##// END OF EJS Templates
dirstate.walk: maintain a list of dirs not found...
Siddharth Agarwal -
r19170:b184e504 default
parent child Browse files
Show More
@@ -1,820 +1,822 b''
1 # dirstate.py - working directory tracking for mercurial
1 # dirstate.py - working directory tracking for mercurial
2 #
2 #
3 # Copyright 2005-2007 Matt Mackall <mpm@selenic.com>
3 # Copyright 2005-2007 Matt Mackall <mpm@selenic.com>
4 #
4 #
5 # This software may be used and distributed according to the terms of the
5 # This software may be used and distributed according to the terms of the
6 # GNU General Public License version 2 or any later version.
6 # GNU General Public License version 2 or any later version.
7 import errno
7 import errno
8
8
9 from node import nullid
9 from node import nullid
10 from i18n import _
10 from i18n import _
11 import scmutil, util, ignore, osutil, parsers, encoding
11 import scmutil, util, ignore, osutil, parsers, encoding
12 import os, stat, errno, gc
12 import os, stat, errno, gc
13
13
14 propertycache = util.propertycache
14 propertycache = util.propertycache
15 filecache = scmutil.filecache
15 filecache = scmutil.filecache
16 _rangemask = 0x7fffffff
16 _rangemask = 0x7fffffff
17
17
18 class repocache(filecache):
18 class repocache(filecache):
19 """filecache for files in .hg/"""
19 """filecache for files in .hg/"""
20 def join(self, obj, fname):
20 def join(self, obj, fname):
21 return obj._opener.join(fname)
21 return obj._opener.join(fname)
22
22
23 class rootcache(filecache):
23 class rootcache(filecache):
24 """filecache for files in the repository root"""
24 """filecache for files in the repository root"""
25 def join(self, obj, fname):
25 def join(self, obj, fname):
26 return obj._join(fname)
26 return obj._join(fname)
27
27
28 class dirstate(object):
28 class dirstate(object):
29
29
30 def __init__(self, opener, ui, root, validate):
30 def __init__(self, opener, ui, root, validate):
31 '''Create a new dirstate object.
31 '''Create a new dirstate object.
32
32
33 opener is an open()-like callable that can be used to open the
33 opener is an open()-like callable that can be used to open the
34 dirstate file; root is the root of the directory tracked by
34 dirstate file; root is the root of the directory tracked by
35 the dirstate.
35 the dirstate.
36 '''
36 '''
37 self._opener = opener
37 self._opener = opener
38 self._validate = validate
38 self._validate = validate
39 self._root = root
39 self._root = root
40 self._rootdir = os.path.join(root, '')
40 self._rootdir = os.path.join(root, '')
41 self._dirty = False
41 self._dirty = False
42 self._dirtypl = False
42 self._dirtypl = False
43 self._lastnormaltime = 0
43 self._lastnormaltime = 0
44 self._ui = ui
44 self._ui = ui
45 self._filecache = {}
45 self._filecache = {}
46
46
47 @propertycache
47 @propertycache
48 def _map(self):
48 def _map(self):
49 '''Return the dirstate contents as a map from filename to
49 '''Return the dirstate contents as a map from filename to
50 (state, mode, size, time).'''
50 (state, mode, size, time).'''
51 self._read()
51 self._read()
52 return self._map
52 return self._map
53
53
54 @propertycache
54 @propertycache
55 def _copymap(self):
55 def _copymap(self):
56 self._read()
56 self._read()
57 return self._copymap
57 return self._copymap
58
58
59 @propertycache
59 @propertycache
60 def _foldmap(self):
60 def _foldmap(self):
61 f = {}
61 f = {}
62 for name, s in self._map.iteritems():
62 for name, s in self._map.iteritems():
63 if s[0] != 'r':
63 if s[0] != 'r':
64 f[util.normcase(name)] = name
64 f[util.normcase(name)] = name
65 for name in self._dirs:
65 for name in self._dirs:
66 f[util.normcase(name)] = name
66 f[util.normcase(name)] = name
67 f['.'] = '.' # prevents useless util.fspath() invocation
67 f['.'] = '.' # prevents useless util.fspath() invocation
68 return f
68 return f
69
69
70 @repocache('branch')
70 @repocache('branch')
71 def _branch(self):
71 def _branch(self):
72 try:
72 try:
73 return self._opener.read("branch").strip() or "default"
73 return self._opener.read("branch").strip() or "default"
74 except IOError, inst:
74 except IOError, inst:
75 if inst.errno != errno.ENOENT:
75 if inst.errno != errno.ENOENT:
76 raise
76 raise
77 return "default"
77 return "default"
78
78
79 @propertycache
79 @propertycache
80 def _pl(self):
80 def _pl(self):
81 try:
81 try:
82 fp = self._opener("dirstate")
82 fp = self._opener("dirstate")
83 st = fp.read(40)
83 st = fp.read(40)
84 fp.close()
84 fp.close()
85 l = len(st)
85 l = len(st)
86 if l == 40:
86 if l == 40:
87 return st[:20], st[20:40]
87 return st[:20], st[20:40]
88 elif l > 0 and l < 40:
88 elif l > 0 and l < 40:
89 raise util.Abort(_('working directory state appears damaged!'))
89 raise util.Abort(_('working directory state appears damaged!'))
