 1159 | # This check only needs to be done here because all the roots |
|
1160 | 1160 | # will start being marked is descendants before the loop. |
|
1161 | 1161 | if n in roots: |
|
1162 | 1162 | # If n was a root, check if it's a 'real' root. |
|
1163 | 1163 | p = tuple(self.parents(n)) |
|
1164 | 1164 | # If any of its parents are descendants, it's not a root. |
|
1165 | 1165 | if (p[0] in descendants) or (p[1] in descendants): |
|
1166 | 1166 | roots.remove(n) |
|
1167 | 1167 | else: |
|
1168 | 1168 | p = tuple(self.parents(n)) |
|
1169 | 1169 | # A node is a descendant if either of its parents are |
|
1170 | 1170 | # descendants. (We seeded the dependents list with the roots |
|
1171 | 1171 | # up there, remember?) |
|
1172 | 1172 | if (p[0] in descendants) or (p[1] in descendants): |
|
1173 | 1173 | descendants.add(n) |
|
1174 | 1174 | isdescendant = True |
|
1175 | 1175 | if isdescendant and ((ancestors is None) or (n in ancestors)): |
|
1176 | 1176 | # Only include nodes that are both descendants and ancestors. |
|
1177 | 1177 | orderedout.append(n) |
|
1178 | 1178 | if (ancestors is not None) and (n in heads): |
|
1179 | 1179 | # We're trying to figure out which heads are reachable |
|
1180 | 1180 | # from roots. |
|
1181 | 1181 | # Mark this head as having been reached |
|
1182 | 1182 | heads[n] = True |
|
1183 | 1183 | elif ancestors is None: |
|
1184 | 1184 | # Otherwise, we're trying to discover the heads. |
|
1185 | 1185 | # Assume this is a head because if it isn't, the next step |
|
1186 | 1186 | # will eventually remove it. |
|
1187 | 1187 | heads[n] = True |
|
1188 | 1188 | # But, obviously its parents aren't. |
|
1189 | 1189 | for p in self.parents(n): |
|
1190 | 1190 | heads.pop(p, None) |
|
1191 | 1191 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] |
|
1192 | 1192 | roots = list(roots) |
|
1193 | 1193 | assert orderedout |
|
1194 | 1194 | assert roots |
|
1195 | 1195 | assert heads |
|
1196 | 1196 | return (orderedout, roots, heads) |
|
1197 | 1197 | |
|
1198 | 1198 | def headrevs(self, revs=None): |
|
1199 | 1199 | if revs is None: |
|
1200 | 1200 | try: |
|
1201 | 1201 | return self.index.headrevs() |
|
1202 | 1202 | except AttributeError: |
|
1203 | 1203 | return self._headrevs() |
|
1204 | 1204 | if rustdagop is not None: |
|
1205 | 1205 | return rustdagop.headrevs(self.index, revs) |
|
1206 | 1206 | return dagop.headrevs(revs, self._uncheckedparentrevs) |
|
1207 | 1207 | |
|
1208 | 1208 | def computephases(self, roots): |
|
1209 | 1209 | return self.index.computephasesmapsets(roots) |
|
1210 | 1210 | |
|
1211 | 1211 | def _headrevs(self): |
|
1212 | 1212 | count = len(self) |
|
1213 | 1213 | if not count: |
|
1214 | 1214 | return [nullrev] |
|
1215 | 1215 | # we won't iter over filtered rev so nobody is a head at start |
|
1216 | 1216 | ishead = [0] * (count + 1) |
|
1217 | 1217 | index = self.index |
|
1218 | 1218 | for r in self: |
|
1219 | 1219 | ishead[r] = 1 # I may be an head |
|
1220 | 1220 | e = index[r] |
|
1221 | 1221 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not |
|
1222 | 1222 | return [r for r, val in enumerate(ishead) if val] |
|
1223 | 1223 | |
|
1224 | 1224 | def heads(self, start=None, stop=None): |
|
1225 | 1225 | """return the list of all nodes that have no children |
|
1226 | 1226 | |
|
1227 | 1227 | if start is specified, only heads that are descendants of |
|
1228 | 1228 | start will be returned |
|
1229 | 1229 | if stop is specified, it will consider all the revs from stop |
|
1230 | 1230 | as if they had no children |
|
1231 | 1231 | """ |
|
1232 | 1232 | if start is None and stop is None: |
|
1233 | 1233 | if not len(self): |
|
1234 | 1234 | return [nullid] |
|
1235 | 1235 | return [self.node(r) for r in self.headrevs()] |
|
1236 | 1236 | |
|
1237 | 1237 | if start is None: |
|
1238 | 1238 | start = nullrev |
|
1239 | 1239 | else: |
|
1240 | 1240 | start = self.rev(start) |
|
1241 | 1241 | |
|
1242 | 1242 | stoprevs = set(self.rev(n) for n in stop or []) |
|
1243 | 1243 | |
|
1244 | 1244 | revs = dagop.headrevssubset( |
|
1245 | 1245 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs |
|
1246 | 1246 | ) |
|
1247 | 1247 | |
|
1248 | 1248 | return [self.node(rev) for rev in revs] |
|
1249 | 1249 | |
|
1250 | 1250 | def children(self, node): |
|
1251 | 1251 | """find the children of a given node""" |
|
1252 | 1252 | c = [] |
|
1253 | 1253 | p = self.rev(node) |
|
1254 | 1254 | for r in self.revs(start=p + 1): |
|
1255 | 1255 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
1256 | 1256 | if prevs: |
|
1257 | 1257 | for pr in prevs: |
|
1258 | 1258 | if pr == p: |
|
1259 | 1259 | c.append(self.node(r)) |
|
1260 | 1260 | elif p == nullrev: |
|
1261 | 1261 | c.append(self.node(r)) |
|
1262 | 1262 | return c |
|
1263 | 1263 | |
|
1264 | 1264 | def commonancestorsheads(self, a, b): |
|
1265 | 1265 | """calculate all the heads of the common ancestors of nodes a and b""" |
|
1266 | 1266 | a, b = self.rev(a), self.rev(b) |
|
1267 | 1267 | ancs = self._commonancestorsheads(a, b) |
|
1268 | 1268 | return pycompat.maplist(self.node, ancs) |
|
1269 | 1269 | |
|
1270 | 1270 | def _commonancestorsheads(self, *revs): |
|
1271 | 1271 | """calculate all the heads of the common ancestors of revs""" |
|
1272 | 1272 | try: |
|
1273 | 1273 | ancs = self.index.commonancestorsheads(*revs) |
|
1274 | 1274 | except (AttributeError, OverflowError): # C implementation failed |
|
1275 | 1275 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) |
|
1276 | 1276 | return ancs |
|
1277 | 1277 | |
|
1278 | 1278 | def isancestor(self, a, b): |
|
1279 | 1279 | """return True if node a is an ancestor of node b |
|
1280 | 1280 | |
|
1281 | 1281 | A revision is considered an ancestor of itself.""" |
|
1282 | 1282 | a, b = self.rev(a), self.rev(b) |
|
1283 | 1283 | return self.isancestorrev(a, b) |
|
1284 | 1284 | |
|
1285 | 1285 | def isancestorrev(self, a, b): |
|
1286 | 1286 | """return True if revision a is an ancestor of revision b |
|
1287 | 1287 | |
|
1288 | 1288 | A revision is considered an ancestor of itself. |
|
1289 | 1289 | |
|
1290 | 1290 | The implementation of this is trivial but the use of |
|
1291 | 1291 | reachableroots is not.""" |
|
1292 | 1292 | if a == nullrev: |
|
1293 | 1293 | return True |
|
1294 | 1294 | elif a == b: |
|
1295 | 1295 | return True |
|
1296 | 1296 | elif a > b: |
|
1297 | 1297 | return False |
|
1298 | 1298 | return bool(self.reachableroots(a, [b], [a], includepath=False)) |
|
1299 | 1299 | |
|
1300 | 1300 | def reachableroots(self, minroot, heads, roots, includepath=False): |
|
1301 | 1301 | """return (heads(::(<roots> and <roots>::<heads>))) |
|
1302 | 1302 | |
|
1303 | 1303 | If includepath is True, return (<roots>::<heads>).""" |
|
1304 | 1304 | try: |
|
1305 | 1305 | return self.index.reachableroots2( |
|
1306 | 1306 | minroot, heads, roots, includepath |
|
1307 | 1307 | ) |
|
1308 | 1308 | except AttributeError: |
|
1309 | 1309 | return dagop._reachablerootspure( |
|
1310 | 1310 | self.parentrevs, minroot, roots, heads, includepath |
|
1311 | 1311 | ) |
|
1312 | 1312 | |
|
1313 | 1313 | def ancestor(self, a, b): |
|
1314 | 1314 | """calculate the "best" common ancestor of nodes a and b""" |
|
1315 | 1315 | |
|
1316 | 1316 | a, b = self.rev(a), self.rev(b) |
|
1317 | 1317 | try: |
|
1318 | 1318 | ancs = self.index.ancestors(a, b) |
|
1319 | 1319 | except (AttributeError, OverflowError): |
|
1320 | 1320 | ancs = ancestor.ancestors(self.parentrevs, a, b) |
|
1321 | 1321 | if ancs: |
|
1322 | 1322 | # choose a consistent winner when there's a tie |
|
1323 | 1323 | return min(map(self.node, ancs)) |
|
1324 | 1324 | return nullid |
|
1325 | 1325 | |
|
1326 | 1326 | def _match(self, id): |
|
1327 | 1327 | if isinstance(id, int): |
|
1328 | 1328 | # rev |
|
1329 | 1329 | return self.node(id) |
|
1330 | 1330 | if len(id) == 20: |
|
1331 | 1331 | # possibly a binary node |
|
1332 | 1332 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
1333 | 1333 | try: |
|
1334 | 1334 | node = id |
|
1335 | 1335 | self.rev(node) # quick search the index |
|
1336 | 1336 | return node |
|
1337 | 1337 | except error.LookupError: |
|
1338 | 1338 | pass # may be partial hex id |
|
1339 | 1339 | try: |
|
1340 | 1340 | # str(rev) |
|
1341 | 1341 | rev = int(id) |
|
1342 | 1342 | if b"%d" % rev != id: |
|
1343 | 1343 | raise ValueError |
|
1344 | 1344 | if rev < 0: |
|
1345 | 1345 | rev = len(self) + rev |
|
1346 | 1346 | if rev < 0 or rev >= len(self): |
|
1347 | 1347 | raise ValueError |
|
1348 | 1348 | return self.node(rev) |
|
1349 | 1349 | except (ValueError, OverflowError): |
|
1350 | 1350 | pass |
|
1351 | 1351 | if len(id) == 40: |
|
1352 | 1352 | try: |
|
1353 | 1353 | # a full hex nodeid? |
|
1354 | 1354 | node = bin(id) |
|
1355 | 1355 | self.rev(node) |
|
1356 | 1356 | return node |
|
1357 | 1357 | except (TypeError, error.LookupError): |
|
1358 | 1358 | pass |
|
1359 | 1359 | |
|
1360 | 1360 | def _partialmatch(self, id): |
|
1361 | 1361 | # we don't care wdirfilenodeids as they should be always full hash |
|
1362 | 1362 | maybewdir = wdirhex.startswith(id) |
|
1363 | 1363 | try: |
|
1364 | 1364 | partial = self.index.partialmatch(id) |
|
1365 | 1365 | if partial and self.hasnode(partial): |
|
1366 | 1366 | if maybewdir: |
|
1367 | 1367 | # single 'ff...' match in radix tree, ambiguous with wdir |
|
1368 | 1368 | raise error.RevlogError |
|
1369 | 1369 | return partial |
|
1370 | 1370 | if maybewdir: |
|
1371 | 1371 | # no 'ff...' match in radix tree, wdir identified |
|
1372 | 1372 | raise error.WdirUnsupported |
|
1373 | 1373 | return None |
|
1374 | 1374 | except error.RevlogError: |
|
1375 | 1375 | # parsers.c radix tree lookup gave multiple matches |
|
1376 | 1376 | # fast path: for unfiltered changelog, radix tree is accurate |
|
1377 | 1377 | if not getattr(self, 'filteredrevs', None): |
|
1378 | 1378 | raise error.AmbiguousPrefixLookupError( |
|
1379 | 1379 | id, self.indexfile, _(b'ambiguous identifier') |
|
1380 | 1380 | ) |
|
1381 | 1381 | # fall through to slow path that filters hidden revisions |
|
1382 | 1382 | except (AttributeError, ValueError): |
|
1383 | 1383 | # we are pure python, or key was too short to search radix tree |
|
1384 | 1384 | pass |
|
1385 | 1385 | |
|
1386 | 1386 | if id in self._pcache: |
|
1387 | 1387 | return self._pcache[id] |
|
1388 | 1388 | |
|
1389 | 1389 | if len(id) <= 40: |
|
1390 | 1390 | try: |
|
1391 | 1391 | # hex(node)[:...] |
|
1392 | 1392 | l = len(id) // 2 # grab an even number of digits |
|
1393 | 1393 | prefix = bin(id[: l * 2]) |
|
1394 | 1394 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] |
|
1395 | 1395 | nl = [ |
|
1396 | 1396 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) |
|
1397 | 1397 | ] |
|
1398 | 1398 | if nullhex.startswith(id): |
|
1399 | 1399 | nl.append(nullid) |
|
1400 | 1400 | if len(nl) > 0: |
|
1401 | 1401 | if len(nl) == 1 and not maybewdir: |
|
1402 | 1402 | self._pcache[id] = nl[0] |
|
1403 | 1403 | return nl[0] |
|
1404 | 1404 | raise error.AmbiguousPrefixLookupError( |
|
1405 | 1405 | id, self.indexfile, _(b'ambiguous identifier') |
|
1406 | 1406 | ) |
|
1407 | 1407 | if maybewdir: |
|
1408 | 1408 | raise error.WdirUnsupported |
|
1409 | 1409 | return None |
|
1410 | 1410 | except TypeError: |
|
1411 | 1411 | pass |
|
1412 | 1412 | |
|
1413 | 1413 | def lookup(self, id): |
|
1414 | 1414 | """locate a node based on: |
|
1415 | 1415 | - revision number or str(revision number) |
|
1416 | 1416 | - nodeid or subset of hex nodeid |
|
1417 | 1417 | """ |
|
1418 | 1418 | n = self._match(id) |
|
1419 | 1419 | if n is not None: |
|
1420 | 1420 | return n |
|
1421 | 1421 | n = self._partialmatch(id) |
|
1422 | 1422 | if n: |
|
1423 | 1423 | return n |
|
1424 | 1424 | |
|
1425 | 1425 | raise error.LookupError(id, self.indexfile, _(b'no match found')) |
|
1426 | 1426 | |
|
1427 | 1427 | def shortest(self, node, minlength=1): |
|
1428 | 1428 | """Find the shortest unambiguous prefix that matches node.""" |
|
1429 | 1429 | |
|
1430 | 1430 | def isvalid(prefix): |
|
1431 | 1431 | try: |
|
1432 | 1432 | matchednode = self._partialmatch(prefix) |
|
1433 | 1433 | except error.AmbiguousPrefixLookupError: |
|
1434 | 1434 | return False |
|
1435 | 1435 | except error.WdirUnsupported: |
|
1436 | 1436 | # single 'ff...' match |
|
1437 | 1437 | return True |
|
1438 | 1438 | if matchednode is None: |
|
1439 | 1439 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
1440 | 1440 | return True |
|
1441 | 1441 | |
|
1442 | 1442 | def maybewdir(prefix): |
|
1443 | 1443 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) |
|
1444 | 1444 | |
|
1445 | 1445 | hexnode = hex(node) |
|
1446 | 1446 | |
|
1447 | 1447 | def disambiguate(hexnode, minlength): |
|
1448 | 1448 | """Disambiguate against wdirid.""" |
|
1449 | 1449 | for length in range(minlength, 41): |
|
1450 | 1450 | prefix = hexnode[:length] |
|
1451 | 1451 | if not maybewdir(prefix): |
|
1452 | 1452 | return prefix |
|
1453 | 1453 | |
|
1454 | 1454 | if not getattr(self, 'filteredrevs', None): |
|
1455 | 1455 | try: |
|
1456 | 1456 | length = max(self.index.shortest(node), minlength) |
|
1457 | 1457 | return disambiguate(hexnode, length) |
|
1458 | 1458 | except error.RevlogError: |
|
1459 | 1459 | if node != wdirid: |
|
1460 | 1460 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
1461 | 1461 | except AttributeError: |
|
1462 | 1462 | # Fall through to pure code |
|
1463 | 1463 | pass |
|
1464 | 1464 | |
|
1465 | 1465 | if node == wdirid: |
|
1466 | 1466 | for length in range(minlength, 41): |
|
1467 | 1467 | prefix = hexnode[:length] |
|
1468 | 1468 | if isvalid(prefix): |
|
1469 | 1469 | return prefix |
|
1470 | 1470 | |
|
1471 | 1471 | for length in range(minlength, 41): |
|
1472 | 1472 | prefix = hexnode[:length] |
|
1473 | 1473 | if isvalid(prefix): |
|
1474 | 1474 | return disambiguate(hexnode, length) |
|
1475 | 1475 | |
|
1476 | 1476 | def cmp(self, node, text): |
|
1477 | 1477 | """compare text with a given file revision |
|
1478 | 1478 | |
|
1479 | 1479 | returns True if text is different than what is stored. |
|
1480 | 1480 | """ |
|
1481 | 1481 | p1, p2 = self.parents(node) |
|
1482 | 1482 | return storageutil.hashrevisionsha1(text, p1, p2) != node |
|
1483 | 1483 | |
|
1484 | 1484 | def _cachesegment(self, offset, data): |
|
1485 | 1485 | """Add a segment to the revlog cache. |
|
1486 | 1486 | |
|
1487 | 1487 | Accepts an absolute offset and the data that is at that location. |
|
1488 | 1488 | """ |
|
1489 | 1489 | o, d = self._chunkcache |
|
1490 | 1490 | # try to add to existing cache |
|
1491 | 1491 | if o + len(d) == offset and len(d) + len(data) < _chunksize: |
|
1492 | 1492 | self._chunkcache = o, d + data |
|
1493 | 1493 | else: |
|
1494 | 1494 | self._chunkcache = offset, data |
|
1495 | 1495 | |
|
1496 | 1496 | def _readsegment(self, offset, length, df=None): |
|
1497 | 1497 | """Load a segment of raw data from the revlog. |
|
1498 | 1498 | |
|
1499 | 1499 | Accepts an absolute offset, length to read, and an optional existing |
|
1500 | 1500 | file handle to read from. |
|
1501 | 1501 | |
|
1502 | 1502 | If an existing file handle is passed, it will be seeked and the |
|
1503 | 1503 | original seek position will NOT be restored. |
|
1504 | 1504 | |
|
1505 | 1505 | Returns a str or buffer of raw byte data. |
|
1506 | 1506 | |
|
1507 | 1507 | Raises if the requested number of bytes could not be read. |
|
1508 | 1508 | """ |
|
1509 | 1509 | # Cache data both forward and backward around the requested |
|
1510 | 1510 | # data, in a fixed size window. This helps speed up operations |
|
1511 | 1511 | # involving reading the revlog backwards. |
|
1512 | 1512 | cachesize = self._chunkcachesize |
|
1513 | 1513 | realoffset = offset & ~(cachesize - 1) |
|
1514 | 1514 | reallength = ( |
|
1515 | 1515 | (offset + length + cachesize) & ~(cachesize - 1) |
|
1516 | 1516 | ) - realoffset |
|
1517 | 1517 | with self._datareadfp(df) as df: |
|
1518 | 1518 | df.seek(realoffset) |
|
1519 | 1519 | d = df.read(reallength) |
|
1520 | 1520 | |
|
1521 | 1521 | self._cachesegment(realoffset, d) |
|
1522 | 1522 | if offset != realoffset or reallength != length: |
|
1523 | 1523 | startoffset = offset - realoffset |
|
1524 | 1524 | if len(d) - startoffset < length: |
|
1525 | 1525 | raise error.RevlogError( |
|
1526 | 1526 | _( |
|
1527 | 1527 | b'partial read of revlog %s; expected %d bytes from ' |
|
1528 | 1528 | b'offset %d, got %d' |
|
1529 | 1529 | ) |
|
1530 | 1530 | % ( |
|
1531 | 1531 | self.indexfile if self._inline else self.datafile, |
|
1532 | 1532 | length, |
|
1533 | 1533 | realoffset, |
|
1534 | 1534 | len(d) - startoffset, |
|
1535 | 1535 | ) |
|
1536 | 1536 | ) |
|
1537 | 1537 | |
|
1538 | 1538 | return util.buffer(d, startoffset, length) |
|
1539 | 1539 | |
|
1540 | 1540 | if len(d) < length: |
|
1541 | 1541 | raise error.RevlogError( |
|
1542 | 1542 | _( |
|
1543 | 1543 | b'partial read of revlog %s; expected %d bytes from offset ' |
|
1544 | 1544 | b'%d, got %d' |
|
1545 | 1545 | ) |
|
1546 | 1546 | % ( |
|
1547 | 1547 | self.indexfile if self._inline else self.datafile, |
|
1548 | 1548 | length, |
|
1549 | 1549 | offset, |
|
1550 | 1550 | len(d), |
|
1551 | 1551 | ) |
|
1552 | 1552 | ) |
|
1553 | 1553 | |
|
1554 | 1554 | return d |
|
1555 | 1555 | |
|
1556 | 1556 | def _getsegment(self, offset, length, df=None): |
|
1557 | 1557 | """Obtain a segment of raw data from the revlog. |
|
1558 | 1558 | |
|
1559 | 1559 | Accepts an absolute offset, length of bytes to obtain, and an |
|
1560 | 1560 | optional file handle to the already-opened revlog. If the file |
|
1561 | 1561 | handle is used, it's original seek position will not be preserved. |
|
1562 | 1562 | |
|
1563 | 1563 | Requests for data may be returned from a cache. |
|
1564 | 1564 | |
|
1565 | 1565 | Returns a str or a buffer instance of raw byte data. |
|
1566 | 1566 | """ |
|
1567 | 1567 | o, d = self._chunkcache |
|
1568 | 1568 | l = len(d) |
|
1569 | 1569 | |
|
1570 | 1570 | # is it in the cache? |
|
1571 | 1571 | cachestart = offset - o |
|
1572 | 1572 | cacheend = cachestart + length |
|
1573 | 1573 | if cachestart >= 0 and cacheend <= l: |
|
1574 | 1574 | if cachestart == 0 and cacheend == l: |
|
1575 | 1575 | return d # avoid a copy |
|
1576 | 1576 | return util.buffer(d, cachestart, cacheend - cachestart) |
|
1577 | 1577 | |
|
1578 | 1578 | return self._readsegment(offset, length, df=df) |
|
1579 | 1579 | |
|
1580 | 1580 | def _getsegmentforrevs(self, startrev, endrev, df=None): |
|
1581 | 1581 | """Obtain a segment of raw data corresponding to a range of revisions. |
|
1582 | 1582 | |
|
1583 | 1583 | Accepts the start and end revisions and an optional already-open |
|
1584 | 1584 | file handle to be used for reading. If the file handle is read, its |
|
1585 | 1585 | seek position will not be preserved. |
|
1586 | 1586 | |
|
1587 | 1587 | Requests for data may be satisfied by a cache. |
|
1588 | 1588 | |
|
1589 | 1589 | Returns a 2-tuple of (offset, data) for the requested range of |
|
1590 | 1590 | revisions. Offset is the integer offset from the beginning of the |
|
1591 | 1591 | revlog and data is a str or buffer of the raw byte data. |
|
1592 | 1592 | |
|
1593 | 1593 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` |
|
1594 | 1594 | to determine where each revision's data begins and ends. |
|
1595 | 1595 | """ |
|
1596 | 1596 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons |
|
1597 | 1597 | # (functions are expensive). |
|
1598 | 1598 | index = self.index |
|
1599 | 1599 | istart = index[startrev] |
