Show More
@@ -1,473 +1,476 b'' | |||
|
1 | 1 | """ |
|
2 | 2 | dirstate.py - working directory tracking for mercurial |
|
3 | 3 | |
|
4 | 4 | Copyright 2005 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 | import struct, os |
|
11 | 11 | from node import * |
|
12 | 12 | from i18n import gettext as _ |
|
13 | 13 | from demandload import * |
|
14 | 14 | demandload(globals(), "time bisect stat util re errno") |
|
15 | 15 | |
|
16 | 16 | class dirstate(object): |
|
17 | format = ">cllll" | |
|
18 | ||
|
17 | 19 | def __init__(self, opener, ui, root): |
|
18 | 20 | self.opener = opener |
|
19 | 21 | self.root = root |
|
20 | 22 | self.dirty = 0 |
|
21 | 23 | self.ui = ui |
|
22 | 24 | self.map = None |
|
23 | 25 | self.pl = None |
|
24 | 26 | self.copies = {} |
|
25 | 27 | self.ignorefunc = None |
|
26 | 28 | self.blockignore = False |
|
27 | 29 | |
|
28 | 30 | def wjoin(self, f): |
|
29 | 31 | return os.path.join(self.root, f) |
|
30 | 32 | |
|
31 | 33 | def getcwd(self): |
|
32 | 34 | cwd = os.getcwd() |
|
33 | 35 | if cwd == self.root: return '' |
|
34 | 36 | return cwd[len(self.root) + 1:] |
|
35 | 37 | |
|
36 | 38 | def hgignore(self): |
|
37 | 39 | '''return the contents of .hgignore files as a list of patterns. |
|
38 | 40 | |
|
39 | 41 | the files parsed for patterns include: |
|
40 | 42 | .hgignore in the repository root |
|
41 | 43 | any additional files specified in the [ui] section of ~/.hgrc |
|
42 | 44 | |
|
43 | 45 | trailing white space is dropped. |
|
44 | 46 | the escape character is backslash. |
|
45 | 47 | comments start with #. |
|
46 | 48 | empty lines are skipped. |
|
47 | 49 | |
|
48 | 50 | lines can be of the following formats: |
|
49 | 51 | |
|
50 | 52 | syntax: regexp # defaults following lines to non-rooted regexps |
|
51 | 53 | syntax: glob # defaults following lines to non-rooted globs |
|
52 | 54 | re:pattern # non-rooted regular expression |
|
53 | 55 | glob:pattern # non-rooted glob |
|
54 | 56 | pattern # pattern of the current default type''' |
|
55 | 57 | syntaxes = {'re': 'relre:', 'regexp': 'relre:', 'glob': 'relglob:'} |
|
56 | 58 | def parselines(fp): |
|
57 | 59 | for line in fp: |
|
58 | 60 | escape = False |
|
59 | 61 | for i in xrange(len(line)): |
|
60 | 62 | if escape: escape = False |
|
61 | 63 | elif line[i] == '\\': escape = True |
|
62 | 64 | elif line[i] == '#': break |
|
63 | 65 | line = line[:i].rstrip() |
|
64 | 66 | if line: yield line |
|
65 | 67 | repoignore = self.wjoin('.hgignore') |
|
66 | 68 | files = [repoignore] |
|
67 | 69 | files.extend(self.ui.hgignorefiles()) |
|
68 | 70 | pats = {} |
|
69 | 71 | for f in files: |
|
70 | 72 | try: |
|
71 | 73 | pats[f] = [] |
|
72 | 74 | fp = open(f) |
|
73 | 75 | syntax = 'relre:' |
|
74 | 76 | for line in parselines(fp): |
|
75 | 77 | if line.startswith('syntax:'): |
|
76 | 78 | s = line[7:].strip() |
|
77 | 79 | try: |
|
78 | 80 | syntax = syntaxes[s] |
|
79 | 81 | except KeyError: |
|
80 | 82 | self.ui.warn(_("%s: ignoring invalid " |
|
81 | 83 | "syntax '%s'\n") % (f, s)) |
