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
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
):
41 def __init__(self
, *cmd
):
42 """Initialise the Run object and insert the 'git' command name.
44 Run
.__init__(self
, 'git', *cmd
)
51 """An author, committer, etc."""
52 def __init__(self
, name
= None, email
= None, date
= '',
54 self
.name
= self
.email
= self
.date
= None
55 if name
or email
or date
:
61 assert not (name
or email
or date
)
63 m
= re
.match(r
'^(.+)<(.+)>(.*)$', s
)
65 return [x
.strip() or None for x
in m
.groups()]
66 self
.name
, self
.email
, self
.date
= parse_desc(desc
)
67 def set_name(self
, val
):
70 def set_email(self
, val
):
73 def set_date(self
, val
):
77 if self
.name
and self
.email
:
78 return '%s <%s>' %
(self
.name
, self
.email
)
80 raise GitException
, 'not enough identity data'
83 """Handle the commit objects
85 def __init__(self
, id_hash
):
86 self
.__id_hash
= id_hash
88 lines
= GRun('cat-file', 'commit', id_hash
).output_lines()
89 for i
in range(len(lines
)):
92 break # we've seen all the header fields
93 key
, val
= line
.split(' ', 1)
98 elif key
== 'committer':
99 self
.__committer
= val
101 pass # ignore other headers
102 self
.__log
= '\n'.join(lines
[i
+1:])
104 def get_id_hash(self
):
105 return self
.__id_hash
110 def get_parent(self
):
111 parents
= self
.get_parents()
117 def get_parents(self
):
118 return GRun('rev-list', '--parents', '--max-count=1', self
.__id_hash
119 ).output_one_line().split()[1:]
121 def get_author(self
):
124 def get_committer(self
):
125 return self
.__committer
131 return self
.get_id_hash()
133 # dictionary of Commit objects, used to avoid multiple calls to git
140 def get_commit(id_hash
):
141 """Commit objects factory. Save/look-up them in the __commits
146 if id_hash
in __commits
:
147 return __commits
[id_hash
]
149 commit
= Commit(id_hash
)
150 __commits
[id_hash
] = commit
154 """Return the list of file conflicts
157 for line
in GRun('ls-files', '-z', '--unmerged'
158 ).raw_output().split('\0')[:-1]:
159 stat
, path
= line
.split('\t', 1)
164 files
= [os
.path
.join(basedir
.get(), 'info', 'exclude')]
165 user_exclude
= config
.get('core.excludesfile')
167 files
.append(user_exclude
)
170 def ls_files(files
, tree
= 'HEAD', full_name
= True):
171 """Return the files known to GIT or raise an error otherwise. It also
172 converts the file to the full path relative the the .git directory.
179 args
.append('--with-tree=%s' % tree
)
181 args
.append('--full-name')
185 # use a set to avoid file names duplication due to different stages
186 fileset
= set(GRun('ls-files', '--error-unmatch', *args
).output_lines())
187 except GitRunException
:
188 # just hide the details of the 'git ls-files' command we use
189 raise GitException
, \
190 'Some of the given paths are either missing or not known to GIT'
193 def parse_git_ls(output
):
194 """Parse the output of git diff-index, diff-files, etc. Doesn't handle
195 rename/copy output, so don't feed it output generated with the -M
198 for line
in output
.split('\0'):
200 # There's a zero byte at the end of the output, which
201 # gives us an empty string as the last "line".
204 mode_a
, mode_b
, sha1_a
, sha1_b
, t
= line
.split(' ')
209 def tree_status(files
= None, tree_id
= 'HEAD', unknown
= False,
210 noexclude
= True, verbose
= False):
211 """Get the status of all changed files, or of a selected set of
212 files. Returns a list of pairs - (status, filename).
