##// END OF EJS Templates
dirstate.walk: fold findfiles into main walk loop
Matt Mackall -
r6820:639d9cb9 default
parent child Browse files
Show More
@@ -1,623 +1,626 b''
1 1 """
2 2 dirstate.py - working directory tracking for mercurial
3 3
4 4 Copyright 2005-2007 Matt Mackall <mpm@selenic.com>
5 5
6 6 This software may be used and distributed according to the terms
7 7 of the GNU General Public License, incorporated herein by reference.
8 8 """
9 9
10 10 from node import nullid
11 11 from i18n import _
12 12 import struct, os, bisect, stat, util, errno, ignore
13 13 import cStringIO, osutil, sys
14 14
15 15 _unknown = ('?', 0, 0, 0)
16 16 _format = ">cllll"
17 17
18 18 def _finddirs(path):
19 19 pos = len(path)
20 20 while 1:
21 21 pos = path.rfind('/', 0, pos)
22 22 if pos == -1:
23 23 break
24 24 yield path[:pos]
25 25
26 26 class dirstate(object):
27 27
28 28 def __init__(self, opener, ui, root):
29 29 self._opener = opener
30 30 self._root = root
31 31 self._dirty = False
32 32 self._dirtypl = False
33 33 self._ui = ui
34 34
35 35 def __getattr__(self, name):
36 36 if name == '_map':
37 37 self._read()
38 38 return self._map
39 39 elif name == '_copymap':
40 40 self._read()
41 41 return self._copymap
42 42 elif name == '_foldmap':
43 43 _foldmap = {}
44 44 for name in self._map:
45 45 norm = os.path.normcase(os.path.normpath(name))
46 46 _foldmap[norm] = name
47 47 self._foldmap = _foldmap
48 48 return self._foldmap
49 49 elif name == '_branch':
50 50 try:
51 51 self._branch = (self._opener("branch").read().strip()
52 52 or "default")
53 53 except IOError:
54 54 self._branch = "default"
55 55 return self._branch
56 56 elif name == '_pl':
57 57 self._pl = [nullid, nullid]
58 58 try:
59 59 st = self._opener("dirstate").read(40)
60 60 if len(st) == 40:
61 61 self._pl = st[:20], st[20:40]
62 62 except IOError, err:
63 63 if err.errno != errno.ENOENT: raise
64 64 return self._pl
65 65 elif name == '_dirs':
66 66 dirs = {}
67 67 for f,s in self._map.items():
68 68 if s[0] != 'r':
69 69 for base in _finddirs(f):
70 70 dirs[base] = dirs.get(base, 0) + 1
71 71 self._dirs = dirs
72 72 return self._dirs
73 73 elif name == '_ignore':
74 74 files = [self._join('.hgignore')]
75 75 for name, path in self._ui.configitems("ui"):
76 76 if name == 'ignore' or name.startswith('ignore.'):
77 77 files.append(os.path.expanduser(path))
78 78 self._ignore = ignore.ignore(self._root, files, self._ui.warn)
79 79 return self._ignore
80 80 elif name == '_slash':
81 81 self._slash = self._ui.configbool('ui', 'slash') and os.sep != '/'
82 82 return self._slash
83 83 elif name == '_checklink':
84 84 self._checklink = util.checklink(self._root)
85 85 return self._checklink
86 86 elif name == '_checkexec':
87 87 self._checkexec = util.checkexec(self._root)
88 88 return self._checkexec
89 89 elif name == '_checkcase':
90 90 self._checkcase = not util.checkcase(self._join('.hg'))
91 91 return self._checkcase
92 92 elif name == 'normalize':
93 93 if self._checkcase:
94 94 self.normalize = self._normalize
95 95 else:
96 96 self.normalize = lambda x: x
97 97 return self.normalize
98 98 else:
99 99 raise AttributeError, name
100 100
101 101 def _join(self, f):
102 102 return os.path.join(self._root, f)
103 103
104 104 def flagfunc(self, fallback):
105 105 if self._checklink:
106 106 if self._checkexec:
107 107 def f(x):
108 108 p = os.path.join(self._root, x)
109 109 if os.path.islink(p):
110 110 return 'l'
111 111 if util.is_exec(p):
112 112 return 'x'
113 113 return ''
114 114 return f
115 115 def f(x):
116 116 if os.path.islink(os.path.join(self._root, x)):
117 117 return 'l'
118 118 if 'x' in fallback(x):
119 119 return 'x'
120 120 return ''
121 121 return f
122 122 if self._checkexec:
123 123 def f(x):
124 124 if 'l' in fallback(x):
125 125 return 'l'
126 126 if util.is_exec(os.path.join(self._root, x)):
127 127 return 'x'
128 128 return ''
129 129 return f
130 130 return fallback
131 131
132 132 def getcwd(self):
133 133 cwd = os.getcwd()
134 134 if cwd == self._root: return ''
135 135 # self._root ends with a path separator if self._root is '/' or 'C:\'
136 136 rootsep = self._root
137 137 if not util.endswithsep(rootsep):
138 138 rootsep += os.sep
139 139 if cwd.startswith(rootsep):
140 140 return cwd[len(rootsep):]
141 141 else:
142 142 # we're outside the repo. return an absolute path.
143 143 return cwd
144 144
145 145 def pathto(self, f, cwd=None):
146 146 if cwd is None:
147 147 cwd = self.getcwd()
148 148 path = util.pathto(self._root, cwd, f)
149 149 if self._slash:
150 150 return util.normpath(path)
151 151 return path
152 152
153 153 def __getitem__(self, key):
154 154 ''' current states:
155 155 n normal
156 156 m needs merging
157 157 r marked for removal
158 158 a marked for addition
159 159 ? not tracked'''
160 160 return self._map.get(key, ("?",))[0]
161 161
162 162 def __contains__(self, key):
163 163 return key in self._map
164 164
165 165 def __iter__(self):
166 166 for x in util.sort(self._map):
167 167 yield x
168 168
169 169 def parents(self):
170 170 return self._pl
171 171
172 172 def branch(self):
173 173 return self._branch
174 174
175 175 def setparents(self, p1, p2=nullid):
176 176 self._dirty = self._dirtypl = True
177 177 self._pl = p1, p2
178 178
179 179 def setbranch(self, branch):
180 180 self._branch = branch
181 181 self._opener("branch", "w").write(branch + '\n')
182 182
183 183 def _read(self):
184 184 self._map = {}
185 185 self._copymap = {}
186 186 if not self._dirtypl:
187 187 self._pl = [nullid, nullid]
188 188 try:
189 189 st = self._opener("dirstate").read()
190 190 except IOError, err:
191 191 if err.errno != errno.ENOENT: raise
192 192 return
193 193 if not st:
194 194 return
195 195
196 196 if not self._dirtypl:
197 197 self._pl = [st[:20], st[20: 40]]
198 198
199 199 # deref fields so they will be local in loop
200 200 dmap = self._map
201 201 copymap = self._copymap
202 202 unpack = struct.unpack
203 203 e_size = struct.calcsize(_format)
204 204 pos1 = 40
205 205 l = len(st)
206 206
207 207 # the inner loop
208 208 while pos1 < l:
209 209 pos2 = pos1 + e_size
210 210 e = unpack(">cllll", st[pos1:pos2]) # a literal here is faster
211 211 pos1 = pos2 + e[4]
212 212 f = st[pos2:pos1]
213 213 if '\0' in f:
214 214 f, c = f.split('\0')
215 215 copymap[f] = c
216 216 dmap[f] = e # we hold onto e[4] because making a subtuple is slow
217 217
218 218 def invalidate(self):
219 219 for a in "_map _copymap _foldmap _branch _pl _dirs _ignore".split():
220 220 if a in self.__dict__:
221 221 delattr(self, a)
222 222 self._dirty = False
223 223
224 224 def copy(self, source, dest):