|
82 | 84 | continue |
|
83 | 85 | pat = syntax + line |
|
84 | 86 | for s in syntaxes.values(): |
|
85 | 87 | if line.startswith(s): |
|
86 | 88 | pat = line |
|
87 | 89 | break |
|
88 | 90 | pats[f].append(pat) |
|
89 | 91 | except IOError, inst: |
|
90 | 92 | if f != repoignore: |
|
91 | 93 | self.ui.warn(_("skipping unreadable ignore file" |
|
92 | 94 | " '%s': %s\n") % (f, inst.strerror)) |
|
93 | 95 | return pats |
|
94 | 96 | |
|
95 | 97 | def ignore(self, fn): |
|
96 | 98 | '''default match function used by dirstate and |
|
97 | 99 | localrepository. this honours the repository .hgignore file |
|
98 | 100 | and any other files specified in the [ui] section of .hgrc.''' |
|
99 | 101 | if self.blockignore: |
|
100 | 102 | return False |
|
101 | 103 | if not self.ignorefunc: |
|
102 | 104 | ignore = self.hgignore() |
|
103 | 105 | allpats = [] |
|
104 | 106 | [allpats.extend(patlist) for patlist in ignore.values()] |
|
105 | 107 | if allpats: |
|
106 | 108 | try: |
|
107 | 109 | files, self.ignorefunc, anypats = ( |
|
108 | 110 | util.matcher(self.root, inc=allpats, src='.hgignore')) |
|
109 | 111 | except util.Abort: |
|
110 | 112 | # Re-raise an exception where the src is the right file |
|
111 | 113 | for f, patlist in ignore.items(): |
|
112 | 114 | files, self.ignorefunc, anypats = ( |
|
113 | 115 | util.matcher(self.root, inc=patlist, src=f)) |
|
114 | 116 | else: |
|
115 | 117 | self.ignorefunc = util.never |
|
116 | 118 | return self.ignorefunc(fn) |
|
117 | 119 | |
|
118 | 120 | def __del__(self): |
|
119 | 121 | if self.dirty: |
|
120 | 122 | self.write() |
|
121 | 123 | |
|
122 | 124 | def __getitem__(self, key): |
|
123 | 125 | try: |
|
124 | 126 | return self.map[key] |
|
125 | 127 | except TypeError: |
|
126 | 128 | self.lazyread() |
|
127 | 129 | return self[key] |
|
128 | 130 | |
|
129 | 131 | def __contains__(self, key): |
|
130 | 132 | self.lazyread() |
|
131 | 133 | return key in self.map |
|
132 | 134 | |
|
133 | 135 | def parents(self): |
|
134 | 136 | self.lazyread() |
|
135 | 137 | return self.pl |
|
136 | 138 | |
|
137 | 139 | def markdirty(self): |
|
138 | 140 | if not self.dirty: |
|
139 | 141 | self.dirty = 1 |
|
140 | 142 | |
|
141 | 143 | def setparents(self, p1, p2=nullid): |
|
142 | 144 | self.lazyread() |
|
143 | 145 | self.markdirty() |
|
144 | 146 | self.pl = p1, p2 |
|
145 | 147 | |
|
146 | 148 | def state(self, key): |
|
147 | 149 | try: |
|
148 | 150 | return self[key][0] |
|
149 | 151 | except KeyError: |
|
150 | 152 | return "?" |
|
151 | 153 | |
|
152 | 154 | def lazyread(self): |
|
153 | 155 | if self.map is None: |
|
154 | 156 | self.read() |
|
155 | 157 | |
|
156 | 158 | def read(self): |
|
157 | 159 | self.map = {} |
|
158 | 160 | self.pl = [nullid, nullid] |
|
159 | 161 | try: |
|
160 | 162 | st = self.opener("dirstate").read() |
|
161 | 163 | if not st: return |
|
162 | 164 | except: return |
|
163 | 165 | |
|
164 | 166 | self.pl = [st[:20], st[20: 40]] |
|
165 | 167 | |
|
166 | 168 | pos = 40 |
|
169 | e_size = struct.calcsize(self.format) | |
