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.exception import *
  25 from stgit import basedir
  26 from stgit.utils import *
  27 from stgit.out import *
  28 from stgit.run import *
  29 from stgit.config import config
  30
  31 # git exception class
  32 class GitException(StgException):
  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     def __init__(self, *cmd):
  42         """Initialise the Run object and insert the 'git' command name.
  43         """
  44         Run.__init__(self, 'git', *cmd)
  45
  46
  47 #
  48 # Classes
  49 #
  50
  51 class Person:
  52     """An author, committer, etc."""
  53     def __init__(self, name = None, email = None, date = '',
  54                  desc = None):
  55         self.name = self.email = self.date = None
  56         if name or email or date:
  57             assert not desc
  58             self.name = name
  59             self.email = email
  60             self.date = date
  61         elif desc:
  62             assert not (name or email or date)
  63             def parse_desc(s):
  64                 m = re.match(r'^(.+)<(.+)>(.*)$', s)
  65                 assert m
  66                 return [x.strip() or None for x in m.groups()]
  67             self.name, self.email, self.date = parse_desc(desc)
  68     def set_name(self, val):
  69         if val:
  70             self.name = val
  71     def set_email(self, val):
  72         if val:
  73             self.email = val
  74     def set_date(self, val):
  75         if val:
  76             self.date = val
  77     def __str__(self):
  78         if self.name and self.email:
  79             return '%s <%s>' % (self.name, self.email)
  80         else:
  81             raise GitException, 'not enough identity data'
  82
  83 class Commit:
  84     """Handle the commit objects
  85     """
  86     def __init__(self, id_hash):
  87         self.__id_hash = id_hash
  88
  89         lines = GRun('cat-file', 'commit', id_hash).output_lines()
  90         for i in range(len(lines)):
  91             line = lines[i]
  92             if not line:
  93                 break # we've seen all the header fields
  94             key, val = line.split(' ', 1)
  95             if key == 'tree':
  96                 self.__tree = val
  97             elif key == 'author':
  98                 self.__author = val
  99             elif key == 'committer':
 100                 self.__committer = val
 101             else:
 102                 pass # ignore other headers
 103         self.__log = '\n'.join(lines[i+1:])
 104
 105     def get_id_hash(self):
 106         return self.__id_hash
 107
 108     def get_tree(self):
 109         return self.__tree
 110
 111     def get_parent(self):
 112         parents = self.get_parents()
 113         if parents:
 114             return parents[0]
 115         else:
 116             return None
 117
 118     def get_parents(self):
 119         return GRun('rev-list', '--parents', '--max-count=1', self.__id_hash
 120                     ).output_one_line().split()[1:]
 121
 122     def get_author(self):
 123         return self.__author
 124
 125     def get_committer(self):
 126         return self.__committer
 127
 128     def get_log(self):
 129         return self.__log
 130
 131     def __str__(self):
 132         return self.get_id_hash()
 133
 134 # dictionary of Commit objects, used to avoid multiple calls to git
 135 __commits = dict()
 136
 137 #
 138 # Functions
 139 #
 140
 141 def get_commit(id_hash):
 142     """Commit objects factory. Save/look-up them in the __commits
 143     dictionary
 144     """
 145     global __commits
 146
 147     if id_hash in __commits:
 148         return __commits[id_hash]
 149     else:
 150         commit = Commit(id_hash)
 151         __commits[id_hash] = commit
 152         return commit
 153
 154 def get_conflicts():
 155     """Return the list of file conflicts
 156     """
 157     conflicts_file = os.path.join(basedir.get(), 'conflicts')
 158     if os.path.isfile(conflicts_file):
 159         f = file(conflicts_file)
 160         names = [line.strip() for line in f.readlines()]
 161         f.close()
 162         return names
 163     else:
 164         return None
 165
 166 def exclude_files():
 167     files = [os.path.join(basedir.get(), 'info', 'exclude')]
 168     user_exclude = config.get('core.excludesfile')
 169     if user_exclude:
 170         files.append(user_exclude)
 171     return files
 172
 173 def ls_files(files, tree = None, full_name = True):
 174     """Return the files known to GIT or raise an error otherwise. It also
 175     converts the file to the full path relative the the .git directory.
