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
if len(ud) == 0: # Identical contents, nothing to do
91
# work-around for difflib being too smart for its own good
92
# if /dev/null is "1,0", patch won't recognize it as /dev/null
94
ud[2] = ud[2].replace('-1,0', '-0,0')
96
ud[2] = ud[2].replace('+1,0', '+0,0')
97
# work around for difflib emitting random spaces after the label
98
ud[0] = ud[0][:-2] + '\n'
99
ud[1] = ud[1][:-2] + '\n'
103
if not line.endswith('\n'):
104
to_file.write("\n\\ No newline at end of file\n")
108
def _spawn_external_diff(diffcmd, capture_errors=True):
109
"""Spawn the externall diff process, and return the child handle.
111
:param diffcmd: The command list to spawn
112
:param capture_errors: Capture stderr as well as setting LANG=C
113
and LC_ALL=C. This lets us read and understand the output of diff,
114
and respond to any errors.
115
:return: A Popen object.
118
# construct minimal environment
120
path = os.environ.get('PATH')
123
env['LANGUAGE'] = 'C' # on win32 only LANGUAGE has effect
126
stderr = subprocess.PIPE
132
pipe = subprocess.Popen(diffcmd,
133
stdin=subprocess.PIPE,
134
stdout=subprocess.PIPE,
138
if e.errno == errno.ENOENT:
139
raise errors.NoDiff(str(e))
145
def external_diff(old_filename, oldlines, new_filename, newlines, to_file,
147
"""Display a diff by calling out to the external diff program."""
148
# make sure our own output is properly ordered before the diff
151
oldtmp_fd, old_abspath = tempfile.mkstemp(prefix='bzr-diff-old-')
152
newtmp_fd, new_abspath = tempfile.mkstemp(prefix='bzr-diff-new-')
153
oldtmpf = os.fdopen(oldtmp_fd, 'wb')
154
newtmpf = os.fdopen(newtmp_fd, 'wb')
157
# TODO: perhaps a special case for comparing to or from the empty
158
# sequence; can just use /dev/null on Unix
160
# TODO: if either of the files being compared already exists as a
161
# regular named file (e.g. in the working directory) then we can
162
# compare directly to that, rather than copying it.
164
oldtmpf.writelines(oldlines)
165
newtmpf.writelines(newlines)
173
'--label', old_filename,
175
'--label', new_filename,
180
# diff only allows one style to be specified; they don't override.
181
# note that some of these take optargs, and the optargs can be
182
# directly appended to the options.
183
# this is only an approximate parser; it doesn't properly understand
185
for s in ['-c', '-u', '-C', '-U',
190
'-y', '--side-by-side',
202
diffcmd.extend(diff_opts)
204
pipe = _spawn_external_diff(diffcmd, capture_errors=True)
205
out,err = pipe.communicate()
208
# internal_diff() adds a trailing newline, add one here for consistency
211
# 'diff' gives retcode == 2 for all sorts of errors
212
# one of those is 'Binary files differ'.
213
# Bad options could also be the problem.
214
# 'Binary files' is not a real error, so we suppress that error.
217
# Since we got here, we want to make sure to give an i18n error
218
pipe = _spawn_external_diff(diffcmd, capture_errors=False)
219
out, err = pipe.communicate()
221
# Write out the new i18n diff response
222
to_file.write(out+'\n')
223
if pipe.returncode != 2:
224
raise errors.BzrError(
225
'external diff failed with exit code 2'
226
' when run with LANG=C and LC_ALL=C,'
227
' but not when run natively: %r' % (diffcmd,))
229
first_line = lang_c_out.split('\n', 1)[0]