|
167 | 170 | while pos < len(st): |
|
168 |
e = struct.unpack( |
|
|
171 | e = struct.unpack(self.format, st[pos:pos+e_size]) | |
|
169 | 172 | l = e[4] |
|
170 |
pos += |
|
|
173 | pos += e_size | |
|
171 | 174 | f = st[pos:pos + l] |
|
172 | 175 | if '\0' in f: |
|
173 | 176 | f, c = f.split('\0') |
|
174 | 177 | self.copies[f] = c |
|
175 | 178 | self.map[f] = e[:4] |
|
176 | 179 | pos += l |
|
177 | 180 | |
|
178 | 181 | def copy(self, source, dest): |
|
179 | 182 | self.lazyread() |
|
180 | 183 | self.markdirty() |
|
181 | 184 | self.copies[dest] = source |
|
182 | 185 | |
|
183 | 186 | def copied(self, file): |
|
184 | 187 | return self.copies.get(file, None) |
|
185 | 188 | |
|
186 | 189 | def update(self, files, state, **kw): |
|
187 | 190 | ''' current states: |
|
188 | 191 | n normal |
|
189 | 192 | m needs merging |
|
190 | 193 | r marked for removal |
|
191 | 194 | a marked for addition''' |
|
192 | 195 | |
|
193 | 196 | if not files: return |
|
194 | 197 | self.lazyread() |
|
195 | 198 | self.markdirty() |
|
196 | 199 | for f in files: |
|
197 | 200 | if state == "r": |
|
198 | 201 | self.map[f] = ('r', 0, 0, 0) |
|
199 | 202 | else: |
|
200 | 203 | s = os.lstat(self.wjoin(f)) |
|
201 | 204 | st_size = kw.get('st_size', s.st_size) |
|
202 | 205 | st_mtime = kw.get('st_mtime', s.st_mtime) |
|
203 | 206 | self.map[f] = (state, s.st_mode, st_size, st_mtime) |
|
204 | 207 | if self.copies.has_key(f): |
|
205 | 208 | del self.copies[f] |
|
206 | 209 | |
|
207 | 210 | def forget(self, files): |
|
208 | 211 | if not files: return |
|
209 | 212 | self.lazyread() |
|
210 | 213 | self.markdirty() |
|
211 | 214 | for f in files: |
|
212 | 215 | try: |
|
213 | 216 | del self.map[f] |
|
214 | 217 | except KeyError: |
|
215 | 218 | self.ui.warn(_("not in dirstate: %s!\n") % f) |
|
216 | 219 | pass |
|
217 | 220 | |
|
218 | 221 | def clear(self): |
|
219 | 222 | self.map = {} |
|
220 | 223 | self.copies = {} |
|
221 | 224 | self.markdirty() |
|
222 | 225 | |
|
223 | 226 | def rebuild(self, parent, files): |
|
224 | 227 | self.clear() |
|
225 | 228 | umask = os.umask(0) |
|
226 | 229 | os.umask(umask) |
|
227 | 230 | for f, mode in files: |
|
228 | 231 | if mode: |
|
229 | 232 | self.map[f] = ('n', ~umask, -1, 0) |
|
230 | 233 | else: |
|
231 | 234 | self.map[f] = ('n', ~umask & 0666, -1, 0) |
|
232 | 235 | self.pl = (parent, nullid) |
|
233 | 236 | self.markdirty() |
|
234 | 237 | |
|
235 | 238 | def write(self): |
|
236 | 239 | if not self.dirty: |
|
237 | 240 | return |
|
238 | 241 | st = self.opener("dirstate", "w", atomic=True) |
|
239 | 242 | st.write("".join(self.pl)) |
|
240 | 243 | for f, e in self.map.items(): |
|
241 | 244 | c = self.copied(f) |
|
242 | 245 | if c: |
|
243 | 246 | f = f + "\0" + c |
|
244 |
e = struct.pack( |
|
|
247 | e = struct.pack(self.format, e[0], e[1], e[2], e[3], len(f)) | |
|
245 | 248 | st.write(e + f) |
|
246 | 249 | self.dirty = 0 |
|
247 | 250 | |
|
248 | 251 | def filterfiles(self, files): |
|
249 | 252 | ret = {} |
|