90 except IOError, err:
90 except IOError, err:
91 if err.errno != errno.ENOENT:
91 if err.errno != errno.ENOENT:
92 raise
92 raise
93 return [nullid, nullid]
93 return [nullid, nullid]
94
94
95 @propertycache
95 @propertycache
96 def _dirs(self):
96 def _dirs(self):
97 return scmutil.dirs(self._map, 'r')
97 return scmutil.dirs(self._map, 'r')
98
98
99 def dirs(self):
99 def dirs(self):
100 return self._dirs
100 return self._dirs
101
101
102 @rootcache('.hgignore')
102 @rootcache('.hgignore')
103 def _ignore(self):
103 def _ignore(self):
104 files = [self._join('.hgignore')]
104 files = [self._join('.hgignore')]
105 for name, path in self._ui.configitems("ui"):
105 for name, path in self._ui.configitems("ui"):
106 if name == 'ignore' or name.startswith('ignore.'):
106 if name == 'ignore' or name.startswith('ignore.'):
107 files.append(util.expandpath(path))
107 files.append(util.expandpath(path))
108 return ignore.ignore(self._root, files, self._ui.warn)
108 return ignore.ignore(self._root, files, self._ui.warn)
109
109
110 @propertycache
110 @propertycache
111 def _slash(self):
111 def _slash(self):
112 return self._ui.configbool('ui', 'slash') and os.sep != '/'
112 return self._ui.configbool('ui', 'slash') and os.sep != '/'
113
113
114 @propertycache
114 @propertycache
115 def _checklink(self):
115 def _checklink(self):
116 return util.checklink(self._root)
116 return util.checklink(self._root)
117
117
118 @propertycache
118 @propertycache
119 def _checkexec(self):
119 def _checkexec(self):
120 return util.checkexec(self._root)
120 return util.checkexec(self._root)
121
121
122 @propertycache
122 @propertycache
123 def _checkcase(self):
123 def _checkcase(self):
124 return not util.checkcase(self._join('.hg'))
124 return not util.checkcase(self._join('.hg'))
125
125
126 def _join(self, f):
126 def _join(self, f):
127 # much faster than os.path.join()
127 # much faster than os.path.join()
128 # it's safe because f is always a relative path
128 # it's safe because f is always a relative path
129 return self._rootdir + f
129 return self._rootdir + f
130
130
131 def flagfunc(self, buildfallback):
131 def flagfunc(self, buildfallback):
132 if self._checklink and self._checkexec:
132 if self._checklink and self._checkexec:
133 def f(x):
133 def f(x):
134 try:
134 try:
135 st = os.lstat(self._join(x))
135 st = os.lstat(self._join(x))
136 if util.statislink(st):
136 if util.statislink(st):
137 return 'l'
137 return 'l'
138 if util.statisexec(st):
138 if util.statisexec(st):
139 return 'x'
139 return 'x'
140 except OSError:
140 except OSError:
141 pass
141 pass
142 return ''
142 return ''
143 return f
143 return f
144
144
145 fallback = buildfallback()
145 fallback = buildfallback()
146 if self._checklink:
146 if self._checklink:
147 def f(x):
147 def f(x):
148 if os.path.islink(self._join(x)):
148 if os.path.islink(self._join(x)):
149 return 'l'
149 return 'l'
150 if 'x' in fallback(x):
150 if 'x' in fallback(x):
151 return 'x'
151 return 'x'
152 return ''
152 return ''
153 return f
153 return f
154 if self._checkexec:
154 if self._checkexec:
155 def f(x):
155 def f(x):
156 if 'l' in fallback(x):
156 if 'l' in fallback(x):
157 return 'l'
157 return 'l'
158 if util.isexec(self._join(x)):
158 if util.isexec(self._join(x)):
159 return 'x'
159 return 'x'
160 return ''
160 return ''
161 return f
161 return f
162 else:
162 else:
163 return fallback
163 return fallback
164
164
165 def getcwd(self):
165 def getcwd(self):
166 cwd = os.getcwd()
166 cwd = os.getcwd()
167 if cwd == self._root:
167 if cwd == self._root:
168 return ''
168 return ''
169 # self._root ends with a path separator if self._root is '/' or 'C:\'
169 # self._root ends with a path separator if self._root is '/' or 'C:\'
170 rootsep = self._root
170 rootsep = self._root
171 if not util.endswithsep(rootsep):
171 if not util.endswithsep(rootsep):
172 rootsep += os.sep
172 rootsep += os.sep
173 if cwd.startswith(rootsep):
173 if cwd.startswith(rootsep):
174 return cwd[len(rootsep):]
174 return cwd[len(rootsep):]
175 else:
175 else:
176 # we're outside the repo. return an absolute path.
176 # we're outside the repo. return an absolute path.
177 return cwd
177 return cwd
178
178
179 def pathto(self, f, cwd=None):
179 def pathto(self, f, cwd=None):
180 if cwd is None:
180 if cwd is None:
181 cwd = self.getcwd()
181 cwd = self.getcwd()
182 path = util.pathto(self._root, cwd, f)
182 path = util.pathto(self._root, cwd, f)
183 if self._slash:
183 if self._slash:
184 return util.normpath(path)
184 return util.normpath(path)
185 return path
185 return path
186
186
187 def __getitem__(self, key):
187 def __getitem__(self, key):
188 '''Return the current state of key (a filename) in the dirstate.
188 '''Return the current state of key (a filename) in the dirstate.
189
189
190 States are:
190 States are:
191 n normal
191 n normal
192 m needs merging
192 m needs merging
193 r marked for removal
193 r marked for removal
194 a marked for addition
194 a marked for addition
195 ? not tracked
195 ? not tracked
196 '''
196 '''
197 return self._map.get(key, ("?",))[0]
197 return self._map.get(key, ("?",))[0]
198
198
199 def __contains__(self, key):
199 def __contains__(self, key):
200 return key in self._map
200 return key in self._map
201
201
202 def __iter__(self):
202 def __iter__(self):
203 for x in sorted(self._map):
203 for x in sorted(self._map):
204 yield x
204 yield x
205
205
206 def iteritems(self):
206 def iteritems(self):
207 return self._map.iteritems()
207 return self._map.iteritems()
208
208
209 def parents(self):
209 def parents(self):
210 return [self._validate(p) for p in self._pl]
210 return [self._validate(p) for p in self._pl]
211
211
212 def p1(self):
212 def p1(self):
213 return self._validate(self._pl[0])
213 return self._validate(self._pl[0])
214
214
215 def p2(self):
215 def p2(self):
216 return self._validate(self._pl[1])
216 return self._validate(self._pl[1])
217
217
218 def branch(self):
218 def branch(self):
219 return encoding.tolocal(self._branch)
219 return encoding.tolocal(self._branch)
220
220
221 def setparents(self, p1, p2=nullid):
221 def setparents(self, p1, p2=nullid):
222 """Set dirstate parents to p1 and p2.
222 """Set dirstate parents to p1 and p2.
223
223
224 When moving from two parents to one, 'm' merged entries a
224 When moving from two parents to one, 'm' merged entries a
225 adjusted to normal and previous copy records discarded and
225 adjusted to normal and previous copy records discarded and
226 returned by the call.
226 returned by the call.