225 225 if source == dest:
226 226 return
227 227 self._dirty = True
228 228 self._copymap[dest] = source
229 229
230 230 def copied(self, file):
231 231 return self._copymap.get(file, None)
232 232
233 233 def copies(self):
234 234 return self._copymap
235 235
236 236 def _droppath(self, f):
237 237 if self[f] not in "?r" and "_dirs" in self.__dict__:
238 238 dirs = self._dirs
239 239 for base in _finddirs(f):
240 240 if dirs[base] == 1:
241 241 del dirs[base]
242 242 else:
243 243 dirs[base] -= 1
244 244
245 245 def _addpath(self, f, check=False):
246 246 oldstate = self[f]
247 247 if check or oldstate == "r":
248 248 if '\r' in f or '\n' in f:
249 249 raise util.Abort(
250 250 _("'\\n' and '\\r' disallowed in filenames: %r") % f)
251 251 if f in self._dirs:
252 252 raise util.Abort(_('directory %r already in dirstate') % f)
253 253 # shadows
254 254 for d in _finddirs(f):
255 255 if d in self._dirs:
256 256 break
257 257 if d in self._map and self[d] != 'r':
258 258 raise util.Abort(
259 259 _('file %r in dirstate clashes with %r') % (d, f))
260 260 if oldstate in "?r" and "_dirs" in self.__dict__:
261 261 dirs = self._dirs
262 262 for base in _finddirs(f):
263 263 dirs[base] = dirs.get(base, 0) + 1
264 264
265 265 def normal(self, f):
266 266 'mark a file normal and clean'
267 267 self._dirty = True
268 268 self._addpath(f)
269 269 s = os.lstat(self._join(f))
270 270 self._map[f] = ('n', s.st_mode, s.st_size, s.st_mtime, 0)
271 271 if f in self._copymap:
272 272 del self._copymap[f]
273 273
274 274 def normallookup(self, f):
275 275 'mark a file normal, but possibly dirty'
276 276 if self._pl[1] != nullid and f in self._map:
277 277 # if there is a merge going on and the file was either
278 278 # in state 'm' or dirty before being removed, restore that state.
279 279 entry = self._map[f]
280 280 if entry[0] == 'r' and entry[2] in (-1, -2):
281 281 source = self._copymap.get(f)
282 282 if entry[2] == -1:
283 283 self.merge(f)
284 284 elif entry[2] == -2:
285 285 self.normaldirty(f)
286 286 if source:
287 287 self.copy(source, f)
288 288 return
289 289 if entry[0] == 'm' or entry[0] == 'n' and entry[2] == -2:
290 290 return
291 291 self._dirty = True
292 292 self._addpath(f)
293 293 self._map[f] = ('n', 0, -1, -1, 0)
294 294 if f in self._copymap:
295 295 del self._copymap[f]
296 296
297 297 def normaldirty(self, f):
298 298 'mark a file normal, but dirty'
299 299 self._dirty = True
300 300 self._addpath(f)
301 301 self._map[f] = ('n', 0, -2, -1, 0)
302 302 if f in self._copymap:
303 303 del self._copymap[f]
304 304
305 305 def add(self, f):
306 306 'mark a file added'
307 307 self._dirty = True
308 308 self._addpath(f, True)
309 309 self._map[f] = ('a', 0, -1, -1, 0)
310 310 if f in self._copymap:
311 311 del self._copymap[f]
312 312
313 313 def remove(self, f):
314 314 'mark a file removed'
315 315 self._dirty = True
316 316 self._droppath(f)
317 317 size = 0
318 318 if self._pl[1] != nullid and f in self._map:
319 319 entry = self._map[f]
320 320 if entry[0] == 'm':
321 321 size = -1
322 322 elif entry[0] == 'n' and entry[2] == -2:
323 323 size = -2
324 324 self._map[f] = ('r', 0, size, 0, 0)
325 325 if size == 0 and f in self._copymap:
326 326 del self._copymap[f]
327 327
328 328 def merge(self, f):
329 329 'mark a file merged'
330 330 self._dirty = True
