mdw@git.distorted.org.uk Git - stgit/blob - stgit/git.py

   1 """Python GIT interface
   2 """
   3
   4 __copyright__ = """
   5 Copyright (C) 2005, Catalin Marinas <catalin.marinas@gmail.com>
   6
   7 This program is free software; you can redistribute it and/or modify
   8 it under the terms of the GNU General Public License version 2 as
   9 published by the Free Software Foundation.
  10
  11 This program is distributed in the hope that it will be useful,
  12 but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 GNU General Public License for more details.
  15
  16 You should have received a copy of the GNU General Public License
  17 along with this program; if not, write to the Free Software
  18 Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
  19 """
  20
  21 import sys, os, re, gitmergeonefile
  22 from shutil import copyfile
  23
  24 from stgit import basedir
  25 from stgit.utils import *
  26 from stgit.out import *
  27 from stgit.run import *
  28 from stgit.config import config
  29 from sets import Set
  30
  31 # git exception class
  32 class GitException(Exception):
  33     pass
  34
  35 # When a subprocess has a problem, we want the exception to be a
  36 # subclass of GitException.
  37 class GitRunException(GitException):
  38     pass
  39 class GRun(Run):
  40     exc = GitRunException
  41
  42
  43 #
  44 # Classes
  45 #
  46
  47 class Person:
  48     """An author, committer, etc."""
  49     def __init__(self, name = None, email = None, date = '',
  50                  desc = None):
  51         self.name = self.email = self.date = None
  52         if name or email or date:
  53             assert not desc
  54             self.name = name
  55             self.email = email
  56             self.date = date
  57         elif desc:
  58             assert not (name or email or date)
  59             def parse_desc(s):
  60                 m = re.match(r'^(.+)<(.+)>(.*)$', s)
  61                 assert m
  62                 return [x.strip() or None for x in m.groups()]
  63             self.name, self.email, self.date = parse_desc(desc)
  64     def set_name(self, val):
  65         if val:
  66             self.name = val
  67     def set_email(self, val):
  68         if val:
  69             self.email = val
  70     def set_date(self, val):
  71         if val:
  72             self.date = val
  73     def __str__(self):
  74         if self.name and self.email:
  75             return '%s <%s>' % (self.name, self.email)
  76         else:
  77             raise GitException, 'not enough identity data'
  78
  79 class Commit:
  80     """Handle the commit objects
  81     """
  82     def __init__(self, id_hash):
  83         self.__id_hash = id_hash
  84
  85         lines = GRun('git-cat-file', 'commit', id_hash).output_lines()
  86         for i in range(len(lines)):
  87             line = lines[i]
  88             if not line:
  89                 break # we've seen all the header fields
  90             key, val = line.split(' ', 1)
  91             if key == 'tree':
  92                 self.__tree = val
  93             elif key == 'author':
  94                 self.__author = val
  95             elif key == 'committer':
  96                 self.__committer = val
  97             else:
  98                 pass # ignore other headers
  99         self.__log = '\n'.join(lines[i+1:])
 100
 101     def get_id_hash(self):
 102         return self.__id_hash
 103
 104     def get_tree(self):
 105         return self.__tree
 106
 107     def get_parent(self):
 108         parents = self.get_parents()
 109         if parents:
 110             return parents[0]
 111         else:
 112             return None
 113
 114     def get_parents(self):
 115         return GRun('git-rev-list', '--parents', '--max-count=1', self.__id_hash
 116                     ).output_one_line().split()[1:]
 117
 118     def get_author(self):
 119         return self.__author
 120
 121     def get_committer(self):
 122         return self.__committer
 123
 124     def get_log(self):
 125         return self.__log
 126
 127     def __str__(self):
 128         return self.get_id_hash()
 129
 130 # dictionary of Commit objects, used to avoid multiple calls to git
 131 __commits = dict()
 132
 133 #
 134 # Functions
 135 #
 136
 137 def get_commit(id_hash):
 138     """Commit objects factory. Save/look-up them in the __commits
 139     dictionary
 140     """
 141     global __commits
 142
 143     if id_hash in __commits:
 144         return __commits[id_hash]
 145     else:
 146         commit = Commit(id_hash)