230
# Starting with diffutils 2.8.4 the word "binary" was dropped.
231
m = re.match('^(binary )?files.*differ$', first_line, re.I)
233
raise errors.BzrError('external diff failed with exit code 2;'
234
' command: %r' % (diffcmd,))
236
# Binary files differ, just return
239
# If we got to here, we haven't written out the output of diff
243
# returns 1 if files differ; that's OK
245
msg = 'signal %d' % (-rc)
247
msg = 'exit code %d' % rc
249
raise errors.BzrError('external diff failed with %s; command: %r'
254
oldtmpf.close() # and delete
256
# Clean up. Warn in case the files couldn't be deleted
257
# (in case windows still holds the file open, but not
258
# if the files have already been deleted)
260
os.remove(old_abspath)
262
if e.errno not in (errno.ENOENT,):
263
warning('Failed to delete temporary file: %s %s',
266
os.remove(new_abspath)
268
if e.errno not in (errno.ENOENT,):
269
warning('Failed to delete temporary file: %s %s',
273
def diff_cmd_helper(tree, specific_files, external_diff_options,
274
old_revision_spec=None, new_revision_spec=None,
276
old_label='a/', new_label='b/'):
277
"""Helper for cmd_diff.
282
:param specific_files:
283
The specific files to compare, or None
285
:param external_diff_options:
286
If non-None, run an external diff, and pass it these options
288
:param old_revision_spec:
289
If None, use basis tree as old revision, otherwise use the tree for
290
the specified revision.
292
:param new_revision_spec:
293
If None, use working tree as new revision, otherwise use the tree for
294
the specified revision.
296
:param revision_specs:
297
Zero, one or two RevisionSpecs from the command line, saying what revisions
298
to compare. This can be passed as an alternative to the old_revision_spec
299
and new_revision_spec parameters.
301
The more general form is show_diff_trees(), where the caller
302
supplies any two trees.
305
# TODO: perhaps remove the old parameters old_revision_spec and
306
# new_revision_spec, since this is only really for use from cmd_diff and
307
# it now always passes through a sequence of revision_specs -- mbp
312
revision = spec.in_store(tree.branch)
314
revision = spec.in_store(None)
315
revision_id = revision.rev_id
316
branch = revision.branch
317
return branch.repository.revision_tree(revision_id)
319
if revision_specs is not None:
320
assert (old_revision_spec is None
321
and new_revision_spec is None)
322
if len(revision_specs) > 0:
323
old_revision_spec = revision_specs[0]
324
if len(revision_specs) > 1:
325
new_revision_spec = revision_specs[1]
327
if old_revision_spec is None:
328
old_tree = tree.basis_tree()
330
old_tree = spec_tree(old_revision_spec)
332
if (new_revision_spec is None
333
or new_revision_spec.spec is None):
336
new_tree = spec_tree(new_revision_spec)
338
if new_tree is not tree:
339
extra_trees = (tree,)
343
return show_diff_trees(old_tree, new_tree, sys.stdout, specific_files,
344
external_diff_options,
345
old_label=old_label, new_label=new_label,
346
extra_trees=extra_trees)
349
def show_diff_trees(old_tree, new_tree, to_file, specific_files=None,
350
external_diff_options=None,
351
old_label='a/', new_label='b/',
353
path_encoding='utf8'):
354
"""Show in text form the changes from one tree to another.
357
If set, include only changes to these files.
359
external_diff_options
360
If set, use an external GNU diff and pass these options.
363
If set, more Trees to use for looking up file ids
366
If set, the path will be encoded as specified, otherwise is supposed
371
if extra_trees is not None:
372
for tree in extra_trees:
376
differ = DiffTree.from_trees_options(old_tree, new_tree, to_file,
378
external_diff_options,
379
old_label, new_label)
380
return differ.show_diff(specific_files, extra_trees)
383
if extra_trees is not None:
384
for tree in extra_trees:
390
def _patch_header_date(tree, file_id, path):
391
"""Returns a timestamp suitable for use in a patch header."""
392
mtime = tree.get_file_mtime(file_id, path)
393
assert mtime is not None, \
394
"got an mtime of None for file-id %s, path %s in tree %s" % (
396
return timestamp.format_patch_date(mtime)
399
def _raise_if_nonexistent(paths, old_tree, new_tree):
400
"""Complain if paths are not in either inventory or tree.