227
227
228 See localrepo.setparents()
228 See localrepo.setparents()
229 """
229 """
230 self._dirty = self._dirtypl = True
230 self._dirty = self._dirtypl = True
231 oldp2 = self._pl[1]
231 oldp2 = self._pl[1]
232 self._pl = p1, p2
232 self._pl = p1, p2
233 copies = {}
233 copies = {}
234 if oldp2 != nullid and p2 == nullid:
234 if oldp2 != nullid and p2 == nullid:
235 # Discard 'm' markers when moving away from a merge state
235 # Discard 'm' markers when moving away from a merge state
236 for f, s in self._map.iteritems():
236 for f, s in self._map.iteritems():
237 if s[0] == 'm':
237 if s[0] == 'm':
238 if f in self._copymap:
238 if f in self._copymap:
239 copies[f] = self._copymap[f]
239 copies[f] = self._copymap[f]
240 self.normallookup(f)
240 self.normallookup(f)
241 return copies
241 return copies
242
242
243 def setbranch(self, branch):
243 def setbranch(self, branch):
244 self._branch = encoding.fromlocal(branch)
244 self._branch = encoding.fromlocal(branch)
245 f = self._opener('branch', 'w', atomictemp=True)
245 f = self._opener('branch', 'w', atomictemp=True)
246 try:
246 try:
247 f.write(self._branch + '\n')
247 f.write(self._branch + '\n')
248 f.close()
248 f.close()
249
249
250 # make sure filecache has the correct stat info for _branch after
250 # make sure filecache has the correct stat info for _branch after
251 # replacing the underlying file
251 # replacing the underlying file
252 ce = self._filecache['_branch']
252 ce = self._filecache['_branch']
253 if ce:
253 if ce:
254 ce.refresh()
254 ce.refresh()
255 except: # re-raises
255 except: # re-raises
256 f.discard()
256 f.discard()
257 raise
257 raise
258
258
259 def _read(self):
259 def _read(self):
260 self._map = {}
260 self._map = {}
261 self._copymap = {}
261 self._copymap = {}
262 try:
262 try:
263 st = self._opener.read("dirstate")
263 st = self._opener.read("dirstate")
264 except IOError, err:
264 except IOError, err:
265 if err.errno != errno.ENOENT:
265 if err.errno != errno.ENOENT:
266 raise
266 raise
267 return
267 return
268 if not st:
268 if not st:
269 return
269 return
270
270
271 # Python's garbage collector triggers a GC each time a certain number
271 # Python's garbage collector triggers a GC each time a certain number
272 # of container objects (the number being defined by
272 # of container objects (the number being defined by
273 # gc.get_threshold()) are allocated. parse_dirstate creates a tuple
273 # gc.get_threshold()) are allocated. parse_dirstate creates a tuple
274 # for each file in the dirstate. The C version then immediately marks
274 # for each file in the dirstate. The C version then immediately marks
275 # them as not to be tracked by the collector. However, this has no
275 # them as not to be tracked by the collector. However, this has no
276 # effect on when GCs are triggered, only on what objects the GC looks
276 # effect on when GCs are triggered, only on what objects the GC looks
277 # into. This means that O(number of files) GCs are unavoidable.
277 # into. This means that O(number of files) GCs are unavoidable.
278 # Depending on when in the process's lifetime the dirstate is parsed,
278 # Depending on when in the process's lifetime the dirstate is parsed,
279 # this can get very expensive. As a workaround, disable GC while
279 # this can get very expensive. As a workaround, disable GC while
280 # parsing the dirstate.
280 # parsing the dirstate.
281 gcenabled = gc.isenabled()
281 gcenabled = gc.isenabled()
282 gc.disable()
282 gc.disable()
283 try:
283 try:
284 p = parsers.parse_dirstate(self._map, self._copymap, st)
284 p = parsers.parse_dirstate(self._map, self._copymap, st)
285 finally:
285 finally:
286 if gcenabled:
286 if gcenabled:
287 gc.enable()
287 gc.enable()
288 if not self._dirtypl:
288 if not self._dirtypl:
289 self._pl = p
289 self._pl = p
290
290
291 def invalidate(self):
291 def invalidate(self):
292 for a in ("_map", "_copymap", "_foldmap", "_branch", "_pl", "_dirs",
292 for a in ("_map", "_copymap", "_foldmap", "_branch", "_pl", "_dirs",
293 "_ignore"):
293 "_ignore"):
294 if a in self.__dict__:
294 if a in self.__dict__:
295 delattr(self, a)
295 delattr(self, a)
296 self._lastnormaltime = 0
296 self._lastnormaltime = 0
297 self._dirty = False
297 self._dirty = False
298
298
299 def copy(self, source, dest):
299 def copy(self, source, dest):
300 """Mark dest as a copy of source. Unmark dest if source is None."""
300 """Mark dest as a copy of source. Unmark dest if source is None."""
301 if source == dest:
301 if source == dest:
302 return
302 return
303 self._dirty = True
303 self._dirty = True
304 if source is not None:
304 if source is not None:
305 self._copymap[dest] = source
305 self._copymap[dest] = source
306 elif dest in self._copymap:
306 elif dest in self._copymap:
307 del self._copymap[dest]
307 del self._copymap[dest]
308
308
309 def copied(self, file):
309 def copied(self, file):
310 return self._copymap.get(file, None)
310 return self._copymap.get(file, None)
311
311
312 def copies(self):
312 def copies(self):
313 return self._copymap
313 return self._copymap
314
314
315 def _droppath(self, f):
315 def _droppath(self, f):
316 if self[f] not in "?r" and "_dirs" in self.__dict__:
316 if self[f] not in "?r" and "_dirs" in self.__dict__:
317 self._dirs.delpath(f)
317 self._dirs.delpath(f)
318
318
319 def _addpath(self, f, state, mode, size, mtime):
319 def _addpath(self, f, state, mode, size, mtime):
320 oldstate = self[f]
320 oldstate = self[f]
321 if state == 'a' or oldstate == 'r':
321 if state == 'a' or oldstate == 'r':
322 scmutil.checkfilename(f)
322 scmutil.checkfilename(f)
323 if f in self._dirs:
323 if f in self._dirs:
324 raise util.Abort(_('directory %r already in dirstate') % f)
324 raise util.Abort(_('directory %r already in dirstate') % f)
325 # shadows
325 # shadows
326 for d in scmutil.finddirs(f):
326 for d in scmutil.finddirs(f):
327 if d in self._dirs:
327 if d in self._dirs:
328 break
328 break
329 if d in self._map and self[d] != 'r':
329 if d in self._map and self[d] != 'r':
330 raise util.Abort(
330 raise util.Abort(
331 _('file %r in dirstate clashes with %r') % (d, f))
331 _('file %r in dirstate clashes with %r') % (d, f))
332 if oldstate in "?r" and "_dirs" in self.__dict__:
332 if oldstate in "?r" and "_dirs" in self.__dict__:
333 self._dirs.addpath(f)
333 self._dirs.addpath(f)
334 self._dirty = True
334 self._dirty = True
335 self._map[f] = (state, mode, size, mtime)
335 self._map[f] = (state, mode, size, mtime)
336
336
337 def normal(self, f):
337 def normal(self, f):
338 '''Mark a file normal and clean.'''