250 | 253 | unknown = [] |
|
251 | 254 | |
|
252 | 255 | for x in files: |
|
253 | 256 | if x == '.': |
|
254 | 257 | return self.map.copy() |
|
255 | 258 | if x not in self.map: |
|
256 | 259 | unknown.append(x) |
|
257 | 260 | else: |
|
258 | 261 | ret[x] = self.map[x] |
|
259 | 262 | |
|
260 | 263 | if not unknown: |
|
261 | 264 | return ret |
|
262 | 265 | |
|
263 | 266 | b = self.map.keys() |
|
264 | 267 | b.sort() |
|
265 | 268 | blen = len(b) |
|
266 | 269 | |
|
267 | 270 | for x in unknown: |
|
268 | 271 | bs = bisect.bisect(b, x) |
|
269 | 272 | if bs != 0 and b[bs-1] == x: |
|
270 | 273 | ret[x] = self.map[x] |
|
271 | 274 | continue |
|
272 | 275 | while bs < blen: |
|
273 | 276 | s = b[bs] |
|
274 | 277 | if len(s) > len(x) and s.startswith(x) and s[len(x)] == '/': |
|
275 | 278 | ret[s] = self.map[s] |
|
276 | 279 | else: |
|
277 | 280 | break |
|
278 | 281 | bs += 1 |
|
279 | 282 | return ret |
|
280 | 283 | |
|
281 | 284 | def supported_type(self, f, st, verbose=False): |
|
282 | 285 | if stat.S_ISREG(st.st_mode): |
|
283 | 286 | return True |
|
284 | 287 | if verbose: |
|
285 | 288 | kind = 'unknown' |
|
286 | 289 | if stat.S_ISCHR(st.st_mode): kind = _('character device') |
|
287 | 290 | elif stat.S_ISBLK(st.st_mode): kind = _('block device') |
|
288 | 291 | elif stat.S_ISFIFO(st.st_mode): kind = _('fifo') |
|
289 | 292 | elif stat.S_ISLNK(st.st_mode): kind = _('symbolic link') |
|
290 | 293 | elif stat.S_ISSOCK(st.st_mode): kind = _('socket') |
|
291 | 294 | elif stat.S_ISDIR(st.st_mode): kind = _('directory') |
|
292 | 295 | self.ui.warn(_('%s: unsupported file type (type is %s)\n') % ( |
|
293 | 296 | util.pathto(self.getcwd(), f), |
|
294 | 297 | kind)) |
|
295 | 298 | return False |
|
296 | 299 | |
|
297 | 300 | def statwalk(self, files=None, match=util.always, dc=None, ignored=False, |
|
298 | 301 | badmatch=None): |
|
299 | 302 | self.lazyread() |
|
300 | 303 | |
|
301 | 304 | # walk all files by default |
|
302 | 305 | if not files: |
|
303 | 306 | files = [self.root] |
|
304 | 307 | if not dc: |
|
305 | 308 | dc = self.map.copy() |
|
306 | 309 | elif not dc: |
|
307 | 310 | dc = self.filterfiles(files) |
|
308 | 311 | |
|
309 | 312 | def statmatch(file_, stat): |
|
310 | 313 | file_ = util.pconvert(file_) |
|
311 | 314 | if not ignored and file_ not in dc and self.ignore(file_): |
|
312 | 315 | return False |
|
313 | 316 | return match(file_) |
|
314 | 317 | |
|
315 | 318 | return self.walkhelper(files=files, statmatch=statmatch, dc=dc, |
|
316 | 319 | badmatch=badmatch) |
|
317 | 320 | |
|
318 | 321 | def walk(self, files=None, match=util.always, dc=None, badmatch=None): |
|
319 | 322 | # filter out the stat |
|
320 | 323 | for src, f, st in self.statwalk(files, match, dc, badmatch=badmatch): |
|
321 | 324 | yield src, f |
|
322 | 325 | |
|
323 | 326 | # walk recursively through the directory tree, finding all files |
|
324 | 327 | # matched by the statmatch function |
|
325 | 328 | # |
|
326 | 329 | # results are yielded in a tuple (src, filename, st), where src |
