1 """Python GIT interface
5 Copyright (C) 2005, Catalin Marinas <catalin.marinas@gmail.com>
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.
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.
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
21 import sys, os, re, gitmergeonefile
22 from shutil import copyfile
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
32 class GitException(StgException):
35 # When a subprocess has a problem, we want the exception to be a
36 # subclass of GitException.
37 class GitRunException(GitException):
48 """An author, committer, etc."""
49 def __init__(self, name = None, email = None, date = '',
51 self.name = self.email = self.date = None
52 if name or email or date:
58 assert not (name or email or date)
60 m = re.match(r'^(.+)<(.+)>(.*)$', s)
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):
67 def set_email(self, val):
70 def set_date(self, val):
74 if self.name and self.email:
75 return '%s <%s>' % (self.name, self.email)
77 raise GitException, 'not enough identity data'
80 """Handle the commit objects
82 def __init__(self, id_hash):
83 self.__id_hash = id_hash
85 lines = GRun('git-cat-file', 'commit', id_hash).output_lines()
86 for i in range(len(lines)):
89 break # we've seen all the header fields
90 key, val = line.split(' ', 1)
95 elif key == 'committer':
96 self.__committer = val
98 pass # ignore other headers
99 self.__log = '\n'.join(lines[i+1:])
101 def get_id_hash(self):
102 return self.__id_hash
107 def get_parent(self):
108 parents = self.get_parents()
114 def get_parents(self):
115 return GRun('git-rev-list', '--parents', '--max-count=1', self.__id_hash
116 ).output_one_line().split()[1:]
118 def get_author(self):
121 def get_committer(self):
122 return self.__committer
128 return self.get_id_hash()
130 # dictionary of Commit objects, used to avoid multiple calls to git
137 def get_commit(id_hash):
138 """Commit objects factory. Save/look-up them in the __commits
143 if id_hash in __commits:
144 return __commits[id_hash]
146 commit = Commit(id_hash)
147 __commits[id_hash] = commit
151 """Return the list of file conflicts
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()]
163 files = [os.path.join(basedir.get(), 'info', 'exclude')]
164 user_exclude = config.get('core.excludesfile')
166 files.append(user_exclude)
169 def tree_status(files = None, tree_id = 'HEAD', unknown = False,
170 noexclude = True, verbose = False, diff_flags = []):
171 """Get the status of all changed files, or of a selected set of
172 files. Returns a list of pairs - (status, filename).
174 If 'files' is None, it will check all files, and optionally all
175 unknown files. If 'files' is a list, it will only check the files
178 assert files == None or not unknown
181 out.start('Checking for changes in the working directory')
189 cmd = ['git-ls-files', '-z', '--others', '--directory',
190 '--no-empty-directory']
192 cmd += ['--exclude=%s' % s for s in
193 ['*.[ao]', '*.pyc', '.*', '*~', '#*', 'TAGS', 'tags']]
194 cmd += ['--exclude-per-directory=.gitignore']
195 cmd += ['--exclude-from=%s' % fn
196 for fn in exclude_files()
197 if os.path.exists(fn)]
199 lines = GRun(*cmd).raw_output().split('\0')
200 cache_files += [('?', line) for line in lines if line]
203 conflicts = get_conflicts()
206 cache_files += [('C', filename) for filename in conflicts
207 if files == None or filename in files]
210 args = diff_flags + [tree_id]
212 args += ['--'] + files
213 for line in GRun('git-diff-index', *args).output_lines():
214 fs = tuple(line.rstrip().split(' ',4)[-1].split('\t',1))
215 if fs[1] not in conflicts:
216 cache_files.append(fs)
221 assert files == None or set(f for s,f in cache_files) <= set(files)
224 def local_changes(verbose = True):
225 """Return true if there are local changes in the tree
227 return len(tree_status(verbose = verbose)) != 0
231 hr = re.compile(r'^[0-9a-f]{40} refs/heads/(.+)$')
232 for line in GRun('git-show-ref', '--heads').output_lines():
234 heads.append(m.group(1))
241 """Verifies the HEAD and returns the SHA1 id that represents it
246 __head = rev_parse('HEAD')
250 """Returns the name of the file pointed to by the HEAD link
252 return strip_prefix('refs/heads/',
253 GRun('git-symbolic-ref', 'HEAD').output_one_line())
255 def set_head_file(ref):
256 """Resets HEAD to point to a new ref
258 # head cache flushing is needed since we might have a different value
262 GRun('git-symbolic-ref', 'HEAD', 'refs/heads/%s' % ref).run()
263 except GitRunException:
264 raise GitException, 'Could not set head to "%s"' % ref
266 def set_ref(ref, val):
267 """Point ref at a new commit object."""
