1
# Copyright (C) 2004, 2005, 2006 Canonical Ltd.
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
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
44
# TODO: Rather than building a changeset object, we should probably
45
# invoke callbacks on an object. That object can either accumulate a
46
# 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'):
61
# FIXME: difflib is wrong if there is no trailing newline.
62
# The syntax used by patch seems to be "\ No newline at
63
# end of file" following the last diff line from that
64
# file. This is not trivial to insert into the
65
# unified_diff output and it might be better to just fix
66
# or replace that function.
68
# In the meantime we at least make sure the patch isn't
72
# Special workaround for Python2.3, where difflib fails if
73
# both sequences are empty.
74
if not oldlines and not newlines:
77
if allow_binary is False:
78
textfile.check_text_lines(oldlines)
79
textfile.check_text_lines(newlines)
81
if sequence_matcher is None:
82
sequence_matcher = patiencediff.PatienceSequenceMatcher
83
ud = patiencediff.unified_diff(oldlines, newlines,
84
fromfile=old_filename.encode(path_encoding),
85
tofile=new_filename.encode(path_encoding),
86
sequencematcher=sequence_matcher)
89
# work-around for difflib being too smart for its own good
90
# if /dev/null is "1,0", patch won't recognize it as /dev/null
92
ud[2] = ud[2].replace('-1,0', '-0,0')
94
ud[2] = ud[2].replace('+1,0', '+0,0')
95
# work around for difflib emitting random spaces after the label
96
ud[0] = ud[0][:-2] + '\n'
97
ud[1] = ud[1][:-2] + '\n'
101
if not line.endswith('\n'):
102
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
def external_diff(old_filename, oldlines, new_filename, newlines, to_file,
145
"""Display a diff by calling out to the external diff program."""
146
# 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')
155
# TODO: perhaps a special case for comparing to or from the empty
156
# sequence; can just use /dev/null on Unix
158
# TODO: if either of the files being compared already exists as a
159
# regular named file (e.g. in the working directory) then we can
160
# compare directly to that, rather than copying it.
162
oldtmpf.writelines(oldlines)
163
newtmpf.writelines(newlines)
171
'--label', old_filename,
173
'--label', new_filename,
178
# diff only allows one style to be specified; they don't override.
179
# note that some of these take optargs, and the optargs can be
180
# directly appended to the options.
181
# this is only an approximate parser; it doesn't properly understand
183
for s in ['-c', '-u', '-C', '-U',
188
'-y', '--side-by-side',
200
diffcmd.extend(diff_opts)
202
pipe = _spawn_external_diff(diffcmd, capture_errors=True)
203
out,err = pipe.communicate()
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
241
# returns 1 if files differ; that's OK
243
msg = 'signal %d' % (-rc)
245
msg = 'exit code %d' % rc
247
raise errors.BzrError('external diff failed with %s; command: %r'
252
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.
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
The more general form is show_diff_trees(), where the caller
300
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)
347
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'):
352
"""Show in text form the changes from one tree to another.
355
If set, include only changes to these files.
357
external_diff_options
358
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
differ = DiffTree.from_trees_options(old_tree, new_tree, to_file,
376
external_diff_options,
377
old_label, new_label)
378
return differ.show_diff(specific_files, extra_trees)
381
if extra_trees is not None:
382
for tree in extra_trees:
388
def _patch_header_date(tree, file_id, path):
389
"""Returns a timestamp suitable for use in a patch header."""
390
mtime = tree.get_file_mtime(file_id, path)
391
assert mtime is not None, \
392
"got an mtime of None for file-id %s, path %s in tree %s" % (
394
return timestamp.format_patch_date(mtime)
397
def _raise_if_nonexistent(paths, old_tree, new_tree):
398
"""Complain if paths are not in either inventory or tree.
400
It's OK with the files exist in either tree's inventory, or
401
if they exist in the tree but are not versioned.
