1
# Copyright (C) 2005-2014 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
17
from __future__ import absolute_import
25
from .lazy_import import lazy_import
26
lazy_import(globals(), """
42
from breezy.workingtree import WorkingTree
43
from breezy.i18n import gettext
46
from .registry import (
49
from .sixish import text_type
50
from .trace import mutter, note, warning
51
from .tree import FileTimestampUnavailable
54
DEFAULT_CONTEXT_AMOUNT = 3
57
class AtTemplate(string.Template):
58
"""Templating class that uses @ instead of $."""
63
# TODO: Rather than building a changeset object, we should probably
64
# invoke callbacks on an object. That object can either accumulate a
65
# list, write them out directly, etc etc.
68
class _PrematchedMatcher(difflib.SequenceMatcher):
69
"""Allow SequenceMatcher operations to use predetermined blocks"""
71
def __init__(self, matching_blocks):
72
difflib.SequenceMatcher(self, None, None)
73
self.matching_blocks = matching_blocks
77
def internal_diff(old_label, oldlines, new_label, newlines, to_file,
78
allow_binary=False, sequence_matcher=None,
79
path_encoding='utf8', context_lines=DEFAULT_CONTEXT_AMOUNT):
80
# FIXME: difflib is wrong if there is no trailing newline.
81
# The syntax used by patch seems to be "\ No newline at
82
# end of file" following the last diff line from that
83
# file. This is not trivial to insert into the
84
# unified_diff output and it might be better to just fix
85
# or replace that function.
87
# In the meantime we at least make sure the patch isn't
90
if allow_binary is False:
91
textfile.check_text_lines(oldlines)
92
textfile.check_text_lines(newlines)
94
if sequence_matcher is None:
95
sequence_matcher = patiencediff.PatienceSequenceMatcher
96
ud = patiencediff.unified_diff_bytes(oldlines, newlines,
97
fromfile=old_label.encode(
98
path_encoding, 'replace'),
99
tofile=new_label.encode(
100
path_encoding, 'replace'),
101
n=context_lines, sequencematcher=sequence_matcher)
104
if len(ud) == 0: # Identical contents, nothing to do
106
# work-around for difflib being too smart for its own good
107
# if /dev/null is "1,0", patch won't recognize it as /dev/null
109
ud[2] = ud[2].replace(b'-1,0', b'-0,0')
111
ud[2] = ud[2].replace(b'+1,0', b'+0,0')
115
if not line.endswith(b'\n'):
116
to_file.write(b"\n\\ No newline at end of file\n")
120
def _spawn_external_diff(diffcmd, capture_errors=True):
121
"""Spawn the external diff process, and return the child handle.
123
:param diffcmd: The command list to spawn
124
:param capture_errors: Capture stderr as well as setting LANG=C
125
and LC_ALL=C. This lets us read and understand the output of diff,
126
and respond to any errors.
127
:return: A Popen object.
130
# construct minimal environment
132
path = os.environ.get('PATH')
135
env['LANGUAGE'] = 'C' # on win32 only LANGUAGE has effect
138
stderr = subprocess.PIPE
144
pipe = subprocess.Popen(diffcmd,
145
stdin=subprocess.PIPE,
146
stdout=subprocess.PIPE,
150
if e.errno == errno.ENOENT:
151
raise errors.NoDiff(str(e))
157
# diff style options as of GNU diff v3.2
158
style_option_list = ['-c', '-C', '--context',
160
'-f', '--forward-ed',
164
'-u', '-U', '--unified',
165
'-y', '--side-by-side',
169
def default_style_unified(diff_opts):
170
"""Default to unified diff style if alternative not specified in diff_opts.
172
diff only allows one style to be specified; they don't override.
173
Note that some of these take optargs, and the optargs can be
174
directly appended to the options.
175
This is only an approximate parser; it doesn't properly understand
178
:param diff_opts: List of options for external (GNU) diff.