331 331 s = os.lstat(self._join(f))
332 332 self._addpath(f)
333 333 self._map[f] = ('m', s.st_mode, s.st_size, s.st_mtime, 0)
334 334 if f in self._copymap:
335 335 del self._copymap[f]
336 336
337 337 def forget(self, f):
338 338 'forget a file'
339 339 self._dirty = True
340 340 try:
341 341 self._droppath(f)
342 342 del self._map[f]
343 343 except KeyError:
344 344 self._ui.warn(_("not in dirstate: %s\n") % f)
345 345
346 346 def _normalize(self, path):
347 347 normpath = os.path.normcase(os.path.normpath(path))
348 348 if normpath in self._foldmap:
349 349 return self._foldmap[normpath]
350 350 elif os.path.exists(path):
351 351 self._foldmap[normpath] = util.fspath(path, self._root)
352 352 return self._foldmap[normpath]
353 353 else:
354 354 return path
355 355
356 356 def clear(self):
357 357 self._map = {}
358 358 if "_dirs" in self.__dict__:
359 359 delattr(self, "_dirs");
360 360 self._copymap = {}
361 361 self._pl = [nullid, nullid]
362 362 self._dirty = True
363 363
364 364 def rebuild(self, parent, files):
365 365 self.clear()
366 366 for f in files:
367 367 if 'x' in files.flags(f):
368 368 self._map[f] = ('n', 0777, -1, 0, 0)
369 369 else:
370 370 self._map[f] = ('n', 0666, -1, 0, 0)
371 371 self._pl = (parent, nullid)
372 372 self._dirty = True
373 373
374 374 def write(self):
375 375 if not self._dirty:
376 376 return
377 377 st = self._opener("dirstate", "w", atomictemp=True)
378 378
379 379 try:
380 380 gran = int(self._ui.config('dirstate', 'granularity', 1))
381 381 except ValueError:
382 382 gran = 1
383 383 limit = sys.maxint
384 384 if gran > 0:
385 385 limit = util.fstat(st).st_mtime - gran
386 386
387 387 cs = cStringIO.StringIO()
388 388 copymap = self._copymap
389 389 pack = struct.pack
390 390 write = cs.write
391 391 write("".join(self._pl))
392 392 for f, e in self._map.iteritems():
393 393 if f in copymap:
394 394 f = "%s\0%s" % (f, copymap[f])
395 395 if e[3] > limit and e[0] == 'n':
396 396 e = (e[0], 0, -1, -1, 0)
397 397 e = pack(_format, e[0], e[1], e[2], e[3], len(f))
398 398 write(e)
399 399 write(f)
400 400 st.write(cs.getvalue())
401 401 st.rename()
402 402 self._dirty = self._dirtypl = False
403 403
404 404 def _supported(self, f, mode, verbose=False):
405 405 if stat.S_ISREG(mode) or stat.S_ISLNK(mode):
406 406 return True
407 407 if verbose:
408 408 kind = 'unknown'
409 409 if stat.S_ISCHR(mode): kind = _('character device')
410 410 elif stat.S_ISBLK(mode): kind = _('block device')
411 411 elif stat.S_ISFIFO(mode): kind = _('fifo')
412 412 elif stat.S_ISSOCK(mode): kind = _('socket')
413 413 elif stat.S_ISDIR(mode): kind = _('directory')
414 414 self._ui.warn(_('%s: unsupported file type (type is %s)\n')
415 415 % (self.pathto(f), kind))
416 416 return False
417 417
418 418 def _dirignore(self, f):
419 419 if f == '.':
420 420 return False
421 421 if self._ignore(f):
422 422 return True
423 423 for p in _finddirs(f):
424 424 if self._ignore(p):
425 425 return True
426 426 return False
427 427
428 428 def walk(self, match, unknown, ignored):
429 429 '''
430 430 walk recursively through the directory tree, finding all files
431 431 matched by the match function
432 432
433 433 results are yielded in a tuple (filename, stat), where stat
434 434 and st is the stat result if the file was found in the directory.
