darcs2git.py

   1 #! /usr/bin/python
   2
   3 """
   4
   5     darcs2git -- Darcs to git converter.
   6
   7     Copyright (c) 2007 Han-Wen Nienhuys <hanwen@xs4all.nl>
   8
   9     This program is free software; you can redistribute it and/or modify
  10     it under the terms of the GNU General Public License as published by
  11     the Free Software Foundation; either version 2, or (at your option)
  12     any later version.
  13
  14     This program is distributed in the hope that it will be useful,
  15     but WITHOUT ANY WARRANTY; without even the implied warranty of
  16     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  17     GNU General Public License for more details.
  18
  19     You should have received a copy of the GNU General Public License
  20     along with this program; if not, write to the Free Software
  21     Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  22
  23 """
  24
  25 # TODO:
  26 #
  27 # - time zones
  28 #
  29 # - file modes
  30 #
  31 # - use binary search to find from-patch in case of conflict.
  32 #
  33
  34 import sha
  35 from datetime import datetime
  36 from time import strptime
  37 import urlparse
  38 import distutils.version
  39 import glob
  40 import os
  41 import sys
  42 import time
  43 import xml.dom.minidom
  44 import re
  45 import gzip
  46 import optparse
  47
  48
  49 ################################################################
  50 # globals
  51
  52
  53 log_file = None
  54 options = None
  55 mail_to_name_dict = {}
  56 pending_patches = {}
  57 git_commits = {}
  58 used_tags = {}
  59
  60 ################################################################
  61 # utils
  62
  63 class PullConflict (Exception):
  64     pass
  65 class CommandFailed (Exception):
  66     pass
  67
  68 def progress (s):
  69     sys.stderr.write (s + '\n')
  70
  71 def get_cli_options ():
  72     class MyOP(optparse.OptionParser):
  73         def print_help(self):
  74             optparse.OptionParser.print_help (self)
  75             print '''
  76 DESCRIPTION
  77
  78 This tool is a conversion utility for Darcs repositories, importing
  79 them in chronological order.  It requires a Git version that has
  80 git-fast-import.  It does not support incremental updating.
  81
  82 BUGS
  83
  84  * repositories with skewed timestamps, or different patches with
  85  equal timestamps will confuse darcs2git.
  86  * does not respect file modes or time zones.
  87  * too slow. See source code for instructions to speed it up.
  88  * probably doesn\'t work on partial repositories
  89
  90 Report new bugs to hanwen@xs4all.nl
  91
  92 LICENSE
  93
  94 Copyright (c) 2007 Han-Wen Nienhuys <hanwen@xs4all.nl>.
  95 Distributed under terms of the GNU General Public License
  96 This program comes with NO WARRANTY.
  97 '''
  98
  99     p = MyOP ()
 100
 101     p.usage='''darcs2git [OPTIONS] DARCS-REPO'''
 102     p.description='''Convert darcs repo to git.'''
 103
 104     def update_map (option, opt, value, parser):
 105         for l in open (value).readlines ():
 106             (mail, name) = tuple (l.strip ().split ('='))
 107             mail_to_name_dict[mail] = name
 108
 109     p.add_option ('-a', '--authors', action='callback',
 110                   callback=update_map,
 111                   type='string',
 112                   nargs=1,
 113                   help='read a text file, containing EMAIL=NAME lines')
 114
 115     p.add_option ('--checkpoint-frequency', action='store',
 116                   dest='checkpoint_frequency',
 117                   type='int',
 118                   default=0,
 119                   help='how often should the git importer be synced?\n'
 120                   'Default is 0 (no limit)'
 121                   )
 122
 123     p.add_option ('-d', '--destination', action='store',
 124                   type='string',
 125                   default='',
 126                   dest='target_git_repo',
 127                   help='where to put the resulting Git repo.')
 128
 129     p.add_option ('--verbose', action='store_true',
 130                   dest='verbose',
 131                   default=False,
 132                   help='show commands as they are invoked')
 133
 134     p.add_option ('--history-window', action='store',
 135                   dest='history_window',
 136                   type='int',
 137                   default=0,
 138                   help='Look back this many patches as conflict ancestors.\n'
 139                   'Default is 0 (no limit)')
 140
 141     p.add_option ('--debug', action='store_true',
 142                   dest='debug',
 143                   default=False,
 144                   help="""add patch numbers to commit messages;
 145 don\'t clean conversion repo;
 146 test end result.""")
