1
# Copyright (C) 2005-2010 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(), """
31
branch as _mod_branch,
43
from bzrlib.workingtree import WorkingTree
46
from bzrlib.registry import (
49
from bzrlib.symbol_versioning import (
53
from bzrlib.trace import mutter, note, warning
56
class AtTemplate(string.Template):
57
"""Templating class that uses @ instead of $."""
62
# TODO: Rather than building a changeset object, we should probably
63
# invoke callbacks on an object. That object can either accumulate a
64
# list, write them out directly, etc etc.
67
class _PrematchedMatcher(difflib.SequenceMatcher):
68
"""Allow SequenceMatcher operations to use predetermined blocks"""
70
def __init__(self, matching_blocks):
71
difflib.SequenceMatcher(self, None, None)
72
self.matching_blocks = matching_blocks
76
def internal_diff(old_filename, oldlines, new_filename, newlines, to_file,
77
allow_binary=False, sequence_matcher=None,
78
path_encoding='utf8'):
79
# FIXME: difflib is wrong if there is no trailing newline.
80
# The syntax used by patch seems to be "\ No newline at
81
# end of file" following the last diff line from that
82
# file. This is not trivial to insert into the
83
# unified_diff output and it might be better to just fix
84
# or replace that function.
86
# In the meantime we at least make sure the patch isn't
90
# Special workaround for Python2.3, where difflib fails if
91
# both sequences are empty.
92
if not oldlines and not newlines:
95
if allow_binary is False:
96
textfile.check_text_lines(oldlines)
97
textfile.check_text_lines(newlines)
99
if sequence_matcher is None:
100
sequence_matcher = patiencediff.PatienceSequenceMatcher
101
ud = patiencediff.unified_diff(oldlines, newlines,
102
fromfile=old_filename.encode(path_encoding, 'replace'),
103
tofile=new_filename.encode(path_encoding, 'replace'),
104
sequencematcher=sequence_matcher)
107
if len(ud) == 0: # Identical contents, nothing to do
109
# work-around for difflib being too smart for its own good
110
# if /dev/null is "1,0", patch won't recognize it as /dev/null
112
ud[2] = ud[2].replace('-1,0', '-0,0')
114
ud[2] = ud[2].replace('+1,0', '+0,0')
118
if not line.endswith('\n'):
119
to_file.write("\n\\ No newline at end of file\n")
123
def _spawn_external_diff(diffcmd, capture_errors=True):
124
"""Spawn the externall diff process, and return the child handle.
126
:param diffcmd: The command list to spawn
127
:param capture_errors: Capture stderr as well as setting LANG=C
128
and LC_ALL=C. This lets us read and understand the output of diff,
129
and respond to any errors.
130
:return: A Popen object.
133
# construct minimal environment
135
path = os.environ.get('PATH')
138
env['LANGUAGE'] = 'C' # on win32 only LANGUAGE has effect
141
stderr = subprocess.PIPE
147
pipe = subprocess.Popen(diffcmd,
148
stdin=subprocess.PIPE,
149
stdout=subprocess.PIPE,
153
if e.errno == errno.ENOENT:
154
raise errors.NoDiff(str(e))
160
def external_diff(old_filename, oldlines, new_filename, newlines, to_file,
162
"""Display a diff by calling out to the external diff program."""
163
# make sure our own output is properly ordered before the diff
166
oldtmp_fd, old_abspath = tempfile.mkstemp(prefix='bzr-diff-old-')
167
newtmp_fd, new_abspath = tempfile.mkstemp(prefix='bzr-diff-new-')
168
oldtmpf = os.fdopen(oldtmp_fd, 'wb')
169
newtmpf = os.fdopen(newtmp_fd, 'wb')
172
# TODO: perhaps a special case for comparing to or from the empty
173
# sequence; can just use /dev/null on Unix
175
# TODO: if either of the files being compared already exists as a
176
# regular named file (e.g. in the working directory) then we can
177
# compare directly to that, rather than copying it.
179
oldtmpf.writelines(oldlines)
180
newtmpf.writelines(newlines)
187
if sys.platform == 'win32':
188
# Popen doesn't do the proper encoding for external commands
189
# Since we are dealing with an ANSI api, use mbcs encoding
190
old_filename = old_filename.encode('mbcs')
191
new_filename = new_filename.encode('mbcs')