402
It's OK with the files exist in either tree's inventory, or
403
if they exist in the tree but are not versioned.
405
This can be used by operations such as bzr status that can accept
406
unknown or ignored files.
408
mutter("check paths: %r", paths)
411
s = old_tree.filter_unversioned_files(paths)
412
s = new_tree.filter_unversioned_files(s)
413
s = [path for path in s if not new_tree.has_filename(path)]
415
raise errors.PathsDoNotExist(sorted(s))
418
def get_prop_change(meta_modified):
420
return " (properties changed)"
425
class DiffPath(object):
426
"""Base type for command object that compare files"""
428
# The type or contents of the file were unsuitable for diffing
429
CANNOT_DIFF = 'CANNOT_DIFF'
430
# The file has changed in a semantic way
432
# The file content may have changed, but there is no semantic change
433
UNCHANGED = 'UNCHANGED'
435
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8'):
438
:param old_tree: The tree to show as the old tree in the comparison
439
:param new_tree: The tree to show as new in the comparison
440
:param to_file: The file to write comparison data to
441
:param path_encoding: The character encoding to write paths in
443
self.old_tree = old_tree
444
self.new_tree = new_tree
445
self.to_file = to_file
446
self.path_encoding = path_encoding
449
def from_diff_tree(klass, diff_tree):
450
return klass(diff_tree.old_tree, diff_tree.new_tree,
451
diff_tree.to_file, diff_tree.path_encoding)
454
def _diff_many(differs, file_id, old_path, new_path, old_kind, new_kind):
455
for file_differ in differs:
456
result = file_differ.diff(file_id, old_path, new_path, old_kind,
458
if result is not DiffPath.CANNOT_DIFF:
461
return DiffPath.CANNOT_DIFF
464
class DiffKindChange(object):
465
"""Special differ for file kind changes.
467
Represents kind change as deletion + creation. Uses the other differs
470
def __init__(self, differs):
471
self.differs = differs
474
def from_diff_tree(klass, diff_tree):
475
return klass(diff_tree.differs)
477
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
478
"""Perform comparison
480
:param file_id: The file_id of the file to compare
481
:param old_path: Path of the file in the old tree
482
:param new_path: Path of the file in the new tree
483
:param old_kind: Old file-kind of the file
484
:param new_kind: New file-kind of the file
486
if None in (old_kind, new_kind):
487
return DiffPath.CANNOT_DIFF
488
result = DiffPath._diff_many(self.differs, file_id, old_path,
489
new_path, old_kind, None)
490
if result is DiffPath.CANNOT_DIFF:
492
return DiffPath._diff_many(self.differs, file_id, old_path, new_path,
496
class DiffDirectory(DiffPath):
498
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
499
"""Perform comparison between two directories. (dummy)
502
if 'directory' not in (old_kind, new_kind):
503
return self.CANNOT_DIFF
504
if old_kind not in ('directory', None):
505
return self.CANNOT_DIFF
506
if new_kind not in ('directory', None):
507
return self.CANNOT_DIFF
511
class DiffSymlink(DiffPath):
513
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
514
"""Perform comparison between two symlinks
516
:param file_id: The file_id of the file to compare
517
:param old_path: Path of the file in the old tree
518
:param new_path: Path of the file in the new tree
519
:param old_kind: Old file-kind of the file
520
:param new_kind: New file-kind of the file
522
if 'symlink' not in (old_kind, new_kind):
523
return self.CANNOT_DIFF
524
if old_kind == 'symlink':
525
old_target = self.old_tree.get_symlink_target(file_id)
526
elif old_kind is None:
529
return self.CANNOT_DIFF
530
if new_kind == 'symlink':
531
new_target = self.new_tree.get_symlink_target(file_id)
532
elif new_kind is None:
535
return self.CANNOT_DIFF
536
return self.diff_symlink(old_target, new_target)
538
def diff_symlink(self, old_target, new_target):
539
if old_target is None:
540
self.to_file.write('=== target is %r\n' % new_target)
541
elif new_target is None:
542
self.to_file.write('=== target was %r\n' % old_target)
544
self.to_file.write('=== target changed %r => %r\n' %
545
(old_target, new_target))
549
class DiffText(DiffPath):
