|
1 # Mercurial built-in replacement for cvsps. |
|
2 # |
|
3 # Copyright 2008, Frank Kingswood <frank@kingswood-consulting.co.uk> |
|
4 # |
|
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. |
|
7 |
|
8 import os |
|
9 import re |
|
10 import cPickle as pickle |
|
11 from mercurial import util |
|
12 from mercurial.i18n import _ |
|
13 from mercurial import hook |
|
14 |
|
15 class logentry(object): |
|
16 '''Class logentry has the following attributes: |
|
17 .author - author name as CVS knows it |
|
18 .branch - name of branch this revision is on |
|
19 .branches - revision tuple of branches starting at this revision |
|
20 .comment - commit message |
|
21 .date - the commit date as a (time, tz) tuple |
|
22 .dead - true if file revision is dead |
|
23 .file - Name of file |
|
24 .lines - a tuple (+lines, -lines) or None |
|
25 .parent - Previous revision of this entry |
|
26 .rcs - name of file as returned from CVS |
|
27 .revision - revision number as tuple |
|
28 .tags - list of tags on the file |
|
29 .synthetic - is this a synthetic "file ... added on ..." revision? |
|
30 .mergepoint- the branch that has been merged from |
|
31 (if present in rlog output) |
|
32 .branchpoints- the branches that start at the current entry |
|
33 ''' |
|
34 def __init__(self, **entries): |
|
35 self.synthetic = False |
|
36 self.__dict__.update(entries) |
|
37 |
|
38 def __repr__(self): |
|
39 return "<%s at 0x%x: %s %s>" % (self.__class__.__name__, |
|
40 id(self), |
|
41 self.file, |
|
42 ".".join(map(str, self.revision))) |
|
43 |
|
44 class logerror(Exception): |
|
45 pass |
|
46 |
|
47 def getrepopath(cvspath): |
|
48 """Return the repository path from a CVS path. |
|
49 |
|
50 >>> getrepopath('/foo/bar') |
|
51 '/foo/bar' |
|
52 >>> getrepopath('c:/foo/bar') |
|
53 'c:/foo/bar' |
|
54 >>> getrepopath(':pserver:10/foo/bar') |
|
55 '/foo/bar' |
|
56 >>> getrepopath(':pserver:10c:/foo/bar') |
|
57 '/foo/bar' |
|
58 >>> getrepopath(':pserver:/foo/bar') |
|
59 '/foo/bar' |
|
60 >>> getrepopath(':pserver:c:/foo/bar') |
|
61 'c:/foo/bar' |
|
62 >>> getrepopath(':pserver:truc@foo.bar:/foo/bar') |
|
63 '/foo/bar' |
|
64 >>> getrepopath(':pserver:truc@foo.bar:c:/foo/bar') |
|
65 'c:/foo/bar' |
|
66 """ |
|
67 # According to CVS manual, CVS paths are expressed like: |
|
68 # [:method:][[user][:password]@]hostname[:[port]]/path/to/repository |
|
69 # |
|
70 # Unfortunately, Windows absolute paths start with a drive letter |
|
71 # like 'c:' making it harder to parse. Here we assume that drive |
|
72 # letters are only one character long and any CVS component before |
|
73 # the repository path is at least 2 characters long, and use this |
|
74 # to disambiguate. |
|
75 parts = cvspath.split(':') |
|
76 if len(parts) == 1: |
|
77 return parts[0] |
|
78 # Here there is an ambiguous case if we have a port number |
|
79 # immediately followed by a Windows driver letter. We assume this |
|
80 # never happens and decide it must be CVS path component, |
|
81 # therefore ignoring it. |
|
82 if len(parts[-2]) > 1: |
|
83 return parts[-1].lstrip('0123456789') |
|
84 return parts[-2] + ':' + parts[-1] |
|
85 |
|
86 def createlog(ui, directory=None, root="", rlog=True, cache=None): |
|
87 '''Collect the CVS rlog''' |
|
88 |
|
89 # Because we store many duplicate commit log messages, reusing strings |
|
90 # saves a lot of memory and pickle storage space. |
|
91 _scache = {} |
|
92 def scache(s): |
|
93 "return a shared version of a string" |
|
94 return _scache.setdefault(s, s) |
|
95 |
|
96 ui.status(_('collecting CVS rlog\n')) |
|
97 |
|