269 GRun('git-update-ref', ref, val).run()
270 except GitRunException:
271 raise GitException, 'Could not update %s to "%s".' % (ref, val)
273 def set_branch(branch, val):
274 set_ref('refs/heads/%s' % branch, val)
277 """Sets the HEAD value
281 if not __head or __head != val:
285 # only allow SHA1 hashes
286 assert(len(__head) == 40)
288 def __clear_head_cache():
289 """Sets the __head to None so that a re-read is forced
296 """Refresh index with stat() information from the working directory.
298 GRun('git-update-index', '-q', '--unmerged', '--refresh').run()
300 def rev_parse(git_id):
301 """Parse the string and return a verified SHA1 id
304 return GRun('git-rev-parse', '--verify', git_id
305 ).discard_stderr().output_one_line()
306 except GitRunException:
307 raise GitException, 'Unknown revision: %s' % git_id
316 def branch_exists(branch):
317 return ref_exists('refs/heads/%s' % branch)
319 def create_branch(new_branch, tree_id = None):
320 """Create a new branch in the git repository
322 if branch_exists(new_branch):
323 raise GitException, 'Branch "%s" already exists' % new_branch
325 current_head = get_head()
326 set_head_file(new_branch)
327 __set_head(current_head)
329 # a checkout isn't needed if new branch points to the current head
333 if os.path.isfile(os.path.join(basedir.get(), 'MERGE_HEAD')):
334 os.remove(os.path.join(basedir.get(), 'MERGE_HEAD'))
336 def switch_branch(new_branch):
337 """Switch to a git branch
341 if not branch_exists(new_branch):
342 raise GitException, 'Branch "%s" does not exist' % new_branch
344 tree_id = rev_parse('refs/heads/%s^{commit}' % new_branch)
345 if tree_id != get_head():
348 GRun('git-read-tree', '-u', '-m', get_head(), tree_id).run()
349 except GitRunException:
350 raise GitException, 'git-read-tree failed (local changes maybe?)'
352 set_head_file(new_branch)
354 if os.path.isfile(os.path.join(basedir.get(), 'MERGE_HEAD')):
355 os.remove(os.path.join(basedir.get(), 'MERGE_HEAD'))
358 if not ref_exists(ref):
359 raise GitException, '%s does not exist' % ref
360 sha1 = GRun('git-show-ref', '-s', ref).output_one_line()
362 GRun('git-update-ref', '-d', ref, sha1).run()
363 except GitRunException:
364 raise GitException, 'Failed to delete ref %s' % ref
366 def delete_branch(name):
367 delete_ref('refs/heads/%s' % name)
369 def rename_ref(from_ref, to_ref):
370 if not ref_exists(from_ref):
371 raise GitException, '"%s" does not exist' % from_ref
372 if ref_exists(to_ref):
373 raise GitException, '"%s" already exists' % to_ref
375 sha1 = GRun('git-show-ref', '-s', from_ref).output_one_line()
377 GRun('git-update-ref', to_ref, sha1, '0'*40).run()
378 except GitRunException:
379 raise GitException, 'Failed to create new ref %s' % to_ref
381 GRun('git-update-ref', '-d', from_ref, sha1).run()
382 except GitRunException:
383 raise GitException, 'Failed to delete ref %s' % from_ref
385 def rename_branch(from_name, to_name):
386 """Rename a git branch."""