403
This can be used by operations such as bzr status that can accept
404
unknown or ignored files.
406
mutter("check paths: %r", paths)
409
s = old_tree.filter_unversioned_files(paths)
410
s = new_tree.filter_unversioned_files(s)
411
s = [path for path in s if not new_tree.has_filename(path)]
413
raise errors.PathsDoNotExist(sorted(s))
416
def get_prop_change(meta_modified):
418
return " (properties changed)"
423
class DiffPath(object):
424
"""Base type for command object that compare files"""
426
# The type or contents of the file were unsuitable for diffing
427
CANNOT_DIFF = 'CANNOT_DIFF'
428
# The file has changed in a semantic way
430
# The file content may have changed, but there is no semantic change
431
UNCHANGED = 'UNCHANGED'
433
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8'):
436
:param old_tree: The tree to show as the old tree in the comparison
437
:param new_tree: The tree to show as new in the comparison
438
:param to_file: The file to write comparison data to
439
:param path_encoding: The character encoding to write paths in
441
self.old_tree = old_tree
442
self.new_tree = new_tree
443
self.to_file = to_file
444
self.path_encoding = path_encoding
447
def from_diff_tree(klass, diff_tree):
448
return klass(diff_tree.old_tree, diff_tree.new_tree,
449
diff_tree.to_file, diff_tree.path_encoding)
452
def _diff_many(differs, file_id, old_path, new_path, old_kind, new_kind):
453
for file_differ in differs:
454
result = file_differ.diff(file_id, old_path, new_path, old_kind,
456
if result is not DiffPath.CANNOT_DIFF:
459
return DiffPath.CANNOT_DIFF
462
class DiffKindChange(object):
463
"""Special differ for file kind changes.
465
Represents kind change as deletion + creation. Uses the other differs
468
def __init__(self, differs):
469
self.differs = differs
472
def from_diff_tree(klass, diff_tree):
473
return klass(diff_tree.differs)
475
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
476
"""Perform comparison
478
:param file_id: The file_id of the file to compare
479
:param old_path: Path of the file in the old tree
480
:param new_path: Path of the file in the new tree
481
:param old_kind: Old file-kind of the file
482
:param new_kind: New file-kind of the file
484
if None in (old_kind, new_kind):
485
return DiffPath.CANNOT_DIFF
486
result = DiffPath._diff_many(self.differs, file_id, old_path,
487
new_path, old_kind, None)
488
if result is DiffPath.CANNOT_DIFF:
490
return DiffPath._diff_many(self.differs, file_id, old_path, new_path,
494
class DiffDirectory(DiffPath):
496
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
497
"""Perform comparison between two directories. (dummy)
500
if 'directory' not in (old_kind, new_kind):
501
return self.CANNOT_DIFF
502
if old_kind not in ('directory', None):
503
return self.CANNOT_DIFF
504
if new_kind not in ('directory', None):
505
return self.CANNOT_DIFF
509
class DiffSymlink(DiffPath):
511
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
512
"""Perform comparison between two symlinks
514
:param file_id: The file_id of the file to compare
515
:param old_path: Path of the file in the old tree
516
:param new_path: Path of the file in the new tree
517
:param old_kind: Old file-kind of the file
518
:param new_kind: New file-kind of the file
520
if 'symlink' not in (old_kind, new_kind):
521
return self.CANNOT_DIFF
522
if old_kind == 'symlink':
523
old_target = self.old_tree.get_symlink_target(file_id)
524
elif old_kind is None:
527
return self.CANNOT_DIFF
528
if new_kind == 'symlink':
529
new_target = self.new_tree.get_symlink_target(file_id)
530
elif new_kind is None:
533
return self.CANNOT_DIFF
534
return self.diff_symlink(old_target, new_target)
536
def diff_symlink(self, old_target, new_target):
537
if old_target is None:
538
self.to_file.write('=== target is %r\n' % new_target)
539
elif new_target is None:
540
self.to_file.write('=== target was %r\n' % old_target)
542
self.to_file.write('=== target changed %r => %r\n' %
543
(old_target, new_target))
547
class DiffText(DiffPath):
