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(), """
43
from breezy.workingtree import WorkingTree
44
from breezy.i18n import gettext
47
from .registry import (
50
from .trace import mutter, note, warning
51
from .tree import FileTimestampUnavailable
54
DEFAULT_CONTEXT_AMOUNT = 3
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', context_lines=DEFAULT_CONTEXT_AMOUNT):
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
n=context_lines, 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 external 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))
159
# diff style options as of GNU diff v3.2
160
style_option_list = ['-c', '-C', '--context',
162
'-f', '--forward-ed',
166
'-u', '-U', '--unified',
167
'-y', '--side-by-side',
170
def default_style_unified(diff_opts):
171
"""Default to unified diff style if alternative not specified in diff_opts.
173
diff only allows one style to be specified; they don't override.
174
Note that some of these take optargs, and the optargs can be
175
directly appended to the options.
176
This is only an approximate parser; it doesn't properly understand
179
:param diff_opts: List of options for external (GNU) diff.
180
:return: List of options with default style=='unified'.
182
for s in style_option_list:
190
diff_opts.append('-u')
194
def external_diff(old_filename, oldlines, new_filename, newlines, to_file,
196
"""Display a diff by calling out to the external diff program."""
197
# make sure our own output is properly ordered before the diff
200
oldtmp_fd, old_abspath = tempfile.mkstemp(prefix='brz-diff-old-')
201
newtmp_fd, new_abspath = tempfile.mkstemp(prefix='brz-diff-new-')
202
oldtmpf = os.fdopen(oldtmp_fd, 'wb')
203
newtmpf = os.fdopen(newtmp_fd, 'wb')
206
# TODO: perhaps a special case for comparing to or from the empty
207
# sequence; can just use /dev/null on Unix
209
# TODO: if either of the files being compared already exists as a
210
# regular named file (e.g. in the working directory) then we can
211
# compare directly to that, rather than copying it.
213
oldtmpf.writelines(oldlines)
214
newtmpf.writelines(newlines)
221
if sys.platform == 'win32':
222
# Popen doesn't do the proper encoding for external commands
223
# Since we are dealing with an ANSI api, use mbcs encoding
224
old_filename = old_filename.encode('mbcs')
225
new_filename = new_filename.encode('mbcs')
227
'--label', old_filename,
229
'--label', new_filename,
234
diff_opts = default_style_unified(diff_opts)
237
diffcmd.extend(diff_opts)
239
pipe = _spawn_external_diff(diffcmd, capture_errors=True)
240
out,err = pipe.communicate()
243
# internal_diff() adds a trailing newline, add one here for consistency
246
# 'diff' gives retcode == 2 for all sorts of errors
247
# one of those is 'Binary files differ'.
248
# Bad options could also be the problem.
249
# 'Binary files' is not a real error, so we suppress that error.
252
# Since we got here, we want to make sure to give an i18n error
253
pipe = _spawn_external_diff(diffcmd, capture_errors=False)
254
out, err = pipe.communicate()
256
# Write out the new i18n diff response
257
to_file.write(out+'\n')
258
if pipe.returncode != 2:
259
raise errors.BzrError(
260
'external diff failed with exit code 2'
261
' when run with LANG=C and LC_ALL=C,'
262
' but not when run natively: %r' % (diffcmd,))
264
first_line = lang_c_out.split('\n', 1)[0]
265
# Starting with diffutils 2.8.4 the word "binary" was dropped.
266
m = re.match('^(binary )?files.*differ$', first_line, re.I)
268
raise errors.BzrError('external diff failed with exit code 2;'
269
' command: %r' % (diffcmd,))
271
# Binary files differ, just return
274
# If we got to here, we haven't written out the output of diff
278
# returns 1 if files differ; that's OK
280
msg = 'signal %d' % (-rc)
282
msg = 'exit code %d' % rc
284
raise errors.BzrError('external diff failed with %s; command: %r'
289
oldtmpf.close() # and delete
293
# Warn in case the file couldn't be deleted (in case windows still
294
# holds the file open, but not if the files have already been
299
if e.errno not in (errno.ENOENT,):
300
warning('Failed to delete temporary file: %s %s', path, e)
306
def get_trees_and_branches_to_diff_locked(
307
path_list, revision_specs, old_url, new_url, add_cleanup, apply_view=True):
308
"""Get the trees and specific files to diff given a list of paths.
