stg diff / files: don't update directory cache
[stgit] / 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, glob, popen2
22
23 from stgit.utils import *
24
25 # git exception class
26 class GitException(Exception):
27 pass
28
29
30 # Different start-up variables read from the environment
31 if 'GIT_DIR' in os.environ:
32 base_dir = os.environ['GIT_DIR']
33 else:
34 base_dir = '.git'
35
36 head_link = os.path.join(base_dir, 'HEAD')
37
38 #
39 # Classes
40 #
41 class Commit:
42 """Handle the commit objects
43 """
44 def __init__(self, id_hash):
45 self.__id_hash = id_hash
46
47 lines = _output_lines('git-cat-file commit %s' % id_hash)
48 self.__parents = []
49 for i in range(len(lines)):
50 line = lines[i]
51 if line == '\n':
52 break
53 field = line.strip().split(' ', 1)
54 if field[0] == 'tree':
55 self.__tree = field[1]
56 elif field[0] == 'parent':
57 self.__parents.append(field[1])
58 if field[0] == 'author':
59 self.__author = field[1]
60 if field[0] == 'committer':
61 self.__committer = field[1]
62 self.__log = ''.join(lines[i:])
63
64 def get_id_hash(self):
65 return self.__id_hash
66
67 def get_tree(self):
68 return self.__tree
69
70 def get_parent(self):
71 return self.__parents[0]
72
73 def get_parents(self):
74 return self.__parents
75
76 def get_author(self):
77 return self.__author
78
79 def get_committer(self):
80 return self.__committer
81
82 def get_log(self):
83 return self.__log
84
85 # dictionary of Commit objects, used to avoid multiple calls to git
86 __commits = dict()
87
88 #
89 # Functions
90 #
91 def get_commit(id_hash):
92 """Commit objects factory. Save/look-up them in the __commits
93 dictionary
94 """
95 if id_hash in __commits:
96 return __commits[id_hash]
97 else:
98 commit = Commit(id_hash)
99 __commits[id_hash] = commit
100 return commit
101
102 def get_conflicts():
103 """Return the list of file conflicts
104 """
105 conflicts_file = os.path.join(base_dir, 'conflicts')
106 if os.path.isfile(conflicts_file):
107 f = file(conflicts_file)
108 names = [line.strip() for line in f.readlines()]
109 f.close()
110 return names
111 else:
112 return None
113
114 def _input(cmd, file_desc):
115 p = popen2.Popen3(cmd)
116 for line in file_desc:
117 p.tochild.write(line)
118 p.tochild.close()
119 if p.wait():
120 raise GitException, '%s failed' % str(cmd)
121
122 def _output(cmd):
123 p=popen2.Popen3(cmd)
124 string = p.fromchild.read()
125 if p.wait():
126 raise GitException, '%s failed' % str(cmd)
127 return string
128
129 def _output_one_line(cmd):
130 p=popen2.Popen3(cmd)
131 string = p.fromchild.readline().strip()
132 if p.wait():
133 raise GitException, '%s failed' % str(cmd)
134 return string
135
136 def _output_lines(cmd):
137 p=popen2.Popen3(cmd)
138 lines = p.fromchild.readlines()
139 if p.wait():
140 raise GitException, '%s failed' % str(cmd)
141 return lines
142
143 def __run(cmd, args=None):
144 """__run: runs cmd using spawnvp.
145
146 Runs cmd using spawnvp. The shell is avoided so it won't mess up
147 our arguments. If args is very large, the command is run multiple
148 times; args is split xargs style: cmd is passed on each
149 invocation. Unlike xargs, returns immediately if any non-zero
150 return code is received.
151 """
152
153 args_l=cmd.split()
154 if args is None:
155 args = []
156 for i in range(0, len(args)+1, 100):
157 r=os.spawnvp(os.P_WAIT, args_l[0], args_l + args[i:min(i+100, len(args))])
158 if r:
159 return r
160 return 0
161
162 def __check_base_dir():
163 return os.path.isdir(base_dir)
164
165 def __tree_status(files = [], tree_id = 'HEAD', unknown = False,
166 noexclude = True):
167 """Returns a list of pairs - [status, filename]
168 """
169 os.system('git-update-cache --refresh > /dev/null')
170
171 cache_files = []
172
173 # unknown files
174 if unknown:
175 exclude_file = os.path.join(base_dir, 'info', 'exclude')
176 base_exclude = ['--exclude=%s' % s for s in
177 ['*.[ao]', '*.pyc', '.*', '*~', '#*', 'TAGS', 'tags']]
178 base_exclude.append('--exclude-per-directory=.gitignore')
179
180 if os.path.exists(exclude_file):
181 extra_exclude = '--exclude-from=%s' % exclude_file
182 else:
183 extra_exclude = []
184 if noexclude:
185 extra_exclude = base_exclude = []
186
187 lines = _output_lines(['git-ls-files', '--others'] + base_exclude
188 + extra_exclude)
189 cache_files += [('?', line.strip()) for line in lines]
190
191 # conflicted files
192 conflicts = get_conflicts()
193 if not conflicts:
194 conflicts = []
195 cache_files += [('C', filename) for filename in conflicts]
196
197 # the rest
198 for line in _output_lines(['git-diff-cache', '-r', tree_id] + files):
199 fs = tuple(line.rstrip().split(' ',4)[-1].split('\t',1))
200 if fs[1] not in conflicts:
201 cache_files.append(fs)
202
203 return cache_files
204
205 def local_changes():
206 """Return true if there are local changes in the tree
207 """
208 return len(__tree_status()) != 0
209
210 def get_head():
211 """Returns a string representing the HEAD
212 """
213 return read_string(head_link)
214
215 def get_head_file():
216 """Returns the name of the file pointed to by the HEAD link
217 """
218 # valid link
219 if os.path.islink(head_link) and os.path.isfile(head_link):
220 return os.path.basename(os.readlink(head_link))
221 else:
222 raise GitException, 'Invalid .git/HEAD link. Git tree not initialised?'