338 '''Mark a file normal and clean.'''
339 s = os.lstat(self._join(f))
339 s = os.lstat(self._join(f))
340 mtime = int(s.st_mtime)
340 mtime = int(s.st_mtime)
341 self._addpath(f, 'n', s.st_mode,
341 self._addpath(f, 'n', s.st_mode,
342 s.st_size & _rangemask, mtime & _rangemask)
342 s.st_size & _rangemask, mtime & _rangemask)
343 if f in self._copymap:
343 if f in self._copymap:
344 del self._copymap[f]
344 del self._copymap[f]
345 if mtime > self._lastnormaltime:
345 if mtime > self._lastnormaltime:
346 # Remember the most recent modification timeslot for status(),
346 # Remember the most recent modification timeslot for status(),
347 # to make sure we won't miss future size-preserving file content
347 # to make sure we won't miss future size-preserving file content
348 # modifications that happen within the same timeslot.
348 # modifications that happen within the same timeslot.
349 self._lastnormaltime = mtime
349 self._lastnormaltime = mtime
350
350
351 def normallookup(self, f):
351 def normallookup(self, f):
352 '''Mark a file normal, but possibly dirty.'''
352 '''Mark a file normal, but possibly dirty.'''
353 if self._pl[1] != nullid and f in self._map:
353 if self._pl[1] != nullid and f in self._map:
354 # if there is a merge going on and the file was either
354 # if there is a merge going on and the file was either
355 # in state 'm' (-1) or coming from other parent (-2) before
355 # in state 'm' (-1) or coming from other parent (-2) before
356 # being removed, restore that state.
356 # being removed, restore that state.
357 entry = self._map[f]
357 entry = self._map[f]
358 if entry[0] == 'r' and entry[2] in (-1, -2):
358 if entry[0] == 'r' and entry[2] in (-1, -2):
359 source = self._copymap.get(f)
359 source = self._copymap.get(f)
360 if entry[2] == -1:
360 if entry[2] == -1:
361 self.merge(f)
361 self.merge(f)
362 elif entry[2] == -2:
362 elif entry[2] == -2:
363 self.otherparent(f)
363 self.otherparent(f)
364 if source:
364 if source:
365 self.copy(source, f)
365 self.copy(source, f)
366 return
366 return
367 if entry[0] == 'm' or entry[0] == 'n' and entry[2] == -2:
367 if entry[0] == 'm' or entry[0] == 'n' and entry[2] == -2:
368 return
368 return
369 self._addpath(f, 'n', 0, -1, -1)
369 self._addpath(f, 'n', 0, -1, -1)
370 if f in self._copymap:
370 if f in self._copymap:
371 del self._copymap[f]
371 del self._copymap[f]
372
372
373 def otherparent(self, f):
373 def otherparent(self, f):
374 '''Mark as coming from the other parent, always dirty.'''
374 '''Mark as coming from the other parent, always dirty.'''
375 if self._pl[1] == nullid:
375 if self._pl[1] == nullid:
376 raise util.Abort(_("setting %r to other parent "
376 raise util.Abort(_("setting %r to other parent "
377 "only allowed in merges") % f)
377 "only allowed in merges") % f)
378 self._addpath(f, 'n', 0, -2, -1)
378 self._addpath(f, 'n', 0, -2, -1)
379 if f in self._copymap:
379 if f in self._copymap:
380 del self._copymap[f]
380 del self._copymap[f]
381
381
382 def add(self, f):
382 def add(self, f):
383 '''Mark a file added.'''
383 '''Mark a file added.'''
384 self._addpath(f, 'a', 0, -1, -1)
384 self._addpath(f, 'a', 0, -1, -1)
385 if f in self._copymap:
385 if f in self._copymap:
386 del self._copymap[f]
386 del self._copymap[f]
387
387
388 def remove(self, f):
388 def remove(self, f):
389 '''Mark a file removed.'''
389 '''Mark a file removed.'''
390 self._dirty = True
390 self._dirty = True
391 self._droppath(f)
391 self._droppath(f)
392 size = 0
392 size = 0
393 if self._pl[1] != nullid and f in self._map:
393 if self._pl[1] != nullid and f in self._map:
394 # backup the previous state
394 # backup the previous state
395 entry = self._map[f]
395 entry = self._map[f]
396 if entry[0] == 'm': # merge
396 if entry[0] == 'm': # merge
397 size = -1
397 size = -1
398 elif entry[0] == 'n' and entry[2] == -2: # other parent
398 elif entry[0] == 'n' and entry[2] == -2: # other parent
399 size = -2
399 size = -2
400 self._map[f] = ('r', 0, size, 0)
400 self._map[f] = ('r', 0, size, 0)
401 if size == 0 and f in self._copymap:
401 if size == 0 and f in self._copymap:
402 del self._copymap[f]
402 del self._copymap[f]
403
403
404 def merge(self, f):
404 def merge(self, f):
405 '''Mark a file merged.'''
405 '''Mark a file merged.'''