387 rename_ref('refs/heads/%s' % from_name, 'refs/heads/%s' % to_name)
388 if get_head_file() == from_name:
389 set_head_file(to_name)
390 reflog_dir = os.path.join(basedir.get(), 'logs', 'refs', 'heads')
391 if os.path.exists(reflog_dir) \
392 and os.path.exists(os.path.join(reflog_dir, from_name)):
393 rename(reflog_dir, from_name, to_name)
396 """Add the files or recursively add the directory contents
398 # generate the file list
401 if not os.path.exists(i):
402 raise GitException, 'Unknown file or directory: %s' % i
405 # recursive search. We only add files
406 for root, dirs, local_files in os.walk(i):
407 for name in [os.path.join(root, f) for f in local_files]:
408 if os.path.isfile(name):
409 files.append(os.path.normpath(name))
410 elif os.path.isfile(i):
411 files.append(os.path.normpath(i))
413 raise GitException, '%s is not a file or directory' % i
417 GRun('git-update-index', '--add', '--').xargs(files)
418 except GitRunException:
419 raise GitException, 'Unable to add file'
421 def __copy_single(source, target, target2=''):
422 """Copy file or dir named 'source' to name target+target2"""
424 # "source" (file or dir) must match one or more git-controlled file
425 realfiles = GRun('git-ls-files', source).output_lines()
426 if len(realfiles) == 0:
427 raise GitException, '"%s" matches no git-controled files' % source
429 if os.path.isdir(source):
430 # physically copy the files, and record them to add them in one run
432 re_string='^'+source+'/(.*)$'
433 prefix_regexp = re.compile(re_string)
434 for f in [f.strip() for f in realfiles]:
435 m = prefix_regexp.match(f)
437 raise Exception, '"%s" does not match "%s"' % (f, re_string)
438 newname = target+target2+'/'+m.group(1)
439 if not os.path.exists(os.path.dirname(newname)):
440 os.makedirs(os.path.dirname(newname))
442 newfiles.append(newname)
445 else: # files, symlinks, ...
446 newname = target+target2
447 copyfile(source, newname)
451 def copy(filespecs, target):
452 if os.path.isdir(target):
453 # target is a directory: copy each entry on the command line,
454 # with the same name, into the target
455 target = target.rstrip('/')
457 # first, check that none of the children of the target
458 # matching the command line aleady exist
459 for filespec in filespecs:
460 entry = target+ '/' + os.path.basename(filespec.rstrip('/'))
461 if os.path.exists(entry):
462 raise GitException, 'Target "%s" already exists' % entry
464 for filespec in filespecs:
465 filespec = filespec.rstrip('/')
466 basename = '/' + os.path.basename(filespec)
467 __copy_single(filespec, target, basename)
469 elif os.path.exists(target):
470 raise GitException, 'Target "%s" exists but is not a directory' % target
471 elif len(filespecs) != 1:
472 raise GitException, 'Cannot copy more than one file to non-directory'
475 # at this point: len(filespecs)==1 and target does not exist
477 # check target directory
478 targetdir = os.path.dirname(target)
479 if targetdir != '' and not os.path.isdir(targetdir):
480 raise GitException, 'Target directory "%s" does not exist' % targetdir
482 __copy_single(filespecs[0].rstrip('/'), target)
485 def rm(files, force = False):
486 """Remove a file from the repository
490 if os.path.exists(f):
491 raise GitException, '%s exists. Remove it first' %f
493 GRun('git-update-index', '--remove', '--').xargs(files)
496 GRun('git-update-index', '--force-remove', '--').xargs(files)
504 """Return the user information.
508 name=config.get('user.name')
509 email=config.get('user.email')
510 __user = Person(name, email)
514 """Return the author information.
519 # the environment variables take priority over config
521 date = os.environ['GIT_AUTHOR_DATE']
524 __author = Person(os.environ['GIT_AUTHOR_NAME'],
525 os.environ['GIT_AUTHOR_EMAIL'],
532 """Return the author information.