435 435 '''
436 436
437 437 def fwarn(f, msg):
438 438 self._ui.warn('%s: %s\n' % (self.pathto(ff), msg))
439 439 return False
440 440 badfn = fwarn
441 441 if hasattr(match, 'bad'):
442 442 badfn = match.bad
443 443
444 444 files = util.unique(match.files())
445 445 if not files or '.' in files:
446 446 files = ['']
447 447 dc = self._map
448 448
449 449 def imatch(file_):
450 450 if file_ not in dc and self._ignore(file_):
451 451 return False
452 452 return match(file_)
453 453
454 454 # TODO: don't walk unknown directories if unknown and ignored are False
455 455 ignore = self._ignore
456 456 dirignore = self._dirignore
457 457 if ignored:
458 458 imatch = match
459 459 ignore = util.never
460 460 dirignore = util.never
461 461
462 462 # self._root may end with a path separator when self._root == '/'
463 463 common_prefix_len = len(self._root)
464 464 if not util.endswithsep(self._root):
465 465 common_prefix_len += 1
466 466
467 467 normpath = util.normpath
468 468 listdir = osutil.listdir
469 469 lstat = os.lstat
470 470 bisect_left = bisect.bisect_left
471 471 isdir = os.path.isdir
472 472 pconvert = util.pconvert
473 473 join = os.path.join
474 474 s_isdir = stat.S_ISDIR
475 475 supported = self._supported
476 476 _join = self._join
477 work = []
478 wadd = work.append
479 found = []
480 add = found.append
481
477 482 known = {'.hg': 1}
478 483
479 # recursion free walker, faster than os.walk.
480 def findfiles(s):
481 work = [s]
482 wadd = work.append
483 found = []
484 add = found.append
484 # step one, find all files that match our criteria
485 for ff in util.sort(files):
486 nf = normpath(ff)
487 nn = self.normalize(nf)
488 f = _join(ff)
489 if nn in known:
490 continue
491
492 try:
493 st = lstat(f)
494 except OSError, inst:
495 keep = False
496 for fn in dc:
497 if nf == fn or (fn.startswith(nf) and fn[len(nf)] == '/'):
498 keep = True
499 break
500 if not keep:
501 if inst.errno != errno.ENOENT:
502 fwarn(ff, inst.strerror)
503 elif badfn(ff, inst.strerror) and imatch(nf):
504 yield ff, None
505 continue
506
507 if not s_isdir(st.st_mode):
508 known[nn] = 1
509 if supported(ff, st.st_mode, verbose=True):
510 yield self.normalize(nf), st
511 elif ff in dc:
512 yield nf, None
513 continue
514
515 if dirignore(nf):
516 continue
517
485 518 if hasattr(match, 'dir'):
486 match.dir(normpath(s[common_prefix_len:]))
519 match.dir(normpath(f[common_prefix_len:]))
520 wadd(f)
487 521 while work:
488 522 top = work.pop()
489 523 entries = listdir(top, stat=True)
490 524 # nd is the top of the repository dir tree
491 525 nd = normpath(top[common_prefix_len:])
492 526 if nd == '.':
493 527 nd = ''
494 528 else:
495 529 # do not recurse into a repo contained in this
496 530 # one. use bisect to find .hg directory so speed
497 531 # is good on big directory.