 147
 148     global options
 149     options, args = p.parse_args ()
 150     if not args:
 151         p.print_help ()
 152         sys.exit (2)
 153
 154     if len(urlparse.urlparse(args[0])) == 0:
 155         raise NotImplementedError, "We support local DARCS repos only."
 156
 157     git_version = distutils.version.LooseVersion(
 158         os.popen("git --version","r").read().strip().split(" ")[-1])
 159     ideal_version = distutils.version.LooseVersion("1.5.0")
 160     if git_version<ideal_version:
 161         raise RuntimeError,"You need git >= 1.5.0 for this."
 162
 163     options.basename = os.path.basename (
 164         os.path.normpath (args[0])).replace ('.darcs', '')
 165     if not options.target_git_repo:
 166         options.target_git_repo = options.basename + '.git'
 167
 168     if options.debug:
 169         global log_file
 170         name = options.target_git_repo.replace ('.git', '.log')
 171         if name == options.target_git_repo:
 172             name += '.log'
 173
 174         progress ("Shell log to %s" % name)
 175         log_file = open (name, 'w')
 176
 177     return (options, args)
 178
 179 def read_pipe (cmd, ignore_errors=False):
 180     if options.verbose:
 181         progress ('pipe %s' % cmd)
 182     pipe = os.popen (cmd)
 183
 184     val = pipe.read ()
 185     if pipe.close () and not ignore_errors:
 186         raise CommandFailed ("Pipe failed: %s" % cmd)
 187
 188     return val
 189
 190 def system (c, ignore_error=0, timed=0):
 191     if timed:
 192         c = "time " + c
 193     if options.verbose:
 194         progress (c)
 195
 196     if log_file:
 197         log_file.write ('%s\n' % c)
 198         log_file.flush ()
 199
 200     if os.system (c) and not ignore_error:
 201         raise CommandFailed ("Command failed: %s" % c)
 202
 203 def darcs_date_to_git (x):
 204     t = time.strptime (x, '%Y%m%d%H%M%S')
 205     return '%d' % int (time.mktime (t))
 206
 207 def darcs_timezone (x) :
 208     time.strptime (x, '%a %b %d %H:%M:%S %Z %Y')
 209
 210     # todo
 211     return "+0100"
 212
 213 ################################################################
 214 # darcs
 215
 216 ## http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/521889
 217
 218 PATCH_DATE_FORMAT = '%Y%m%d%H%M%S'
 219
 220 patch_pattern = r"""
 221    \[                                   # Patch start indicator
 222    (?P<name>[^\n]+)\n                   # Patch name (rest of same line)
 223    (?P<author>[^\*]+)                   # Patch author
 224    \*                                   # Author/date separator
 225    (?P<inverted>[-\*])                  # Inverted patch indicator
 226    (?P<date>\d{14})                     # Patch date
 227    (?:\n(?P<comment>(?:^\ [^\n]*\n)+))? # Optional long comment
 228    \]                                   # Patch end indicator
 229    """
 230 patch_re = re.compile(patch_pattern, re.VERBOSE | re.MULTILINE)
 231 tidy_comment_re = re.compile(r'^ ', re.MULTILINE)
 232
 233 def parse_inventory(inventory):
 234     """
 235     Given the contents of a darcs inventory file, generates ``Patch``
 236     objects representing contained patch details.
 237     """
 238     for match in patch_re.finditer(inventory):
 239         attrs = match.groupdict(None)
 240         attrs['inverted'] = (attrs['inverted'] == '-')
 241         if attrs['comment'] is not None:
 242             attrs['comment'] = tidy_comment_re.sub('', attrs['comment']).strip()
 243         yield InventoryPatch(**attrs)
 244
 245 def fix_braindead_darcs_escapes(s):
 246     def insert_hibit(match):
 247         return chr(int(match.group(1), 16))
 248
 249     return re.sub(r'\[_\\([0-9a-f][0-9a-f])_\]',
 250            insert_hibit, str(s))
 251
 252 class InventoryPatch:
 253     """
 254     Patch details, as defined in a darcs inventory file.
 255
 256     Attribute names match those generated by the
 257     ``darcs changes --xml-output`` command.