223
224 def __set_head(val):
225 """Sets the HEAD value
226 """
227 write_string(head_link, val)
228
229 def add(names):
230 """Add the files or recursively add the directory contents
231 """
232 # generate the file list
233 files = []
234 for i in names:
235 if not os.path.exists(i):
236 raise GitException, 'Unknown file or directory: %s' % i
237
238 if os.path.isdir(i):
239 # recursive search. We only add files
240 for root, dirs, local_files in os.walk(i):
241 for name in [os.path.join(root, f) for f in local_files]:
242 if os.path.isfile(name):
243 files.append(os.path.normpath(name))
244 elif os.path.isfile(i):
245 files.append(os.path.normpath(i))
246 else:
247 raise GitException, '%s is not a file or directory' % i
248
249 if files:
250 if __run('git-update-cache --add --', files):
251 raise GitException, 'Unable to add file'
252
253 def rm(files, force = False):
254 """Remove a file from the repository
255 """
256 if force:
257 git_opt = '--force-remove'
258 else:
259 git_opt = '--remove'
260
261 if not force:
262 for f in files:
263 if os.path.exists(f):
264 raise GitException, '%s exists. Remove it first' %f
265 if files:
266 __run('git-update-cache --remove --', files)
267 else:
268 if files:
269 __run('git-update-cache --force-remove --', files)
270
271 def update_cache(files = [], force = False):
272 """Update the cache information for the given files
273 """
274 cache_files = __tree_status(files)
275
276 # everything is up-to-date
277 if len(cache_files) == 0:
278 return False
279
280 # check for unresolved conflicts
281 if not force and [x for x in cache_files
282 if x[0] not in ['M', 'N', 'A', 'D']]:
283 raise GitException, 'Updating cache failed: unresolved conflicts'
284
285 # update the cache
286 add_files = [x[1] for x in cache_files if x[0] in ['N', 'A']]
287 rm_files = [x[1] for x in cache_files if x[0] in ['D']]
288 m_files = [x[1] for x in cache_files if x[0] in ['M']]
289
290 if add_files and __run('git-update-cache --add --', add_files) != 0:
291 raise GitException, 'Failed git-update-cache --add'
292 if rm_files and __run('git-update-cache --force-remove --', rm_files) != 0:
293 raise GitException, 'Failed git-update-cache --rm'
294 if m_files and __run('git-update-cache --', m_files) != 0:
295 raise GitException, 'Failed git-update-cache'
296
297 return True
298
299 def commit(message, files = [], parents = [], allowempty = False,
300 cache_update = True,
301 author_name = None, author_email = None, author_date = None,
302 committer_name = None, committer_email = None):
303 """Commit the current tree to repository
304 """
305 # Get the tree status
306 if cache_update and parents != []:
307 changes = update_cache(files)
308 if not changes and not allowempty:
309 raise GitException, 'No changes to commit'
310
311 # get the commit message
312 f = file('.commitmsg', 'w+')
313 if message[-1:] == '\n':
314 f.write(message)
315 else:
316 print >> f, message
317 f.close()
318
319 # write the index to repository
320 tree_id = _output_one_line('git-write-tree')
321
322 # the commit
323 cmd = ''
324 if author_name:
325 cmd += 'GIT_AUTHOR_NAME="%s" ' % author_name
326 if author_email:
327 cmd += 'GIT_AUTHOR_EMAIL="%s" ' % author_email
328 if author_date:
329 cmd += 'GIT_AUTHOR_DATE="%s" ' % author_date
330 if committer_name:
331 cmd += 'GIT_COMMITTER_NAME="%s" ' % committer_name
332 if committer_email:
333 cmd += 'GIT_COMMITTER_EMAIL="%s" ' % committer_email
334 cmd += 'git-commit-tree %s' % tree_id
335
336 # get the parents
337 for p in parents:
338 cmd += ' -p %s' % p
339
340 cmd += ' < .commitmsg'
341
342 commit_id = _output_one_line(cmd)
343 __set_head(commit_id)
344 os.remove('.commitmsg')
345
346 return commit_id
347
348 def merge(base, head1, head2):
349 """Perform a 3-way merge between base, head1 and head2 into the
350 local tree
351 """
352 if __run('git-read-tree -u -m', [base, head1, head2]) != 0:
353 raise GitException, 'git-read-tree failed (local changes maybe?)'