193
'--label', old_filename,
195
'--label', new_filename,
200
# diff only allows one style to be specified; they don't override.
201
# note that some of these take optargs, and the optargs can be
202
# directly appended to the options.
203
# this is only an approximate parser; it doesn't properly understand
205
for s in ['-c', '-u', '-C', '-U',
210
'-y', '--side-by-side',
222
diffcmd.extend(diff_opts)
224
pipe = _spawn_external_diff(diffcmd, capture_errors=True)
225
out,err = pipe.communicate()
228
# internal_diff() adds a trailing newline, add one here for consistency
231
# 'diff' gives retcode == 2 for all sorts of errors
232
# one of those is 'Binary files differ'.
233
# Bad options could also be the problem.
234
# 'Binary files' is not a real error, so we suppress that error.
237
# Since we got here, we want to make sure to give an i18n error
238
pipe = _spawn_external_diff(diffcmd, capture_errors=False)
239
out, err = pipe.communicate()
241
# Write out the new i18n diff response
242
to_file.write(out+'\n')
243
if pipe.returncode != 2:
244
raise errors.BzrError(
245
'external diff failed with exit code 2'
246
' when run with LANG=C and LC_ALL=C,'
247
' but not when run natively: %r' % (diffcmd,))
249
first_line = lang_c_out.split('\n', 1)[0]
250
# Starting with diffutils 2.8.4 the word "binary" was dropped.
251
m = re.match('^(binary )?files.*differ$', first_line, re.I)
253
raise errors.BzrError('external diff failed with exit code 2;'
254
' command: %r' % (diffcmd,))
256
# Binary files differ, just return
259
# If we got to here, we haven't written out the output of diff
263
# returns 1 if files differ; that's OK
265
msg = 'signal %d' % (-rc)
267
msg = 'exit code %d' % rc
269
raise errors.BzrError('external diff failed with %s; command: %r'
274
oldtmpf.close() # and delete
276
# Clean up. Warn in case the files couldn't be deleted
277
# (in case windows still holds the file open, but not
278
# if the files have already been deleted)
280
os.remove(old_abspath)
282
if e.errno not in (errno.ENOENT,):
283
warning('Failed to delete temporary file: %s %s',
286
os.remove(new_abspath)
288
if e.errno not in (errno.ENOENT,):
289
warning('Failed to delete temporary file: %s %s',
293
@deprecated_function(deprecated_in((2, 2, 0)))
294
def get_trees_and_branches_to_diff(path_list, revision_specs, old_url, new_url,
296
"""Get the trees and specific files to diff given a list of paths.
298
This method works out the trees to be diff'ed and the files of
299
interest within those trees.
302
the list of arguments passed to the diff command
303
:param revision_specs:
304
Zero, one or two RevisionSpecs from the diff command line,
305
saying what revisions to compare.
307
The url of the old branch or tree. If None, the tree to use is
308
taken from the first path, if any, or the current working tree.
310
The url of the new branch or tree. If None, the tree to use is
311
taken from the first path, if any, or the current working tree.
313
if True and a view is set, apply the view or check that the paths
316
a tuple of (old_tree, new_tree, old_branch, new_branch,
317
specific_files, extra_trees) where extra_trees is a sequence of
318
additional trees to search in for file-ids. The trees and branches
321
op = cleanup.OperationWithCleanups(get_trees_and_branches_to_diff_locked)
322
return op.run_simple(path_list, revision_specs, old_url, new_url,
323
op.add_cleanup, apply_view=apply_view)
326
def get_trees_and_branches_to_diff_locked(
327
path_list, revision_specs, old_url, new_url, add_cleanup, apply_view=True):
328
"""Get the trees and specific files to diff given a list of paths.
330
This method works out the trees to be diff'ed and the files of
331
interest within those trees.
334
the list of arguments passed to the diff command
335
:param revision_specs:
336
Zero, one or two RevisionSpecs from the diff command line,
337
saying what revisions to compare.
339
The url of the old branch or tree. If None, the tree to use is
340
taken from the first path, if any, or the current working tree.
342
The url of the new branch or tree. If None, the tree to use is
343
taken from the first path, if any, or the current working tree.
345
a callable like Command.add_cleanup. get_trees_and_branches_to_diff
346
will register cleanups that must be run to unlock the trees, etc.