98 log = [] # list of logentry objects containing the CVS state |
|
99 |
|
100 # patterns to match in CVS (r)log output, by state of use |
|
101 re_00 = re.compile('RCS file: (.+)$') |
|
102 re_01 = re.compile('cvs \\[r?log aborted\\]: (.+)$') |
|
103 re_02 = re.compile('cvs (r?log|server): (.+)\n$') |
|
104 re_03 = re.compile("(Cannot access.+CVSROOT)|" |
|
105 "(can't create temporary directory.+)$") |
|
106 re_10 = re.compile('Working file: (.+)$') |
|
107 re_20 = re.compile('symbolic names:') |
|
108 re_30 = re.compile('\t(.+): ([\\d.]+)$') |
|
109 re_31 = re.compile('----------------------------$') |
|
110 re_32 = re.compile('=======================================' |
|
111 '======================================$') |
|
112 re_50 = re.compile('revision ([\\d.]+)(\s+locked by:\s+.+;)?$') |
|
113 re_60 = re.compile(r'date:\s+(.+);\s+author:\s+(.+);\s+state:\s+(.+?);' |
|
114 r'(\s+lines:\s+(\+\d+)?\s+(-\d+)?;)?' |
|
115 r'(.*mergepoint:\s+([^;]+);)?') |
|
116 re_70 = re.compile('branches: (.+);$') |
|
117 |
|
118 file_added_re = re.compile(r'file [^/]+ was (initially )?added on branch') |
|
119 |
|
120 prefix = '' # leading path to strip of what we get from CVS |
|
121 |
|
122 if directory is None: |
|
123 # Current working directory |
|
124 |
|
125 # Get the real directory in the repository |
|
126 try: |
|
127 prefix = open(os.path.join('CVS','Repository')).read().strip() |
|
128 directory = prefix |
|
129 if prefix == ".": |
|
130 prefix = "" |
|
131 except IOError: |
|
132 raise logerror(_('not a CVS sandbox')) |
|
133 |
|
134 if prefix and not prefix.endswith(os.sep): |
|
135 prefix += os.sep |
|
136 |
|
137 # Use the Root file in the sandbox, if it exists |
|
138 try: |
|
139 root = open(os.path.join('CVS','Root')).read().strip() |
|
140 except IOError: |
|
141 pass |
|
142 |
|
143 if not root: |
|
144 root = os.environ.get('CVSROOT', '') |
|
145 |
|
146 # read log cache if one exists |
|
147 oldlog = [] |
|
148 date = None |
|
149 |
|
150 if cache: |
|
151 cachedir = os.path.expanduser('~/.hg.cvsps') |
|
152 if not os.path.exists(cachedir): |
|
153 os.mkdir(cachedir) |
|
154 |
|
155 # The cvsps cache pickle needs a uniquified name, based on the |
|
156 # repository location. The address may have all sort of nasties |
|
157 # in it, slashes, colons and such. So here we take just the |
|
158 # alphanumerics, concatenated in a way that does not mix up the |
|
159 # various components, so that |
|
160 # :pserver:user@server:/path |
|
161 # and |
|
162 # /pserver/user/server/path |
|
163 # are mapped to different cache file names. |
|
164 cachefile = root.split(":") + [directory, "cache"] |
|
165 cachefile = ['-'.join(re.findall(r'\w+', s)) for s in cachefile if s] |
|
166 cachefile = os.path.join(cachedir, |
|
167 '.'.join([s for s in cachefile if s])) |
|
168 |
|
169 if cache == 'update': |
|
170 try: |
|
171 ui.note(_('reading cvs log cache %s\n') % cachefile) |
|
172 oldlog = pickle.load(open(cachefile)) |
|
173 ui.note(_('cache has %d log entries\n') % len(oldlog)) |
|
174 except Exception, e: |
|
175 ui.note(_('error reading cache: %r\n') % e) |
|
176 |
|
177 if oldlog: |
|
178 date = oldlog[-1].date # last commit date as a (time,tz) tuple |
|
179 date = util.datestr(date, '%Y/%m/%d %H:%M:%S %1%2') |
|
180 |
|
181 # build the CVS commandline |
|
182 cmd = ['cvs', '-q'] |
|
183 if root: |
|
184 cmd.append('-d%s' % root) |
|
185 p = util.normpath(getrepopath(root)) |
|
186 if not p.endswith('/'): |
|
187 p += '/' |
|
188 if prefix: |
|
189 # looks like normpath replaces "" by "." |
|
190 prefix = p + util.normpath(prefix) |