 176     """
 177     if not files:
 178         return []
 179
 180     args = []
 181     if tree:
 182         args.append('--with-tree=%s' % tree)
 183     if full_name:
 184         args.append('--full-name')
 185     args.append('--')
 186     args.extend(files)
 187     try:
 188         return GRun('ls-files', '--error-unmatch', *args).output_lines()
 189     except GitRunException:
 190         # just hide the details of the 'git ls-files' command we use
 191         raise GitException, \
 192             'Some of the given paths are either missing or not known to GIT'
 193
 194 def tree_status(files = None, tree_id = 'HEAD', unknown = False,
 195                   noexclude = True, verbose = False, diff_flags = []):
 196     """Get the status of all changed files, or of a selected set of
 197     files. Returns a list of pairs - (status, filename).
 198
 199     If 'not files', it will check all files, and optionally all
 200     unknown files.  If 'files' is a list, it will only check the files
 201     in the list.
 202     """
 203     assert not files or not unknown
 204
 205     if verbose:
 206         out.start('Checking for changes in the working directory')
 207
 208     refresh_index()
 209
 210     cache_files = []
 211
 212     # unknown files
 213     if unknown:
 214         cmd = ['ls-files', '-z', '--others', '--directory',
 215                '--no-empty-directory']
 216         if not noexclude:
 217             cmd += ['--exclude=%s' % s for s in
 218                     ['*.[ao]', '*.pyc', '.*', '*~', '#*', 'TAGS', 'tags']]
 219             cmd += ['--exclude-per-directory=.gitignore']
 220             cmd += ['--exclude-from=%s' % fn
 221                     for fn in exclude_files()
 222                     if os.path.exists(fn)]
 223
 224         lines = GRun(*cmd).raw_output().split('\0')
 225         cache_files += [('?', line) for line in lines if line]
 226
 227     # conflicted files
 228     conflicts = get_conflicts()
 229     if not conflicts:
 230         conflicts = []
 231     cache_files += [('C', filename) for filename in conflicts
 232                     if not files or filename in files]
 233     reported_files = set(conflicts)
 234
 235     # files in the index
 236     args = diff_flags + [tree_id]
 237     if files:
 238         args += ['--'] + files
 239     for line in GRun('diff-index', *args).output_lines():
 240         fs = tuple(line.rstrip().split(' ',4)[-1].split('\t',1))
 241         if fs[1] not in reported_files:
 242             cache_files.append(fs)
 243             reported_files.add(fs[1])
 244
 245     # files in the index but changed on (or removed from) disk
 246     args = list(diff_flags)
 247     if files:
 248         args += ['--'] + files
 249     for line in GRun('diff-files', *args).output_lines():
 250         fs = tuple(line.rstrip().split(' ',4)[-1].split('\t',1))
 251         if fs[1] not in reported_files:
 252             cache_files.append(fs)
 253             reported_files.add(fs[1])
 254
 255     if verbose:
 256         out.done()
 257
 258     return cache_files
 259
 260 def local_changes(verbose = True):
 261     """Return true if there are local changes in the tree
 262     """
 263     return len(tree_status(verbose = verbose)) != 0
 264
 265 def get_heads():
 266     heads = []
 267     hr = re.compile(r'^[0-9a-f]{40} refs/heads/(.+)$')
 268     for line in GRun('show-ref', '--heads').output_lines():
 269         m = hr.match(line)
 270         heads.append(m.group(1))
 271     return heads
 272
 273 # HEAD value cached
 274 __head = None
 275
 276 def get_head():
 277     """Verifies the HEAD and returns the SHA1 id that represents it
 278     """
 279     global __head
 280
 281     if not __head:
 282         __head = rev_parse('HEAD')
 283     return __head
 284
 285 class DetachedHeadException(GitException):
 286     def __init__(self):
 287         GitException.__init__(self, 'Not on any branch')
 288
 289 def get_head_file():
 290     """Return the name of the file pointed to by the HEAD symref.
 291     Throw an exception if HEAD is detached."""
 292     try:
 293         return strip_prefix(
 294             'refs/heads/', GRun('symbolic-ref', '-q', 'HEAD'
 295                                 ).output_one_line())
 296     except GitRunException:
 297         raise DetachedHeadException()
 298
 299 def set_head_file(ref):
 300     """Resets HEAD to point to a new ref
 301     """
 302     # head cache flushing is needed since we might have a different value
 303     # in the new head
 304     __clear_head_cache()
 305     try:
 306         GRun('symbolic-ref', 'HEAD', 'refs/heads/%s' % ref).run()
 307     except GitRunException:
 308         raise GitException, 'Could not set head to "%s"' % ref
 309
 310 def set_ref(ref, val):
 311     """Point ref at a new commit object."""