|
1600 | 1600 | start = int(istart[0] >> 16) |
|
1601 | 1601 | if startrev == endrev: |
|
1602 | 1602 | end = start + istart[1] |
|
1603 | 1603 | else: |
|
1604 | 1604 | iend = index[endrev] |
|
1605 | 1605 | end = int(iend[0] >> 16) + iend[1] |
|
1606 | 1606 | |
|
1607 | 1607 | if self._inline: |
|
1608 | 1608 | start += (startrev + 1) * self._io.size |
|
1609 | 1609 | end += (endrev + 1) * self._io.size |
|
1610 | 1610 | length = end - start |
|
1611 | 1611 | |
|
1612 | 1612 | return start, self._getsegment(start, length, df=df) |
|
1613 | 1613 | |
|
1614 | 1614 | def _chunk(self, rev, df=None): |
|
1615 | 1615 | """Obtain a single decompressed chunk for a revision. |
|
1616 | 1616 | |
|
1617 | 1617 | Accepts an integer revision and an optional already-open file handle |
|
1618 | 1618 | to be used for reading. If used, the seek position of the file will not |
|
1619 | 1619 | be preserved. |
|
1620 | 1620 | |
|
1621 | 1621 | Returns a str holding uncompressed data for the requested revision. |
|
1622 | 1622 | """ |
|
1623 | 1623 | return self.decompress(self._getsegmentforrevs(rev, rev, df=df)[1]) |
|
1624 | 1624 | |
|
1625 | 1625 | def _chunks(self, revs, df=None, targetsize=None): |
|
1626 | 1626 | """Obtain decompressed chunks for the specified revisions. |
|
1627 | 1627 | |
|
1628 | 1628 | Accepts an iterable of numeric revisions that are assumed to be in |
|
1629 | 1629 | ascending order. Also accepts an optional already-open file handle |
|
1630 | 1630 | to be used for reading. If used, the seek position of the file will |
|
1631 | 1631 | not be preserved. |
|
1632 | 1632 | |
|
1633 | 1633 | This function is similar to calling ``self._chunk()`` multiple times, |
|
1634 | 1634 | but is faster. |
|
1635 | 1635 | |
|
1636 | 1636 | Returns a list with decompressed data for each requested revision. |
|
1637 | 1637 | """ |
|
1638 | 1638 | if not revs: |
|
1639 | 1639 | return [] |
|
1640 | 1640 | start = self.start |
|
1641 | 1641 | length = self.length |
|
1642 | 1642 | inline = self._inline |
|
1643 | 1643 | iosize = self._io.size |
|
1644 | 1644 | buffer = util.buffer |
|
1645 | 1645 | |
|
1646 | 1646 | l = [] |
|
1647 | 1647 | ladd = l.append |
|
1648 | 1648 | |
|
1649 | 1649 | if not self._withsparseread: |
|
1650 | 1650 | slicedchunks = (revs,) |
|
1651 | 1651 | else: |
|
1652 | 1652 | slicedchunks = deltautil.slicechunk( |
|
1653 | 1653 | self, revs, targetsize=targetsize |
|
1654 | 1654 | ) |
|
1655 | 1655 | |
|
1656 | 1656 | for revschunk in slicedchunks: |
|
1657 | 1657 | firstrev = revschunk[0] |
|
1658 | 1658 | # Skip trailing revisions with empty diff |
|
1659 | 1659 | for lastrev in revschunk[::-1]: |
|
1660 | 1660 | if length(lastrev) != 0: |
|
1661 | 1661 | break |
|
1662 | 1662 | |
|
1663 | 1663 | try: |
|
1664 | 1664 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) |
|
1665 | 1665 | except OverflowError: |
|
1666 | 1666 | # issue4215 - we can't cache a run of chunks greater than |
|
1667 | 1667 | # 2G on Windows |
|
1668 | 1668 | return [self._chunk(rev, df=df) for rev in revschunk] |
|
1669 | 1669 | |
|
1670 | 1670 | decomp = self.decompress |
|
1671 | 1671 | for rev in revschunk: |
|
1672 | 1672 | chunkstart = start(rev) |
|
1673 | 1673 | if inline: |
|
1674 | 1674 | chunkstart += (rev + 1) * iosize |
|
1675 | 1675 | chunklength = length(rev) |
|
1676 | 1676 | ladd(decomp(buffer(data, chunkstart - offset, chunklength))) |
|
1677 | 1677 | |
|
1678 | 1678 | return l |
|
1679 | 1679 | |
|
1680 | 1680 | def _chunkclear(self): |
|
1681 | 1681 | """Clear the raw chunk cache.""" |
|
1682 | 1682 | self._chunkcache = (0, b'') |
|
1683 | 1683 | |
|
1684 | 1684 | def deltaparent(self, rev): |
|
1685 | 1685 | """return deltaparent of the given revision""" |
|
1686 | 1686 | base = self.index[rev][3] |
|
1687 | 1687 | if base == rev: |
|
1688 | 1688 | return nullrev |
|
1689 | 1689 | elif self._generaldelta: |
|
1690 | 1690 | return base |
|
1691 | 1691 | else: |
|
1692 | 1692 | return rev - 1 |
|
1693 | 1693 | |
|
1694 | 1694 | def issnapshot(self, rev): |
|
1695 | 1695 | """tells whether rev is a snapshot |
|
1696 | 1696 | """ |
|
1697 | 1697 | if not self._sparserevlog: |
|
1698 | 1698 | return self.deltaparent(rev) == nullrev |
|
1699 | 1699 | elif util.safehasattr(self.index, b'issnapshot'): |
|
1700 | 1700 | # directly assign the method to cache the testing and access |
|
1701 | 1701 | self.issnapshot = self.index.issnapshot |
|
1702 | 1702 | return self.issnapshot(rev) |
|
1703 | 1703 | if rev == nullrev: |
|
1704 | 1704 | return True |
|
1705 | 1705 | entry = self.index[rev] |
|
1706 | 1706 | base = entry[3] |
|
1707 | 1707 | if base == rev: |
|
1708 | 1708 | return True |
|
1709 | 1709 | if base == nullrev: |
|
1710 | 1710 | return True |
|
1711 | 1711 | p1 = entry[5] |
|
1712 | 1712 | p2 = entry[6] |
|
1713 | 1713 | if base == p1 or base == p2: |
|
1714 | 1714 | return False |
|
1715 | 1715 | return self.issnapshot(base) |
|
1716 | 1716 | |
|
1717 | 1717 | def snapshotdepth(self, rev): |
|
1718 | 1718 | """number of snapshot in the chain before this one""" |
|
1719 | 1719 | if not self.issnapshot(rev): |
|
1720 | 1720 | raise error.ProgrammingError(b'revision %d not a snapshot') |
|
1721 | 1721 | return len(self._deltachain(rev)[0]) - 1 |
|
1722 | 1722 | |
|
1723 | 1723 | def revdiff(self, rev1, rev2): |
|
1724 | 1724 | """return or calculate a delta between two revisions |
|
1725 | 1725 | |
|
1726 | 1726 | The delta calculated is in binary form and is intended to be written to |
|
1727 | 1727 | revlog data directly. So this function needs raw revision data. |
|
1728 | 1728 | """ |
|
1729 | 1729 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: |
|
1730 | 1730 | return bytes(self._chunk(rev2)) |
|
1731 | 1731 | |
|
1732 | 1732 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) |
|
1733 | 1733 | |
|
1734 | 1734 | def _processflags(self, text, flags, operation, raw=False): |
|
1735 | 1735 | """deprecated entry point to access flag processors""" |
|
1736 | 1736 | msg = b'_processflag(...) use the specialized variant' |
|
1737 | 1737 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1738 | 1738 | if raw: |
|
1739 | 1739 | return text, flagutil.processflagsraw(self, text, flags) |
|
1740 | 1740 | elif operation == b'read': |
|
1741 | 1741 | return flagutil.processflagsread(self, text, flags) |
|
1742 | 1742 | else: # write operation |
|
1743 | 1743 | return flagutil.processflagswrite(self, text, flags) |
|
1744 | 1744 | |
|
1745 | 1745 | def revision(self, nodeorrev, _df=None, raw=False): |
|
1746 | 1746 | """return an uncompressed revision of a given node or revision |
|
1747 | 1747 | number. |
|
1748 | 1748 | |
|
1749 | 1749 | _df - an existing file handle to read from. (internal-only) |
|
1750 | 1750 | raw - an optional argument specifying if the revision data is to be |
|
1751 | 1751 | treated as raw data when applying flag transforms. 'raw' should be set |
|
1752 | 1752 | to True when generating changegroups or in debug commands. |
|
1753 | 1753 | """ |
|
1754 | 1754 | if raw: |
|
1755 | 1755 | msg = ( |
|
1756 | 1756 | b'revlog.revision(..., raw=True) is deprecated, ' |
|
1757 | 1757 | b'use revlog.rawdata(...)' |
|
1758 | 1758 | ) |
|
1759 | 1759 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1760 | 1760 | return self._revisiondata(nodeorrev, _df, raw=raw)[0] |
|
1761 | 1761 | |
|
1762 | 1762 | def sidedata(self, nodeorrev, _df=None): |
|
1763 | 1763 | """a map of extra data related to the changeset but not part of the hash |
|
1764 | 1764 | |
|
1765 | 1765 | This function currently return a dictionary. However, more advanced |
|
1766 | 1766 | mapping object will likely be used in the future for a more |
|
1767 | 1767 | efficient/lazy code. |
|
1768 | 1768 | """ |
|
1769 | 1769 | return self._revisiondata(nodeorrev, _df)[1] |
|
1770 | 1770 | |
|
1771 | 1771 | def _revisiondata(self, nodeorrev, _df=None, raw=False): |
|
1772 | 1772 | # deal with <nodeorrev> argument type |
|
1773 | 1773 | if isinstance(nodeorrev, int): |
|
1774 | 1774 | rev = nodeorrev |
|
1775 | 1775 | node = self.node(rev) |
|
1776 | 1776 | else: |
|
1777 | 1777 | node = nodeorrev |
|
1778 | 1778 | rev = None |
|
1779 | 1779 | |
|
1780 | 1780 | # fast path the special `nullid` rev |
|
1781 | 1781 | if node == nullid: |
|
1782 | 1782 | return b"", {} |
|
1783 | 1783 | |
|
1784 | 1784 | # ``rawtext`` is the text as stored inside the revlog. Might be the |
|
1785 | 1785 | # revision or might need to be processed to retrieve the revision. |
|
1786 | 1786 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) |
|
1787 | 1787 | |
|
1788 | 1788 | if raw and validated: |
|
1789 | 1789 | # if we don't want to process the raw text and that raw |
|
1790 | 1790 | # text is cached, we can exit early. |
|
1791 | 1791 | return rawtext, {} |
|
1792 | 1792 | if rev is None: |
|
1793 | 1793 | rev = self.rev(node) |
|
1794 | 1794 | # the revlog's flag for this revision |
|
1795 | 1795 | # (usually alter its state or content) |
|
1796 | 1796 | flags = self.flags(rev) |
|
1797 | 1797 | |
|
1798 | 1798 | if validated and flags == REVIDX_DEFAULT_FLAGS: |
|
1799 | 1799 | # no extra flags set, no flag processor runs, text = rawtext |
|
1800 | 1800 | return rawtext, {} |
|
1801 | 1801 | |
|
1802 | 1802 | sidedata = {} |
|
1803 | 1803 | if raw: |
|
1804 | 1804 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
1805 | 1805 | text = rawtext |
|
1806 | 1806 | else: |
|
1807 | 1807 | try: |
|
1808 | 1808 | r = flagutil.processflagsread(self, rawtext, flags) |
|
1809 | 1809 | except error.SidedataHashError as exc: |
|
1810 | 1810 | msg = _(b"integrity check failed on %s:%s sidedata key %d") |
|
1811 | 1811 | msg %= (self.indexfile, pycompat.bytestr(rev), exc.sidedatakey) |
|
1812 | 1812 | raise error.RevlogError(msg) |
|
1813 | 1813 | text, validatehash, sidedata = r |
|
1814 | 1814 | if validatehash: |
|
1815 | 1815 | self.checkhash(text, node, rev=rev) |
|
1816 | 1816 | if not validated: |
|
1817 | 1817 | self._revisioncache = (node, rev, rawtext) |
|
1818 | 1818 | |
|
1819 | 1819 | return text, sidedata |
|
1820 | 1820 | |
|
1821 | 1821 | def _rawtext(self, node, rev, _df=None): |
|
1822 | 1822 | """return the possibly unvalidated rawtext for a revision |
|
1823 | 1823 | |
|
1824 | 1824 | returns (rev, rawtext, validated) |
|
1825 | 1825 | """ |
|
1826 | 1826 | |
|
1827 | 1827 | # revision in the cache (could be useful to apply delta) |
|
1828 | 1828 | cachedrev = None |
|
1829 | 1829 | # An intermediate text to apply deltas to |
|
1830 | 1830 | basetext = None |
|
1831 | 1831 | |
|
1832 | 1832 | # Check if we have the entry in cache |
|
1833 | 1833 | # The cache entry looks like (node, rev, rawtext) |
|
1834 | 1834 | if self._revisioncache: |
|
1835 | 1835 | if self._revisioncache[0] == node: |
|
1836 | 1836 | return (rev, self._revisioncache[2], True) |
|
1837 | 1837 | cachedrev = self._revisioncache[1] |
|
1838 | 1838 | |
|
1839 | 1839 | if rev is None: |
|
1840 | 1840 | rev = self.rev(node) |
|
1841 | 1841 | |
|
1842 | 1842 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) |
|
1843 | 1843 | if stopped: |
|
1844 | 1844 | basetext = self._revisioncache[2] |
|
1845 | 1845 | |
|
1846 | 1846 | # drop cache to save memory, the caller is expected to |
|
1847 | 1847 | # update self._revisioncache after validating the text |
|
1848 | 1848 | self._revisioncache = None |
|
1849 | 1849 | |
|
1850 | 1850 | targetsize = None |
|
1851 | 1851 | rawsize = self.index[rev][2] |
|
1852 | 1852 | if 0 <= rawsize: |
|
1853 | 1853 | targetsize = 4 * rawsize |
|
1854 | 1854 | |
|
1855 | 1855 | bins = self._chunks(chain, df=_df, targetsize=targetsize) |
|
1856 | 1856 | if basetext is None: |
|
1857 | 1857 | basetext = bytes(bins[0]) |
|
1858 | 1858 | bins = bins[1:] |
|
1859 | 1859 | |
|
1860 | 1860 | rawtext = mdiff.patches(basetext, bins) |
|
1861 | 1861 | del basetext # let us have a chance to free memory early |
|
1862 | 1862 | return (rev, rawtext, False) |
|
1863 | 1863 | |
|
1864 | 1864 | def rawdata(self, nodeorrev, _df=None): |
|
1865 | 1865 | """return an uncompressed raw data of a given node or revision number. |
|
1866 | 1866 | |
|
1867 | 1867 | _df - an existing file handle to read from. (internal-only) |
|
1868 | 1868 | """ |
|
1869 | 1869 | return self._revisiondata(nodeorrev, _df, raw=True)[0] |
|
1870 | 1870 | |
|
1871 | 1871 | def hash(self, text, p1, p2): |
|
1872 | 1872 | """Compute a node hash. |
|
1873 | 1873 | |
|
1874 | 1874 | Available as a function so that subclasses can replace the hash |
|
1875 | 1875 | as needed. |
|
1876 | 1876 | """ |
|
1877 | 1877 | return storageutil.hashrevisionsha1(text, p1, p2) |
|
1878 | 1878 | |
|
1879 | 1879 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
1880 | 1880 | """Check node hash integrity. |
|
1881 | 1881 | |
|
1882 | 1882 | Available as a function so that subclasses can extend hash mismatch |
|
1883 | 1883 | behaviors as needed. |
|
1884 | 1884 | """ |
|
1885 | 1885 | try: |
|
1886 | 1886 | if p1 is None and p2 is None: |
|
1887 | 1887 | p1, p2 = self.parents(node) |
|
1888 | 1888 | if node != self.hash(text, p1, p2): |
|
1889 | 1889 | # Clear the revision cache on hash failure. The revision cache |
|
1890 | 1890 | # only stores the raw revision and clearing the cache does have |
|
1891 | 1891 | # the side-effect that we won't have a cache hit when the raw |
|
1892 | 1892 | # revision data is accessed. But this case should be rare and |
|
1893 | 1893 | # it is extra work to teach the cache about the hash |
|
1894 | 1894 | # verification state. |
|
1895 | 1895 | if self._revisioncache and self._revisioncache[0] == node: |
|
1896 | 1896 | self._revisioncache = None |
|
1897 | 1897 | |
|
1898 | 1898 | revornode = rev |
|
1899 | 1899 | if revornode is None: |
|
1900 | 1900 | revornode = templatefilters.short(hex(node)) |
|
1901 | 1901 | raise error.RevlogError( |
|
1902 | 1902 | _(b"integrity check failed on %s:%s") |
|
1903 | 1903 | % (self.indexfile, pycompat.bytestr(revornode)) |
|
1904 | 1904 | ) |
|
1905 | 1905 | except error.RevlogError: |
|
1906 | 1906 | if self._censorable and storageutil.iscensoredtext(text): |
|
1907 | 1907 | raise error.CensoredNodeError(self.indexfile, node, text) |
|
1908 | 1908 | raise |
|
1909 | 1909 | |
|
1910 | 1910 | def _enforceinlinesize(self, tr, fp=None): |
|
1911 | 1911 | """Check if the revlog is too big for inline and convert if so. |
|
1912 | 1912 | |
|
1913 | 1913 | This should be called after revisions are added to the revlog. If the |
|
1914 | 1914 | revlog has grown too large to be an inline revlog, it will convert it |
|
1915 | 1915 | to use multiple index and data files. |
|
1916 | 1916 | """ |
|
1917 | 1917 | tiprev = len(self) - 1 |
|
1918 | 1918 | if ( |
|
1919 | 1919 | not self._inline |
|
1920 | 1920 | or (self.start(tiprev) + self.length(tiprev)) < _maxinline |
|
1921 | 1921 | ): |
|
1922 | 1922 | return |
|
1923 | 1923 | |
|
1924 | 1924 | trinfo = tr.find(self.indexfile) |
|
1925 | 1925 | if trinfo is None: |
|
1926 | 1926 | raise error.RevlogError( |
|
1927 | 1927 | _(b"%s not found in the transaction") % self.indexfile |
|
1928 | 1928 | ) |
|
1929 | 1929 | |
|
1930 | 1930 | trindex = trinfo[2] |
|
1931 | 1931 | if trindex is not None: |
|
1932 | 1932 | dataoff = self.start(trindex) |
|
1933 | 1933 | else: |
|
1934 | 1934 | # revlog was stripped at start of transaction, use all leftover data |
|
1935 | 1935 | trindex = len(self) - 1 |
|
1936 | 1936 | dataoff = self.end(tiprev) |
|
1937 | 1937 | |
|
1938 | 1938 | tr.add(self.datafile, dataoff) |
|
1939 | 1939 | |
|
1940 | 1940 | if fp: |
|
1941 | 1941 | fp.flush() |
|
1942 | 1942 | fp.close() |
|
1943 | 1943 | # We can't use the cached file handle after close(). So prevent |
|
1944 | 1944 | # its usage. |
|
1945 | 1945 | self._writinghandles = None |
|
1946 | 1946 | |
|
1947 | 1947 | with self._indexfp(b'r') as ifh, self._datafp(b'w') as dfh: |
|
1948 | 1948 | for r in self: |
|
1949 | 1949 | dfh.write(self._getsegmentforrevs(r, r, df=ifh)[1]) |
|
1950 | 1950 | |
|
1951 | 1951 | with self._indexfp(b'w') as fp: |
|
1952 | 1952 | self.version &= ~FLAG_INLINE_DATA |
|
1953 | 1953 | self._inline = False |
|
1954 | 1954 | io = self._io |
|
1955 | 1955 | for i in self: |
|
1956 | 1956 | e = io.packentry(self.index[i], self.node, self.version, i) |
|
1957 | 1957 | fp.write(e) |
|
1958 | 1958 | |
|
1959 | 1959 | # the temp file replace the real index when we exit the context |
|
1960 | 1960 | # manager |
|
1961 | 1961 | |
|
1962 | 1962 | tr.replace(self.indexfile, trindex * self._io.size) |
|
1963 | 1963 | self._chunkclear() |
|
1964 | 1964 | |
|
1965 | 1965 | def _nodeduplicatecallback(self, transaction, node): |
|
1966 | 1966 | """called when trying to add a node already stored. |
|
1967 | 1967 | """ |
|
1968 | 1968 | |
|
1969 | 1969 | def addrevision( |
|
1970 | 1970 | self, |
|
1971 | 1971 | text, |
|
1972 | 1972 | transaction, |
|
1973 | 1973 | link, |
|
1974 | 1974 | p1, |
|
1975 | 1975 | p2, |
|
1976 | 1976 | cachedelta=None, |
|
1977 | 1977 | node=None, |
|
1978 | 1978 | flags=REVIDX_DEFAULT_FLAGS, |
|
1979 | 1979 | deltacomputer=None, |
|
1980 | 1980 | sidedata=None, |
|
1981 | 1981 | ): |
|
1982 | 1982 | """add a revision to the log |
|
1983 | 1983 | |
|
1984 | 1984 | text - the revision data to add |
|
1985 | 1985 | transaction - the transaction object used for rollback |
|
1986 | 1986 | link - the linkrev data to add |
|
1987 | 1987 | p1, p2 - the parent nodeids of the revision |
|
1988 | 1988 | cachedelta - an optional precomputed delta |
|
1989 | 1989 | node - nodeid of revision; typically node is not specified, and it is |
|
1990 | 1990 | computed by default as hash(text, p1, p2), however subclasses might |
|
1991 | 1991 | use different hashing method (and override checkhash() in such case) |
|
1992 | 1992 | flags - the known flags to set on the revision |
|
1993 | 1993 | deltacomputer - an optional deltacomputer instance shared between |
|
1994 | 1994 | multiple calls |
|
1995 | 1995 | """ |
|
1996 | 1996 | if link == nullrev: |
|
1997 | 1997 | raise error.RevlogError( |
|
1998 | 1998 | _(b"attempted to add linkrev -1 to %s") % self.indexfile |
|
1999 | 1999 | ) |
|
2000 | 2000 | |
|
2001 | 2001 | if sidedata is None: |
|
2002 | 2002 | sidedata = {} |
|
2003 | 2003 | flags = flags & ~REVIDX_SIDEDATA |
|
2004 | 2004 | elif not self.hassidedata: |
|
2005 | 2005 | raise error.ProgrammingError( |
|
2006 | 2006 | _(b"trying to add sidedata to a revlog who don't support them") |
|
2007 | 2007 | ) |
|
2008 | 2008 | else: |
|
2009 | 2009 | flags |= REVIDX_SIDEDATA |
|
2010 | 2010 | |
|
2011 | 2011 | if flags: |
|
2012 | 2012 | node = node or self.hash(text, p1, p2) |
|
2013 | 2013 | |
|
2014 | 2014 | rawtext, validatehash = flagutil.processflagswrite( |
|
2015 | 2015 | self, text, flags, sidedata=sidedata |
|
2016 | 2016 | ) |
|
2017 | 2017 | |
|
2018 | 2018 | # If the flag processor modifies the revision data, ignore any provided |
|
2019 | 2019 | # cachedelta. |
|
2020 | 2020 | if rawtext != text: |
|
2021 | 2021 | cachedelta = None |
|
2022 | 2022 | |
|
2023 | 2023 | if len(rawtext) > _maxentrysize: |
|
2024 | 2024 | raise error.RevlogError( |
|
2025 | 2025 | _( |
|
2026 | 2026 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" |
|
2027 | 2027 | ) |
|
2028 | 2028 | % (self.indexfile, len(rawtext)) |
|
2029 | 2029 | ) |
|
2030 | 2030 | |
|
2031 | 2031 | node = node or self.hash(rawtext, p1, p2) |
|
2032 | 2032 | if self.index.has_node(node): |
|
2033 | 2033 | return node |
|
2034 | 2034 | |
|
2035 | 2035 | if validatehash: |
|
