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, popen2
  22
  23 from stgit.utils import *
  24
  25 # git exception class
  26 class GitException(Exception):
  27     pass
  28
  29
  30
  31 #
  32 # Classes
  33 #
  34 class Commit:
  35     """Handle the commit objects
  36     """
  37     def __init__(self, id_hash):
  38         self.__id_hash = id_hash
  39
  40         lines = _output_lines('git-cat-file commit %s' % id_hash)
  41         self.__parents = []
  42         for i in range(len(lines)):
  43             line = lines[i]
  44             if line == '\n':
  45                 break
  46             field = line.strip().split(' ', 1)
  47             if field[0] == 'tree':
  48                 self.__tree = field[1]
  49             elif field[0] == 'parent':
  50                 self.__parents.append(field[1])
  51             if field[0] == 'author':
  52                 self.__author = field[1]
  53             if field[0] == 'committer':
  54                 self.__committer = field[1]
  55         self.__log = ''.join(lines[i+1:])
  56
  57     def get_id_hash(self):
  58         return self.__id_hash
  59
  60     def get_tree(self):
  61         return self.__tree
  62
  63     def get_parent(self):
  64         return self.__parents[0]
  65
  66     def get_parents(self):
  67         return self.__parents
  68
  69     def get_author(self):
  70         return self.__author
  71
  72     def get_committer(self):
  73         return self.__committer
  74
  75     def get_log(self):
  76         return self.__log
  77
  78 # dictionary of Commit objects, used to avoid multiple calls to git
  79 __commits = dict()
  80
  81 #
  82 # Functions
  83 #
  84
  85 def get_base_dir():
  86     """Different start-up variables read from the environment
  87     """
  88     if 'GIT_DIR' in os.environ:
  89         return os.environ['GIT_DIR']
  90     else:
  91         return _output_one_line('git-rev-parse --git-dir')
  92
  93 def get_commit(id_hash):
  94     """Commit objects factory. Save/look-up them in the __commits
  95     dictionary
  96     """
  97     global __commits
  98
  99     if id_hash in __commits:
 100         return __commits[id_hash]
 101     else:
 102         commit = Commit(id_hash)
 103         __commits[id_hash] = commit
 104         return commit
 105
 106 def get_conflicts():
 107     """Return the list of file conflicts
 108     """
 109     conflicts_file = os.path.join(get_base_dir(), 'conflicts')
 110     if os.path.isfile(conflicts_file):
 111         f = file(conflicts_file)
 112         names = [line.strip() for line in f.readlines()]
 113         f.close()
 114         return names
 115     else:
 116         return None
 117
 118 def _input(cmd, file_desc):
 119     p = popen2.Popen3(cmd, True)
 120     while True:
 121         line = file_desc.readline()
 122         if not line:
 123             break
 124         p.tochild.write(line)
 125     p.tochild.close()
 126     if p.wait():
 127         raise GitException, '%s failed' % str(cmd)
 128
 129 def _output(cmd):
 130     p=popen2.Popen3(cmd, True)
 131     output = p.fromchild.read()
 132     if p.wait():
 133         raise GitException, '%s failed' % str(cmd)
 134     return output
 135
 136 def _output_one_line(cmd, file_desc = None):
 137     p=popen2.Popen3(cmd, True)
 138     if file_desc != None:
 139         for line in file_desc:
 140             p.tochild.write(line)
 141         p.tochild.close()
 142     output = p.fromchild.readline().strip()
 143     if p.wait():
 144         raise GitException, '%s failed' % str(cmd)
 145     return output
 146
 147 def _output_lines(cmd):
 148     p=popen2.Popen3(cmd, True)
 149     lines = p.fromchild.readlines()
 150     if p.wait():
 151         raise GitException, '%s failed' % str(cmd)
 152     return lines
 153
 154 def __run(cmd, args=None):
 155     """__run: runs cmd using spawnvp.
 156
 157     Runs cmd using spawnvp.  The shell is avoided so it won't mess up
 158     our arguments.  If args is very large, the command is run multiple
 159     times; args is split xargs style: cmd is passed on each
 160     invocation.  Unlike xargs, returns immediately if any non-zero
 161     return code is received.