348
if True and a view is set, apply the view or check that the paths
351
a tuple of (old_tree, new_tree, old_branch, new_branch,
352
specific_files, extra_trees) where extra_trees is a sequence of
353
additional trees to search in for file-ids. The trees and branches
354
will be read-locked until the cleanups registered via the add_cleanup
357
# Get the old and new revision specs
358
old_revision_spec = None
359
new_revision_spec = None
360
if revision_specs is not None:
361
if len(revision_specs) > 0:
362
old_revision_spec = revision_specs[0]
364
old_url = old_revision_spec.get_branch()
365
if len(revision_specs) > 1:
366
new_revision_spec = revision_specs[1]
368
new_url = new_revision_spec.get_branch()
371
make_paths_wt_relative = True
372
consider_relpath = True
373
if path_list is None or len(path_list) == 0:
374
# If no path is given, the current working tree is used
375
default_location = u'.'
376
consider_relpath = False
377
elif old_url is not None and new_url is not None:
378
other_paths = path_list
379
make_paths_wt_relative = False
381
default_location = path_list[0]
382
other_paths = path_list[1:]
384
def lock_tree_or_branch(wt, br):
387
add_cleanup(wt.unlock)
390
add_cleanup(br.unlock)
392
# Get the old location
395
old_url = default_location
396
working_tree, branch, relpath = \
397
bzrdir.BzrDir.open_containing_tree_or_branch(old_url)
398
lock_tree_or_branch(working_tree, branch)
399
if consider_relpath and relpath != '':
400
if working_tree is not None and apply_view:
401
views.check_path_in_view(working_tree, relpath)
402
specific_files.append(relpath)
403
old_tree = _get_tree_to_diff(old_revision_spec, working_tree, branch)
406
# Get the new location
408
new_url = default_location
409
if new_url != old_url:
410
working_tree, branch, relpath = \
411
bzrdir.BzrDir.open_containing_tree_or_branch(new_url)
412
lock_tree_or_branch(working_tree, branch)
413
if consider_relpath and relpath != '':
414
if working_tree is not None and apply_view:
415
views.check_path_in_view(working_tree, relpath)
416
specific_files.append(relpath)
417
new_tree = _get_tree_to_diff(new_revision_spec, working_tree, branch,
418
basis_is_default=working_tree is None)
421
# Get the specific files (all files is None, no files is [])
422
if make_paths_wt_relative and working_tree is not None:
423
other_paths = working_tree.safe_relpath_files(
425
apply_view=apply_view)
426
specific_files.extend(other_paths)
427
if len(specific_files) == 0:
428
specific_files = None
429
if (working_tree is not None and working_tree.supports_views()
431
view_files = working_tree.views.lookup_view()
433
specific_files = view_files
434
view_str = views.view_display_str(view_files)
435
note("*** Ignoring files outside view. View is %s" % view_str)
437
# Get extra trees that ought to be searched for file-ids
439
if working_tree is not None and working_tree not in (old_tree, new_tree):
440
extra_trees = (working_tree,)
441
return old_tree, new_tree, old_branch, new_branch, specific_files, extra_trees
444
def _get_tree_to_diff(spec, tree=None, branch=None, basis_is_default=True):
445
if branch is None and tree is not None:
447
if spec is None or spec.spec is None:
450
return tree.basis_tree()
452
return branch.basis_tree()
455
return spec.as_tree(branch)
458
def show_diff_trees(old_tree, new_tree, to_file, specific_files=None,
459
external_diff_options=None,
460
old_label='a/', new_label='b/',
462
path_encoding='utf8',
465
"""Show in text form the changes from one tree to another.
467
:param to_file: The output stream.
468
:param specific_files:Include only changes to these files - None for all
470
:param external_diff_options: If set, use an external GNU diff and pass
472
:param extra_trees: If set, more Trees to use for looking up file ids
473
:param path_encoding: If set, the path will be encoded as specified,
474
otherwise is supposed to be utf8
475
:param format_cls: Formatter class (DiffTree subclass)
477
if format_cls is None:
478
format_cls = DiffTree
481
if extra_trees is not None:
482
for tree in extra_trees:
486
differ = format_cls.from_trees_options(old_tree, new_tree, to_file,
488
external_diff_options,
489
old_label, new_label, using)
490
return differ.show_diff(specific_files, extra_trees)
493
if extra_trees is not None:
494
for tree in extra_trees:
500
def _patch_header_date(tree, file_id, path):
501
"""Returns a timestamp suitable for use in a patch header."""