310
This method works out the trees to be diff'ed and the files of
311
interest within those trees.
314
the list of arguments passed to the diff command
315
:param revision_specs:
316
Zero, one or two RevisionSpecs from the diff command line,
317
saying what revisions to compare.
319
The url of the old branch or tree. If None, the tree to use is
320
taken from the first path, if any, or the current working tree.
322
The url of the new branch or tree. If None, the tree to use is
323
taken from the first path, if any, or the current working tree.
325
a callable like Command.add_cleanup. get_trees_and_branches_to_diff
326
will register cleanups that must be run to unlock the trees, etc.
328
if True and a view is set, apply the view or check that the paths
331
a tuple of (old_tree, new_tree, old_branch, new_branch,
332
specific_files, extra_trees) where extra_trees is a sequence of
333
additional trees to search in for file-ids. The trees and branches
334
will be read-locked until the cleanups registered via the add_cleanup
337
# Get the old and new revision specs
338
old_revision_spec = None
339
new_revision_spec = None
340
if revision_specs is not None:
341
if len(revision_specs) > 0:
342
old_revision_spec = revision_specs[0]
344
old_url = old_revision_spec.get_branch()
345
if len(revision_specs) > 1:
346
new_revision_spec = revision_specs[1]
348
new_url = new_revision_spec.get_branch()
351
make_paths_wt_relative = True
352
consider_relpath = True
353
if path_list is None or len(path_list) == 0:
354
# If no path is given, the current working tree is used
355
default_location = u'.'
356
consider_relpath = False
357
elif old_url is not None and new_url is not None:
358
other_paths = path_list
359
make_paths_wt_relative = False
361
default_location = path_list[0]
362
other_paths = path_list[1:]
364
def lock_tree_or_branch(wt, br):
367
add_cleanup(wt.unlock)
370
add_cleanup(br.unlock)
372
# Get the old location
375
old_url = default_location
376
working_tree, branch, relpath = \
377
controldir.ControlDir.open_containing_tree_or_branch(old_url)
378
lock_tree_or_branch(working_tree, branch)
379
if consider_relpath and relpath != '':
380
if working_tree is not None and apply_view:
381
views.check_path_in_view(working_tree, relpath)
382
specific_files.append(relpath)
383
old_tree = _get_tree_to_diff(old_revision_spec, working_tree, branch)
386
# Get the new location
388
new_url = default_location
389
if new_url != old_url:
390
working_tree, branch, relpath = \
391
controldir.ControlDir.open_containing_tree_or_branch(new_url)
392
lock_tree_or_branch(working_tree, branch)
393
if consider_relpath and relpath != '':
394
if working_tree is not None and apply_view:
395
views.check_path_in_view(working_tree, relpath)
396
specific_files.append(relpath)
397
new_tree = _get_tree_to_diff(new_revision_spec, working_tree, branch,
398
basis_is_default=working_tree is None)
401
# Get the specific files (all files is None, no files is [])
402
if make_paths_wt_relative and working_tree is not None:
403
other_paths = working_tree.safe_relpath_files(
405
apply_view=apply_view)
406
specific_files.extend(other_paths)
407
if len(specific_files) == 0:
408
specific_files = None
409
if (working_tree is not None and working_tree.supports_views()
411
view_files = working_tree.views.lookup_view()
413
specific_files = view_files
414
view_str = views.view_display_str(view_files)
415
note(gettext("*** Ignoring files outside view. View is %s") % view_str)
417
# Get extra trees that ought to be searched for file-ids
419
if working_tree is not None and working_tree not in (old_tree, new_tree):
420
extra_trees = (working_tree,)
421
return (old_tree, new_tree, old_branch, new_branch,
422
specific_files, extra_trees)
425
def _get_tree_to_diff(spec, tree=None, branch=None, basis_is_default=True):
426
if branch is None and tree is not None:
428
if spec is None or spec.spec is None:
431
return tree.basis_tree()
433
return branch.basis_tree()
436
return spec.as_tree(branch)
439
def show_diff_trees(old_tree, new_tree, to_file, specific_files=None,
440
external_diff_options=None,
441
old_label='a/', new_label='b/',
443
path_encoding='utf8',
446
context=DEFAULT_CONTEXT_AMOUNT):
447
"""Show in text form the changes from one tree to another.