 312     try:
 313         GRun('update-ref', ref, val).run()
 314     except GitRunException:
 315         raise GitException, 'Could not update %s to "%s".' % (ref, val)
 316
 317 def set_branch(branch, val):
 318     set_ref('refs/heads/%s' % branch, val)
 319
 320 def __set_head(val):
 321     """Sets the HEAD value
 322     """
 323     global __head
 324
 325     if not __head or __head != val:
 326         set_ref('HEAD', val)
 327         __head = val
 328
 329     # only allow SHA1 hashes
 330     assert(len(__head) == 40)
 331
 332 def __clear_head_cache():
 333     """Sets the __head to None so that a re-read is forced
 334     """
 335     global __head
 336
 337     __head = None
 338
 339 def refresh_index():
 340     """Refresh index with stat() information from the working directory.
 341     """
 342     GRun('update-index', '-q', '--unmerged', '--refresh').run()
 343
 344 def rev_parse(git_id):
 345     """Parse the string and return a verified SHA1 id
 346     """
 347     try:
 348         return GRun('rev-parse', '--verify', git_id
 349                     ).discard_stderr().output_one_line()
 350     except GitRunException:
 351         raise GitException, 'Unknown revision: %s' % git_id
 352
 353 def ref_exists(ref):
 354     try:
 355         rev_parse(ref)
 356         return True
 357     except GitException:
 358         return False
 359
 360 def branch_exists(branch):
 361     return ref_exists('refs/heads/%s' % branch)
 362
 363 def create_branch(new_branch, tree_id = None):
 364     """Create a new branch in the git repository
 365     """
 366     if branch_exists(new_branch):
 367         raise GitException, 'Branch "%s" already exists' % new_branch
 368
 369     current_head_file = get_head_file()
 370     current_head = get_head()
 371     set_head_file(new_branch)
 372     __set_head(current_head)
 373
 374     # a checkout isn't needed if new branch points to the current head
 375     if tree_id:
 376         try:
 377             switch(tree_id)
 378         except GitException:
 379             # Tree switching failed. Revert the head file
 380             set_head_file(current_head_file)
 381             delete_branch(new_branch)
 382             raise
 383
 384     if os.path.isfile(os.path.join(basedir.get(), 'MERGE_HEAD')):
 385         os.remove(os.path.join(basedir.get(), 'MERGE_HEAD'))
 386
 387 def switch_branch(new_branch):
 388     """Switch to a git branch
 389     """
 390     global __head
 391
 392     if not branch_exists(new_branch):
 393         raise GitException, 'Branch "%s" does not exist' % new_branch
 394
 395     tree_id = rev_parse('refs/heads/%s^{commit}' % new_branch)
 396     if tree_id != get_head():
 397         refresh_index()
 398         try:
 399             GRun('read-tree', '-u', '-m', get_head(), tree_id).run()
 400         except GitRunException:
 401             raise GitException, 'read-tree failed (local changes maybe?)'
 402         __head = tree_id
 403     set_head_file(new_branch)
 404
 405     if os.path.isfile(os.path.join(basedir.get(), 'MERGE_HEAD')):
 406         os.remove(os.path.join(basedir.get(), 'MERGE_HEAD'))
 407
 408 def delete_ref(ref):
 409     if not ref_exists(ref):
 410         raise GitException, '%s does not exist' % ref
 411     sha1 = GRun('show-ref', '-s', ref).output_one_line()
 412     try:
 413         GRun('update-ref', '-d', ref, sha1).run()
 414     except GitRunException:
 415         raise GitException, 'Failed to delete ref %s' % ref
 416
 417 def delete_branch(name):
 418     delete_ref('refs/heads/%s' % name)
 419
 420 def rename_ref(from_ref, to_ref):
 421     if not ref_exists(from_ref):
 422         raise GitException, '"%s" does not exist' % from_ref
 423     if ref_exists(to_ref):
 424         raise GitException, '"%s" already exists' % to_ref
 425
 426     sha1 = GRun('show-ref', '-s', from_ref).output_one_line()
 427     try:
 428         GRun('update-ref', to_ref, sha1, '0'*40).run()
 429     except GitRunException:
 430         raise GitException, 'Failed to create new ref %s' % to_ref
 431     try:
 432         GRun('update-ref', '-d', from_ref, sha1).run()
 433     except GitRunException:
 434         raise GitException, 'Failed to delete ref %s' % from_ref
 435
 436 def rename_branch(from_name, to_name):
 437     """Rename a git branch."""