 258     """
 259
 260     def __init__(self, name, author, date, inverted, comment=None):
 261         self.name = name
 262         self.author = author
 263         self.date = datetime(*strptime(date, PATCH_DATE_FORMAT)[:6])
 264         self.inverted = inverted
 265         self.comment = comment
 266
 267     def __str__(self):
 268         return self.name
 269
 270     @property
 271     def complete_patch_details(self):
 272         date_str = self.date.strftime(PATCH_DATE_FORMAT)
 273         return '%s%s%s%s%s' % (
 274             self.name, self.author, date_str,
 275             self.comment and ''.join([l.rstrip() for l in self.comment.split('\n')]) or '',
 276             self.inverted and 't' or 'f')
 277
 278     def short_id (self):
 279         inv = '*'
 280         if self.inverted:
 281             inv = '-'
 282
 283         return unicode('%s%s*%s%s' % (self.name, self.author, inv, self.hash.split ('-')[0]), 'UTF-8')
 284
 285     @property
 286     def hash(self):
 287         """
 288         Calculates the filename of the gzipped file containing patch
 289         contents in the repository's ``patches`` directory.
 290
 291         This consists of the patch date, a partial SHA-1 hash of the
 292         patch author and a full SHA-1 hash of the complete patch
 293         details.
 294         """
 295
 296         date_str = self.date.strftime(PATCH_DATE_FORMAT)
 297         return '%s-%s-%s.gz' % (date_str,
 298                                 sha.new(self.author).hexdigest()[:5],
 299                                 sha.new(self.complete_patch_details).hexdigest())
 300
 301 ## http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/521889
 302
 303 class DarcsConversionRepo:
 304     """Representation of a Darcs repo.
 305
 306 The repo is thought to be ordered, and supports methods for
 307 going back (obliterate) and forward (pull).
 308
 309     """
 310
 311     def __init__ (self, dir, patches):
 312         self.dir = os.path.abspath (dir)
 313         self.patches = patches
 314         self._current_number = -1
 315         self._is_valid = -1
 316         self._inventory_dict = None
 317         self._short_id_dict = dict ((p.short_id (), p) for p in patches)
 318
 319     def __del__ (self):
 320         if not options.debug:
 321             system ('rm -fr %s' % self.dir)
 322
 323     def is_contiguous (self):
 324         return (len (self.inventory_dict ()) == self._current_number + 1
 325                 and self.contains_contiguous (self._current_number))
 326
 327     def contains_contiguous (self, num):
 328         if not self._is_valid:
 329             return False
 330
 331         darcs_dir = self.dir + '/_darcs'
 332         if not os.path.exists (darcs_dir):
 333             return False
 334
 335         for p in self.patches[:num + 1]:
 336             if not self.has_patch (p):
 337                 return False
 338
 339         return True
 340
 341     def has_patch (self, p):
 342         assert self._is_valid
 343
 344         return p.short_id () in self.inventory_dict ()
 345
 346     def pristine_tree (self):
 347         return self.dir + '/_darcs/pristine'
 348
 349     def go_back_to (self, dest):
 350
 351         # at 4, len = 5, go to 2: count == 2
 352         count = len (self.inventory_dict()) - dest - 1
 353
 354         assert self._is_valid
 355         assert count > 0
 356
 357         self.checkout ()
 358         dir = self.dir
 359
 360         progress ('Rewinding %d patches' % count)
 361         system ('cd %(dir)s && echo ay|darcs obliterate --ignore-times --last %(count)d' % locals ())
 362         d = self.inventory_dict ()
 363         for p in self.patches[dest+1:self._current_number+1]:
 364             try:
 365                 del d[p.short_id ()]
 366             except KeyError:
 367                 pass
 368
 369         self._current_number = dest
 370
 371     def clean (self):
 372         system ('rm -rf %s' % self.dir)
 373
 374     def checkout (self):
 375         dir = self.dir
 376         system ('rsync -a  %(dir)s/_darcs/pristine/ %(dir)s/' % locals ())
 377
 378     def pull (self, patch):
 379         id = patch.attributes['hash']
 380         source_repo = patch.dir
 381         dir = self.dir
 382
 383         progress ('Pull patch %d' % patch.number)
 384         system ('cd %(dir)s && darcs pull --ignore-times --quiet --all --match "hash %(id)s" %(source_repo)s ' % locals ())
 385
 386         self._current_number = patch.number
 387
 388         ## must reread: the pull may have pulled in others.