449
:param to_file: The output stream.
450
:param specific_files: Include only changes to these files - None for all
452
:param external_diff_options: If set, use an external GNU diff and pass
454
:param extra_trees: If set, more Trees to use for looking up file ids
455
:param path_encoding: If set, the path will be encoded as specified,
456
otherwise is supposed to be utf8
457
:param format_cls: Formatter class (DiffTree subclass)
460
context = DEFAULT_CONTEXT_AMOUNT
461
if format_cls is None:
462
format_cls = DiffTree
463
with old_tree.lock_read():
464
if extra_trees is not None:
465
for tree in extra_trees:
469
differ = format_cls.from_trees_options(old_tree, new_tree, to_file,
471
external_diff_options,
472
old_label, new_label, using,
473
context_lines=context)
474
return differ.show_diff(specific_files, extra_trees)
477
if extra_trees is not None:
478
for tree in extra_trees:
482
def _patch_header_date(tree, file_id, path):
483
"""Returns a timestamp suitable for use in a patch header."""
485
mtime = tree.get_file_mtime(file_id, path)
486
except FileTimestampUnavailable:
488
return timestamp.format_patch_date(mtime)
491
def get_executable_change(old_is_x, new_is_x):
492
descr = { True:"+x", False:"-x", None:"??" }
493
if old_is_x != new_is_x:
494
return ["%s to %s" % (descr[old_is_x], descr[new_is_x],)]
499
class DiffPath(object):
500
"""Base type for command object that compare files"""
502
# The type or contents of the file were unsuitable for diffing
503
CANNOT_DIFF = 'CANNOT_DIFF'
504
# The file has changed in a semantic way
506
# The file content may have changed, but there is no semantic change
507
UNCHANGED = 'UNCHANGED'
509
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8'):
512
:param old_tree: The tree to show as the old tree in the comparison
513
:param new_tree: The tree to show as new in the comparison
514
:param to_file: The file to write comparison data to
515
:param path_encoding: The character encoding to write paths in
517
self.old_tree = old_tree
518
self.new_tree = new_tree
519
self.to_file = to_file
520
self.path_encoding = path_encoding
526
def from_diff_tree(klass, diff_tree):
527
return klass(diff_tree.old_tree, diff_tree.new_tree,
528
diff_tree.to_file, diff_tree.path_encoding)
531
def _diff_many(differs, file_id, old_path, new_path, old_kind, new_kind):
532
for file_differ in differs:
533
result = file_differ.diff(file_id, old_path, new_path, old_kind,
535
if result is not DiffPath.CANNOT_DIFF:
538
return DiffPath.CANNOT_DIFF
541
class DiffKindChange(object):
542
"""Special differ for file kind changes.
