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     p.add_option ('-e', '--encoding', action='store',
 149                   type='string',
 150                   default='UTF-8',
 151                   nargs=1,
 152                   help='specify encoding, for example ISO-8859-2')
 153
 154     global options
 155     options, args = p.parse_args ()
 156     if not args:
 157         p.print_help ()
 158         sys.exit (2)
 159
 160     if len(urlparse.urlparse(args[0])) == 0:
 161         raise NotImplementedError, "We support local DARCS repos only."
 162
 163     git_version = distutils.version.LooseVersion(
 164         os.popen("git --version","r").read().strip().split(" ")[-1])
 165     ideal_version = distutils.version.LooseVersion("1.5.0")
 166     if git_version<ideal_version:
 167         raise RuntimeError,"You need git >= 1.5.0 for this."
 168
 169     options.basename = os.path.basename (
 170         os.path.normpath (args[0])).replace ('.darcs', '')
 171     if not options.target_git_repo:
 172         options.target_git_repo = options.basename + '.git'
 173
 174     if options.debug:
 175         global log_file
 176         name = options.target_git_repo.replace ('.git', '.log')
 177         if name == options.target_git_repo:
 178             name += '.log'
 179
 180         progress ("Shell log to %s" % name)
 181         log_file = open (name, 'w')
 182
 183     return (options, args)
 184
 185 def read_pipe (cmd, ignore_errors=False):
 186     if options.verbose:
 187         progress ('pipe %s' % cmd)
 188     pipe = os.popen (cmd)
 189
 190     val = pipe.read ()
 191     if pipe.close () and not ignore_errors:
 192         raise CommandFailed ("Pipe failed: %s" % cmd)
 193
 194     return val
 195
 196 def system (c, ignore_error=0, timed=0):
 197     if timed:
 198         c = "time " + c
 199     if options.verbose:
 200         progress (c)
 201
 202     if log_file:
 203         log_file.write ('%s\n' % c)
 204         log_file.flush ()
 205
 206     if os.system (c) and not ignore_error:
 207         raise CommandFailed ("Command failed: %s" % c)
 208
 209 def darcs_date_to_git (x):
 210     t = time.strptime (x, '%Y%m%d%H%M%S')
 211     return '%d' % int (time.mktime (t))
 212
 213 def darcs_timezone (x) :
 214     time.strptime (x, '%a %b %d %H:%M:%S %Z %Y')
 215
 216     # todo
 217     return "+0100"
 218
 219 ################################################################
 220 # darcs
 221
 222 ## http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/521889
 223
 224 PATCH_DATE_FORMAT = '%Y%m%d%H%M%S'
 225
 226 patch_pattern = r"""
 227    \[                                   # Patch start indicator
 228    (?P<name>[^\n]+)\n                   # Patch name (rest of same line)
 229    (?P<author>[^\*]+)                   # Patch author
 230    \*                                   # Author/date separator
 231    (?P<inverted>[-\*])                  # Inverted patch indicator
 232    (?P<date>\d{14})                     # Patch date
 233    (?:\n(?P<comment>(?:^\ [^\n]*\n)+))? # Optional long comment
 234    \]                                   # Patch end indicator
 235    """
 236 patch_re = re.compile(patch_pattern, re.VERBOSE | re.MULTILINE)
 237 tidy_comment_re = re.compile(r'^ ', re.MULTILINE)
 238
 239 def parse_inventory(inventory):
 240     """
 241     Given the contents of a darcs inventory file, generates ``Patch``
 242     objects representing contained patch details.
 243     """
 244     for match in patch_re.finditer(inventory):
 245         attrs = match.groupdict(None)
 246         attrs['inverted'] = (attrs['inverted'] == '-')
 247         if attrs['comment'] is not None:
 248             attrs['comment'] = tidy_comment_re.sub('', attrs['comment']).strip()
 249         yield InventoryPatch(**attrs)
 250
 251 def fix_braindead_darcs_escapes(s):
 252     def insert_hibit(match):
 253         return chr(int(match.group(1), 16))
 254
 255     return re.sub(r'\[_\\([0-9a-f][0-9a-f])_\]',
 256            insert_hibit, str(s))
 257
 258 class InventoryPatch:
 259     """
 260     Patch details, as defined in a darcs inventory file.
 261
 262     Attribute names match those generated by the
 263     ``darcs changes --xml-output`` command.