|
191 else: |
|
192 prefix = p |
|
193 cmd.append(['log', 'rlog'][rlog]) |
|
194 if date: |
|
195 # no space between option and date string |
|
196 cmd.append('-d>%s' % date) |
|
197 cmd.append(directory) |
|
198 |
|
199 # state machine begins here |
|
200 tags = {} # dictionary of revisions on current file with their tags |
|
201 branchmap = {} # mapping between branch names and revision numbers |
|
202 state = 0 |
|
203 store = False # set when a new record can be appended |
|
204 |
|
205 cmd = [util.shellquote(arg) for arg in cmd] |
|
206 ui.note(_("running %s\n") % (' '.join(cmd))) |
|
207 ui.debug("prefix=%r directory=%r root=%r\n" % (prefix, directory, root)) |
|
208 |
|
209 pfp = util.popen(' '.join(cmd)) |
|
210 peek = pfp.readline() |
|
211 while True: |
|
212 line = peek |
|
213 if line == '': |
|
214 break |
|
215 peek = pfp.readline() |
|
216 if line.endswith('\n'): |
|
217 line = line[:-1] |
|
218 #ui.debug('state=%d line=%r\n' % (state, line)) |
|
219 |
|
220 if state == 0: |
|
221 # initial state, consume input until we see 'RCS file' |
|
222 match = re_00.match(line) |
|
223 if match: |
|
224 rcs = match.group(1) |
|
225 tags = {} |
|
226 if rlog: |
|
227 filename = util.normpath(rcs[:-2]) |
|
228 if filename.startswith(prefix): |
|
229 filename = filename[len(prefix):] |
|
230 if filename.startswith('/'): |
|
231 filename = filename[1:] |
|
232 if filename.startswith('Attic/'): |
|
233 filename = filename[6:] |
|
234 else: |
|
235 filename = filename.replace('/Attic/', '/') |
|
236 state = 2 |
|
237 continue |
|
238 state = 1 |
|
239 continue |
|
240 match = re_01.match(line) |
|
241 if match: |
|
242 raise logerror(match.group(1)) |
|
243 match = re_02.match(line) |
|
244 if match: |
|
245 raise logerror(match.group(2)) |
|
246 if re_03.match(line): |
|
247 raise logerror(line) |
|
248 |
|
249 elif state == 1: |
|
250 # expect 'Working file' (only when using log instead of rlog) |
|
251 match = re_10.match(line) |
|
252 assert match, _('RCS file must be followed by working file') |
|
253 filename = util.normpath(match.group(1)) |
|
254 state = 2 |
|
255 |
|
256 elif state == 2: |
|
257 # expect 'symbolic names' |
|
258 if re_20.match(line): |
|
259 branchmap = {} |
|
260 state = 3 |
|
261 |
|
262 elif state == 3: |
|
263 # read the symbolic names and store as tags |
|
264 match = re_30.match(line) |
|
265 if match: |
|
266 rev = [int(x) for x in match.group(2).split('.')] |
|
267 |
|
268 # Convert magic branch number to an odd-numbered one |
|
269 revn = len(rev) |
|
270 if revn > 3 and (revn % 2) == 0 and rev[-2] == 0: |
|
271 rev = rev[:-2] + rev[-1:] |
|
272 rev = tuple(rev) |
|
273 |
|
274 if rev not in tags: |
|
275 tags[rev] = [] |
|
276 tags[rev].append(match.group(1)) |
|
277 branchmap[match.group(1)] = match.group(2) |
|
278 |
|
279 elif re_31.match(line): |
|
280 state = 5 |
|
281 elif re_32.match(line): |
|
282 state = 0 |
|
283 |
|
284 elif state == 4: |
|
285 # expecting '------' separator before first revision |
|
286 if re_31.match(line): |
|
287 state = 5 |
|
288 else: |
|
289 assert not re_32.match(line), _('must have at least ' |
|
290 'some revisions') |
|
291 |
|
292 elif state == 5: |
|
293 # expecting revision number and possibly (ignored) lock indication |
|
294 # we create the logentry here from values stored in states 0 to 4, |
|
295 # as this state is re-entered for subsequent revisions of a file. |
|
296 match = re_50.match(line) |
|
297 assert match, _('expected revision number') |
|
298 e = logentry(rcs=scache(rcs), file=scache(filename), |
|