406 if self._pl[1] == nullid:
406 if self._pl[1] == nullid:
407 return self.normallookup(f)
407 return self.normallookup(f)
408 s = os.lstat(self._join(f))
408 s = os.lstat(self._join(f))
409 self._addpath(f, 'm', s.st_mode,
409 self._addpath(f, 'm', s.st_mode,
410 s.st_size & _rangemask, int(s.st_mtime) & _rangemask)
410 s.st_size & _rangemask, int(s.st_mtime) & _rangemask)
411 if f in self._copymap:
411 if f in self._copymap:
412 del self._copymap[f]
412 del self._copymap[f]
413
413
414 def drop(self, f):
414 def drop(self, f):
415 '''Drop a file from the dirstate'''
415 '''Drop a file from the dirstate'''
416 if f in self._map:
416 if f in self._map:
417 self._dirty = True
417 self._dirty = True
418 self._droppath(f)
418 self._droppath(f)
419 del self._map[f]
419 del self._map[f]
420
420
421 def _normalize(self, path, isknown, ignoremissing=False, exists=None):
421 def _normalize(self, path, isknown, ignoremissing=False, exists=None):
422 normed = util.normcase(path)
422 normed = util.normcase(path)
423 folded = self._foldmap.get(normed, None)
423 folded = self._foldmap.get(normed, None)
424 if folded is None:
424 if folded is None:
425 if isknown:
425 if isknown:
426 folded = path
426 folded = path
427 else:
427 else:
428 if exists is None:
428 if exists is None:
429 exists = os.path.lexists(os.path.join(self._root, path))
429 exists = os.path.lexists(os.path.join(self._root, path))
430 if not exists:
430 if not exists:
431 # Maybe a path component exists
431 # Maybe a path component exists
432 if not ignoremissing and '/' in path:
432 if not ignoremissing and '/' in path:
433 d, f = path.rsplit('/', 1)
433 d, f = path.rsplit('/', 1)
434 d = self._normalize(d, isknown, ignoremissing, None)
434 d = self._normalize(d, isknown, ignoremissing, None)
435 folded = d + "/" + f
435 folded = d + "/" + f
436 else:
436 else:
437 # No path components, preserve original case
437 # No path components, preserve original case
438 folded = path
438 folded = path
439 else:
439 else:
440 # recursively normalize leading directory components
440 # recursively normalize leading directory components
441 # against dirstate
441 # against dirstate
442 if '/' in normed:
442 if '/' in normed:
443 d, f = normed.rsplit('/', 1)
443 d, f = normed.rsplit('/', 1)
444 d = self._normalize(d, isknown, ignoremissing, True)
444 d = self._normalize(d, isknown, ignoremissing, True)
445 r = self._root + "/" + d
445 r = self._root + "/" + d
446 folded = d + "/" + util.fspath(f, r)
446 folded = d + "/" + util.fspath(f, r)
447 else:
447 else:
448 folded = util.fspath(normed, self._root)
448 folded = util.fspath(normed, self._root)
449 self._foldmap[normed] = folded
449 self._foldmap[normed] = folded
450
450
451 return folded
451 return folded
452
452
453 def normalize(self, path, isknown=False, ignoremissing=False):
453 def normalize(self, path, isknown=False, ignoremissing=False):
454 '''
454 '''
455 normalize the case of a pathname when on a casefolding filesystem
455 normalize the case of a pathname when on a casefolding filesystem
456
456
457 isknown specifies whether the filename came from walking the
457 isknown specifies whether the filename came from walking the
458 disk, to avoid extra filesystem access.
458 disk, to avoid extra filesystem access.
459
459
460 If ignoremissing is True, missing path are returned
460 If ignoremissing is True, missing path are returned
461 unchanged. Otherwise, we try harder to normalize possibly
461 unchanged. Otherwise, we try harder to normalize possibly
462 existing path components.
462 existing path components.
463
463
464 The normalized case is determined based on the following precedence:
464 The normalized case is determined based on the following precedence:
465
465
466 - version of name already stored in the dirstate
466 - version of name already stored in the dirstate
467 - version of name stored on disk
467 - version of name stored on disk
468 - version provided via command arguments
468 - version provided via command arguments
469 '''
469 '''
470
470
471 if self._checkcase:
471 if self._checkcase:
472 return self._normalize(path, isknown, ignoremissing)
472 return self._normalize(path, isknown, ignoremissing)
473 return path
473 return path
474
474
475 def clear(self):
475 def clear(self):
476 self._map = {}
476 self._map = {}
477 if "_dirs" in self.__dict__:
477 if "_dirs" in self.__dict__:
478 delattr(self, "_dirs")
478 delattr(self, "_dirs")
479 self._copymap = {}
479 self._copymap = {}
480 self._pl = [nullid, nullid]
480 self._pl = [nullid, nullid]
481 self._lastnormaltime = 0
481 self._lastnormaltime = 0
482 self._dirty = True
482 self._dirty = True
483
483
484 def rebuild(self, parent, allfiles, changedfiles=None):
484 def rebuild(self, parent, allfiles, changedfiles=None):
485 changedfiles = changedfiles or allfiles
485 changedfiles = changedfiles or allfiles
486 oldmap = self._map
486 oldmap = self._map
487 self.clear()
487 self.clear()
488 for f in allfiles:
488 for f in allfiles:
489 if f not in changedfiles:
489 if f not in changedfiles:
490 self._map[f] = oldmap[f]
490 self._map[f] = oldmap[f]
491 else:
491 else:
492 if 'x' in allfiles.flags(f):
492 if 'x' in allfiles.flags(f):
493 self._map[f] = ('n', 0777, -1, 0)
493 self._map[f] = ('n', 0777, -1, 0)
494 else:
494 else:
495 self._map[f] = ('n', 0666, -1, 0)
495 self._map[f] = ('n', 0666, -1, 0)
496 self._pl = (parent, nullid)
496 self._pl = (parent, nullid)
497 self._dirty = True
497 self._dirty = True
498
498
499 def write(self):
499 def write(self):
500 if not self._dirty:
500 if not self._dirty:
501 return
501 return
502 st = self._opener("dirstate", "w", atomictemp=True)
502 st = self._opener("dirstate", "w", atomictemp=True)
503
503
504 def finish(s):
504 def finish(s):
505 st.write(s)
505 st.write(s)
506 st.close()
506 st.close()
507 self._lastnormaltime = 0
507 self._lastnormaltime = 0
508 self._dirty = self._dirtypl = False
508 self._dirty = self._dirtypl = False
509
509
510 # use the modification time of the newly created temporary file as the
510 # use the modification time of the newly created temporary file as the
511 # filesystem's notion of 'now'
511 # filesystem's notion of 'now'
512 now = util.fstat(st).st_mtime
512 now = util.fstat(st).st_mtime
513 finish(parsers.pack_dirstate(self._map, self._copymap, self._pl, now))
513 finish(parsers.pack_dirstate(self._map, self._copymap, self._pl, now))
514
514
515 def _dirignore(self, f):
515 def _dirignore(self, f):
516 if f == '.':
516 if f == '.':
517 return False
517 return False
518 if self._ignore(f):
518 if self._ignore(f):
519 return True
519 return True
520 for p in scmutil.finddirs(f):
520 for p in scmutil.finddirs(f):
521 if self._ignore(p):
521 if self._ignore(p):
522 return True
522 return True
523 return False
523 return False
524
524
525 def walk(self, match, subrepos, unknown, ignored):
525 def walk(self, match, subrepos, unknown, ignored):
526 '''
526 '''
527 Walk recursively through the directory tree, finding all files
527 Walk recursively through the directory tree, finding all files
528 matched by match.
