Show More
@@ -1,316 +1,318 b'' | |||
|
1 | 1 | # narrowspec.py - methods for working with a narrow view of a repository |
|
2 | 2 | # |
|
3 | 3 | # Copyright 2017 Google, 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 errno |
|
11 | 11 | |
|
12 | 12 | from .i18n import _ |
|
13 | 13 | from . import ( |
|
14 | 14 | error, |
|
15 | 15 | match as matchmod, |
|
16 | 16 | merge, |
|
17 | 17 | repository, |
|
18 | scmutil, | |
|
18 | 19 | sparse, |
|
19 | 20 | util, |
|
20 | 21 | ) |
|
21 | 22 | |
|
22 | 23 | # The file in .hg/store/ that indicates which paths exit in the store |
|
23 | 24 | FILENAME = 'narrowspec' |
|
24 | 25 | # The file in .hg/ that indicates which paths exit in the dirstate |
|
25 | 26 | DIRSTATE_FILENAME = 'narrowspec.dirstate' |
|
26 | 27 | |
|
27 | 28 | # Pattern prefixes that are allowed in narrow patterns. This list MUST |
|
28 | 29 | # only contain patterns that are fast and safe to evaluate. Keep in mind |
|
29 | 30 | # that patterns are supplied by clients and executed on remote servers |
|
30 | 31 | # as part of wire protocol commands. That means that changes to this |
|
31 | 32 | # data structure influence the wire protocol and should not be taken |
|
32 | 33 | # lightly - especially removals. |
|
33 | 34 | VALID_PREFIXES = ( |
|
34 | 35 | b'path:', |
|
35 | 36 | b'rootfilesin:', |
|
36 | 37 | ) |
|
37 | 38 | |
|
38 | 39 | def normalizesplitpattern(kind, pat): |
|
39 | 40 | """Returns the normalized version of a pattern and kind. |
|
40 | 41 | |
|
41 | 42 | Returns a tuple with the normalized kind and normalized pattern. |
|
42 | 43 | """ |
|
43 | 44 | pat = pat.rstrip('/') |
|
44 | 45 | _validatepattern(pat) |
|
45 | 46 | return kind, pat |
|
46 | 47 | |
|
47 | 48 | def _numlines(s): |
|
48 | 49 | """Returns the number of lines in s, including ending empty lines.""" |
|
49 | 50 | # We use splitlines because it is Unicode-friendly and thus Python 3 |
|
50 | 51 | # compatible. However, it does not count empty lines at the end, so trick |
|
51 | 52 | # it by adding a character at the end. |
|
52 | 53 | return len((s + 'x').splitlines()) |
|
53 | 54 | |
|
54 | 55 | def _validatepattern(pat): |
|
55 | 56 | """Validates the pattern and aborts if it is invalid. |
|
56 | 57 | |
|
57 | 58 | Patterns are stored in the narrowspec as newline-separated |
|
58 | 59 | POSIX-style bytestring paths. There's no escaping. |
|
59 | 60 | """ |
|
60 | 61 | |
|
61 | 62 | # We use newlines as separators in the narrowspec file, so don't allow them |
|
62 | 63 | # in patterns. |
|
63 | 64 | if _numlines(pat) > 1: |
|
64 | 65 | raise error.Abort(_('newlines are not allowed in narrowspec paths')) |
|
65 | 66 | |
|
66 | 67 | components = pat.split('/') |
|
67 | 68 | if '.' in components or '..' in components: |
|
68 | 69 | raise error.Abort(_('"." and ".." are not allowed in narrowspec paths')) |
|
69 | 70 | |
|
70 | 71 | def normalizepattern(pattern, defaultkind='path'): |
|
71 | 72 | """Returns the normalized version of a text-format pattern. |
|
72 | 73 | |
|
73 | 74 | If the pattern has no kind, the default will be added. |
|
74 | 75 | """ |
|
75 | 76 | kind, pat = matchmod._patsplit(pattern, defaultkind) |
|
76 | 77 | return '%s:%s' % normalizesplitpattern(kind, pat) |
|
77 | 78 | |
|
78 | 79 | def parsepatterns(pats): |
|
79 | 80 | """Parses an iterable of patterns into a typed pattern set. |
|
80 | 81 | |
|
81 | 82 | Patterns are assumed to be ``path:`` if no prefix is present. |
|
82 | 83 | For safety and performance reasons, only some prefixes are allowed. |
|
83 | 84 | See ``validatepatterns()``. |
|
84 | 85 | |
|
85 | 86 | This function should be used on patterns that come from the user to |
|
86 | 87 | normalize and validate them to the internal data structure used for |