214 If 'not files', it will check all files, and optionally all
215 unknown files. If 'files' is a list, it will only check the files
218 assert not files
or not unknown
221 out
.start('Checking for changes in the working directory')
231 cmd
= ['ls-files', '-z', '--others', '--directory',
232 '--no-empty-directory']
234 cmd
+= ['--exclude=%s' % s for s in
235 ['*.[ao]', '*.pyc', '.*', '*~', '#*', 'TAGS', 'tags']]
236 cmd
+= ['--exclude-per-directory=.gitignore']
237 cmd
+= ['--exclude-from=%s' % fn
238 for fn
in exclude_files()
239 if os
.path
.exists(fn
)]
241 lines
= GRun(*cmd
).raw_output().split('\0')
242 cache_files
+= [('?', line
) for line
in lines
if line
]
245 conflicts
= get_conflicts()
246 cache_files
+= [('C', filename
) for filename
in conflicts
247 if not files
or filename
in files
]
248 reported_files
= set(conflicts
)
249 files_left
= [f
for f
in files
if f
not in reported_files
]
251 # files in the index. Only execute this code if no files were
252 # specified when calling the function (i.e. report all files) or
253 # files were specified but already found in the previous step
254 if not files
or files_left
:
257 args
+= ['--'] + files_left
258 for t
, fn
in parse_git_ls(GRun('diff-index', '-z', *args
).raw_output()):
259 # the condition is needed in case files is emtpy and
260 # diff-index lists those already reported
261 if not fn
in reported_files
:
262 cache_files
.append((t
, fn
))
263 reported_files
.add(fn
)
264 files_left
= [f
for f
in files
if f
not in reported_files
]
266 # files in the index but changed on (or removed from) disk. Only
267 # execute this code if no files were specified when calling the
268 # function (i.e. report all files) or files were specified but
269 # already found in the previous step
270 if not files
or files_left
:
273 args
+= ['--'] + files_left
274 for t
, fn
in parse_git_ls(GRun('diff-files', '-z', *args
).raw_output()):
275 # the condition is needed in case files is empty and
276 # diff-files lists those already reported
277 if not fn
in reported_files
:
278 cache_files
.append((t
, fn
))
279 reported_files
.add(fn
)
286 def local_changes(verbose
= True):
287 """Return true if there are local changes in the tree
289 return len(tree_status(verbose
= verbose
)) != 0
293 hr
= re
.compile(r
'^[0-9a-f]{40} refs/heads/(.+)$')
294 for line
in GRun('show-ref', '--heads').output_lines():
296 heads
.append(m
.group(1))
303 """Verifies the HEAD and returns the SHA1 id that represents it
308 __head
= rev_parse('HEAD')
311 class DetachedHeadException(GitException
):
313 GitException
.__init__(self
, 'Not on any branch')
316 """Return the name of the file pointed to by the HEAD symref.
317 Throw an exception if HEAD is detached."""
320 'refs/heads/', GRun('symbolic-ref', '-q', 'HEAD'
322 except GitRunException
:
323 raise DetachedHeadException()
325 def set_head_file(ref
):
326 """Resets HEAD to point to a new ref
328 # head cache flushing is needed since we might have a different value
332 GRun('symbolic-ref', 'HEAD', 'refs/heads/%s' % ref
).run()
333 except GitRunException
:
334 raise GitException
, 'Could not set head to "%s"' % ref
336 def set_ref(ref
, val
):
337 """Point ref at a new commit object."""
339 GRun('update-ref', ref
, val
).run()
340 except GitRunException
:
341 raise GitException
, 'Could not update %s to "%s".' %
(ref
, val
)
343 def set_branch(branch
, val
):
344 set_ref('refs/heads/%s' % branch
, val
)
347 """Sets the HEAD value
351 if not __head
or __head
!= val
:
355 # only allow SHA1 hashes
356 assert(len(__head
) == 40)
358 def __clear_head_cache():
359 """Sets the __head to None so that a re-read is forced
366 """Refresh index with stat() information from the working directory.