299 revision=tuple([int(x) for x in match.group(1).split('.')]), |
|
300 branches=[], parent=None) |
|
301 state = 6 |
|
302 |
|
303 elif state == 6: |
|
304 # expecting date, author, state, lines changed |
|
305 match = re_60.match(line) |
|
306 assert match, _('revision must be followed by date line') |
|
307 d = match.group(1) |
|
308 if d[2] == '/': |
|
309 # Y2K |
|
310 d = '19' + d |
|
311 |
|
312 if len(d.split()) != 3: |
|
313 # cvs log dates always in GMT |
|
314 d = d + ' UTC' |
|
315 e.date = util.parsedate(d, ['%y/%m/%d %H:%M:%S', |
|
316 '%Y/%m/%d %H:%M:%S', |
|
317 '%Y-%m-%d %H:%M:%S']) |
|
318 e.author = scache(match.group(2)) |
|
319 e.dead = match.group(3).lower() == 'dead' |
|
320 |
|
321 if match.group(5): |
|
322 if match.group(6): |
|
323 e.lines = (int(match.group(5)), int(match.group(6))) |
|
324 else: |
|
325 e.lines = (int(match.group(5)), 0) |
|
326 elif match.group(6): |
|
327 e.lines = (0, int(match.group(6))) |
|
328 else: |
|
329 e.lines = None |
|
330 |
|
331 if match.group(7): # cvsnt mergepoint |
|
332 myrev = match.group(8).split('.') |
|
333 if len(myrev) == 2: # head |
|
334 e.mergepoint = 'HEAD' |
|
335 else: |
|
336 myrev = '.'.join(myrev[:-2] + ['0', myrev[-2]]) |
|
337 branches = [b for b in branchmap if branchmap[b] == myrev] |
|
338 assert len(branches) == 1, 'unknown branch: %s' % e.mergepoint |
|
339 e.mergepoint = branches[0] |
|
340 else: |
|
341 e.mergepoint = None |
|
342 e.comment = [] |
|
343 state = 7 |
|
344 |
|
345 elif state == 7: |
|
346 # read the revision numbers of branches that start at this revision |
|
347 # or store the commit log message otherwise |
|
348 m = re_70.match(line) |
|
349 if m: |
|
350 e.branches = [tuple([int(y) for y in x.strip().split('.')]) |
|
351 for x in m.group(1).split(';')] |
|
352 state = 8 |
|
353 elif re_31.match(line) and re_50.match(peek): |
|
354 state = 5 |
|
355 store = True |
|
356 elif re_32.match(line): |
|
357 state = 0 |
|
358 store = True |
|
359 else: |
|
360 e.comment.append(line) |
|
361 |
|
362 elif state == 8: |
|
363 # store commit log message |
|
364 if re_31.match(line): |
|
365 state = 5 |
|
366 store = True |
|
367 elif re_32.match(line): |
|
368 state = 0 |
|
369 store = True |
|
370 else: |
|
371 e.comment.append(line) |
|
372 |
|
373 # When a file is added on a branch B1, CVS creates a synthetic |
|
374 # dead trunk revision 1.1 so that the branch has a root. |
|
375 # Likewise, if you merge such a file to a later branch B2 (one |
|
376 # that already existed when the file was added on B1), CVS |
|
377 # creates a synthetic dead revision 1.1.x.1 on B2. Don't drop |
|
378 # these revisions now, but mark them synthetic so |
|
379 # createchangeset() can take care of them. |
|
380 if (store and |
|
381 e.dead and |
|
382 e.revision[-1] == 1 and # 1.1 or 1.1.x.1 |
|
383 len(e.comment) == 1 and |
|
384 file_added_re.match(e.comment[0])): |
|
385 ui.debug('found synthetic revision in %s: %r\n' |
|
386 % (e.rcs, e.comment[0])) |
|
387 e.synthetic = True |
|
388 |
|
389 if store: |
|
390 # clean up the results and save in the log. |
|
391 store = False |
|
392 e.tags = sorted([scache(x) for x in tags.get(e.revision, [])]) |
|
393 e.comment = scache('\n'.join(e.comment)) |
|
394 |
|
395 revn = len(e.revision) |
|
396 if revn > 3 and (revn % 2) == 0: |
|
397 e.branch = tags.get(e.revision[:-1], [None])[0] |
|
398 else: |
|
399 e.branch = None |
|
400 |
|
401 # find the branches starting from this revision |
|
402 branchpoints = set() |
|
403 for branch, revision in branchmap.iteritems(): |
|
404 revparts = tuple([int(i) for i in revision.split('.')]) |
|