528 matched by match.
529
529
530 Return a dict mapping filename to stat-like object (either
530 Return a dict mapping filename to stat-like object (either
531 mercurial.osutil.stat instance or return value of os.stat()).
531 mercurial.osutil.stat instance or return value of os.stat()).
532 '''
532 '''
533
533
534 def fwarn(f, msg):
534 def fwarn(f, msg):
535 self._ui.warn('%s: %s\n' % (self.pathto(f), msg))
535 self._ui.warn('%s: %s\n' % (self.pathto(f), msg))
536 return False
536 return False
537
537
538 def badtype(mode):
538 def badtype(mode):
539 kind = _('unknown')
539 kind = _('unknown')
540 if stat.S_ISCHR(mode):
540 if stat.S_ISCHR(mode):
541 kind = _('character device')
541 kind = _('character device')
542 elif stat.S_ISBLK(mode):
542 elif stat.S_ISBLK(mode):
543 kind = _('block device')
543 kind = _('block device')
544 elif stat.S_ISFIFO(mode):
544 elif stat.S_ISFIFO(mode):
545 kind = _('fifo')
545 kind = _('fifo')
546 elif stat.S_ISSOCK(mode):
546 elif stat.S_ISSOCK(mode):
547 kind = _('socket')
547 kind = _('socket')
548 elif stat.S_ISDIR(mode):
548 elif stat.S_ISDIR(mode):
549 kind = _('directory')
549 kind = _('directory')
550 return _('unsupported file type (type is %s)') % kind
550 return _('unsupported file type (type is %s)') % kind
551
551
552 ignore = self._ignore
552 ignore = self._ignore
553 dirignore = self._dirignore
553 dirignore = self._dirignore
554 if ignored:
554 if ignored:
555 ignore = util.never
555 ignore = util.never
556 dirignore = util.never
556 dirignore = util.never
557 elif not unknown:
557 elif not unknown:
558 # if unknown and ignored are False, skip step 2
558 # if unknown and ignored are False, skip step 2
559 ignore = util.always
559 ignore = util.always
560 dirignore = util.always
560 dirignore = util.always
561
561
562 matchfn = match.matchfn
562 matchfn = match.matchfn
563 matchalways = match.always()
563 matchalways = match.always()
564 matchedir = match.explicitdir
564 matchedir = match.explicitdir
565 matchtdir = match.traversedir
565 matchtdir = match.traversedir
566 badfn = match.bad
566 badfn = match.bad
567 dmap = self._map
567 dmap = self._map
568 normpath = util.normpath
568 normpath = util.normpath
569 listdir = osutil.listdir
569 listdir = osutil.listdir
570 lstat = os.lstat
570 lstat = os.lstat
571 getkind = stat.S_IFMT
571 getkind = stat.S_IFMT
572 dirkind = stat.S_IFDIR
572 dirkind = stat.S_IFDIR
573 regkind = stat.S_IFREG
573 regkind = stat.S_IFREG
574 lnkkind = stat.S_IFLNK
574 lnkkind = stat.S_IFLNK
575 join = self._join
575 join = self._join
576 work = []
576 work = []
577 wadd = work.append
577 wadd = work.append
578
578
579 exact = skipstep3 = False
579 exact = skipstep3 = False
580 if matchfn == match.exact: # match.exact
580 if matchfn == match.exact: # match.exact
581 exact = True
581 exact = True
582 dirignore = util.always # skip step 2
582 dirignore = util.always # skip step 2
583 elif match.files() and not match.anypats(): # match.match, no patterns
583 elif match.files() and not match.anypats(): # match.match, no patterns
584 skipstep3 = True
584 skipstep3 = True
585
585
586 if not exact and self._checkcase:
586 if not exact and self._checkcase:
587 normalize = self._normalize
587 normalize = self._normalize
588 skipstep3 = False
588 skipstep3 = False
589 else:
589 else:
590 normalize = None
590 normalize = None
591
591
592 files = sorted(match.files())
592 files = sorted(match.files())
593 subrepos.sort()
593 subrepos.sort()
594 i, j = 0, 0
594 i, j = 0, 0
595 while i < len(files) and j < len(subrepos):
595 while i < len(files) and j < len(subrepos):
596 subpath = subrepos[j] + "/"
596 subpath = subrepos[j] + "/"
597 if files[i] < subpath:
597 if files[i] < subpath:
598 i += 1
598 i += 1
599 continue
599 continue
600 while i < len(files) and files[i].startswith(subpath):
600 while i < len(files) and files[i].startswith(subpath):
601 del files[i]
601 del files[i]
602 j += 1
602 j += 1
603
603
604 if not files or '.' in files:
604 if not files or '.' in files:
605 files = ['']
605 files = ['']
606 results = dict.fromkeys(subrepos)
606 results = dict.fromkeys(subrepos)
607 results['.hg'] = None
607 results['.hg'] = None
608
608
609 # step 1: find all explicit files
609 # step 1: find all explicit files
610 dirsnotfound = []
610 for ff in files:
611 for ff in files:
611 if normalize:
612 if normalize:
612 nf = normalize(normpath(ff), False, True)
613 nf = normalize(normpath(ff), False, True)
613 else:
614 else:
614 nf = normpath(ff)
615 nf = normpath(ff)
615 if nf in results:
616 if nf in results:
616 continue
617 continue
617
618
618 try:
619 try:
619 st = lstat(join(nf))
620 st = lstat(join(nf))
620 kind = getkind(st.st_mode)
621 kind = getkind(st.st_mode)
621 if kind == dirkind:
622 if kind == dirkind:
622 skipstep3 = False
623 skipstep3 = False
623 if nf in dmap:
624 if nf in dmap:
624 #file deleted on disk but still in dirstate
625 #file deleted on disk but still in dirstate
625 results[nf] = None
626 results[nf] = None
626 if matchedir:
627 if matchedir:
627 matchedir(nf)
628 matchedir(nf)
628 if not dirignore(nf):
629 if not dirignore(nf):
629 wadd(nf)
630 wadd(nf)
630 elif kind == regkind or kind == lnkkind:
631 elif kind == regkind or kind == lnkkind:
631 results[nf] = st
632 results[nf] = st
632 else:
633 else:
633 badfn(ff, badtype(kind))
634 badfn(ff, badtype(kind))
634 if nf in dmap:
635 if nf in dmap:
635 results[nf] = None
636 results[nf] = None
636 except OSError, inst:
637 except OSError, inst:
637 if nf in dmap: # does it exactly match a file?