 264     """
 265
 266     def __init__(self, name, author, date, inverted, comment=None):
 267         self.name = name
 268         self.author = author
 269         self.date = datetime(*strptime(date, PATCH_DATE_FORMAT)[:6])
 270         self.inverted = inverted
 271         self.comment = comment
 272
 273     def __str__(self):
 274         return self.name
 275
 276     @property
 277     def complete_patch_details(self):
 278         date_str = self.date.strftime(PATCH_DATE_FORMAT)
 279         return '%s%s%s%s%s' % (
 280             self.name, self.author, date_str,
 281             self.comment and ''.join([l.rstrip() for l in self.comment.split('\n')]) or '',
 282             self.inverted and 't' or 'f')
 283
 284     def short_id (self):
 285         inv = '*'
 286         if self.inverted:
 287             inv = '-'
 288
 289         return unicode('%s%s*%s%s' % (self.name, self.author, inv, self.hash.split ('-')[0]), options.encoding)
 290
 291     @property
 292     def hash(self):
 293         """
 294         Calculates the filename of the gzipped file containing patch
 295         contents in the repository's ``patches`` directory.
 296
 297         This consists of the patch date, a partial SHA-1 hash of the
 298         patch author and a full SHA-1 hash of the complete patch
 299         details.
 300         """
 301
 302         date_str = self.date.strftime(PATCH_DATE_FORMAT)
 303         return '%s-%s-%s.gz' % (date_str,
 304                                 sha.new(self.author).hexdigest()[:5],
 305                                 sha.new(self.complete_patch_details).hexdigest())
 306
 307 ## http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/521889
 308
 309 class DarcsConversionRepo:
 310     """Representation of a Darcs repo.
 311
 312 The repo is thought to be ordered, and supports methods for
 313 going back (obliterate) and forward (pull).
 314
 315     """
 316
 317     def __init__ (self, dir, patches):
 318         self.dir = os.path.abspath (dir)
 319         self.patches = patches
 320         self._current_number = -1
 321         self._is_valid = -1
 322         self._inventory_dict = None
 323         self._short_id_dict = dict ((p.short_id (), p) for p in patches)
 324
 325     def __del__ (self):
 326         if not options.debug:
 327             system ('rm -fr %s' % self.dir)
 328
 329     def is_contiguous (self):
 330         return (len (self.inventory_dict ()) == self._current_number + 1
 331                 and self.contains_contiguous (self._current_number))
 332
 333     def contains_contiguous (self, num):
 334         if not self._is_valid:
 335             return False
 336
 337         darcs_dir = self.dir + '/_darcs'
 338         if not os.path.exists (darcs_dir):
 339             return False
 340
 341         for p in self.patches[:num + 1]:
 342             if not self.has_patch (p):
 343                 return False
 344
 345         return True
 346
 347     def has_patch (self, p):
 348         assert self._is_valid
 349
 350         return p.short_id () in self.inventory_dict ()
 351
 352     def pristine_tree (self):
 353         return self.dir + '/_darcs/pristine'
 354
 355     def go_back_to (self, dest):
 356
 357         # at 4, len = 5, go to 2: count == 2
 358         count = len (self.inventory_dict()) - dest - 1
 359
 360         assert self._is_valid
 361         assert count > 0
 362
 363         self.checkout ()
 364         dir = self.dir
 365
 366         progress ('Rewinding %d patches' % count)
 367         system ('cd %(dir)s && darcs revert --all' % locals())
 368         system ('cd %(dir)s && yes|darcs obliterate -a --ignore-times --last %(count)d' % locals ())
 369         system ('cd %(dir)s && darcs revert -a' % locals())
 370         d = self.inventory_dict ()
 371         for p in self.patches[dest+1:self._current_number+1]:
 372             try:
 373                 del d[p.short_id ()]
 374             except KeyError:
 375                 pass
 376
 377         self._current_number = dest
 378
 379     def clean (self):
 380         system ('rm -rf %s' % self.dir)
 381
 382     def checkout (self):
 383         dir = self.dir
 384         system ('rsync -a  %(dir)s/_darcs/pristine/ %(dir)s/' % locals ())
 385
 386     def pull (self, patch):
 387         id = patch.attributes['hash']
 388         source_repo = patch.dir
 389         dir = self.dir
 390
 391         progress ('Pull patch %d' % patch.number)
 392         system ('cd %(dir)s && darcs revert --all' % locals())
 393         system ('cd %(dir)s && darcs pull --ignore-times --quiet --all --match "hash %(id)s" %(source_repo)s ' % locals ())
 394
 395         self._current_number = patch.number
 396
 397         ## must reread: the pull may have pulled in others.