405 if len(revparts) < 2: # bad tags |
|
406 continue |
|
407 if revparts[-2] == 0 and revparts[-1] % 2 == 0: |
|
408 # normal branch |
|
409 if revparts[:-2] == e.revision: |
|
410 branchpoints.add(branch) |
|
411 elif revparts == (1, 1, 1): # vendor branch |
|
412 if revparts in e.branches: |
|
413 branchpoints.add(branch) |
|
414 e.branchpoints = branchpoints |
|
415 |
|
416 log.append(e) |
|
417 |
|
418 if len(log) % 100 == 0: |
|
419 ui.status(util.ellipsis('%d %s' % (len(log), e.file), 80)+'\n') |
|
420 |
|
421 log.sort(key=lambda x: (x.rcs, x.revision)) |
|
422 |
|
423 # find parent revisions of individual files |
|
424 versions = {} |
|
425 for e in log: |
|
426 branch = e.revision[:-1] |
|
427 p = versions.get((e.rcs, branch), None) |
|
428 if p is None: |
|
429 p = e.revision[:-2] |
|
430 e.parent = p |
|
431 versions[(e.rcs, branch)] = e.revision |
|
432 |
|
433 # update the log cache |
|
434 if cache: |
|
435 if log: |
|
436 # join up the old and new logs |
|
437 log.sort(key=lambda x: x.date) |
|
438 |
|
439 if oldlog and oldlog[-1].date >= log[0].date: |
|
440 raise logerror(_('log cache overlaps with new log entries,' |
|
441 ' re-run without cache.')) |
|
442 |
|
443 log = oldlog + log |
|
444 |
|
445 # write the new cachefile |
|
446 ui.note(_('writing cvs log cache %s\n') % cachefile) |
|
447 pickle.dump(log, open(cachefile, 'w')) |
|
448 else: |
|
449 log = oldlog |
|
450 |
|
451 ui.status(_('%d log entries\n') % len(log)) |
|
452 |
|
453 hook.hook(ui, None, "cvslog", True, log=log) |
|
454 |
|
455 return log |
|
456 |
|
457 |
|
458 class changeset(object): |
|
459 '''Class changeset has the following attributes: |
|
460 .id - integer identifying this changeset (list index) |
|
461 .author - author name as CVS knows it |
|
462 .branch - name of branch this changeset is on, or None |
|
463 .comment - commit message |
|
464 .date - the commit date as a (time,tz) tuple |
|
465 .entries - list of logentry objects in this changeset |
|
466 .parents - list of one or two parent changesets |
|
467 .tags - list of tags on this changeset |
|
468 .synthetic - from synthetic revision "file ... added on branch ..." |
|
469 .mergepoint- the branch that has been merged from |
|
470 (if present in rlog output) |
|
471 .branchpoints- the branches that start at the current entry |
|
472 ''' |
|
473 def __init__(self, **entries): |
|
474 self.synthetic = False |
|
475 self.__dict__.update(entries) |
|
476 |
|
477 def __repr__(self): |
|
478 return "<%s at 0x%x: %s>" % (self.__class__.__name__, |
|
479 id(self), |
|
480 getattr(self, 'id', "(no id)")) |
|
481 |
|
482 def createchangeset(ui, log, fuzz=60, mergefrom=None, mergeto=None): |
|
483 '''Convert log into changesets.''' |
|
484 |
|
485 ui.status(_('creating changesets\n')) |
|
486 |
|
487 # Merge changesets |
|
488 |
|
489 log.sort(key=lambda x: (x.comment, x.author, x.branch, x.date)) |
|
490 |
|
491 changesets = [] |
|
492 files = set() |
|
493 c = None |
|
494 for i, e in enumerate(log): |
|
495 |
|
496 # Check if log entry belongs to the current changeset or not. |
|
497 |
|
498 # Since CVS is file centric, two different file revisions with |
|
499 # different branchpoints should be treated as belonging to two |
|
500 # different changesets (and the ordering is important and not |
|
501 # honoured by cvsps at this point). |
|
502 # |
|
503 # Consider the following case: |
|
504 # foo 1.1 branchpoints: [MYBRANCH] |
|
505 # bar 1.1 branchpoints: [MYBRANCH, MYBRANCH2] |
|
506 # |
|
507 # Here foo is part only of MYBRANCH, but not MYBRANCH2, e.g. a |
|