503
mtime = tree.get_file_mtime(file_id, path)
504
except errors.FileTimestampUnavailable:
506
return timestamp.format_patch_date(mtime)
509
def get_executable_change(old_is_x, new_is_x):
510
descr = { True:"+x", False:"-x", None:"??" }
511
if old_is_x != new_is_x:
512
return ["%s to %s" % (descr[old_is_x], descr[new_is_x],)]
517
class DiffPath(object):
518
"""Base type for command object that compare files"""
520
# The type or contents of the file were unsuitable for diffing
521
CANNOT_DIFF = 'CANNOT_DIFF'
522
# The file has changed in a semantic way
524
# The file content may have changed, but there is no semantic change
525
UNCHANGED = 'UNCHANGED'
527
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8'):
530
:param old_tree: The tree to show as the old tree in the comparison
531
:param new_tree: The tree to show as new in the comparison
532
:param to_file: The file to write comparison data to
533
:param path_encoding: The character encoding to write paths in
535
self.old_tree = old_tree
536
self.new_tree = new_tree
537
self.to_file = to_file
538
self.path_encoding = path_encoding
544
def from_diff_tree(klass, diff_tree):
545
return klass(diff_tree.old_tree, diff_tree.new_tree,
546
diff_tree.to_file, diff_tree.path_encoding)
549
def _diff_many(differs, file_id, old_path, new_path, old_kind, new_kind):
550
for file_differ in differs:
551
result = file_differ.diff(file_id, old_path, new_path, old_kind,
553
if result is not DiffPath.CANNOT_DIFF:
556
return DiffPath.CANNOT_DIFF
559
class DiffKindChange(object):
560
"""Special differ for file kind changes.
562
Represents kind change as deletion + creation. Uses the other differs
565
def __init__(self, differs):
566
self.differs = differs
572
def from_diff_tree(klass, diff_tree):
573
return klass(diff_tree.differs)
575
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
576
"""Perform comparison
578
:param file_id: The file_id of the file to compare
579
:param old_path: Path of the file in the old tree
580
:param new_path: Path of the file in the new tree
581
:param old_kind: Old file-kind of the file
582
:param new_kind: New file-kind of the file
584
if None in (old_kind, new_kind):
585
return DiffPath.CANNOT_DIFF
586
result = DiffPath._diff_many(self.differs, file_id, old_path,
587
new_path, old_kind, None)
588
if result is DiffPath.CANNOT_DIFF:
590
return DiffPath._diff_many(self.differs, file_id, old_path, new_path,
594
class DiffDirectory(DiffPath):
596
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
597
"""Perform comparison between two directories. (dummy)
600
if 'directory' not in (old_kind, new_kind):
601
return self.CANNOT_DIFF
602
if old_kind not in ('directory', None):
603
return self.CANNOT_DIFF
604
if new_kind not in ('directory', None):
605
return self.CANNOT_DIFF
609
class DiffSymlink(DiffPath):
611
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
612
"""Perform comparison between two symlinks
614
:param file_id: The file_id of the file to compare
615
:param old_path: Path of the file in the old tree
616
:param new_path: Path of the file in the new tree
617
:param old_kind: Old file-kind of the file
618
:param new_kind: New file-kind of the file
620
if 'symlink' not in (old_kind, new_kind):
621
return self.CANNOT_DIFF
622
if old_kind == 'symlink':
623
old_target = self.old_tree.get_symlink_target(file_id)
624
elif old_kind is None:
627
return self.CANNOT_DIFF
628
if new_kind == 'symlink':
629
new_target = self.new_tree.get_symlink_target(file_id)
630
elif new_kind is None:
633
return self.CANNOT_DIFF
634
return self.diff_symlink(old_target, new_target)
636
def diff_symlink(self, old_target, new_target):
637
if old_target is None:
638
self.to_file.write('=== target is %r\n' % new_target)
639
elif new_target is None:
640
self.to_file.write('=== target was %r\n' % old_target)
642
self.to_file.write('=== target changed %r => %r\n' %
643
(old_target, new_target))
647
class DiffText(DiffPath):