551
# GNU Patch uses the epoch date to detect files that are being added
552
# or removed in a diff.
553
EPOCH_DATE = '1970-01-01 00:00:00 +0000'
555
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
556
old_label='', new_label='', text_differ=internal_diff):
557
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
558
self.text_differ = text_differ
559
self.old_label = old_label
560
self.new_label = new_label
561
self.path_encoding = path_encoding
563
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
564
"""Compare two files in unified diff format
566
:param file_id: The file_id of the file to compare
567
:param old_path: Path of the file in the old tree
568
:param new_path: Path of the file in the new tree
569
:param old_kind: Old file-kind of the file
570
:param new_kind: New file-kind of the file
572
if 'file' not in (old_kind, new_kind):
573
return self.CANNOT_DIFF
574
from_file_id = to_file_id = file_id
575
if old_kind == 'file':
576
old_date = _patch_header_date(self.old_tree, file_id, old_path)
577
elif old_kind is None:
578
old_date = self.EPOCH_DATE
581
return self.CANNOT_DIFF
582
if new_kind == 'file':
583
new_date = _patch_header_date(self.new_tree, file_id, new_path)
584
elif new_kind is None:
585
new_date = self.EPOCH_DATE
588
return self.CANNOT_DIFF
589
from_label = '%s%s\t%s' % (self.old_label, old_path, old_date)
590
to_label = '%s%s\t%s' % (self.new_label, new_path, new_date)
591
return self.diff_text(from_file_id, to_file_id, from_label, to_label)
593
def diff_text(self, from_file_id, to_file_id, from_label, to_label):
594
"""Diff the content of given files in two trees
596
:param from_file_id: The id of the file in the from tree. If None,
597
the file is not present in the from tree.
598
:param to_file_id: The id of the file in the to tree. This may refer
599
to a different file from from_file_id. If None,
600
the file is not present in the to tree.
602
def _get_text(tree, file_id):
603
if file_id is not None:
604
return tree.get_file(file_id).readlines()
608
from_text = _get_text(self.old_tree, from_file_id)
609
to_text = _get_text(self.new_tree, to_file_id)
610
self.text_differ(from_label, from_text, to_label, to_text,
612
except errors.BinaryFile:
614
("Binary files %s and %s differ\n" %
615
(from_label, to_label)).encode(self.path_encoding))
619
class DiffTree(object):
620
"""Provides textual representations of the difference between two trees.
622
A DiffTree examines two trees and where a file-id has altered
623
between them, generates a textual representation of the difference.
624
DiffTree uses a sequence of DiffPath objects which are each
625
given the opportunity to handle a given altered fileid. The list
626
of DiffPath objects can be extended globally by appending to
627
DiffTree.diff_factories, or for a specific diff operation by
628
supplying the extra_factories option to the appropriate method.
631
# list of factories that can provide instances of DiffPath objects
632
# may be extended by plugins.
633
diff_factories = [DiffSymlink.from_diff_tree,
634
DiffDirectory.from_diff_tree]
636
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
637
diff_text=None, extra_factories=None):
640
:param old_tree: Tree to show as old in the comparison
641
:param new_tree: Tree to show as new in the comparison
642
:param to_file: File to write comparision to
643
:param path_encoding: Character encoding to write paths in
644
:param diff_text: DiffPath-type object to use as a last resort for
646
:param extra_factories: Factories of DiffPaths to try before any other
648
if diff_text is None:
649
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
650
'', '', internal_diff)
651
self.old_tree = old_tree
652
self.new_tree = new_tree
653
self.to_file = to_file
654
self.path_encoding = path_encoding
656
if extra_factories is not None:
657
self.differs.extend(f(self) for f in extra_factories)
658
self.differs.extend(f(self) for f in self.diff_factories)
659
self.differs.extend([diff_text, DiffKindChange.from_diff_tree(self)])
662
def from_trees_options(klass, old_tree, new_tree, to_file,
663
path_encoding, external_diff_options, old_label,
665
"""Factory for producing a DiffTree.