368 GRun('update-index', '-q', '--unmerged', '--refresh').run()
370 def rev_parse(git_id
):
371 """Parse the string and return a verified SHA1 id
374 return GRun('rev-parse', '--verify', git_id
375 ).discard_stderr().output_one_line()
376 except GitRunException
:
377 raise GitException
, 'Unknown revision: %s' % git_id
386 def branch_exists(branch
):
387 return ref_exists('refs/heads/%s' % branch
)
389 def create_branch(new_branch
, tree_id
= None):
390 """Create a new branch in the git repository
392 if branch_exists(new_branch
):
393 raise GitException
, 'Branch "%s" already exists' % new_branch
395 current_head_file
= get_head_file()
396 current_head
= get_head()
397 set_head_file(new_branch
)
398 __set_head(current_head
)
400 # a checkout isn't needed if new branch points to the current head
405 # Tree switching failed. Revert the head file
406 set_head_file(current_head_file
)
407 delete_branch(new_branch
)
410 if os
.path
.isfile(os
.path
.join(basedir
.get(), 'MERGE_HEAD')):
411 os
.remove(os
.path
.join(basedir
.get(), 'MERGE_HEAD'))
413 def switch_branch(new_branch
):
414 """Switch to a git branch
418 if not branch_exists(new_branch
):
419 raise GitException
, 'Branch "%s" does not exist' % new_branch
421 tree_id
= rev_parse('refs/heads/%s^{commit}' % new_branch
)
422 if tree_id
!= get_head():
425 GRun('read-tree', '-u', '-m', get_head(), tree_id
).run()
426 except GitRunException
:
427 raise GitException
, 'read-tree failed (local changes maybe?)'
429 set_head_file(new_branch
)
431 if os
.path
.isfile(os
.path
.join(basedir
.get(), 'MERGE_HEAD')):
432 os
.remove(os
.path
.join(basedir
.get(), 'MERGE_HEAD'))
435 if not ref_exists(ref
):
436 raise GitException
, '%s does not exist' % ref
437 sha1
= GRun('show-ref', '-s', ref
).output_one_line()
439 GRun('update-ref', '-d', ref
, sha1
).run()
440 except GitRunException
:
441 raise GitException
, 'Failed to delete ref %s' % ref
443 def delete_branch(name
):
444 delete_ref('refs/heads/%s' % name
)
446 def rename_ref(from_ref
, to_ref
):
447 if not ref_exists(from_ref
):
448 raise GitException
, '"%s" does not exist' % from_ref
449 if ref_exists(to_ref
):
450 raise GitException
, '"%s" already exists' % to_ref
452 sha1
= GRun('show-ref', '-s', from_ref
).output_one_line()
454 GRun('update-ref', to_ref
, sha1
, '0'*40).run()
455 except GitRunException
:
456 raise GitException
, 'Failed to create new ref %s' % to_ref
458 GRun('update-ref', '-d', from_ref
, sha1
).run()
459 except GitRunException
:
460 raise GitException
, 'Failed to delete ref %s' % from_ref
462 def rename_branch(from_name
, to_name
):
463 """Rename a git branch."""
464 rename_ref('refs/heads/%s' % from_name
, 'refs/heads/%s' % to_name
)
466 if get_head_file() == from_name
:
467 set_head_file(to_name
)
468 except DetachedHeadException
:
469 pass # detached HEAD, so the renamee can't be the current branch
470 reflog_dir
= os
.path
.join(basedir
.get(), 'logs', 'refs', 'heads')
471 if os
.path
.exists(reflog_dir
) \
472 and os
.path
.exists(os
.path
.join(reflog_dir
, from_name
)):
473 rename(reflog_dir
, from_name
, to_name
)
481 """Return the user information.
485 name
=config
.get('user.name')
486 email
=config
.get('user.email')
487 __user
= Person(name
, email
)
491 """Return the author information.
496 # the environment variables take priority over config
498 date
= os
.environ
['GIT_AUTHOR_DATE']
501 __author
= Person(os
.environ
['GIT_AUTHOR_NAME'],
502 os
.environ
['GIT_AUTHOR_EMAIL'],
509 """Return the author information.