 162     """
 163
 164     args_l=cmd.split()
 165     if args is None:
 166         args = []
 167     for i in range(0, len(args)+1, 100):
 168         r=os.spawnvp(os.P_WAIT, args_l[0], args_l + args[i:min(i+100, len(args))])
 169     if r:
 170         return r
 171     return 0
 172
 173 def __tree_status(files = None, tree_id = 'HEAD', unknown = False,
 174                   noexclude = True):
 175     """Returns a list of pairs - [status, filename]
 176     """
 177     refresh_index()
 178
 179     if not files:
 180         files = []
 181     cache_files = []
 182
 183     # unknown files
 184     if unknown:
 185         exclude_file = os.path.join(get_base_dir(), 'info', 'exclude')
 186         base_exclude = ['--exclude=%s' % s for s in
 187                         ['*.[ao]', '*.pyc', '.*', '*~', '#*', 'TAGS', 'tags']]
 188         base_exclude.append('--exclude-per-directory=.gitignore')
 189
 190         if os.path.exists(exclude_file):
 191             extra_exclude = ['--exclude-from=%s' % exclude_file]
 192         else:
 193             extra_exclude = []
 194         if noexclude:
 195             extra_exclude = base_exclude = []
 196
 197         lines = _output_lines(['git-ls-files', '--others'] + base_exclude
 198                         + extra_exclude)
 199         cache_files += [('?', line.strip()) for line in lines]
 200
 201     # conflicted files
 202     conflicts = get_conflicts()
 203     if not conflicts:
 204         conflicts = []
 205     cache_files += [('C', filename) for filename in conflicts]
 206
 207     # the rest
 208     for line in _output_lines(['git-diff-index', tree_id] + files):
 209         fs = tuple(line.rstrip().split(' ',4)[-1].split('\t',1))
 210         if fs[1] not in conflicts:
 211             cache_files.append(fs)
 212
 213     return cache_files
 214
 215 def local_changes():
 216     """Return true if there are local changes in the tree
 217     """
 218     return len(__tree_status()) != 0
 219
 220 # HEAD value cached
 221 __head = None
 222
 223 def get_head():
 224     """Verifies the HEAD and returns the SHA1 id that represents it
 225     """
 226     global __head
 227
 228     if not __head:
 229         __head = rev_parse('HEAD')
 230     return __head
 231
 232 def get_head_file():
 233     """Returns the name of the file pointed to by the HEAD link
 234     """
 235     return os.path.basename(_output_one_line('git-symbolic-ref HEAD'))
 236
 237 def set_head_file(ref):
 238     """Resets HEAD to point to a new ref
 239     """
 240     # head cache flushing is needed since we might have a different value
 241     # in the new head
 242     __clear_head_cache()
 243     if __run('git-symbolic-ref HEAD', [ref]) != 0:
 244         raise GitException, 'Could not set head to "%s"' % ref
 245
 246 def __set_head(val):
 247     """Sets the HEAD value
 248     """
 249     global __head
 250
 251     if not __head or __head != val:
 252         if __run('git-update-ref HEAD', [val]) != 0:
 253             raise GitException, 'Could not update HEAD to "%s".' % val
 254         __head = val
 255
 256 def __clear_head_cache():
 257     """Sets the __head to None so that a re-read is forced
 258     """
 259     global __head
 260
 261     __head = None
 262
 263 def refresh_index():
 264     """Refresh index with stat() information from the working directory.
 265     """
 266     __run('git-update-index -q --unmerged --refresh')
 267
 268 def rev_parse(git_id):
 269     """Parse the string and return a verified SHA1 id
 270     """
 271     try:
 272         return _output_one_line(['git-rev-parse', '--verify', git_id])
 273     except GitException:
 274         raise GitException, 'Unknown revision: %s' % git_id
 275
 276 def branch_exists(branch):
 277     """Existance check for the named branch
 278     """
 279     for line in _output_lines(['git-rev-parse', '--symbolic', '--all']):
 280         if line.strip() == branch:
 281             return True
 282     return False
 283
 284 def create_branch(new_branch, tree_id = None):
 285     """Create a new branch in the git repository
 286     """
 287     new_head = os.path.join('refs', 'heads', new_branch)
 288     if branch_exists(new_head):
 289         raise GitException, 'Branch "%s" already exists' % new_branch
 290
 291     current_head = get_head()
 292     set_head_file(new_head)
 293     __set_head(current_head)
 294
 295     # a checkout isn't needed if new branch points to the current head
 296     if tree_id:
 297         switch(tree_id)
 298
 299     if os.path.isfile(os.path.join(get_base_dir(), 'MERGE_HEAD')):
 300         os.remove(os.path.join(get_base_dir(), 'MERGE_HEAD'))
 301
 302 def switch_branch(name):
 303     """Switch to a git branch
 304     """
 305     global __head
 306
 307     new_head = os.path.join('refs', 'heads', name)
 308     if not branch_exists(new_head):
 309         raise GitException, 'Branch "%s" does not exist' % name
 310
 311     tree_id = rev_parse(new_head + '^0')
 312     if tree_id != get_head():
 313         refresh_index()
 314         if __run('git-read-tree -u -m', [get_head(), tree_id]) != 0:
 315             raise GitException, 'git-read-tree failed (local changes maybe?)'