 389         self._inventory_dict = None
 390
 391     def go_forward_to (self, num):
 392         d = self.inventory_dict ()
 393
 394         pull_me = []
 395
 396         ## ugh
 397         for p in self.patches[0:num+1]:
 398             if not d.has_key (p.short_id ()):
 399                 pull_me.append (p)
 400                 d[p.short_id ()] = p
 401
 402         pull_str = ' || '.join (['hash %s' % p.id () for p in pull_me])
 403         dir = self.dir
 404         src = self.patches[0].dir
 405
 406         progress ('Pulling %d patches to go to %d' % (len (pull_me), num))
 407         system ('darcs pull --all --repo %(dir)s --match "%(pull_str)s" %(src)s' % locals ())
 408
 409     def create_fresh (self):
 410         dir = self.dir
 411         system ('rm -rf %(dir)s && mkdir %(dir)s && darcs init --repo  %(dir)s'
 412                 % locals ())
 413         self._is_valid = True
 414         self._current_number = -1
 415         self._inventory_dict = {}
 416
 417     def inventory (self):
 418         darcs_dir = self.dir + '/_darcs'
 419         i = ''
 420         for f in [darcs_dir + '/inventory'] + glob.glob (darcs_dir + '/inventories/*'):
 421             i += open (f).read ()
 422         return i
 423
 424     def inventory_dict (self):
 425
 426         if type (self._inventory_dict) != type ({}):
 427             self._inventory_dict = {}
 428
 429             for p in parse_inventory(self.inventory()):
 430                 key = p.short_id()
 431
 432                 try:
 433                     self._inventory_dict[key] = self._short_id_dict[key]
 434                 except KeyError:
 435                     print 'key not found', key
 436                     print self._short_id_dict
 437                     raise
 438
 439         return self._inventory_dict
 440
 441     def start_at (self, num):
 442         """Move the repo to NUM.
 443
 444         This uses the fishy technique of writing the inventory and
 445         constructing the pristine tree with 'darcs repair'
 446         """
 447         progress ('Starting afresh at %d' % num)
 448
 449         self.create_fresh ()
 450         dir = self.dir
 451         iv = open (dir + '/_darcs/inventory', 'w')
 452         if log_file:
 453             log_file.write ("# messing with _darcs/inventory")
 454
 455         for p in self.patches[:num+1]:
 456             os.link (p.filename (), dir + '/_darcs/patches/' + os.path.basename (p.filename ()))
 457             iv.write (p.header ())
 458             self._inventory_dict[p.short_id ()] = p
 459         iv.close ()
 460
 461         system ('darcs repair --repo %(dir)s --quiet' % locals ())
 462         self.checkout ()
 463         self._current_number = num
 464         self._is_valid = True
 465
 466     def go_to (self, dest):
 467         if not self._is_valid:
 468             self.start_at (dest)
 469         elif dest == self._current_number and self.is_contiguous ():
 470             pass
 471         elif (self.contains_contiguous (dest)):
 472             self.go_back_to (dest)
 473         elif dest - len (self.inventory_dict ()) < dest / 100:
 474             self.go_forward_to (dest)
 475         else:
 476             self.start_at (dest)
 477
 478
 479     def go_from_to (self, from_patch, to_patch):
 480
 481         """Move the repo to FROM_PATCH, then go to TO_PATCH. Raise
 482         PullConflict if conflict is detected
 483         """
 484
 485         progress ('Trying %s -> %s' % (from_patch, to_patch))
 486         dir = self.dir
 487         source = to_patch.dir
 488
 489         if from_patch:
 490             self.go_to (from_patch.number)
 491         else:
 492             self.create_fresh ()
 493
 494         try:
 495             self.pull (to_patch)
 496             success = 'No conflicts to resolve' in read_pipe ('cd %(dir)s && echo y|darcs resolve' % locals  ())
 497         except CommandFailed:
 498             self._is_valid = False
 499             raise PullConflict ()
 500
 501         if not success:
 502             raise PullConflict ()
 503
 504 class DarcsPatch:
 505     def __repr__ (self):
 506         return 'patch %d' % self.number
 507
 508     def __init__ (self, xml, dir):
 509         self.xml = xml
 510         self.dir = dir
 511         self.number = -1
 512         self.attributes = {}
 513         self._contents = None
 514         for (nm, value) in xml.attributes.items():
 515             self.attributes[nm] = value
 516
 517         # fixme: ugh attributes vs. methods.
