1
# Copyright (C) 2005-2011 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23
from bzrlib.lazy_import import lazy_import
24
lazy_import(globals(), """
41
from bzrlib.workingtree import WorkingTree
44
from bzrlib.registry import (
47
from bzrlib.trace import mutter, note, warning
50
class AtTemplate(string.Template):
51
"""Templating class that uses @ instead of $."""
56
# TODO: Rather than building a changeset object, we should probably
57
# invoke callbacks on an object. That object can either accumulate a
58
# list, write them out directly, etc etc.
61
class _PrematchedMatcher(difflib.SequenceMatcher):
62
"""Allow SequenceMatcher operations to use predetermined blocks"""
64
def __init__(self, matching_blocks):
65
difflib.SequenceMatcher(self, None, None)
66
self.matching_blocks = matching_blocks
70
def internal_diff(old_filename, oldlines, new_filename, newlines, to_file,
71
allow_binary=False, sequence_matcher=None,
72
path_encoding='utf8'):
73
# FIXME: difflib is wrong if there is no trailing newline.
74
# The syntax used by patch seems to be "\ No newline at
75
# end of file" following the last diff line from that
76
# file. This is not trivial to insert into the
77
# unified_diff output and it might be better to just fix
78
# or replace that function.
80
# In the meantime we at least make sure the patch isn't
84
# Special workaround for Python2.3, where difflib fails if
85
# both sequences are empty.
86
if not oldlines and not newlines:
89
if allow_binary is False:
90
textfile.check_text_lines(oldlines)
91
textfile.check_text_lines(newlines)
93
if sequence_matcher is None:
94
sequence_matcher = patiencediff.PatienceSequenceMatcher
95
ud = patiencediff.unified_diff(oldlines, newlines,
96
fromfile=old_filename.encode(path_encoding, 'replace'),
97
tofile=new_filename.encode(path_encoding, 'replace'),
98
sequencematcher=sequence_matcher)
101
if len(ud) == 0: # Identical contents, nothing to do
103
# work-around for difflib being too smart for its own good
104
# if /dev/null is "1,0", patch won't recognize it as /dev/null
106
ud[2] = ud[2].replace('-1,0', '-0,0')
108
ud[2] = ud[2].replace('+1,0', '+0,0')
112
if not line.endswith('\n'):
113
to_file.write("\n\\ No newline at end of file\n")
117
def _spawn_external_diff(diffcmd, capture_errors=True):
118
"""Spawn the externall diff process, and return the child handle.
120
:param diffcmd: The command list to spawn
121
:param capture_errors: Capture stderr as well as setting LANG=C
122
and LC_ALL=C. This lets us read and understand the output of diff,
123
and respond to any errors.
124
:return: A Popen object.
127
# construct minimal environment
129
path = os.environ.get('PATH')
132
env['LANGUAGE'] = 'C' # on win32 only LANGUAGE has effect
135
stderr = subprocess.PIPE
141
pipe = subprocess.Popen(diffcmd,
142
stdin=subprocess.PIPE,
143
stdout=subprocess.PIPE,
147
if e.errno == errno.ENOENT:
148
raise errors.NoDiff(str(e))
154
def external_diff(old_filename, oldlines, new_filename, newlines, to_file,
156
"""Display a diff by calling out to the external diff program."""
157
# make sure our own output is properly ordered before the diff
160
oldtmp_fd, old_abspath = tempfile.mkstemp(prefix='bzr-diff-old-')
161
newtmp_fd, new_abspath = tempfile.mkstemp(prefix='bzr-diff-new-')
162
oldtmpf = os.fdopen(oldtmp_fd, 'wb')
163
newtmpf = os.fdopen(newtmp_fd, 'wb')
166
# TODO: perhaps a special case for comparing to or from the empty
167
# sequence; can just use /dev/null on Unix
169
# TODO: if either of the files being compared already exists as a
170
# regular named file (e.g. in the working directory) then we can
171
# compare directly to that, rather than copying it.
173
oldtmpf.writelines(oldlines)
174
newtmpf.writelines(newlines)
181
if sys.platform == 'win32':
182
# Popen doesn't do the proper encoding for external commands
183
# Since we are dealing with an ANSI api, use mbcs encoding
184
old_filename = old_filename.encode('mbcs')
185
new_filename = new_filename.encode('mbcs')