649
# GNU Patch uses the epoch date to detect files that are being added
650
# or removed in a diff.
651
EPOCH_DATE = '1970-01-01 00:00:00 +0000'
653
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
654
old_label='', new_label='', text_differ=internal_diff):
655
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
656
self.text_differ = text_differ
657
self.old_label = old_label
658
self.new_label = new_label
659
self.path_encoding = path_encoding
661
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
662
"""Compare two files in unified diff format
664
:param file_id: The file_id of the file to compare
665
:param old_path: Path of the file in the old tree
666
:param new_path: Path of the file in the new tree
667
:param old_kind: Old file-kind of the file
668
:param new_kind: New file-kind of the file
670
if 'file' not in (old_kind, new_kind):
671
return self.CANNOT_DIFF
672
from_file_id = to_file_id = file_id
673
if old_kind == 'file':
674
old_date = _patch_header_date(self.old_tree, file_id, old_path)
675
elif old_kind is None:
676
old_date = self.EPOCH_DATE
679
return self.CANNOT_DIFF
680
if new_kind == 'file':
681
new_date = _patch_header_date(self.new_tree, file_id, new_path)
682
elif new_kind is None:
683
new_date = self.EPOCH_DATE
686
return self.CANNOT_DIFF
687
from_label = '%s%s\t%s' % (self.old_label, old_path, old_date)
688
to_label = '%s%s\t%s' % (self.new_label, new_path, new_date)
689
return self.diff_text(from_file_id, to_file_id, from_label, to_label,
692
def diff_text(self, from_file_id, to_file_id, from_label, to_label,
693
from_path=None, to_path=None):
694
"""Diff the content of given files in two trees
696
:param from_file_id: The id of the file in the from tree. If None,
697
the file is not present in the from tree.
698
:param to_file_id: The id of the file in the to tree. This may refer
699
to a different file from from_file_id. If None,
700
the file is not present in the to tree.
701
:param from_path: The path in the from tree or None if unknown.
702
:param to_path: The path in the to tree or None if unknown.
704
def _get_text(tree, file_id, path):
705
if file_id is not None:
706
return tree.get_file_lines(file_id, path)
710
from_text = _get_text(self.old_tree, from_file_id, from_path)
711
to_text = _get_text(self.new_tree, to_file_id, to_path)
712
self.text_differ(from_label, from_text, to_label, to_text,
713
self.to_file, path_encoding=self.path_encoding)
714
except errors.BinaryFile:
716
("Binary files %s and %s differ\n" %
717
(from_label, to_label)).encode(self.path_encoding,'replace'))
721
class DiffFromTool(DiffPath):
723
def __init__(self, command_template, old_tree, new_tree, to_file,
724
path_encoding='utf-8'):
725
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
726
self.command_template = command_template
727
self._root = osutils.mkdtemp(prefix='bzr-diff-')
730
def from_string(klass, command_string, old_tree, new_tree, to_file,
731
path_encoding='utf-8'):
732
command_template = cmdline.split(command_string)
733
if '@' not in command_string:
734
command_template.extend(['@old_path', '@new_path'])
735
return klass(command_template, old_tree, new_tree, to_file,
739
def make_from_diff_tree(klass, command_string, external_diff_options=None):
740
def from_diff_tree(diff_tree):
741
full_command_string = [command_string]
742
if external_diff_options is not None:
743
full_command_string += ' ' + external_diff_options
744
return klass.from_string(full_command_string, diff_tree.old_tree,
745
diff_tree.new_tree, diff_tree.to_file)
746
return from_diff_tree
748
def _get_command(self, old_path, new_path):
749
my_map = {'old_path': old_path, 'new_path': new_path}
750
return [AtTemplate(t).substitute(my_map) for t in
751
self.command_template]
753
def _execute(self, old_path, new_path):
754
command = self._get_command(old_path, new_path)
756
proc = subprocess.Popen(command, stdout=subprocess.PIPE,
759
if e.errno == errno.ENOENT:
760
raise errors.ExecutableMissing(command[0])
763
self.to_file.write(proc.stdout.read())
766
def _try_symlink_root(self, tree, prefix):
767
if (getattr(tree, 'abspath', None) is None
768
or not osutils.host_os_dereferences_symlinks()):
771
os.symlink(tree.abspath(''), osutils.pathjoin(self._root, prefix))
773
if e.errno != errno.EEXIST:
777
def _write_file(self, file_id, tree, prefix, relpath, force_temp=False,