549
# GNU Patch uses the epoch date to detect files that are being added
550
# or removed in a diff.
551
EPOCH_DATE = '1970-01-01 00:00:00 +0000'
553
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
554
old_label='', new_label='', text_differ=internal_diff):
555
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
556
self.text_differ = text_differ
557
self.old_label = old_label
558
self.new_label = new_label
559
self.path_encoding = path_encoding
561
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
562
"""Compare two files in unified diff format
564
:param file_id: The file_id of the file to compare
565
:param old_path: Path of the file in the old tree
566
:param new_path: Path of the file in the new tree
567
:param old_kind: Old file-kind of the file
568
:param new_kind: New file-kind of the file
570
if 'file' not in (old_kind, new_kind):
571
return self.CANNOT_DIFF
572
from_file_id = to_file_id = file_id
573
if old_kind == 'file':
574
old_date = _patch_header_date(self.old_tree, file_id, old_path)
575
elif old_kind is None:
576
old_date = self.EPOCH_DATE
579
return self.CANNOT_DIFF
580
if new_kind == 'file':
581
new_date = _patch_header_date(self.new_tree, file_id, new_path)
582
elif new_kind is None:
583
new_date = self.EPOCH_DATE
586
return self.CANNOT_DIFF
587
from_label = '%s%s\t%s' % (self.old_label, old_path, old_date)
588
to_label = '%s%s\t%s' % (self.new_label, new_path, new_date)
589
return self.diff_text(from_file_id, to_file_id, from_label, to_label)
591
def diff_text(self, from_file_id, to_file_id, from_label, to_label):
592
"""Diff the content of given files in two trees
594
:param from_file_id: The id of the file in the from tree. If None,
595
the file is not present in the from tree.
596
:param to_file_id: The id of the file in the to tree. This may refer
597
to a different file from from_file_id. If None,
598
the file is not present in the to tree.
600
def _get_text(tree, file_id):
601
if file_id is not None:
602
return tree.get_file(file_id).readlines()
606
from_text = _get_text(self.old_tree, from_file_id)
607
to_text = _get_text(self.new_tree, to_file_id)
608
self.text_differ(from_label, from_text, to_label, to_text,
610
except errors.BinaryFile:
612
("Binary files %s and %s differ\n" %
613
(from_label, to_label)).encode(self.path_encoding))
617
class DiffTree(object):
618
"""Provides textual representations of the difference between two trees.
620
A DiffTree examines two trees and where a file-id has altered
621
between them, generates a textual representation of the difference.
622
DiffTree uses a sequence of DiffPath objects which are each
623
given the opportunity to handle a given altered fileid. The list
624
of DiffPath objects can be extended globally by appending to
625
DiffTree.diff_factories, or for a specific diff operation by
626
supplying the extra_factories option to the appropriate method.
629
# list of factories that can provide instances of DiffPath objects
630
# may be extended by plugins.
631
diff_factories = [DiffSymlink.from_diff_tree,
632
DiffDirectory.from_diff_tree]
634
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
635
diff_text=None, extra_factories=None):
638
:param old_tree: Tree to show as old in the comparison
639
:param new_tree: Tree to show as new in the comparison
640
:param to_file: File to write comparision to
641
:param path_encoding: Character encoding to write paths in
642
:param diff_text: DiffPath-type object to use as a last resort for
644
:param extra_factories: Factories of DiffPaths to try before any other
646
if diff_text is None:
647
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
648
'', '', internal_diff)
649
self.old_tree = old_tree
650
self.new_tree = new_tree
651
self.to_file = to_file
652
self.path_encoding = path_encoding
654
if extra_factories is not None:
655
self.differs.extend(f(self) for f in extra_factories)
656
self.differs.extend(f(self) for f in self.diff_factories)
657
self.differs.extend([diff_text, DiffKindChange.from_diff_tree(self)])
660
def from_trees_options(klass, old_tree, new_tree, to_file,
661
path_encoding, external_diff_options, old_label,
663
"""Factory for producing a DiffTree.