187
'--label', old_filename,
189
'--label', new_filename,
194
# diff only allows one style to be specified; they don't override.
195
# note that some of these take optargs, and the optargs can be
196
# directly appended to the options.
197
# this is only an approximate parser; it doesn't properly understand
199
for s in ['-c', '-u', '-C', '-U',
204
'-y', '--side-by-side',
216
diffcmd.extend(diff_opts)
218
pipe = _spawn_external_diff(diffcmd, capture_errors=True)
219
out,err = pipe.communicate()
222
# internal_diff() adds a trailing newline, add one here for consistency
225
# 'diff' gives retcode == 2 for all sorts of errors
226
# one of those is 'Binary files differ'.
227
# Bad options could also be the problem.
228
# 'Binary files' is not a real error, so we suppress that error.
231
# Since we got here, we want to make sure to give an i18n error
232
pipe = _spawn_external_diff(diffcmd, capture_errors=False)
233
out, err = pipe.communicate()
235
# Write out the new i18n diff response
236
to_file.write(out+'\n')
237
if pipe.returncode != 2:
238
raise errors.BzrError(
239
'external diff failed with exit code 2'
240
' when run with LANG=C and LC_ALL=C,'
241
' but not when run natively: %r' % (diffcmd,))
243
first_line = lang_c_out.split('\n', 1)[0]
244
# Starting with diffutils 2.8.4 the word "binary" was dropped.
245
m = re.match('^(binary )?files.*differ$', first_line, re.I)
247
raise errors.BzrError('external diff failed with exit code 2;'
248
' command: %r' % (diffcmd,))
250
# Binary files differ, just return
253
# If we got to here, we haven't written out the output of diff
257
# returns 1 if files differ; that's OK
259
msg = 'signal %d' % (-rc)
261
msg = 'exit code %d' % rc
263
raise errors.BzrError('external diff failed with %s; command: %r'
268
oldtmpf.close() # and delete
270
# Clean up. Warn in case the files couldn't be deleted
271
# (in case windows still holds the file open, but not
272
# if the files have already been deleted)
274
os.remove(old_abspath)
276
if e.errno not in (errno.ENOENT,):
277
warning('Failed to delete temporary file: %s %s',
280
os.remove(new_abspath)
282
if e.errno not in (errno.ENOENT,):
283
warning('Failed to delete temporary file: %s %s',
287
def get_trees_and_branches_to_diff_locked(
288
path_list, revision_specs, old_url, new_url, add_cleanup, apply_view=True):
289
"""Get the trees and specific files to diff given a list of paths.
291
This method works out the trees to be diff'ed and the files of
292
interest within those trees.
295
the list of arguments passed to the diff command
296
:param revision_specs:
297
Zero, one or two RevisionSpecs from the diff command line,
298
saying what revisions to compare.
300
The url of the old branch or tree. If None, the tree to use is
301
taken from the first path, if any, or the current working tree.
303
The url of the new branch or tree. If None, the tree to use is
304
taken from the first path, if any, or the current working tree.
306
a callable like Command.add_cleanup. get_trees_and_branches_to_diff
307
will register cleanups that must be run to unlock the trees, etc.
309
if True and a view is set, apply the view or check that the paths
312
a tuple of (old_tree, new_tree, old_branch, new_branch,
313
specific_files, extra_trees) where extra_trees is a sequence of
314
additional trees to search in for file-ids. The trees and branches
315
will be read-locked until the cleanups registered via the add_cleanup
318
# Get the old and new revision specs
319
old_revision_spec = None
320
new_revision_spec = None
321
if revision_specs is not None:
322
if len(revision_specs) > 0:
323
old_revision_spec = revision_specs[0]
325
old_url = old_revision_spec.get_branch()
326
if len(revision_specs) > 1:
327
new_revision_spec = revision_specs[1]
329
new_url = new_revision_spec.get_branch()
332
make_paths_wt_relative = True
333
consider_relpath = True
334
if path_list is None or len(path_list) == 0:
335
# If no path is given, the current working tree is used
336
default_location = u'.'