 398         self._inventory_dict = None
 399
 400     def go_forward_to (self, num):
 401         d = self.inventory_dict ()
 402
 403         pull_me = []
 404
 405         ## ugh
 406         for p in self.patches[0:num+1]:
 407             if not d.has_key (p.short_id ()):
 408                 pull_me.append (p)
 409                 d[p.short_id ()] = p
 410
 411         pull_str = ' || '.join (['hash %s' % p.id () for p in pull_me])
 412         dir = self.dir
 413         src = self.patches[0].dir
 414
 415         progress ('Pulling %d patches to go to %d' % (len (pull_me), num))
 416         system ('darcs revert --repo %(dir)s  --all' % locals ())
 417         system ('darcs pull --all --repo %(dir)s --match "%(pull_str)s" %(src)s' % locals ())
 418
 419     def create_fresh (self):
 420         dir = self.dir
 421         system ('rm -rf %(dir)s && mkdir %(dir)s && darcs init --repo  %(dir)s'
 422                 % locals ())
 423         self._is_valid = True
 424         self._current_number = -1
 425         self._inventory_dict = {}
 426
 427     def inventory (self):
 428         darcs_dir = self.dir + '/_darcs'
 429         i = ''
 430         for f in [darcs_dir + '/inventory'] + glob.glob (darcs_dir + '/inventories/*'):
 431             i += open (f).read ()
 432         return i
 433
 434     def inventory_dict (self):
 435
 436         if type (self._inventory_dict) != type ({}):
 437             self._inventory_dict = {}
 438
 439             for p in parse_inventory(self.inventory()):
 440                 key = p.short_id()
 441
 442                 try:
 443                     self._inventory_dict[key] = self._short_id_dict[key]
 444                 except KeyError:
 445                     print 'key not found', key
 446                     print self._short_id_dict
 447                     raise
 448
 449         return self._inventory_dict
 450
 451     def start_at (self, num):
 452         """Move the repo to NUM.
 453
 454         This uses the fishy technique of writing the inventory and
 455         constructing the pristine tree with 'darcs repair'
 456         """
 457         progress ('Starting afresh at %d' % num)
 458
 459         self.create_fresh ()
 460         dir = self.dir
 461         iv = open (dir + '/_darcs/inventory', 'w')
 462         if log_file:
 463             log_file.write ("# messing with _darcs/inventory")
 464
 465         for p in self.patches[:num+1]:
 466             os.link (p.filename (), dir + '/_darcs/patches/' + os.path.basename (p.filename ()))
 467             iv.write (p.header ())
 468             self._inventory_dict[p.short_id ()] = p
 469         iv.close ()
 470
 471         system ('darcs revert --repo %(dir)s --all' % locals())
 472         system ('darcs repair --repo %(dir)s --quiet' % locals ())
 473         self.checkout ()
 474         self._current_number = num
 475         self._is_valid = True
 476
 477     def go_to (self, dest):
 478         if not self._is_valid:
 479             self.start_at (dest)
 480         elif dest == self._current_number and self.is_contiguous ():
 481             pass
 482         elif (self.contains_contiguous (dest)):
 483             self.go_back_to (dest)
 484         elif dest - len (self.inventory_dict ()) < dest / 100:
 485             self.go_forward_to (dest)
 486         else:
 487             self.start_at (dest)
 488
 489
 490     def go_from_to (self, from_patch, to_patch):
 491
 492         """Move the repo to FROM_PATCH, then go to TO_PATCH. Raise
 493         PullConflict if conflict is detected
 494         """
 495
 496         progress ('Trying %s -> %s' % (from_patch, to_patch))
 497         dir = self.dir
 498         source = to_patch.dir
 499
 500         if from_patch:
 501             self.go_to (from_patch.number)
 502         else:
 503             self.create_fresh ()
 504
 505         try:
 506             self.pull (to_patch)
 507             success = 'No conflicts to resolve' in read_pipe ('cd %(dir)s && echo y|darcs resolve' % locals  ())
 508         except CommandFailed:
 509             self._is_valid = False
 510             raise PullConflict ()
 511
 512         if not success:
 513             raise PullConflict ()
 514
 515 class DarcsPatch:
 516     def __repr__ (self):
 517         return 'patch %d' % self.number
 518
 519     def __init__ (self, xml, dir):
 520         self.xml = xml
 521         self.dir = dir
 522         self.number = -1
 523         self.attributes = {}
 524         self._contents = None
 525         for (nm, value) in xml.attributes.items():
 526             self.attributes[nm] = value
 527
 528         # fixme: ugh attributes vs. methods.