179
:return: List of options with default style=='unified'.
181
for s in style_option_list:
189
diff_opts.append('-u')
193
def external_diff(old_label, oldlines, new_label, newlines, to_file,
195
"""Display a diff by calling out to the external diff program."""
196
# make sure our own output is properly ordered before the diff
199
oldtmp_fd, old_abspath = tempfile.mkstemp(prefix='brz-diff-old-')
200
newtmp_fd, new_abspath = tempfile.mkstemp(prefix='brz-diff-new-')
201
oldtmpf = os.fdopen(oldtmp_fd, 'wb')
202
newtmpf = os.fdopen(newtmp_fd, 'wb')
205
# TODO: perhaps a special case for comparing to or from the empty
206
# sequence; can just use /dev/null on Unix
208
# TODO: if either of the files being compared already exists as a
209
# regular named file (e.g. in the working directory) then we can
210
# compare directly to that, rather than copying it.
212
oldtmpf.writelines(oldlines)
213
newtmpf.writelines(newlines)
220
if sys.platform == 'win32':
221
# Popen doesn't do the proper encoding for external commands
222
# Since we are dealing with an ANSI api, use mbcs encoding
223
old_label = old_label.encode('mbcs')
224
new_label = new_label.encode('mbcs')
226
'--label', old_label,
228
'--label', new_label,
233
diff_opts = default_style_unified(diff_opts)
236
diffcmd.extend(diff_opts)
238
pipe = _spawn_external_diff(diffcmd, capture_errors=True)
239
out, err = pipe.communicate()
242
# internal_diff() adds a trailing newline, add one here for consistency
245
# 'diff' gives retcode == 2 for all sorts of errors
246
# one of those is 'Binary files differ'.
247
# Bad options could also be the problem.
248
# 'Binary files' is not a real error, so we suppress that error.
251
# Since we got here, we want to make sure to give an i18n error
252
pipe = _spawn_external_diff(diffcmd, capture_errors=False)
253
out, err = pipe.communicate()
255
# Write out the new i18n diff response
256
to_file.write(out + b'\n')
257
if pipe.returncode != 2:
258
raise errors.BzrError(
259
'external diff failed with exit code 2'
260
' when run with LANG=C and LC_ALL=C,'
261
' but not when run natively: %r' % (diffcmd,))
263
first_line = lang_c_out.split(b'\n', 1)[0]
264
# Starting with diffutils 2.8.4 the word "binary" was dropped.
265
m = re.match(b'^(binary )?files.*differ$', first_line, re.I)
267
raise errors.BzrError('external diff failed with exit code 2;'
268
' command: %r' % (diffcmd,))
270
# Binary files differ, just return
273
# If we got to here, we haven't written out the output of diff
277
# returns 1 if files differ; that's OK
279
msg = 'signal %d' % (-rc)
281
msg = 'exit code %d' % rc
283
raise errors.BzrError('external diff failed with %s; command: %r'
287
oldtmpf.close() # and delete
291
# Warn in case the file couldn't be deleted (in case windows still
292
# holds the file open, but not if the files have already been
297
if e.errno not in (errno.ENOENT,):
298
warning('Failed to delete temporary file: %s %s', path, e)
304
def get_trees_and_branches_to_diff_locked(
305
path_list, revision_specs, old_url, new_url, add_cleanup, apply_view=True):
306
"""Get the trees and specific files to diff given a list of paths.
308
This method works out the trees to be diff'ed and the files of
309
interest within those trees.
312
the list of arguments passed to the diff command
313
:param revision_specs:
314
Zero, one or two RevisionSpecs from the diff command line,
315
saying what revisions to compare.
317
The url of the old branch or tree. If None, the tree to use is
318
taken from the first path, if any, or the current working tree.
320
The url of the new branch or tree. If None, the tree to use is
321
taken from the first path, if any, or the current working tree.