 518         self.extract_author ()
 519         self.extract_message ()
 520         self.extract_time ()
 521
 522     def id (self):
 523         return self.attributes['hash']
 524
 525     def short_id (self):
 526         inv = '*'
 527         if self.attributes['inverted'] == 'True':
 528             inv = '-'
 529
 530         return '%s%s*%s%s' % (self.name(), self.attributes['author'], inv, self.attributes['hash'].split ('-')[0])
 531
 532     def filename (self):
 533         return self.dir + '/_darcs/patches/' + self.attributes['hash']
 534
 535     def contents (self):
 536         if type (self._contents) != type (''):
 537             f = gzip.open (self.filename ())
 538             self._contents = f.read ()
 539
 540         return self._contents
 541
 542     def header (self):
 543         lines = self.contents ().split ('\n')
 544
 545         name = lines[0]
 546         committer = lines[1] + '\n'
 547         committer = re.sub ('] {\n$', ']\n', committer)
 548         committer = re.sub ('] *\n$', ']\n', committer)
 549         comment = ''
 550         if not committer.endswith (']\n'):
 551             for l in lines[2:]:
 552                 if l[0] == ']':
 553                     comment += ']\n'
 554                     break
 555                 comment += l + '\n'
 556
 557         header = name  + '\n' + committer
 558         if comment:
 559             header += comment
 560
 561         assert header[-1] == '\n'
 562         return header
 563
 564     def extract_author (self):
 565         mail = self.attributes['author']
 566         name = ''
 567         m = re.search ("^(.*) <(.*)>$", mail)
 568
 569         if m:
 570             name = m.group (1)
 571             mail = m.group (2)
 572         else:
 573             try:
 574                 name = mail_to_name_dict[mail]
 575             except KeyError:
 576                 name = mail.split ('@')[0]
 577
 578         self.author_name = name
 579
 580         # mail addresses should be plain strings.
 581         self.author_mail = mail.encode('utf-8')
 582
 583     def extract_time (self):
 584         self.date = darcs_date_to_git (self.attributes['date']) + ' ' + darcs_timezone (self.attributes['local_date'])
 585
 586     def name (self):
 587         patch_name = ''
 588         try:
 589             name_elt = self.xml.getElementsByTagName ('name')[0]
 590             patch_name = unicode(fix_braindead_darcs_escapes(str(name_elt.childNodes[0].data)), 'UTF-8')
 591         except IndexError:
 592             pass
 593         return patch_name
 594
 595     def extract_message (self):
 596         patch_name = self.name ()
 597         comment_elts = self.xml.getElementsByTagName ('comment')
 598         comment = ''
 599         if comment_elts:
 600             comment = comment_elts[0].childNodes[0].data
 601
 602         if self.attributes['inverted'] == 'True':
 603             patch_name = 'UNDO: ' + patch_name
 604
 605         self.message = '%s\n\n%s' % (patch_name, comment)
 606
 607     def tag_name (self):
 608         patch_name = self.name ()
 609         if patch_name.startswith ("TAG "):
 610            tag = patch_name[4:]
 611            tag = re.sub (r'\s', '_', tag).strip ()
 612            tag = re.sub (r':', '_', tag).strip ()
 613            return tag
 614         return ''
 615
 616 def get_darcs_patches (darcs_repo):
 617     progress ('reading patches.')