508 # later version of foo may be in MYBRANCH2, so foo should be the |
|
509 # first changeset and bar the next and MYBRANCH and MYBRANCH2 |
|
510 # should both start off of the bar changeset. No provisions are |
|
511 # made to ensure that this is, in fact, what happens. |
|
512 if not (c and |
|
513 e.comment == c.comment and |
|
514 e.author == c.author and |
|
515 e.branch == c.branch and |
|
516 (not hasattr(e, 'branchpoints') or |
|
517 not hasattr (c, 'branchpoints') or |
|
518 e.branchpoints == c.branchpoints) and |
|
519 ((c.date[0] + c.date[1]) <= |
|
520 (e.date[0] + e.date[1]) <= |
|
521 (c.date[0] + c.date[1]) + fuzz) and |
|
522 e.file not in files): |
|
523 c = changeset(comment=e.comment, author=e.author, |
|
524 branch=e.branch, date=e.date, entries=[], |
|
525 mergepoint=getattr(e, 'mergepoint', None), |
|
526 branchpoints=getattr(e, 'branchpoints', set())) |
|
527 changesets.append(c) |
|
528 files = set() |
|
529 if len(changesets) % 100 == 0: |
|
530 t = '%d %s' % (len(changesets), repr(e.comment)[1:-1]) |
|
531 ui.status(util.ellipsis(t, 80) + '\n') |
|
532 |
|
533 c.entries.append(e) |
|
534 files.add(e.file) |
|
535 c.date = e.date # changeset date is date of latest commit in it |
|
536 |
|
537 # Mark synthetic changesets |
|
538 |
|
539 for c in changesets: |
|
540 # Synthetic revisions always get their own changeset, because |
|
541 # the log message includes the filename. E.g. if you add file3 |
|
542 # and file4 on a branch, you get four log entries and three |
|
543 # changesets: |
|
544 # "File file3 was added on branch ..." (synthetic, 1 entry) |
|
545 # "File file4 was added on branch ..." (synthetic, 1 entry) |
|
546 # "Add file3 and file4 to fix ..." (real, 2 entries) |
|
547 # Hence the check for 1 entry here. |
|
548 c.synthetic = len(c.entries) == 1 and c.entries[0].synthetic |
|
549 |
|
550 # Sort files in each changeset |
|
551 |
|
552 for c in changesets: |
|
553 def pathcompare(l, r): |
|
554 'Mimic cvsps sorting order' |
|
555 l = l.split('/') |
|
556 r = r.split('/') |
|
557 nl = len(l) |
|
558 nr = len(r) |
|
559 n = min(nl, nr) |
|
560 for i in range(n): |
|
561 if i + 1 == nl and nl < nr: |
|
562 return -1 |
|
563 elif i + 1 == nr and nl > nr: |
|
564 return +1 |
|
565 elif l[i] < r[i]: |
|
566 return -1 |
|
567 elif l[i] > r[i]: |
|
568 return +1 |
|
569 return 0 |
|
570 def entitycompare(l, r): |
|
571 return pathcompare(l.file, r.file) |
|
572 |
|
573 c.entries.sort(entitycompare) |
|
574 |
|
575 # Sort changesets by date |
|
576 |
|
577 def cscmp(l, r): |
|
578 d = sum(l.date) - sum(r.date) |
|
579 if d: |
|
580 return d |
|
581 |
|
582 # detect vendor branches and initial commits on a branch |
|
583 le = {} |
|
584 for e in l.entries: |
|
585 le[e.rcs] = e.revision |
|
586 re = {} |
|
587 for e in r.entries: |
|
588 re[e.rcs] = e.revision |
|
589 |
|
590 d = 0 |
|
591 for e in l.entries: |
|
592 if re.get(e.rcs, None) == e.parent: |
|
593 assert not d |
|
594 d = 1 |
|
595 break |
|
596 |
|
597 for e in r.entries: |
|
598 if le.get(e.rcs, None) == e.parent: |
|
599 assert not d |
|
600 d = -1 |
|
601 break |
|
602 |
|
603 return d |
|
604 |
|
605 changesets.sort(cscmp) |
|
606 |
|
607 # Collect tags |
|
608 |
|
609 globaltags = {} |
|
610 for c in changesets: |
|
611 for e in c.entries: |
|
612 for tag in e.tags: |
|
613 # remember which is the latest changeset to have this tag |
|
614 globaltags[tag] = c |
|
615 |
|
616 for c in changesets: |
|
617 tags = set() |
|
618 for e in c.entries: |
|
619 tags.update(e.tags) |
|
620 # remember tags only if this is the latest changeset to have it |
|