|
87 | 88 | representing patterns. |
|
88 | 89 | """ |
|
89 | 90 | res = {normalizepattern(orig) for orig in pats} |
|
90 | 91 | validatepatterns(res) |
|
91 | 92 | return res |
|
92 | 93 | |
|
93 | 94 | def validatepatterns(pats): |
|
94 | 95 | """Validate that patterns are in the expected data structure and format. |
|
95 | 96 | |
|
96 | 97 | And that is a set of normalized patterns beginning with ``path:`` or |
|
97 | 98 | ``rootfilesin:``. |
|
98 | 99 | |
|
99 | 100 | This function should be used to validate internal data structures |
|
100 | 101 | and patterns that are loaded from sources that use the internal, |
|
101 | 102 | prefixed pattern representation (but can't necessarily be fully trusted). |
|
102 | 103 | """ |
|
103 | 104 | if not isinstance(pats, set): |
|
104 | 105 | raise error.ProgrammingError('narrow patterns should be a set; ' |
|
105 | 106 | 'got %r' % pats) |
|
106 | 107 | |
|
107 | 108 | for pat in pats: |
|
108 | 109 | if not pat.startswith(VALID_PREFIXES): |
|
109 | 110 | # Use a Mercurial exception because this can happen due to user |
|
110 | 111 | # bugs (e.g. manually updating spec file). |
|
111 | 112 | raise error.Abort(_('invalid prefix on narrow pattern: %s') % pat, |
|
112 | 113 | hint=_('narrow patterns must begin with one of ' |
|
113 | 114 | 'the following: %s') % |
|
114 | 115 | ', '.join(VALID_PREFIXES)) |
|
115 | 116 | |
|
116 | 117 | def format(includes, excludes): |
|
117 | 118 | output = '[include]\n' |
|
118 | 119 | for i in sorted(includes - excludes): |
|
119 | 120 | output += i + '\n' |
|
120 | 121 | output += '[exclude]\n' |
|
121 | 122 | for e in sorted(excludes): |
|
122 | 123 | output += e + '\n' |
|
123 | 124 | return output |
|
124 | 125 | |
|
125 | 126 | def match(root, include=None, exclude=None): |
|
126 | 127 | if not include: |
|
127 | 128 | # Passing empty include and empty exclude to matchmod.match() |
|
128 | 129 | # gives a matcher that matches everything, so explicitly use |
|
129 | 130 | # the nevermatcher. |
|
130 | 131 | return matchmod.never() |
|
131 | 132 | return matchmod.match(root, '', [], include=include or [], |
|
132 | 133 | exclude=exclude or []) |
|
133 | 134 | |
|
134 | 135 | def parseconfig(ui, spec): |
|
135 | 136 | # maybe we should care about the profiles returned too |
|
136 | 137 | includepats, excludepats, profiles = sparse.parseconfig(ui, spec, 'narrow') |
|
137 | 138 | if profiles: |
|
138 | 139 | raise error.Abort(_("including other spec files using '%include' is not" |
|
139 | 140 | " supported in narrowspec")) |
|
140 | 141 | |
|
141 | 142 | validatepatterns(includepats) |
|
142 | 143 | validatepatterns(excludepats) |
|
143 | 144 | |
|
144 | 145 | return includepats, excludepats |
|
145 | 146 | |
|
146 | 147 | def load(repo): |
|
147 | 148 | try: |
|
148 | 149 | spec = repo.svfs.read(FILENAME) |
|
149 | 150 | except IOError as e: |
|
150 | 151 | # Treat "narrowspec does not exist" the same as "narrowspec file exists |
|
151 | 152 | # and is empty". |
|
152 | 153 | if e.errno == errno.ENOENT: |
|
153 | 154 | return set(), set() |
|
154 | 155 | raise |
|
155 | 156 | |
|
156 | 157 | return parseconfig(repo.ui, spec) |
|
157 | 158 | |
|
158 | 159 | def save(repo, includepats, excludepats): |
|
159 | 160 | validatepatterns(includepats) |
|
160 | 161 | validatepatterns(excludepats) |
|
161 | 162 | spec = format(includepats, excludepats) |
|
162 | 163 | repo.svfs.write(FILENAME, spec) |
|
163 | 164 | |
|
164 | 165 | def copytoworkingcopy(repo): |
|
165 | 166 | spec = repo.svfs.read(FILENAME) |
|
166 | 167 | repo.vfs.write(DIRSTATE_FILENAME, spec) |
|
167 | 168 | |
|
168 | 169 | def savebackup(repo, backupname): |
|