 618
 619     xml_string = read_pipe ('darcs changes --xml --reverse --repo ' + darcs_repo)
 620
 621     dom = xml.dom.minidom.parseString(xml_string)
 622     xmls = dom.documentElement.getElementsByTagName('patch')
 623
 624     patches = [DarcsPatch (x, darcs_repo) for x in xmls]
 625
 626     n = 0
 627     for p in patches:
 628         p.number = n
 629         n += 1
 630
 631     return patches
 632
 633 ################################################################
 634 # GIT export
 635
 636 class GitCommit:
 637     def __init__ (self, parent, darcs_patch):
 638         self.parent = parent
 639         self.darcs_patch = darcs_patch
 640         if parent:
 641             self.depth = parent.depth + 1
 642         else:
 643             self.depth = 0
 644
 645     def number (self):
 646         return self.darcs_patch.number
 647
 648     def parent_patch (self):
 649         if self.parent:
 650             return self.parent.darcs_patch
 651         else:
 652             return None
 653
 654 def common_ancestor (a, b):
 655     while 1:
 656         if a.depth < b.depth:
 657             b = b.parent
 658         elif a.depth > b.depth:
 659             a = a.parent
 660         else:
 661             break
 662
 663     while a and b:
 664         if a == b:
 665             return a
 666
 667         a = a.parent
 668         b = b.parent
 669
 670     return None
 671
 672 def export_checkpoint (gfi):
 673     gfi.write ('checkpoint\n\n')
 674
 675 def export_tree (tree, gfi):
 676     tree = os.path.normpath (tree)
 677     gfi.write ('deleteall\n')
 678     for (root, dirs, files) in os.walk (tree):
 679         for f in files:
 680             rf = os.path.normpath (os.path.join (root, f))
 681             s = open (rf).read ()
 682             rf = rf.replace (tree + '/', '')
 683
 684             gfi.write ('M 644 inline %s\n' % rf)
 685             gfi.write ('data %d\n%s\n' % (len (s), s))
 686     gfi.write ('\n')
 687
 688
 689 def export_commit (repo, patch, last_patch, gfi):
 690     gfi.write ('commit refs/heads/darcstmp%d\n' % patch.number)
 691     gfi.write ('mark :%d\n' % (patch.number + 1))
 692
 693     raw_name = patch.author_name
 694     gfi.write ('committer %s <%s> %s\n' % (raw_name,
 695                                            patch.author_mail,
 696                                            patch.date))
 697
 698     msg = patch.message
 699     if options.debug:
 700         msg += '\n\n#%d\n' % patch.number
 701
 702     msg = msg.encode('utf-8')
 703     gfi.write ('data %d\n%s\n' % (len (msg), msg))
 704     mergers = []
 705     for (n, p) in pending_patches.items ():
 706         if repo.has_patch (p):
 707             mergers.append (n)
 708             del pending_patches[n]
 709
 710     if (last_patch
 711         and mergers == []
 712         and git_commits.has_key (last_patch.number)):
 713         mergers = [last_patch.number]
 714
 715     if mergers:
 716         gfi.write ('from :%d\n' % (mergers[0] + 1))
 717         for m in mergers[1:]:
 718             gfi.write ('merge :%d\n' % (m + 1))
 719
 720     pending_patches[patch.number] = patch
 721     export_tree (repo.pristine_tree (), gfi)
 722
 723     n = -1
 724     if last_patch:
 725         n = last_patch.number
 726     git_commits[patch.number] = GitCommit (git_commits.get (n, None),
 727                                            patch)
 728
 729 def export_pending (gfi):
 730     if len (pending_patches.items ()) == 1:
 731         gfi.write ('reset refs/heads/master\n')
 732         gfi.write ('from :%d\n\n' % (pending_patches.values()[0].number+1))
 733
 734         progress ("Creating branch master")
 735         return
 736
 737     for (n, p) in pending_patches.items ():
 738         gfi.write ('reset refs/heads/master%d\n' % n)
 739         gfi.write ('from :%d\n\n' % (n+1))
 740
 741         progress ("Creating branch master%d" % n)
 742
 743     patches = pending_patches.values()
 744     patch = patches[0]
 745     gfi.write ('commit refs/heads/master\n')
 746     gfi.write ('committer %s <%s> %s\n' % (patch.author_name,
 747                                            patch.author_mail,
 748                                            patch.date))
 749     msg = 'tie together'
 750     gfi.write ('data %d\n%s\n' % (len(msg), msg))
 751     gfi.write ('from :%d\n' % (patch.number + 1))
 752     for p in patches[1:]:
 753         gfi.write ('merge :%d\n' % (p.number + 1))
 754     gfi.write ('\n')
 755
 756 def export_tag (patch, gfi):
 757     gfi.write ('tag %s\n' % patch.tag_name ())
 758     gfi.write ('from :%d\n' % (patch.number + 1))
 759     gfi.write ('tagger %s <%s> %s\n' % (patch.author_name,
 760                                         patch.author_mail,
 761                                         patch.date))
 762
 763     raw_message = patch.message.encode('utf-8')
 764     gfi.write ('data %d\n%s\n' % (len (raw_message),
 765                                   raw_message))
 766
 767 ################################################################
 768 # main.