|
327 | 330 | # is one of: |
|
328 | 331 | # 'f' the file was found in the directory tree |
|
329 | 332 | # 'm' the file was only in the dirstate and not in the tree |
|
330 | 333 | # and st is the stat result if the file was found in the directory. |
|
331 | 334 | # |
|
332 | 335 | # dc is an optional arg for the current dirstate. dc is not modified |
|
333 | 336 | # directly by this function, but might be modified by your statmatch call. |
|
334 | 337 | # |
|
335 | 338 | def walkhelper(self, files, statmatch, dc, badmatch=None): |
|
336 | 339 | # recursion free walker, faster than os.walk. |
|
337 | 340 | def findfiles(s): |
|
338 | 341 | work = [s] |
|
339 | 342 | while work: |
|
340 | 343 | top = work.pop() |
|
341 | 344 | names = os.listdir(top) |
|
342 | 345 | names.sort() |
|
343 | 346 | # nd is the top of the repository dir tree |
|
344 | 347 | nd = util.normpath(top[len(self.root) + 1:]) |
|
345 | 348 | if nd == '.': |
|
346 | 349 | nd = '' |
|
347 | 350 | else: |
|
348 | 351 | # do not recurse into a repo contained in this |
|
349 | 352 | # one. use bisect to find .hg directory so speed |
|
350 | 353 | # is good on big directory. |
|
351 | 354 | hg = bisect.bisect_left(names, '.hg') |
|
352 | 355 | if hg < len(names) and names[hg] == '.hg': |
|
353 | 356 | if os.path.isdir(os.path.join(top, '.hg')): |
|
354 | 357 | continue |
|
355 | 358 | for f in names: |
|
356 | 359 | np = util.pconvert(os.path.join(nd, f)) |
|
357 | 360 | if seen(np): |
|
358 | 361 | continue |
|
359 | 362 | p = os.path.join(top, f) |
|
360 | 363 | # don't trip over symlinks |
|
361 | 364 | st = os.lstat(p) |
|
362 | 365 | if stat.S_ISDIR(st.st_mode): |
|
363 | 366 | ds = os.path.join(nd, f +'/') |
|
364 | 367 | if statmatch(ds, st): |
|
365 | 368 | work.append(p) |
|
366 | 369 | if statmatch(np, st) and np in dc: |
|
367 | 370 | yield 'm', np, st |
|
368 | 371 | elif statmatch(np, st): |
|
369 | 372 | if self.supported_type(np, st): |
|
370 | 373 | yield 'f', np, st |
|
371 | 374 | elif np in dc: |
|
372 | 375 | yield 'm', np, st |
|
373 | 376 | |
|
374 | 377 | known = {'.hg': 1} |
|
375 | 378 | def seen(fn): |
|
376 | 379 | if fn in known: return True |
|
377 | 380 | known[fn] = 1 |
|
378 | 381 | |
|
379 | 382 | # step one, find all files that match our criteria |
|
380 | 383 | files.sort() |
|
381 | 384 | for ff in util.unique(files): |
|
382 | 385 | f = self.wjoin(ff) |
|
383 | 386 | try: |
|
384 | 387 | st = os.lstat(f) |
|
385 | 388 | except OSError, inst: |
|
386 | 389 | nf = util.normpath(ff) |
|
387 | 390 | found = False |
|
388 | 391 | for fn in dc: |
|
389 | 392 | if nf == fn or (fn.startswith(nf) and fn[len(nf)] == '/'): |
|
390 | 393 | found = True |
|
391 | 394 | break |
|
392 | 395 | if not found: |
|
393 | 396 | if inst.errno != errno.ENOENT or not badmatch: |
|
394 | 397 | self.ui.warn('%s: %s\n' % ( |
|
395 | 398 | util.pathto(self.getcwd(), ff), |
|
396 | 399 | inst.strerror)) |
|
397 | 400 | elif badmatch and badmatch(ff) and statmatch(ff, None): |
|
398 | 401 | yield 'b', ff, None |