779
if not force_temp and isinstance(tree, WorkingTree):
780
return tree.abspath(tree.id2path(file_id))
782
full_path = osutils.pathjoin(self._root, prefix, relpath)
783
if not force_temp and self._try_symlink_root(tree, prefix):
785
parent_dir = osutils.dirname(full_path)
787
os.makedirs(parent_dir)
789
if e.errno != errno.EEXIST:
791
source = tree.get_file(file_id, relpath)
793
target = open(full_path, 'wb')
795
osutils.pumpfile(source, target)
801
mtime = tree.get_file_mtime(file_id)
802
except errors.FileTimestampUnavailable:
805
os.utime(full_path, (mtime, mtime))
807
osutils.make_readonly(full_path)
810
def _prepare_files(self, file_id, old_path, new_path, force_temp=False,
811
allow_write_new=False):
812
old_disk_path = self._write_file(file_id, self.old_tree, 'old',
813
old_path, force_temp)
814
new_disk_path = self._write_file(file_id, self.new_tree, 'new',
815
new_path, force_temp,
816
allow_write=allow_write_new)
817
return old_disk_path, new_disk_path
821
osutils.rmtree(self._root)
823
if e.errno != errno.ENOENT:
824
mutter("The temporary directory \"%s\" was not "
825
"cleanly removed: %s." % (self._root, e))
827
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
828
if (old_kind, new_kind) != ('file', 'file'):
829
return DiffPath.CANNOT_DIFF
830
(old_disk_path, new_disk_path) = self._prepare_files(
831
file_id, old_path, new_path)
832
self._execute(old_disk_path, new_disk_path)
834
def edit_file(self, file_id):
835
"""Use this tool to edit a file.
837
A temporary copy will be edited, and the new contents will be
840
:param file_id: The id of the file to edit.
841
:return: The new contents of the file.
843
old_path = self.old_tree.id2path(file_id)
844
new_path = self.new_tree.id2path(file_id)
845
new_abs_path = self._prepare_files(file_id, old_path, new_path,
846
allow_write_new=True,
848
command = self._get_command(osutils.pathjoin('old', old_path),
849
osutils.pathjoin('new', new_path))
850
subprocess.call(command, cwd=self._root)
851
new_file = open(new_abs_path, 'r')
853
return new_file.read()
858
class DiffTree(object):
859
"""Provides textual representations of the difference between two trees.
861
A DiffTree examines two trees and where a file-id has altered
862
between them, generates a textual representation of the difference.
863
DiffTree uses a sequence of DiffPath objects which are each
864
given the opportunity to handle a given altered fileid. The list
865
of DiffPath objects can be extended globally by appending to
866
DiffTree.diff_factories, or for a specific diff operation by
867
supplying the extra_factories option to the appropriate method.
870
# list of factories that can provide instances of DiffPath objects
871
# may be extended by plugins.
872
diff_factories = [DiffSymlink.from_diff_tree,
873
DiffDirectory.from_diff_tree]
875
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
876
diff_text=None, extra_factories=None):
879
:param old_tree: Tree to show as old in the comparison
880
:param new_tree: Tree to show as new in the comparison
881
:param to_file: File to write comparision to
882
:param path_encoding: Character encoding to write paths in
883
:param diff_text: DiffPath-type object to use as a last resort for
885
:param extra_factories: Factories of DiffPaths to try before any other
887
if diff_text is None:
888
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
889
'', '', internal_diff)
890
self.old_tree = old_tree
891
self.new_tree = new_tree
892
self.to_file = to_file
893
self.path_encoding = path_encoding
895
if extra_factories is not None:
896
self.differs.extend(f(self) for f in extra_factories)
897
self.differs.extend(f(self) for f in self.diff_factories)
898
self.differs.extend([diff_text, DiffKindChange.from_diff_tree(self)])
901
def from_trees_options(klass, old_tree, new_tree, to_file,
902
path_encoding, external_diff_options, old_label,
904
"""Factory for producing a DiffTree.
906
Designed to accept options used by show_diff_trees.
907
:param old_tree: The tree to show as old in the comparison
908
:param new_tree: The tree to show as new in the comparison
909
:param to_file: File to write comparisons to
910
:param path_encoding: Character encoding to use for writing paths
911
:param external_diff_options: If supplied, use the installed diff
912
binary to perform file comparison, using supplied options.