 529         self.extract_author ()
 530         self.extract_message ()
 531         self.extract_time ()
 532
 533     def id (self):
 534         return self.attributes['hash']
 535
 536     def short_id (self):
 537         inv = '*'
 538         if self.attributes['inverted'] == 'True':
 539             inv = '-'
 540
 541         return '%s%s*%s%s' % (self.name(), self.attributes['author'], inv, self.attributes['hash'].split ('-')[0])
 542
 543     def filename (self):
 544         return self.dir + '/_darcs/patches/' + self.attributes['hash']
 545
 546     def contents (self):
 547         if type (self._contents) != type (''):
 548             f = gzip.open (self.filename ())
 549             self._contents = f.read ()
 550
 551         return self._contents
 552
 553     def header (self):
 554         lines = self.contents ().split ('\n')
 555
 556         name = lines[0]
 557         committer = lines[1] + '\n'
 558         committer = re.sub ('] {\n$', ']\n', committer)
 559         committer = re.sub ('] *\n$', ']\n', committer)
 560         comment = ''
 561         if not committer.endswith (']\n'):
 562             for l in lines[2:]:
 563                 if l[0] == ']':
 564                     comment += ']\n'
 565                     break
 566                 comment += l + '\n'
 567
 568         header = name  + '\n' + committer
 569         if comment:
 570             header += comment
 571
 572         assert header[-1] == '\n'
 573         return header
 574
 575     def extract_author (self):
 576         mail = self.attributes['author']
 577         name = ''
 578         m = re.search ("^(.*) <(.*)>$", mail)
 579
 580         if m:
 581             name = m.group (1)
 582             mail = m.group (2)
 583         else:
 584             try:
 585                 name = mail_to_name_dict[mail]
 586             except KeyError:
 587                 name = mail.split ('@')[0]
 588
 589         self.author_name = name
 590
 591         # mail addresses should be plain strings.
 592         self.author_mail = mail.encode('utf-8')
 593
 594     def extract_time (self):
 595         self.date = darcs_date_to_git (self.attributes['date']) + ' ' + darcs_timezone (self.attributes['local_date'])
 596
 597     def name (self):
 598         patch_name = ''
 599         try:
 600             name_elt = self.xml.getElementsByTagName ('name')[0]
 601             patch_name = unicode(fix_braindead_darcs_escapes(str(name_elt.childNodes[0].data)), options.encoding)
 602         except IndexError:
 603             pass
 604         return patch_name
 605
 606     def extract_message (self):
 607         patch_name = self.name ()
 608         comment_elts = self.xml.getElementsByTagName ('comment')
 609         comment = ''
 610         if comment_elts:
 611             comment = comment_elts[0].childNodes[0].data
 612
 613         if self.attributes['inverted'] == 'True':
 614             patch_name = 'UNDO: ' + patch_name
 615
 616         self.message = '%s\n\n%s' % (patch_name, comment)
 617
 618     def tag_name (self):
 619         patch_name = self.name ()
 620         if patch_name.startswith ("TAG "):
 621            tag = patch_name[4:]
 622            tag = re.sub (r'\s', '_', tag).strip ()
 623            tag = re.sub (r':', '_', tag).strip ()
 624            return tag
 625         return ''
 626
 627 def get_darcs_patches (darcs_repo):
 628     progress ('reading patches.')