621 c.tags = sorted(tag for tag in tags if globaltags[tag] is c) |
|
622 |
|
623 # Find parent changesets, handle {{mergetobranch BRANCHNAME}} |
|
624 # by inserting dummy changesets with two parents, and handle |
|
625 # {{mergefrombranch BRANCHNAME}} by setting two parents. |
|
626 |
|
627 if mergeto is None: |
|
628 mergeto = r'{{mergetobranch ([-\w]+)}}' |
|
629 if mergeto: |
|
630 mergeto = re.compile(mergeto) |
|
631 |
|
632 if mergefrom is None: |
|
633 mergefrom = r'{{mergefrombranch ([-\w]+)}}' |
|
634 if mergefrom: |
|
635 mergefrom = re.compile(mergefrom) |
|
636 |
|
637 versions = {} # changeset index where we saw any particular file version |
|
638 branches = {} # changeset index where we saw a branch |
|
639 n = len(changesets) |
|
640 i = 0 |
|
641 while i < n: |
|
642 c = changesets[i] |
|
643 |
|
644 for f in c.entries: |
|
645 versions[(f.rcs, f.revision)] = i |
|
646 |
|
647 p = None |
|
648 if c.branch in branches: |
|
649 p = branches[c.branch] |
|
650 else: |
|
651 # first changeset on a new branch |
|
652 # the parent is a changeset with the branch in its |
|
653 # branchpoints such that it is the latest possible |
|
654 # commit without any intervening, unrelated commits. |
|
655 |
|
656 for candidate in xrange(i): |
|
657 if c.branch not in changesets[candidate].branchpoints: |
|
658 if p is not None: |
|
659 break |
|
660 continue |
|
661 p = candidate |
|
662 |
|
663 c.parents = [] |
|
664 if p is not None: |
|
665 p = changesets[p] |
|
666 |
|
667 # Ensure no changeset has a synthetic changeset as a parent. |
|
668 while p.synthetic: |
|
669 assert len(p.parents) <= 1, \ |
|
670 _('synthetic changeset cannot have multiple parents') |
|
671 if p.parents: |
|
672 p = p.parents[0] |
|
673 else: |
|
674 p = None |
|
675 break |
|
676 |
|
677 if p is not None: |
|
678 c.parents.append(p) |
|
679 |
|
680 if c.mergepoint: |
|
681 if c.mergepoint == 'HEAD': |
|
682 c.mergepoint = None |
|
683 c.parents.append(changesets[branches[c.mergepoint]]) |
|
684 |
|
685 if mergefrom: |
|
686 m = mergefrom.search(c.comment) |
|
687 if m: |
|
688 m = m.group(1) |
|
689 if m == 'HEAD': |
|
690 m = None |
|
691 try: |
|
692 candidate = changesets[branches[m]] |
|
693 except KeyError: |
|
694 ui.warn(_("warning: CVS commit message references " |
|
695 "non-existent branch %r:\n%s\n") |
|
696 % (m, c.comment)) |
|
697 if m in branches and c.branch != m and not candidate.synthetic: |
|
698 c.parents.append(candidate) |
|
699 |
|
700 if mergeto: |
|
701 m = mergeto.search(c.comment) |
|
702 if m: |
|
703 try: |
|
704 m = m.group(1) |
|
705 if m == 'HEAD': |
|
706 m = None |
|
707 except: |
|
708 m = None # if no group found then merge to HEAD |
|
709 if m in branches and c.branch != m: |
|
710 # insert empty changeset for merge |
|
711 cc = changeset( |
|
712 author=c.author, branch=m, date=c.date, |
|
713 comment='convert-repo: CVS merge from branch %s' |
|
714 % c.branch, |
|
715 entries=[], tags=[], |
|
716 parents=[changesets[branches[m]], c]) |
|
717 changesets.insert(i + 1, cc) |
|
718 branches[m] = i + 1 |
|
719 |
|
720 # adjust our loop counters now we have inserted a new entry |
|
721 n += 1 |
|
722 i += 2 |
|
723 continue |
|
724 |
|
725 branches[c.branch] = i |
|
726 i += 1 |
|
727 |
|
728 # Drop synthetic changesets (safe now that we have ensured no other |
|
729 # changesets can have them as parents). |
|
730 i = 0 |
|
731 while i < len(changesets): |
|
732 if changesets[i].synthetic: |
|
733 del changesets[i] |
|
734 else: |
|
735 i += 1 |
|
736 |
|
737 # Number changesets |
|
738 |
|
739 for i, c in enumerate(changesets): |