 147         __commits[id_hash] = commit
 148         return commit
 149
 150 def get_conflicts():
 151     """Return the list of file conflicts
 152     """
 153     conflicts_file = os.path.join(basedir.get(), 'conflicts')
 154     if os.path.isfile(conflicts_file):
 155         f = file(conflicts_file)
 156         names = [line.strip() for line in f.readlines()]
 157         f.close()
 158         return names
 159     else:
 160         return None
 161
 162 def exclude_files():
 163     files = [os.path.join(basedir.get(), 'info', 'exclude')]
 164     user_exclude = config.get('core.excludesfile')
 165     if user_exclude:
 166         files.append(user_exclude)
 167     return files
 168
 169 def tree_status(files = None, tree_id = 'HEAD', unknown = False,
 170                   noexclude = True, verbose = False, diff_flags = []):
 171     """Returns a list of pairs - (status, filename)
 172     """
 173     if verbose:
 174         out.start('Checking for changes in the working directory')
 175
 176     refresh_index()
 177
 178     if not files:
 179         files = []
 180     cache_files = []
 181
 182     # unknown files
 183     if unknown:
 184         cmd = ['git-ls-files', '-z', '--others', '--directory',
 185                '--no-empty-directory']
 186         if not noexclude:
 187             cmd += ['--exclude=%s' % s for s in
 188                     ['*.[ao]', '*.pyc', '.*', '*~', '#*', 'TAGS', 'tags']]
 189             cmd += ['--exclude-per-directory=.gitignore']
 190             cmd += ['--exclude-from=%s' % fn
 191                     for fn in exclude_files()
 192                     if os.path.exists(fn)]
 193
 194         lines = GRun(*cmd).raw_output().split('\0')
 195         cache_files += [('?', line) for line in lines if line]
 196
 197     # conflicted files
 198     conflicts = get_conflicts()
 199     if not conflicts:
 200         conflicts = []
 201     cache_files += [('C', filename) for filename in conflicts]
 202
 203     # the rest
 204     for line in GRun('git-diff-index', *(diff_flags + [tree_id, '--'] + files)
 205                      ).output_lines():
 206         fs = tuple(line.rstrip().split(' ',4)[-1].split('\t',1))
 207         if fs[1] not in conflicts:
 208             cache_files.append(fs)
 209
 210     if verbose:
 211         out.done()
 212
 213     return cache_files
 214
 215 def local_changes(verbose = True):
 216     """Return true if there are local changes in the tree
 217     """
 218     return len(tree_status(verbose = verbose)) != 0
 219
 220 def get_heads():
 221     heads = []
 222     hr = re.compile(r'^[0-9a-f]{40} refs/heads/(.+)$')
 223     for line in GRun('git-show-ref', '--heads').output_lines():
 224         m = hr.match(line)
 225         heads.append(m.group(1))
 226     return heads
 227
 228 # HEAD value cached
 229 __head = None
 230
 231 def get_head():
 232     """Verifies the HEAD and returns the SHA1 id that represents it
 233     """
 234     global __head
 235
 236     if not __head:
 237         __head = rev_parse('HEAD')
 238     return __head
 239
 240 def get_head_file():
 241     """Returns the name of the file pointed to by the HEAD link
 242     """
 243     return strip_prefix('refs/heads/',
 244                         GRun('git-symbolic-ref', 'HEAD').output_one_line())
 245
 246 def set_head_file(ref):
 247     """Resets HEAD to point to a new ref
 248     """
 249     # head cache flushing is needed since we might have a different value
 250     # in the new head
 251     __clear_head_cache()
 252     try:
 253         GRun('git-symbolic-ref', 'HEAD', 'refs/heads/%s' % ref).run()
 254     except GitRunException:
 255         raise GitException, 'Could not set head to "%s"' % ref
 256
 257 def set_ref(ref, val):
 258     """Point ref at a new commit object."""
 259     try:
 260         GRun('git-update-ref', ref, val).run()
 261     except GitRunException:
 262         raise GitException, 'Could not update %s to "%s".' % (ref, val)
 263
 264 def set_branch(branch, val):
 265     set_ref('refs/heads/%s' % branch, val)
 266
 267 def __set_head(val):
 268     """Sets the HEAD value
 269     """
 270     global __head
 271
 272     if not __head or __head != val:
 273         set_ref('HEAD', val)
 274         __head = val
 275
 276     # only allow SHA1 hashes
 277     assert(len(__head) == 40)
 278
 279 def __clear_head_cache():
 280     """Sets the __head to None so that a re-read is forced
 281     """
 282     global __head
 283
 284     __head = None
 285
 286 def refresh_index():
 287     """Refresh index with stat() information from the working directory.