 769
 770 def test_conversion (darcs_repo, git_repo):
 771     pristine = '%(darcs_repo)s/_darcs/pristine' % locals ()
 772     if not os.path.exists (pristine):
 773         progress ("darcs repository does not contain pristine tree?!")
 774         return
 775
 776     gd = options.basename + '.checkouttmp.git'
 777     system ('rm -rf %(gd)s && git clone %(git_repo)s %(gd)s' % locals ())
 778     diff_cmd = 'diff --exclude .git -urN %(gd)s %(pristine)s' % locals ()
 779     diff = read_pipe (diff_cmd, ignore_errors=True)
 780     system ('rm -rf %(gd)s' % locals ())
 781
 782     if diff:
 783         if len (diff) > 1024:
 784             diff = diff[:512] + '\n...\n'  +  diff[-512:]
 785
 786         progress ("Conversion introduced changes: %s" % diff)
 787         raise 'fdsa'
 788     else:
 789         progress ("Checkout matches pristine darcs tree.")
 790
 791 def main ():
 792     (options, args) = get_cli_options ()
 793
 794     darcs_repo = os.path.abspath (args[0])
 795     git_repo = os.path.abspath (options.target_git_repo)
 796
 797     if os.path.exists (git_repo):
 798         system ('rm -rf %(git_repo)s' % locals ())
 799
 800     system ('mkdir %(git_repo)s && cd %(git_repo)s && git --bare init' % locals ())
 801     system ('git --git-dir %(git_repo)s  repo-config core.logAllRefUpdates false' % locals ())
 802
 803     os.environ['GIT_DIR'] = git_repo
 804
 805     quiet = ' --quiet'
 806     if options.verbose:
 807         quiet = ' '
 808
 809     gfi = os.popen ('git-fast-import %s' % quiet, 'w')
 810
 811     patches = get_darcs_patches (darcs_repo)
 812     conv_repo = DarcsConversionRepo (options.basename + ".tmpdarcs", patches)
 813     conv_repo.start_at (-1)
 814
 815     for p in patches:
 816         parent_patch = None
 817         parent_number = -1
 818
 819         combinations = [(v, w) for v in pending_patches.values ()
 820                         for w in pending_patches.values ()]
 821         candidates = [common_ancestor (git_commits[c[0].number], git_commits[c[1].number]) for c in combinations]
 822         candidates = sorted ([(-a.darcs_patch.number, a) for a in candidates])
 823         for (depth, c) in candidates:
 824             q = c.darcs_patch
 825             try:
 826                 conv_repo.go_from_to (q, p)
 827
 828                 parent_patch = q
 829                 parent_number = q.number
 830                 progress ('Found existing common parent as predecessor')
 831                 break
 832
 833             except PullConflict:
 834                 pass
 835
 836         ## no branches found where we could attach.
 837         ## try previous commits one by one.
 838         if not parent_patch:
 839             parent_number = p.number - 2
 840             while 1:
 841                 if parent_number >= 0:
 842                     parent_patch = patches[parent_number]
 843
 844                 try:
 845                     conv_repo.go_from_to (parent_patch, p)
 846                     break
 847                 except PullConflict:
 848
 849                     ## simplistic, may not be enough.
 850                     progress ('conflict, going one back')
 851                     parent_number -= 1
 852
 853                     if parent_number < 0:
 854                         break
 855
 856                     if (options.history_window
 857                         and parent_number < p.number - options.history_window):
 858
 859                         parent_number = -2
 860                         break
 861
 862         if parent_number >= 0 or p.number == 0:
 863             progress ('Export %d -> %d (total %d)' % (parent_number,
 864                                                       p.number, len (patches)))
 865             export_commit (conv_repo, p, parent_patch, gfi)
 866             if p.tag_name ():
 867                 export_tag (p, gfi)
 868
 869             if options.checkpoint_frequency and p.number % options.checkpoint_frequency == 0:
 870                 export_checkpoint (gfi)
 871         else:
 872             progress ("Can't import patch %d, need conflict resolution patch?"
 873                       % p.number)
 874
 875     export_pending (gfi)
 876     gfi.close ()
 877     for f in glob.glob ('%(git_repo)s/refs/heads/darcstmp*' % locals ()):
 878         os.unlink (f)
 879
 880     test_conversion (darcs_repo, git_repo)
 881
 882     if not options.debug:
 883         conv_repo.clean ()
 884
 885 main ()