544
Represents kind change as deletion + creation. Uses the other differs
547
def __init__(self, differs):
548
self.differs = differs
554
def from_diff_tree(klass, diff_tree):
555
return klass(diff_tree.differs)
557
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
558
"""Perform comparison
560
:param file_id: The file_id of the file to compare
561
:param old_path: Path of the file in the old tree
562
:param new_path: Path of the file in the new tree
563
:param old_kind: Old file-kind of the file
564
:param new_kind: New file-kind of the file
566
if None in (old_kind, new_kind):
567
return DiffPath.CANNOT_DIFF
568
result = DiffPath._diff_many(self.differs, file_id, old_path,
569
new_path, old_kind, None)
570
if result is DiffPath.CANNOT_DIFF:
572
return DiffPath._diff_many(self.differs, file_id, old_path, new_path,
576
class DiffDirectory(DiffPath):
578
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
579
"""Perform comparison between two directories. (dummy)
582
if 'directory' not in (old_kind, new_kind):
583
return self.CANNOT_DIFF
584
if old_kind not in ('directory', None):
585
return self.CANNOT_DIFF
586
if new_kind not in ('directory', None):
587
return self.CANNOT_DIFF
591
class DiffSymlink(DiffPath):
593
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
594
"""Perform comparison between two symlinks
596
:param file_id: The file_id of the file to compare
597
:param old_path: Path of the file in the old tree
598
:param new_path: Path of the file in the new tree
599
:param old_kind: Old file-kind of the file
600
:param new_kind: New file-kind of the file
602
if 'symlink' not in (old_kind, new_kind):
603
return self.CANNOT_DIFF
604
if old_kind == 'symlink':
605
old_target = self.old_tree.get_symlink_target(file_id)
606
elif old_kind is None:
609
return self.CANNOT_DIFF
610
if new_kind == 'symlink':
611
new_target = self.new_tree.get_symlink_target(file_id)
612
elif new_kind is None:
615
return self.CANNOT_DIFF
616
return self.diff_symlink(old_target, new_target)
618
def diff_symlink(self, old_target, new_target):
619
if old_target is None:
620
self.to_file.write('=== target is %r\n' % new_target)
621
elif new_target is None:
622
self.to_file.write('=== target was %r\n' % old_target)
624
self.to_file.write('=== target changed %r => %r\n' %
625
(old_target, new_target))
629
class DiffText(DiffPath):
631
# GNU Patch uses the epoch date to detect files that are being added
632
# or removed in a diff.
633
EPOCH_DATE = '1970-01-01 00:00:00 +0000'
635
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
636
old_label='', new_label='', text_differ=internal_diff,
637
context_lines=DEFAULT_CONTEXT_AMOUNT):
638
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
639
self.text_differ = text_differ
640
self.old_label = old_label
641
self.new_label = new_label
642
self.path_encoding = path_encoding
643
self.context_lines = context_lines
645
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
646
"""Compare two files in unified diff format
648
:param file_id: The file_id of the file to compare
649
:param old_path: Path of the file in the old tree
650
:param new_path: Path of the file in the new tree
651
:param old_kind: Old file-kind of the file
652
:param new_kind: New file-kind of the file
654
if 'file' not in (old_kind, new_kind):
655
return self.CANNOT_DIFF
656
from_file_id = to_file_id = file_id
657
if old_kind == 'file':
658
old_date = _patch_header_date(self.old_tree, file_id, old_path)
659
elif old_kind is None:
660
old_date = self.EPOCH_DATE
663
return self.CANNOT_DIFF
664
if new_kind == 'file':
665
new_date = _patch_header_date(self.new_tree, file_id, new_path)
666
elif new_kind is None:
667
new_date = self.EPOCH_DATE
670
return self.CANNOT_DIFF
671
from_label = '%s%s\t%s' % (self.old_label, old_path, old_date)
672
to_label = '%s%s\t%s' % (self.new_label, new_path, new_date)
673
return self.diff_text(from_file_id, to_file_id, from_label, to_label,
676
def diff_text(self, from_file_id, to_file_id, from_label, to_label,
677
from_path=None, to_path=None):
678
"""Diff the content of given files in two trees
680
:param from_file_id: The id of the file in the from tree. If None,
681
the file is not present in the from tree.
682
:param to_file_id: The id of the file in the to tree. This may refer
683
to a different file from from_file_id. If None,
684
the file is not present in the to tree.
685
:param from_path: The path in the from tree or None if unknown.
686
:param to_path: The path in the to tree or None if unknown.