2036 | 2036 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
2037 | 2037 | |
|
2038 | 2038 | return self.addrawrevision( |
|
2039 | 2039 | rawtext, |
|
2040 | 2040 | transaction, |
|
2041 | 2041 | link, |
|
2042 | 2042 | p1, |
|
2043 | 2043 | p2, |
|
2044 | 2044 | node, |
|
2045 | 2045 | flags, |
|
2046 | 2046 | cachedelta=cachedelta, |
|
2047 | 2047 | deltacomputer=deltacomputer, |
|
2048 | 2048 | ) |
|
2049 | 2049 | |
|
2050 | 2050 | def addrawrevision( |
|
2051 | 2051 | self, |
|
2052 | 2052 | rawtext, |
|
2053 | 2053 | transaction, |
|
2054 | 2054 | link, |
|
2055 | 2055 | p1, |
|
2056 | 2056 | p2, |
|
2057 | 2057 | node, |
|
2058 | 2058 | flags, |
|
2059 | 2059 | cachedelta=None, |
|
2060 | 2060 | deltacomputer=None, |
|
2061 | 2061 | ): |
|
2062 | 2062 | """add a raw revision with known flags, node and parents |
|
2063 | 2063 | useful when reusing a revision not stored in this revlog (ex: received |
|
2064 | 2064 | over wire, or read from an external bundle). |
|
2065 | 2065 | """ |
|
2066 | 2066 | dfh = None |
|
2067 | 2067 | if not self._inline: |
|
2068 | 2068 | dfh = self._datafp(b"a+") |
|
2069 | 2069 | ifh = self._indexfp(b"a+") |
|
2070 | 2070 | try: |
|
2071 | 2071 | return self._addrevision( |
|
2072 | 2072 | node, |
|
2073 | 2073 | rawtext, |
|
2074 | 2074 | transaction, |
|
2075 | 2075 | link, |
|
2076 | 2076 | p1, |
|
2077 | 2077 | p2, |
|
2078 | 2078 | flags, |
|
2079 | 2079 | cachedelta, |
|
2080 | 2080 | ifh, |
|
2081 | 2081 | dfh, |
|
2082 | 2082 | deltacomputer=deltacomputer, |
|
2083 | 2083 | ) |
|
2084 | 2084 | finally: |
|
2085 | 2085 | if dfh: |
|
2086 | 2086 | dfh.close() |
|
2087 | 2087 | ifh.close() |
|
2088 | 2088 | |
|
2089 | 2089 | def compress(self, data): |
|
2090 | 2090 | """Generate a possibly-compressed representation of data.""" |
|
2091 | 2091 | if not data: |
|
2092 | 2092 | return b'', data |
|
2093 | 2093 | |
|
2094 | 2094 | compressed = self._compressor.compress(data) |
|
2095 | 2095 | |
|
2096 | 2096 | if compressed: |
|
2097 | 2097 | # The revlog compressor added the header in the returned data. |
|
2098 | 2098 | return b'', compressed |
|
2099 | 2099 | |
|
2100 | 2100 | if data[0:1] == b'\0': |
|
2101 | 2101 | return b'', data |
|
2102 | 2102 | return b'u', data |
|
2103 | 2103 | |
|
2104 | 2104 | def decompress(self, data): |
|
2105 | 2105 | """Decompress a revlog chunk. |
|
2106 | 2106 | |
|
2107 | 2107 | The chunk is expected to begin with a header identifying the |
|
2108 | 2108 | format type so it can be routed to an appropriate decompressor. |
|
2109 | 2109 | """ |
|
2110 | 2110 | if not data: |
|
2111 | 2111 | return data |
|
2112 | 2112 | |
|
2113 | 2113 | # Revlogs are read much more frequently than they are written and many |
|
2114 | 2114 | # chunks only take microseconds to decompress, so performance is |
|
2115 | 2115 | # important here. |
|
2116 | 2116 | # |
|
2117 | 2117 | # We can make a few assumptions about revlogs: |
|
2118 | 2118 | # |
|
2119 | 2119 | # 1) the majority of chunks will be compressed (as opposed to inline |
|
2120 | 2120 | # raw data). |
|
2121 | 2121 | # 2) decompressing *any* data will likely by at least 10x slower than |
|
2122 | 2122 | # returning raw inline data. |
|
2123 | 2123 | # 3) we want to prioritize common and officially supported compression |
|
2124 | 2124 | # engines |
|
2125 | 2125 | # |
|
2126 | 2126 | # It follows that we want to optimize for "decompress compressed data |
|
2127 | 2127 | # when encoded with common and officially supported compression engines" |
|
2128 | 2128 | # case over "raw data" and "data encoded by less common or non-official |
|
2129 | 2129 | # compression engines." That is why we have the inline lookup first |
|
2130 | 2130 | # followed by the compengines lookup. |
|
2131 | 2131 | # |
|
2132 | 2132 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib |
|
2133 | 2133 | # compressed chunks. And this matters for changelog and manifest reads. |
|
2134 | 2134 | t = data[0:1] |
|
2135 | 2135 | |
|
2136 | 2136 | if t == b'x': |
|
2137 | 2137 | try: |
|
2138 | 2138 | return _zlibdecompress(data) |
|
2139 | 2139 | except zlib.error as e: |
|
2140 | 2140 | raise error.RevlogError( |
|
2141 | 2141 | _(b'revlog decompress error: %s') |
|
2142 | 2142 | % stringutil.forcebytestr(e) |
|
2143 | 2143 | ) |
|
2144 | 2144 | # '\0' is more common than 'u' so it goes first. |
|
2145 | 2145 | elif t == b'\0': |
|
2146 | 2146 | return data |
|
2147 | 2147 | elif t == b'u': |
|
2148 | 2148 | return util.buffer(data, 1) |
|
2149 | 2149 | |
|
2150 | 2150 | try: |
|
2151 | 2151 | compressor = self._decompressors[t] |
|
2152 | 2152 | except KeyError: |
|
2153 | 2153 | try: |
|
2154 | 2154 | engine = util.compengines.forrevlogheader(t) |
|
2155 | 2155 | compressor = engine.revlogcompressor(self._compengineopts) |
|
2156 | 2156 | self._decompressors[t] = compressor |
|
2157 | 2157 | except KeyError: |
|
2158 | 2158 | raise error.RevlogError(_(b'unknown compression type %r') % t) |
|
2159 | 2159 | |
|
2160 | 2160 | return compressor.decompress(data) |
|
2161 | 2161 | |
|
2162 | 2162 | def _addrevision( |
|
2163 | 2163 | self, |
|
2164 | 2164 | node, |
|
2165 | 2165 | rawtext, |
|
2166 | 2166 | transaction, |
|
2167 | 2167 | link, |
|
2168 | 2168 | p1, |
|
2169 | 2169 | p2, |
|
2170 | 2170 | flags, |
|
2171 | 2171 | cachedelta, |
|
2172 | 2172 | ifh, |
|
2173 | 2173 | dfh, |
|
2174 | 2174 | alwayscache=False, |
|
2175 | 2175 | deltacomputer=None, |
|
2176 | 2176 | ): |
|
2177 | 2177 | """internal function to add revisions to the log |
|
2178 | 2178 | |
|
2179 | 2179 | see addrevision for argument descriptions. |
|
2180 | 2180 | |
|
2181 | 2181 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. |
|
2182 | 2182 | |
|
2183 | 2183 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will |
|
2184 | 2184 | be used. |
|
2185 | 2185 | |
|
2186 | 2186 | invariants: |
|
2187 | 2187 | - rawtext is optional (can be None); if not set, cachedelta must be set. |
|
2188 | 2188 | if both are set, they must correspond to each other. |
|
2189 | 2189 | """ |
|
2190 | 2190 | if node == nullid: |
|
2191 | 2191 | raise error.RevlogError( |
|
2192 | 2192 | _(b"%s: attempt to add null revision") % self.indexfile |
|
2193 | 2193 | ) |
|
2194 | 2194 | if node == wdirid or node in wdirfilenodeids: |
|
2195 | 2195 | raise error.RevlogError( |
|
2196 | 2196 | _(b"%s: attempt to add wdir revision") % self.indexfile |
|
2197 | 2197 | ) |
|
2198 | 2198 | |
|
2199 | 2199 | if self._inline: |
|
2200 | 2200 | fh = ifh |
|
2201 | 2201 | else: |
|
2202 | 2202 | fh = dfh |
|
2203 | 2203 | |
|
2204 | 2204 | btext = [rawtext] |
|
2205 | 2205 | |
|
2206 | 2206 | curr = len(self) |
|
2207 | 2207 | prev = curr - 1 |
|
2208 | 2208 | offset = self.end(prev) |
|
2209 | 2209 | p1r, p2r = self.rev(p1), self.rev(p2) |
|
2210 | 2210 | |
|
2211 | 2211 | # full versions are inserted when the needed deltas |
|
2212 | 2212 | # become comparable to the uncompressed text |
|
2213 | 2213 | if rawtext is None: |
|
2214 | 2214 | # need rawtext size, before changed by flag processors, which is |
|
2215 | 2215 | # the non-raw size. use revlog explicitly to avoid filelog's extra |
|
2216 | 2216 | # logic that might remove metadata size. |
|
2217 | 2217 | textlen = mdiff.patchedsize( |
|
2218 | 2218 | revlog.size(self, cachedelta[0]), cachedelta[1] |
|
2219 | 2219 | ) |
|
2220 | 2220 | else: |
|
2221 | 2221 | textlen = len(rawtext) |
|
2222 | 2222 | |
|
2223 | 2223 | if deltacomputer is None: |
|
2224 | 2224 | deltacomputer = deltautil.deltacomputer(self) |
|
2225 | 2225 | |
|
2226 | 2226 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) |
|
2227 | 2227 | |
|
2228 | 2228 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) |
|
2229 | 2229 | |
|
2230 | 2230 | e = ( |
|
2231 | 2231 | offset_type(offset, flags), |
|
2232 | 2232 | deltainfo.deltalen, |
|
2233 | 2233 | textlen, |
|
2234 | 2234 | deltainfo.base, |
|
2235 | 2235 | link, |
|
2236 | 2236 | p1r, |
|
2237 | 2237 | p2r, |
|
2238 | 2238 | node, |
|
2239 | 2239 | ) |
|
2240 | 2240 | self.index.append(e) |
|
2241 | 2241 | |
|
2242 | 2242 | entry = self._io.packentry(e, self.node, self.version, curr) |
|
2243 | 2243 | self._writeentry( |
|
2244 | 2244 | transaction, ifh, dfh, entry, deltainfo.data, link, offset |
|
2245 | 2245 | ) |
|
2246 | 2246 | |
|
2247 | 2247 | rawtext = btext[0] |
|
2248 | 2248 | |
|
2249 | 2249 | if alwayscache and rawtext is None: |
|
2250 | 2250 | rawtext = deltacomputer.buildtext(revinfo, fh) |
|
2251 | 2251 | |
|
2252 | 2252 | if type(rawtext) == bytes: # only accept immutable objects |
|
2253 | 2253 | self._revisioncache = (node, curr, rawtext) |
|
2254 | 2254 | self._chainbasecache[curr] = deltainfo.chainbase |
|
2255 | 2255 | return node |
|
2256 | 2256 | |
|
2257 | 2257 | def _writeentry(self, transaction, ifh, dfh, entry, data, link, offset): |
|
2258 | 2258 | # Files opened in a+ mode have inconsistent behavior on various |
|
2259 | 2259 | # platforms. Windows requires that a file positioning call be made |
|
2260 | 2260 | # when the file handle transitions between reads and writes. See |
|
2261 | 2261 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
2262 | 2262 | # platforms, Python or the platform itself can be buggy. Some versions |
|
2263 | 2263 | # of Solaris have been observed to not append at the end of the file |
|
2264 | 2264 | # if the file was seeked to before the end. See issue4943 for more. |
|
2265 | 2265 | # |
|
2266 | 2266 | # We work around this issue by inserting a seek() before writing. |
|
2267 | 2267 | # Note: This is likely not necessary on Python 3. However, because |
|
2268 | 2268 | # the file handle is reused for reads and may be seeked there, we need |
|
2269 | 2269 | # to be careful before changing this. |
|
2270 | 2270 | ifh.seek(0, os.SEEK_END) |
|
2271 | 2271 | if dfh: |
|
2272 | 2272 | dfh.seek(0, os.SEEK_END) |
|
2273 | 2273 | |
|
2274 | 2274 | curr = len(self) - 1 |
|
2275 | 2275 | if not self._inline: |
|
2276 | 2276 | transaction.add(self.datafile, offset) |
|
2277 | 2277 | transaction.add(self.indexfile, curr * len(entry)) |
|
2278 | 2278 | if data[0]: |
|
2279 | 2279 | dfh.write(data[0]) |
|
2280 | 2280 | dfh.write(data[1]) |
|
2281 | 2281 | ifh.write(entry) |
|
2282 | 2282 | else: |
|
2283 | 2283 | offset += curr * self._io.size |
|
2284 | 2284 | transaction.add(self.indexfile, offset, curr) |
|
2285 | 2285 | ifh.write(entry) |
|
2286 | 2286 | ifh.write(data[0]) |
|
2287 | 2287 | ifh.write(data[1]) |
|
2288 | 2288 | self._enforceinlinesize(transaction, ifh) |
|
2289 | 2289 | |
|
2290 | 2290 | def addgroup(self, deltas, linkmapper, transaction, addrevisioncb=None): |
|
2291 | 2291 | """ |
|
2292 | 2292 | add a delta group |
|
2293 | 2293 | |
|
2294 | 2294 | given a set of deltas, add them to the revision log. the |
|
2295 | 2295 | first delta is against its parent, which should be in our |
|
2296 | 2296 | log, the rest are against the previous delta. |
|
2297 | 2297 | |
|
2298 | 2298 | If ``addrevisioncb`` is defined, it will be called with arguments of |
|
2299 | 2299 | this revlog and the node that was added. |
|
2300 | 2300 | """ |
|
2301 | 2301 | |
|
2302 | 2302 | if self._writinghandles: |
|
2303 | 2303 | raise error.ProgrammingError(b'cannot nest addgroup() calls') |
|
2304 | 2304 | |
|
2305 | 2305 | nodes = [] |
|
2306 | 2306 | |
|
2307 | 2307 | r = len(self) |
|
2308 | 2308 | end = 0 |
|
2309 | 2309 | if r: |
|
2310 | 2310 | end = self.end(r - 1) |
|
2311 | 2311 | ifh = self._indexfp(b"a+") |
|
2312 | 2312 | isize = r * self._io.size |
|
2313 | 2313 | if self._inline: |
|
2314 | 2314 | transaction.add(self.indexfile, end + isize, r) |
|
2315 | 2315 | dfh = None |
|
2316 | 2316 | else: |
|
2317 | 2317 | transaction.add(self.indexfile, isize, r) |
|
2318 | 2318 | transaction.add(self.datafile, end) |
|
2319 | 2319 | dfh = self._datafp(b"a+") |
|
2320 | 2320 | |
|
2321 | 2321 | def flush(): |
|
2322 | 2322 | if dfh: |
|
2323 | 2323 | dfh.flush() |
|
2324 | 2324 | ifh.flush() |
|
2325 | 2325 | |
|
2326 | 2326 | self._writinghandles = (ifh, dfh) |
|
2327 | 2327 | |
|
2328 | 2328 | try: |
|
2329 | 2329 | deltacomputer = deltautil.deltacomputer(self) |
|
2330 | 2330 | # loop through our set of deltas |
|
2331 | 2331 | for data in deltas: |
|
2332 | 2332 | node, p1, p2, linknode, deltabase, delta, flags = data |
|
2333 | 2333 | link = linkmapper(linknode) |
|
2334 | 2334 | flags = flags or REVIDX_DEFAULT_FLAGS |
|
2335 | 2335 | |
|
2336 | 2336 | nodes.append(node) |
|
2337 | 2337 | |
|
2338 | 2338 | if self.index.has_node(node): |
|
2339 | 2339 | self._nodeduplicatecallback(transaction, node) |
|
2340 | 2340 | # this can happen if two branches make the same change |
|
2341 | 2341 | continue |
|
2342 | 2342 | |
|
2343 | 2343 | for p in (p1, p2): |
|
2344 | 2344 | if not self.index.has_node(p): |
|
2345 | 2345 | raise error.LookupError( |
|
2346 | 2346 | p, self.indexfile, _(b'unknown parent') |
|
2347 | 2347 | ) |
|
2348 | 2348 | |
|
2349 | 2349 | if not self.index.has_node(deltabase): |
|
2350 | 2350 | raise error.LookupError( |
|
2351 | 2351 | deltabase, self.indexfile, _(b'unknown delta base') |
|
2352 | 2352 | ) |
|
2353 | 2353 | |
|
2354 | 2354 | baserev = self.rev(deltabase) |
|
2355 | 2355 | |
|
2356 | 2356 | if baserev != nullrev and self.iscensored(baserev): |
|
2357 | 2357 | # if base is censored, delta must be full replacement in a |
|
2358 | 2358 | # single patch operation |
|
2359 | 2359 | hlen = struct.calcsize(b">lll") |
|
2360 | 2360 | oldlen = self.rawsize(baserev) |
|
2361 | 2361 | newlen = len(delta) - hlen |
|
2362 | 2362 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): |
|
2363 | 2363 | raise error.CensoredBaseError( |
|
2364 | 2364 | self.indexfile, self.node(baserev) |
|
2365 | 2365 | ) |
|
2366 | 2366 | |
|
2367 | 2367 | if not flags and self._peek_iscensored(baserev, delta, flush): |
|
2368 | 2368 | flags |= REVIDX_ISCENSORED |
|
2369 | 2369 | |
|
2370 | 2370 | # We assume consumers of addrevisioncb will want to retrieve |
|
2371 | 2371 | # the added revision, which will require a call to |
|
2372 | 2372 | # revision(). revision() will fast path if there is a cache |
|
2373 | 2373 | # hit. So, we tell _addrevision() to always cache in this case. |
|
2374 | 2374 | # We're only using addgroup() in the context of changegroup |
|
2375 | 2375 | # generation so the revision data can always be handled as raw |
|
2376 | 2376 | # by the flagprocessor. |
|
2377 | 2377 | self._addrevision( |
|
2378 | 2378 | node, |
|
2379 | 2379 | None, |
|
2380 | 2380 | transaction, |
|
2381 | 2381 | link, |
|
2382 | 2382 | p1, |
|
2383 | 2383 | p2, |
|
2384 | 2384 | flags, |
|
2385 | 2385 | (baserev, delta), |
|
2386 | 2386 | ifh, |
|
2387 | 2387 | dfh, |
|
2388 | 2388 | alwayscache=bool(addrevisioncb), |
|
2389 | 2389 | deltacomputer=deltacomputer, |
|
2390 | 2390 | ) |
|
2391 | 2391 | |
|
2392 | 2392 | if addrevisioncb: |
|
2393 | 2393 | addrevisioncb(self, node) |
|
2394 | 2394 | |
|
2395 | 2395 | if not dfh and not self._inline: |
|
2396 | 2396 | # addrevision switched from inline to conventional |
|
2397 | 2397 | # reopen the index |
|
2398 | 2398 | ifh.close() |
|
2399 | 2399 | dfh = self._datafp(b"a+") |
|
2400 | 2400 | ifh = self._indexfp(b"a+") |
|
2401 | 2401 | self._writinghandles = (ifh, dfh) |
|
2402 | 2402 | finally: |
|
2403 | 2403 | self._writinghandles = None |
|
2404 | 2404 | |
|
2405 | 2405 | if dfh: |
|
2406 | 2406 | dfh.close() |
|
2407 | 2407 | ifh.close() |
|
2408 | 2408 | |
|
2409 | 2409 | return nodes |
|
2410 | 2410 | |
|
2411 | 2411 | def iscensored(self, rev): |
|
2412 | 2412 | """Check if a file revision is censored.""" |
|
2413 | 2413 | if not self._censorable: |
|
2414 | 2414 | return False |
|
2415 | 2415 | |
|
2416 | 2416 | return self.flags(rev) & REVIDX_ISCENSORED |
|
2417 | 2417 | |
|
2418 | 2418 | def _peek_iscensored(self, baserev, delta, flush): |
|
2419 | 2419 | """Quickly check if a delta produces a censored revision.""" |
|
2420 | 2420 | if not self._censorable: |
|
2421 | 2421 | return False |
|
2422 | 2422 | |
|
2423 | 2423 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) |
|
2424 | 2424 | |
|
2425 | 2425 | def getstrippoint(self, minlink): |
|
2426 | 2426 | """find the minimum rev that must be stripped to strip the linkrev |
|
2427 | 2427 | |
|
2428 | 2428 | Returns a tuple containing the minimum rev and a set of all revs that |
|
2429 | 2429 | have linkrevs that will be broken by this strip. |
|
2430 | 2430 | """ |
|
2431 | 2431 | return storageutil.resolvestripinfo( |
|
2432 | 2432 | minlink, |
|
2433 | 2433 | len(self) - 1, |
|
2434 | 2434 | self.headrevs(), |
|
2435 | 2435 | self.linkrev, |
|
2436 | 2436 | self.parentrevs, |
|
2437 | 2437 | ) |
|
2438 | 2438 | |
|
2439 | 2439 | def strip(self, minlink, transaction): |
|
2440 | 2440 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
2441 | 2441 | |
|
2442 | 2442 | This function is called when we're stripping revision minlink and |
|
2443 | 2443 | its descendants from the repository. |
|
2444 | 2444 | |
|
2445 | 2445 | We have to remove all revisions with linkrev >= minlink, because |
|
2446 | 2446 | the equivalent changelog revisions will be renumbered after the |
|
2447 | 2447 | strip. |
|
2448 | 2448 | |
|
2449 | 2449 | So we truncate the revlog on the first of these revisions, and |
|
2450 | 2450 | trust that the caller has saved the revisions that shouldn't be |
|
2451 | 2451 | removed and that it'll re-add them after this truncation. |
|
2452 | 2452 | """ |
|
2453 | 2453 | if len(self) == 0: |
|
2454 | 2454 | return |
|
2455 | 2455 | |
|
2456 | 2456 | rev, _ = self.getstrippoint(minlink) |
|
2457 | 2457 | if rev == len(self): |
|
2458 | 2458 | return |
|
2459 | 2459 | |
|
2460 | 2460 | # first truncate the files on disk |
|
2461 | 2461 | end = self.start(rev) |
|
2462 | 2462 | if not self._inline: |
|
2463 | 2463 | transaction.add(self.datafile, end) |
|
2464 | 2464 | end = rev * self._io.size |
|
2465 | 2465 | else: |
|
2466 | 2466 | end += rev * self._io.size |
|
2467 | 2467 | |
|
2468 | 2468 | transaction.add(self.indexfile, end) |
|
2469 | 2469 | |
|
2470 | 2470 | # then reset internal state in memory to forget those revisions |
|
2471 | 2471 | self._revisioncache = None |
|
2472 | 2472 | self._chaininfocache = {} |
|
2473 | 2473 | self._chunkclear() |
|
2474 | 2474 | |
|
2475 | 2475 | del self.index[rev:-1] |
|
2476 | 2476 | |
|
2477 | 2477 | def checksize(self): |
|
2478 | 2478 | """Check size of index and data files |
|
2479 | 2479 | |
|
2480 | 2480 | return a (dd, di) tuple. |
|
2481 | 2481 | - dd: extra bytes for the "data" file |
|
2482 | 2482 | - di: extra bytes for the "index" file |
|
2483 | 2483 | |
|
2484 | 2484 | A healthy revlog will return (0, 0). |
|
2485 | 2485 | """ |
|
2486 | 2486 | expected = 0 |
|
2487 | 2487 | if len(self): |
|
2488 | 2488 | expected = max(0, self.end(len(self) - 1)) |
|
2489 | 2489 | |
|
2490 | 2490 | try: |
|
2491 | 2491 | with self._datafp() as f: |
|
2492 | 2492 | f.seek(0, io.SEEK_END) |
|
2493 | 2493 | actual = f.tell() |
|
2494 | 2494 | dd = actual - expected |
|
2495 | 2495 | except IOError as inst: |
|
2496 | 2496 | if inst.errno != errno.ENOENT: |
|
2497 | 2497 | raise |
|
2498 | 2498 | dd = 0 |
|
2499 | 2499 | |
|
2500 | 2500 | try: |
|
2501 | 2501 | f = self.opener(self.indexfile) |
|
2502 | 2502 | f.seek(0, io.SEEK_END) |
|
2503 | 2503 | actual = f.tell() |
|
2504 | 2504 | f.close() |
|
2505 | 2505 | s = self._io.size |
|
2506 | 2506 | i = max(0, actual // s) |
|
2507 | 2507 | di = actual - (i * s) |
|
2508 | 2508 | if self._inline: |
|