 288     """
 289     GRun('git-update-index', '-q', '--unmerged', '--refresh').run()
 290
 291 def rev_parse(git_id):
 292     """Parse the string and return a verified SHA1 id
 293     """
 294     try:
 295         return GRun('git-rev-parse', '--verify', git_id).output_one_line()
 296     except GitRunException:
 297         raise GitException, 'Unknown revision: %s' % git_id
 298
 299 def ref_exists(ref):
 300     try:
 301         rev_parse(ref)
 302         return True
 303     except GitException:
 304         return False
 305
 306 def branch_exists(branch):
 307     return ref_exists('refs/heads/%s' % branch)
 308
 309 def create_branch(new_branch, tree_id = None):
 310     """Create a new branch in the git repository
 311     """
 312     if branch_exists(new_branch):
 313         raise GitException, 'Branch "%s" already exists' % new_branch
 314
 315     current_head = get_head()
 316     set_head_file(new_branch)
 317     __set_head(current_head)
 318
 319     # a checkout isn't needed if new branch points to the current head
 320     if tree_id:
 321         switch(tree_id)
 322
 323     if os.path.isfile(os.path.join(basedir.get(), 'MERGE_HEAD')):
 324         os.remove(os.path.join(basedir.get(), 'MERGE_HEAD'))
 325
 326 def switch_branch(new_branch):
 327     """Switch to a git branch
 328     """
 329     global __head
 330
 331     if not branch_exists(new_branch):
 332         raise GitException, 'Branch "%s" does not exist' % new_branch
 333
 334     tree_id = rev_parse('refs/heads/%s^{commit}' % new_branch)
 335     if tree_id != get_head():
 336         refresh_index()
 337         try:
 338             GRun('git-read-tree', '-u', '-m', get_head(), tree_id).run()
 339         except GitRunException:
 340             raise GitException, 'git-read-tree failed (local changes maybe?)'
 341         __head = tree_id
 342     set_head_file(new_branch)
 343
 344     if os.path.isfile(os.path.join(basedir.get(), 'MERGE_HEAD')):
 345         os.remove(os.path.join(basedir.get(), 'MERGE_HEAD'))
 346
 347 def delete_ref(ref):
 348     if not ref_exists(ref):
 349         raise GitException, '%s does not exist' % ref
 350     sha1 = GRun('git-show-ref', '-s', ref).output_one_line()
 351     try:
 352         GRun('git-update-ref', '-d', ref, sha1).run()
 353     except GitRunException:
 354         raise GitException, 'Failed to delete ref %s' % ref
 355
 356 def delete_branch(name):
 357     delete_ref('refs/heads/%s' % name)
 358
 359 def rename_ref(from_ref, to_ref):
 360     if not ref_exists(from_ref):
 361         raise GitException, '"%s" does not exist' % from_ref
 362     if ref_exists(to_ref):
 363         raise GitException, '"%s" already exists' % to_ref
 364
 365     sha1 = GRun('git-show-ref', '-s', from_ref).output_one_line()
 366     try:
 367         GRun('git-update-ref', to_ref, sha1, '0'*40).run()
 368     except GitRunException:
 369         raise GitException, 'Failed to create new ref %s' % to_ref
 370     try:
 371         GRun('git-update-ref', '-d', from_ref, sha1).run()
 372     except GitRunException:
 373         raise GitException, 'Failed to delete ref %s' % from_ref
 374
 375 def rename_branch(from_name, to_name):
 376     """Rename a git branch."""