337
consider_relpath = False
338
elif old_url is not None and new_url is not None:
339
other_paths = path_list
340
make_paths_wt_relative = False
342
default_location = path_list[0]
343
other_paths = path_list[1:]
345
def lock_tree_or_branch(wt, br):
348
add_cleanup(wt.unlock)
351
add_cleanup(br.unlock)
353
# Get the old location
356
old_url = default_location
357
working_tree, branch, relpath = \
358
bzrdir.BzrDir.open_containing_tree_or_branch(old_url)
359
lock_tree_or_branch(working_tree, branch)
360
if consider_relpath and relpath != '':
361
if working_tree is not None and apply_view:
362
views.check_path_in_view(working_tree, relpath)
363
specific_files.append(relpath)
364
old_tree = _get_tree_to_diff(old_revision_spec, working_tree, branch)
367
# Get the new location
369
new_url = default_location
370
if new_url != old_url:
371
working_tree, branch, relpath = \
372
bzrdir.BzrDir.open_containing_tree_or_branch(new_url)
373
lock_tree_or_branch(working_tree, branch)
374
if consider_relpath and relpath != '':
375
if working_tree is not None and apply_view:
376
views.check_path_in_view(working_tree, relpath)
377
specific_files.append(relpath)
378
new_tree = _get_tree_to_diff(new_revision_spec, working_tree, branch,
379
basis_is_default=working_tree is None)
382
# Get the specific files (all files is None, no files is [])
383
if make_paths_wt_relative and working_tree is not None:
384
other_paths = working_tree.safe_relpath_files(
386
apply_view=apply_view)
387
specific_files.extend(other_paths)
388
if len(specific_files) == 0:
389
specific_files = None
390
if (working_tree is not None and working_tree.supports_views()
392
view_files = working_tree.views.lookup_view()
394
specific_files = view_files
395
view_str = views.view_display_str(view_files)
396
note("*** Ignoring files outside view. View is %s" % view_str)
398
# Get extra trees that ought to be searched for file-ids
400
if working_tree is not None and working_tree not in (old_tree, new_tree):
401
extra_trees = (working_tree,)
402
return (old_tree, new_tree, old_branch, new_branch,
403
specific_files, extra_trees)
406
def _get_tree_to_diff(spec, tree=None, branch=None, basis_is_default=True):
407
if branch is None and tree is not None:
409
if spec is None or spec.spec is None:
412
return tree.basis_tree()
414
return branch.basis_tree()
417
return spec.as_tree(branch)
420
def show_diff_trees(old_tree, new_tree, to_file, specific_files=None,
421
external_diff_options=None,
422
old_label='a/', new_label='b/',
424
path_encoding='utf8',
427
"""Show in text form the changes from one tree to another.
429
:param to_file: The output stream.
430
:param specific_files: Include only changes to these files - None for all
432
:param external_diff_options: If set, use an external GNU diff and pass
434
:param extra_trees: If set, more Trees to use for looking up file ids
435
:param path_encoding: If set, the path will be encoded as specified,
436
otherwise is supposed to be utf8
437
:param format_cls: Formatter class (DiffTree subclass)
439
if format_cls is None:
440
format_cls = DiffTree
443
if extra_trees is not None:
444
for tree in extra_trees:
448
differ = format_cls.from_trees_options(old_tree, new_tree, to_file,
450
external_diff_options,
451
old_label, new_label, using)
452
return differ.show_diff(specific_files, extra_trees)
455
if extra_trees is not None:
456
for tree in extra_trees:
462
def _patch_header_date(tree, file_id, path):
463
"""Returns a timestamp suitable for use in a patch header."""
465
mtime = tree.get_file_mtime(file_id, path)
466
except errors.FileTimestampUnavailable:
468
return timestamp.format_patch_date(mtime)
471
def get_executable_change(old_is_x, new_is_x):
472
descr = { True:"+x", False:"-x", None:"??" }
473
if old_is_x != new_is_x:
474
return ["%s to %s" % (descr[old_is_x], descr[new_is_x],)]
479
class DiffPath(object):
480
"""Base type for command object that compare files"""
482
# The type or contents of the file were unsuitable for diffing
483
CANNOT_DIFF = 'CANNOT_DIFF'
484
# The file has changed in a semantic way
486
# The file content may have changed, but there is no semantic change
487
UNCHANGED = 'UNCHANGED'
489
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8'):
492
:param old_tree: The tree to show as the old tree in the comparison
493
:param new_tree: The tree to show as new in the comparison
494
:param to_file: The file to write comparison data to
495
:param path_encoding: The character encoding to write paths in
497
self.old_tree = old_tree
498
self.new_tree = new_tree
499
self.to_file = to_file
500
self.path_encoding = path_encoding
506
def from_diff_tree(klass, diff_tree):
507
return klass(diff_tree.old_tree, diff_tree.new_tree,
508
diff_tree.to_file, diff_tree.path_encoding)
511
def _diff_many(differs, file_id, old_path, new_path, old_kind, new_kind):
512
for file_differ in differs:
513
result = file_differ.diff(file_id, old_path, new_path, old_kind,
515
if result is not DiffPath.CANNOT_DIFF:
518
return DiffPath.CANNOT_DIFF
521
class DiffKindChange(object):
522
"""Special differ for file kind changes.