665
Designed to accept options used by show_diff_trees.
666
:param old_tree: The tree to show as old in the comparison
667
:param new_tree: The tree to show as new in the comparison
668
:param to_file: File to write comparisons to
669
:param path_encoding: Character encoding to use for writing paths
670
:param external_diff_options: If supplied, use the installed diff
671
binary to perform file comparison, using supplied options.
672
:param old_label: Prefix to use for old file labels
673
:param new_label: Prefix to use for new file labels
675
if external_diff_options:
676
assert isinstance(external_diff_options, basestring)
677
opts = external_diff_options.split()
678
def diff_file(olab, olines, nlab, nlines, to_file):
679
external_diff(olab, olines, nlab, nlines, to_file, opts)
681
diff_file = internal_diff
682
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
683
old_label, new_label, diff_file)
684
return klass(old_tree, new_tree, to_file, path_encoding, diff_text)
686
def show_diff(self, specific_files, extra_trees=None):
687
"""Write tree diff to self.to_file
689
:param sepecific_files: the specific files to compare (recursive)
690
:param extra_trees: extra trees to use for mapping paths to file_ids
692
# TODO: Generation of pseudo-diffs for added/deleted files could
693
# be usefully made into a much faster special case.
695
delta = self.new_tree.changes_from(self.old_tree,
696
specific_files=specific_files,
697
extra_trees=extra_trees, require_versioned=True)
700
for path, file_id, kind in delta.removed:
702
path_encoded = path.encode(self.path_encoding, "replace")
703
self.to_file.write("=== removed %s '%s'\n" % (kind, path_encoded))
704
self.diff(file_id, path, path)
706
for path, file_id, kind in delta.added:
708
path_encoded = path.encode(self.path_encoding, "replace")
709
self.to_file.write("=== added %s '%s'\n" % (kind, path_encoded))
710
self.diff(file_id, path, path)
711
for (old_path, new_path, file_id, kind,
712
text_modified, meta_modified) in delta.renamed:
714
prop_str = get_prop_change(meta_modified)
715
oldpath_encoded = old_path.encode(self.path_encoding, "replace")
716
newpath_encoded = new_path.encode(self.path_encoding, "replace")
717
self.to_file.write("=== renamed %s '%s' => '%s'%s\n" % (kind,
718
oldpath_encoded, newpath_encoded, prop_str))
720
self.diff(file_id, old_path, new_path)
721
for path, file_id, kind, text_modified, meta_modified in\
724
prop_str = get_prop_change(meta_modified)
725
path_encoded = path.encode(self.path_encoding, "replace")
726
self.to_file.write("=== modified %s '%s'%s\n" % (kind,
727
path_encoded, prop_str))
728
# The file may be in a different location in the old tree (because
729
# the containing dir was renamed, but the file itself was not)
731
old_path = self.old_tree.id2path(file_id)
732
self.diff(file_id, old_path, path)
735
def diff(self, file_id, old_path, new_path):
736
"""Perform a diff of a single file
738
:param file_id: file-id of the file
739
:param old_path: The path of the file in the old tree
740
:param new_path: The path of the file in the new tree
743
old_kind = self.old_tree.kind(file_id)
744
except errors.NoSuchId:
747
new_kind = self.new_tree.kind(file_id)
748
except errors.NoSuchId:
751
result = DiffPath._diff_many(self.differs, file_id, old_path,
752
new_path, old_kind, new_kind)
753
if result is DiffPath.CANNOT_DIFF:
754
error_path = new_path
755
if error_path is None:
756
error_path = old_path
757
raise errors.NoDiffFound(error_path)
added
removed
bzrlib/diff.py
