##// END OF EJS Templates
dirstate.walk: eliminate filter function...
Matt Mackall -
r6819:78624d74 default
parent child Browse files
Show More
@@ -1,658 +1,623 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 def _filter(self, files):
405 ret = {}
406 unknown = []
407
408 for x in files:
409 if x == '.':
410 return self._map.copy()
411 if x not in self._map:
412 unknown.append(x)
413 else:
414 ret[x] = self._map[x]
415
416 if not unknown:
417 return ret
418
419 b = util.sort(self._map)
420 blen = len(b)
421
422 for x in unknown:
423 bs = bisect.bisect(b, "%s%s" % (x, '/'))
424 while bs < blen:
425 s = b[bs]
426 if len(s) > len(x) and s.startswith(x):
427 ret[s] = self._map[s]
428 else:
429 break
430 bs += 1
431 return ret
432
433 404 def _supported(self, f, mode, verbose=False):
434 405 if stat.S_ISREG(mode) or stat.S_ISLNK(mode):
435 406 return True
436 407 if verbose:
437 408 kind = 'unknown'
438 409 if stat.S_ISCHR(mode): kind = _('character device')
439 410 elif stat.S_ISBLK(mode): kind = _('block device')
440 411 elif stat.S_ISFIFO(mode): kind = _('fifo')
441 412 elif stat.S_ISSOCK(mode): kind = _('socket')
442 413 elif stat.S_ISDIR(mode): kind = _('directory')
443 414 self._ui.warn(_('%s: unsupported file type (type is %s)\n')
444 415 % (self.pathto(f), kind))
445 416 return False
446 417
447 418 def _dirignore(self, f):
448 419 if f == '.':
449 420 return False
450 421 if self._ignore(f):
451 422 return True
452 423 for p in _finddirs(f):
453 424 if self._ignore(p):
454 425 return True
455 426 return False
456 427
457 428 def walk(self, match, unknown, ignored):
458 429 '''
459 430 walk recursively through the directory tree, finding all files
460 431 matched by the match function
461 432
462 433 results are yielded in a tuple (filename, stat), where stat
463 434 and st is the stat result if the file was found in the directory.
464 435 '''
465 436
466 437 def fwarn(f, msg):
467 438 self._ui.warn('%s: %s\n' % (self.pathto(ff), msg))
468 439 return False
469 440 badfn = fwarn
470 441 if hasattr(match, 'bad'):
471 442 badfn = match.bad
472 443
473 # walk all files by default
474 files = match.files()
475 if not files:
476 files = ['.']
477 dc = self._map.copy()
478 else:
479 files = util.unique(files)
480 dc = self._filter(files)
444 files = util.unique(match.files())
445 if not files or '.' in files:
446 files = ['']
447 dc = self._map
481 448
482 449 def imatch(file_):
483 450 if file_ not in dc and self._ignore(file_):
484 451 return False
485 452 return match(file_)
486 453
487 454 # TODO: don't walk unknown directories if unknown and ignored are False
488 455 ignore = self._ignore
489 456 dirignore = self._dirignore
490 457 if ignored:
491 458 imatch = match
492 459 ignore = util.never
493 460 dirignore = util.never
494 461
495 462 # self._root may end with a path separator when self._root == '/'
496 463 common_prefix_len = len(self._root)
497 464 if not util.endswithsep(self._root):
498 465 common_prefix_len += 1
499 466
500 467 normpath = util.normpath
501 468 listdir = osutil.listdir
502 469 lstat = os.lstat
503 470 bisect_left = bisect.bisect_left
504 471 isdir = os.path.isdir
505 472 pconvert = util.pconvert
506 473 join = os.path.join
507 474 s_isdir = stat.S_ISDIR
508 475 supported = self._supported
509 476 _join = self._join
510 477 known = {'.hg': 1}
511 478
512 479 # recursion free walker, faster than os.walk.
513 480 def findfiles(s):
514 481 work = [s]
515 482 wadd = work.append
516 483 found = []
517 484 add = found.append
518 485 if hasattr(match, 'dir'):
519 486 match.dir(normpath(s[common_prefix_len:]))
520 487 while work:
521 488 top = work.pop()
522 489 entries = listdir(top, stat=True)
523 490 # nd is the top of the repository dir tree
524 491 nd = normpath(top[common_prefix_len:])
525 492 if nd == '.':
526 493 nd = ''
527 494 else:
528 495 # do not recurse into a repo contained in this
529 496 # one. use bisect to find .hg directory so speed
530 497 # is good on big directory.