524
Represents kind change as deletion + creation. Uses the other differs
527
def __init__(self, differs):
528
self.differs = differs
534
def from_diff_tree(klass, diff_tree):
535
return klass(diff_tree.differs)
537
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
538
"""Perform comparison
540
:param file_id: The file_id of the file to compare
541
:param old_path: Path of the file in the old tree
542
:param new_path: Path of the file in the new tree
543
:param old_kind: Old file-kind of the file
544
:param new_kind: New file-kind of the file
546
if None in (old_kind, new_kind):
547
return DiffPath.CANNOT_DIFF
548
result = DiffPath._diff_many(self.differs, file_id, old_path,
549
new_path, old_kind, None)
550
if result is DiffPath.CANNOT_DIFF:
552
return DiffPath._diff_many(self.differs, file_id, old_path, new_path,
556
class DiffDirectory(DiffPath):
558
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
559
"""Perform comparison between two directories. (dummy)
562
if 'directory' not in (old_kind, new_kind):
563
return self.CANNOT_DIFF
564
if old_kind not in ('directory', None):
565
return self.CANNOT_DIFF
566
if new_kind not in ('directory', None):
567
return self.CANNOT_DIFF
571
class DiffSymlink(DiffPath):
573
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
574
"""Perform comparison between two symlinks
576
:param file_id: The file_id of the file to compare
577
:param old_path: Path of the file in the old tree
578
:param new_path: Path of the file in the new tree
579
:param old_kind: Old file-kind of the file
580
:param new_kind: New file-kind of the file
582
if 'symlink' not in (old_kind, new_kind):
583
return self.CANNOT_DIFF
584
if old_kind == 'symlink':
585
old_target = self.old_tree.get_symlink_target(file_id)
586
elif old_kind is None:
589
return self.CANNOT_DIFF
590
if new_kind == 'symlink':
591
new_target = self.new_tree.get_symlink_target(file_id)
592
elif new_kind is None:
595
return self.CANNOT_DIFF
596
return self.diff_symlink(old_target, new_target)
598
def diff_symlink(self, old_target, new_target):
599
if old_target is None:
600
self.to_file.write('=== target is %r\n' % new_target)
601
elif new_target is None:
602
self.to_file.write('=== target was %r\n' % old_target)
604
self.to_file.write('=== target changed %r => %r\n' %
605
(old_target, new_target))
609
class DiffText(DiffPath):
611
# GNU Patch uses the epoch date to detect files that are being added
612
# or removed in a diff.
613
EPOCH_DATE = '1970-01-01 00:00:00 +0000'
615
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
616
old_label='', new_label='', text_differ=internal_diff):
617
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
618
self.text_differ = text_differ
619
self.old_label = old_label
620
self.new_label = new_label
621
self.path_encoding = path_encoding
623
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
624
"""Compare two files in unified diff format
626
:param file_id: The file_id of the file to compare
627
:param old_path: Path of the file in the old tree
628
:param new_path: Path of the file in the new tree
629
:param old_kind: Old file-kind of the file
630
:param new_kind: New file-kind of the file
632
if 'file' not in (old_kind, new_kind):
633
return self.CANNOT_DIFF
634
from_file_id = to_file_id = file_id
635
if old_kind == 'file':
636
old_date = _patch_header_date(self.old_tree, file_id, old_path)
637
elif old_kind is None:
638
old_date = self.EPOCH_DATE
641
return self.CANNOT_DIFF
642
if new_kind == 'file':
643
new_date = _patch_header_date(self.new_tree, file_id, new_path)
644
elif new_kind is None:
645
new_date = self.EPOCH_DATE
648
return self.CANNOT_DIFF
649
from_label = '%s%s\t%s' % (self.old_label, old_path, old_date)
650
to_label = '%s%s\t%s' % (self.new_label, new_path, new_date)
651
return self.diff_text(from_file_id, to_file_id, from_label, to_label,
654
def diff_text(self, from_file_id, to_file_id, from_label, to_label,
655
from_path=None, to_path=None):
656
"""Diff the content of given files in two trees
658
:param from_file_id: The id of the file in the from tree. If None,
659
the file is not present in the from tree.