 438     rename_ref('refs/heads/%s' % from_name, 'refs/heads/%s' % to_name)
 439     try:
 440         if get_head_file() == from_name:
 441             set_head_file(to_name)
 442     except DetachedHeadException:
 443         pass # detached HEAD, so the renamee can't be the current branch
 444     reflog_dir = os.path.join(basedir.get(), 'logs', 'refs', 'heads')
 445     if os.path.exists(reflog_dir) \
 446            and os.path.exists(os.path.join(reflog_dir, from_name)):
 447         rename(reflog_dir, from_name, to_name)
 448
 449 def add(names):
 450     """Add the files or recursively add the directory contents
 451     """
 452     # generate the file list
 453     files = []
 454     for i in names:
 455         if not os.path.exists(i):
 456             raise GitException, 'Unknown file or directory: %s' % i
 457
 458         if os.path.isdir(i):
 459             # recursive search. We only add files
 460             for root, dirs, local_files in os.walk(i):
 461                 for name in [os.path.join(root, f) for f in local_files]:
 462                     if os.path.isfile(name):
 463                         files.append(os.path.normpath(name))
 464         elif os.path.isfile(i):
 465             files.append(os.path.normpath(i))
 466         else:
 467             raise GitException, '%s is not a file or directory' % i
 468
 469     if files:
 470         try:
 471             GRun('update-index', '--add', '--').xargs(files)
 472         except GitRunException:
 473             raise GitException, 'Unable to add file'
 474
 475 def __copy_single(source, target, target2=''):
 476     """Copy file or dir named 'source' to name target+target2"""
 477
 478     # "source" (file or dir) must match one or more git-controlled file
 479     realfiles = GRun('ls-files', source).output_lines()
 480     if len(realfiles) == 0:
 481         raise GitException, '"%s" matches no git-controled files' % source
 482
 483     if os.path.isdir(source):
 484         # physically copy the files, and record them to add them in one run
 485         newfiles = []
 486         re_string='^'+source+'/(.*)$'
 487         prefix_regexp = re.compile(re_string)
 488         for f in [f.strip() for f in realfiles]:
 489             m = prefix_regexp.match(f)
 490             if not m:
 491                 raise Exception, '"%s" does not match "%s"' % (f, re_string)
 492             newname = target+target2+'/'+m.group(1)
 493             if not os.path.exists(os.path.dirname(newname)):
 494                 os.makedirs(os.path.dirname(newname))
 495             copyfile(f, newname)
 496             newfiles.append(newname)
 497
 498         add(newfiles)
 499     else: # files, symlinks, ...
 500         newname = target+target2
 501         copyfile(source, newname)
 502         add([newname])
 503
 504
 505 def copy(filespecs, target):
 506     if os.path.isdir(target):
 507         # target is a directory: copy each entry on the command line,
 508         # with the same name, into the target
 509         target = target.rstrip('/')
 510
 511         # first, check that none of the children of the target
 512         # matching the command line aleady exist
 513         for filespec in filespecs:
 514             entry = target+ '/' + os.path.basename(filespec.rstrip('/'))
 515             if os.path.exists(entry):
 516                 raise GitException, 'Target "%s" already exists' % entry
 517
 518         for filespec in filespecs:
 519             filespec = filespec.rstrip('/')
 520             basename = '/' + os.path.basename(filespec)
 521             __copy_single(filespec, target, basename)
 522
 523     elif os.path.exists(target):
 524         raise GitException, 'Target "%s" exists but is not a directory' % target
 525     elif len(filespecs) != 1:
 526         raise GitException, 'Cannot copy more than one file to non-directory'
 527
 528     else:
 529         # at this point: len(filespecs)==1 and target does not exist
 530
 531         # check target directory
 532         targetdir = os.path.dirname(target)
 533         if targetdir != '' and not os.path.isdir(targetdir):
 534             raise GitException, 'Target directory "%s" does not exist' % targetdir
 535
 536         __copy_single(filespecs[0].rstrip('/'), target)
 537
 538
 539 def rm(files, force = False):
 540     """Remove a file from the repository
 541     """
 542     if not force:
 543         for f in files:
 544             if os.path.exists(f):
 545                 raise GitException, '%s exists. Remove it first' %f
 546         if files:
 547             GRun('update-index', '--remove', '--').xargs(files)
 548     else:
 549         if files:
 550             GRun('update-index', '--force-remove', '--').xargs(files)
 551
 552 # Persons caching
 553 __user = None
 554 __author = None
 555 __committer = None
 556
 557 def user():
 558     """Return the user information.