323
a callable like Command.add_cleanup. get_trees_and_branches_to_diff
324
will register cleanups that must be run to unlock the trees, etc.
326
if True and a view is set, apply the view or check that the paths
329
a tuple of (old_tree, new_tree, old_branch, new_branch,
330
specific_files, extra_trees) where extra_trees is a sequence of
331
additional trees to search in for file-ids. The trees and branches
332
will be read-locked until the cleanups registered via the add_cleanup
335
# Get the old and new revision specs
336
old_revision_spec = None
337
new_revision_spec = None
338
if revision_specs is not None:
339
if len(revision_specs) > 0:
340
old_revision_spec = revision_specs[0]
342
old_url = old_revision_spec.get_branch()
343
if len(revision_specs) > 1:
344
new_revision_spec = revision_specs[1]
346
new_url = new_revision_spec.get_branch()
349
make_paths_wt_relative = True
350
consider_relpath = True
351
if path_list is None or len(path_list) == 0:
352
# If no path is given, the current working tree is used
353
default_location = u'.'
354
consider_relpath = False
355
elif old_url is not None and new_url is not None:
356
other_paths = path_list
357
make_paths_wt_relative = False
359
default_location = path_list[0]
360
other_paths = path_list[1:]
362
def lock_tree_or_branch(wt, br):
365
add_cleanup(wt.unlock)
368
add_cleanup(br.unlock)
370
# Get the old location
373
old_url = default_location
374
working_tree, branch, relpath = \
375
controldir.ControlDir.open_containing_tree_or_branch(old_url)
376
lock_tree_or_branch(working_tree, branch)
377
if consider_relpath and relpath != '':
378
if working_tree is not None and apply_view:
379
views.check_path_in_view(working_tree, relpath)
380
specific_files.append(relpath)
381
old_tree = _get_tree_to_diff(old_revision_spec, working_tree, branch)
384
# Get the new location
386
new_url = default_location
387
if new_url != old_url:
388
working_tree, branch, relpath = \
389
controldir.ControlDir.open_containing_tree_or_branch(new_url)
390
lock_tree_or_branch(working_tree, branch)
391
if consider_relpath and relpath != '':
392
if working_tree is not None and apply_view:
393
views.check_path_in_view(working_tree, relpath)
394
specific_files.append(relpath)
395
new_tree = _get_tree_to_diff(new_revision_spec, working_tree, branch,
396
basis_is_default=working_tree is None)
399
# Get the specific files (all files is None, no files is [])
400
if make_paths_wt_relative and working_tree is not None:
401
other_paths = working_tree.safe_relpath_files(
403
apply_view=apply_view)
404
specific_files.extend(other_paths)
405
if len(specific_files) == 0:
406
specific_files = None
407
if (working_tree is not None and working_tree.supports_views() and
409
view_files = working_tree.views.lookup_view()
411
specific_files = view_files
412
view_str = views.view_display_str(view_files)
413
note(gettext("*** Ignoring files outside view. View is %s") % view_str)
415
# Get extra trees that ought to be searched for file-ids
417
if working_tree is not None and working_tree not in (old_tree, new_tree):
418
extra_trees = (working_tree,)
419
return (old_tree, new_tree, old_branch, new_branch,
420
specific_files, extra_trees)
423
def _get_tree_to_diff(spec, tree=None, branch=None, basis_is_default=True):
424
if branch is None and tree is not None:
426
if spec is None or spec.spec is None:
429
return tree.basis_tree()
431
return branch.basis_tree()
434
return spec.as_tree(branch)
437
def show_diff_trees(old_tree, new_tree, to_file, specific_files=None,
438
external_diff_options=None,
439
old_label='a/', new_label='b/',
441
path_encoding='utf8',
444
context=DEFAULT_CONTEXT_AMOUNT):
445
"""Show in text form the changes from one tree to another.
447
:param to_file: The output stream.