514 # the environment variables take priority over config
516 date
= os
.environ
['GIT_COMMITTER_DATE']
519 __committer
= Person(os
.environ
['GIT_COMMITTER_NAME'],
520 os
.environ
['GIT_COMMITTER_EMAIL'],
526 def update_cache(files
= None, force
= False):
527 """Update the cache information for the given files
529 cache_files
= tree_status(files
, verbose
= False)
531 # everything is up-to-date
532 if len(cache_files
) == 0:
535 # check for unresolved conflicts
536 if not force
and [x
for x
in cache_files
537 if x
[0] not in ['M', 'N', 'A', 'D']]:
538 raise GitException
, 'Updating cache failed: unresolved conflicts'
541 add_files
= [x
[1] for x
in cache_files
if x
[0] in ['N', 'A']]
542 rm_files
= [x
[1] for x
in cache_files
if x
[0] in ['D']]
543 m_files
= [x
[1] for x
in cache_files
if x
[0] in ['M']]
545 GRun('update-index', '--add', '--').xargs(add_files
)
546 GRun('update-index', '--force-remove', '--').xargs(rm_files
)
547 GRun('update-index', '--').xargs(m_files
)
551 def commit(message
, files
= None, parents
= None, allowempty
= False,
552 cache_update
= True, tree_id
= None, set_head
= False,
553 author_name
= None, author_email
= None, author_date
= None,
554 committer_name
= None, committer_email
= None):
555 """Commit the current tree to repository
560 # Get the tree status
561 if cache_update
and parents
!= []:
562 changes
= update_cache(files
)
563 if not changes
and not allowempty
:
564 raise GitException
, 'No changes to commit'
566 # get the commit message
569 elif message
[-1:] != '\n':
572 # write the index to repository
574 tree_id
= GRun('write-tree').output_one_line()
580 env
['GIT_AUTHOR_NAME'] = author_name
582 env
['GIT_AUTHOR_EMAIL'] = author_email
584 env
['GIT_AUTHOR_DATE'] = author_date
586 env
['GIT_COMMITTER_NAME'] = committer_name
588 env
['GIT_COMMITTER_EMAIL'] = committer_email
589 commit_id
= GRun('commit-tree', tree_id
,
590 *sum([['-p', p
] for p
in parents
], [])
591 ).env(env
).raw_input(message
).output_one_line()
593 __set_head(commit_id
)
597 def apply_diff(rev1
, rev2
, check_index
= True, files
= None):
598 """Apply the diff between rev1 and rev2 onto the current
599 index. This function doesn't need to raise an exception since it
600 is only used for fast-pushing a patch. If this operation fails,
601 the pushing would fall back to the three-way merge.
604 index_opt
= ['--index']
611 diff_str
= diff(files
, rev1
, rev2
)
614 GRun('apply', *index_opt
).raw_input(
615 diff_str
).discard_stderr().no_output()
616 except GitRunException
:
621 stages_re
= re
.compile('^([0-7]+) ([0-9a-f]{40}) ([1-3])\t(.*)$', re
.S
)
623 def merge_recursive(base
, head1
, head2
):
624 """Perform a 3-way merge between base, head1 and head2 into the
628 p
= GRun('merge-recursive', base
, '--', head1
, head2
).env(
629 { 'GITHEAD_%s' % base
: 'ancestor',
630 'GITHEAD_%s' % head1
: 'current',
631 'GITHEAD_%s' % head2
: 'patched'}).returns([0, 1])
632 output
= p
.output_lines()
634 # There were conflicts
635 if config
.get('stgit.autoimerge') == 'yes':
638 conflicts
= [l
for l
in output
if l
.startswith('CONFLICT')]
640 raise GitException
, "%d conflict(s)" %
len(conflicts
)
642 def mergetool(files
= ()):
643 """Invoke 'git mergetool' to resolve any outstanding conflicts. If 'not
644 files', all the files in an unmerged state will be processed."""