 559     """
 560     global __user
 561     if not __user:
 562         name=config.get('user.name')
 563         email=config.get('user.email')
 564         __user = Person(name, email)
 565     return __user;
 566
 567 def author():
 568     """Return the author information.
 569     """
 570     global __author
 571     if not __author:
 572         try:
 573             # the environment variables take priority over config
 574             try:
 575                 date = os.environ['GIT_AUTHOR_DATE']
 576             except KeyError:
 577                 date = ''
 578             __author = Person(os.environ['GIT_AUTHOR_NAME'],
 579                               os.environ['GIT_AUTHOR_EMAIL'],
 580                               date)
 581         except KeyError:
 582             __author = user()
 583     return __author
 584
 585 def committer():
 586     """Return the author information.
 587     """
 588     global __committer
 589     if not __committer:
 590         try:
 591             # the environment variables take priority over config
 592             try:
 593                 date = os.environ['GIT_COMMITTER_DATE']
 594             except KeyError:
 595                 date = ''
 596             __committer = Person(os.environ['GIT_COMMITTER_NAME'],
 597                                  os.environ['GIT_COMMITTER_EMAIL'],
 598                                  date)
 599         except KeyError:
 600             __committer = user()
 601     return __committer
 602
 603 def update_cache(files = None, force = False):
 604     """Update the cache information for the given files
 605     """
 606     cache_files = tree_status(files, verbose = False)
 607
 608     # everything is up-to-date
 609     if len(cache_files) == 0:
 610         return False
 611
 612     # check for unresolved conflicts
 613     if not force and [x for x in cache_files
 614                       if x[0] not in ['M', 'N', 'A', 'D']]:
 615         raise GitException, 'Updating cache failed: unresolved conflicts'
 616
 617     # update the cache
 618     add_files = [x[1] for x in cache_files if x[0] in ['N', 'A']]
 619     rm_files =  [x[1] for x in cache_files if x[0] in ['D']]
 620     m_files =   [x[1] for x in cache_files if x[0] in ['M']]
 621
 622     GRun('update-index', '--add', '--').xargs(add_files)
 623     GRun('update-index', '--force-remove', '--').xargs(rm_files)
 624     GRun('update-index', '--').xargs(m_files)
 625
 626     return True
 627
 628 def commit(message, files = None, parents = None, allowempty = False,
 629            cache_update = True, tree_id = None, set_head = False,
 630            author_name = None, author_email = None, author_date = None,
 631            committer_name = None, committer_email = None):
 632     """Commit the current tree to repository
 633     """
 634     if not parents:
 635         parents = []
 636
 637     # Get the tree status
 638     if cache_update and parents != []:
 639         changes = update_cache(files)
 640         if not changes and not allowempty:
 641             raise GitException, 'No changes to commit'
 642
 643     # get the commit message
 644     if not message:
 645         message = '\n'
 646     elif message[-1:] != '\n':
 647         message += '\n'
 648
 649     # write the index to repository
 650     if tree_id == None:
 651         tree_id = GRun('write-tree').output_one_line()
 652         set_head = True
 653
 654     # the commit
 655     env = {}
 656     if author_name:
 657         env['GIT_AUTHOR_NAME'] = author_name
 658     if author_email:
 659         env['GIT_AUTHOR_EMAIL'] = author_email
 660     if author_date:
 661         env['GIT_AUTHOR_DATE'] = author_date
 662     if committer_name:
 663         env['GIT_COMMITTER_NAME'] = committer_name
 664     if committer_email:
 665         env['GIT_COMMITTER_EMAIL'] = committer_email
 666     commit_id = GRun('commit-tree', tree_id,
 667                      *sum([['-p', p] for p in parents], [])
 668                      ).env(env).raw_input(message).output_one_line()
 669     if set_head:
 670         __set_head(commit_id)
 671
 672     return commit_id
 673
 674 def apply_diff(rev1, rev2, check_index = True, files = None):
 675     """Apply the diff between rev1 and rev2 onto the current
 676     index. This function doesn't need to raise an exception since it
 677     is only used for fast-pushing a patch. If this operation fails,
 678     the pushing would fall back to the three-way merge.