448
:param specific_files: Include only changes to these files - None for all
450
:param external_diff_options: If set, use an external GNU diff and pass
452
:param extra_trees: If set, more Trees to use for looking up file ids
453
:param path_encoding: If set, the path will be encoded as specified,
454
otherwise is supposed to be utf8
455
:param format_cls: Formatter class (DiffTree subclass)
458
context = DEFAULT_CONTEXT_AMOUNT
459
if format_cls is None:
460
format_cls = DiffTree
461
with old_tree.lock_read():
462
if extra_trees is not None:
463
for tree in extra_trees:
467
differ = format_cls.from_trees_options(old_tree, new_tree, to_file,
469
external_diff_options,
470
old_label, new_label, using,
471
context_lines=context)
472
return differ.show_diff(specific_files, extra_trees)
475
if extra_trees is not None:
476
for tree in extra_trees:
480
def _patch_header_date(tree, path):
481
"""Returns a timestamp suitable for use in a patch header."""
483
mtime = tree.get_file_mtime(path)
484
except FileTimestampUnavailable:
486
return timestamp.format_patch_date(mtime)
489
def get_executable_change(old_is_x, new_is_x):
490
descr = {True: b"+x", False: b"-x", None: b"??"}
491
if old_is_x != new_is_x:
492
return [b"%s to %s" % (descr[old_is_x], descr[new_is_x],)]
497
class DiffPath(object):
498
"""Base type for command object that compare files"""
500
# The type or contents of the file were unsuitable for diffing
501
CANNOT_DIFF = 'CANNOT_DIFF'
502
# The file has changed in a semantic way
504
# The file content may have changed, but there is no semantic change
505
UNCHANGED = 'UNCHANGED'
507
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8'):
510
:param old_tree: The tree to show as the old tree in the comparison
511
:param new_tree: The tree to show as new in the comparison
512
:param to_file: The file to write comparison data to
513
:param path_encoding: The character encoding to write paths in
515
self.old_tree = old_tree
516
self.new_tree = new_tree
517
self.to_file = to_file
518
self.path_encoding = path_encoding
524
def from_diff_tree(klass, diff_tree):
525
return klass(diff_tree.old_tree, diff_tree.new_tree,
526
diff_tree.to_file, diff_tree.path_encoding)
529
def _diff_many(differs, old_path, new_path, old_kind, new_kind):
530
for file_differ in differs:
531
result = file_differ.diff(old_path, new_path, old_kind, new_kind)
532
if result is not DiffPath.CANNOT_DIFF:
535
return DiffPath.CANNOT_DIFF
538
class DiffKindChange(object):
539
"""Special differ for file kind changes.
541
Represents kind change as deletion + creation. Uses the other differs
545
def __init__(self, differs):
546
self.differs = differs
552
def from_diff_tree(klass, diff_tree):
553
return klass(diff_tree.differs)
555
def diff(self, old_path, new_path, old_kind, new_kind):
556
"""Perform comparison
558
:param old_path: Path of the file in the old tree
559
:param new_path: Path of the file in the new tree
560
:param old_kind: Old file-kind of the file
561
:param new_kind: New file-kind of the file
563
if None in (old_kind, new_kind):
564
return DiffPath.CANNOT_DIFF
565
result = DiffPath._diff_many(
566
self.differs, old_path, new_path, old_kind, None)
567
if result is DiffPath.CANNOT_DIFF:
569
return DiffPath._diff_many(
570
self.differs, old_path, new_path, None, new_kind)
573
class DiffDirectory(DiffPath):
575
def diff(self, old_path, new_path, old_kind, new_kind):