688
def _get_text(tree, file_id, path):
689
if file_id is not None:
690
return tree.get_file_lines(file_id, path)
694
from_text = _get_text(self.old_tree, from_file_id, from_path)
695
to_text = _get_text(self.new_tree, to_file_id, to_path)
696
self.text_differ(from_label, from_text, to_label, to_text,
697
self.to_file, path_encoding=self.path_encoding,
698
context_lines=self.context_lines)
699
except errors.BinaryFile:
701
("Binary files %s and %s differ\n" %
702
(from_label, to_label)).encode(self.path_encoding,'replace'))
706
class DiffFromTool(DiffPath):
708
def __init__(self, command_template, old_tree, new_tree, to_file,
709
path_encoding='utf-8'):
710
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
711
self.command_template = command_template
712
self._root = osutils.mkdtemp(prefix='brz-diff-')
715
def from_string(klass, command_string, old_tree, new_tree, to_file,
716
path_encoding='utf-8'):
717
command_template = cmdline.split(command_string)
718
if '@' not in command_string:
719
command_template.extend(['@old_path', '@new_path'])
720
return klass(command_template, old_tree, new_tree, to_file,
724
def make_from_diff_tree(klass, command_string, external_diff_options=None):
725
def from_diff_tree(diff_tree):
726
full_command_string = [command_string]
727
if external_diff_options is not None:
728
full_command_string += ' ' + external_diff_options
729
return klass.from_string(full_command_string, diff_tree.old_tree,
730
diff_tree.new_tree, diff_tree.to_file)
731
return from_diff_tree
733
def _get_command(self, old_path, new_path):
734
my_map = {'old_path': old_path, 'new_path': new_path}
735
command = [AtTemplate(t).substitute(my_map) for t in
736
self.command_template]
737
if sys.platform == 'win32': # Popen doesn't accept unicode on win32
740
if isinstance(c, unicode):
741
command_encoded.append(c.encode('mbcs'))
743
command_encoded.append(c)
744
return command_encoded
748
def _execute(self, old_path, new_path):
749
command = self._get_command(old_path, new_path)
751
proc = subprocess.Popen(command, stdout=subprocess.PIPE,
754
if e.errno == errno.ENOENT:
755
raise errors.ExecutableMissing(command[0])
758
self.to_file.write(proc.stdout.read())
761
def _try_symlink_root(self, tree, prefix):
762
if (getattr(tree, 'abspath', None) is None
763
or not osutils.host_os_dereferences_symlinks()):
766
os.symlink(tree.abspath(''), osutils.pathjoin(self._root, prefix))
768
if e.errno != errno.EEXIST:
774
"""Returns safe encoding for passing file path to diff tool"""
775
if sys.platform == 'win32':
778
# Don't fallback to 'utf-8' because subprocess may not be able to
779
# handle utf-8 correctly when locale is not utf-8.
780
return sys.getfilesystemencoding() or 'ascii'
782
def _is_safepath(self, path):
783
"""Return true if `path` may be able to pass to subprocess."""
786
return path == path.encode(fenc).decode(fenc)
790
def _safe_filename(self, prefix, relpath):
791
"""Replace unsafe character in `relpath` then join `self._root`,
792
`prefix` and `relpath`."""
