1
# Copyright (C) 2004, 2005, 2006 Canonical Ltd.
1
# -*- coding: UTF-8 -*-
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
5
5
# the Free Software Foundation; either version 2 of the License, or
6
6
# (at your option) any later version.
8
8
# This program is distributed in the hope that it will be useful,
9
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
11
# GNU General Public License for more details.
13
13
# You should have received a copy of the GNU General Public License
14
14
# along with this program; if not, write to the Free Software
15
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
22
from bzrlib.lazy_import import lazy_import
23
lazy_import(globals(), """
38
from bzrlib.symbol_versioning import (
41
from bzrlib.trace import mutter, warning
17
from bzrlib.trace import mutter
18
from bzrlib.errors import BzrError
19
from bzrlib.delta import compare_trees
44
21
# TODO: Rather than building a changeset object, we should probably
45
22
# invoke callbacks on an object. That object can either accumulate a
46
23
# list, write them out directly, etc etc.
49
class _PrematchedMatcher(difflib.SequenceMatcher):
50
"""Allow SequenceMatcher operations to use predetermined blocks"""
52
def __init__(self, matching_blocks):
53
difflib.SequenceMatcher(self, None, None)
54
self.matching_blocks = matching_blocks
58
def internal_diff(old_filename, oldlines, new_filename, newlines, to_file,
59
allow_binary=False, sequence_matcher=None,
60
path_encoding='utf8'):
25
def internal_diff(old_filename, oldlines, new_filename, newlines, to_file):
61
28
# FIXME: difflib is wrong if there is no trailing newline.
62
29
# The syntax used by patch seems to be "\ No newline at
63
30
# end of file" following the last diff line from that
100
60
to_file.write(line)
101
61
if not line.endswith('\n'):
102
62
to_file.write("\n\\ No newline at end of file\n")
106
def _spawn_external_diff(diffcmd, capture_errors=True):
107
"""Spawn the externall diff process, and return the child handle.
109
:param diffcmd: The command list to spawn
110
:param capture_errors: Capture stderr as well as setting LANG=C
111
and LC_ALL=C. This lets us read and understand the output of diff,
112
and respond to any errors.
113
:return: A Popen object.
116
# construct minimal environment
118
path = os.environ.get('PATH')
121
env['LANGUAGE'] = 'C' # on win32 only LANGUAGE has effect
124
stderr = subprocess.PIPE
130
pipe = subprocess.Popen(diffcmd,
131
stdin=subprocess.PIPE,
132
stdout=subprocess.PIPE,
136
if e.errno == errno.ENOENT:
137
raise errors.NoDiff(str(e))
143
66
def external_diff(old_filename, oldlines, new_filename, newlines, to_file,
145
68
"""Display a diff by calling out to the external diff program."""
71
if to_file != sys.stdout:
72
raise NotImplementedError("sorry, can't send external diff other than to stdout yet",
146
75
# make sure our own output is properly ordered before the diff
149
oldtmp_fd, old_abspath = tempfile.mkstemp(prefix='bzr-diff-old-')
150
newtmp_fd, new_abspath = tempfile.mkstemp(prefix='bzr-diff-new-')
151
oldtmpf = os.fdopen(oldtmp_fd, 'wb')
152
newtmpf = os.fdopen(newtmp_fd, 'wb')
78
from tempfile import NamedTemporaryFile
81
oldtmpf = NamedTemporaryFile()
82
newtmpf = NamedTemporaryFile()
155
85
# TODO: perhaps a special case for comparing to or from the empty
200
128
diffcmd.extend(diff_opts)
202
pipe = _spawn_external_diff(diffcmd, capture_errors=True)
203
out,err = pipe.communicate()
130
rc = os.spawnvp(os.P_WAIT, 'diff', diffcmd)
206
# internal_diff() adds a trailing newline, add one here for consistency
209
# 'diff' gives retcode == 2 for all sorts of errors
210
# one of those is 'Binary files differ'.
211
# Bad options could also be the problem.
212
# 'Binary files' is not a real error, so we suppress that error.
215
# Since we got here, we want to make sure to give an i18n error
216
pipe = _spawn_external_diff(diffcmd, capture_errors=False)
217
out, err = pipe.communicate()
219
# Write out the new i18n diff response
220
to_file.write(out+'\n')
221
if pipe.returncode != 2:
222
raise errors.BzrError(
223
'external diff failed with exit code 2'
224
' when run with LANG=C and LC_ALL=C,'
225
' but not when run natively: %r' % (diffcmd,))
227
first_line = lang_c_out.split('\n', 1)[0]