576
"""Perform comparison between two directories. (dummy)
579
if 'directory' not in (old_kind, new_kind):
580
return self.CANNOT_DIFF
581
if old_kind not in ('directory', None):
582
return self.CANNOT_DIFF
583
if new_kind not in ('directory', None):
584
return self.CANNOT_DIFF
588
class DiffSymlink(DiffPath):
590
def diff(self, old_path, new_path, old_kind, new_kind):
591
"""Perform comparison between two symlinks
593
:param old_path: Path of the file in the old tree
594
:param new_path: Path of the file in the new tree
595
:param old_kind: Old file-kind of the file
596
:param new_kind: New file-kind of the file
598
if 'symlink' not in (old_kind, new_kind):
599
return self.CANNOT_DIFF
600
if old_kind == 'symlink':
601
old_target = self.old_tree.get_symlink_target(old_path)
602
elif old_kind is None:
605
return self.CANNOT_DIFF
606
if new_kind == 'symlink':
607
new_target = self.new_tree.get_symlink_target(new_path)
608
elif new_kind is None:
611
return self.CANNOT_DIFF
612
return self.diff_symlink(old_target, new_target)
614
def diff_symlink(self, old_target, new_target):
615
if old_target is None:
616
self.to_file.write(b'=== target is \'%s\'\n' %
617
new_target.encode(self.path_encoding, 'replace'))
618
elif new_target is None:
619
self.to_file.write(b'=== target was \'%s\'\n' %
620
old_target.encode(self.path_encoding, 'replace'))
622
self.to_file.write(b'=== target changed \'%s\' => \'%s\'\n' %
623
(old_target.encode(self.path_encoding, 'replace'),
624
new_target.encode(self.path_encoding, 'replace')))
628
class DiffText(DiffPath):
630
# GNU Patch uses the epoch date to detect files that are being added
631
# or removed in a diff.
632
EPOCH_DATE = '1970-01-01 00:00:00 +0000'
634
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
635
old_label='', new_label='', text_differ=internal_diff,
636
context_lines=DEFAULT_CONTEXT_AMOUNT):
637
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
638
self.text_differ = text_differ
639
self.old_label = old_label
640
self.new_label = new_label
641
self.path_encoding = path_encoding
642
self.context_lines = context_lines
644
def diff(self, old_path, new_path, old_kind, new_kind):
645
"""Compare two files in unified diff format
647
:param old_path: Path of the file in the old tree
648
:param new_path: Path of the file in the new tree
649
:param old_kind: Old file-kind of the file
650
:param new_kind: New file-kind of the file
652
if 'file' not in (old_kind, new_kind):
653
return self.CANNOT_DIFF
654
if old_kind == 'file':
655
old_date = _patch_header_date(self.old_tree, old_path)
656
elif old_kind is None:
657
old_date = self.EPOCH_DATE
659
return self.CANNOT_DIFF
660
if new_kind == 'file':
661
new_date = _patch_header_date(self.new_tree, new_path)
662
elif new_kind is None:
663
new_date = self.EPOCH_DATE
665
return self.CANNOT_DIFF
666
from_label = '%s%s\t%s' % (self.old_label, old_path,
668
to_label = '%s%s\t%s' % (self.new_label, new_path,
670
return self.diff_text(old_path, new_path, from_label, to_label)
672
def diff_text(self, from_path, to_path, from_label, to_label):
673
"""Diff the content of given files in two trees
675
:param from_path: The path in the from tree. If None,
676
the file is not present in the from tree.
677
:param to_path: The path in the to tree. This may refer
678
to a different file from from_path. If None,
679
the file is not present in the to tree.