660
:param to_file_id: The id of the file in the to tree. This may refer
661
to a different file from from_file_id. If None,
662
the file is not present in the to tree.
663
:param from_path: The path in the from tree or None if unknown.
664
:param to_path: The path in the to tree or None if unknown.
666
def _get_text(tree, file_id, path):
667
if file_id is not None:
668
return tree.get_file_lines(file_id, path)
672
from_text = _get_text(self.old_tree, from_file_id, from_path)
673
to_text = _get_text(self.new_tree, to_file_id, to_path)
674
self.text_differ(from_label, from_text, to_label, to_text,
675
self.to_file, path_encoding=self.path_encoding)
676
except errors.BinaryFile:
678
("Binary files %s and %s differ\n" %
679
(from_label, to_label)).encode(self.path_encoding,'replace'))
683
class DiffFromTool(DiffPath):
685
def __init__(self, command_template, old_tree, new_tree, to_file,
686
path_encoding='utf-8'):
687
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
688
self.command_template = command_template
689
self._root = osutils.mkdtemp(prefix='bzr-diff-')
692
def from_string(klass, command_string, old_tree, new_tree, to_file,
693
path_encoding='utf-8'):
694
command_template = cmdline.split(command_string)
695
if '@' not in command_string:
696
command_template.extend(['@old_path', '@new_path'])
697
return klass(command_template, old_tree, new_tree, to_file,
701
def make_from_diff_tree(klass, command_string, external_diff_options=None):
702
def from_diff_tree(diff_tree):
703
full_command_string = [command_string]
704
if external_diff_options is not None:
705
full_command_string += ' ' + external_diff_options
706
return klass.from_string(full_command_string, diff_tree.old_tree,
707
diff_tree.new_tree, diff_tree.to_file)
708
return from_diff_tree
710
def _get_command(self, old_path, new_path):
711
my_map = {'old_path': old_path, 'new_path': new_path}
712
command = [AtTemplate(t).substitute(my_map) for t in
713
self.command_template]
714
if sys.platform == 'win32': # Popen doesn't accept unicode on win32
717
if isinstance(c, unicode):
718
command_encoded.append(c.encode('mbcs'))
720
command_encoded.append(c)
721
return command_encoded
725
def _execute(self, old_path, new_path):
726
command = self._get_command(old_path, new_path)
728
proc = subprocess.Popen(command, stdout=subprocess.PIPE,
731
if e.errno == errno.ENOENT:
732
raise errors.ExecutableMissing(command[0])
735
self.to_file.write(proc.stdout.read())
738
def _try_symlink_root(self, tree, prefix):
739
if (getattr(tree, 'abspath', None) is None
740
or not osutils.host_os_dereferences_symlinks()):
743
os.symlink(tree.abspath(''), osutils.pathjoin(self._root, prefix))
745
if e.errno != errno.EEXIST:
751
"""Returns safe encoding for passing file path to diff tool"""
752
if sys.platform == 'win32':
755
# Don't fallback to 'utf-8' because subprocess may not be able to
756
# handle utf-8 correctly when locale is not utf-8.
757
return sys.getfilesystemencoding() or 'ascii'
759
def _is_safepath(self, path):
760
"""Return true if `path` may be able to pass to subprocess."""
763
return path == path.encode(fenc).decode(fenc)
767
def _safe_filename(self, prefix, relpath):
768
"""Replace unsafe character in `relpath` then join `self._root`,
769
`prefix` and `relpath`."""