2509 | 2509 | databytes = 0 |
|
2510 | 2510 | for r in self: |
|
2511 | 2511 | databytes += max(0, self.length(r)) |
|
2512 | 2512 | dd = 0 |
|
2513 | 2513 | di = actual - len(self) * s - databytes |
|
2514 | 2514 | except IOError as inst: |
|
2515 | 2515 | if inst.errno != errno.ENOENT: |
|
2516 | 2516 | raise |
|
2517 | 2517 | di = 0 |
|
2518 | 2518 | |
|
2519 | 2519 | return (dd, di) |
|
2520 | 2520 | |
|
2521 | 2521 | def files(self): |
|
2522 | 2522 | res = [self.indexfile] |
|
2523 | 2523 | if not self._inline: |
|
2524 | 2524 | res.append(self.datafile) |
|
2525 | 2525 | return res |
|
2526 | 2526 | |
|
2527 | 2527 | def emitrevisions( |
|
2528 | 2528 | self, |
|
2529 | 2529 | nodes, |
|
2530 | 2530 | nodesorder=None, |
|
2531 | 2531 | revisiondata=False, |
|
2532 | 2532 | assumehaveparentrevisions=False, |
|
2533 | 2533 | deltamode=repository.CG_DELTAMODE_STD, |
|
2534 | 2534 | ): |
|
2535 | 2535 | if nodesorder not in (b'nodes', b'storage', b'linear', None): |
|
2536 | 2536 | raise error.ProgrammingError( |
|
2537 | 2537 | b'unhandled value for nodesorder: %s' % nodesorder |
|
2538 | 2538 | ) |
|
2539 | 2539 | |
|
2540 | 2540 | if nodesorder is None and not self._generaldelta: |
|
2541 | 2541 | nodesorder = b'storage' |
|
2542 | 2542 | |
|
2543 | 2543 | if ( |
|
2544 | 2544 | not self._storedeltachains |
|
2545 | 2545 | and deltamode != repository.CG_DELTAMODE_PREV |
|
2546 | 2546 | ): |
|
2547 | 2547 | deltamode = repository.CG_DELTAMODE_FULL |
|
2548 | 2548 | |
|
2549 | 2549 | return storageutil.emitrevisions( |
|
2550 | 2550 | self, |
|
2551 | 2551 | nodes, |
|
2552 | 2552 | nodesorder, |
|
2553 | 2553 | revlogrevisiondelta, |
|
2554 | 2554 | deltaparentfn=self.deltaparent, |
|
2555 | 2555 | candeltafn=self.candelta, |
|
2556 | 2556 | rawsizefn=self.rawsize, |
|
2557 | 2557 | revdifffn=self.revdiff, |
|
2558 | 2558 | flagsfn=self.flags, |
|
2559 | 2559 | deltamode=deltamode, |
|
2560 | 2560 | revisiondata=revisiondata, |
|
2561 | 2561 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
2562 | 2562 | ) |
|
2563 | 2563 | |
|
2564 | 2564 | DELTAREUSEALWAYS = b'always' |
|
2565 | 2565 | DELTAREUSESAMEREVS = b'samerevs' |
|
2566 | 2566 | DELTAREUSENEVER = b'never' |
|
2567 | 2567 | |
|
2568 | 2568 | DELTAREUSEFULLADD = b'fulladd' |
|
2569 | 2569 | |
|
2570 | 2570 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} |
|
2571 | 2571 | |
|
2572 | 2572 | def clone( |
|
2573 | 2573 | self, |
|
2574 | 2574 | tr, |
|
2575 | 2575 | destrevlog, |
|
2576 | 2576 | addrevisioncb=None, |
|
2577 | 2577 | deltareuse=DELTAREUSESAMEREVS, |
|
2578 | 2578 | forcedeltabothparents=None, |
|
2579 | 2579 | sidedatacompanion=None, |
|
2580 | 2580 | ): |
|
2581 | 2581 | """Copy this revlog to another, possibly with format changes. |
|
2582 | 2582 | |
|
2583 | 2583 | The destination revlog will contain the same revisions and nodes. |
|
2584 | 2584 | However, it may not be bit-for-bit identical due to e.g. delta encoding |
|
2585 | 2585 | differences. |
|
2586 | 2586 | |
|
2587 | 2587 | The ``deltareuse`` argument control how deltas from the existing revlog |
|
2588 | 2588 | are preserved in the destination revlog. The argument can have the |
|
2589 | 2589 | following values: |
|
2590 | 2590 | |
|
2591 | 2591 | DELTAREUSEALWAYS |
|
2592 | 2592 | Deltas will always be reused (if possible), even if the destination |
|
2593 | 2593 | revlog would not select the same revisions for the delta. This is the |
|
2594 | 2594 | fastest mode of operation. |
|
2595 | 2595 | DELTAREUSESAMEREVS |
|
2596 | 2596 | Deltas will be reused if the destination revlog would pick the same |
|
2597 | 2597 | revisions for the delta. This mode strikes a balance between speed |
|
2598 | 2598 | and optimization. |
|
2599 | 2599 | DELTAREUSENEVER |
|
2600 | 2600 | Deltas will never be reused. This is the slowest mode of execution. |
|
2601 | 2601 | This mode can be used to recompute deltas (e.g. if the diff/delta |
|
2602 | 2602 | algorithm changes). |
|
2603 | 2603 | DELTAREUSEFULLADD |
|
2604 | 2604 | Revision will be re-added as if their were new content. This is |
|
2605 | 2605 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. |
|
2606 | 2606 | eg: large file detection and handling. |
|
2607 | 2607 | |
|
2608 | 2608 | Delta computation can be slow, so the choice of delta reuse policy can |
|
2609 | 2609 | significantly affect run time. |
|
2610 | 2610 | |
|
2611 | 2611 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between |
|
2612 | 2612 | two extremes. Deltas will be reused if they are appropriate. But if the |
|
2613 | 2613 | delta could choose a better revision, it will do so. This means if you |
|
2614 | 2614 | are converting a non-generaldelta revlog to a generaldelta revlog, |
|
2615 | 2615 | deltas will be recomputed if the delta's parent isn't a parent of the |
|
2616 | 2616 | revision. |
|
2617 | 2617 | |
|
2618 | 2618 | In addition to the delta policy, the ``forcedeltabothparents`` |
|
2619 | 2619 | argument controls whether to force compute deltas against both parents |
|
2620 | 2620 | for merges. By default, the current default is used. |
|
2621 | 2621 | |
|
2622 | 2622 | If not None, the `sidedatacompanion` is callable that accept two |
|
2623 | 2623 | arguments: |
|
2624 | 2624 | |
|
2625 | 2625 | (srcrevlog, rev) |
|
2626 | 2626 | |
|
2627 | 2627 | and return a triplet that control changes to sidedata content from the |
|
2628 | 2628 | old revision to the new clone result: |
|
2629 | 2629 | |
|
2630 | 2630 | (dropall, filterout, update) |
|
2631 | 2631 | |
|
2632 | 2632 | * if `dropall` is True, all sidedata should be dropped |
|
2633 | 2633 | * `filterout` is a set of sidedata keys that should be dropped |
|
2634 | 2634 | * `update` is a mapping of additionnal/new key -> value |
|
2635 | 2635 | """ |
|
2636 | 2636 | if deltareuse not in self.DELTAREUSEALL: |
|
2637 | 2637 | raise ValueError( |
|
2638 | 2638 | _(b'value for deltareuse invalid: %s') % deltareuse |
|
2639 | 2639 | ) |
|
2640 | 2640 | |
|
2641 | 2641 | if len(destrevlog): |
|
2642 | 2642 | raise ValueError(_(b'destination revlog is not empty')) |
|
2643 | 2643 | |
|
2644 | 2644 | if getattr(self, 'filteredrevs', None): |
|
2645 | 2645 | raise ValueError(_(b'source revlog has filtered revisions')) |
|
2646 | 2646 | if getattr(destrevlog, 'filteredrevs', None): |
|
2647 | 2647 | raise ValueError(_(b'destination revlog has filtered revisions')) |
|
2648 | 2648 | |
|
2649 | 2649 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, |
|
2650 | 2650 | # if possible. |
|
2651 | 2651 | oldlazydelta = destrevlog._lazydelta |
|
2652 | 2652 | oldlazydeltabase = destrevlog._lazydeltabase |
|
2653 | 2653 | oldamd = destrevlog._deltabothparents |
|
2654 | 2654 | |
|
2655 | 2655 | try: |
|
2656 | 2656 | if deltareuse == self.DELTAREUSEALWAYS: |
|
2657 | 2657 | destrevlog._lazydeltabase = True |
|
2658 | 2658 | destrevlog._lazydelta = True |
|
2659 | 2659 | elif deltareuse == self.DELTAREUSESAMEREVS: |
|
2660 | 2660 | destrevlog._lazydeltabase = False |
|
2661 | 2661 | destrevlog._lazydelta = True |
|
2662 | 2662 | elif deltareuse == self.DELTAREUSENEVER: |
|
2663 | 2663 | destrevlog._lazydeltabase = False |
|
2664 | 2664 | destrevlog._lazydelta = False |
|
2665 | 2665 | |
|
2666 | 2666 | destrevlog._deltabothparents = forcedeltabothparents or oldamd |
|
2667 | 2667 | |
|
2668 | 2668 | self._clone( |
|
2669 | 2669 | tr, |
|
2670 | 2670 | destrevlog, |
|
2671 | 2671 | addrevisioncb, |
|
2672 | 2672 | deltareuse, |
|
2673 | 2673 | forcedeltabothparents, |
|
2674 | 2674 | sidedatacompanion, |
|
2675 | 2675 | ) |
|
2676 | 2676 | |
|
2677 | 2677 | finally: |
|
2678 | 2678 | destrevlog._lazydelta = oldlazydelta |
|
2679 | 2679 | destrevlog._lazydeltabase = oldlazydeltabase |
|
2680 | 2680 | destrevlog._deltabothparents = oldamd |
|
2681 | 2681 | |
|
2682 | 2682 | def _clone( |
|
2683 | 2683 | self, |
|
2684 | 2684 | tr, |
|
2685 | 2685 | destrevlog, |
|
2686 | 2686 | addrevisioncb, |
|
2687 | 2687 | deltareuse, |
|
2688 | 2688 | forcedeltabothparents, |
|
2689 | 2689 | sidedatacompanion, |
|
2690 | 2690 | ): |
|
2691 | 2691 | """perform the core duty of `revlog.clone` after parameter processing""" |
|
2692 | 2692 | deltacomputer = deltautil.deltacomputer(destrevlog) |
|
2693 | 2693 | index = self.index |
|
2694 | 2694 | for rev in self: |
|
2695 | 2695 | entry = index[rev] |
|
2696 | 2696 | |
|
2697 | 2697 | # Some classes override linkrev to take filtered revs into |
|
2698 | 2698 | # account. Use raw entry from index. |
|
2699 | 2699 | flags = entry[0] & 0xFFFF |
|
2700 | 2700 | linkrev = entry[4] |
|
2701 | 2701 | p1 = index[entry[5]][7] |
|
2702 | 2702 | p2 = index[entry[6]][7] |
|
2703 | 2703 | node = entry[7] |
|
2704 | 2704 | |
|
2705 | 2705 | sidedataactions = (False, [], {}) |
|
2706 | 2706 | if sidedatacompanion is not None: |
|
2707 | 2707 | sidedataactions = sidedatacompanion(self, rev) |
|
2708 | 2708 | |
|
2709 | 2709 | # (Possibly) reuse the delta from the revlog if allowed and |
|
2710 | 2710 | # the revlog chunk is a delta. |
|
2711 | 2711 | cachedelta = None |
|
2712 | 2712 | rawtext = None |
|
2713 | 2713 | if any(sidedataactions) or deltareuse == self.DELTAREUSEFULLADD: |
|
2714 | 2714 | dropall, filterout, update = sidedataactions |
|
2715 | 2715 | text, sidedata = self._revisiondata(rev) |
|
2716 | 2716 | if dropall: |
|
2717 | 2717 | sidedata = {} |
|
2718 | 2718 | for key in filterout: |
|
2719 | 2719 | sidedata.pop(key, None) |
|
2720 | 2720 | sidedata.update(update) |
|
2721 | 2721 | if not sidedata: |
|
2722 | 2722 | sidedata = None |
|
2723 | 2723 | destrevlog.addrevision( |
|
2724 | 2724 | text, |
|
2725 | 2725 | tr, |
|
2726 | 2726 | linkrev, |
|
2727 | 2727 | p1, |
|
2728 | 2728 | p2, |
|
2729 | 2729 | cachedelta=cachedelta, |
|
2730 | 2730 | node=node, |
|
2731 | 2731 | flags=flags, |
|
2732 | 2732 | deltacomputer=deltacomputer, |
|
2733 | 2733 | sidedata=sidedata, |
|
2734 | 2734 | ) |
|
2735 | 2735 | else: |
|
2736 | 2736 | if destrevlog._lazydelta: |
|
2737 | 2737 | dp = self.deltaparent(rev) |
|
2738 | 2738 | if dp != nullrev: |
|
2739 | 2739 | cachedelta = (dp, bytes(self._chunk(rev))) |
|
2740 | 2740 | |
|
2741 | 2741 | if not cachedelta: |
|
2742 | 2742 | rawtext = self.rawdata(rev) |
|
2743 | 2743 | |
|
2744 | 2744 | ifh = destrevlog.opener( |
|
2745 | 2745 | destrevlog.indexfile, b'a+', checkambig=False |
|
2746 | 2746 | ) |
|
2747 | 2747 | dfh = None |
|
2748 | 2748 | if not destrevlog._inline: |
|
2749 | 2749 | dfh = destrevlog.opener(destrevlog.datafile, b'a+') |
|
2750 | 2750 | try: |
|
2751 | 2751 | destrevlog._addrevision( |
|
2752 | 2752 | node, |
|
2753 | 2753 | rawtext, |
|
2754 | 2754 | tr, |
|
2755 | 2755 | linkrev, |
|
2756 | 2756 | p1, |
|
2757 | 2757 | p2, |
|
2758 | 2758 | flags, |
|
2759 | 2759 | cachedelta, |
|
2760 | 2760 | ifh, |
|
2761 | 2761 | dfh, |
|
2762 | 2762 | deltacomputer=deltacomputer, |
|
2763 | 2763 | ) |
|
2764 | 2764 | finally: |
|
2765 | 2765 | if dfh: |
|
2766 | 2766 | dfh.close() |
|
2767 | 2767 | ifh.close() |
|
2768 | 2768 | |
|
2769 | 2769 | if addrevisioncb: |
|
2770 | 2770 | addrevisioncb(self, rev, node) |
|
2771 | 2771 | |
|
2772 | 2772 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
2773 | 2773 | if (self.version & 0xFFFF) == REVLOGV0: |
|
2774 | 2774 | raise error.RevlogError( |
|
2775 | 2775 | _(b'cannot censor with version %d revlogs') % self.version |
|
2776 | 2776 | ) |
|
2777 | 2777 | |
|
2778 | 2778 | censorrev = self.rev(censornode) |
|
2779 | 2779 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') |
|
2780 | 2780 | |
|
2781 | 2781 | if len(tombstone) > self.rawsize(censorrev): |
|
2782 | 2782 | raise error.Abort( |
|
2783 | 2783 | _(b'censor tombstone must be no longer than censored data') |
|
2784 | 2784 | ) |
|
2785 | 2785 | |
|
2786 | 2786 | # Rewriting the revlog in place is hard. Our strategy for censoring is |
|
2787 | 2787 | # to create a new revlog, copy all revisions to it, then replace the |
|
2788 | 2788 | # revlogs on transaction close. |
|
2789 | 2789 | |
|
2790 | 2790 | newindexfile = self.indexfile + b'.tmpcensored' |
|
2791 | 2791 | newdatafile = self.datafile + b'.tmpcensored' |
|
2792 | 2792 | |
|
2793 | 2793 | # This is a bit dangerous. We could easily have a mismatch of state. |
|
2794 | 2794 | newrl = revlog(self.opener, newindexfile, newdatafile, censorable=True) |
|
2795 | 2795 | newrl.version = self.version |
|
2796 | 2796 | newrl._generaldelta = self._generaldelta |
|
2797 | 2797 | newrl._io = self._io |
|
2798 | 2798 | |
|
2799 | 2799 | for rev in self.revs(): |
|
2800 | 2800 | node = self.node(rev) |
|
2801 | 2801 | p1, p2 = self.parents(node) |
|
2802 | 2802 | |
|
2803 | 2803 | if rev == censorrev: |
|
2804 | 2804 | newrl.addrawrevision( |
|
2805 | 2805 | tombstone, |
|
2806 | 2806 | tr, |
|
2807 | 2807 | self.linkrev(censorrev), |
|
2808 | 2808 | p1, |
|
2809 | 2809 | p2, |
|
2810 | 2810 | censornode, |
|
2811 | 2811 | REVIDX_ISCENSORED, |
|
2812 | 2812 | ) |
|
2813 | 2813 | |
|
2814 | 2814 | if newrl.deltaparent(rev) != nullrev: |
|
2815 | 2815 | raise error.Abort( |
|
2816 | 2816 | _( |
|
2817 | 2817 | b'censored revision stored as delta; ' |
|
2818 | 2818 | b'cannot censor' |
|
2819 | 2819 | ), |
|
2820 | 2820 | hint=_( |
|
2821 | 2821 | b'censoring of revlogs is not ' |
|
2822 | 2822 | b'fully implemented; please report ' |
|
2823 | 2823 | b'this bug' |
|
2824 | 2824 | ), |
|
2825 | 2825 | ) |
|
2826 | 2826 | continue |
|
2827 | 2827 | |
|
2828 | 2828 | if self.iscensored(rev): |
|
2829 | 2829 | if self.deltaparent(rev) != nullrev: |
|
2830 | 2830 | raise error.Abort( |
|
2831 | 2831 | _( |
|
2832 | 2832 | b'cannot censor due to censored ' |
|
2833 | 2833 | b'revision having delta stored' |
|
2834 | 2834 | ) |
|
2835 | 2835 | ) |
|
2836 | 2836 | rawtext = self._chunk(rev) |
|
2837 | 2837 | else: |
|
2838 | 2838 | rawtext = self.rawdata(rev) |
|
2839 | 2839 | |
|
2840 | 2840 | newrl.addrawrevision( |
|
2841 | 2841 | rawtext, tr, self.linkrev(rev), p1, p2, node, self.flags(rev) |
|
2842 | 2842 | ) |
|
2843 | 2843 | |
|
2844 | 2844 | tr.addbackup(self.indexfile, location=b'store') |
|
2845 | 2845 | if not self._inline: |
|
2846 | 2846 | tr.addbackup(self.datafile, location=b'store') |
|
2847 | 2847 | |
|
2848 | 2848 | self.opener.rename(newrl.indexfile, self.indexfile) |
|
2849 | 2849 | if not self._inline: |
|
2850 | 2850 | self.opener.rename(newrl.datafile, self.datafile) |
|
2851 | 2851 | |
|
2852 | 2852 | self.clearcaches() |
|
2853 | 2853 | self._loadindex() |
|
2854 | 2854 | |
|
2855 | 2855 | def verifyintegrity(self, state): |
|
2856 | 2856 | """Verifies the integrity of the revlog. |
|
2857 | 2857 | |
|
2858 | 2858 | Yields ``revlogproblem`` instances describing problems that are |
|
2859 | 2859 | found. |
|
2860 | 2860 | """ |
|
2861 | 2861 | dd, di = self.checksize() |
|
2862 | 2862 | if dd: |
|
2863 | 2863 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) |
|
2864 | 2864 | if di: |
|
2865 | 2865 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) |
|
2866 | 2866 | |
|
2867 | 2867 | version = self.version & 0xFFFF |
|
2868 | 2868 | |
|
2869 | 2869 | # The verifier tells us what version revlog we should be. |
|
2870 | 2870 | if version != state[b'expectedversion']: |
|
2871 | 2871 | yield revlogproblem( |
|
2872 | 2872 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") |
|
2873 | 2873 | % (self.indexfile, version, state[b'expectedversion']) |
|
2874 | 2874 | ) |
|
2875 | 2875 | |
|
2876 | 2876 | state[b'skipread'] = set() |
|
2877 | state[b'safe_renamed'] = set() | |
|
2877 | 2878 | |
|
2878 | 2879 | for rev in self: |
|
2879 | 2880 | node = self.node(rev) |
|
2880 | 2881 | |
|
2881 | 2882 | # Verify contents. 4 cases to care about: |
|
2882 | 2883 | # |
|
2883 | 2884 | # common: the most common case |
|
2884 | 2885 | # rename: with a rename |
|
2885 | 2886 | # meta: file content starts with b'\1\n', the metadata |
|
2886 | 2887 | # header defined in filelog.py, but without a rename |
|
2887 | 2888 | # ext: content stored externally |
|
2888 | 2889 | # |
|
2889 | 2890 | # More formally, their differences are shown below: |
|
2890 | 2891 | # |
|
2891 | 2892 | # | common | rename | meta | ext |
|
2892 | 2893 | # ------------------------------------------------------- |
|
2893 | 2894 | # flags() | 0 | 0 | 0 | not 0 |
|
2894 | 2895 | # renamed() | False | True | False | ? |
|
2895 | 2896 | # rawtext[0:2]=='\1\n'| False | True | True | ? |
|
2896 | 2897 | # |
|
2897 | 2898 | # "rawtext" means the raw text stored in revlog data, which |
|
2898 | 2899 | # could be retrieved by "rawdata(rev)". "text" |
|
2899 | 2900 | # mentioned below is "revision(rev)". |
|
2900 | 2901 | # |
|
2901 | 2902 | # There are 3 different lengths stored physically: |
|
2902 | 2903 | # 1. L1: rawsize, stored in revlog index |
|
2903 | 2904 | # 2. L2: len(rawtext), stored in revlog data |
|
2904 | 2905 | # 3. L3: len(text), stored in revlog data if flags==0, or |
|
2905 | 2906 | # possibly somewhere else if flags!=0 |
|
2906 | 2907 | # |
|
2907 | 2908 | # L1 should be equal to L2. L3 could be different from them. |
|
2908 | 2909 | # "text" may or may not affect commit hash depending on flag |
|
2909 | 2910 | # processors (see flagutil.addflagprocessor). |
|
2910 | 2911 | # |
|
2911 | 2912 | # | common | rename | meta | ext |
|
2912 | 2913 | # ------------------------------------------------- |
|
2913 | 2914 | # rawsize() | L1 | L1 | L1 | L1 |
|
2914 | 2915 | # size() | L1 | L2-LM | L1(*) | L1 (?) |
|
2915 | 2916 | # len(rawtext) | L2 | L2 | L2 | L2 |
|
2916 | 2917 | # len(text) | L2 | L2 | L2 | L3 |
|
2917 | 2918 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) |
|
2918 | 2919 | # |
|
2919 | 2920 | # LM: length of metadata, depending on rawtext |
|
2920 | 2921 | # (*): not ideal, see comment in filelog.size |
|
2921 | 2922 | # (?): could be "- len(meta)" if the resolved content has |
|
2922 | 2923 | # rename metadata |
|
2923 | 2924 | # |
|
2924 | 2925 | # Checks needed to be done: |
|
2925 | 2926 | # 1. length check: L1 == L2, in all cases. |
|
2926 | 2927 | # 2. hash check: depending on flag processor, we may need to |
|
2927 | 2928 | # use either "text" (external), or "rawtext" (in revlog). |
|
2928 | 2929 | |
|
2929 | 2930 | try: |
|
2930 | 2931 | skipflags = state.get(b'skipflags', 0) |
|
2931 | 2932 | if skipflags: |
|
2932 | 2933 | skipflags &= self.flags(rev) |
|
2933 | 2934 | |
|
2934 | 2935 | _verify_revision(self, skipflags, state, node) |
|
2935 | 2936 | |
|
2936 | 2937 | l1 = self.rawsize(rev) |
|
2937 | 2938 | l2 = len(self.rawdata(node)) |
|
2938 | 2939 | |
|
2939 | 2940 | if l1 != l2: |
|
2940 | 2941 | yield revlogproblem( |
|
2941 | 2942 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), |
|
2942 | 2943 | node=node, |
|
2943 | 2944 | ) |
|
2944 | 2945 | |
|
2945 | 2946 | except error.CensoredNodeError: |
|
2946 | 2947 | if state[b'erroroncensored']: |
|
2947 | 2948 | yield revlogproblem( |
|
2948 | 2949 | error=_(b'censored file data'), node=node |
|
2949 | 2950 | ) |
|
2950 | 2951 | state[b'skipread'].add(node) |
|
2951 | 2952 | except Exception as e: |
|
2952 | 2953 | yield revlogproblem( |
|
2953 | 2954 | error=_(b'unpacking %s: %s') |
|
2954 | 2955 | % (short(node), stringutil.forcebytestr(e)), |
|
2955 | 2956 | node=node, |
|
2956 | 2957 | ) |
|
2957 | 2958 | state[b'skipread'].add(node) |
|
2958 | 2959 | |
|
2959 | 2960 | def storageinfo( |
|
2960 | 2961 | self, |
|
2961 | 2962 | exclusivefiles=False, |
|
2962 | 2963 | sharedfiles=False, |
|
2963 | 2964 | revisionscount=False, |
|
2964 | 2965 | trackedsize=False, |
|