498 532 names = [e[0] for e in entries]
499 533 hg = bisect_left(names, '.hg')
500 534 if hg < len(names) and names[hg] == '.hg':
501 535 if isdir(join(top, '.hg')):
502 536 continue
503 537 for f, kind, st in entries:
504 538 np = pconvert(join(nd, f))
505 539 nn = self.normalize(np)
506 540 if np in known:
507 541 continue
508 542 known[nn] = 1
509 543 p = join(top, f)
510 544 # don't trip over symlinks
511 545 if kind == stat.S_IFDIR:
512 546 if not ignore(np):
513 547 wadd(p)
514 548 if hasattr(match, 'dir'):
515 549 match.dir(np)
516 550 if np in dc and match(np):
517 551 add((nn, None))
518 552 elif imatch(np):
519 553 if supported(np, st.st_mode):
520 554 add((nn, st))
521 555 elif np in dc:
522 556 add((nn, None))
523 return util.sort(found)
524
525 # step one, find all files that match our criteria
526 for ff in util.sort(files):
527 nf = normpath(ff)
528 nn = self.normalize(nf)
529 f = _join(ff)
530 try:
531 st = lstat(f)
532 except OSError, inst:
533 found = False
534 for fn in dc:
535 if nf == fn or (fn.startswith(nf) and fn[len(nf)] == '/'):
536 found = True
537 break
538 if not found:
539 if inst.errno != errno.ENOENT:
540 fwarn(ff, inst.strerror)
541 elif badfn(ff, inst.strerror) and imatch(nf):
542 yield ff, None
543 continue
544 if s_isdir(st.st_mode):
545 if not dirignore(nf):
546 for e in findfiles(f):
547 yield e
548 else:
549 if nn in known:
550 continue
551 known[nn] = 1
552 if supported(ff, st.st_mode, verbose=True):
553 yield self.normalize(nf), st
554 elif ff in dc:
555 yield nf, None
557 for e in util.sort(found):
558 yield e
556 559
557 560 # step two run through anything left in the dc hash and yield
558 561 # if we haven't already seen it
559 562 for f in util.sort(dc):
560 563 if f in known or not match(f):
561 564 continue
562 565 known[f] = 1
563 566 try:
564 567 st = lstat(_join(f))
565 568 if supported(f, st.st_mode):
566 569 yield f, st
567 570 continue
568 571 except OSError, inst:
569 572 if inst.errno not in (errno.ENOENT, errno.ENOTDIR):
570 573 raise
571 574 yield f, None
572 575
573 576 def status(self, match, ignored, clean, unknown):
574 577 listignored, listclean, listunknown = ignored, clean, unknown
575 578 lookup, modified, added, unknown, ignored = [], [], [], [], []
576 579 removed, deleted, clean = [], [], []
577 580
578 581 _join = self._join
579 582 lstat = os.lstat
580 583 cmap = self._copymap
581 584 dmap = self._map
582 585 ladd = lookup.append
583 586 madd = modified.append
584 587 aadd = added.append
585 588 uadd = unknown.append
586 589 iadd = ignored.append
587 590 radd = removed.append
588 591 dadd = deleted.append
589 592 cadd = clean.append
590 593
591 594 for fn, st in self.walk(match, listunknown, listignored):
592 595 if fn not in dmap:
593 596 if (listignored or match.exact(fn)) and self._dirignore(fn):
594 597 if listignored:
595 598 iadd(fn)
596 599 elif listunknown:
597 600 uadd(fn)
598 601 continue
599 602
600 603 state, mode, size, time, foo = dmap[fn]
601 604
602 605 if not st and state in "nma":
603 606 dadd(fn)
604 607 elif state == 'n':
605 608 if (size >= 0 and
606 609 (size != st.st_size
607 610 or ((mode ^ st.st_mode) & 0100 and self._checkexec))
608 611 or size == -2
609 612 or fn in self._copymap):
610 613 madd(fn)
611 614 elif time != int(st.st_mtime):
612 615 ladd(fn)
613 616 elif listclean:
614 617 cadd(fn)
615 618 elif state == 'm':
616 619 madd(fn)
617 620 elif state == 'a':
618 621 aadd(fn)
619 622 elif state == 'r':
620 623 radd(fn)
621 624
622 625 return (lookup, modified, added, removed, deleted, unknown, ignored,
623 626 clean)
General Comments 0
You need to be logged in to leave comments. Login now