354
355 # this can fail if there are conflicts
356 if os.system('git-merge-cache -o -q gitmergeonefile.py -a') != 0:
357 raise GitException, 'git-merge-cache failed (possible conflicts)'
358
359 def status(files = [], modified = False, new = False, deleted = False,
360 conflict = False, unknown = False, noexclude = False):
361 """Show the tree status
362 """
363 cache_files = __tree_status(files, unknown = True, noexclude = noexclude)
364 all = not (modified or new or deleted or conflict or unknown)
365
366 if not all:
367 filestat = []
368 if modified:
369 filestat.append('M')
370 if new:
371 filestat.append('A')
372 filestat.append('N')
373 if deleted:
374 filestat.append('D')
375 if conflict:
376 filestat.append('C')
377 if unknown:
378 filestat.append('?')
379 cache_files = [x for x in cache_files if x[0] in filestat]
380
381 for fs in cache_files:
382 if all:
383 print '%s %s' % (fs[0], fs[1])
384 else:
385 print '%s' % fs[1]
386
387 def diff(files = [], rev1 = 'HEAD', rev2 = None, out_fd = None):
388 """Show the diff between rev1 and rev2
389 """
390
391 if rev2:
392 diff_str = _output(['git-diff-tree', '-p', rev1, rev2] + files)
393 else:
394 os.system('git-update-cache --refresh > /dev/null')
395 diff_str = _output(['git-diff-cache', '-p', rev1] + files)
396
397 if out_fd:
398 out_fd.write(diff_str)
399 else:
400 return diff_str
401
402 def diffstat(files = [], rev1 = 'HEAD', rev2 = None):
403 """Return the diffstat between rev1 and rev2
404 """
405
406 p=popen2.Popen3('git-apply --stat')
407 diff(files, rev1, rev2, p.tochild)
408 p.tochild.close()
409 str = p.fromchild.read().rstrip()
410 if p.wait():
411 raise GitException, 'git.diffstat failed'
412 return str
413
414 def files(rev1, rev2):
415 """Return the files modified between rev1 and rev2
416 """
417
418 str = ''
419 for line in _output_lines('git-diff-tree -r %s %s' % (rev1, rev2)):
420 str += '%s %s\n' % tuple(line.rstrip().split(' ',4)[-1].split('\t',1))
421
422 return str.rstrip()
423
424 def checkout(files = [], tree_id = None, force = False):
425 """Check out the given or all files
426 """
427 if tree_id and __run('git-read-tree -m', [tree_id]) != 0:
428 raise GitException, 'Failed git-read-tree -m %s' % tree_id
429
430 checkout_cmd = 'git-checkout-cache -q -u'
431 if force:
432 checkout_cmd += ' -f'
433 if len(files) == 0:
434 checkout_cmd += ' -a'
435 else:
436 checkout_cmd += ' --'
437
438 if __run(checkout_cmd, files) != 0:
439 raise GitException, 'Failed git-checkout-cache'
440
441 def switch(tree_id):
442 """Switch the tree to the given id
443 """
444 if __run('git-read-tree -u -m', [get_head(), tree_id]) != 0:
445 raise GitException, 'git-read-tree failed (local changes maybe?)'
446
447 __set_head(tree_id)
448
449 def reset(tree_id = None):
450 """Revert the tree changes relative to the given tree_id. It removes
451 any local changes
452 """
453 if not tree_id:
454 tree_id = get_head()
455
456 cache_files = __tree_status(tree_id = tree_id)
457 rm_files = [x[1] for x in cache_files if x[0] in ['D']]
458
459 checkout(tree_id = tree_id, force = True)
460 __set_head(tree_id)
461
462 # checkout doesn't remove files
463 map(os.remove, rm_files)
464
465 def pull(repository = 'origin', refspec = None):
466 """Pull changes from the remote repository. At the moment, just
467 use the 'git pull' command
468 """
469 args = [repository]
470 if refspec:
471 args.append(refspec)
472
473 if __run('git pull', args) != 0:
474 raise GitException, 'Failed "git pull %s"' % repository
475
476 def apply_patch(filename = None):
477 """Apply a patch onto the current index. There must not be any
478 local changes in the tree, otherwise the command fails
479 """
480 os.system('git-update-cache --refresh > /dev/null')
481
482 if filename:
483 if __run('git-apply --index', [filename]) != 0:
484 raise GitException, 'Patch does not apply cleanly'
485 else:
486 _input('git-apply --index', sys.stdin)
487
488 def clone(repository, local_dir):
489 """Clone a remote repository. At the moment, just use the
490 'git clone' script
491 """
492 if __run('git clone', [repository, local_dir]) != 0:
493 raise GitException, 'Failed "git clone %s %s"' \
494 % (repository, local_dir)