681
def _get_text(tree, path):
685
return tree.get_file_lines(path)
686
except errors.NoSuchFile:
689
from_text = _get_text(self.old_tree, from_path)
690
to_text = _get_text(self.new_tree, to_path)
691
self.text_differ(from_label, from_text, to_label, to_text,
692
self.to_file, path_encoding=self.path_encoding,
693
context_lines=self.context_lines)
694
except errors.BinaryFile:
696
("Binary files %s and %s differ\n" %
697
(from_label, to_label)).encode(self.path_encoding, 'replace'))
701
class DiffFromTool(DiffPath):
703
def __init__(self, command_template, old_tree, new_tree, to_file,
704
path_encoding='utf-8'):
705
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
706
self.command_template = command_template
707
self._root = osutils.mkdtemp(prefix='brz-diff-')
710
def from_string(klass, command_string, old_tree, new_tree, to_file,
711
path_encoding='utf-8'):
712
command_template = cmdline.split(command_string)
713
if '@' not in command_string:
714
command_template.extend(['@old_path', '@new_path'])
715
return klass(command_template, old_tree, new_tree, to_file,
719
def make_from_diff_tree(klass, command_string, external_diff_options=None):
720
def from_diff_tree(diff_tree):
721
full_command_string = [command_string]
722
if external_diff_options is not None:
723
full_command_string += ' ' + external_diff_options
724
return klass.from_string(full_command_string, diff_tree.old_tree,
725
diff_tree.new_tree, diff_tree.to_file)
726
return from_diff_tree
728
def _get_command(self, old_path, new_path):
729
my_map = {'old_path': old_path, 'new_path': new_path}
730
command = [AtTemplate(t).substitute(my_map) for t in
731
self.command_template]
732
if sys.platform == 'win32': # Popen doesn't accept unicode on win32
735
if isinstance(c, text_type):
736
command_encoded.append(c.encode('mbcs'))
738
command_encoded.append(c)
739
return command_encoded
743
def _execute(self, old_path, new_path):
744
command = self._get_command(old_path, new_path)
746
proc = subprocess.Popen(command, stdout=subprocess.PIPE,
749
if e.errno == errno.ENOENT:
750
raise errors.ExecutableMissing(command[0])
753
self.to_file.write(proc.stdout.read())
757
def _try_symlink_root(self, tree, prefix):
758
if (getattr(tree, 'abspath', None) is None or
759
not osutils.host_os_dereferences_symlinks()):
762
os.symlink(tree.abspath(''), osutils.pathjoin(self._root, prefix))
764
if e.errno != errno.EEXIST:
770
"""Returns safe encoding for passing file path to diff tool"""
771
if sys.platform == 'win32':
774
# Don't fallback to 'utf-8' because subprocess may not be able to
775
# handle utf-8 correctly when locale is not utf-8.
776
return sys.getfilesystemencoding() or 'ascii'
778
def _is_safepath(self, path):
779
"""Return true if `path` may be able to pass to subprocess."""
782
return path == path.encode(fenc).decode(fenc)
786
def _safe_filename(self, prefix, relpath):
787
"""Replace unsafe character in `relpath` then join `self._root`,
788
`prefix` and `relpath`."""
790
# encoded_str.replace('?', '_') may break multibyte char.
791
# So we should encode, decode, then replace(u'?', u'_')
792
relpath_tmp = relpath.encode(fenc, 'replace').decode(fenc, 'replace')
793
relpath_tmp = relpath_tmp.replace(u'?', u'_')
794
return osutils.pathjoin(self._root, prefix, relpath_tmp)
796
def _write_file(self, relpath, tree, prefix, force_temp=False,
798
if not force_temp and isinstance(tree, WorkingTree):
799
full_path = tree.abspath(relpath)
800
if self._is_safepath(full_path):
803
full_path = self._safe_filename(prefix, relpath)
804
if not force_temp and self._try_symlink_root(tree, prefix):
806
parent_dir = osutils.dirname(full_path)
808
os.makedirs(parent_dir)
810
if e.errno != errno.EEXIST:
812
source = tree.get_file(relpath)
814
with open(full_path, 'wb') as target:
815
osutils.pumpfile(source, target)
819
mtime = tree.get_file_mtime(relpath)
820
except FileTimestampUnavailable:
823
os.utime(full_path, (mtime, mtime))
825
osutils.make_readonly(full_path)
828
def _prepare_files(self, old_path, new_path, force_temp=False,
829
allow_write_new=False):
830
old_disk_path = self._write_file(
831
old_path, self.old_tree, 'old', force_temp)
832
new_disk_path = self._write_file(
833
new_path, self.new_tree, 'new', force_temp,
834
allow_write=allow_write_new)
835
return old_disk_path, new_disk_path
839
osutils.rmtree(self._root)
841
if e.errno != errno.ENOENT:
842
mutter("The temporary directory \"%s\" was not "
843
"cleanly removed: %s." % (self._root, e))
845
def diff(self, old_path, new_path, old_kind, new_kind):
846
if (old_kind, new_kind) != ('file', 'file'):
847
return DiffPath.CANNOT_DIFF
848
(old_disk_path, new_disk_path) = self._prepare_files(
850
self._execute(old_disk_path, new_disk_path)
852
def edit_file(self, old_path, new_path):
853
"""Use this tool to edit a file.