537 # the environment variables take priority over config
539 date = os.environ['GIT_COMMITTER_DATE']
542 __committer = Person(os.environ['GIT_COMMITTER_NAME'],
543 os.environ['GIT_COMMITTER_EMAIL'],
549 def update_cache(files = None, force = False):
550 """Update the cache information for the given files
552 cache_files = tree_status(files, verbose = False)
554 # everything is up-to-date
555 if len(cache_files) == 0:
558 # check for unresolved conflicts
559 if not force and [x for x in cache_files
560 if x[0] not in ['M', 'N', 'A', 'D']]:
561 raise GitException, 'Updating cache failed: unresolved conflicts'
564 add_files = [x[1] for x in cache_files if x[0] in ['N', 'A']]
565 rm_files = [x[1] for x in cache_files if x[0] in ['D']]
566 m_files = [x[1] for x in cache_files if x[0] in ['M']]
568 GRun('git-update-index', '--add', '--').xargs(add_files)
569 GRun('git-update-index', '--force-remove', '--').xargs(rm_files)
570 GRun('git-update-index', '--').xargs(m_files)
574 def commit(message, files = None, parents = None, allowempty = False,
575 cache_update = True, tree_id = None, set_head = False,
576 author_name = None, author_email = None, author_date = None,
577 committer_name = None, committer_email = None):
578 """Commit the current tree to repository
583 # Get the tree status
584 if cache_update and parents != []:
585 changes = update_cache(files)
586 if not changes and not allowempty:
587 raise GitException, 'No changes to commit'
589 # get the commit message
592 elif message[-1:] != '\n':
595 # write the index to repository
597 tree_id = GRun('git-write-tree').output_one_line()
603 env['GIT_AUTHOR_NAME'] = author_name
605 env['GIT_AUTHOR_EMAIL'] = author_email
607 env['GIT_AUTHOR_DATE'] = author_date
609 env['GIT_COMMITTER_NAME'] = committer_name
611 env['GIT_COMMITTER_EMAIL'] = committer_email
612 commit_id = GRun('git-commit-tree', tree_id,
613 *sum([['-p', p] for p in parents], [])
614 ).env(env).raw_input(message).output_one_line()
616 __set_head(commit_id)
620 def apply_diff(rev1, rev2, check_index = True, files = None):
621 """Apply the diff between rev1 and rev2 onto the current
622 index. This function doesn't need to raise an exception since it
623 is only used for fast-pushing a patch. If this operation fails,
624 the pushing would fall back to the three-way merge.
627 index_opt = ['--index']
634 diff_str = diff(files, rev1, rev2)
637 GRun('git-apply', *index_opt).raw_input(
638 diff_str).discard_stderr().no_output()
639 except GitRunException:
644 def merge(base, head1, head2, recursive = False):
645 """Perform a 3-way merge between base, head1 and head2 into the
652 # this operation tracks renames but it is slower (used in
653 # general when pushing or picking patches)
655 # discard output to mask the verbose prints of the tool
656 GRun('git-merge-recursive', base, '--', head1, head2
658 except GitRunException, ex:
662 # the fast case where we don't track renames (used when the
663 # distance between base and heads is small, i.e. folding or
664 # synchronising patches)
666 GRun('git-read-tree', '-u', '-m', '--aggressive',
667 base, head1, head2).run()
668 except GitRunException:
669 raise GitException, 'git-read-tree failed (local changes maybe?)'