 316         __head = tree_id
 317     set_head_file(new_head)
 318
 319     if os.path.isfile(os.path.join(get_base_dir(), 'MERGE_HEAD')):
 320         os.remove(os.path.join(get_base_dir(), 'MERGE_HEAD'))
 321
 322 def delete_branch(name):
 323     """Delete a git branch
 324     """
 325     branch_head = os.path.join('refs', 'heads', name)
 326     if not branch_exists(branch_head):
 327         raise GitException, 'Branch "%s" does not exist' % name
 328     os.remove(os.path.join(get_base_dir(), branch_head))
 329
 330 def rename_branch(from_name, to_name):
 331     """Rename a git branch
 332     """
 333     from_head = os.path.join('refs', 'heads', from_name)
 334     if not branch_exists(from_head):
 335         raise GitException, 'Branch "%s" does not exist' % from_name
 336     to_head = os.path.join('refs', 'heads', to_name)
 337     if branch_exists(to_head):
 338         raise GitException, 'Branch "%s" already exists' % to_name
 339
 340     if get_head_file() == from_name:
 341         set_head_file(to_head)
 342     os.rename(os.path.join(get_base_dir(), from_head), \
 343               os.path.join(get_base_dir(), to_head))
 344
 345 def add(names):
 346     """Add the files or recursively add the directory contents
 347     """
 348     # generate the file list
 349     files = []
 350     for i in names:
 351         if not os.path.exists(i):
 352             raise GitException, 'Unknown file or directory: %s' % i
 353
 354         if os.path.isdir(i):
 355             # recursive search. We only add files
 356             for root, dirs, local_files in os.walk(i):
 357                 for name in [os.path.join(root, f) for f in local_files]:
 358                     if os.path.isfile(name):
 359                         files.append(os.path.normpath(name))
 360         elif os.path.isfile(i):
 361             files.append(os.path.normpath(i))
 362         else:
 363             raise GitException, '%s is not a file or directory' % i
 364
 365     if files:
 366         if __run('git-update-index --add --', files):
 367             raise GitException, 'Unable to add file'
 368
 369 def rm(files, force = False):
 370     """Remove a file from the repository
 371     """
 372     if not force:
 373         for f in files:
 374             if os.path.exists(f):
 375                 raise GitException, '%s exists. Remove it first' %f
 376         if files:
 377             __run('git-update-index --remove --', files)
 378     else:
 379         if files:
 380             __run('git-update-index --force-remove --', files)
 381
 382 def update_cache(files = None, force = False):
 383     """Update the cache information for the given files
 384     """
 385     if not files:
 386         files = []
 387
 388     cache_files = __tree_status(files)
 389
 390     # everything is up-to-date
 391     if len(cache_files) == 0:
 392         return False
 393
 394     # check for unresolved conflicts
 395     if not force and [x for x in cache_files
 396                       if x[0] not in ['M', 'N', 'A', 'D']]:
 397         raise GitException, 'Updating cache failed: unresolved conflicts'
 398
 399     # update the cache
 400     add_files = [x[1] for x in cache_files if x[0] in ['N', 'A']]
 401     rm_files =  [x[1] for x in cache_files if x[0] in ['D']]
 402     m_files =   [x[1] for x in cache_files if x[0] in ['M']]
 403
 404     if add_files and __run('git-update-index --add --', add_files) != 0:
 405         raise GitException, 'Failed git-update-index --add'
 406     if rm_files and __run('git-update-index --force-remove --', rm_files) != 0:
 407         raise GitException, 'Failed git-update-index --rm'