 629
 630     xml_string = read_pipe ('darcs changes --xml --reverse --repo ' + darcs_repo)
 631
 632     dom = xml.dom.minidom.parseString(xml_string)
 633     xmls = dom.documentElement.getElementsByTagName('patch')
 634
 635     patches = [DarcsPatch (x, darcs_repo) for x in xmls]
 636
 637     n = 0
 638     for p in patches:
 639         p.number = n
 640         n += 1
 641
 642     return patches
 643
 644 ################################################################
 645 # GIT export
 646
 647 class GitCommit:
 648     def __init__ (self, parent, darcs_patch):
 649         self.parent = parent
 650         self.darcs_patch = darcs_patch
 651         if parent:
 652             self.depth = parent.depth + 1
 653         else:
 654             self.depth = 0
 655
 656     def number (self):
 657         return self.darcs_patch.number
 658
 659     def parent_patch (self):
 660         if self.parent:
 661             return self.parent.darcs_patch
 662         else:
 663             return None
 664
 665 def common_ancestor (a, b):
 666     while 1:
 667         if a.depth < b.depth:
 668             b = b.parent
 669         elif a.depth > b.depth:
 670             a = a.parent
 671         else:
 672             break
 673
 674     while a and b:
 675         if a == b:
 676             return a
 677
 678         a = a.parent
 679         b = b.parent
 680
 681     return None
 682
 683 def export_checkpoint (gfi):
 684     gfi.write ('checkpoint\n\n')
 685
 686 def export_tree (tree, gfi):
 687     tree = os.path.normpath (tree)
 688     gfi.write ('deleteall\n')
 689     for (root, dirs, files) in os.walk (tree):
 690         for f in files:
 691             rf = os.path.normpath (os.path.join (root, f))
 692             s = open (rf).read ()
 693             rf = rf.replace (tree + '/', '')
 694
 695             gfi.write ('M 644 inline %s\n' % rf)
 696             gfi.write ('data %d\n%s\n' % (len (s), s))
 697     gfi.write ('\n')
 698
 699
 700 def export_commit (repo, patch, last_patch, gfi):
 701     gfi.write ('commit refs/heads/darcstmp%d\n' % patch.number)
 702     gfi.write ('mark :%d\n' % (patch.number + 1))
 703
 704     raw_name = patch.author_name
 705     gfi.write ('committer %s <%s> %s\n' % (raw_name,
 706                                            patch.author_mail,
 707                                            patch.date))
 708
 709     msg = patch.message
 710     if options.debug:
 711         msg += '\n\n#%d\n' % patch.number
 712
 713     msg = msg.encode('utf-8')
 714     gfi.write ('data %d\n%s\n' % (len (msg), msg))
 715     mergers = []
 716     for (n, p) in pending_patches.items ():
 717         if repo.has_patch (p):
 718             mergers.append (n)
 719             del pending_patches[n]
 720
 721     if (last_patch
 722         and mergers == []
 723         and git_commits.has_key (last_patch.number)):
 724         mergers = [last_patch.number]
 725
 726     if mergers:
 727         gfi.write ('from :%d\n' % (mergers[0] + 1))
 728         for m in mergers[1:]:
 729             gfi.write ('merge :%d\n' % (m + 1))
 730
 731     pending_patches[patch.number] = patch
 732     export_tree (repo.pristine_tree (), gfi)
 733
 734     n = -1
 735     if last_patch:
 736         n = last_patch.number
 737     git_commits[patch.number] = GitCommit (git_commits.get (n, None),
 738                                            patch)
 739
 740 def export_pending (gfi):
 741     if len (pending_patches.items ()) == 1:
 742         gfi.write ('reset refs/heads/master\n')
 743         gfi.write ('from :%d\n\n' % (pending_patches.values()[0].number+1))
 744
 745         progress ("Creating branch master")
 746         return
 747
 748     for (n, p) in pending_patches.items ():
 749         gfi.write ('reset refs/heads/master%d\n' % n)
 750         gfi.write ('from :%d\n\n' % (n+1))
 751
 752         progress ("Creating branch master%d" % n)
 753
 754     patches = pending_patches.values()
 755     patch = patches[0]
 756     gfi.write ('commit refs/heads/master\n')
 757     gfi.write ('committer %s <%s> %s\n' % (patch.author_name,
 758                                            patch.author_mail,
 759                                            patch.date))
 760     msg = 'tie together'
 761     gfi.write ('data %d\n%s\n' % (len(msg), msg))
 762     gfi.write ('from :%d\n' % (patch.number + 1))
 763     for p in patches[1:]:
 764         gfi.write ('merge :%d\n' % (p.number + 1))
 765     gfi.write ('\n')
 766
 767 def export_tag (patch, gfi):
 768     gfi.write ('tag %s\n' % patch.tag_name ())
 769     gfi.write ('from :%d\n' % (patch.number + 1))
 770     gfi.write ('tagger %s <%s> %s\n' % (patch.author_name,
 771                                         patch.author_mail,
 772                                         patch.date))
 773
 774     raw_message = patch.message.encode('utf-8')
 775     gfi.write ('data %d\n%s\n' % (len (raw_message),
 776                                   raw_message))
 777
 778 ################################################################
 779 # main.