 377     rename_ref('refs/heads/%s' % from_name, 'refs/heads/%s' % to_name)
 378     if get_head_file() == from_name:
 379         set_head_file(to_name)
 380     reflog_dir = os.path.join(basedir.get(), 'logs', 'refs', 'heads')
 381     if os.path.exists(reflog_dir) \
 382            and os.path.exists(os.path.join(reflog_dir, from_name)):
 383         rename(reflog_dir, from_name, to_name)
 384
 385 def add(names):
 386     """Add the files or recursively add the directory contents
 387     """
 388     # generate the file list
 389     files = []
 390     for i in names:
 391         if not os.path.exists(i):
 392             raise GitException, 'Unknown file or directory: %s' % i
 393
 394         if os.path.isdir(i):
 395             # recursive search. We only add files
 396             for root, dirs, local_files in os.walk(i):
 397                 for name in [os.path.join(root, f) for f in local_files]:
 398                     if os.path.isfile(name):
 399                         files.append(os.path.normpath(name))
 400         elif os.path.isfile(i):
 401             files.append(os.path.normpath(i))
 402         else:
 403             raise GitException, '%s is not a file or directory' % i
 404
 405     if files:
 406         try:
 407             GRun('git-update-index', '--add', '--').xargs(files)
 408         except GitRunException:
 409             raise GitException, 'Unable to add file'
 410
 411 def __copy_single(source, target, target2=''):
 412     """Copy file or dir named 'source' to name target+target2"""
 413
 414     # "source" (file or dir) must match one or more git-controlled file
 415     realfiles = GRun('git-ls-files', source).output_lines()
 416     if len(realfiles) == 0:
 417         raise GitException, '"%s" matches no git-controled files' % source
 418
 419     if os.path.isdir(source):
 420         # physically copy the files, and record them to add them in one run
 421         newfiles = []
 422         re_string='^'+source+'/(.*)$'
 423         prefix_regexp = re.compile(re_string)
 424         for f in [f.strip() for f in realfiles]:
 425             m = prefix_regexp.match(f)
 426             if not m:
 427                 raise Exception, '"%s" does not match "%s"' % (f, re_string)
 428             newname = target+target2+'/'+m.group(1)
 429             if not os.path.exists(os.path.dirname(newname)):
 430                 os.makedirs(os.path.dirname(newname))
 431             copyfile(f, newname)
 432             newfiles.append(newname)
 433
 434         add(newfiles)
 435     else: # files, symlinks, ...
 436         newname = target+target2
 437         copyfile(source, newname)
 438         add([newname])
 439
 440
 441 def copy(filespecs, target):
 442     if os.path.isdir(target):
 443         # target is a directory: copy each entry on the command line,
 444         # with the same name, into the target
 445         target = target.rstrip('/')
 446
 447         # first, check that none of the children of the target
 448         # matching the command line aleady exist
 449         for filespec in filespecs:
 450             entry = target+ '/' + os.path.basename(filespec.rstrip('/'))
 451             if os.path.exists(entry):
 452                 raise GitException, 'Target "%s" already exists' % entry
 453
 454         for filespec in filespecs:
 455             filespec = filespec.rstrip('/')
 456             basename = '/' + os.path.basename(filespec)
 457             __copy_single(filespec, target, basename)
 458
 459     elif os.path.exists(target):
 460         raise GitException, 'Target "%s" exists but is not a directory' % target
 461     elif len(filespecs) != 1:
 462         raise GitException, 'Cannot copy more than one file to non-directory'
 463
 464     else:
 465         # at this point: len(filespecs)==1 and target does not exist
 466
 467         # check target directory
 468         targetdir = os.path.dirname(target)
 469         if targetdir != '' and not os.path.isdir(targetdir):
 470             raise GitException, 'Target directory "%s" does not exist' % targetdir
 471
 472         __copy_single(filespecs[0].rstrip('/'), target)
 473
 474
 475 def rm(files, force = False):
 476     """Remove a file from the repository
 477     """
 478     if not force:
 479         for f in files:
 480             if os.path.exists(f):
 481                 raise GitException, '%s exists. Remove it first' %f
 482         if files:
 483             GRun('git-update-index', '--remove', '--').xargs(files)
 484     else:
 485         if files:
 486             GRun('git-update-index', '--force-remove', '--').xargs(files)
 487
 488 # Persons caching
 489 __user = None
 490 __author = None
 491 __committer = None
 492
 493 def user():
 494     """Return the user information.
 495     """
 496     global __user
 497     if not __user:
 498         name=config.get('user.name')
 499         email=config.get('user.email')
 500         __user = Person(name, email)
 501     return __user;
 502
 503 def author():
 504     """Return the author information.
 505     """
 506     global __author
 507     if not __author:
 508         try:
 509             # the environment variables take priority over config
 510             try:
 511                 date = os.environ['GIT_AUTHOR_DATE']
 512             except KeyError:
 513                 date = ''
 514             __author = Person(os.environ['GIT_AUTHOR_NAME'],
 515                               os.environ['GIT_AUTHOR_EMAIL'],
 516                               date)
 517         except KeyError:
 518             __author = user()
 519     return __author
 520
 521 def committer():
 522     """Return the author information.