638 if nf in dmap: # does it exactly match a file?
638 results[nf] = None
639 results[nf] = None
639 else: # does it match a directory?
640 else: # does it match a directory?
640 prefix = nf + "/"
641 prefix = nf + "/"
641 for fn in dmap:
642 for fn in dmap:
642 if fn.startswith(prefix):
643 if fn.startswith(prefix):
643 if matchedir:
644 if matchedir:
644 matchedir(nf)
645 matchedir(nf)
646 dirsnotfound.append(nf)
645 skipstep3 = False
647 skipstep3 = False
646 break
648 break
647 else:
649 else:
648 badfn(ff, inst.strerror)
650 badfn(ff, inst.strerror)
649
651
650 # step 2: visit subdirectories
652 # step 2: visit subdirectories
651 while work:
653 while work:
652 nd = work.pop()
654 nd = work.pop()
653 skip = None
655 skip = None
654 if nd == '.':
656 if nd == '.':
655 nd = ''
657 nd = ''
656 else:
658 else:
657 skip = '.hg'
659 skip = '.hg'
658 try:
660 try:
659 entries = listdir(join(nd), stat=True, skip=skip)
661 entries = listdir(join(nd), stat=True, skip=skip)
660 except OSError, inst:
662 except OSError, inst:
661 if inst.errno in (errno.EACCES, errno.ENOENT):
663 if inst.errno in (errno.EACCES, errno.ENOENT):
662 fwarn(nd, inst.strerror)
664 fwarn(nd, inst.strerror)
663 continue
665 continue
664 raise
666 raise
665 for f, kind, st in entries:
667 for f, kind, st in entries:
666 if normalize:
668 if normalize:
667 nf = normalize(nd and (nd + "/" + f) or f, True, True)
669 nf = normalize(nd and (nd + "/" + f) or f, True, True)
668 else:
670 else:
669 nf = nd and (nd + "/" + f) or f
671 nf = nd and (nd + "/" + f) or f
670 if nf not in results:
672 if nf not in results:
671 if kind == dirkind:
673 if kind == dirkind:
672 if not ignore(nf):
674 if not ignore(nf):
673 if matchtdir:
675 if matchtdir:
674 matchtdir(nf)
676 matchtdir(nf)
675 wadd(nf)
677 wadd(nf)
676 if nf in dmap and (matchalways or matchfn(nf)):
678 if nf in dmap and (matchalways or matchfn(nf)):
677 results[nf] = None
679 results[nf] = None
678 elif kind == regkind or kind == lnkkind:
680 elif kind == regkind or kind == lnkkind:
679 if nf in dmap:
681 if nf in dmap:
680 if matchalways or matchfn(nf):
682 if matchalways or matchfn(nf):
681 results[nf] = st
683 results[nf] = st
682 elif (matchalways or matchfn(nf)) and not ignore(nf):
684 elif (matchalways or matchfn(nf)) and not ignore(nf):
683 results[nf] = st
685 results[nf] = st
684 elif nf in dmap and (matchalways or matchfn(nf)):
686 elif nf in dmap and (matchalways or matchfn(nf)):
685 results[nf] = None
687 results[nf] = None
686
688
687 for s in subrepos:
689 for s in subrepos:
688 del results[s]
690 del results[s]
689 del results['.hg']
691 del results['.hg']
690
692
691 # step 3: report unseen items in the dmap hash
693 # step 3: report unseen items in the dmap hash
692 if not skipstep3 and not exact:
694 if not skipstep3 and not exact:
693 if not results and matchalways:
695 if not results and matchalways:
694 visit = dmap.keys()
696 visit = dmap.keys()
695 else:
697 else:
696 visit = [f for f in dmap if f not in results and matchfn(f)]
698 visit = [f for f in dmap if f not in results and matchfn(f)]
697 visit.sort()
699 visit.sort()
698
700
699 if unknown:
701 if unknown:
700 # unknown == True means we walked the full directory tree above.
702 # unknown == True means we walked the full directory tree above.
701 # So if a file is not seen it was either a) not matching matchfn
703 # So if a file is not seen it was either a) not matching matchfn
702 # b) ignored, c) missing, or d) under a symlink directory.
704 # b) ignored, c) missing, or d) under a symlink directory.
703 audit_path = scmutil.pathauditor(self._root)
705 audit_path = scmutil.pathauditor(self._root)
704
706
705 for nf in iter(visit):
707 for nf in iter(visit):
706 # Report ignored items in the dmap as long as they are not
708 # Report ignored items in the dmap as long as they are not
707 # under a symlink directory.
709 # under a symlink directory.
708 if audit_path.check(nf):
710 if audit_path.check(nf):
709 try:
711 try:
710 results[nf] = lstat(join(nf))
712 results[nf] = lstat(join(nf))
711 except OSError:
713 except OSError:
712 # file doesn't exist
714 # file doesn't exist
713 results[nf] = None
715 results[nf] = None
714 else:
716 else:
715 # It's either missing or under a symlink directory
717 # It's either missing or under a symlink directory
716 results[nf] = None
718 results[nf] = None
717 else:
719 else:
718 # We may not have walked the full directory tree above,
720 # We may not have walked the full directory tree above,
719 # so stat everything we missed.
721 # so stat everything we missed.