771
# encoded_str.replace('?', '_') may break multibyte char.
772
# So we should encode, decode, then replace(u'?', u'_')
773
relpath_tmp = relpath.encode(fenc, 'replace').decode(fenc, 'replace')
774
relpath_tmp = relpath_tmp.replace(u'?', u'_')
775
return osutils.pathjoin(self._root, prefix, relpath_tmp)
777
def _write_file(self, file_id, tree, prefix, relpath, force_temp=False,
779
if not force_temp and isinstance(tree, WorkingTree):
780
full_path = tree.abspath(tree.id2path(file_id))
781
if self._is_safepath(full_path):
784
full_path = self._safe_filename(prefix, relpath)
785
if not force_temp and self._try_symlink_root(tree, prefix):
787
parent_dir = osutils.dirname(full_path)
789
os.makedirs(parent_dir)
791
if e.errno != errno.EEXIST:
793
source = tree.get_file(file_id, relpath)
795
target = open(full_path, 'wb')
797
osutils.pumpfile(source, target)
803
mtime = tree.get_file_mtime(file_id)
804
except errors.FileTimestampUnavailable:
807
os.utime(full_path, (mtime, mtime))
809
osutils.make_readonly(full_path)
812
def _prepare_files(self, file_id, old_path, new_path, force_temp=False,
813
allow_write_new=False):
814
old_disk_path = self._write_file(file_id, self.old_tree, 'old',
815
old_path, force_temp)
816
new_disk_path = self._write_file(file_id, self.new_tree, 'new',
817
new_path, force_temp,
818
allow_write=allow_write_new)
819
return old_disk_path, new_disk_path
823
osutils.rmtree(self._root)
825
if e.errno != errno.ENOENT:
826
mutter("The temporary directory \"%s\" was not "
827
"cleanly removed: %s." % (self._root, e))
829
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
830
if (old_kind, new_kind) != ('file', 'file'):
831
return DiffPath.CANNOT_DIFF
832
(old_disk_path, new_disk_path) = self._prepare_files(
833
file_id, old_path, new_path)
834
self._execute(old_disk_path, new_disk_path)
836
def edit_file(self, file_id):
837
"""Use this tool to edit a file.
839
A temporary copy will be edited, and the new contents will be
842
:param file_id: The id of the file to edit.
843
:return: The new contents of the file.
845
old_path = self.old_tree.id2path(file_id)
846
new_path = self.new_tree.id2path(file_id)
847
old_abs_path, new_abs_path = self._prepare_files(
848
file_id, old_path, new_path,
849
allow_write_new=True,
851
command = self._get_command(old_abs_path, new_abs_path)
852
subprocess.call(command, cwd=self._root)
853
new_file = open(new_abs_path, 'rb')
855
return new_file.read()
860
class DiffTree(object):
861
"""Provides textual representations of the difference between two trees.
863
A DiffTree examines two trees and where a file-id has altered
864
between them, generates a textual representation of the difference.
865
DiffTree uses a sequence of DiffPath objects which are each
866
given the opportunity to handle a given altered fileid. The list
867
of DiffPath objects can be extended globally by appending to
868
DiffTree.diff_factories, or for a specific diff operation by
869
supplying the extra_factories option to the appropriate method.
872
# list of factories that can provide instances of DiffPath objects
873
# may be extended by plugins.
874
diff_factories = [DiffSymlink.from_diff_tree,
875
DiffDirectory.from_diff_tree]
877
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
878
diff_text=None, extra_factories=None):
881
:param old_tree: Tree to show as old in the comparison
882
:param new_tree: Tree to show as new in the comparison
883
:param to_file: File to write comparision to
884
:param path_encoding: Character encoding to write paths in
885
:param diff_text: DiffPath-type object to use as a last resort for
887
:param extra_factories: Factories of DiffPaths to try before any other
889
if diff_text is None:
890
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
891
'', '', internal_diff)
892
self.old_tree = old_tree
893
self.new_tree = new_tree
894
self.to_file = to_file
895
self.path_encoding = path_encoding
897
if extra_factories is not None:
898
self.differs.extend(f(self) for f in extra_factories)
899
self.differs.extend(f(self) for f in self.diff_factories)
900
self.differs.extend([diff_text, DiffKindChange.from_diff_tree(self)])
903
def from_trees_options(klass, old_tree, new_tree, to_file,
904
path_encoding, external_diff_options, old_label,
906
"""Factory for producing a DiffTree.
908
Designed to accept options used by show_diff_trees.
910
:param old_tree: The tree to show as old in the comparison
911
:param new_tree: The tree to show as new in the comparison
912
:param to_file: File to write comparisons to
913
:param path_encoding: Character encoding to use for writing paths
914
:param external_diff_options: If supplied, use the installed diff
915
binary to perform file comparison, using supplied options.