|
399 | 402 | continue |
|
400 | 403 | if stat.S_ISDIR(st.st_mode): |
|
401 | 404 | cmp1 = (lambda x, y: cmp(x[1], y[1])) |
|
402 | 405 | sorted_ = [ x for x in findfiles(f) ] |
|
403 | 406 | sorted_.sort(cmp1) |
|
404 | 407 | for e in sorted_: |
|
405 | 408 | yield e |
|
406 | 409 | else: |
|
407 | 410 | ff = util.normpath(ff) |
|
408 | 411 | if seen(ff): |
|
409 | 412 | continue |
|
410 | 413 | self.blockignore = True |
|
411 | 414 | if statmatch(ff, st): |
|
412 | 415 | if self.supported_type(ff, st, verbose=True): |
|
413 | 416 | yield 'f', ff, st |
|
414 | 417 | elif ff in dc: |
|
415 | 418 | yield 'm', ff, st |
|
416 | 419 | self.blockignore = False |
|
417 | 420 | |
|
418 | 421 | # step two run through anything left in the dc hash and yield |
|
419 | 422 | # if we haven't already seen it |
|
420 | 423 | ks = dc.keys() |
|
421 | 424 | ks.sort() |
|
422 | 425 | for k in ks: |
|
423 | 426 | if not seen(k) and (statmatch(k, None)): |
|
424 | 427 | yield 'm', k, None |
|
425 | 428 | |
|
426 | 429 | def changes(self, files=None, match=util.always, show_ignored=None): |
|
427 | 430 | lookup, modified, added, unknown, ignored = [], [], [], [], [] |
|
428 | 431 | removed, deleted = [], [] |
|
429 | 432 | |
|
430 | 433 | for src, fn, st in self.statwalk(files, match, ignored=show_ignored): |
|
431 | 434 | try: |
|
432 | 435 | type_, mode, size, time = self[fn] |
|
433 | 436 | except KeyError: |
|
434 | 437 | if show_ignored and self.ignore(fn): |
|
435 | 438 | ignored.append(fn) |
|
436 | 439 | else: |
|
437 | 440 | unknown.append(fn) |
|
438 | 441 | continue |
|
439 | 442 | if src == 'm': |
|
440 | 443 | nonexistent = True |
|
441 | 444 | if not st: |
|
442 | 445 | try: |
|
443 | 446 | f = self.wjoin(fn) |
|
444 | 447 | st = os.lstat(f) |
|
445 | 448 | except OSError, inst: |
|
446 | 449 | if inst.errno != errno.ENOENT: |
|
447 | 450 | raise |
|
448 | 451 | st = None |
|
449 | 452 | # We need to re-check that it is a valid file |
|
450 | 453 | if st and self.supported_type(fn, st): |
|
451 | 454 | nonexistent = False |
|
452 | 455 | # XXX: what to do with file no longer present in the fs |
|
453 | 456 | # who are not removed in the dirstate ? |
|
454 | 457 | if nonexistent and type_ in "nm": |
|
455 | 458 | deleted.append(fn) |
|
456 | 459 | continue |
|
457 | 460 | # check the common case first |
|
458 | 461 | if type_ == 'n': |
|
459 | 462 | if not st: |
|
460 | 463 | st = os.stat(fn) |
|
461 | 464 | if size >= 0 and (size != st.st_size |
|
462 | 465 | or (mode ^ st.st_mode) & 0100): |
|
463 | 466 | modified.append(fn) |
|
464 | 467 | elif time != st.st_mtime: |
|
465 | 468 | lookup.append(fn) |
|
466 | 469 | elif type_ == 'm': |
|
467 | 470 | modified.append(fn) |
|
468 | 471 | elif type_ == 'a': |
|
469 | 472 | added.append(fn) |
|
470 | 473 | elif type_ == 'r': |
|
471 | 474 | removed.append(fn) |
|
472 | 475 | |
|
473 | 476 | return (lookup, modified, added, removed, deleted, unknown, ignored) |
General Comments 0
You need to be logged in to leave comments.
Login now