2965 | 2966 | storedsize=False, |
|
2966 | 2967 | ): |
|
2967 | 2968 | d = {} |
|
2968 | 2969 | |
|
2969 | 2970 | if exclusivefiles: |
|
2970 | 2971 | d[b'exclusivefiles'] = [(self.opener, self.indexfile)] |
|
2971 | 2972 | if not self._inline: |
|
2972 | 2973 | d[b'exclusivefiles'].append((self.opener, self.datafile)) |
|
2973 | 2974 | |
|
2974 | 2975 | if sharedfiles: |
|
2975 | 2976 | d[b'sharedfiles'] = [] |
|
2976 | 2977 | |
|
2977 | 2978 | if revisionscount: |
|
2978 | 2979 | d[b'revisionscount'] = len(self) |
|
2979 | 2980 | |
|
2980 | 2981 | if trackedsize: |
|
2981 | 2982 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) |
|
2982 | 2983 | |
|
2983 | 2984 | if storedsize: |
|
2984 | 2985 | d[b'storedsize'] = sum( |
|
2985 | 2986 | self.opener.stat(path).st_size for path in self.files() |
|
2986 | 2987 | ) |
|
2987 | 2988 | |
|
2988 | 2989 | return d |
@@ -1,626 +1,628 b'' | |||
|
1 | 1 | # verify.py - repository integrity checking for Mercurial |
|
2 | 2 | # |
|
3 | 3 | # Copyright 2006, 2007 Matt Mackall <mpm@selenic.com> |
|
4 | 4 | # |
|
5 | 5 | # This software may be used and distributed according to the terms of the |
|
6 | 6 | # GNU General Public License version 2 or any later version. |
|
7 | 7 | |
|
8 | 8 | from __future__ import absolute_import |
|
9 | 9 | |
|
10 | 10 | import os |
|
11 | 11 | |
|
12 | 12 | from .i18n import _ |
|
13 | 13 | from .node import ( |
|
14 | 14 | nullid, |
|
15 | 15 | short, |
|
16 | 16 | ) |
|
17 | 17 | |
|
18 | 18 | from . import ( |
|
19 | 19 | error, |
|
20 | 20 | pycompat, |
|
21 | 21 | revlog, |
|
22 | 22 | util, |
|
23 | 23 | ) |
|
24 | 24 | |
|
25 | 25 | VERIFY_DEFAULT = 0 |
|
26 | 26 | VERIFY_FULL = 1 |
|
27 | 27 | |
|
28 | 28 | |
|
29 | 29 | def verify(repo, level=None): |
|
30 | 30 | with repo.lock(): |
|
31 | 31 | v = verifier(repo, level) |
|
32 | 32 | return v.verify() |
|
33 | 33 | |
|
34 | 34 | |
|
35 | 35 | def _normpath(f): |
|
36 | 36 | # under hg < 2.4, convert didn't sanitize paths properly, so a |
|
37 | 37 | # converted repo may contain repeated slashes |
|
38 | 38 | while b'//' in f: |
|
39 | 39 | f = f.replace(b'//', b'/') |
|
40 | 40 | return f |
|
41 | 41 | |
|
42 | 42 | |
|
43 | 43 | class verifier(object): |
|
44 | 44 | def __init__(self, repo, level=None): |
|
45 | 45 | self.repo = repo.unfiltered() |
|
46 | 46 | self.ui = repo.ui |
|
47 | 47 | self.match = repo.narrowmatch() |
|
48 | 48 | if level is None: |
|
49 | 49 | level = VERIFY_DEFAULT |
|
50 | 50 | self._level = level |
|
51 | 51 | self.badrevs = set() |
|
52 | 52 | self.errors = 0 |
|
53 | 53 | self.warnings = 0 |
|
54 | 54 | self.havecl = len(repo.changelog) > 0 |
|
55 | 55 | self.havemf = len(repo.manifestlog.getstorage(b'')) > 0 |
|
56 | 56 | self.revlogv1 = repo.changelog.version != revlog.REVLOGV0 |
|
57 | 57 | self.lrugetctx = util.lrucachefunc(repo.__getitem__) |
|
58 | 58 | self.refersmf = False |
|
59 | 59 | self.fncachewarned = False |
|
60 | 60 | # developer config: verify.skipflags |
|
61 | 61 | self.skipflags = repo.ui.configint(b'verify', b'skipflags') |
|
62 | 62 | self.warnorphanstorefiles = True |
|
63 | 63 | |
|
64 | 64 | def _warn(self, msg): |
|
65 | 65 | """record a "warning" level issue""" |
|
66 | 66 | self.ui.warn(msg + b"\n") |
|
67 | 67 | self.warnings += 1 |
|
68 | 68 | |
|
69 | 69 | def _err(self, linkrev, msg, filename=None): |
|
70 | 70 | """record a "error" level issue""" |
|
71 | 71 | if linkrev is not None: |
|
72 | 72 | self.badrevs.add(linkrev) |
|
73 | 73 | linkrev = b"%d" % linkrev |
|
74 | 74 | else: |
|
75 | 75 | linkrev = b'?' |
|
76 | 76 | msg = b"%s: %s" % (linkrev, msg) |
|
77 | 77 | if filename: |
|
78 | 78 | msg = b"%s@%s" % (filename, msg) |
|
79 | 79 | self.ui.warn(b" " + msg + b"\n") |
|
80 | 80 | self.errors += 1 |
|
81 | 81 | |
|
82 | 82 | def _exc(self, linkrev, msg, inst, filename=None): |
|
83 | 83 | """record exception raised during the verify process""" |
|
84 | 84 | fmsg = pycompat.bytestr(inst) |
|
85 | 85 | if not fmsg: |
|
86 | 86 | fmsg = pycompat.byterepr(inst) |
|
87 | 87 | self._err(linkrev, b"%s: %s" % (msg, fmsg), filename) |
|
88 | 88 | |
|
89 | 89 | def _checkrevlog(self, obj, name, linkrev): |
|
90 | 90 | """verify high level property of a revlog |
|
91 | 91 | |
|
92 | 92 | - revlog is present, |
|
93 | 93 | - revlog is non-empty, |
|
94 | 94 | - sizes (index and data) are correct, |
|
95 | 95 | - revlog's format version is correct. |
|
96 | 96 | """ |
|
97 | 97 | if not len(obj) and (self.havecl or self.havemf): |
|
98 | 98 | self._err(linkrev, _(b"empty or missing %s") % name) |
|
99 | 99 | return |
|
100 | 100 | |
|
101 | 101 | d = obj.checksize() |
|
102 | 102 | if d[0]: |
|
103 | 103 | self._err(None, _(b"data length off by %d bytes") % d[0], name) |
|
104 | 104 | if d[1]: |
|
105 | 105 | self._err(None, _(b"index contains %d extra bytes") % d[1], name) |
|
106 | 106 | |
|
107 | 107 | if obj.version != revlog.REVLOGV0: |
|
108 | 108 | if not self.revlogv1: |
|
109 | 109 | self._warn(_(b"warning: `%s' uses revlog format 1") % name) |
|
110 | 110 | elif self.revlogv1: |
|
111 | 111 | self._warn(_(b"warning: `%s' uses revlog format 0") % name) |
|
112 | 112 | |
|
113 | 113 | def _checkentry(self, obj, i, node, seen, linkrevs, f): |
|
114 | 114 | """verify a single revlog entry |
|
115 | 115 | |
|
116 | 116 | arguments are: |
|
117 | 117 | - obj: the source revlog |
|
118 | 118 | - i: the revision number |
|
119 | 119 | - node: the revision node id |
|
120 | 120 | - seen: nodes previously seen for this revlog |
|
121 | 121 | - linkrevs: [changelog-revisions] introducing "node" |
|
122 | 122 | - f: string label ("changelog", "manifest", or filename) |
|
123 | 123 | |
|
124 | 124 | Performs the following checks: |
|
125 | 125 | - linkrev points to an existing changelog revision, |
|
126 | 126 | - linkrev points to a changelog revision that introduces this revision, |
|
127 | 127 | - linkrev points to the lowest of these changesets, |
|
128 | 128 | - both parents exist in the revlog, |
|
129 | 129 | - the revision is not duplicated. |
|
130 | 130 | |
|
131 | 131 | Return the linkrev of the revision (or None for changelog's revisions). |
|
132 | 132 | """ |
|
133 | 133 | lr = obj.linkrev(obj.rev(node)) |
|
134 | 134 | if lr < 0 or (self.havecl and lr not in linkrevs): |
|
135 | 135 | if lr < 0 or lr >= len(self.repo.changelog): |
|
136 | 136 | msg = _(b"rev %d points to nonexistent changeset %d") |
|
137 | 137 | else: |
|
138 | 138 | msg = _(b"rev %d points to unexpected changeset %d") |
|
139 | 139 | self._err(None, msg % (i, lr), f) |
|
140 | 140 | if linkrevs: |
|
141 | 141 | if f and len(linkrevs) > 1: |
|
142 | 142 | try: |
|
143 | 143 | # attempt to filter down to real linkrevs |
|
144 | 144 | linkrevs = [ |
|
145 | 145 | l |
|
146 | 146 | for l in linkrevs |
|
147 | 147 | if self.lrugetctx(l)[f].filenode() == node |
|
148 | 148 | ] |
|
149 | 149 | except Exception: |
|
150 | 150 | pass |
|
151 | 151 | self._warn( |
|
152 | 152 | _(b" (expected %s)") |
|
153 | 153 | % b" ".join(map(pycompat.bytestr, linkrevs)) |
|
154 | 154 | ) |
|
155 | 155 | lr = None # can't be trusted |
|
156 | 156 | |
|
157 | 157 | try: |
|
158 | 158 | p1, p2 = obj.parents(node) |
|
159 | 159 | if p1 not in seen and p1 != nullid: |
|
160 | 160 | self._err( |
|
161 | 161 | lr, |
|
162 | 162 | _(b"unknown parent 1 %s of %s") % (short(p1), short(node)), |
|
163 | 163 | f, |
|
164 | 164 | ) |
|
165 | 165 | if p2 not in seen and p2 != nullid: |
|
166 | 166 | self._err( |
|
167 | 167 | lr, |
|
168 | 168 | _(b"unknown parent 2 %s of %s") % (short(p2), short(node)), |
|
169 | 169 | f, |
|
170 | 170 | ) |
|
171 | 171 | except Exception as inst: |
|
172 | 172 | self._exc(lr, _(b"checking parents of %s") % short(node), inst, f) |
|
173 | 173 | |
|
174 | 174 | if node in seen: |
|
175 | 175 | self._err(lr, _(b"duplicate revision %d (%d)") % (i, seen[node]), f) |
|
176 | 176 | seen[node] = i |
|
177 | 177 | return lr |
|
178 | 178 | |
|
179 | 179 | def verify(self): |
|
180 | 180 | """verify the content of the Mercurial repository |
|
181 | 181 | |
|
182 | 182 | This method run all verifications, displaying issues as they are found. |
|
183 | 183 | |
|
184 | 184 | return 1 if any error have been encountered, 0 otherwise.""" |
|
185 | 185 | # initial validation and generic report |
|
186 | 186 | repo = self.repo |
|
187 | 187 | ui = repo.ui |
|
188 | 188 | if not repo.url().startswith(b'file:'): |
|
189 | 189 | raise error.Abort(_(b"cannot verify bundle or remote repos")) |
|
190 | 190 | |
|
191 | 191 | if os.path.exists(repo.sjoin(b"journal")): |
|
192 | 192 | ui.warn(_(b"abandoned transaction found - run hg recover\n")) |
|
193 | 193 | |
|
194 | 194 | if ui.verbose or not self.revlogv1: |
|
195 | 195 | ui.status( |
|
196 | 196 | _(b"repository uses revlog format %d\n") |
|
197 | 197 | % (self.revlogv1 and 1 or 0) |
|
198 | 198 | ) |
|
199 | 199 | |
|
200 | 200 | # data verification |
|
201 | 201 | mflinkrevs, filelinkrevs = self._verifychangelog() |
|
202 | 202 | filenodes = self._verifymanifest(mflinkrevs) |
|
203 | 203 | del mflinkrevs |
|
204 | 204 | self._crosscheckfiles(filelinkrevs, filenodes) |
|
205 | 205 | totalfiles, filerevisions = self._verifyfiles(filenodes, filelinkrevs) |
|
206 | 206 | |
|
207 | 207 | # final report |
|
208 | 208 | ui.status( |
|
209 | 209 | _(b"checked %d changesets with %d changes to %d files\n") |
|
210 | 210 | % (len(repo.changelog), filerevisions, totalfiles) |
|
211 | 211 | ) |
|
212 | 212 | if self.warnings: |
|
213 | 213 | ui.warn(_(b"%d warnings encountered!\n") % self.warnings) |
|
214 | 214 | if self.fncachewarned: |
|
215 | 215 | ui.warn( |
|
216 | 216 | _( |
|
217 | 217 | b'hint: run "hg debugrebuildfncache" to recover from ' |
|
218 | 218 | b'corrupt fncache\n' |
|
219 | 219 | ) |
|
220 | 220 | ) |
|
221 | 221 | if self.errors: |
|
222 | 222 | ui.warn(_(b"%d integrity errors encountered!\n") % self.errors) |
|
223 | 223 | if self.badrevs: |
|
224 | 224 | ui.warn( |
|
225 | 225 | _(b"(first damaged changeset appears to be %d)\n") |
|
226 | 226 | % min(self.badrevs) |
|
227 | 227 | ) |
|
228 | 228 | return 1 |
|
229 | 229 | return 0 |
|
230 | 230 | |
|
231 | 231 | def _verifychangelog(self): |
|
232 | 232 | """verify the changelog of a repository |
|
233 | 233 | |
|
234 | 234 | The following checks are performed: |
|
235 | 235 | - all of `_checkrevlog` checks, |
|
236 | 236 | - all of `_checkentry` checks (for each revisions), |
|
237 | 237 | - each revision can be read. |
|
238 | 238 | |
|
239 | 239 | The function returns some of the data observed in the changesets as a |
|
240 | 240 | (mflinkrevs, filelinkrevs) tuples: |
|
241 | 241 | - mflinkrevs: is a { manifest-node -> [changelog-rev] } mapping |
|
242 | 242 | - filelinkrevs: is a { file-path -> [changelog-rev] } mapping |
|
243 | 243 | |
|
244 | 244 | If a matcher was specified, filelinkrevs will only contains matched |
|
245 | 245 | files. |
|
246 | 246 | """ |
|
247 | 247 | ui = self.ui |
|
248 | 248 | repo = self.repo |
|
249 | 249 | match = self.match |
|
250 | 250 | cl = repo.changelog |
|
251 | 251 | |
|
252 | 252 | ui.status(_(b"checking changesets\n")) |
|
253 | 253 | mflinkrevs = {} |
|
254 | 254 | filelinkrevs = {} |
|
255 | 255 | seen = {} |
|
256 | 256 | self._checkrevlog(cl, b"changelog", 0) |
|
257 | 257 | progress = ui.makeprogress( |
|
258 | 258 | _(b'checking'), unit=_(b'changesets'), total=len(repo) |
|
259 | 259 | ) |
|
260 | 260 | for i in repo: |
|
261 | 261 | progress.update(i) |
|
262 | 262 | n = cl.node(i) |
|
263 | 263 | self._checkentry(cl, i, n, seen, [i], b"changelog") |
|
264 | 264 | |
|
265 | 265 | try: |
|
266 | 266 | changes = cl.read(n) |
|
267 | 267 | if changes[0] != nullid: |
|
268 | 268 | mflinkrevs.setdefault(changes[0], []).append(i) |
|
269 | 269 | self.refersmf = True |
|
270 | 270 | for f in changes[3]: |
|
271 | 271 | if match(f): |
|
272 | 272 | filelinkrevs.setdefault(_normpath(f), []).append(i) |
|
273 | 273 | except Exception as inst: |
|
274 | 274 | self.refersmf = True |
|
275 | 275 | self._exc(i, _(b"unpacking changeset %s") % short(n), inst) |
|
276 | 276 | progress.complete() |
|
277 | 277 | return mflinkrevs, filelinkrevs |
|
278 | 278 | |
|
279 | 279 | def _verifymanifest( |
|
280 | 280 | self, mflinkrevs, dir=b"", storefiles=None, subdirprogress=None |
|
281 | 281 | ): |
|
282 | 282 | """verify the manifestlog content |
|
283 | 283 | |
|
284 | 284 | Inputs: |
|
285 | 285 | - mflinkrevs: a {manifest-node -> [changelog-revisions]} mapping |
|
286 | 286 | - dir: a subdirectory to check (for tree manifest repo) |
|
287 | 287 | - storefiles: set of currently "orphan" files. |
|
288 | 288 | - subdirprogress: a progress object |
|
289 | 289 | |
|
290 | 290 | This function checks: |
|
291 | 291 | * all of `_checkrevlog` checks (for all manifest related revlogs) |
|
292 | 292 | * all of `_checkentry` checks (for all manifest related revisions) |
|
293 | 293 | * nodes for subdirectory exists in the sub-directory manifest |
|
294 | 294 | * each manifest entries have a file path |
|
295 | 295 | * each manifest node refered in mflinkrevs exist in the manifest log |
|
296 | 296 | |
|
297 | 297 | If tree manifest is in use and a matchers is specified, only the |
|
298 | 298 | sub-directories matching it will be verified. |
|
299 | 299 | |
|
300 | 300 | return a two level mapping: |
|
301 | 301 | {"path" -> { filenode -> changelog-revision}} |
|
302 | 302 | |
|
303 | 303 | This mapping primarily contains entries for every files in the |
|
304 | 304 | repository. In addition, when tree-manifest is used, it also contains |
|
305 | 305 | sub-directory entries. |
|
306 | 306 | |
|
307 | 307 | If a matcher is provided, only matching paths will be included. |
|
308 | 308 | """ |
|
309 | 309 | repo = self.repo |
|
310 | 310 | ui = self.ui |
|
311 | 311 | match = self.match |
|
312 | 312 | mfl = self.repo.manifestlog |
|
313 | 313 | mf = mfl.getstorage(dir) |
|
314 | 314 | |
|
315 | 315 | if not dir: |
|
316 | 316 | self.ui.status(_(b"checking manifests\n")) |
|
317 | 317 | |
|
318 | 318 | filenodes = {} |
|
319 | 319 | subdirnodes = {} |
|
320 | 320 | seen = {} |
|
321 | 321 | label = b"manifest" |
|
322 | 322 | if dir: |
|
323 | 323 | label = dir |
|
324 | 324 | revlogfiles = mf.files() |
|
325 | 325 | storefiles.difference_update(revlogfiles) |
|
326 | 326 | if subdirprogress: # should be true since we're in a subdirectory |
|
327 | 327 | subdirprogress.increment() |
|
328 | 328 | if self.refersmf: |
|
329 | 329 | # Do not check manifest if there are only changelog entries with |
|
330 | 330 | # null manifests. |
|
331 | 331 | self._checkrevlog(mf, label, 0) |
|
332 | 332 | progress = ui.makeprogress( |
|
333 | 333 | _(b'checking'), unit=_(b'manifests'), total=len(mf) |
|
334 | 334 | ) |
|
335 | 335 | for i in mf: |
|
336 | 336 | if not dir: |
|
337 | 337 | progress.update(i) |
|
338 | 338 | n = mf.node(i) |
|
339 | 339 | lr = self._checkentry(mf, i, n, seen, mflinkrevs.get(n, []), label) |
|
340 | 340 | if n in mflinkrevs: |
|
341 | 341 | del mflinkrevs[n] |
|
342 | 342 | elif dir: |
|
343 | 343 | self._err( |
|
344 | 344 | lr, |
|
345 | 345 | _(b"%s not in parent-directory manifest") % short(n), |
|
346 | 346 | label, |
|
347 | 347 | ) |
|
348 | 348 | else: |
|
349 | 349 | self._err(lr, _(b"%s not in changesets") % short(n), label) |
|
350 | 350 | |
|
351 | 351 | try: |
|
352 | 352 | mfdelta = mfl.get(dir, n).readdelta(shallow=True) |
|
353 | 353 | for f, fn, fl in mfdelta.iterentries(): |
|
354 | 354 | if not f: |
|
355 | 355 | self._err(lr, _(b"entry without name in manifest")) |
|
356 | 356 | elif f == b"/dev/null": # ignore this in very old repos |
|
357 | 357 | continue |
|
358 | 358 | fullpath = dir + _normpath(f) |
|
359 | 359 | if fl == b't': |
|
360 | 360 | if not match.visitdir(fullpath): |
|
361 | 361 | continue |
|
362 | 362 | subdirnodes.setdefault(fullpath + b'/', {}).setdefault( |
|
363 | 363 | fn, [] |
|
364 | 364 | ).append(lr) |
|
365 | 365 | else: |
|
366 | 366 | if not match(fullpath): |
|
367 | 367 | continue |
|
368 | 368 | filenodes.setdefault(fullpath, {}).setdefault(fn, lr) |
|
369 | 369 | except Exception as inst: |
|
370 | 370 | self._exc(lr, _(b"reading delta %s") % short(n), inst, label) |
|
371 | 371 | if self._level >= VERIFY_FULL: |
|
372 | 372 | try: |
|
373 | 373 | # Various issues can affect manifest. So we read each full |
|
374 | 374 | # text from storage. This triggers the checks from the core |
|
375 | 375 | # code (eg: hash verification, filename are ordered, etc.) |
|
376 | 376 | mfdelta = mfl.get(dir, n).read() |
|
377 | 377 | except Exception as inst: |
|
378 | 378 | self._exc( |
|
379 | 379 | lr, |
|
380 | 380 | _(b"reading full manifest %s") % short(n), |
|
381 | 381 | inst, |
|
382 | 382 | label, |
|
383 | 383 | ) |
|
384 | 384 | |
|
385 | 385 | if not dir: |
|
386 | 386 | progress.complete() |
|
387 | 387 | |
|
388 | 388 | if self.havemf: |
|
389 | 389 | # since we delete entry in `mflinkrevs` during iteration, any |
|
390 | 390 | # remaining entries are "missing". We need to issue errors for them. |
|
391 | 391 | changesetpairs = [(c, m) for m in mflinkrevs for c in mflinkrevs[m]] |
|
392 | 392 | for c, m in sorted(changesetpairs): |
|
393 | 393 | if dir: |
|
394 | 394 | self._err( |
|
395 | 395 | c, |
|
396 | 396 | _( |
|
397 | 397 | b"parent-directory manifest refers to unknown" |
|
398 | 398 | b" revision %s" |
|
399 | 399 | ) |
|
400 | 400 | % short(m), |
|
401 | 401 | label, |
|
402 | 402 | ) |
|
403 | 403 | else: |
|
404 | 404 | self._err( |
|
405 | 405 | c, |
|
406 | 406 | _(b"changeset refers to unknown revision %s") |
|
407 | 407 | % short(m), |
|
408 | 408 | label, |
|
409 | 409 | ) |
|
410 | 410 | |
|
411 | 411 | if not dir and subdirnodes: |
|
412 | 412 | self.ui.status(_(b"checking directory manifests\n")) |
|
413 | 413 | storefiles = set() |
|
414 | 414 | subdirs = set() |
|
415 | 415 | revlogv1 = self.revlogv1 |
|
416 | 416 | for f, f2, size in repo.store.datafiles(): |
|
417 | 417 | if not f: |
|
418 | 418 | self._err(None, _(b"cannot decode filename '%s'") % f2) |
|
419 | 419 | elif (size > 0 or not revlogv1) and f.startswith(b'meta/'): |
|
420 | 420 | storefiles.add(_normpath(f)) |
|
421 | 421 | subdirs.add(os.path.dirname(f)) |
|
422 | 422 | subdirprogress = ui.makeprogress( |
|
423 | 423 | _(b'checking'), unit=_(b'manifests'), total=len(subdirs) |
|
424 | 424 | ) |
|
425 | 425 | |
|
426 | 426 | for subdir, linkrevs in pycompat.iteritems(subdirnodes): |
|
427 | 427 | subdirfilenodes = self._verifymanifest( |
|
428 | 428 | linkrevs, subdir, storefiles, subdirprogress |
|
429 | 429 | ) |
|
430 | 430 | for f, onefilenodes in pycompat.iteritems(subdirfilenodes): |
|
431 | 431 | filenodes.setdefault(f, {}).update(onefilenodes) |
|
432 | 432 | |
|
433 | 433 | if not dir and subdirnodes: |
|
434 | 434 | subdirprogress.complete() |
|
435 | 435 | if self.warnorphanstorefiles: |
|
436 | 436 | for f in sorted(storefiles): |
|
437 | 437 | self._warn(_(b"warning: orphan data file '%s'") % f) |
|
438 | 438 | |
|
439 | 439 | return filenodes |
|
440 | 440 | |
|
441 | 441 | def _crosscheckfiles(self, filelinkrevs, filenodes): |
|
442 | 442 | repo = self.repo |
|
443 | 443 | ui = self.ui |
|
444 | 444 | ui.status(_(b"crosschecking files in changesets and manifests\n")) |
|
445 | 445 | |
|
446 | 446 | total = len(filelinkrevs) + len(filenodes) |
|
447 | 447 | progress = ui.makeprogress( |
|
448 | 448 | _(b'crosschecking'), unit=_(b'files'), total=total |
|
449 | 449 | ) |
|
450 | 450 | if self.havemf: |
|
451 | 451 | for f in sorted(filelinkrevs): |