228
# Starting with diffutils 2.8.4 the word "binary" was dropped.
229
m = re.match('^(binary )?files.*differ$', first_line, re.I)
231
raise errors.BzrError('external diff failed with exit code 2;'
232
' command: %r' % (diffcmd,))
234
# Binary files differ, just return
237
# If we got to here, we haven't written out the output of diff
132
if rc != 0 and rc != 1:
241
133
# returns 1 if files differ; that's OK
243
135
msg = 'signal %d' % (-rc)
245
137
msg = 'exit code %d' % rc
247
raise errors.BzrError('external diff failed with %s; command: %r'
139
raise BzrError('external diff failed with %s; command: %r' % (rc, diffcmd))
252
141
oldtmpf.close() # and delete
254
# Clean up. Warn in case the files couldn't be deleted
255
# (in case windows still holds the file open, but not
256
# if the files have already been deleted)
258
os.remove(old_abspath)
260
if e.errno not in (errno.ENOENT,):
261
warning('Failed to delete temporary file: %s %s',
264
os.remove(new_abspath)
266
if e.errno not in (errno.ENOENT,):
267
warning('Failed to delete temporary file: %s %s',
271
def diff_cmd_helper(tree, specific_files, external_diff_options,
272
old_revision_spec=None, new_revision_spec=None,
274
old_label='a/', new_label='b/'):
275
"""Helper for cmd_diff.
280
:param specific_files:
281
The specific files to compare, or None
283
:param external_diff_options:
284
If non-None, run an external diff, and pass it these options
286
:param old_revision_spec:
287
If None, use basis tree as old revision, otherwise use the tree for
288
the specified revision.
290
:param new_revision_spec:
291
If None, use working tree as new revision, otherwise use the tree for
292
the specified revision.
144
def show_diff(b, from_spec, specific_files, external_diff_options=None,
145
revision2=None, output=None, b2=None):
146
"""Shortcut for showing the diff to the working tree.
152
None for 'basis tree', or otherwise the old revision to compare against.
294
:param revision_specs:
295
Zero, one or two RevisionSpecs from the command line, saying what revisions
296
to compare. This can be passed as an alternative to the old_revision_spec
297
and new_revision_spec parameters.
299
154
The more general form is show_diff_trees(), where the caller
300
155
supplies any two trees.
303
# TODO: perhaps remove the old parameters old_revision_spec and
304
# new_revision_spec, since this is only really for use from cmd_diff and
305
# it now always passes through a sequence of revision_specs -- mbp
310
revision = spec.in_store(tree.branch)
312
revision = spec.in_store(None)
313
revision_id = revision.rev_id
314
branch = revision.branch
315
return branch.repository.revision_tree(revision_id)
317
if revision_specs is not None:
318
assert (old_revision_spec is None
319
and new_revision_spec is None)
320
if len(revision_specs) > 0:
321
old_revision_spec = revision_specs[0]
322
if len(revision_specs) > 1:
323
new_revision_spec = revision_specs[1]
325
if old_revision_spec is None:
326
old_tree = tree.basis_tree()
328
old_tree = spec_tree(old_revision_spec)
330
if (new_revision_spec is None
331
or new_revision_spec.spec is None):
334
new_tree = spec_tree(new_revision_spec)
336
if new_tree is not tree:
337
extra_trees = (tree,)
341
return show_diff_trees(old_tree, new_tree, sys.stdout, specific_files,
342
external_diff_options,
343
old_label=old_label, new_label=new_label,
344
extra_trees=extra_trees)
161
if from_spec is None:
163
old_tree = b.basis_tree()
165
old_tree = b.working_tree()
167
old_tree = b.revision_tree(from_spec.in_history(b).rev_id)
169
if revision2 is None:
171
new_tree = b.working_tree()
173
new_tree = b2.working_tree()
175
new_tree = b.revision_tree(revision2.in_history(b).rev_id)
177
return show_diff_trees(old_tree, new_tree, output, specific_files,
178
external_diff_options)
347
182
def show_diff_trees(old_tree, new_tree, to_file, specific_files=None,
348
external_diff_options=None,
349
old_label='a/', new_label='b/',
351
path_encoding='utf8'):
183
external_diff_options=None):
352
184
"""Show in text form the changes from one tree to another.
357
189
external_diff_options
358
190
If set, use an external GNU diff and pass these options.
361
If set, more Trees to use for looking up file ids
364
If set, the path will be encoded as specified, otherwise is supposed
369
if extra_trees is not None:
370
for tree in extra_trees:
374
return _show_diff_trees(old_tree, new_tree, to_file,
375
specific_files, external_diff_options,
376
old_label=old_label, new_label=new_label,
377
extra_trees=extra_trees,
378
path_encoding=path_encoding)
381
if extra_trees is not None:
382
for tree in extra_trees:
388
def _show_diff_trees(old_tree, new_tree, to_file,
389
specific_files, external_diff_options, path_encoding,
390
old_label='a/', new_label='b/', extra_trees=None):
392
# GNU Patch uses the epoch date to detect files that are being added
393
# or removed in a diff.
394
EPOCH_DATE = '1970-01-01 00:00:00 +0000'
193
# TODO: Options to control putting on a prefix or suffix, perhaps as a format string
197
DEVNULL = '/dev/null'