671 # check the index for unmerged entries
673 stages_re = re.compile('^([0-7]+) ([0-9a-f]{40}) ([1-3])\t(.*)$', re.S)
675 for line in GRun('git-ls-files', '--unmerged', '--stage', '-z'
676 ).raw_output().split('\0'):
680 mode, hash, stage, path = stages_re.findall(line)[0]
682 if not path in files:
684 files[path]['1'] = ('', '')
685 files[path]['2'] = ('', '')
686 files[path]['3'] = ('', '')
688 files[path][stage] = (mode, hash)
690 if err_output and not files:
691 # if no unmerged files, there was probably a different type of
692 # error and we have to abort the merge
693 raise GitException, err_output
695 # merge the unmerged files
698 # remove additional files that might be generated for some
699 # newer versions of GIT
700 for suffix in [base, head1, head2]:
703 fname = path + '~' + suffix
704 if os.path.exists(fname):
708 if gitmergeonefile.merge(stages['1'][1], stages['2'][1],
709 stages['3'][1], path, stages['1'][0],
710 stages['2'][0], stages['3'][0]) != 0:
714 raise GitException, 'GIT index merging failed (possible conflicts)'
716 def diff(files = None, rev1 = 'HEAD', rev2 = None, diff_flags = []):
717 """Show the diff between rev1 and rev2
723 return GRun('git-diff-tree', '-p',
724 *(diff_flags + [rev1, rev2, '--'] + files)).raw_output()
728 return GRun('git-diff-index', '-p', '-R',
729 *(diff_flags + [rev2, '--'] + files)).raw_output()
731 return GRun('git-diff-index', '-p',
732 *(diff_flags + [rev1, '--'] + files)).raw_output()
736 # TODO: take another parameter representing a diff string as we
737 # usually invoke git.diff() form the calling functions
738 def diffstat(files = None, rev1 = 'HEAD', rev2 = None):
739 """Return the diffstat between rev1 and rev2."""
740 return GRun('git-apply', '--stat', '--summary'
741 ).raw_input(diff(files, rev1, rev2)).raw_output()
743 def files(rev1, rev2, diff_flags = []):
744 """Return the files modified between rev1 and rev2
748 for line in GRun('git-diff-tree', *(diff_flags + ['-r', rev1, rev2])
750 result.append('%s %s' % tuple(line.split(' ', 4)[-1].split('\t', 1)))
752 return '\n'.join(result)
754 def barefiles(rev1, rev2):
755 """Return the files modified between rev1 and rev2, without status info
759 for line in GRun('git-diff-tree', '-r', rev1, rev2).output_lines():
760 result.append(line.split(' ', 4)[-1].split('\t', 1)[-1])
762 return '\n'.join(result)
764 def pretty_commit(commit_id = 'HEAD', diff_flags = []):
765 """Return a given commit (log + diff)
767 return GRun('git-diff-tree',
769 + ['--cc', '--always', '--pretty', '-r', commit_id])
772 def checkout(files = None, tree_id = None, force = False):
773 """Check out the given or all files
777 GRun('git-read-tree', '--reset', tree_id).run()
778 except GitRunException:
779 raise GitException, 'Failed git-read-tree --reset %s' % tree_id
781 cmd = ['git-checkout-index', '-q', '-u']
785 GRun(*(cmd + ['--'])).xargs(files)
787 GRun(*(cmd + ['-a'])).run()
789 def switch(tree_id, keep = False):
790 """Switch the tree to the given id
795 GRun('git-read-tree', '-u', '-m', get_head(), tree_id).run()
796 except GitRunException:
797 raise GitException, 'git-read-tree failed (local changes maybe?)'