 408     if m_files and __run('git-update-index --', m_files) != 0:
 409         raise GitException, 'Failed git-update-index'
 410
 411     return True
 412
 413 def commit(message, files = None, parents = None, allowempty = False,
 414            cache_update = True, tree_id = None,
 415            author_name = None, author_email = None, author_date = None,
 416            committer_name = None, committer_email = None):
 417     """Commit the current tree to repository
 418     """
 419     if not files:
 420         files = []
 421     if not parents:
 422         parents = []
 423
 424     # Get the tree status
 425     if cache_update and parents != []:
 426         changes = update_cache(files)
 427         if not changes and not allowempty:
 428             raise GitException, 'No changes to commit'
 429
 430     # get the commit message
 431     if message[-1:] != '\n':
 432         message += '\n'
 433
 434     must_switch = True
 435     # write the index to repository
 436     if tree_id == None:
 437         tree_id = _output_one_line('git-write-tree')
 438     else:
 439         must_switch = False
 440
 441     # the commit
 442     cmd = ''
 443     if author_name:
 444         cmd += 'GIT_AUTHOR_NAME="%s" ' % author_name
 445     if author_email:
 446         cmd += 'GIT_AUTHOR_EMAIL="%s" ' % author_email
 447     if author_date:
 448         cmd += 'GIT_AUTHOR_DATE="%s" ' % author_date
 449     if committer_name:
 450         cmd += 'GIT_COMMITTER_NAME="%s" ' % committer_name
 451     if committer_email:
 452         cmd += 'GIT_COMMITTER_EMAIL="%s" ' % committer_email
 453     cmd += 'git-commit-tree %s' % tree_id
 454
 455     # get the parents
 456     for p in parents:
 457         cmd += ' -p %s' % p
 458
 459     commit_id = _output_one_line(cmd, message)
 460     if must_switch:
 461         __set_head(commit_id)
 462
 463     return commit_id
 464
 465 def apply_diff(rev1, rev2):
 466     """Apply the diff between rev1 and rev2 onto the current
 467     index. This function doesn't need to raise an exception since it
 468     is only used for fast-pushing a patch. If this operation fails,
 469     the pushing would fall back to the three-way merge.
 470     """
 471     return os.system('git-diff-tree -p %s %s | git-apply --index 2> /dev/null'
 472                      % (rev1, rev2)) == 0
 473
 474 def merge(base, head1, head2):
 475     """Perform a 3-way merge between base, head1 and head2 into the
 476     local tree
 477     """
 478     refresh_index()
 479     if __run('git-read-tree -u -m', [base, head1, head2]) != 0:
 480         raise GitException, 'git-read-tree failed (local changes maybe?)'
 481
 482     # this can fail if there are conflicts
 483     if __run('git-merge-index -o -q gitmergeonefile.py -a') != 0:
 484         raise GitException, 'git-merge-index failed (possible conflicts)'
 485
 486 def status(files = None, modified = False, new = False, deleted = False,
 487            conflict = False, unknown = False, noexclude = False):
 488     """Show the tree status
 489     """
 490     if not files:
 491         files = []
 492
 493     cache_files = __tree_status(files, unknown = True, noexclude = noexclude)
 494     all = not (modified or new or deleted or conflict or unknown)
 495
 496     if not all:
 497         filestat = []
 498         if modified:
 499             filestat.append('M')
 500         if new:
 501             filestat.append('A')
 502             filestat.append('N')
 503         if deleted:
 504             filestat.append('D')
 505         if conflict:
 506             filestat.append('C')
 507         if unknown:
 508             filestat.append('?')