 523     """
 524     global __committer
 525     if not __committer:
 526         try:
 527             # the environment variables take priority over config
 528             try:
 529                 date = os.environ['GIT_COMMITTER_DATE']
 530             except KeyError:
 531                 date = ''
 532             __committer = Person(os.environ['GIT_COMMITTER_NAME'],
 533                                  os.environ['GIT_COMMITTER_EMAIL'],
 534                                  date)
 535         except KeyError:
 536             __committer = user()
 537     return __committer
 538
 539 def update_cache(files = None, force = False):
 540     """Update the cache information for the given files
 541     """
 542     if not files:
 543         files = []
 544
 545     cache_files = tree_status(files, verbose = False)
 546
 547     # everything is up-to-date
 548     if len(cache_files) == 0:
 549         return False
 550
 551     # check for unresolved conflicts
 552     if not force and [x for x in cache_files
 553                       if x[0] not in ['M', 'N', 'A', 'D']]:
 554         raise GitException, 'Updating cache failed: unresolved conflicts'
 555
 556     # update the cache
 557     add_files = [x[1] for x in cache_files if x[0] in ['N', 'A']]
 558     rm_files =  [x[1] for x in cache_files if x[0] in ['D']]
 559     m_files =   [x[1] for x in cache_files if x[0] in ['M']]
 560
 561     GRun('git-update-index', '--add', '--').xargs(add_files)
 562     GRun('git-update-index', '--force-remove', '--').xargs(rm_files)
 563     GRun('git-update-index', '--').xargs(m_files)
 564
 565     return True
 566
 567 def commit(message, files = None, parents = None, allowempty = False,
 568            cache_update = True, tree_id = None, set_head = False,
 569            author_name = None, author_email = None, author_date = None,
 570            committer_name = None, committer_email = None):
 571     """Commit the current tree to repository
 572     """
 573     if not files:
 574         files = []
 575     if not parents:
 576         parents = []
 577
 578     # Get the tree status
 579     if cache_update and parents != []:
 580         changes = update_cache(files)
 581         if not changes and not allowempty:
 582             raise GitException, 'No changes to commit'
 583
 584     # get the commit message
 585     if not message:
 586         message = '\n'
 587     elif message[-1:] != '\n':
 588         message += '\n'
 589
 590     # write the index to repository
 591     if tree_id == None:
 592         tree_id = GRun('git-write-tree').output_one_line()
 593         set_head = True
 594
 595     # the commit
 596     env = {}
 597     if author_name:
 598         env['GIT_AUTHOR_NAME'] = author_name
 599     if author_email:
 600         env['GIT_AUTHOR_EMAIL'] = author_email
 601     if author_date:
 602         env['GIT_AUTHOR_DATE'] = author_date
 603     if committer_name:
 604         env['GIT_COMMITTER_NAME'] = committer_name
 605     if committer_email:
 606         env['GIT_COMMITTER_EMAIL'] = committer_email
 607     commit_id = GRun('git-commit-tree', tree_id,
 608                      *sum([['-p', p] for p in parents], [])
 609                      ).env(env).raw_input(message).output_one_line()
 610     if set_head:
 611         __set_head(commit_id)
 612
 613     return commit_id
 614
 615 def apply_diff(rev1, rev2, check_index = True, files = None):
 616     """Apply the diff between rev1 and rev2 onto the current
 617     index. This function doesn't need to raise an exception since it
 618     is only used for fast-pushing a patch. If this operation fails,
 619     the pushing would fall back to the three-way merge.
 620     """
 621     if check_index:
 622         index_opt = ['--index']
 623     else:
 624         index_opt = []
 625
 626     if not files:
 627         files = []
 628
 629     diff_str = diff(files, rev1, rev2)
 630     if diff_str:
 631         try:
 632             GRun('git-apply', *index_opt).raw_input(diff_str).no_output()
 633         except GitRunException:
 634             return False
 635
 636     return True
 637
 638 def merge(base, head1, head2, recursive = False):
 639     """Perform a 3-way merge between base, head1 and head2 into the
 640     local tree
 641     """
 642     refresh_index()
 643
 644     err_output = None
 645     if recursive:
 646         # this operation tracks renames but it is slower (used in
 647         # general when pushing or picking patches)
 648         try:
 649             # discard output to mask the verbose prints of the tool
 650             GRun('git-merge-recursive', base, '--', head1, head2
 651                  ).discard_output()
 652         except GitRunException, ex:
 653             err_output = str(ex)
 654             pass
 655     else:
 656         # the fast case where we don't track renames (used when the
 657         # distance between base and heads is small, i.e. folding or
 658         # synchronising patches)
 659         try:
 660             GRun('git-read-tree', '-u', '-m', '--aggressive',
 661                  base, head1, head2).run()
 662         except GitRunException:
 663             raise GitException, 'git-read-tree failed (local changes maybe?)'