645 GRun('mergetool', *list(files
)).returns([0, 1]).run()
646 # check for unmerged entries (prepend 'CONFLICT ' for consistency with
648 conflicts
= ['CONFLICT ' + f
for f
in get_conflicts()]
651 raise GitException
, "%d conflict(s)" %
len(conflicts
)
653 def diff(files
= None, rev1
= 'HEAD', rev2
= None, diff_flags
= [],
655 """Show the diff between rev1 and rev2
659 if binary
and '--binary' not in diff_flags
:
660 diff_flags
= diff_flags
+ ['--binary']
663 return GRun('diff-tree', '-p',
664 *(diff_flags
+ [rev1
, rev2
, '--'] + files
)).raw_output()
668 return GRun('diff-index', '-p', '-R',
669 *(diff_flags
+ [rev2
, '--'] + files
)).raw_output()
671 return GRun('diff-index', '-p',
672 *(diff_flags
+ [rev1
, '--'] + files
)).raw_output()
676 def files(rev1
, rev2
, diff_flags
= []):
677 """Return the files modified between rev1 and rev2
681 for line
in GRun('diff-tree', *(diff_flags
+ ['-r', rev1
, rev2
])
683 result
.append('%s %s' %
tuple(line
.split(' ', 4)[-1].split('\t', 1)))
685 return '\n'.join(result
)
687 def barefiles(rev1
, rev2
):
688 """Return the files modified between rev1 and rev2, without status info
692 for line
in GRun('diff-tree', '-r', rev1
, rev2
).output_lines():
693 result
.append(line
.split(' ', 4)[-1].split('\t', 1)[-1])
695 return '\n'.join(result
)
697 def pretty_commit(commit_id
= 'HEAD', flags
= []):
698 """Return a given commit (log + diff)
700 return GRun('show', *(flags
+ [commit_id
])).raw_output()
702 def checkout(files
= None, tree_id
= None, force
= False):
703 """Check out the given or all files
707 GRun('read-tree', '--reset', tree_id
).run()
708 except GitRunException
:
709 raise GitException
, 'Failed "git read-tree" --reset %s' % tree_id
711 cmd
= ['checkout-index', '-q', '-u']
715 GRun(*(cmd
+ ['--'])).xargs(files
)
717 GRun(*(cmd
+ ['-a'])).run()
719 def switch(tree_id
, keep
= False):
720 """Switch the tree to the given id
723 # only update the index while keeping the local changes
724 GRun('read-tree', tree_id
).run()
728 GRun('read-tree', '-u', '-m', get_head(), tree_id
).run()
729 except GitRunException
:
730 raise GitException
, 'read-tree failed (local changes maybe?)'
734 def reset(files
= None, tree_id
= None, check_out
= True):
735 """Revert the tree changes relative to the given tree_id. It removes
742 cache_files
= tree_status(files
, tree_id
)
743 # files which were added but need to be removed
744 rm_files
= [x
[1] for x
in cache_files
if x
[0] in ['A']]
746 checkout(files
, tree_id
, True)
747 # checkout doesn't remove files
748 map(os
.remove
, rm_files
)
750 # if the reset refers to the whole tree, switch the HEAD as well
754 def resolved(filenames
, reset
= None):
756 stage
= {'ancestor': 1, 'current': 2, 'patched': 3}[reset
]
757 GRun('checkout-index', '--no-create', '--stage=%d' % stage
,
758 '--stdin', '-z').input_nulterm(filenames
).no_output()
759 GRun('update-index', '--add', '--').xargs(filenames
)
760 for filename
in filenames
:
761 # update the access and modificatied times
762 os
.utime(filename
, None)
764 def fetch(repository
= 'origin', refspec
= None):
765 """Fetches changes from the remote repository, using 'git fetch'
775 command
= config
.get('branch.%s.stgit.fetchcmd' % get_head_file
()) or \
776 config
.get('stgit.fetchcmd')
777 Run(*(command
.split() + args
)).run()
779 def pull(repository
= 'origin', refspec
= None):
780 """Fetches changes from the remote repository, using 'git pull'
790 command
= config
.get('branch.%s.stgit.pullcmd' % get_head_file
()) or \
791 config
.get('stgit.pullcmd')
792 Run(*(command
.split() + args
)).run()
794 def rebase(tree_id
= None):
795 """Rebase the current tree to the give tree_id. The tree_id