667
Designed to accept options used by show_diff_trees.
668
:param old_tree: The tree to show as old in the comparison
669
:param new_tree: The tree to show as new in the comparison
670
:param to_file: File to write comparisons to
671
:param path_encoding: Character encoding to use for writing paths
672
:param external_diff_options: If supplied, use the installed diff
673
binary to perform file comparison, using supplied options.
674
:param old_label: Prefix to use for old file labels
675
:param new_label: Prefix to use for new file labels
677
if external_diff_options:
678
assert isinstance(external_diff_options, basestring)
679
opts = external_diff_options.split()
680
def diff_file(olab, olines, nlab, nlines, to_file):
681
external_diff(olab, olines, nlab, nlines, to_file, opts)
683
diff_file = internal_diff
684
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
685
old_label, new_label, diff_file)
686
return klass(old_tree, new_tree, to_file, path_encoding, diff_text)
688
def show_diff(self, specific_files, extra_trees=None):
689
"""Write tree diff to self.to_file
691
:param sepecific_files: the specific files to compare (recursive)
692
:param extra_trees: extra trees to use for mapping paths to file_ids
694
# TODO: Generation of pseudo-diffs for added/deleted files could
695
# be usefully made into a much faster special case.
697
delta = self.new_tree.changes_from(self.old_tree,
698
specific_files=specific_files,
699
extra_trees=extra_trees, require_versioned=True)
702
for path, file_id, kind in delta.removed:
704
path_encoded = path.encode(self.path_encoding, "replace")
705
self.to_file.write("=== removed %s '%s'\n" % (kind, path_encoded))
706
self.diff(file_id, path, path)
708
for path, file_id, kind in delta.added:
710
path_encoded = path.encode(self.path_encoding, "replace")
711
self.to_file.write("=== added %s '%s'\n" % (kind, path_encoded))
712
self.diff(file_id, path, path)
713
for (old_path, new_path, file_id, kind,
714
text_modified, meta_modified) in delta.renamed:
716
prop_str = get_prop_change(meta_modified)
717
oldpath_encoded = old_path.encode(self.path_encoding, "replace")
718
newpath_encoded = new_path.encode(self.path_encoding, "replace")
719
self.to_file.write("=== renamed %s '%s' => '%s'%s\n" % (kind,
720
oldpath_encoded, newpath_encoded, prop_str))
722
self.diff(file_id, old_path, new_path)
723
for path, file_id, kind, text_modified, meta_modified in\
726
prop_str = get_prop_change(meta_modified)
727
path_encoded = path.encode(self.path_encoding, "replace")
728
self.to_file.write("=== modified %s '%s'%s\n" % (kind,
729
path_encoded, prop_str))
730
# The file may be in a different location in the old tree (because
731
# the containing dir was renamed, but the file itself was not)
733
old_path = self.old_tree.id2path(file_id)
734
self.diff(file_id, old_path, path)
737
def diff(self, file_id, old_path, new_path):
738
"""Perform a diff of a single file
740
:param file_id: file-id of the file
741
:param old_path: The path of the file in the old tree
742
:param new_path: The path of the file in the new tree
745
old_kind = self.old_tree.kind(file_id)
746
except (errors.NoSuchId, errors.NoSuchFile):
749
new_kind = self.new_tree.kind(file_id)
750
except (errors.NoSuchId, errors.NoSuchFile):
753
result = DiffPath._diff_many(self.differs, file_id, old_path,
754
new_path, old_kind, new_kind)
755
if result is DiffPath.CANNOT_DIFF:
756
error_path = new_path
757
if error_path is None:
758
error_path = old_path
759
raise errors.NoDiffFound(error_path)
added
removed
bzrlib/diff.py