 664
 665     # check the index for unmerged entries
 666     files = {}
 667     stages_re = re.compile('^([0-7]+) ([0-9a-f]{40}) ([1-3])\t(.*)$', re.S)
 668
 669     for line in GRun('git-ls-files', '--unmerged', '--stage', '-z'
 670                      ).raw_output().split('\0'):
 671         if not line:
 672             continue
 673
 674         mode, hash, stage, path = stages_re.findall(line)[0]
 675
 676         if not path in files:
 677             files[path] = {}
 678             files[path]['1'] = ('', '')
 679             files[path]['2'] = ('', '')
 680             files[path]['3'] = ('', '')
 681
 682         files[path][stage] = (mode, hash)
 683
 684     if err_output and not files:
 685         # if no unmerged files, there was probably a different type of
 686         # error and we have to abort the merge
 687         raise GitException, err_output
 688
 689     # merge the unmerged files
 690     errors = False
 691     for path in files:
 692         # remove additional files that might be generated for some
 693         # newer versions of GIT
 694         for suffix in [base, head1, head2]:
 695             if not suffix:
 696                 continue
 697             fname = path + '~' + suffix
 698             if os.path.exists(fname):
 699                 os.remove(fname)
 700
 701         stages = files[path]
 702         if gitmergeonefile.merge(stages['1'][1], stages['2'][1],
 703                                  stages['3'][1], path, stages['1'][0],
 704                                  stages['2'][0], stages['3'][0]) != 0:
 705             errors = True
 706
 707     if errors:
 708         raise GitException, 'GIT index merging failed (possible conflicts)'
 709
 710 def status(files = None, modified = False, new = False, deleted = False,
 711            conflict = False, unknown = False, noexclude = False,
 712            diff_flags = []):
 713     """Show the tree status
 714     """
 715     if not files:
 716         files = []
 717
 718     cache_files = tree_status(files, unknown = True, noexclude = noexclude,
 719                                 diff_flags = diff_flags)
 720     all = not (modified or new or deleted or conflict or unknown)
 721
 722     if not all:
 723         filestat = []
 724         if modified:
 725             filestat.append('M')
 726         if new:
 727             filestat.append('A')
 728             filestat.append('N')
 729         if deleted:
 730             filestat.append('D')
 731         if conflict:
 732             filestat.append('C')
 733         if unknown:
 734             filestat.append('?')
 735         cache_files = [x for x in cache_files if x[0] in filestat]
 736
 737     for fs in cache_files:
 738         if files and not fs[1] in files:
 739             continue
 740         if all:
 741             out.stdout('%s %s' % (fs[0], fs[1]))
 742         else:
 743             out.stdout('%s' % fs[1])
 744
 745 def diff(files = None, rev1 = 'HEAD', rev2 = None, diff_flags = []):
 746     """Show the diff between rev1 and rev2
 747     """
 748     if not files:
 749         files = []
 750
 751     if rev1 and rev2:
 752         return GRun('git-diff-tree', '-p',
 753                     *(diff_flags + [rev1, rev2, '--'] + files)).raw_output()
 754     elif rev1 or rev2:
 755         refresh_index()
 756         if rev2:
 757             return GRun('git-diff-index', '-p', '-R',
 758                         *(diff_flags + [rev2, '--'] + files)).raw_output()
 759         else:
 760             return GRun('git-diff-index', '-p',
 761                         *(diff_flags + [rev1, '--'] + files)).raw_output()
 762     else:
 763         return ''
 764
 765 def diffstat(files = None, rev1 = 'HEAD', rev2 = None):
 766     """Return the diffstat between rev1 and rev2."""