 679     """
 680     if check_index:
 681         index_opt = ['--index']
 682     else:
 683         index_opt = []
 684
 685     if not files:
 686         files = []
 687
 688     diff_str = diff(files, rev1, rev2)
 689     if diff_str:
 690         try:
 691             GRun('apply', *index_opt).raw_input(
 692                 diff_str).discard_stderr().no_output()
 693         except GitRunException:
 694             return False
 695
 696     return True
 697
 698 def merge(base, head1, head2, recursive = False):
 699     """Perform a 3-way merge between base, head1 and head2 into the
 700     local tree
 701     """
 702     refresh_index()
 703
 704     err_output = None
 705     if recursive:
 706         # this operation tracks renames but it is slower (used in
 707         # general when pushing or picking patches)
 708         try:
 709             # discard output to mask the verbose prints of the tool
 710             GRun('merge-recursive', base, '--', head1, head2
 711                  ).discard_output()
 712         except GitRunException, ex:
 713             err_output = str(ex)
 714             pass
 715     else:
 716         # the fast case where we don't track renames (used when the
 717         # distance between base and heads is small, i.e. folding or
 718         # synchronising patches)
 719         try:
 720             GRun('read-tree', '-u', '-m', '--aggressive',
 721                  base, head1, head2).run()
 722         except GitRunException:
 723             raise GitException, 'read-tree failed (local changes maybe?)'
 724
 725     # check the index for unmerged entries
 726     files = {}
 727     stages_re = re.compile('^([0-7]+) ([0-9a-f]{40}) ([1-3])\t(.*)$', re.S)
 728
 729     for line in GRun('ls-files', '--unmerged', '--stage', '-z'
 730                      ).raw_output().split('\0'):
 731         if not line:
 732             continue
 733
 734         mode, hash, stage, path = stages_re.findall(line)[0]
 735
 736         if not path in files:
 737             files[path] = {}
 738             files[path]['1'] = ('', '')
 739             files[path]['2'] = ('', '')
 740             files[path]['3'] = ('', '')
 741
 742         files[path][stage] = (mode, hash)
 743
 744     if err_output and not files:
 745         # if no unmerged files, there was probably a different type of
 746         # error and we have to abort the merge
 747         raise GitException, err_output
 748
 749     # merge the unmerged files
 750     errors = False
 751     for path in files:
 752         # remove additional files that might be generated for some
 753         # newer versions of GIT
 754         for suffix in [base, head1, head2]:
 755             if not suffix:
 756                 continue
 757             fname = path + '~' + suffix
 758             if os.path.exists(fname):
 759                 os.remove(fname)
 760
 761         stages = files[path]
 762         if gitmergeonefile.merge(stages['1'][1], stages['2'][1],
 763                                  stages['3'][1], path, stages['1'][0],
 764                                  stages['2'][0], stages['3'][0]) != 0:
 765             errors = True
 766
 767     if errors:
 768         raise GitException, 'GIT index merging failed (possible conflicts)'
 769
 770 def diff(files = None, rev1 = 'HEAD', rev2 = None, diff_flags = [],
 771          binary = True):
 772     """Show the diff between rev1 and rev2
 773     """
 774     if not files:
 775         files = []
 776     if binary and '--binary' not in diff_flags:
 777         diff_flags = diff_flags + ['--binary']
 778
 779     if rev1 and rev2:
 780         return GRun('diff-tree', '-p',
 781                     *(diff_flags + [rev1, rev2, '--'] + files)).raw_output()
 782     elif rev1 or rev2:
 783         refresh_index()
 784         if rev2:
 785             return GRun('diff-index', '-p', '-R',
 786                         *(diff_flags + [rev2, '--'] + files)).raw_output()
 787         else:
 788             return GRun('diff-index', '-p',
 789                         *(diff_flags + [rev1, '--'] + files)).raw_output()
 790     else:
 791         return ''
 792
 793 # TODO: take another parameter representing a diff string as we
 794 # usually invoke git.diff() form the calling functions
 795 def diffstat(files = None, rev1 = 'HEAD', rev2 = None):
 796     """Return the diffstat between rev1 and rev2."""