855
A temporary copy will be edited, and the new contents will be
858
:return: The new contents of the file.
860
old_abs_path, new_abs_path = self._prepare_files(
861
old_path, new_path, allow_write_new=True, force_temp=True)
862
command = self._get_command(old_abs_path, new_abs_path)
863
subprocess.call(command, cwd=self._root)
864
with open(new_abs_path, 'rb') as new_file:
865
return new_file.read()
868
class DiffTree(object):
869
"""Provides textual representations of the difference between two trees.
871
A DiffTree examines two trees and where a file-id has altered
872
between them, generates a textual representation of the difference.
873
DiffTree uses a sequence of DiffPath objects which are each
874
given the opportunity to handle a given altered fileid. The list
875
of DiffPath objects can be extended globally by appending to
876
DiffTree.diff_factories, or for a specific diff operation by
877
supplying the extra_factories option to the appropriate method.
880
# list of factories that can provide instances of DiffPath objects
881
# may be extended by plugins.
882
diff_factories = [DiffSymlink.from_diff_tree,
883
DiffDirectory.from_diff_tree]
885
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
886
diff_text=None, extra_factories=None):
889
:param old_tree: Tree to show as old in the comparison
890
:param new_tree: Tree to show as new in the comparison
891
:param to_file: File to write comparision to
892
:param path_encoding: Character encoding to write paths in
893
:param diff_text: DiffPath-type object to use as a last resort for
895
:param extra_factories: Factories of DiffPaths to try before any other
897
if diff_text is None:
898
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
899
'', '', internal_diff)
900
self.old_tree = old_tree
901
self.new_tree = new_tree
902
self.to_file = to_file
903
self.path_encoding = path_encoding
905
if extra_factories is not None:
906
self.differs.extend(f(self) for f in extra_factories)
907
self.differs.extend(f(self) for f in self.diff_factories)
908
self.differs.extend([diff_text, DiffKindChange.from_diff_tree(self)])
911
def from_trees_options(klass, old_tree, new_tree, to_file,
912
path_encoding, external_diff_options, old_label,
913
new_label, using, context_lines):
914
"""Factory for producing a DiffTree.
916
Designed to accept options used by show_diff_trees.
918
:param old_tree: The tree to show as old in the comparison
919
:param new_tree: The tree to show as new in the comparison
920
:param to_file: File to write comparisons to
921
:param path_encoding: Character encoding to use for writing paths
922
:param external_diff_options: If supplied, use the installed diff
923
binary to perform file comparison, using supplied options.
924
:param old_label: Prefix to use for old file labels
925
:param new_label: Prefix to use for new file labels
926
:param using: Commandline to use to invoke an external diff tool
928
if using is not None:
929
extra_factories = [DiffFromTool.make_from_diff_tree(
930
using, external_diff_options)]
933
if external_diff_options:
934
opts = external_diff_options.split()
936
def diff_file(olab, olines, nlab, nlines, to_file, path_encoding=None, context_lines=None):
937
""":param path_encoding: not used but required
938
to match the signature of internal_diff.