|
452 | 452 | progress.increment() |
|
453 | 453 | if f not in filenodes: |
|
454 | 454 | lr = filelinkrevs[f][0] |
|
455 | 455 | self._err(lr, _(b"in changeset but not in manifest"), f) |
|
456 | 456 | |
|
457 | 457 | if self.havecl: |
|
458 | 458 | for f in sorted(filenodes): |
|
459 | 459 | progress.increment() |
|
460 | 460 | if f not in filelinkrevs: |
|
461 | 461 | try: |
|
462 | 462 | fl = repo.file(f) |
|
463 | 463 | lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]]) |
|
464 | 464 | except Exception: |
|
465 | 465 | lr = None |
|
466 | 466 | self._err(lr, _(b"in manifest but not in changeset"), f) |
|
467 | 467 | |
|
468 | 468 | progress.complete() |
|
469 | 469 | |
|
470 | 470 | def _verifyfiles(self, filenodes, filelinkrevs): |
|
471 | 471 | repo = self.repo |
|
472 | 472 | ui = self.ui |
|
473 | 473 | lrugetctx = self.lrugetctx |
|
474 | 474 | revlogv1 = self.revlogv1 |
|
475 | 475 | havemf = self.havemf |
|
476 | 476 | ui.status(_(b"checking files\n")) |
|
477 | 477 | |
|
478 | 478 | storefiles = set() |
|
479 | 479 | for f, f2, size in repo.store.datafiles(): |
|
480 | 480 | if not f: |
|
481 | 481 | self._err(None, _(b"cannot decode filename '%s'") % f2) |
|
482 | 482 | elif (size > 0 or not revlogv1) and f.startswith(b'data/'): |
|
483 | 483 | storefiles.add(_normpath(f)) |
|
484 | 484 | |
|
485 | 485 | state = { |
|
486 | 486 | # TODO this assumes revlog storage for changelog. |
|
487 | 487 | b'expectedversion': self.repo.changelog.version & 0xFFFF, |
|
488 | 488 | b'skipflags': self.skipflags, |
|
489 | 489 | # experimental config: censor.policy |
|
490 | 490 | b'erroroncensored': ui.config(b'censor', b'policy') == b'abort', |
|
491 | 491 | } |
|
492 | 492 | |
|
493 | 493 | files = sorted(set(filenodes) | set(filelinkrevs)) |
|
494 | 494 | revisions = 0 |
|
495 | 495 | progress = ui.makeprogress( |
|
496 | 496 | _(b'checking'), unit=_(b'files'), total=len(files) |
|
497 | 497 | ) |
|
498 | 498 | for i, f in enumerate(files): |
|
499 | 499 | progress.update(i, item=f) |
|
500 | 500 | try: |
|
501 | 501 | linkrevs = filelinkrevs[f] |
|
502 | 502 | except KeyError: |
|
503 | 503 | # in manifest but not in changelog |
|
504 | 504 | linkrevs = [] |
|
505 | 505 | |
|
506 | 506 | if linkrevs: |
|
507 | 507 | lr = linkrevs[0] |
|
508 | 508 | else: |
|
509 | 509 | lr = None |
|
510 | 510 | |
|
511 | 511 | try: |
|
512 | 512 | fl = repo.file(f) |
|
513 | 513 | except error.StorageError as e: |
|
514 | 514 | self._err(lr, _(b"broken revlog! (%s)") % e, f) |
|
515 | 515 | continue |
|
516 | 516 | |
|
517 | 517 | for ff in fl.files(): |
|
518 | 518 | try: |
|
519 | 519 | storefiles.remove(ff) |
|
520 | 520 | except KeyError: |
|
521 | 521 | if self.warnorphanstorefiles: |
|
522 | 522 | self._warn( |
|
523 | 523 | _(b" warning: revlog '%s' not in fncache!") % ff |
|
524 | 524 | ) |
|
525 | 525 | self.fncachewarned = True |
|
526 | 526 | |
|
527 | 527 | if not len(fl) and (self.havecl or self.havemf): |
|
528 | 528 | self._err(lr, _(b"empty or missing %s") % f) |
|
529 | 529 | else: |
|
530 | 530 | # Guard against implementations not setting this. |
|
531 | 531 | state[b'skipread'] = set() |
|
532 | state[b'safe_renamed'] = set() | |
|
533 | ||
|
532 | 534 | for problem in fl.verifyintegrity(state): |
|
533 | 535 | if problem.node is not None: |
|
534 | 536 | linkrev = fl.linkrev(fl.rev(problem.node)) |
|
535 | 537 | else: |
|
536 | 538 | linkrev = None |
|
537 | 539 | |
|
538 | 540 | if problem.warning: |
|
539 | 541 | self._warn(problem.warning) |
|
540 | 542 | elif problem.error: |
|
541 | 543 | self._err( |
|
542 | 544 | linkrev if linkrev is not None else lr, |
|
543 | 545 | problem.error, |
|
544 | 546 | f, |
|
545 | 547 | ) |
|
546 | 548 | else: |
|
547 | 549 | raise error.ProgrammingError( |
|
548 | 550 | b'problem instance does not set warning or error ' |
|
549 | 551 | b'attribute: %s' % problem.msg |
|
550 | 552 | ) |
|
551 | 553 | |
|
552 | 554 | seen = {} |
|
553 | 555 | for i in fl: |
|
554 | 556 | revisions += 1 |
|
555 | 557 | n = fl.node(i) |
|
556 | 558 | lr = self._checkentry(fl, i, n, seen, linkrevs, f) |
|
557 | 559 | if f in filenodes: |
|
558 | 560 | if havemf and n not in filenodes[f]: |
|
559 | 561 | self._err(lr, _(b"%s not in manifests") % (short(n)), f) |
|
560 | 562 | else: |
|
561 | 563 | del filenodes[f][n] |
|
562 | 564 | |
|
563 | if n in state[b'skipread']: | |
|
565 | if n in state[b'skipread'] and n not in state[b'safe_renamed']: | |
|
564 | 566 | continue |
|
565 | 567 | |
|
566 | 568 | # check renames |
|
567 | 569 | try: |
|
568 | 570 | # This requires resolving fulltext (at least on revlogs, |
|
569 | 571 | # though not with LFS revisions). We may want |
|
570 | 572 | # ``verifyintegrity()`` to pass a set of nodes with |
|
571 | 573 | # rename metadata as an optimization. |
|
572 | 574 | rp = fl.renamed(n) |
|
573 | 575 | if rp: |
|
574 | 576 | if lr is not None and ui.verbose: |
|
575 | 577 | ctx = lrugetctx(lr) |
|
576 | 578 | if not any(rp[0] in pctx for pctx in ctx.parents()): |
|
577 | 579 | self._warn( |
|
578 | 580 | _( |
|
579 | 581 | b"warning: copy source of '%s' not" |
|
580 | 582 | b" in parents of %s" |
|
581 | 583 | ) |
|
582 | 584 | % (f, ctx) |
|
583 | 585 | ) |
|
584 | 586 | fl2 = repo.file(rp[0]) |
|
585 | 587 | if not len(fl2): |
|
586 | 588 | self._err( |
|
587 | 589 | lr, |
|
588 | 590 | _( |
|
589 | 591 | b"empty or missing copy source revlog " |
|
590 | 592 | b"%s:%s" |
|
591 | 593 | ) |
|
592 | 594 | % (rp[0], short(rp[1])), |
|
593 | 595 | f, |
|
594 | 596 | ) |
|
595 | 597 | elif rp[1] == nullid: |
|
596 | 598 | ui.note( |
|
597 | 599 | _( |
|
598 | 600 | b"warning: %s@%s: copy source" |
|
599 | 601 | b" revision is nullid %s:%s\n" |
|
600 | 602 | ) |
|
601 | 603 | % (f, lr, rp[0], short(rp[1])) |
|
602 | 604 | ) |
|
603 | 605 | else: |
|
604 | 606 | fl2.rev(rp[1]) |
|
605 | 607 | except Exception as inst: |
|
606 | 608 | self._exc( |
|
607 | 609 | lr, _(b"checking rename of %s") % short(n), inst, f |
|
608 | 610 | ) |
|
609 | 611 | |
|
610 | 612 | # cross-check |
|
611 | 613 | if f in filenodes: |
|
612 | 614 | fns = [(v, k) for k, v in pycompat.iteritems(filenodes[f])] |
|
613 | 615 | for lr, node in sorted(fns): |
|
614 | 616 | self._err( |
|
615 | 617 | lr, |
|
616 | 618 | _(b"manifest refers to unknown revision %s") |
|
617 | 619 | % short(node), |
|
618 | 620 | f, |
|
619 | 621 | ) |
|
620 | 622 | progress.complete() |
|
621 | 623 | |
|
622 | 624 | if self.warnorphanstorefiles: |
|
623 | 625 | for f in sorted(storefiles): |
|
624 | 626 | self._warn(_(b"warning: orphan data file '%s'") % f) |
|
625 | 627 | |
|
626 | 628 | return len(files), revisions |
@@ -1,1192 +1,1231 b'' | |||
|
1 | 1 | #require no-reposimplestore no-chg |
|
2 | 2 | |
|
3 | 3 | $ hg init requirements |
|
4 | 4 | $ cd requirements |
|
5 | 5 | |
|
6 | 6 | # LFS not loaded by default. |
|
7 | 7 | |
|
8 | 8 | $ hg config extensions |
|
9 | 9 | [1] |
|
10 | 10 | |
|
11 | 11 | # Adding lfs to requires file will auto-load lfs extension. |
|
12 | 12 | |
|
13 | 13 | $ echo lfs >> .hg/requires |
|
14 | 14 | $ hg config extensions |
|
15 | 15 | extensions.lfs= |
|
16 | 16 | |
|
17 | 17 | # But only if there is no config entry for the extension already. |
|
18 | 18 | |
|
19 | 19 | $ cat > .hg/hgrc << EOF |
|
20 | 20 | > [extensions] |
|
21 | 21 | > lfs=! |
|
22 | 22 | > EOF |
|
23 | 23 | |
|
24 | 24 | $ hg config extensions |
|
25 | 25 | abort: repository requires features unknown to this Mercurial: lfs! |
|
26 | 26 | (see https://mercurial-scm.org/wiki/MissingRequirement for more information) |
|
27 | 27 | [255] |
|
28 | 28 | |
|
29 | 29 | $ cat > .hg/hgrc << EOF |
|
30 | 30 | > [extensions] |
|
31 | 31 | > lfs= |
|
32 | 32 | > EOF |
|
33 | 33 | |
|
34 | 34 | $ hg config extensions |
|
35 | 35 | extensions.lfs= |
|
36 | 36 | |
|
37 | 37 | $ cat > .hg/hgrc << EOF |
|
38 | 38 | > [extensions] |
|
39 | 39 | > lfs = missing.py |
|
40 | 40 | > EOF |
|
41 | 41 | |
|
42 | 42 | $ hg config extensions |
|
43 | 43 | \*\*\* failed to import extension lfs from missing.py: [Errno *] $ENOENT$: 'missing.py' (glob) |
|
44 | 44 | abort: repository requires features unknown to this Mercurial: lfs! |
|
45 | 45 | (see https://mercurial-scm.org/wiki/MissingRequirement for more information) |
|
46 | 46 | [255] |
|
47 | 47 | |
|
48 | 48 | $ cd .. |
|
49 | 49 | |
|
50 | 50 | # Initial setup |
|
51 | 51 | |
|
52 | 52 | $ cat >> $HGRCPATH << EOF |
|
53 | 53 | > [extensions] |
|
54 | 54 | > lfs= |
|
55 | 55 | > [lfs] |
|
56 | 56 | > # Test deprecated config |
|
57 | 57 | > threshold=1000B |
|
58 | 58 | > EOF |
|
59 | 59 | |
|
60 | 60 | $ LONG=AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC |
|
61 | 61 | |
|
62 | 62 | # Prepare server and enable extension |
|
63 | 63 | $ hg init server |
|
64 | 64 | $ hg clone -q server client |
|
65 | 65 | $ cd client |
|
66 | 66 | |
|
67 | 67 | # Commit small file |
|
68 | 68 | $ echo s > smallfile |
|
69 | 69 | $ echo '**.py = LF' > .hgeol |
|
70 | 70 | $ hg --config lfs.track='"size(\">1000B\")"' commit -Aqm "add small file" |
|
71 | 71 | hg: parse error: unsupported file pattern: size(">1000B") |
|
72 | 72 | (paths must be prefixed with "path:") |
|
73 | 73 | [255] |
|
74 | 74 | $ hg --config lfs.track='size(">1000B")' commit -Aqm "add small file" |
|
75 | 75 | |
|
76 | 76 | # Commit large file |
|
77 | 77 | $ echo $LONG > largefile |
|
78 | 78 | $ grep lfs .hg/requires |
|
79 | 79 | [1] |
|
80 | 80 | $ hg commit --traceback -Aqm "add large file" |
|
81 | 81 | $ grep lfs .hg/requires |
|
82 | 82 | lfs |
|
83 | 83 | |
|
84 | 84 | # Ensure metadata is stored |
|
85 | 85 | $ hg debugdata largefile 0 |
|
86 | 86 | version https://git-lfs.github.com/spec/v1 |
|
87 | 87 | oid sha256:f11e77c257047a398492d8d6cb9f6acf3aa7c4384bb23080b43546053e183e4b |
|
88 | 88 | size 1501 |
|
89 | 89 | x-is-binary 0 |
|
90 | 90 | |
|
91 | 91 | # Check the blobstore is populated |
|
92 | 92 | $ find .hg/store/lfs/objects | sort |
|
93 | 93 | .hg/store/lfs/objects |
|
94 | 94 | .hg/store/lfs/objects/f1 |
|
95 | 95 | .hg/store/lfs/objects/f1/1e77c257047a398492d8d6cb9f6acf3aa7c4384bb23080b43546053e183e4b |
|
96 | 96 | |
|
97 | 97 | # Check the blob stored contains the actual contents of the file |
|
98 | 98 | $ cat .hg/store/lfs/objects/f1/1e77c257047a398492d8d6cb9f6acf3aa7c4384bb23080b43546053e183e4b |
|
99 | 99 | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC |
|
100 | 100 | |
|
101 | 101 | # Push changes to the server |
|
102 | 102 | |
|
103 | 103 | $ hg push |
|
104 | 104 | pushing to $TESTTMP/server |
|
105 | 105 | searching for changes |
|
106 | 106 | abort: lfs.url needs to be configured |
|
107 | 107 | [255] |
|
108 | 108 | |
|
109 | 109 | $ cat >> $HGRCPATH << EOF |
|
110 | 110 | > [lfs] |
|
111 | 111 | > url=file:$TESTTMP/dummy-remote/ |
|
112 | 112 | > EOF |
|
113 | 113 | |
|
114 | 114 | Push to a local non-lfs repo with the extension enabled will add the |
|
115 | 115 | lfs requirement |
|
116 | 116 | |
|
117 | 117 | $ grep lfs $TESTTMP/server/.hg/requires |
|
118 | 118 | [1] |
|
119 | 119 | $ hg push -v | egrep -v '^(uncompressed| )' |
|
120 | 120 | pushing to $TESTTMP/server |
|
121 | 121 | searching for changes |
|
122 | 122 | lfs: found f11e77c257047a398492d8d6cb9f6acf3aa7c4384bb23080b43546053e183e4b in the local lfs store |
|
123 | 123 | 2 changesets found |
|
124 | 124 | adding changesets |
|
125 | 125 | adding manifests |
|
126 | 126 | adding file changes |
|
127 | 127 | calling hook pretxnchangegroup.lfs: hgext.lfs.checkrequireslfs |
|
128 | 128 | added 2 changesets with 3 changes to 3 files |
|
129 | 129 | $ grep lfs $TESTTMP/server/.hg/requires |
|
130 | 130 | lfs |
|
131 | 131 | |
|
132 | 132 | # Unknown URL scheme |
|
133 | 133 | |
|
134 | 134 | $ hg push --config lfs.url=ftp://foobar |
|
135 | 135 | abort: lfs: unknown url scheme: ftp |
|
136 | 136 | [255] |
|
137 | 137 | |
|
138 | 138 | $ cd ../ |
|
139 | 139 | |
|
140 | 140 | # Initialize new client (not cloning) and setup extension |
|
141 | 141 | $ hg init client2 |
|
142 | 142 | $ cd client2 |
|
143 | 143 | $ cat >> .hg/hgrc <<EOF |
|
144 | 144 | > [paths] |
|
145 | 145 | > default = $TESTTMP/server |
|
146 | 146 | > EOF |
|
147 | 147 | |
|
148 | 148 | # Pull from server |
|
149 | 149 | |
|
150 | 150 | Pulling a local lfs repo into a local non-lfs repo with the extension |
|
151 | 151 | enabled adds the lfs requirement |
|
152 | 152 | |
|
153 | 153 | $ grep lfs .hg/requires $TESTTMP/server/.hg/requires |
|
154 | 154 | $TESTTMP/server/.hg/requires:lfs |
|
155 | 155 | $ hg pull default |
|
156 | 156 | pulling from $TESTTMP/server |
|
157 | 157 | requesting all changes |
|
158 | 158 | adding changesets |
|
159 | 159 | adding manifests |
|
160 | 160 | adding file changes |
|
161 | 161 | added 2 changesets with 3 changes to 3 files |
|
162 | 162 | new changesets 0ead593177f7:b88141481348 |
|
163 | 163 | (run 'hg update' to get a working copy) |
|
164 | 164 | $ grep lfs .hg/requires $TESTTMP/server/.hg/requires |
|
165 | 165 | .hg/requires:lfs |
|
166 | 166 | $TESTTMP/server/.hg/requires:lfs |
|
167 | 167 | |
|
168 | 168 | # Check the blobstore is not yet populated |
|
169 | 169 | $ [ -d .hg/store/lfs/objects ] |
|
170 | 170 | [1] |
|
171 | 171 | |
|
172 | 172 | # Update to the last revision containing the large file |
|
173 | 173 | $ hg update |
|
174 | 174 | 3 files updated, 0 files merged, 0 files removed, 0 files unresolved |
|
175 | 175 | |
|
176 | 176 | # Check the blobstore has been populated on update |
|
177 | 177 | $ find .hg/store/lfs/objects | sort |
|
178 | 178 | .hg/store/lfs/objects |
|
179 | 179 | .hg/store/lfs/objects/f1 |
|
180 | 180 | .hg/store/lfs/objects/f1/1e77c257047a398492d8d6cb9f6acf3aa7c4384bb23080b43546053e183e4b |
|
181 | 181 | |
|
182 | 182 | # Check the contents of the file are fetched from blobstore when requested |
|
183 | 183 | $ hg cat -r . largefile |
|
184 | 184 | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC |
|
185 | 185 | |
|
186 | 186 | # Check the file has been copied in the working copy |
|
187 | 187 | $ cat largefile |
|
188 | 188 | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC |
|
189 | 189 | |
|
190 | 190 | $ cd .. |
|
191 | 191 | |
|
192 | 192 | # Check rename, and switch between large and small files |
|
193 | 193 | |
|
194 | 194 | $ hg init repo3 |
|
195 | 195 | $ cd repo3 |
|
196 | 196 | $ cat >> .hg/hgrc << EOF |
|
197 | 197 | > [lfs] |
|
198 | 198 | > track=size(">10B") |
|
199 | 199 | > EOF |
|
200 | 200 | |
|
201 | 201 | $ echo LONGER-THAN-TEN-BYTES-WILL-TRIGGER-LFS > large |
|
202 | 202 | $ echo SHORTER > small |
|
203 | 203 | $ hg add . -q |
|
204 | 204 | $ hg commit -m 'commit with lfs content' |
|
205 | 205 | |
|
206 | 206 | $ hg files -r . 'set:added()' |
|
207 | 207 | large |
|
208 | 208 | small |
|
209 | 209 | $ hg files -r . 'set:added() & lfs()' |
|
210 | 210 | large |
|
211 | 211 | |
|
212 | 212 | $ hg mv large l |
|
213 | 213 | $ hg mv small s |
|
214 | 214 | $ hg status 'set:removed()' |
|
215 | 215 | R large |
|
216 | 216 | R small |
|
217 | 217 | $ hg status 'set:removed() & lfs()' |
|
218 | 218 | R large |
|
219 | 219 | $ hg commit -m 'renames' |
|
220 | 220 | |
|
221 | $ hg cat -r . l -T '{rawdata}\n' | |
|
222 | version https://git-lfs.github.com/spec/v1 | |
|
223 | oid sha256:66100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e | |
|
224 | size 39 | |
|
225 | x-hg-copy large | |
|
226 | x-hg-copyrev 2c531e0992ff3107c511b53cb82a91b6436de8b2 | |
|
227 | x-is-binary 0 | |
|
228 | ||
|
229 | ||
|
221 | 230 | $ hg files -r . 'set:copied()' |
|
222 | 231 | l |
|
223 | 232 | s |
|
224 | 233 | $ hg files -r . 'set:copied() & lfs()' |
|
225 | 234 | l |
|
226 | 235 | $ hg status --change . 'set:removed()' |
|
227 | 236 | R large |
|
228 | 237 | R small |
|
229 | 238 | $ hg status --change . 'set:removed() & lfs()' |
|
230 | 239 | R large |
|
231 | 240 | |
|
232 | 241 | $ echo SHORT > l |
|
233 | 242 | $ echo BECOME-LARGER-FROM-SHORTER > s |
|
234 | 243 | $ hg commit -m 'large to small, small to large' |
|
235 | 244 | |
|
236 | 245 | $ echo 1 >> l |
|
237 | 246 | $ echo 2 >> s |
|
238 | 247 | $ hg commit -m 'random modifications' |
|
239 | 248 | |
|
240 | 249 | $ echo RESTORE-TO-BE-LARGE > l |
|
241 | 250 | $ echo SHORTER > s |
|
242 | 251 | $ hg commit -m 'switch large and small again' |
|
243 | 252 | |
|
244 | 253 | # Test lfs_files template |
|
245 | 254 | |
|
246 | 255 | $ hg log -r 'all()' -T '{rev} {join(lfs_files, ", ")}\n' |
|
247 | 256 | 0 large |
|
248 | 257 | 1 l, large |
|
249 | 258 | 2 s |
|
250 | 259 | 3 s |
|
251 | 260 | 4 l |
|
252 | 261 | |
|
253 | 262 | # Push and pull the above repo |
|
254 | 263 | |
|
255 | 264 | $ hg --cwd .. init repo4 |
|
256 | 265 | $ hg push ../repo4 |
|
257 | 266 | pushing to ../repo4 |
|
258 | 267 | searching for changes |
|
259 | 268 | adding changesets |
|
260 | 269 | adding manifests |
|
261 | 270 | adding file changes |
|
262 | 271 | added 5 changesets with 10 changes to 4 files |
|
263 | 272 | |
|
264 | 273 | $ hg --cwd .. init repo5 |
|
265 | 274 | $ hg --cwd ../repo5 pull ../repo3 |
|
266 | 275 | pulling from ../repo3 |
|
267 | 276 | requesting all changes |
|
268 | 277 | adding changesets |
|
269 | 278 | adding manifests |
|
270 | 279 | adding file changes |
|
271 | 280 | added 5 changesets with 10 changes to 4 files |
|
272 | 281 | new changesets fd47a419c4f7:5adf850972b9 |
|
273 | 282 | (run 'hg update' to get a working copy) |
|
274 | 283 | |
|
275 | 284 | $ cd .. |
|
276 | 285 | |
|
277 | 286 | # Test clone |
|
278 | 287 | |
|
279 | 288 | $ hg init repo6 |
|
280 | 289 | $ cd repo6 |
|
281 | 290 | $ cat >> .hg/hgrc << EOF |
|
282 | 291 | > [lfs] |
|
283 | 292 | > track=size(">30B") |
|
284 | 293 | > EOF |
|
285 | 294 | |
|
286 | 295 | $ echo LARGE-BECAUSE-IT-IS-MORE-THAN-30-BYTES > large |
|
287 | 296 | $ echo SMALL > small |
|
288 | 297 | $ hg commit -Aqm 'create a lfs file' large small |
|
289 | 298 | $ hg debuglfsupload -r 'all()' -v |
|
290 | 299 | lfs: found 8e92251415339ae9b148c8da89ed5ec665905166a1ab11b09dca8fad83344738 in the local lfs store |
|
291 | 300 | |
|
292 | 301 | $ cd .. |
|
293 | 302 | |
|
294 | 303 | $ hg clone repo6 repo7 |
|
295 | 304 | updating to branch default |
|
296 | 305 | 2 files updated, 0 files merged, 0 files removed, 0 files unresolved |
|
297 | 306 | $ cd repo7 |
|
298 | 307 | $ cat large |
|
299 | 308 | LARGE-BECAUSE-IT-IS-MORE-THAN-30-BYTES |
|
300 | 309 | $ cat small |
|
301 | 310 | SMALL |
|
302 | 311 | |
|
303 | 312 | $ cd .. |
|
304 | 313 | |
|
305 | 314 | $ hg --config extensions.share= share repo7 sharedrepo |
|
306 | 315 | updating working directory |
|
307 | 316 | 2 files updated, 0 files merged, 0 files removed, 0 files unresolved |
|
308 | 317 | $ grep lfs sharedrepo/.hg/requires |
|
309 | 318 | lfs |
|
310 | 319 | |
|
311 | 320 | # Test rename and status |
|
312 | 321 | |
|
313 | 322 | $ hg init repo8 |
|
314 | 323 | $ cd repo8 |
|
315 | 324 | $ cat >> .hg/hgrc << EOF |
|
316 | 325 | > [lfs] |
|
317 | 326 | > track=size(">10B") |
|
318 | 327 | > EOF |
|
319 | 328 | |
|
320 | 329 | $ echo THIS-IS-LFS-BECAUSE-10-BYTES > a1 |
|
321 | 330 | $ echo SMALL > a2 |
|
322 | 331 | $ hg commit -m a -A a1 a2 |
|
323 | 332 | $ hg status |
|
324 | 333 | $ hg mv a1 b1 |
|
325 | 334 | $ hg mv a2 a1 |
|
326 | 335 | $ hg mv b1 a2 |
|
327 | 336 | $ hg commit -m b |
|
328 | 337 | $ hg status |
|
329 | 338 | >>> with open('a2', 'wb') as f: |
|
330 | 339 | ... f.write(b'\1\nSTART-WITH-HG-FILELOG-METADATA') and None |
|
331 | 340 | >>> with open('a1', 'wb') as f: |
|
332 | 341 | ... f.write(b'\1\nMETA\n') and None |