|
740 c.id = i + 1 |
|
741 |
|
742 ui.status(_('%d changeset entries\n') % len(changesets)) |
|
743 |
|
744 hook.hook(ui, None, "cvschangesets", True, changesets=changesets) |
|
745 |
|
746 return changesets |
|
747 |
|
748 |
|
749 def debugcvsps(ui, *args, **opts): |
|
750 '''Read CVS rlog for current directory or named path in |
|
751 repository, and convert the log to changesets based on matching |
|
752 commit log entries and dates. |
|
753 ''' |
|
754 if opts["new_cache"]: |
|
755 cache = "write" |
|
756 elif opts["update_cache"]: |
|
757 cache = "update" |
|
758 else: |
|
759 cache = None |
|
760 |
|
761 revisions = opts["revisions"] |
|
762 |
|
763 try: |
|
764 if args: |
|
765 log = [] |
|
766 for d in args: |
|
767 log += createlog(ui, d, root=opts["root"], cache=cache) |
|
768 else: |
|
769 log = createlog(ui, root=opts["root"], cache=cache) |
|
770 except logerror, e: |
|
771 ui.write("%r\n"%e) |
|
772 return |
|
773 |
|
774 changesets = createchangeset(ui, log, opts["fuzz"]) |
|
775 del log |
|
776 |
|
777 # Print changesets (optionally filtered) |
|
778 |
|
779 off = len(revisions) |
|
780 branches = {} # latest version number in each branch |
|
781 ancestors = {} # parent branch |
|
782 for cs in changesets: |
|
783 |
|
784 if opts["ancestors"]: |
|
785 if cs.branch not in branches and cs.parents and cs.parents[0].id: |
|
786 ancestors[cs.branch] = (changesets[cs.parents[0].id - 1].branch, |
|
787 cs.parents[0].id) |
|
788 branches[cs.branch] = cs.id |
|
789 |
|
790 # limit by branches |
|
791 if opts["branches"] and (cs.branch or 'HEAD') not in opts["branches"]: |
|
792 continue |
|
793 |
|
794 if not off: |
|
795 # Note: trailing spaces on several lines here are needed to have |
|
796 # bug-for-bug compatibility with cvsps. |
|
797 ui.write('---------------------\n') |
|
798 ui.write('PatchSet %d \n' % cs.id) |
|
799 ui.write('Date: %s\n' % util.datestr(cs.date, |
|
800 '%Y/%m/%d %H:%M:%S %1%2')) |
|
801 ui.write('Author: %s\n' % cs.author) |
|
802 ui.write('Branch: %s\n' % (cs.branch or 'HEAD')) |
|
803 ui.write('Tag%s: %s \n' % (['', 's'][len(cs.tags) > 1], |
|
804 ','.join(cs.tags) or '(none)')) |
|
805 branchpoints = getattr(cs, 'branchpoints', None) |
|
806 if branchpoints: |
|
807 ui.write('Branchpoints: %s \n' % ', '.join(branchpoints)) |
|
808 if opts["parents"] and cs.parents: |
|
809 if len(cs.parents) > 1: |
|
810 ui.write('Parents: %s\n' % |
|
811 (','.join([str(p.id) for p in cs.parents]))) |
|
812 else: |
|
813 ui.write('Parent: %d\n' % cs.parents[0].id) |
|
814 |
|
815 if opts["ancestors"]: |
|
816 b = cs.branch |
|
817 r = [] |
|
818 while b: |
|
819 b, c = ancestors[b] |
|
820 r.append('%s:%d:%d' % (b or "HEAD", c, branches[b])) |
|
821 if r: |
|
822 ui.write('Ancestors: %s\n' % (','.join(r))) |
|
823 |
|
824 ui.write('Log:\n') |
|
825 ui.write('%s\n\n' % cs.comment) |
|
826 ui.write('Members: \n') |
|
827 for f in cs.entries: |
|
828 fn = f.file |
|
829 if fn.startswith(opts["prefix"]): |
|
830 fn = fn[len(opts["prefix"]):] |
|
831 ui.write('\t%s:%s->%s%s \n' % ( |
|
832 fn, '.'.join([str(x) for x in f.parent]) or 'INITIAL', |
|
833 '.'.join([str(x) for x in f.revision]), |
|
834 ['', '(DEAD)'][f.dead])) |
|
835 ui.write('\n') |
|
836 |
|
837 # have we seen the start tag? |
|
838 if revisions and off: |
|
839 if revisions[0] == str(cs.id) or \ |
|
840 revisions[0] in cs.tags: |
|
841 off = False |
|
842 |
|
843 # see if we reached the end tag |
|
844 if len(revisions) > 1 and not off: |
|
845 if revisions[1] == str(cs.id) or \ |
|
846 revisions[1] in cs.tags: |
|
847 break |