720 nf = iter(visit).next
722 nf = iter(visit).next
721 for st in util.statfiles([join(i) for i in visit]):
723 for st in util.statfiles([join(i) for i in visit]):
722 results[nf()] = st
724 results[nf()] = st
723 return results
725 return results
724
726
725 def status(self, match, subrepos, ignored, clean, unknown):
727 def status(self, match, subrepos, ignored, clean, unknown):
726 '''Determine the status of the working copy relative to the
728 '''Determine the status of the working copy relative to the
727 dirstate and return a tuple of lists (unsure, modified, added,
729 dirstate and return a tuple of lists (unsure, modified, added,
728 removed, deleted, unknown, ignored, clean), where:
730 removed, deleted, unknown, ignored, clean), where:
729
731
730 unsure:
732 unsure:
731 files that might have been modified since the dirstate was
733 files that might have been modified since the dirstate was
732 written, but need to be read to be sure (size is the same
734 written, but need to be read to be sure (size is the same
733 but mtime differs)
735 but mtime differs)
734 modified:
736 modified:
735 files that have definitely been modified since the dirstate
737 files that have definitely been modified since the dirstate
736 was written (different size or mode)
738 was written (different size or mode)
737 added:
739 added:
738 files that have been explicitly added with hg add
740 files that have been explicitly added with hg add
739 removed:
741 removed:
740 files that have been explicitly removed with hg remove
742 files that have been explicitly removed with hg remove
741 deleted:
743 deleted:
742 files that have been deleted through other means ("missing")
744 files that have been deleted through other means ("missing")
743 unknown:
745 unknown:
744 files not in the dirstate that are not ignored
746 files not in the dirstate that are not ignored
745 ignored:
747 ignored:
746 files not in the dirstate that are ignored
748 files not in the dirstate that are ignored
747 (by _dirignore())
749 (by _dirignore())
748 clean:
750 clean:
749 files that have definitely not been modified since the
751 files that have definitely not been modified since the
750 dirstate was written
752 dirstate was written
751 '''
753 '''
752 listignored, listclean, listunknown = ignored, clean, unknown
754 listignored, listclean, listunknown = ignored, clean, unknown
753 lookup, modified, added, unknown, ignored = [], [], [], [], []
755 lookup, modified, added, unknown, ignored = [], [], [], [], []
754 removed, deleted, clean = [], [], []
756 removed, deleted, clean = [], [], []
755
757
756 dmap = self._map
758 dmap = self._map
757 ladd = lookup.append # aka "unsure"
759 ladd = lookup.append # aka "unsure"
758 madd = modified.append
760 madd = modified.append
759 aadd = added.append
761 aadd = added.append
760 uadd = unknown.append
762 uadd = unknown.append
761 iadd = ignored.append
763 iadd = ignored.append
762 radd = removed.append
764 radd = removed.append
763 dadd = deleted.append
765 dadd = deleted.append
764 cadd = clean.append
766 cadd = clean.append
765 mexact = match.exact
767 mexact = match.exact
766 dirignore = self._dirignore
768 dirignore = self._dirignore
767 checkexec = self._checkexec
769 checkexec = self._checkexec
768 checklink = self._checklink
770 checklink = self._checklink
769 copymap = self._copymap
771 copymap = self._copymap
770 lastnormaltime = self._lastnormaltime
772 lastnormaltime = self._lastnormaltime
771
773
772 lnkkind = stat.S_IFLNK
774 lnkkind = stat.S_IFLNK
773
775
774 for fn, st in self.walk(match, subrepos, listunknown,
776 for fn, st in self.walk(match, subrepos, listunknown,
775 listignored).iteritems():
777 listignored).iteritems():
776 if fn not in dmap:
778 if fn not in dmap:
777 if (listignored or mexact(fn)) and dirignore(fn):
779 if (listignored or mexact(fn)) and dirignore(fn):
778 if listignored:
780 if listignored:
779 iadd(fn)
781 iadd(fn)
780 elif listunknown:
782 elif listunknown:
781 uadd(fn)
783 uadd(fn)
782 continue
784 continue
783
785
784 state, mode, size, time = dmap[fn]
786 state, mode, size, time = dmap[fn]
785
787
786 if not st and state in "nma":
788 if not st and state in "nma":
787 dadd(fn)
789 dadd(fn)
788 elif state == 'n':
790 elif state == 'n':
789 # The "mode & lnkkind != lnkkind or self._checklink"
791 # The "mode & lnkkind != lnkkind or self._checklink"
790 # lines are an expansion of "islink => checklink"
792 # lines are an expansion of "islink => checklink"
791 # where islink means "is this a link?" and checklink
793 # where islink means "is this a link?" and checklink
792 # means "can we check links?".
794 # means "can we check links?".
793 mtime = int(st.st_mtime)
795 mtime = int(st.st_mtime)
794 if (size >= 0 and
796 if (size >= 0 and
795 ((size != st.st_size and size != st.st_size & _rangemask)
797 ((size != st.st_size and size != st.st_size & _rangemask)
796 or ((mode ^ st.st_mode) & 0100 and checkexec))
798 or ((mode ^ st.st_mode) & 0100 and checkexec))
797 and (mode & lnkkind != lnkkind or checklink)
799 and (mode & lnkkind != lnkkind or checklink)
798 or size == -2 # other parent
800 or size == -2 # other parent
799 or fn in copymap):
801 or fn in copymap):
800 madd(fn)
802 madd(fn)
801 elif ((time != mtime and time != mtime & _rangemask)
803 elif ((time != mtime and time != mtime & _rangemask)
802 and (mode & lnkkind != lnkkind or checklink)):
804 and (mode & lnkkind != lnkkind or checklink)):
803 ladd(fn)
805 ladd(fn)
804 elif mtime == lastnormaltime:
806 elif mtime == lastnormaltime:
805 # fn may have been changed in the same timeslot without
807 # fn may have been changed in the same timeslot without
806 # changing its size. This can happen if we quickly do
808 # changing its size. This can happen if we quickly do
807 # multiple commits in a single transaction.
809 # multiple commits in a single transaction.
808 # Force lookup, so we don't miss such a racy file change.
810 # Force lookup, so we don't miss such a racy file change.
809 ladd(fn)
811 ladd(fn)
810 elif listclean:
812 elif listclean:
811 cadd(fn)
813 cadd(fn)
812 elif state == 'm':
814 elif state == 'm':
813 madd(fn)
815 madd(fn)
814 elif state == 'a':
816 elif state == 'a':
815 aadd(fn)
817 aadd(fn)
816 elif state == 'r':
818 elif state == 'r':
817 radd(fn)
819 radd(fn)
818
820
819 return (lookup, modified, added, removed, deleted, unknown, ignored,
821 return (lookup, modified, added, removed, deleted, unknown, ignored,
820 clean)
822 clean)
General Comments 0
You need to be logged in to leave comments. Login now