169 | 170 | if repository.NARROW_REQUIREMENT not in repo.requirements: |
|
170 | 171 | return |
|
171 | 172 | svfs = repo.svfs |
|
172 | 173 | svfs.tryunlink(backupname) |
|
173 | 174 | util.copyfile(svfs.join(FILENAME), svfs.join(backupname), hardlink=True) |
|
174 | 175 | |
|
175 | 176 | def restorebackup(repo, backupname): |
|
176 | 177 | if repository.NARROW_REQUIREMENT not in repo.requirements: |
|
177 | 178 | return |
|
178 | 179 | util.rename(repo.svfs.join(backupname), repo.svfs.join(FILENAME)) |
|
179 | 180 | |
|
180 | 181 | def savewcbackup(repo, backupname): |
|
181 | 182 | if repository.NARROW_REQUIREMENT not in repo.requirements: |
|
182 | 183 | return |
|
183 | 184 | vfs = repo.vfs |
|
184 | 185 | vfs.tryunlink(backupname) |
|
185 | 186 | # It may not exist in old repos |
|
186 | 187 | if vfs.exists(DIRSTATE_FILENAME): |
|
187 | 188 | util.copyfile(vfs.join(DIRSTATE_FILENAME), vfs.join(backupname), |
|
188 | 189 | hardlink=True) |
|
189 | 190 | |
|
190 | 191 | def restorewcbackup(repo, backupname): |
|
191 | 192 | if repository.NARROW_REQUIREMENT not in repo.requirements: |
|
192 | 193 | return |
|
193 | 194 | # It may not exist in old repos |
|
194 | 195 | if repo.vfs.exists(backupname): |
|
195 | 196 | util.rename(repo.vfs.join(backupname), repo.vfs.join(DIRSTATE_FILENAME)) |
|
196 | 197 | |
|
197 | 198 | def clearwcbackup(repo, backupname): |
|
198 | 199 | if repository.NARROW_REQUIREMENT not in repo.requirements: |
|
199 | 200 | return |
|
200 | 201 | repo.vfs.tryunlink(backupname) |
|
201 | 202 | |
|
202 | 203 | def restrictpatterns(req_includes, req_excludes, repo_includes, repo_excludes): |
|
203 | 204 | r""" Restricts the patterns according to repo settings, |
|
204 | 205 | results in a logical AND operation |
|
205 | 206 | |
|
206 | 207 | :param req_includes: requested includes |
|
207 | 208 | :param req_excludes: requested excludes |
|
208 | 209 | :param repo_includes: repo includes |
|
209 | 210 | :param repo_excludes: repo excludes |
|
210 | 211 | :return: include patterns, exclude patterns, and invalid include patterns. |
|
211 | 212 | |
|
212 | 213 | >>> restrictpatterns({'f1','f2'}, {}, ['f1'], []) |
|
213 | 214 | (set(['f1']), {}, []) |
|
214 | 215 | >>> restrictpatterns({'f1'}, {}, ['f1','f2'], []) |
|
215 | 216 | (set(['f1']), {}, []) |
|
216 | 217 | >>> restrictpatterns({'f1/fc1', 'f3/fc3'}, {}, ['f1','f2'], []) |
|
217 | 218 | (set(['f1/fc1']), {}, []) |
|
218 | 219 | >>> restrictpatterns({'f1_fc1'}, {}, ['f1','f2'], []) |
|
219 | 220 | ([], set(['path:.']), []) |
|
220 | 221 | >>> restrictpatterns({'f1/../f2/fc2'}, {}, ['f1','f2'], []) |
|
221 | 222 | (set(['f2/fc2']), {}, []) |
|
222 | 223 | >>> restrictpatterns({'f1/../f3/fc3'}, {}, ['f1','f2'], []) |
|
223 | 224 | ([], set(['path:.']), []) |
|
224 | 225 | >>> restrictpatterns({'f1/$non_exitent_var'}, {}, ['f1','f2'], []) |
|
225 | 226 | (set(['f1/$non_exitent_var']), {}, []) |
|
226 | 227 | """ |
|
227 | 228 | res_excludes = set(req_excludes) |
|
228 | 229 | res_excludes.update(repo_excludes) |
|
229 | 230 | invalid_includes = [] |
|
230 | 231 | if not req_includes: |
|
231 | 232 | res_includes = set(repo_includes) |
|
232 | 233 | elif 'path:.' not in repo_includes: |
|
233 | 234 | res_includes = [] |
|
234 | 235 | for req_include in req_includes: |
|
235 | 236 | req_include = util.expandpath(util.normpath(req_include)) |
|
236 | 237 | if req_include in repo_includes: |
|
237 | 238 | res_includes.append(req_include) |
|
238 | 239 | continue |
|
239 | 240 | valid = False |
|
240 | 241 | for repo_include in repo_includes: |
|
241 | 242 | if req_include.startswith(repo_include + '/'): |
|
242 | 243 | valid = True |
|
243 | 244 | res_includes.append(req_include) |
|
244 | 245 | break |
|