794
# encoded_str.replace('?', '_') may break multibyte char.
795
# So we should encode, decode, then replace(u'?', u'_')
796
relpath_tmp = relpath.encode(fenc, 'replace').decode(fenc, 'replace')
797
relpath_tmp = relpath_tmp.replace(u'?', u'_')
798
return osutils.pathjoin(self._root, prefix, relpath_tmp)
800
def _write_file(self, file_id, tree, prefix, relpath, force_temp=False,
802
if not force_temp and isinstance(tree, WorkingTree):
803
full_path = tree.abspath(tree.id2path(file_id))
804
if self._is_safepath(full_path):
807
full_path = self._safe_filename(prefix, relpath)
808
if not force_temp and self._try_symlink_root(tree, prefix):
810
parent_dir = osutils.dirname(full_path)
812
os.makedirs(parent_dir)
814
if e.errno != errno.EEXIST:
816
source = tree.get_file(file_id, relpath)
818
target = open(full_path, 'wb')
820
osutils.pumpfile(source, target)
826
mtime = tree.get_file_mtime(file_id)
827
except FileTimestampUnavailable:
830
os.utime(full_path, (mtime, mtime))
832
osutils.make_readonly(full_path)
835
def _prepare_files(self, file_id, old_path, new_path, force_temp=False,
836
allow_write_new=False):
837
old_disk_path = self._write_file(file_id, self.old_tree, 'old',
838
old_path, force_temp)
839
new_disk_path = self._write_file(file_id, self.new_tree, 'new',
840
new_path, force_temp,
841
allow_write=allow_write_new)
842
return old_disk_path, new_disk_path
846
osutils.rmtree(self._root)
848
if e.errno != errno.ENOENT:
849
mutter("The temporary directory \"%s\" was not "
850
"cleanly removed: %s." % (self._root, e))
852
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
853
if (old_kind, new_kind) != ('file', 'file'):
854
return DiffPath.CANNOT_DIFF
855
(old_disk_path, new_disk_path) = self._prepare_files(
856
file_id, old_path, new_path)
857
self._execute(old_disk_path, new_disk_path)
859
def edit_file(self, file_id):
860
"""Use this tool to edit a file.
862
A temporary copy will be edited, and the new contents will be
865
:param file_id: The id of the file to edit.
866
:return: The new contents of the file.
868
old_path = self.old_tree.id2path(file_id)
869
new_path = self.new_tree.id2path(file_id)
870
old_abs_path, new_abs_path = self._prepare_files(
871
file_id, old_path, new_path,
872
allow_write_new=True,
874
command = self._get_command(old_abs_path, new_abs_path)
875
subprocess.call(command, cwd=self._root)
876
new_file = open(new_abs_path, 'rb')
878
return new_file.read()
883
class DiffTree(object):
884
"""Provides textual representations of the difference between two trees.
886
A DiffTree examines two trees and where a file-id has altered
887
between them, generates a textual representation of the difference.
888
DiffTree uses a sequence of DiffPath objects which are each
889
given the opportunity to handle a given altered fileid. The list
890
of DiffPath objects can be extended globally by appending to
891
DiffTree.diff_factories, or for a specific diff operation by
892
supplying the extra_factories option to the appropriate method.
895
# list of factories that can provide instances of DiffPath objects
896
# may be extended by plugins.
897
diff_factories = [DiffSymlink.from_diff_tree,
898
DiffDirectory.from_diff_tree]
900
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
901
diff_text=None, extra_factories=None):
904
:param old_tree: Tree to show as old in the comparison
905
:param new_tree: Tree to show as new in the comparison
906
:param to_file: File to write comparision to
907
:param path_encoding: Character encoding to write paths in
908
:param diff_text: DiffPath-type object to use as a last resort for
910
:param extra_factories: Factories of DiffPaths to try before any other
912
if diff_text is None:
913
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
914
'', '', internal_diff)
915
self.old_tree = old_tree
916
self.new_tree = new_tree
917
self.to_file = to_file
918
self.path_encoding = path_encoding
920
if extra_factories is not None:
921
self.differs.extend(f(self) for f in extra_factories)
922
self.differs.extend(f(self) for f in self.diff_factories)
923
self.differs.extend([diff_text, DiffKindChange.from_diff_tree(self)])
926
def from_trees_options(klass, old_tree, new_tree, to_file,
927
path_encoding, external_diff_options, old_label,
928
new_label, using, context_lines):
929
"""Factory for producing a DiffTree.
931
Designed to accept options used by show_diff_trees.
933
:param old_tree: The tree to show as old in the comparison
934
:param new_tree: The tree to show as new in the comparison
935
:param to_file: File to write comparisons to
936
:param path_encoding: Character encoding to use for writing paths
937
:param external_diff_options: If supplied, use the installed diff
938
binary to perform file comparison, using supplied options.
939
:param old_label: Prefix to use for old file labels
940
:param new_label: Prefix to use for new file labels
941
:param using: Commandline to use to invoke an external diff tool
943
if using is not None:
944
extra_factories = [DiffFromTool.make_from_diff_tree(using, external_diff_options)]
947
if external_diff_options:
948
opts = external_diff_options.split()
949
def diff_file(olab, olines, nlab, nlines, to_file, path_encoding=None, context_lines=None):
950
""":param path_encoding: not used but required
951
to match the signature of internal_diff.