|
333 | 342 | $ hg commit -m meta |
|
334 | 343 | $ hg status |
|
335 | 344 | $ hg log -T '{rev}: {file_copies} | {file_dels} | {file_adds}\n' |
|
336 | 345 | 2: | | |
|
337 | 346 | 1: a1 (a2)a2 (a1) | | |
|
338 | 347 | 0: | | a1 a2 |
|
339 | 348 | |
|
340 | 349 | $ for n in a1 a2; do |
|
341 | 350 | > for r in 0 1 2; do |
|
342 | 351 | > printf '\n%s @ %s\n' $n $r |
|
343 | 352 | > hg debugdata $n $r |
|
344 | 353 | > done |
|
345 | 354 | > done |
|
346 | 355 | |
|
347 | 356 | a1 @ 0 |
|
348 | 357 | version https://git-lfs.github.com/spec/v1 |
|
349 | 358 | oid sha256:5bb8341bee63b3649f222b2215bde37322bea075a30575aa685d8f8d21c77024 |
|
350 | 359 | size 29 |
|
351 | 360 | x-is-binary 0 |
|
352 | 361 | |
|
353 | 362 | a1 @ 1 |
|
354 | 363 | \x01 (esc) |
|
355 | 364 | copy: a2 |
|
356 | 365 | copyrev: 50470ad23cf937b1f4b9f80bfe54df38e65b50d9 |
|
357 | 366 | \x01 (esc) |
|
358 | 367 | SMALL |
|
359 | 368 | |
|
360 | 369 | a1 @ 2 |
|
361 | 370 | \x01 (esc) |
|
362 | 371 | \x01 (esc) |
|
363 | 372 | \x01 (esc) |
|
364 | 373 | META |
|
365 | 374 | |
|
366 | 375 | a2 @ 0 |
|
367 | 376 | SMALL |
|
368 | 377 | |
|
369 | 378 | a2 @ 1 |
|
370 | 379 | version https://git-lfs.github.com/spec/v1 |
|
371 | 380 | oid sha256:5bb8341bee63b3649f222b2215bde37322bea075a30575aa685d8f8d21c77024 |
|
372 | 381 | size 29 |
|
373 | 382 | x-hg-copy a1 |
|
374 | 383 | x-hg-copyrev be23af27908a582af43e5cda209a5a9b319de8d4 |
|
375 | 384 | x-is-binary 0 |
|
376 | 385 | |
|
377 | 386 | a2 @ 2 |
|
378 | 387 | version https://git-lfs.github.com/spec/v1 |
|
379 | 388 | oid sha256:876dadc86a8542f9798048f2c47f51dbf8e4359aed883e8ec80c5db825f0d943 |
|
380 | 389 | size 32 |
|
381 | 390 | x-is-binary 0 |
|
382 | 391 | |
|
383 | 392 | # Verify commit hashes include rename metadata |
|
384 | 393 | |
|
385 | 394 | $ hg log -T '{rev}:{node|short} {desc}\n' |
|
386 | 395 | 2:0fae949de7fa meta |
|
387 | 396 | 1:9cd6bdffdac0 b |
|
388 | 397 | 0:7f96794915f7 a |
|
389 | 398 | |
|
390 | 399 | $ cd .. |
|
391 | 400 | |
|
392 | 401 | # Test bundle |
|
393 | 402 | |
|
394 | 403 | $ hg init repo9 |
|
395 | 404 | $ cd repo9 |
|
396 | 405 | $ cat >> .hg/hgrc << EOF |
|
397 | 406 | > [lfs] |
|
398 | 407 | > track=size(">10B") |
|
399 | 408 | > [diff] |
|
400 | 409 | > git=1 |
|
401 | 410 | > EOF |
|
402 | 411 | |
|
403 | 412 | $ for i in 0 single two three 4; do |
|
404 | 413 | > echo 'THIS-IS-LFS-'$i > a |
|
405 | 414 | > hg commit -m a-$i -A a |
|
406 | 415 | > done |
|
407 | 416 | |
|
408 | 417 | $ hg update 2 -q |
|
409 | 418 | $ echo 'THIS-IS-LFS-2-CHILD' > a |
|
410 | 419 | $ hg commit -m branching -q |
|
411 | 420 | |
|
412 | 421 | $ hg bundle --base 1 bundle.hg -v |
|
413 | 422 | lfs: found 5ab7a3739a5feec94a562d070a14f36dba7cad17e5484a4a89eea8e5f3166888 in the local lfs store |
|
414 | 423 | lfs: found a9c7d1cd6ce2b9bbdf46ed9a862845228717b921c089d0d42e3bcaed29eb612e in the local lfs store |
|
415 | 424 | lfs: found f693890c49c409ec33673b71e53f297681f76c1166daf33b2ad7ebf8b1d3237e in the local lfs store |
|
416 | 425 | lfs: found fda198fea753eb66a252e9856915e1f5cddbe41723bd4b695ece2604ad3c9f75 in the local lfs store |
|
417 | 426 | 4 changesets found |
|
418 | 427 | uncompressed size of bundle content: |
|
419 | 428 | * (changelog) (glob) |
|
420 | 429 | * (manifests) (glob) |
|
421 | 430 | * a (glob) |
|
422 | 431 | $ hg --config extensions.strip= strip -r 2 --no-backup --force -q |
|
423 | 432 | $ hg -R bundle.hg log -p -T '{rev} {desc}\n' a |
|
424 | 433 | 5 branching |
|
425 | 434 | diff --git a/a b/a |
|
426 | 435 | --- a/a |
|
427 | 436 | +++ b/a |
|
428 | 437 | @@ -1,1 +1,1 @@ |
|
429 | 438 | -THIS-IS-LFS-two |
|
430 | 439 | +THIS-IS-LFS-2-CHILD |
|
431 | 440 | |
|
432 | 441 | 4 a-4 |
|
433 | 442 | diff --git a/a b/a |
|
434 | 443 | --- a/a |
|
435 | 444 | +++ b/a |
|
436 | 445 | @@ -1,1 +1,1 @@ |
|
437 | 446 | -THIS-IS-LFS-three |
|
438 | 447 | +THIS-IS-LFS-4 |
|
439 | 448 | |
|
440 | 449 | 3 a-three |
|
441 | 450 | diff --git a/a b/a |
|
442 | 451 | --- a/a |
|
443 | 452 | +++ b/a |
|
444 | 453 | @@ -1,1 +1,1 @@ |
|
445 | 454 | -THIS-IS-LFS-two |
|
446 | 455 | +THIS-IS-LFS-three |
|
447 | 456 | |
|
448 | 457 | 2 a-two |
|
449 | 458 | diff --git a/a b/a |
|
450 | 459 | --- a/a |
|
451 | 460 | +++ b/a |
|
452 | 461 | @@ -1,1 +1,1 @@ |
|
453 | 462 | -THIS-IS-LFS-single |
|
454 | 463 | +THIS-IS-LFS-two |
|
455 | 464 | |
|
456 | 465 | 1 a-single |
|
457 | 466 | diff --git a/a b/a |
|
458 | 467 | --- a/a |
|
459 | 468 | +++ b/a |
|
460 | 469 | @@ -1,1 +1,1 @@ |
|
461 | 470 | -THIS-IS-LFS-0 |
|
462 | 471 | +THIS-IS-LFS-single |
|
463 | 472 | |
|
464 | 473 | 0 a-0 |
|
465 | 474 | diff --git a/a b/a |
|
466 | 475 | new file mode 100644 |
|
467 | 476 | --- /dev/null |
|
468 | 477 | +++ b/a |
|
469 | 478 | @@ -0,0 +1,1 @@ |
|
470 | 479 | +THIS-IS-LFS-0 |
|
471 | 480 | |
|
472 | 481 | $ hg bundle -R bundle.hg --base 1 bundle-again.hg -q |
|
473 | 482 | $ hg -R bundle-again.hg log -p -T '{rev} {desc}\n' a |
|
474 | 483 | 5 branching |
|
475 | 484 | diff --git a/a b/a |
|
476 | 485 | --- a/a |
|
477 | 486 | +++ b/a |
|
478 | 487 | @@ -1,1 +1,1 @@ |
|
479 | 488 | -THIS-IS-LFS-two |
|
480 | 489 | +THIS-IS-LFS-2-CHILD |
|
481 | 490 | |
|
482 | 491 | 4 a-4 |
|
483 | 492 | diff --git a/a b/a |
|
484 | 493 | --- a/a |
|
485 | 494 | +++ b/a |
|
486 | 495 | @@ -1,1 +1,1 @@ |
|
487 | 496 | -THIS-IS-LFS-three |
|
488 | 497 | +THIS-IS-LFS-4 |
|
489 | 498 | |
|
490 | 499 | 3 a-three |
|
491 | 500 | diff --git a/a b/a |
|
492 | 501 | --- a/a |
|
493 | 502 | +++ b/a |
|
494 | 503 | @@ -1,1 +1,1 @@ |
|
495 | 504 | -THIS-IS-LFS-two |
|
496 | 505 | +THIS-IS-LFS-three |
|
497 | 506 | |
|
498 | 507 | 2 a-two |
|
499 | 508 | diff --git a/a b/a |
|
500 | 509 | --- a/a |
|
501 | 510 | +++ b/a |
|
502 | 511 | @@ -1,1 +1,1 @@ |
|
503 | 512 | -THIS-IS-LFS-single |
|
504 | 513 | +THIS-IS-LFS-two |
|
505 | 514 | |
|
506 | 515 | 1 a-single |
|
507 | 516 | diff --git a/a b/a |
|
508 | 517 | --- a/a |
|
509 | 518 | +++ b/a |
|
510 | 519 | @@ -1,1 +1,1 @@ |
|
511 | 520 | -THIS-IS-LFS-0 |
|
512 | 521 | +THIS-IS-LFS-single |
|
513 | 522 | |
|
514 | 523 | 0 a-0 |
|
515 | 524 | diff --git a/a b/a |
|
516 | 525 | new file mode 100644 |
|
517 | 526 | --- /dev/null |
|
518 | 527 | +++ b/a |
|
519 | 528 | @@ -0,0 +1,1 @@ |
|
520 | 529 | +THIS-IS-LFS-0 |
|
521 | 530 | |
|
522 | 531 | $ cd .. |
|
523 | 532 | |
|
524 | 533 | # Test isbinary |
|
525 | 534 | |
|
526 | 535 | $ hg init repo10 |
|
527 | 536 | $ cd repo10 |
|
528 | 537 | $ cat >> .hg/hgrc << EOF |
|
529 | 538 | > [extensions] |
|
530 | 539 | > lfs= |
|
531 | 540 | > [lfs] |
|
532 | 541 | > track=all() |
|
533 | 542 | > EOF |
|
534 | 543 | $ "$PYTHON" <<'EOF' |
|
535 | 544 | > def write(path, content): |
|
536 | 545 | > with open(path, 'wb') as f: |
|
537 | 546 | > f.write(content) |
|
538 | 547 | > write('a', b'\0\0') |
|
539 | 548 | > write('b', b'\1\n') |
|
540 | 549 | > write('c', b'\1\n\0') |
|
541 | 550 | > write('d', b'xx') |
|
542 | 551 | > EOF |
|
543 | 552 | $ hg add a b c d |
|
544 | 553 | $ hg diff --stat |
|
545 | 554 | a | Bin |
|
546 | 555 | b | 1 + |
|
547 | 556 | c | Bin |
|
548 | 557 | d | 1 + |
|
549 | 558 | 4 files changed, 2 insertions(+), 0 deletions(-) |
|
550 | 559 | $ hg commit -m binarytest |
|
551 | 560 | $ cat > $TESTTMP/dumpbinary.py << EOF |
|
552 | 561 | > from mercurial.utils import ( |
|
553 | 562 | > stringutil, |
|
554 | 563 | > ) |
|
555 | 564 | > def reposetup(ui, repo): |
|
556 | 565 | > for n in (b'a', b'b', b'c', b'd'): |
|
557 | 566 | > ui.write((b'%s: binary=%s\n') |
|
558 | 567 | > % (n, stringutil.pprint(repo[b'.'][n].isbinary()))) |
|
559 | 568 | > EOF |
|
560 | 569 | $ hg --config extensions.dumpbinary=$TESTTMP/dumpbinary.py id --trace |
|
561 | 570 | a: binary=True |
|
562 | 571 | b: binary=False |
|
563 | 572 | c: binary=True |
|
564 | 573 | d: binary=False |
|
565 | 574 | b55353847f02 tip |
|
566 | 575 | |
|
567 | 576 | Binary blobs don't need to be present to be skipped in filesets. (And their |
|
568 | 577 | absence doesn't cause an abort.) |
|
569 | 578 | |
|
570 | 579 | $ rm .hg/store/lfs/objects/96/a296d224f285c67bee93c30f8a309157f0daa35dc5b87e410b78630a09cfc7 |
|
571 | 580 | $ rm .hg/store/lfs/objects/92/f76135a4baf4faccb8586a60faf830c2bdfce147cefa188aaf4b790bd01b7e |
|
572 | 581 | |
|
573 | 582 | $ hg files --debug -r . 'set:eol("unix")' --config 'experimental.lfs.disableusercache=True' |
|
574 | 583 | lfs: found c04b5bb1a5b2eb3e9cd4805420dba5a9d133da5b7adeeafb5474c4adae9faa80 in the local lfs store |
|
575 | 584 | 2 b |
|
576 | 585 | lfs: found 5dde896887f6754c9b15bfe3a441ae4806df2fde94001311e08bf110622e0bbe in the local lfs store |
|
577 | 586 | |
|
578 | 587 | $ hg files --debug -r . 'set:binary()' --config 'experimental.lfs.disableusercache=True' |
|
579 | 588 | 2 a |
|
580 | 589 | 3 c |
|
581 | 590 | |
|
582 | 591 | $ cd .. |
|
583 | 592 | |
|
584 | 593 | # Test fctx.cmp fastpath - diff without LFS blobs |
|
585 | 594 | |
|
586 | 595 | $ hg init repo12 |
|
587 | 596 | $ cd repo12 |
|
588 | 597 | $ cat >> .hg/hgrc <<EOF |
|
589 | 598 | > [lfs] |
|
590 | 599 | > threshold=1 |
|
591 | 600 | > EOF |
|
592 | 601 | $ cat > ../patch.diff <<EOF |
|
593 | 602 | > # HG changeset patch |
|
594 | 603 | > 2 |
|
595 | 604 | > |
|
596 | 605 | > diff --git a/a b/a |
|
597 | 606 | > old mode 100644 |
|
598 | 607 | > new mode 100755 |
|
599 | 608 | > EOF |
|
600 | 609 | |
|
601 | 610 | $ for i in 1 2 3; do |
|
602 | 611 | > cp ../repo10/a a |
|
603 | 612 | > if [ $i = 3 ]; then |
|
604 | 613 | > # make a content-only change |
|
605 | 614 | > hg import -q --bypass ../patch.diff |
|
606 | 615 | > hg update -q |
|
607 | 616 | > rm ../patch.diff |
|
608 | 617 | > else |
|
609 | 618 | > echo $i >> a |
|
610 | 619 | > hg commit -m $i -A a |
|
611 | 620 | > fi |
|
612 | 621 | > done |
|
613 | 622 | $ [ -d .hg/store/lfs/objects ] |
|
614 | 623 | |
|
615 | 624 | $ cd .. |
|
616 | 625 | |
|
617 | 626 | $ hg clone repo12 repo13 --noupdate |
|
618 | 627 | $ cd repo13 |
|
619 | 628 | $ hg log --removed -p a -T '{desc}\n' --config diff.nobinary=1 --git |
|
620 | 629 | 2 |
|
621 | 630 | diff --git a/a b/a |
|
622 | 631 | old mode 100644 |
|
623 | 632 | new mode 100755 |
|
624 | 633 | |
|
625 | 634 | 2 |
|
626 | 635 | diff --git a/a b/a |
|
627 | 636 | Binary file a has changed |
|
628 | 637 | |
|
629 | 638 | 1 |
|
630 | 639 | diff --git a/a b/a |
|
631 | 640 | new file mode 100644 |
|
632 | 641 | Binary file a has changed |
|
633 | 642 | |
|
634 | 643 | $ [ -d .hg/store/lfs/objects ] |
|
635 | 644 | [1] |
|
636 | 645 | |
|
637 | 646 | $ cd .. |
|
638 | 647 | |
|
639 | 648 | # Test filter |
|
640 | 649 | |
|
641 | 650 | $ hg init repo11 |
|
642 | 651 | $ cd repo11 |
|
643 | 652 | $ cat >> .hg/hgrc << EOF |
|
644 | 653 | > [lfs] |
|
645 | 654 | > track=(**.a & size(">5B")) | (**.b & !size(">5B")) |
|
646 | 655 | > | (**.c & "path:d" & !"path:d/c.c") | size(">10B") |
|
647 | 656 | > EOF |
|
648 | 657 | |
|
649 | 658 | $ mkdir a |
|
650 | 659 | $ echo aaaaaa > a/1.a |
|
651 | 660 | $ echo a > a/2.a |
|
652 | 661 | $ echo aaaaaa > 1.b |
|
653 | 662 | $ echo a > 2.b |
|
654 | 663 | $ echo a > 1.c |
|
655 | 664 | $ mkdir d |
|
656 | 665 | $ echo a > d/c.c |
|
657 | 666 | $ echo a > d/d.c |
|
658 | 667 | $ echo aaaaaaaaaaaa > x |
|
659 | 668 | $ hg add . -q |
|
660 | 669 | $ hg commit -m files |
|
661 | 670 | |
|
662 | 671 | $ for p in a/1.a a/2.a 1.b 2.b 1.c d/c.c d/d.c x; do |
|
663 | 672 | > if hg debugdata $p 0 2>&1 | grep git-lfs >/dev/null; then |
|
664 | 673 | > echo "${p}: is lfs" |
|
665 | 674 | > else |
|
666 | 675 | > echo "${p}: not lfs" |
|
667 | 676 | > fi |
|
668 | 677 | > done |
|
669 | 678 | a/1.a: is lfs |
|
670 | 679 | a/2.a: not lfs |
|
671 | 680 | 1.b: not lfs |
|
672 | 681 | 2.b: is lfs |
|
673 | 682 | 1.c: not lfs |
|
674 | 683 | d/c.c: not lfs |
|
675 | 684 | d/d.c: is lfs |
|
676 | 685 | x: is lfs |
|
677 | 686 | |
|
678 | 687 | $ cd .. |
|
679 | 688 | |
|
680 | 689 | # Verify the repos |
|
681 | 690 | |
|
682 | 691 | $ cat > $TESTTMP/dumpflog.py << EOF |
|
683 | 692 | > # print raw revision sizes, flags, and hashes for certain files |
|
684 | 693 | > import hashlib |
|
685 | 694 | > from mercurial.node import short |
|
686 | 695 | > from mercurial import ( |
|
687 | 696 | > pycompat, |
|
688 | 697 | > revlog, |
|
689 | 698 | > ) |
|
690 | 699 | > from mercurial.utils import ( |
|
691 | 700 | > stringutil, |
|
692 | 701 | > ) |
|
693 | 702 | > def hash(rawtext): |
|
694 | 703 | > h = hashlib.sha512() |
|
695 | 704 | > h.update(rawtext) |
|
696 | 705 | > return pycompat.sysbytes(h.hexdigest()[:4]) |
|
697 | 706 | > def reposetup(ui, repo): |
|
698 | 707 | > # these 2 files are interesting |
|
699 | 708 | > for name in [b'l', b's']: |
|
700 | 709 | > fl = repo.file(name) |
|
701 | 710 | > if len(fl) == 0: |
|
702 | 711 | > continue |
|
703 | 712 | > sizes = [fl._revlog.rawsize(i) for i in fl] |
|
704 | 713 | > texts = [fl.rawdata(i) for i in fl] |
|
705 | 714 | > flags = [int(fl._revlog.flags(i)) for i in fl] |
|
706 | 715 | > hashes = [hash(t) for t in texts] |
|
707 | 716 | > pycompat.stdout.write(b' %s: rawsizes=%r flags=%r hashes=%s\n' |
|
708 | 717 | > % (name, sizes, flags, stringutil.pprint(hashes))) |
|
709 | 718 | > EOF |
|
710 | 719 | |
|
711 | 720 | $ for i in client client2 server repo3 repo4 repo5 repo6 repo7 repo8 repo9 \ |
|
712 | 721 | > repo10; do |
|
713 | 722 | > echo 'repo:' $i |
|
714 | 723 | > hg --cwd $i verify --config extensions.dumpflog=$TESTTMP/dumpflog.py -q |
|
715 | 724 | > done |
|
716 | 725 | repo: client |
|
717 | 726 | repo: client2 |
|
718 | 727 | repo: server |
|
719 | 728 | repo: repo3 |
|
720 | 729 | l: rawsizes=[211, 6, 8, 141] flags=[8192, 0, 0, 8192] hashes=['d2b8', '948c', 'cc88', '724d'] |
|
721 | 730 | s: rawsizes=[74, 141, 141, 8] flags=[0, 8192, 8192, 0] hashes=['3c80', 'fce0', '874a', '826b'] |
|
722 | 731 | repo: repo4 |
|
723 | 732 | l: rawsizes=[211, 6, 8, 141] flags=[8192, 0, 0, 8192] hashes=['d2b8', '948c', 'cc88', '724d'] |
|
724 | 733 | s: rawsizes=[74, 141, 141, 8] flags=[0, 8192, 8192, 0] hashes=['3c80', 'fce0', '874a', '826b'] |
|
725 | 734 | repo: repo5 |
|
726 | 735 | l: rawsizes=[211, 6, 8, 141] flags=[8192, 0, 0, 8192] hashes=['d2b8', '948c', 'cc88', '724d'] |
|
727 | 736 | s: rawsizes=[74, 141, 141, 8] flags=[0, 8192, 8192, 0] hashes=['3c80', 'fce0', '874a', '826b'] |
|
728 | 737 | repo: repo6 |
|
729 | 738 | repo: repo7 |
|
730 | 739 | repo: repo8 |
|
731 | 740 | repo: repo9 |
|
732 | 741 | repo: repo10 |
|
733 | 742 | |
|
734 | 743 | repo13 doesn't have any cached lfs files and its source never pushed its |
|
735 | 744 | files. Therefore, the files don't exist in the remote store. Use the files in |
|
736 | 745 | the user cache. |
|
737 | 746 | |
|
738 | 747 | $ test -d $TESTTMP/repo13/.hg/store/lfs/objects |
|
739 | 748 | [1] |
|
740 | 749 | |
|
741 | 750 | $ hg --config extensions.share= share repo13 repo14 |
|
742 | 751 | updating working directory |
|
743 | 752 | 1 files updated, 0 files merged, 0 files removed, 0 files unresolved |
|
744 | 753 | $ hg -R repo14 -q verify |
|
745 | 754 | |
|
746 | 755 | $ hg clone repo13 repo15 |
|
747 | 756 | updating to branch default |
|
748 | 757 | 1 files updated, 0 files merged, 0 files removed, 0 files unresolved |
|
749 | 758 | $ hg -R repo15 -q verify |
|
750 | 759 | |
|
751 | 760 | If the source repo doesn't have the blob (maybe it was pulled or cloned with |
|
752 | 761 | --noupdate), the blob is still accessible via the global cache to send to the |
|
753 | 762 | remote store. |
|
754 | 763 | |
|
755 | 764 | $ rm -rf $TESTTMP/repo15/.hg/store/lfs |
|
756 | 765 | $ hg init repo16 |
|
757 | 766 | $ hg -R repo15 push repo16 |
|
758 | 767 | pushing to repo16 |
|
759 | 768 | searching for changes |
|
760 | 769 | adding changesets |
|
761 | 770 | adding manifests |
|
762 | 771 | adding file changes |
|
763 | 772 | added 3 changesets with 2 changes to 1 files |
|
764 | 773 | $ hg -R repo15 -q verify |
|
765 | 774 | |
|
766 | 775 | Test damaged file scenarios. (This also damages the usercache because of the |
|
767 | 776 | hardlinks.) |
|
768 | 777 | |
|
769 | 778 | $ echo 'damage' >> repo5/.hg/store/lfs/objects/66/100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e |
|
770 | 779 | |
|
771 | 780 | Repo with damaged lfs objects in any revision will fail verification. |
|
772 | 781 | |
|
773 | 782 | $ hg -R repo5 verify |
|
774 | 783 | checking changesets |
|
775 | 784 | checking manifests |
|
776 | 785 | crosschecking files in changesets and manifests |
|
777 | 786 | checking files |
|
778 | 787 | l@1: unpacking 46a2f24864bc: integrity check failed on data/l.i:0 |
|
779 | 788 | large@0: unpacking 2c531e0992ff: integrity check failed on data/large.i:0 |
|
780 | 789 | checked 5 changesets with 10 changes to 4 files |
|
781 | 790 | 2 integrity errors encountered! |
|
782 | 791 | (first damaged changeset appears to be 0) |
|
783 | 792 | [1] |
|
784 | 793 | |
|
785 | 794 | Updates work after cloning a damaged repo, if the damaged lfs objects aren't in |
|
786 | 795 | the update destination. Those objects won't be added to the new repo's store |
|
787 | 796 | because they aren't accessed. |
|
788 | 797 | |
|
789 | 798 | $ hg clone -v repo5 fromcorrupt |
|
790 | 799 | updating to branch default |
|
791 | 800 | resolving manifests |
|
792 | 801 | getting l |
|
793 | 802 | lfs: found 22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b in the usercache |
|
794 | 803 | getting s |
|
795 | 804 | 2 files updated, 0 files merged, 0 files removed, 0 files unresolved |
|
796 | 805 | $ test -f fromcorrupt/.hg/store/lfs/objects/66/100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e |
|
797 | 806 | [1] |
|
798 | 807 | |
|
799 | Verify will not try to download lfs blobs, if told not to process lfs content | |
|
808 | Verify will not try to download lfs blobs, if told not to process lfs content. | |
|
809 | The extension makes sure that the filelog.renamed() path is taken on a missing | |
|
810 | blob, and the output shows that it isn't fetched. | |
|
800 | 811 | |
|
801 | $ hg -R fromcorrupt --config lfs.usercache=emptycache verify -v --no-lfs | |
|
812 | $ cat > $TESTTMP/lfsrename.py <<EOF | |
|
813 | > from mercurial import ( | |
|
814 | > exthelper, | |
|
815 | > ) | |
|
816 | > | |
|
817 | > from hgext.lfs import ( | |
|
818 | > pointer, | |
|
819 | > wrapper, | |
|
820 | > ) | |
|
821 | > | |
|
822 | > eh = exthelper.exthelper() | |
|
823 | > uisetup = eh.finaluisetup | |
|
824 | > | |
|
825 | > @eh.wrapfunction(wrapper, b'filelogrenamed') | |
|
826 | > def filelogrenamed(orig, orig1, self, node): | |
|
827 | > ret = orig(orig1, self, node) | |
|
828 | > if wrapper._islfs(self._revlog, node) and ret: | |
|
829 | > rawtext = self._revlog.rawdata(node) | |
|
830 | > metadata = pointer.deserialize(rawtext) | |
|
831 | > print('lfs blob %s renamed %s -> %s' | |
|
832 | > % (metadata[b'oid'], ret[0], self._revlog.filename)) | |
|
833 | > return ret | |
|
834 | > EOF | |
|
835 | ||
|
836 | $ hg -R fromcorrupt --config lfs.usercache=emptycache verify -v --no-lfs \ | |
|
837 | > --config extensions.x=$TESTTMP/lfsrename.py | |
|
802 | 838 | repository uses revlog format 1 |
|
803 | 839 | checking changesets |
|
804 | 840 | checking manifests |
|
805 | 841 | crosschecking files in changesets and manifests |
|
806 | 842 | checking files |
|
807 | 843 | lfs: found 22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b in the local lfs store |
|
844 | lfs blob sha256:66100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e renamed large -> l | |
|