245 | 246 | if not valid: |
|
246 | 247 | invalid_includes.append(req_include) |
|
247 | 248 | if len(res_includes) == 0: |
|
248 | 249 | res_excludes = {'path:.'} |
|
249 | 250 | else: |
|
250 | 251 | res_includes = set(res_includes) |
|
251 | 252 | else: |
|
252 | 253 | res_includes = set(req_includes) |
|
253 | 254 | return res_includes, res_excludes, invalid_includes |
|
254 | 255 | |
|
255 | 256 | # These two are extracted for extensions (specifically for Google's CitC file |
|
256 | 257 | # system) |
|
257 | 258 | def _deletecleanfiles(repo, files): |
|
258 | 259 | for f in files: |
|
259 | 260 | repo.wvfs.unlinkpath(f) |
|
260 | 261 | |
|
261 | 262 | def _writeaddedfiles(repo, pctx, files): |
|
262 | 263 | actions = merge.emptyactions() |
|
263 | 264 | addgaction = actions[merge.ACTION_GET].append |
|
264 | 265 | mf = repo['.'].manifest() |
|
265 | 266 | for f in files: |
|
266 | 267 | if not repo.wvfs.exists(f): |
|
267 | 268 | addgaction((f, (mf.flags(f), False), "narrowspec updated")) |
|
268 | 269 | merge.applyupdates(repo, actions, wctx=repo[None], |
|
269 | 270 | mctx=repo['.'], overwrite=False) |
|
270 | 271 | |
|
271 | 272 | def checkworkingcopynarrowspec(repo): |
|
272 | 273 | storespec = repo.svfs.tryread(FILENAME) |
|
273 | 274 | wcspec = repo.vfs.tryread(DIRSTATE_FILENAME) |
|
274 | 275 | if wcspec != storespec: |
|
275 | 276 | raise error.Abort(_("working copy's narrowspec is stale"), |
|
276 | 277 | hint=_("run 'hg tracked --update-working-copy'")) |
|
277 | 278 | |
|
278 | 279 | def updateworkingcopy(repo, assumeclean=False): |
|
279 | 280 | """updates the working copy and dirstate from the store narrowspec |
|
280 | 281 | |
|
281 | 282 | When assumeclean=True, files that are not known to be clean will also |
|
282 | 283 | be deleted. It is then up to the caller to make sure they are clean. |
|
283 | 284 | """ |
|
284 | 285 | oldspec = repo.vfs.tryread(DIRSTATE_FILENAME) |
|
285 | 286 | newspec = repo.svfs.tryread(FILENAME) |
|
286 | 287 | |
|
287 | 288 | oldincludes, oldexcludes = parseconfig(repo.ui, oldspec) |
|
288 | 289 | newincludes, newexcludes = parseconfig(repo.ui, newspec) |
|
289 | 290 | oldmatch = match(repo.root, include=oldincludes, exclude=oldexcludes) |
|
290 | 291 | newmatch = match(repo.root, include=newincludes, exclude=newexcludes) |
|
291 | 292 | addedmatch = matchmod.differencematcher(newmatch, oldmatch) |
|
292 | 293 | removedmatch = matchmod.differencematcher(oldmatch, newmatch) |
|
293 | 294 | |
|
294 | 295 | ds = repo.dirstate |
|
295 | 296 | lookup, status = ds.status(removedmatch, subrepos=[], ignored=False, |
|
296 | 297 | clean=True, unknown=False) |
|
297 | 298 | trackeddirty = status.modified + status.added |
|
298 | 299 | clean = status.clean |
|
299 | 300 | if assumeclean: |
|
300 | 301 | assert not trackeddirty |
|
301 | 302 | clean.extend(lookup) |
|
302 | 303 | else: |
|
303 | 304 | trackeddirty.extend(lookup) |
|
304 | 305 | _deletecleanfiles(repo, clean) |
|
306 | uipathfn = scmutil.getuipathfn(repo) | |
|
305 | 307 | for f in sorted(trackeddirty): |
|
306 | repo.ui.status(_('not deleting possibly dirty file %s\n') % f) | |
|
308 | repo.ui.status(_('not deleting possibly dirty file %s\n') % uipathfn(f)) | |
|
307 | 309 | for f in clean + trackeddirty: |
|
308 | 310 | ds.drop(f) |
|
309 | 311 | |
|
310 | 312 | repo.narrowpats = newincludes, newexcludes |
|
311 | 313 | repo._narrowmatch = newmatch |
|
312 | 314 | pctx = repo['.'] |
|
313 | 315 | newfiles = [f for f in pctx.manifest().walk(addedmatch) if f not in ds] |
|
314 | 316 | for f in newfiles: |
|
315 | 317 | ds.normallookup(f) |
|
316 | 318 | _writeaddedfiles(repo, pctx, newfiles) |
General Comments 0
You need to be logged in to leave comments.
Login now