913
:param old_label: Prefix to use for old file labels
914
:param new_label: Prefix to use for new file labels
915
:param using: Commandline to use to invoke an external diff tool
917
if using is not None:
918
extra_factories = [DiffFromTool.make_from_diff_tree(using, external_diff_options)]
921
if external_diff_options:
922
opts = external_diff_options.split()
923
def diff_file(olab, olines, nlab, nlines, to_file, path_encoding=None):
924
""":param path_encoding: not used but required
925
to match the signature of internal_diff.
927
external_diff(olab, olines, nlab, nlines, to_file, opts)
929
diff_file = internal_diff
930
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
931
old_label, new_label, diff_file)
932
return klass(old_tree, new_tree, to_file, path_encoding, diff_text,
935
def show_diff(self, specific_files, extra_trees=None):
936
"""Write tree diff to self.to_file
938
:param specific_files: the specific files to compare (recursive)
939
:param extra_trees: extra trees to use for mapping paths to file_ids
942
return self._show_diff(specific_files, extra_trees)
944
for differ in self.differs:
947
def _show_diff(self, specific_files, extra_trees):
948
# TODO: Generation of pseudo-diffs for added/deleted files could
949
# be usefully made into a much faster special case.
950
iterator = self.new_tree.iter_changes(self.old_tree,
951
specific_files=specific_files,
952
extra_trees=extra_trees,
953
require_versioned=True)
955
def changes_key(change):
956
old_path, new_path = change[1]
961
def get_encoded_path(path):
963
return path.encode(self.path_encoding, "replace")
964
for (file_id, paths, changed_content, versioned, parent, name, kind,
965
executable) in sorted(iterator, key=changes_key):
966
# The root does not get diffed, and items with no known kind (that
967
# is, missing) in both trees are skipped as well.
968
if parent == (None, None) or kind == (None, None):
970
oldpath, newpath = paths
971
oldpath_encoded = get_encoded_path(paths[0])
972
newpath_encoded = get_encoded_path(paths[1])
973
old_present = (kind[0] is not None and versioned[0])
974
new_present = (kind[1] is not None and versioned[1])
975
renamed = (parent[0], name[0]) != (parent[1], name[1])
977
properties_changed = []
978
properties_changed.extend(get_executable_change(executable[0], executable[1]))
980
if properties_changed:
981
prop_str = " (properties changed: %s)" % (", ".join(properties_changed),)
985
if (old_present, new_present) == (True, False):
986
self.to_file.write("=== removed %s '%s'\n" %
987
(kind[0], oldpath_encoded))
989
elif (old_present, new_present) == (False, True):
990
self.to_file.write("=== added %s '%s'\n" %
991
(kind[1], newpath_encoded))
994
self.to_file.write("=== renamed %s '%s' => '%s'%s\n" %
995
(kind[0], oldpath_encoded, newpath_encoded, prop_str))
997
# if it was produced by iter_changes, it must be
998
# modified *somehow*, either content or execute bit.
999
self.to_file.write("=== modified %s '%s'%s\n" % (kind[0],
1000
newpath_encoded, prop_str))
1002
self._diff(file_id, oldpath, newpath, kind[0], kind[1])
1008
def diff(self, file_id, old_path, new_path):
1009
"""Perform a diff of a single file
1011
:param file_id: file-id of the file
1012
:param old_path: The path of the file in the old tree
1013
:param new_path: The path of the file in the new tree
1016
old_kind = self.old_tree.kind(file_id)
1017
except (errors.NoSuchId, errors.NoSuchFile):
1020
new_kind = self.new_tree.kind(file_id)
1021
except (errors.NoSuchId, errors.NoSuchFile):
1023
self._diff(file_id, old_path, new_path, old_kind, new_kind)
1026
def _diff(self, file_id, old_path, new_path, old_kind, new_kind):
1027
result = DiffPath._diff_many(self.differs, file_id, old_path,
1028
new_path, old_kind, new_kind)
1029
if result is DiffPath.CANNOT_DIFF:
1030
error_path = new_path
1031
if error_path is None:
1032
error_path = old_path
1033
raise errors.NoDiffFound(error_path)
1036
format_registry = Registry()
1037
format_registry.register('default', DiffTree)
added
removed
bzrlib/diff.py