 767     return GRun('git-apply', '--stat'
 768                 ).raw_input(diff(files, rev1, rev2)).raw_output()
 769
 770 def files(rev1, rev2, diff_flags = []):
 771     """Return the files modified between rev1 and rev2
 772     """
 773
 774     result = []
 775     for line in GRun('git-diff-tree', *(diff_flags + ['-r', rev1, rev2])
 776                      ).output_lines():
 777         result.append('%s %s' % tuple(line.split(' ', 4)[-1].split('\t', 1)))
 778
 779     return '\n'.join(result)
 780
 781 def barefiles(rev1, rev2):
 782     """Return the files modified between rev1 and rev2, without status info
 783     """
 784
 785     result = []
 786     for line in GRun('git-diff-tree', '-r', rev1, rev2).output_lines():
 787         result.append(line.split(' ', 4)[-1].split('\t', 1)[-1])
 788
 789     return '\n'.join(result)
 790
 791 def pretty_commit(commit_id = 'HEAD', diff_flags = []):
 792     """Return a given commit (log + diff)
 793     """
 794     return GRun('git-diff-tree',
 795                 *(diff_flags
 796                   + ['--cc', '--always', '--pretty', '-r', commit_id])
 797                 ).raw_output()
 798
 799 def checkout(files = None, tree_id = None, force = False):
 800     """Check out the given or all files
 801     """
 802     if tree_id:
 803         try:
 804             GRun('git-read-tree', '--reset', tree_id).run()
 805         except GitRunException:
 806             raise GitException, 'Failed git-read-tree --reset %s' % tree_id
 807
 808     cmd = ['git-checkout-index', '-q', '-u']
 809     if force:
 810         cmd.append('-f')
 811     if files:
 812         GRun(*(cmd + ['--'])).xargs(files)
 813     else:
 814         GRun(*(cmd + ['-a'])).run()
 815
 816 def switch(tree_id, keep = False):
 817     """Switch the tree to the given id
 818     """
 819     if not keep:
 820         refresh_index()
 821         try:
 822             GRun('git-read-tree', '-u', '-m', get_head(), tree_id).run()
 823         except GitRunException:
 824             raise GitException, 'git-read-tree failed (local changes maybe?)'
 825
 826     __set_head(tree_id)
 827
 828 def reset(files = None, tree_id = None, check_out = True):
 829     """Revert the tree changes relative to the given tree_id. It removes
 830     any local changes
 831     """
 832     if not tree_id:
 833         tree_id = get_head()
 834
 835     if check_out:
 836         cache_files = tree_status(files, tree_id)
 837         # files which were added but need to be removed
 838         rm_files =  [x[1] for x in cache_files if x[0] in ['A']]
 839
 840         checkout(files, tree_id, True)
 841         # checkout doesn't remove files
 842         map(os.remove, rm_files)
 843
 844     # if the reset refers to the whole tree, switch the HEAD as well
 845     if not files:
 846         __set_head(tree_id)
 847
 848 def fetch(repository = 'origin', refspec = None):
 849     """Fetches changes from the remote repository, using 'git-fetch'
 850     by default.
 851     """
 852     # we update the HEAD
 853     __clear_head_cache()
 854
 855     args = [repository]
 856     if refspec:
 857         args.append(refspec)
 858
 859     command = config.get('branch.%s.stgit.fetchcmd' % get_head_file()) or \
 860               config.get('stgit.fetchcmd')
 861     GRun(*(command.split() + args)).run()
 862
 863 def pull(repository = 'origin', refspec = None):
 864     """Fetches changes from the remote repository, using 'git-pull'
 865     by default.