 780
 781 def test_conversion (darcs_repo, git_repo):
 782     pristine = '%(darcs_repo)s/_darcs/pristine' % locals ()
 783     if not os.path.exists (pristine):
 784         progress ("darcs repository does not contain pristine tree?!")
 785         return
 786
 787     gd = options.basename + '.checkouttmp.git'
 788     system ('rm -rf %(gd)s && git clone %(git_repo)s %(gd)s' % locals ())
 789     diff_cmd = 'diff --exclude .git -urN %(gd)s %(pristine)s' % locals ()
 790     diff = read_pipe (diff_cmd, ignore_errors=True)
 791     system ('rm -rf %(gd)s' % locals ())
 792
 793     if diff:
 794         if len (diff) > 1024:
 795             diff = diff[:512] + '\n...\n'  +  diff[-512:]
 796
 797         progress ("Conversion introduced changes: %s" % diff)
 798         raise 'fdsa'
 799     else:
 800         progress ("Checkout matches pristine darcs tree.")
 801
 802 def main ():
 803     (options, args) = get_cli_options ()
 804
 805     darcs_repo = os.path.abspath (args[0])
 806     git_repo = os.path.abspath (options.target_git_repo)
 807
 808     if os.path.exists (git_repo):
 809         system ('rm -rf %(git_repo)s' % locals ())
 810
 811     system ('mkdir %(git_repo)s && cd %(git_repo)s && git --bare init' % locals ())
 812     system ('git --git-dir %(git_repo)s  repo-config core.logAllRefUpdates false' % locals ())
 813
 814     os.environ['DARCS_DONT_ESCAPE_8BIT'] = '1'
 815     os.environ['GIT_DIR'] = git_repo
 816
 817     quiet = ' --quiet'
 818     if options.verbose:
 819         quiet = ' '
 820
 821     gfi = os.popen ('git fast-import %s' % quiet, 'w')
 822
 823     patches = get_darcs_patches (darcs_repo)
 824     conv_repo = DarcsConversionRepo (options.basename + ".tmpdarcs", patches)
 825     conv_repo.start_at (-1)
 826
 827     for p in patches:
 828         parent_patch = None
 829         parent_number = -1
 830
 831         combinations = [(v, w) for v in pending_patches.values ()
 832                         for w in pending_patches.values ()]
 833         candidates = [common_ancestor (git_commits[c[0].number], git_commits[c[1].number]) for c in combinations]
 834         candidates = sorted ([(-a.darcs_patch.number, a) for a in candidates])
 835         for (depth, c) in candidates:
 836             q = c.darcs_patch
 837             try:
 838                 conv_repo.go_from_to (q, p)
 839
 840                 parent_patch = q
 841                 parent_number = q.number
 842                 progress ('Found existing common parent as predecessor')
 843                 break
 844
 845             except PullConflict:
 846                 pass
 847
 848         ## no branches found where we could attach.
 849         ## try previous commits one by one.
 850         if not parent_patch:
 851             parent_number = p.number - 2
 852             while 1:
 853                 if parent_number >= 0:
 854                     parent_patch = patches[parent_number]
 855
 856                 try:
 857                     conv_repo.go_from_to (parent_patch, p)
 858                     break
 859                 except PullConflict:
 860
 861                     ## simplistic, may not be enough.
 862                     progress ('conflict, going one back')
 863                     parent_number -= 1
 864
 865                     if parent_number < 0:
 866                         break
 867
 868                     if (options.history_window
 869                         and parent_number < p.number - options.history_window):
 870
 871                         parent_number = -2
 872                         break
 873
 874         if parent_number >= 0 or p.number == 0:
 875             progress ('Export %d -> %d (total %d)' % (parent_number,
 876                                                       p.number, len (patches)))
 877             export_commit (conv_repo, p, parent_patch, gfi)
 878             if p.tag_name ():
 879                 export_tag (p, gfi)
 880
 881             if options.checkpoint_frequency and p.number % options.checkpoint_frequency == 0:
 882                 export_checkpoint (gfi)
 883         else:
 884             progress ("Can't import patch %d, need conflict resolution patch?"
 885                       % p.number)
 886
 887     export_pending (gfi)
 888     gfi.close ()
 889     for f in glob.glob ('%(git_repo)s/refs/heads/darcstmp*' % locals ()):
 890         os.unlink (f)
 891
 892     test_conversion (darcs_repo, git_repo)
 893
 894     if not options.debug:
 895         conv_repo.clean ()
 896
 897 main ()