 509         cache_files = [x for x in cache_files if x[0] in filestat]
 510
 511     for fs in cache_files:
 512         if all:
 513             print '%s %s' % (fs[0], fs[1])
 514         else:
 515             print '%s' % fs[1]
 516
 517 def diff(files = None, rev1 = 'HEAD', rev2 = None, out_fd = None):
 518     """Show the diff between rev1 and rev2
 519     """
 520     if not files:
 521         files = []
 522
 523     if rev1 and rev2:
 524         diff_str = _output(['git-diff-tree', '-p', rev1, rev2] + files)
 525     elif rev1 or rev2:
 526         refresh_index()
 527         if rev2:
 528             diff_str = _output(['git-diff-index', '-p', '-R', rev2] + files)
 529         else:
 530             diff_str = _output(['git-diff-index', '-p', rev1] + files)
 531     else:
 532         diff_str = ''
 533
 534     if out_fd:
 535         out_fd.write(diff_str)
 536     else:
 537         return diff_str
 538
 539 def diffstat(files = None, rev1 = 'HEAD', rev2 = None):
 540     """Return the diffstat between rev1 and rev2
 541     """
 542     if not files:
 543         files = []
 544
 545     p=popen2.Popen3('git-apply --stat')
 546     diff(files, rev1, rev2, p.tochild)
 547     p.tochild.close()
 548     diff_str = p.fromchild.read().rstrip()
 549     if p.wait():
 550         raise GitException, 'git.diffstat failed'
 551     return diff_str
 552
 553 def files(rev1, rev2):
 554     """Return the files modified between rev1 and rev2
 555     """
 556
 557     result = ''
 558     for line in _output_lines('git-diff-tree -r %s %s' % (rev1, rev2)):
 559         result += '%s %s\n' % tuple(line.rstrip().split(' ',4)[-1].split('\t',1))
 560
 561     return result.rstrip()
 562
 563 def barefiles(rev1, rev2):
 564     """Return the files modified between rev1 and rev2, without status info
 565     """
 566
 567     result = ''
 568     for line in _output_lines('git-diff-tree -r %s %s' % (rev1, rev2)):
 569         result += '%s\n' % line.rstrip().split(' ',4)[-1].split('\t',1)[-1]
 570
 571     return result.rstrip()
 572
 573 def checkout(files = None, tree_id = None, force = False):
 574     """Check out the given or all files
 575     """
 576     if not files:
 577         files = []
 578
 579     if tree_id and __run('git-read-tree -m', [tree_id]) != 0:
 580         raise GitException, 'Failed git-read-tree -m %s' % tree_id
 581
 582     checkout_cmd = 'git-checkout-index -q -u'
 583     if force:
 584         checkout_cmd += ' -f'
 585     if len(files) == 0:
 586         checkout_cmd += ' -a'
 587     else:
 588         checkout_cmd += ' --'
 589
 590     if __run(checkout_cmd, files) != 0:
 591         raise GitException, 'Failed git-checkout-index'
 592
 593 def switch(tree_id):
 594     """Switch the tree to the given id
 595     """
 596     refresh_index()
 597     if __run('git-read-tree -u -m', [get_head(), tree_id]) != 0:
 598         raise GitException, 'git-read-tree failed (local changes maybe?)'
 599
 600     __set_head(tree_id)
 601
 602 def reset(files = None, tree_id = None):
 603     """Revert the tree changes relative to the given tree_id. It removes
 604     any local changes
 605     """
 606     checkout(files, tree_id, True)
 607
 608     # if the reset refers to the whole tree, switch the HEAD as well
 609     if tree_id and not files:
 610         __set_head(tree_id)
 611
 612 def pull(repository = 'origin', refspec = None):
 613     """Pull changes from the remote repository. At the moment, just
 614     use the 'git pull' command
 615     """
 616     # 'git pull' updates the HEAD
 617     __clear_head_cache()
 618
 619     args = [repository]
 620     if refspec:
 621         args.append(refspec)
 622
 623     if __run('git pull', args) != 0:
 624         raise GitException, 'Failed "git pull %s"' % repository
 625
 626 def apply_patch(filename = None, base = None):
 627     """Apply a patch onto the current or given index. There must not
 628     be any local changes in the tree, otherwise the command fails
 629     """
 630     def __apply_patch():
 631         if filename:
 632             return __run('git-apply --index', [filename]) == 0
 633         else:
 634             try:
 635                 _input('git-apply --index', sys.stdin)
 636             except GitException:
 637                 return False
 638             return True
 639
 640     if base:
 641         orig_head = get_head()
 642         switch(base)
 643     else:
 644         refresh_index()         # needed since __apply_patch() doesn't do it
 645
 646     if not __apply_patch():
 647         if base:
 648             switch(orig_head)
 649         raise GitException, 'Patch does not apply cleanly'
 650     elif base:
 651         top = commit(message = 'temporary commit used for applying a patch',
 652                      parents = [base])
 653         switch(orig_head)
 654         merge(base, orig_head, top)
 655
 656 def clone(repository, local_dir):
 657     """Clone a remote repository. At the moment, just use the
 658     'git clone' script
 659     """
 660     if __run('git clone', [repository, local_dir]) != 0:
 661         raise GitException, 'Failed "git clone %s %s"' \
 662               % (repository, local_dir)
 663
 664 def modifying_revs(files, base_rev):
 665     """Return the revisions from the list modifying the given files
 666     """
 667     cmd = ['git-rev-list', '%s..' % base_rev, '--']
 668     revs = [line.strip() for line in _output_lines(cmd + files)]
 669
 670     return revs