 866     """
 867     # we update the HEAD
 868     __clear_head_cache()
 869
 870     args = [repository]
 871     if refspec:
 872         args.append(refspec)
 873
 874     command = config.get('branch.%s.stgit.pullcmd' % get_head_file()) or \
 875               config.get('stgit.pullcmd')
 876     GRun(*(command.split() + args)).run()
 877
 878 def repack():
 879     """Repack all objects into a single pack
 880     """
 881     GRun('git-repack', '-a', '-d', '-f').run()
 882
 883 def apply_patch(filename = None, diff = None, base = None,
 884                 fail_dump = True):
 885     """Apply a patch onto the current or given index. There must not
 886     be any local changes in the tree, otherwise the command fails
 887     """
 888     if diff is None:
 889         if filename:
 890             f = file(filename)
 891         else:
 892             f = sys.stdin
 893         diff = f.read()
 894         if filename:
 895             f.close()
 896
 897     if base:
 898         orig_head = get_head()
 899         switch(base)
 900     else:
 901         refresh_index()
 902
 903     try:
 904         GRun('git-apply', '--index').raw_input(diff).no_output()
 905     except GitRunException:
 906         if base:
 907             switch(orig_head)
 908         if fail_dump:
 909             # write the failed diff to a file
 910             f = file('.stgit-failed.patch', 'w+')
 911             f.write(diff)
 912             f.close()
 913             out.warn('Diff written to the .stgit-failed.patch file')
 914
 915         raise
 916
 917     if base:
 918         top = commit(message = 'temporary commit used for applying a patch',
 919                      parents = [base])
 920         switch(orig_head)
 921         merge(base, orig_head, top)
 922
 923 def clone(repository, local_dir):
 924     """Clone a remote repository. At the moment, just use the
 925     'git-clone' script
 926     """
 927     GRun('git-clone', repository, local_dir).run()
 928
 929 def modifying_revs(files, base_rev, head_rev):
 930     """Return the revisions from the list modifying the given files."""
 931     return GRun('git-rev-list', '%s..%s' % (base_rev, head_rev), '--', *files
 932                 ).output_lines()
 933
 934 def refspec_localpart(refspec):
 935     m = re.match('^[^:]*:([^:]*)$', refspec)
 936     if m:
 937         return m.group(1)
 938     else:
 939         raise GitException, 'Cannot parse refspec "%s"' % line
 940
 941 def refspec_remotepart(refspec):
 942     m = re.match('^([^:]*):[^:]*$', refspec)
 943     if m:
 944         return m.group(1)
 945     else:
 946         raise GitException, 'Cannot parse refspec "%s"' % line
 947
 948
 949 def __remotes_from_config():
 950     return config.sections_matching(r'remote\.(.*)\.url')
 951
 952 def __remotes_from_dir(dir):
 953     d = os.path.join(basedir.get(), dir)
 954     if os.path.exists(d):
 955         return os.listdir(d)
 956     else:
 957         return None
 958
 959 def remotes_list():
 960     """Return the list of remotes in the repository
 961     """
 962
 963     return Set(__remotes_from_config()) | \
 964            Set(__remotes_from_dir('remotes')) | \
 965            Set(__remotes_from_dir('branches'))
 966
 967 def remotes_local_branches(remote):
 968     """Returns the list of local branches fetched from given remote
 969     """
 970
 971     branches = []
 972     if remote in __remotes_from_config():
 973         for line in config.getall('remote.%s.fetch' % remote):
 974             branches.append(refspec_localpart(line))
 975     elif remote in __remotes_from_dir('remotes'):
 976         stream = open(os.path.join(basedir.get(), 'remotes', remote), 'r')
 977         for line in stream:
 978             # Only consider Pull lines
 979             m = re.match('^Pull: (.*)\n$', line)
 980             if m:
 981                 branches.append(refspec_localpart(m.group(1)))
 982         stream.close()
 983     elif remote in __remotes_from_dir('branches'):
 984         # old-style branches only declare one branch
 985         branches.append('refs/heads/'+remote);
 986     else:
 987         raise GitException, 'Unknown remote "%s"' % remote
 988
 989     return branches
 990
 991 def identify_remote(branchname):
 992     """Return the name for the remote to pull the given branchname
 993     from, or None if we believe it is a local branch.
 994     """
 995
 996     for remote in remotes_list():
 997         if branchname in remotes_local_branches(remote):
 998             return remote
 999
1000     # if we get here we've found nothing, the branch is a local one
1001     return None
1002
1003 def fetch_head():
1004     """Return the git id for the tip of the parent branch as left by
1005     'git fetch'.
1006     """
1007
1008     fetch_head=None
1009     stream = open(os.path.join(basedir.get(), 'FETCH_HEAD'), "r")
1010     for line in stream:
1011         # Only consider lines not tagged not-for-merge
1012         m = re.match('^([^\t]*)\t\t', line)
1013         if m:
1014             if fetch_head:
1015                 raise GitException, "StGit does not support multiple FETCH_HEAD"
1016             else:
1017                 fetch_head=m.group(1)
1018     stream.close()
1019
1020     # here we are sure to have a single fetch_head
1021     return fetch_head
1022
1023 def all_refs():
1024     """Return a list of all refs in the current repository.
1025     """
1026
1027     return [line.split()[1] for line in GRun('git-show-ref').output_lines()]