940
external_diff(olab, olines, nlab, nlines, to_file, opts)
942
diff_file = internal_diff
943
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
944
old_label, new_label, diff_file, context_lines=context_lines)
945
return klass(old_tree, new_tree, to_file, path_encoding, diff_text,
948
def show_diff(self, specific_files, extra_trees=None):
949
"""Write tree diff to self.to_file
951
:param specific_files: the specific files to compare (recursive)
952
:param extra_trees: extra trees to use for mapping paths to file_ids
955
return self._show_diff(specific_files, extra_trees)
957
for differ in self.differs:
960
def _show_diff(self, specific_files, extra_trees):
961
# TODO: Generation of pseudo-diffs for added/deleted files could
962
# be usefully made into a much faster special case.
963
iterator = self.new_tree.iter_changes(self.old_tree,
964
specific_files=specific_files,
965
extra_trees=extra_trees,
966
require_versioned=True)
969
def changes_key(change):
970
old_path, new_path = change[1]
976
def get_encoded_path(path):
978
return path.encode(self.path_encoding, "replace")
979
for (file_id, paths, changed_content, versioned, parent, name, kind,
980
executable) in sorted(iterator, key=changes_key):
981
# The root does not get diffed, and items with no known kind (that
982
# is, missing) in both trees are skipped as well.
983
if parent == (None, None) or kind == (None, None):
985
oldpath, newpath = paths
986
oldpath_encoded = get_encoded_path(paths[0])
987
newpath_encoded = get_encoded_path(paths[1])
988
old_present = (kind[0] is not None and versioned[0])
989
new_present = (kind[1] is not None and versioned[1])
990
renamed = (parent[0], name[0]) != (parent[1], name[1])
992
properties_changed = []
993
properties_changed.extend(
994
get_executable_change(executable[0], executable[1]))
996
if properties_changed:
997
prop_str = b" (properties changed: %s)" % (
998
b", ".join(properties_changed),)
1002
if (old_present, new_present) == (True, False):
1003
self.to_file.write(b"=== removed %s '%s'\n" %
1004
(kind[0].encode('ascii'), oldpath_encoded))
1006
elif (old_present, new_present) == (False, True):
1007
self.to_file.write(b"=== added %s '%s'\n" %
1008
(kind[1].encode('ascii'), newpath_encoded))
1011
self.to_file.write(b"=== renamed %s '%s' => '%s'%s\n" %
1012
(kind[0].encode('ascii'), oldpath_encoded, newpath_encoded, prop_str))
1014
# if it was produced by iter_changes, it must be
1015
# modified *somehow*, either content or execute bit.
1016
self.to_file.write(b"=== modified %s '%s'%s\n" % (kind[0].encode('ascii'),
1017
newpath_encoded, prop_str))
1019
self._diff(oldpath, newpath, kind[0], kind[1])
1025
def diff(self, old_path, new_path):
1026
"""Perform a diff of a single file
1028
:param old_path: The path of the file in the old tree
1029
:param new_path: The path of the file in the new tree
1031
if old_path is None:
1034
old_kind = self.old_tree.kind(old_path)
1035
if new_path is None:
1038
new_kind = self.new_tree.kind(new_path)
1039
self._diff(old_path, new_path, old_kind, new_kind)
1041
def _diff(self, old_path, new_path, old_kind, new_kind):
1042
result = DiffPath._diff_many(
1043
self.differs, old_path, new_path, old_kind, new_kind)
1044
if result is DiffPath.CANNOT_DIFF:
1045
error_path = new_path
1046
if error_path is None:
1047
error_path = old_path
1048
raise errors.NoDiffFound(error_path)
1051
format_registry = Registry()
1052
format_registry.register('default', DiffTree)
added
removed
breezy/diff.py