198
# Windows users, don't panic about this filename -- it is a
199
# special signal to GNU patch that the file should be created or
200
# deleted respectively.
396
202
# TODO: Generation of pseudo-diffs for added/deleted files could
397
203
# be usefully made into a much faster special case.
405
211
diff_file = internal_diff
407
delta = new_tree.changes_from(old_tree,
408
specific_files=specific_files,
409
extra_trees=extra_trees, require_versioned=True)
214
delta = compare_trees(old_tree, new_tree, want_unchanged=False,
215
specific_files=specific_files)
412
218
for path, file_id, kind in delta.removed:
414
path_encoded = path.encode(path_encoding, "replace")
415
to_file.write("=== removed %s '%s'\n" % (kind, path_encoded))
416
old_name = '%s%s\t%s' % (old_label, path,
417
_patch_header_date(old_tree, file_id, path))
418
new_name = '%s%s\t%s' % (new_label, path, EPOCH_DATE)
419
old_tree.inventory[file_id].diff(diff_file, old_name, old_tree,
420
new_name, None, None, to_file)
220
print >>to_file, '=== removed %s %r' % (kind, path)
221
old_tree.inventory[file_id].diff(diff_file, old_label + path, old_tree,
222
DEVNULL, None, None, to_file)
421
223
for path, file_id, kind in delta.added:
423
path_encoded = path.encode(path_encoding, "replace")
424
to_file.write("=== added %s '%s'\n" % (kind, path_encoded))
425
old_name = '%s%s\t%s' % (old_label, path, EPOCH_DATE)
426
new_name = '%s%s\t%s' % (new_label, path,
427
_patch_header_date(new_tree, file_id, path))
428
new_tree.inventory[file_id].diff(diff_file, new_name, new_tree,
429
old_name, None, None, to_file,
225
print >>to_file, '=== added %s %r' % (kind, path)
226
new_tree.inventory[file_id].diff(diff_file, new_label + path, new_tree,
227
DEVNULL, None, None, to_file,
431
229
for (old_path, new_path, file_id, kind,
432
230
text_modified, meta_modified) in delta.renamed:
434
232
prop_str = get_prop_change(meta_modified)
435
oldpath_encoded = old_path.encode(path_encoding, "replace")
436
newpath_encoded = new_path.encode(path_encoding, "replace")
437
to_file.write("=== renamed %s '%s' => '%s'%s\n" % (kind,
438
oldpath_encoded, newpath_encoded, prop_str))
439
old_name = '%s%s\t%s' % (old_label, old_path,
440
_patch_header_date(old_tree, file_id,
442
new_name = '%s%s\t%s' % (new_label, new_path,
443
_patch_header_date(new_tree, file_id,
445
_maybe_diff_file_or_symlink(old_name, old_tree, file_id,
233
print >>to_file, '=== renamed %s %r => %r%s' % (
234
kind, old_path, new_path, prop_str)
235
_maybe_diff_file_or_symlink(old_label, old_path, old_tree, file_id,
236
new_label, new_path, new_tree,
447
237
text_modified, kind, to_file, diff_file)
448
238
for path, file_id, kind, text_modified, meta_modified in delta.modified:
450
240
prop_str = get_prop_change(meta_modified)
451
path_encoded = path.encode(path_encoding, "replace")
452
to_file.write("=== modified %s '%s'%s\n" % (kind,
453
path_encoded, prop_str))
454
# The file may be in a different location in the old tree (because
455
# the containing dir was renamed, but the file itself was not)
456
old_path = old_tree.id2path(file_id)
457
old_name = '%s%s\t%s' % (old_label, old_path,
458
_patch_header_date(old_tree, file_id, old_path))
459
new_name = '%s%s\t%s' % (new_label, path,
460
_patch_header_date(new_tree, file_id, path))
241
print >>to_file, '=== modified %s %r%s' % (kind, path, prop_str)
461
242
if text_modified:
462
_maybe_diff_file_or_symlink(old_name, old_tree, file_id,
243
_maybe_diff_file_or_symlink(old_label, path, old_tree, file_id,
244
new_label, path, new_tree,
464
245
True, kind, to_file, diff_file)
466
246
return has_changes
469
def _patch_header_date(tree, file_id, path):
470
"""Returns a timestamp suitable for use in a patch header."""
471
mtime = tree.get_file_mtime(file_id, path)
472
assert mtime is not None, \
473
"got an mtime of None for file-id %s, path %s in tree %s" % (
475
return timestamp.format_patch_date(mtime)
478
def _raise_if_nonexistent(paths, old_tree, new_tree):
479
"""Complain if paths are not in either inventory or tree.
481
It's OK with the files exist in either tree's inventory, or
482
if they exist in the tree but are not versioned.
484
This can be used by operations such as bzr status that can accept
485
unknown or ignored files.
487
mutter("check paths: %r", paths)
490
s = old_tree.filter_unversioned_files(paths)
491
s = new_tree.filter_unversioned_files(s)
492
s = [path for path in s if not new_tree.has_filename(path)]
494
raise errors.PathsDoNotExist(sorted(s))
497
249
def get_prop_change(meta_modified):
498
250
if meta_modified:
added
removed
bzrlib/diff.py