953
external_diff(olab, olines, nlab, nlines, to_file, opts)
955
diff_file = internal_diff
956
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
957
old_label, new_label, diff_file, context_lines=context_lines)
958
return klass(old_tree, new_tree, to_file, path_encoding, diff_text,
961
def show_diff(self, specific_files, extra_trees=None):
962
"""Write tree diff to self.to_file
964
:param specific_files: the specific files to compare (recursive)
965
:param extra_trees: extra trees to use for mapping paths to file_ids
968
return self._show_diff(specific_files, extra_trees)
970
for differ in self.differs:
973
def _show_diff(self, specific_files, extra_trees):
974
# TODO: Generation of pseudo-diffs for added/deleted files could
975
# be usefully made into a much faster special case.
976
iterator = self.new_tree.iter_changes(self.old_tree,
977
specific_files=specific_files,
978
extra_trees=extra_trees,
979
require_versioned=True)
981
def changes_key(change):
982
old_path, new_path = change[1]
987
def get_encoded_path(path):
989
return path.encode(self.path_encoding, "replace")
990
for (file_id, paths, changed_content, versioned, parent, name, kind,
991
executable) in sorted(iterator, key=changes_key):
992
# The root does not get diffed, and items with no known kind (that
993
# is, missing) in both trees are skipped as well.
994
if parent == (None, None) or kind == (None, None):
996
oldpath, newpath = paths
997
oldpath_encoded = get_encoded_path(paths[0])
998
newpath_encoded = get_encoded_path(paths[1])
999
old_present = (kind[0] is not None and versioned[0])
1000
new_present = (kind[1] is not None and versioned[1])
1001
renamed = (parent[0], name[0]) != (parent[1], name[1])
1003
properties_changed = []
1004
properties_changed.extend(get_executable_change(executable[0], executable[1]))
1006
if properties_changed:
1007
prop_str = " (properties changed: %s)" % (", ".join(properties_changed),)
1011
if (old_present, new_present) == (True, False):
1012
self.to_file.write("=== removed %s '%s'\n" %
1013
(kind[0], oldpath_encoded))
1015
elif (old_present, new_present) == (False, True):
1016
self.to_file.write("=== added %s '%s'\n" %
1017
(kind[1], newpath_encoded))
1020
self.to_file.write("=== renamed %s '%s' => '%s'%s\n" %
1021
(kind[0], oldpath_encoded, newpath_encoded, prop_str))
1023
# if it was produced by iter_changes, it must be
1024
# modified *somehow*, either content or execute bit.
1025
self.to_file.write("=== modified %s '%s'%s\n" % (kind[0],
1026
newpath_encoded, prop_str))
1028
self._diff(file_id, oldpath, newpath, kind[0], kind[1])
1034
def diff(self, file_id, old_path, new_path):
1035
"""Perform a diff of a single file
1037
:param file_id: file-id of the file
1038
:param old_path: The path of the file in the old tree
1039
:param new_path: The path of the file in the new tree
1042
old_kind = self.old_tree.kind(file_id)
1043
except (errors.NoSuchId, errors.NoSuchFile):
1046
new_kind = self.new_tree.kind(file_id)
1047
except (errors.NoSuchId, errors.NoSuchFile):
1049
self._diff(file_id, old_path, new_path, old_kind, new_kind)
1052
def _diff(self, file_id, old_path, new_path, old_kind, new_kind):
1053
result = DiffPath._diff_many(self.differs, file_id, old_path,
1054
new_path, old_kind, new_kind)
1055
if result is DiffPath.CANNOT_DIFF:
1056
error_path = new_path
1057
if error_path is None:
1058
error_path = old_path
1059
raise errors.NoDiffFound(error_path)
1062
format_registry = Registry()
1063
format_registry.register('default', DiffTree)
added
removed
breezy/diff.py