 797     return GRun('apply', '--stat', '--summary'
 798                 ).raw_input(diff(files, rev1, rev2)).raw_output()
 799
 800 def files(rev1, rev2, diff_flags = []):
 801     """Return the files modified between rev1 and rev2
 802     """
 803
 804     result = []
 805     for line in GRun('diff-tree', *(diff_flags + ['-r', rev1, rev2])
 806                      ).output_lines():
 807         result.append('%s %s' % tuple(line.split(' ', 4)[-1].split('\t', 1)))
 808
 809     return '\n'.join(result)
 810
 811 def barefiles(rev1, rev2):
 812     """Return the files modified between rev1 and rev2, without status info
 813     """
 814
 815     result = []
 816     for line in GRun('diff-tree', '-r', rev1, rev2).output_lines():
 817         result.append(line.split(' ', 4)[-1].split('\t', 1)[-1])
 818
 819     return '\n'.join(result)
 820
 821 def pretty_commit(commit_id = 'HEAD', flags = []):
 822     """Return a given commit (log + diff)
 823     """
 824     return GRun('show', *(flags + [commit_id])).raw_output()
 825
 826 def checkout(files = None, tree_id = None, force = False):
 827     """Check out the given or all files
 828     """
 829     if tree_id:
 830         try:
 831             GRun('read-tree', '--reset', tree_id).run()
 832         except GitRunException:
 833             raise GitException, 'Failed "git read-tree" --reset %s' % tree_id
 834
 835     cmd = ['checkout-index', '-q', '-u']
 836     if force:
 837         cmd.append('-f')
 838     if files:
 839         GRun(*(cmd + ['--'])).xargs(files)
 840     else:
 841         GRun(*(cmd + ['-a'])).run()
 842
 843 def switch(tree_id, keep = False):
 844     """Switch the tree to the given id
 845     """
 846     if keep:
 847         # only update the index while keeping the local changes
 848         GRun('read-tree', tree_id).run()
 849     else:
 850         refresh_index()
 851         try:
 852             GRun('read-tree', '-u', '-m', get_head(), tree_id).run()
 853         except GitRunException:
 854             raise GitException, 'read-tree failed (local changes maybe?)'
 855
 856     __set_head(tree_id)
 857
 858 def reset(files = None, tree_id = None, check_out = True):
 859     """Revert the tree changes relative to the given tree_id. It removes
 860     any local changes
 861     """
 862     if not tree_id:
 863         tree_id = get_head()
 864
 865     if check_out:
 866         cache_files = tree_status(files, tree_id)
 867         # files which were added but need to be removed
 868         rm_files =  [x[1] for x in cache_files if x[0] in ['A']]
 869
 870         checkout(files, tree_id, True)
 871         # checkout doesn't remove files
 872         map(os.remove, rm_files)
 873
 874     # if the reset refers to the whole tree, switch the HEAD as well
 875     if not files:
 876         __set_head(tree_id)
 877
 878 def fetch(repository = 'origin', refspec = None):
 879     """Fetches changes from the remote repository, using 'git fetch'
 880     by default.
 881     """
 882     # we update the HEAD
 883     __clear_head_cache()
 884
 885     args = [repository]
 886     if refspec:
 887         args.append(refspec)
 888
 889     command = config.get('branch.%s.stgit.fetchcmd' % get_head_file()) or \
 890               config.get('stgit.fetchcmd')
 891     Run(*(command.split() + args)).run()
 892
 893 def pull(repository = 'origin', refspec = None):
 894     """Fetches changes from the remote repository, using 'git pull'
 895     by default.
 896     """
 897     # we update the HEAD
 898     __clear_head_cache()
 899
 900     args = [repository]
 901     if refspec:
 902         args.append(refspec)
 903
 904     command = config.get('branch.%s.stgit.pullcmd' % get_head_file()) or \
 905               config.get('stgit.pullcmd')
 906     Run(*(command.split() + args)).run()
 907
 908 def rebase(tree_id = None):
 909     """Rebase the current tree to the give tree_id. The tree_id
 910     argument may be something other than a GIT id if an external
 911     command is invoked.
 912     """
 913     command = config.get('branch.%s.stgit.rebasecmd' % get_head_file()) \
 914                 or config.get('stgit.rebasecmd')
 915     if tree_id:
 916         args = [tree_id]
 917     elif command:
 918         args = []
 919     else:
 920         raise GitException, 'Default rebasing requires a commit id'
 921     if command:
 922         # clear the HEAD cache as the custom rebase command will update it
 923         __clear_head_cache()
 924         Run(*(command.split() + args)).run()
 925     else:
 926         # default rebasing
 927         reset(tree_id = tree_id)
 928
 929 def repack():
 930     """Repack all objects into a single pack
 931     """
 932     GRun('repack', '-a', '-d', '-f').run()
 933
 934 def apply_patch(filename = None, diff = None, base = None,
 935                 fail_dump = True):
 936     """Apply a patch onto the current or given index. There must not
 937     be any local changes in the tree, otherwise the command fails
 938     """
 939     if diff is None:
 940         if filename:
 941             f = file(filename)
 942         else:
 943             f = sys.stdin
 944         diff = f.read()
 945         if filename:
 946             f.close()
 947
 948     if base:
 949         orig_head = get_head()
 950         switch(base)
 951     else:
 952         refresh_index()
 953
 954     try:
 955         GRun('apply', '--index').raw_input(diff).no_output()
 956     except GitRunException:
 957         if base:
 958             switch(orig_head)
 959         if fail_dump:
 960             # write the failed diff to a file
 961             f = file('.stgit-failed.patch', 'w+')
 962             f.write(diff)
 963             f.close()
 964             out.warn('Diff written to the .stgit-failed.patch file')
 965
 966         raise
 967
 968     if base:
 969         top = commit(message = 'temporary commit used for applying a patch',
 970                      parents = [base])
 971         switch(orig_head)
 972         merge(base, orig_head, top)
 973
 974 def clone(repository, local_dir):
 975     """Clone a remote repository. At the moment, just use the
 976     'git clone' script
 977     """
 978     GRun('clone', repository, local_dir).run()
 979
 980 def modifying_revs(files, base_rev, head_rev):
 981     """Return the revisions from the list modifying the given files."""