916
:param old_label: Prefix to use for old file labels
917
:param new_label: Prefix to use for new file labels
918
:param using: Commandline to use to invoke an external diff tool
920
if using is not None:
921
extra_factories = [DiffFromTool.make_from_diff_tree(using, external_diff_options)]
924
if external_diff_options:
925
opts = external_diff_options.split()
926
def diff_file(olab, olines, nlab, nlines, to_file, path_encoding=None):
927
""":param path_encoding: not used but required
928
to match the signature of internal_diff.
930
external_diff(olab, olines, nlab, nlines, to_file, opts)
932
diff_file = internal_diff
933
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
934
old_label, new_label, diff_file)
935
return klass(old_tree, new_tree, to_file, path_encoding, diff_text,
938
def show_diff(self, specific_files, extra_trees=None):
939
"""Write tree diff to self.to_file
941
:param specific_files: the specific files to compare (recursive)
942
:param extra_trees: extra trees to use for mapping paths to file_ids
945
return self._show_diff(specific_files, extra_trees)
947
for differ in self.differs:
950
def _show_diff(self, specific_files, extra_trees):
951
# TODO: Generation of pseudo-diffs for added/deleted files could
952
# be usefully made into a much faster special case.
953
iterator = self.new_tree.iter_changes(self.old_tree,
954
specific_files=specific_files,
955
extra_trees=extra_trees,
956
require_versioned=True)
958
def changes_key(change):
959
old_path, new_path = change[1]
964
def get_encoded_path(path):
966
return path.encode(self.path_encoding, "replace")
967
for (file_id, paths, changed_content, versioned, parent, name, kind,
968
executable) in sorted(iterator, key=changes_key):
969
# The root does not get diffed, and items with no known kind (that
970
# is, missing) in both trees are skipped as well.
971
if parent == (None, None) or kind == (None, None):
973
oldpath, newpath = paths
974
oldpath_encoded = get_encoded_path(paths[0])
975
newpath_encoded = get_encoded_path(paths[1])
976
old_present = (kind[0] is not None and versioned[0])
977
new_present = (kind[1] is not None and versioned[1])
978
renamed = (parent[0], name[0]) != (parent[1], name[1])
980
properties_changed = []
981
properties_changed.extend(get_executable_change(executable[0], executable[1]))
983
if properties_changed:
984
prop_str = " (properties changed: %s)" % (", ".join(properties_changed),)
988
if (old_present, new_present) == (True, False):
989
self.to_file.write("=== removed %s '%s'\n" %
990
(kind[0], oldpath_encoded))
992
elif (old_present, new_present) == (False, True):
993
self.to_file.write("=== added %s '%s'\n" %
994
(kind[1], newpath_encoded))
997
self.to_file.write("=== renamed %s '%s' => '%s'%s\n" %
998
(kind[0], oldpath_encoded, newpath_encoded, prop_str))
1000
# if it was produced by iter_changes, it must be
1001
# modified *somehow*, either content or execute bit.
1002
self.to_file.write("=== modified %s '%s'%s\n" % (kind[0],
1003
newpath_encoded, prop_str))
1005
self._diff(file_id, oldpath, newpath, kind[0], kind[1])
1011
def diff(self, file_id, old_path, new_path):
1012
"""Perform a diff of a single file
1014
:param file_id: file-id of the file
1015
:param old_path: The path of the file in the old tree
1016
:param new_path: The path of the file in the new tree
1019
old_kind = self.old_tree.kind(file_id)
1020
except (errors.NoSuchId, errors.NoSuchFile):
1023
new_kind = self.new_tree.kind(file_id)
1024
except (errors.NoSuchId, errors.NoSuchFile):
1026
self._diff(file_id, old_path, new_path, old_kind, new_kind)
1029
def _diff(self, file_id, old_path, new_path, old_kind, new_kind):
1030
result = DiffPath._diff_many(self.differs, file_id, old_path,
1031
new_path, old_kind, new_kind)
1032
if result is DiffPath.CANNOT_DIFF:
1033
error_path = new_path
1034
if error_path is None:
1035
error_path = old_path
1036
raise errors.NoDiffFound(error_path)
1039
format_registry = Registry()
1040
format_registry.register('default', DiffTree)
added
removed
bzrlib/diff.py