801 def reset(files = None, tree_id = None, check_out = True):
802 """Revert the tree changes relative to the given tree_id. It removes
809 cache_files = tree_status(files, tree_id)
810 # files which were added but need to be removed
811 rm_files = [x[1] for x in cache_files if x[0] in ['A']]
813 checkout(files, tree_id, True)
814 # checkout doesn't remove files
815 map(os.remove, rm_files)
817 # if the reset refers to the whole tree, switch the HEAD as well
821 def fetch(repository = 'origin', refspec = None):
822 """Fetches changes from the remote repository, using 'git-fetch'
832 command = config.get('branch.%s.stgit.fetchcmd' % get_head_file()) or \
833 config.get('stgit.fetchcmd')
834 GRun(*(command.split() + args)).run()
836 def pull(repository = 'origin', refspec = None):
837 """Fetches changes from the remote repository, using 'git-pull'
847 command = config.get('branch.%s.stgit.pullcmd' % get_head_file()) or \
848 config.get('stgit.pullcmd')
849 GRun(*(command.split() + args)).run()
851 def rebase(tree_id = None):
852 """Rebase the current tree to the give tree_id. The tree_id
853 argument may be something other than a GIT id if an external
856 command = config.get('branch.%s.stgit.rebasecmd' % get_head_file()) \
857 or config.get('stgit.rebasecmd')
863 raise GitException, 'Default rebasing requires a commit id'
865 # clear the HEAD cache as the custom rebase command will update it
867 GRun(*(command.split() + args)).run()
870 reset(tree_id = tree_id)
873 """Repack all objects into a single pack
875 GRun('git-repack', '-a', '-d', '-f').run()
877 def apply_patch(filename = None, diff = None, base = None,
879 """Apply a patch onto the current or given index. There must not
880 be any local changes in the tree, otherwise the command fails
892 orig_head = get_head()
898 GRun('git-apply', '--index').raw_input(diff).no_output()
899 except GitRunException:
903 # write the failed diff to a file
904 f = file('.stgit-failed.patch', 'w+')
907 out.warn('Diff written to the .stgit-failed.patch file')
912 top = commit(message = 'temporary commit used for applying a patch',
915 merge(base, orig_head, top)
917 def clone(repository, local_dir):
918 """Clone a remote repository. At the moment, just use the
921 GRun('git-clone', repository, local_dir).run()
923 def modifying_revs(files, base_rev, head_rev):
924 """Return the revisions from the list modifying the given files."""
925 return GRun('git-rev-list', '%s..%s' % (base_rev, head_rev), '--', *files
928 def refspec_localpart(refspec):
929 m = re.match('^[^:]*:([^:]*)$', refspec)
933 raise GitException, 'Cannot parse refspec "%s"' % line
935 def refspec_remotepart(refspec):
936 m = re.match('^([^:]*):[^:]*$', refspec)
940 raise GitException, 'Cannot parse refspec "%s"' % line
943 def __remotes_from_config():
944 return config.sections_matching(r'remote\.(.*)\.url')
946 def __remotes_from_dir(dir):
947 d = os.path.join(basedir.get(), dir)
948 if os.path.exists(d):
954 """Return the list of remotes in the repository
956 return (set(__remotes_from_config())
957 | set(__remotes_from_dir('remotes'))
958 | set(__remotes_from_dir('branches')))
960 def remotes_local_branches(remote):
961 """Returns the list of local branches fetched from given remote
965 if remote in __remotes_from_config():
966 for line in config.getall('remote.%s.fetch' % remote):
967 branches.append(refspec_localpart(line))
968 elif remote in __remotes_from_dir('remotes'):
969 stream = open(os.path.join(basedir.get(), 'remotes', remote), 'r')
971 # Only consider Pull lines
972 m = re.match('^Pull: (.*)\n$', line)
974 branches.append(refspec_localpart(m.group(1)))
976 elif remote in __remotes_from_dir('branches'):
977 # old-style branches only declare one branch
978 branches.append('refs/heads/'+remote);
980 raise GitException, 'Unknown remote "%s"' % remote
984 def identify_remote(branchname):
985 """Return the name for the remote to pull the given branchname
986 from, or None if we believe it is a local branch.
989 for remote in remotes_list():
990 if branchname in remotes_local_branches(remote):
993 # if we get here we've found nothing, the branch is a local one
997 """Return the git id for the tip of the parent branch as left by
1002 stream = open(os.path.join(basedir.get(), 'FETCH_HEAD'), "r")
1004 # Only consider lines not tagged not-for-merge
1005 m = re.match('^([^\t]*)\t\t', line)
1008 raise GitException, 'StGit does not support multiple FETCH_HEAD'
1010 fetch_head=m.group(1)
1014 raise GitException, 'No for-merge remote head found in FETCH_HEAD'
1016 # here we are sure to have a single fetch_head
1020 """Return a list of all refs in the current repository.
1023 return [line.split()[1] for line in GRun('git-show-ref').output_lines()]