 982     return GRun('rev-list', '%s..%s' % (base_rev, head_rev), '--', *files
 983                 ).output_lines()
 984
 985 def refspec_localpart(refspec):
 986     m = re.match('^[^:]*:([^:]*)$', refspec)
 987     if m:
 988         return m.group(1)
 989     else:
 990         raise GitException, 'Cannot parse refspec "%s"' % line
 991
 992 def refspec_remotepart(refspec):
 993     m = re.match('^([^:]*):[^:]*$', refspec)
 994     if m:
 995         return m.group(1)
 996     else:
 997         raise GitException, 'Cannot parse refspec "%s"' % line
 998
 999
1000 def __remotes_from_config():
1001     return config.sections_matching(r'remote\.(.*)\.url')
1002
1003 def __remotes_from_dir(dir):
1004     d = os.path.join(basedir.get(), dir)
1005     if os.path.exists(d):
1006         return os.listdir(d)
1007     else:
1008         return []
1009
1010 def remotes_list():
1011     """Return the list of remotes in the repository
1012     """
1013     return (set(__remotes_from_config())
1014             | set(__remotes_from_dir('remotes'))
1015             | set(__remotes_from_dir('branches')))
1016
1017 def remotes_local_branches(remote):
1018     """Returns the list of local branches fetched from given remote
1019     """
1020
1021     branches = []
1022     if remote in __remotes_from_config():
1023         for line in config.getall('remote.%s.fetch' % remote):
1024             branches.append(refspec_localpart(line))
1025     elif remote in __remotes_from_dir('remotes'):
1026         stream = open(os.path.join(basedir.get(), 'remotes', remote), 'r')
1027         for line in stream:
1028             # Only consider Pull lines
1029             m = re.match('^Pull: (.*)\n$', line)
1030             if m:
1031                 branches.append(refspec_localpart(m.group(1)))
1032         stream.close()
1033     elif remote in __remotes_from_dir('branches'):
1034         # old-style branches only declare one branch
1035         branches.append('refs/heads/'+remote);
1036     else:
1037         raise GitException, 'Unknown remote "%s"' % remote
1038
1039     return branches
1040
1041 def identify_remote(branchname):
1042     """Return the name for the remote to pull the given branchname
1043     from, or None if we believe it is a local branch.
1044     """
1045
1046     for remote in remotes_list():
1047         if branchname in remotes_local_branches(remote):
1048             return remote
1049
1050     # if we get here we've found nothing, the branch is a local one
1051     return None
1052
1053 def fetch_head():
1054     """Return the git id for the tip of the parent branch as left by
1055     'git fetch'.
1056     """
1057
1058     fetch_head=None
1059     stream = open(os.path.join(basedir.get(), 'FETCH_HEAD'), "r")
1060     for line in stream:
1061         # Only consider lines not tagged not-for-merge
1062         m = re.match('^([^\t]*)\t\t', line)
1063         if m:
1064             if fetch_head:
1065                 raise GitException, 'StGit does not support multiple FETCH_HEAD'
1066             else:
1067                 fetch_head=m.group(1)
1068     stream.close()
1069
1070     if not fetch_head:
1071         out.warn('No for-merge remote head found in FETCH_HEAD')
1072
1073     # here we are sure to have a single fetch_head
1074     return fetch_head
1075
1076 def all_refs():
1077     """Return a list of all refs in the current repository.
1078     """
1079
1080     return [line.split()[1] for line in GRun('show-ref').output_lines()]