808 | 845 | checked 5 changesets with 10 changes to 4 files |
|
809 | 846 | |
|
810 | 847 | Verify will not try to download lfs blobs, if told not to by the config option |
|
811 | 848 | |
|
812 | 849 | $ hg -R fromcorrupt --config lfs.usercache=emptycache verify -v \ |
|
813 | > --config verify.skipflags=8192 | |
|
850 | > --config verify.skipflags=8192 \ | |
|
851 | > --config extensions.x=$TESTTMP/lfsrename.py | |
|
814 | 852 | repository uses revlog format 1 |
|
815 | 853 | checking changesets |
|
816 | 854 | checking manifests |
|
817 | 855 | crosschecking files in changesets and manifests |
|
818 | 856 | checking files |
|
819 | 857 | lfs: found 22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b in the local lfs store |
|
858 | lfs blob sha256:66100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e renamed large -> l | |
|
820 | 859 | checked 5 changesets with 10 changes to 4 files |
|
821 | 860 | |
|
822 | 861 |
|
|
823 | 862 | present. Bypass the corrupted usercache to show that verify works when fed by |
|
824 | 863 | the (uncorrupted) remote store. |
|
825 | 864 | |
|
826 | 865 | $ hg -R fromcorrupt --config lfs.usercache=emptycache verify -v |
|
827 | 866 | repository uses revlog format 1 |
|
828 | 867 | checking changesets |
|
829 | 868 | checking manifests |
|
830 | 869 | crosschecking files in changesets and manifests |
|
831 | 870 | checking files |
|
832 | 871 | lfs: adding 66100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e to the usercache |
|
833 | 872 | lfs: found 66100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e in the local lfs store |
|
834 | 873 | lfs: found 22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b in the local lfs store |
|
835 | 874 | lfs: found 66100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e in the local lfs store |
|
836 | 875 | lfs: adding 89b6070915a3d573ff3599d1cda305bc5e38549b15c4847ab034169da66e1ca8 to the usercache |
|
837 | 876 | lfs: found 89b6070915a3d573ff3599d1cda305bc5e38549b15c4847ab034169da66e1ca8 in the local lfs store |
|
838 | 877 | lfs: adding b1a6ea88da0017a0e77db139a54618986e9a2489bee24af9fe596de9daac498c to the usercache |
|
839 | 878 | lfs: found b1a6ea88da0017a0e77db139a54618986e9a2489bee24af9fe596de9daac498c in the local lfs store |
|
840 | 879 | checked 5 changesets with 10 changes to 4 files |
|
841 | 880 | |
|
842 | 881 | Verify will not copy/link a corrupted file from the usercache into the local |
|
843 | 882 | store, and poison it. (The verify with a good remote now works.) |
|
844 | 883 | |
|
845 | 884 | $ rm -r fromcorrupt/.hg/store/lfs/objects/66/100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e |
|
846 | 885 | $ hg -R fromcorrupt verify -v |
|
847 | 886 | repository uses revlog format 1 |
|
848 | 887 | checking changesets |
|
849 | 888 | checking manifests |
|
850 | 889 | crosschecking files in changesets and manifests |
|
851 | 890 | checking files |
|
852 | 891 | l@1: unpacking 46a2f24864bc: integrity check failed on data/l.i:0 |
|
853 | 892 | lfs: found 22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b in the local lfs store |
|
854 | 893 | large@0: unpacking 2c531e0992ff: integrity check failed on data/large.i:0 |
|
855 | 894 | lfs: found 89b6070915a3d573ff3599d1cda305bc5e38549b15c4847ab034169da66e1ca8 in the local lfs store |
|
856 | 895 | lfs: found b1a6ea88da0017a0e77db139a54618986e9a2489bee24af9fe596de9daac498c in the local lfs store |
|
857 | 896 | checked 5 changesets with 10 changes to 4 files |
|
858 | 897 | 2 integrity errors encountered! |
|
859 | 898 | (first damaged changeset appears to be 0) |
|
860 | 899 | [1] |
|
861 | 900 | $ hg -R fromcorrupt --config lfs.usercache=emptycache verify -v |
|
862 | 901 | repository uses revlog format 1 |
|
863 | 902 | checking changesets |
|
864 | 903 | checking manifests |
|
865 | 904 | crosschecking files in changesets and manifests |
|
866 | 905 | checking files |
|
867 | 906 | lfs: found 66100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e in the usercache |
|
868 | 907 | lfs: found 22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b in the local lfs store |
|
869 | 908 | lfs: found 66100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e in the local lfs store |
|
870 | 909 | lfs: found 89b6070915a3d573ff3599d1cda305bc5e38549b15c4847ab034169da66e1ca8 in the local lfs store |
|
871 | 910 | lfs: found b1a6ea88da0017a0e77db139a54618986e9a2489bee24af9fe596de9daac498c in the local lfs store |
|
872 | 911 | checked 5 changesets with 10 changes to 4 files |
|
873 | 912 | |
|
874 | 913 | Damaging a file required by the update destination fails the update. |
|
875 | 914 | |
|
876 | 915 |
|
|
877 | 916 | $ hg --config lfs.usercache=emptycache clone -v repo5 fromcorrupt2 |
|
878 | 917 | updating to branch default |
|
879 | 918 | resolving manifests |
|
880 | 919 | abort: corrupt remote lfs object: 22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b |
|
881 | 920 | [255] |
|
882 | 921 | |
|
883 | 922 | A corrupted lfs blob is not transferred from a file://remotestore to the |
|
884 | 923 | usercache or local store. |
|
885 | 924 | |
|
886 | 925 | $ test -f emptycache/22/f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b |
|
887 | 926 | [1] |
|
888 | 927 | $ test -f fromcorrupt2/.hg/store/lfs/objects/22/f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b |
|
889 | 928 | [1] |
|
890 | 929 | |
|
891 | 930 | $ hg -R fromcorrupt2 verify |
|
892 | 931 | checking changesets |
|
893 | 932 | checking manifests |
|
894 | 933 | crosschecking files in changesets and manifests |
|
895 | 934 | checking files |
|
896 | 935 | l@1: unpacking 46a2f24864bc: integrity check failed on data/l.i:0 |
|
897 | 936 | large@0: unpacking 2c531e0992ff: integrity check failed on data/large.i:0 |
|
898 | 937 | checked 5 changesets with 10 changes to 4 files |
|
899 | 938 | 2 integrity errors encountered! |
|
900 | 939 | (first damaged changeset appears to be 0) |
|
901 | 940 | [1] |
|
902 | 941 | |
|
903 | 942 | Corrupt local files are not sent upstream. (The alternate dummy remote |
|
904 | 943 | avoids the corrupt lfs object in the original remote.) |
|
905 | 944 | |
|
906 | 945 | $ mkdir $TESTTMP/dummy-remote2 |
|
907 | 946 | $ hg init dest |
|
908 | 947 | $ hg -R fromcorrupt2 --config lfs.url=file:///$TESTTMP/dummy-remote2 push -v dest |
|
909 | 948 | pushing to dest |
|
910 | 949 | searching for changes |
|
911 | 950 | lfs: found 22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b in the local lfs store |
|
912 | 951 | abort: detected corrupt lfs object: 66100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e |
|
913 | 952 | (run hg verify) |
|
914 | 953 | [255] |
|
915 | 954 | |
|
916 | 955 | $ hg -R fromcorrupt2 --config lfs.url=file:///$TESTTMP/dummy-remote2 verify -v |
|
917 | 956 | repository uses revlog format 1 |
|
918 | 957 | checking changesets |
|
919 | 958 | checking manifests |
|
920 | 959 | crosschecking files in changesets and manifests |
|
921 | 960 | checking files |
|
922 | 961 | l@1: unpacking 46a2f24864bc: integrity check failed on data/l.i:0 |
|
923 | 962 | lfs: found 22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b in the local lfs store |
|
924 | 963 | large@0: unpacking 2c531e0992ff: integrity check failed on data/large.i:0 |
|
925 | 964 | lfs: found 89b6070915a3d573ff3599d1cda305bc5e38549b15c4847ab034169da66e1ca8 in the local lfs store |
|
926 | 965 | lfs: found b1a6ea88da0017a0e77db139a54618986e9a2489bee24af9fe596de9daac498c in the local lfs store |
|
927 | 966 | checked 5 changesets with 10 changes to 4 files |
|
928 | 967 | 2 integrity errors encountered! |
|
929 | 968 | (first damaged changeset appears to be 0) |
|
930 | 969 | [1] |
|
931 | 970 | |
|
932 | 971 | $ cat $TESTTMP/dummy-remote2/22/f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b | $TESTDIR/f --sha256 |
|
933 | 972 | sha256=22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b |
|
934 | 973 | $ cat fromcorrupt2/.hg/store/lfs/objects/22/f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b | $TESTDIR/f --sha256 |
|
935 | 974 | sha256=22f66a3fc0b9bf3f012c814303995ec07099b3a9ce02a7af84b5970811074a3b |
|
936 | 975 | $ test -f $TESTTMP/dummy-remote2/66/100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e |
|
937 | 976 | [1] |
|
938 | 977 | |
|
939 | 978 | Accessing a corrupt file will complain |
|
940 | 979 | |
|
941 | 980 | $ hg --cwd fromcorrupt2 cat -r 0 large |
|
942 | 981 | abort: integrity check failed on data/large.i:0! |
|
943 | 982 | [255] |
|
944 | 983 | |
|
945 | 984 |
|
|
946 | 985 | predicate on the command line will override whatever is configured globally and |
|
947 | 986 | locally, and ensures everything converts to a regular file. For lfs -> normal, |
|
948 | 987 |
|
|
949 | 988 | |
|
950 | 989 |
|
|
951 | 990 | > convert repo8 convert_normal |
|
952 | 991 | initializing destination convert_normal repository |
|
953 | 992 | scanning source... |
|
954 | 993 | sorting... |
|
955 | 994 | converting... |
|
956 | 995 | 2 a |
|
957 | 996 | 1 b |
|
958 | 997 | 0 meta |
|
959 | 998 |
|
|
960 | 999 | [1] |
|
961 | 1000 |
|
|
962 | 1001 | THIS-IS-LFS-BECAUSE-10-BYTES |
|
963 | 1002 | |
|
964 | 1003 | $ hg --config extensions.convert= --config lfs.threshold=10B \ |
|
965 | 1004 | > convert convert_normal convert_lfs |
|
966 | 1005 | initializing destination convert_lfs repository |
|
967 | 1006 | scanning source... |
|
968 | 1007 | sorting... |
|
969 | 1008 | converting... |
|
970 | 1009 | 2 a |
|
971 | 1010 | 1 b |
|
972 | 1011 | 0 meta |
|
973 | 1012 | |
|
974 | 1013 |
|
|
975 | 1014 | version https://git-lfs.github.com/spec/v1 |
|
976 | 1015 | oid sha256:5bb8341bee63b3649f222b2215bde37322bea075a30575aa685d8f8d21c77024 |
|
977 | 1016 | size 29 |
|
978 | 1017 | x-is-binary 0 |
|
979 | 1018 | $ hg --cwd convert_lfs debugdata a1 0 |
|
980 | 1019 | version https://git-lfs.github.com/spec/v1 |
|
981 | 1020 | oid sha256:5bb8341bee63b3649f222b2215bde37322bea075a30575aa685d8f8d21c77024 |
|
982 | 1021 | size 29 |
|
983 | 1022 | x-is-binary 0 |
|
984 | 1023 |
|
|
985 | 1024 | version=https://git-lfs.github.com/spec/v1 |
|
986 | 1025 | oid=sha256:5bb8341bee63b3649f222b2215bde37322bea075a30575aa685d8f8d21c77024 |
|
987 | 1026 | size=29 |
|
988 | 1027 | x-is-binary=0 |
|
989 | 1028 | $ hg --cwd convert_lfs log -r 0 \ |
|
990 | 1029 |
> -T '{lfs_files % "{get(lfspointer, "oid")}\n"}{lfs_files % "{lfspointer.oid}\n"} |
|
991 | 1030 | sha256:5bb8341bee63b3649f222b2215bde37322bea075a30575aa685d8f8d21c77024 |
|
992 | 1031 | sha256:5bb8341bee63b3649f222b2215bde37322bea075a30575aa685d8f8d21c77024 |
|
993 | 1032 |
$ hg --cwd convert_lfs log -r 0 -T '{lfs_files % "{lfspointer}\n"} |
|
994 | 1033 | version=https://git-lfs.github.com/spec/v1 oid=sha256:5bb8341bee63b3649f222b2215bde37322bea075a30575aa685d8f8d21c77024 size=29 x-is-binary=0 |
|
995 | 1034 | $ hg --cwd convert_lfs \ |
|
996 | 1035 |
> log -r |
|
997 | 1036 | 0: a1: 5bb8341bee63b3649f222b2215bde37322bea075a30575aa685d8f8d21c77024 |
|
998 | 1037 | 1: a2: 5bb8341bee63b3649f222b2215bde37322bea075a30575aa685d8f8d21c77024 |
|
999 | 1038 | 2: a2: 876dadc86a8542f9798048f2c47f51dbf8e4359aed883e8ec80c5db825f0d943 |
|
1000 | 1039 | |
|
1001 | 1040 |
|
|
1002 | 1041 | lfs |
|
1003 | 1042 | |
|
1004 | 1043 | The hashes in all stages of the conversion are unchanged. |
|
1005 | 1044 | |
|
1006 | 1045 |
|
|
1007 | 1046 | 0fae949de7fa |
|
1008 | 1047 | 9cd6bdffdac0 |
|
1009 | 1048 | 7f96794915f7 |
|
1010 | 1049 |
|
|
1011 | 1050 | 0fae949de7fa |
|
1012 | 1051 | 9cd6bdffdac0 |
|
1013 | 1052 | 7f96794915f7 |
|
1014 | 1053 |
|
|
1015 | 1054 | 0fae949de7fa |
|
1016 | 1055 | 9cd6bdffdac0 |
|
1017 | 1056 | 7f96794915f7 |
|
1018 | 1057 | |
|
1019 | 1058 | This convert is trickier, because it contains deleted files (via `hg mv`) |
|
1020 | 1059 | |
|
1021 | 1060 | $ hg --config extensions.convert= --config lfs.threshold=1000M \ |
|
1022 | 1061 | > convert repo3 convert_normal2 |
|
1023 | 1062 | initializing destination convert_normal2 repository |
|
1024 | 1063 | scanning source... |
|
1025 | 1064 | sorting... |
|
1026 | 1065 | converting... |
|
1027 | 1066 | 4 commit with lfs content |
|
1028 | 1067 | 3 renames |
|
1029 | 1068 | 2 large to small, small to large |
|
1030 | 1069 | 1 random modifications |
|
1031 | 1070 | 0 switch large and small again |
|
1032 | 1071 |
|
|
1033 | 1072 | [1] |
|
1034 | 1073 | $ hg --cwd convert_normal2 debugdata large 0 |
|
1035 | 1074 | LONGER-THAN-TEN-BYTES-WILL-TRIGGER-LFS |
|
1036 | 1075 | |
|
1037 | 1076 | $ hg --config extensions.convert= --config lfs.threshold=10B \ |
|
1038 | 1077 | > convert convert_normal2 convert_lfs2 |
|
1039 | 1078 | initializing destination convert_lfs2 repository |
|
1040 | 1079 | scanning source... |
|
1041 | 1080 | sorting... |
|
1042 | 1081 | converting... |
|
1043 | 1082 | 4 commit with lfs content |
|
1044 | 1083 | 3 renames |
|
1045 | 1084 | 2 large to small, small to large |
|
1046 | 1085 | 1 random modifications |
|
1047 | 1086 | 0 switch large and small again |
|
1048 | 1087 |
|
|
1049 | 1088 | lfs |
|
1050 | 1089 | $ hg --cwd convert_lfs2 debugdata large 0 |
|
1051 | 1090 | version https://git-lfs.github.com/spec/v1 |
|
1052 | 1091 | oid sha256:66100b384bf761271b407d79fc30cdd0554f3b2c5d944836e936d584b88ce88e |
|
1053 | 1092 | size 39 |
|
1054 | 1093 | x-is-binary 0 |
|
1055 | 1094 | |
|
1056 | 1095 | Committing deleted files works: |
|
1057 | 1096 | |
|
1058 | 1097 | $ hg init $TESTTMP/repo-del |
|
1059 | 1098 | $ cd $TESTTMP/repo-del |
|
1060 | 1099 | $ echo 1 > A |
|
1061 | 1100 |
$ hg commit -m 'add A |
|
1062 | 1101 | $ hg rm A |
|
1063 | 1102 |
$ hg commit -m ' |
|
1064 | 1103 | |
|
1065 | 1104 | Bad .hglfs files will block the commit with a useful message |
|
1066 | 1105 | |
|
1067 | 1106 | $ cat > .hglfs << EOF |
|
1068 | 1107 | > [track] |
|
1069 | 1108 | > **.test = size(">5B") |
|
1070 | 1109 | > bad file ... no commit |
|
1071 | 1110 | > EOF |
|
1072 | 1111 | |
|
1073 | 1112 | $ echo x > file.txt |
|
1074 | 1113 | $ hg ci -Aqm 'should fail' |
|
1075 | 1114 | hg: parse error at .hglfs:3: bad file ... no commit |
|
1076 | 1115 | [255] |
|
1077 | 1116 | |
|
1078 | 1117 | $ cat > .hglfs << EOF |
|
1079 | 1118 | > [track] |
|
1080 | 1119 | > **.test = size(">5B") |
|
1081 | 1120 | > ** = nonexistent() |
|
1082 | 1121 | > EOF |
|
1083 | 1122 | |
|
1084 | 1123 | $ hg ci -Aqm 'should fail' |
|
1085 | 1124 | abort: parse error in .hglfs: unknown identifier: nonexistent |
|
1086 | 1125 | [255] |
|
1087 | 1126 | |
|
1088 | 1127 | '**' works out to mean all files. |
|
1089 | 1128 | |
|
1090 | 1129 | $ cat > .hglfs << EOF |
|
1091 | 1130 | > [track] |
|
1092 | 1131 | > path:.hglfs = none() |
|
1093 | 1132 | > **.test = size(">5B") |
|
1094 | 1133 | > **.exclude = none() |
|
1095 | 1134 | > ** = size(">10B") |
|
1096 | 1135 | > EOF |
|
1097 | 1136 | |
|
1098 | 1137 | The LFS policy takes effect without tracking the .hglfs file |
|
1099 | 1138 | |
|
1100 | 1139 | $ echo 'largefile' > lfs.test |
|
1101 | 1140 | $ echo '012345678901234567890' > nolfs.exclude |
|
1102 | 1141 | $ echo '01234567890123456' > lfs.catchall |
|
1103 | 1142 | $ hg add * |
|
1104 | 1143 | $ hg ci -qm 'before add .hglfs' |
|
1105 | 1144 | $ hg log -r . -T '{rev}: {lfs_files % "{file}: {lfsoid}\n"}\n' |
|
1106 | 1145 | 2: lfs.catchall: d4ec46c2869ba22eceb42a729377432052d9dd75d82fc40390ebaadecee87ee9 |
|
1107 | 1146 | lfs.test: 5489e6ced8c36a7b267292bde9fd5242a5f80a7482e8f23fa0477393dfaa4d6c |
|
1108 | 1147 | |
|
1109 | 1148 | The .hglfs file works when tracked |
|
1110 | 1149 | |
|
1111 | 1150 | $ echo 'largefile2' > lfs.test |
|
1112 | 1151 | $ echo '012345678901234567890a' > nolfs.exclude |
|
1113 | 1152 | $ echo '01234567890123456a' > lfs.catchall |
|
1114 | 1153 | $ hg ci -Aqm 'after adding .hglfs' |
|
1115 | 1154 | $ hg log -r . -T '{rev}: {lfs_files % "{file}: {lfsoid}\n"}\n' |
|
1116 | 1155 | 3: lfs.catchall: 31f43b9c62b540126b0ad5884dc013d21a61c9329b77de1fceeae2fc58511573 |
|
1117 | 1156 | lfs.test: 8acd23467967bc7b8cc5a280056589b0ba0b17ff21dbd88a7b6474d6290378a6 |
|
1118 | 1157 | |
|
1119 | 1158 | The LFS policy stops when the .hglfs is gone |
|
1120 | 1159 | |
|
1121 | 1160 | $ mv .hglfs .hglfs_ |
|
1122 | 1161 | $ echo 'largefile3' > lfs.test |
|
1123 | 1162 | $ echo '012345678901234567890abc' > nolfs.exclude |
|
1124 | 1163 | $ echo '01234567890123456abc' > lfs.catchall |
|
1125 | 1164 | $ hg ci -qm 'file test' -X .hglfs |
|
1126 | 1165 | $ hg log -r . -T '{rev}: {lfs_files % "{file}: {lfsoid}\n"}\n' |
|
1127 | 1166 | 4: |
|
1128 | 1167 | |
|
1129 | 1168 | $ mv .hglfs_ .hglfs |
|
1130 | 1169 | $ echo '012345678901234567890abc' > lfs.test |
|
1131 | 1170 | $ hg ci -m 'back to lfs' |
|
1132 | 1171 | $ hg rm lfs.test |
|
1133 | 1172 | $ hg ci -qm 'remove lfs' |
|
1134 | 1173 | |
|
1135 | 1174 | {lfs_files} will list deleted files too |
|
1136 | 1175 | |
|
1137 | 1176 | $ hg log -T "{lfs_files % '{rev} {file}: {lfspointer.oid}\n'}" |
|
1138 | 1177 | 6 lfs.test: |
|
1139 | 1178 | 5 lfs.test: sha256:43f8f41171b6f62a6b61ba4ce98a8a6c1649240a47ebafd43120aa215ac9e7f6 |
|
1140 | 1179 | 3 lfs.catchall: sha256:31f43b9c62b540126b0ad5884dc013d21a61c9329b77de1fceeae2fc58511573 |
|
1141 | 1180 | 3 lfs.test: sha256:8acd23467967bc7b8cc5a280056589b0ba0b17ff21dbd88a7b6474d6290378a6 |
|
1142 | 1181 | 2 lfs.catchall: sha256:d4ec46c2869ba22eceb42a729377432052d9dd75d82fc40390ebaadecee87ee9 |
|
1143 | 1182 | 2 lfs.test: sha256:5489e6ced8c36a7b267292bde9fd5242a5f80a7482e8f23fa0477393dfaa4d6c |
|
1144 | 1183 | |
|
1145 | 1184 | $ hg log -r 'file("set:lfs()")' -T '{rev} {join(lfs_files, ", ")}\n' |
|
1146 | 1185 | 2 lfs.catchall, lfs.test |
|
1147 | 1186 | 3 lfs.catchall, lfs.test |
|
1148 | 1187 | 5 lfs.test |
|
1149 | 1188 | 6 lfs.test |
|
1150 | 1189 | |
|
1151 | 1190 | $ cd .. |
|
1152 | 1191 | |
|
1153 | 1192 | Unbundling adds a requirement to a non-lfs repo, if necessary. |
|
1154 | 1193 | |
|
1155 | 1194 | $ hg bundle -R $TESTTMP/repo-del -qr 0 --base null nolfs.hg |
|
1156 | 1195 | $ hg bundle -R convert_lfs2 -qr tip --base null lfs.hg |
|
1157 | 1196 | $ hg init unbundle |
|
1158 | 1197 | $ hg pull -R unbundle -q nolfs.hg |
|
1159 | 1198 | $ grep lfs unbundle/.hg/requires |
|
1160 | 1199 | [1] |
|
1161 | 1200 | $ hg pull -R unbundle -q lfs.hg |
|
1162 | 1201 | $ grep lfs unbundle/.hg/requires |
|
1163 | 1202 | lfs |
|
1164 | 1203 | |
|
1165 | 1204 | $ hg init no_lfs |
|
1166 | 1205 | $ cat >> no_lfs/.hg/hgrc <<EOF |
|
1167 | 1206 | > [experimental] |
|
1168 | 1207 | > changegroup3 = True |
|
1169 | 1208 | > [extensions] |
|
1170 | 1209 | > lfs=! |
|
1171 | 1210 | > EOF |
|
1172 | 1211 | $ cp -R no_lfs no_lfs2 |
|
1173 | 1212 | |
|
1174 | 1213 | Pushing from a local lfs repo to a local repo without an lfs requirement and |
|
1175 | 1214 | with lfs disabled, fails. |
|
1176 | 1215 | |
|
1177 | 1216 | $ hg push -R convert_lfs2 no_lfs |
|
1178 | 1217 | pushing to no_lfs |
|
1179 | 1218 | abort: required features are not supported in the destination: lfs |
|
1180 | 1219 | [255] |
|
1181 | 1220 | $ grep lfs no_lfs/.hg/requires |
|
1182 | 1221 | [1] |
|
1183 | 1222 | |
|
1184 | 1223 | Pulling from a local lfs repo to a local repo without an lfs requirement and |
|
1185 | 1224 | with lfs disabled, fails. |
|
1186 | 1225 | |
|
1187 | 1226 | $ hg pull -R no_lfs2 convert_lfs2 |
|
1188 | 1227 | pulling from convert_lfs2 |
|
1189 | 1228 | abort: required features are not supported in the destination: lfs |
|
1190 | 1229 | [255] |
|
1191 | 1230 | $ grep lfs no_lfs2/.hg/requires |
|
1192 | 1231 | [1] |
General Comments 0
You need to be logged in to leave comments.
Login now