796 argument may be something other than a GIT id if an external
799 command
= config
.get('branch.%s.stgit.rebasecmd' % get_head_file
()) \
800 or config
.get('stgit.rebasecmd')
806 raise GitException
, 'Default rebasing requires a commit id'
808 # clear the HEAD cache as the custom rebase command will update it
810 Run(*(command
.split() + args
)).run()
813 reset(tree_id
= tree_id
)
816 """Repack all objects into a single pack
818 GRun('repack', '-a', '-d', '-f').run()
820 def apply_patch(filename
= None, diff
= None, base
= None,
821 fail_dump
= True, reject
= False):
822 """Apply a patch onto the current or given index. There must not
823 be any local changes in the tree, otherwise the command fails
835 orig_head
= get_head()
840 cmd
= ['apply', '--index']
844 GRun(*cmd
).raw_input(diff
).no_output()
845 except GitRunException
:
849 # write the failed diff to a file
850 f
= file('.stgit-failed.patch', 'w+')
853 out
.warn('Diff written to the .stgit-failed.patch file')
858 top
= commit(message
= 'temporary commit used for applying a patch',
861 merge_recursive(base
, orig_head
, top
)
863 def clone(repository
, local_dir
):
864 """Clone a remote repository. At the moment, just use the
867 GRun('clone', repository
, local_dir
).run()
869 def modifying_revs(files
, base_rev
, head_rev
):
870 """Return the revisions from the list modifying the given files."""
871 return GRun('rev-list', '%s..%s' %
(base_rev
, head_rev
), '--', *files
874 def refspec_localpart(refspec
):
875 m
= re
.match('^[^:]*:([^:]*)$', refspec
)
879 raise GitException
, 'Cannot parse refspec "%s"' % line
881 def refspec_remotepart(refspec
):
882 m
= re
.match('^([^:]*):[^:]*$', refspec
)
886 raise GitException
, 'Cannot parse refspec "%s"' % line
888 def __remotes_from_config():
889 return config
.sections_matching(r
'remote\.(.*)\.url')
891 def __remotes_from_dir(dir):
892 d
= os
.path
.join(basedir
.get(), dir)
893 if os
.path
.exists(d
):
899 """Return the list of remotes in the repository
901 return (set(__remotes_from_config())
902 |
set(__remotes_from_dir('remotes'))
903 |
set(__remotes_from_dir('branches')))
905 def remotes_local_branches(remote
):
906 """Returns the list of local branches fetched from given remote
910 if remote
in __remotes_from_config():
911 for line
in config
.getall('remote.%s.fetch' % remote
):
912 branches
.append(refspec_localpart(line
))
913 elif remote
in __remotes_from_dir('remotes'):
914 stream
= open(os
.path
.join(basedir
.get(), 'remotes', remote
), 'r')
916 # Only consider Pull lines
917 m
= re
.match('^Pull: (.*)\n$', line
)
919 branches
.append(refspec_localpart(m
.group(1)))
921 elif remote
in __remotes_from_dir('branches'):
922 # old-style branches only declare one branch
923 branches
.append('refs/heads/'+remote
);
925 raise GitException
, 'Unknown remote "%s"' % remote
929 def identify_remote(branchname
):
930 """Return the name for the remote to pull the given branchname
931 from, or None if we believe it is a local branch.
934 for remote
in remotes_list():
935 if branchname
in remotes_local_branches(remote
):
938 # if we get here we've found nothing, the branch is a local one
942 """Return the git id for the tip of the parent branch as left by
947 stream
= open(os
.path
.join(basedir
.get(), 'FETCH_HEAD'), "r")
949 # Only consider lines not tagged not-for-merge
950 m
= re
.match('^([^\t]*)\t\t', line
)
953 raise GitException
, 'StGit does not support multiple FETCH_HEAD'
955 fetch_head
=m
.group(1)
959 out
.warn('No for-merge remote head found in FETCH_HEAD')
961 # here we are sure to have a single fetch_head
965 """Return a list of all refs in the current repository.
968 return [line
.split()[1] for line
in GRun('show-ref').output_lines()]