531 498 names = [e[0] for e in entries]
532 499 hg = bisect_left(names, '.hg')
533 500 if hg < len(names) and names[hg] == '.hg':
534 501 if isdir(join(top, '.hg')):
535 502 continue
536 503 for f, kind, st in entries:
537 504 np = pconvert(join(nd, f))
538 505 nn = self.normalize(np)
539 506 if np in known:
540 507 continue
541 508 known[nn] = 1
542 509 p = join(top, f)
543 510 # don't trip over symlinks
544 511 if kind == stat.S_IFDIR:
545 512 if not ignore(np):
546 513 wadd(p)
547 514 if hasattr(match, 'dir'):
548 515 match.dir(np)
549 516 if np in dc and match(np):
550 517 add((nn, None))
551 518 elif imatch(np):
552 519 if supported(np, st.st_mode):
553 520 add((nn, st))
554 521 elif np in dc:
555 522 add((nn, None))
556 523 return util.sort(found)
557 524
558 525 # step one, find all files that match our criteria
559 526 for ff in util.sort(files):
560 527 nf = normpath(ff)
561 528 nn = self.normalize(nf)
562 529 f = _join(ff)
563 530 try:
564 531 st = lstat(f)
565 532 except OSError, inst:
566 533 found = False
567 534 for fn in dc:
568 535 if nf == fn or (fn.startswith(nf) and fn[len(nf)] == '/'):
569 536 found = True
570 537 break
571 538 if not found:
572 539 if inst.errno != errno.ENOENT:
573 540 fwarn(ff, inst.strerror)
574 541 elif badfn(ff, inst.strerror) and imatch(nf):
575 542 yield ff, None
576 543 continue
577 544 if s_isdir(st.st_mode):
578 545 if not dirignore(nf):
579 546 for e in findfiles(f):
580 547 yield e
581 548 else:
582 549 if nn in known:
583 550 continue
584 551 known[nn] = 1
585 if match(nf):
586 if supported(ff, st.st_mode, verbose=True):
587 yield nn, st
588 elif ff in dc:
589 yield nf, None
552 if supported(ff, st.st_mode, verbose=True):
553 yield self.normalize(nf), st
554 elif ff in dc:
555 yield nf, None
590 556
591 557 # step two run through anything left in the dc hash and yield
592 558 # if we haven't already seen it
593 559 for f in util.sort(dc):
594 if f in known:
560 if f in known or not match(f):
595 561 continue
596 562 known[f] = 1
597 if imatch(f):
598 try:
599 st = lstat(_join(f))
600 if supported(f, st.st_mode):
601 yield f, st
602 continue
603 except OSError, inst:
604 if inst.errno not in (errno.ENOENT, errno.ENOTDIR):
605 raise
606 yield f, None
563 try:
564 st = lstat(_join(f))
565 if supported(f, st.st_mode):
566 yield f, st
567 continue
568 except OSError, inst:
569 if inst.errno not in (errno.ENOENT, errno.ENOTDIR):
570 raise
571 yield f, None
607 572
608 573 def status(self, match, ignored, clean, unknown):
609 574 listignored, listclean, listunknown = ignored, clean, unknown
610 575 lookup, modified, added, unknown, ignored = [], [], [], [], []
611 576 removed, deleted, clean = [], [], []
612 577
613 578 _join = self._join
614 579 lstat = os.lstat
615 580 cmap = self._copymap
616 581 dmap = self._map
617 582 ladd = lookup.append
618 583 madd = modified.append
619 584 aadd = added.append
620 585 uadd = unknown.append
621 586 iadd = ignored.append
622 587 radd = removed.append
623 588 dadd = deleted.append
624 589 cadd = clean.append
625 590
626 591 for fn, st in self.walk(match, listunknown, listignored):
627 592 if fn not in dmap:
628 593 if (listignored or match.exact(fn)) and self._dirignore(fn):
629 594 if listignored:
630 595 iadd(fn)
631 596 elif listunknown:
632 597 uadd(fn)
633 598 continue
634 599
635 600 state, mode, size, time, foo = dmap[fn]
636 601
637 602 if not st and state in "nma":
638 603 dadd(fn)
639 604 elif state == 'n':
640 605 if (size >= 0 and
641 606 (size != st.st_size
642 607 or ((mode ^ st.st_mode) & 0100 and self._checkexec))
643 608 or size == -2
644 609 or fn in self._copymap):
645 610 madd(fn)
646 611 elif time != int(st.st_mtime):
647 612 ladd(fn)
648 613 elif listclean:
649 614 cadd(fn)
650 615 elif state == 'm':
651 616 madd(fn)
652 617 elif state == 'a':
653 618 aadd(fn)
654 619 elif state == 'r':
655 620 radd(fn)
656 621
657 622 return (lookup, modified, added, removed, deleted, unknown, ignored,
658 623 clean)
General Comments 0
You need to be logged in to leave comments. Login now