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 = unified_diff_bytes(
98
fromfile=old_label.encode(path_encoding, 'replace'),
99
tofile=new_label.encode(path_encoding, 'replace'),
100
n=context_lines, sequencematcher=sequence_matcher)
103
if len(ud) == 0: # Identical contents, nothing to do
105
# work-around for difflib being too smart for its own good
106
# if /dev/null is "1,0", patch won't recognize it as /dev/null
108
ud[2] = ud[2].replace(b'-1,0', b'-0,0')
110
ud[2] = ud[2].replace(b'+1,0', b'+0,0')
114
if not line.endswith(b'\n'):
115
to_file.write(b"\n\\ No newline at end of file\n")
119
def unified_diff_bytes(a, b, fromfile=b'', tofile=b'', fromfiledate=b'',
120
tofiledate=b'', n=3, lineterm=b'\n', sequencematcher=None):
122
Compare two sequences of lines; generate the delta as a unified diff.
124
Unified diffs are a compact way of showing line changes and a few
125
lines of context. The number of context lines is set by 'n' which
128
By default, the diff control lines (those with ---, +++, or @@) are
129
created with a trailing newline. This is helpful so that inputs
130
created from file.readlines() result in diffs that are suitable for
131
file.writelines() since both the inputs and outputs have trailing
134
For inputs that do not have trailing newlines, set the lineterm
135
argument to "" so that the output will be uniformly newline free.
137
The unidiff format normally has a header for filenames and modification
138
times. Any or all of these may be specified using strings for
139
'fromfile', 'tofile', 'fromfiledate', and 'tofiledate'. The modification
140
times are normally expressed in the format returned by time.ctime().
144
>>> for line in bytes_unified_diff(b'one two three four'.split(),
145
... b'zero one tree four'.split(), b'Original', b'Current',
146
... b'Sat Jan 26 23:30:50 1991', b'Fri Jun 06 10:20:52 2003',
149
--- Original Sat Jan 26 23:30:50 1991
150
+++ Current Fri Jun 06 10:20:52 2003
159
if sequencematcher is None:
160
sequencematcher = difflib.SequenceMatcher
163
fromfiledate = b'\t' + bytes(fromfiledate)
165
tofiledate = b'\t' + bytes(tofiledate)
168
for group in sequencematcher(None, a, b).get_grouped_opcodes(n):
170
yield b'--- %s%s%s' % (fromfile, fromfiledate, lineterm)
171
yield b'+++ %s%s%s' % (tofile, tofiledate, lineterm)
173
i1, i2, j1, j2 = group[0][1], group[-1][2], group[0][3], group[-1][4]
174
yield b"@@ -%d,%d +%d,%d @@%s" % (i1 + 1, i2 - i1, j1 + 1, j2 - j1, lineterm)
175
for tag, i1, i2, j1, j2 in group:
177
for line in a[i1:i2]:
180
if tag == 'replace' or tag == 'delete':
181
for line in a[i1:i2]:
183
if tag == 'replace' or tag == 'insert':
184
for line in b[j1:j2]:
188
def _spawn_external_diff(diffcmd, capture_errors=True):
189
"""Spawn the external diff process, and return the child handle.
191
:param diffcmd: The command list to spawn
192
:param capture_errors: Capture stderr as well as setting LANG=C
193
and LC_ALL=C. This lets us read and understand the output of diff,
194
and respond to any errors.
195
:return: A Popen object.
198
# construct minimal environment
200
path = os.environ.get('PATH')
203
env['LANGUAGE'] = 'C' # on win32 only LANGUAGE has effect
206
stderr = subprocess.PIPE
212
pipe = subprocess.Popen(diffcmd,
213
stdin=subprocess.PIPE,
214
stdout=subprocess.PIPE,
218
if e.errno == errno.ENOENT:
219
raise errors.NoDiff(str(e))
225
# diff style options as of GNU diff v3.2
226
style_option_list = ['-c', '-C', '--context',
228
'-f', '--forward-ed',
232
'-u', '-U', '--unified',
233
'-y', '--side-by-side',
237
def default_style_unified(diff_opts):
238
"""Default to unified diff style if alternative not specified in diff_opts.
240
diff only allows one style to be specified; they don't override.
241
Note that some of these take optargs, and the optargs can be
242
directly appended to the options.
243
This is only an approximate parser; it doesn't properly understand
246
:param diff_opts: List of options for external (GNU) diff.
247
:return: List of options with default style=='unified'.
249
for s in style_option_list:
257
diff_opts.append('-u')
261
def external_diff(old_label, oldlines, new_label, newlines, to_file,
263
"""Display a diff by calling out to the external diff program."""
264
# make sure our own output is properly ordered before the diff
267
oldtmp_fd, old_abspath = tempfile.mkstemp(prefix='brz-diff-old-')
268
newtmp_fd, new_abspath = tempfile.mkstemp(prefix='brz-diff-new-')
269
oldtmpf = os.fdopen(oldtmp_fd, 'wb')
270
newtmpf = os.fdopen(newtmp_fd, 'wb')
273
# TODO: perhaps a special case for comparing to or from the empty
274
# sequence; can just use /dev/null on Unix
276
# TODO: if either of the files being compared already exists as a
277
# regular named file (e.g. in the working directory) then we can
278
# compare directly to that, rather than copying it.
280
oldtmpf.writelines(oldlines)
281
newtmpf.writelines(newlines)
288
if sys.platform == 'win32':
289
# Popen doesn't do the proper encoding for external commands
290
# Since we are dealing with an ANSI api, use mbcs encoding
291
old_label = old_label.encode('mbcs')
292
new_label = new_label.encode('mbcs')
294
'--label', old_label,
296
'--label', new_label,
301
diff_opts = default_style_unified(diff_opts)
304
diffcmd.extend(diff_opts)
306
pipe = _spawn_external_diff(diffcmd, capture_errors=True)
307
out, err = pipe.communicate()
310
# internal_diff() adds a trailing newline, add one here for consistency
313
# 'diff' gives retcode == 2 for all sorts of errors
314
# one of those is 'Binary files differ'.
315
# Bad options could also be the problem.
316
# 'Binary files' is not a real error, so we suppress that error.
319
# Since we got here, we want to make sure to give an i18n error
320
pipe = _spawn_external_diff(diffcmd, capture_errors=False)
321
out, err = pipe.communicate()
323
# Write out the new i18n diff response
324
to_file.write(out + b'\n')
325
if pipe.returncode != 2:
326
raise errors.BzrError(
327
'external diff failed with exit code 2'
328
' when run with LANG=C and LC_ALL=C,'
329
' but not when run natively: %r' % (diffcmd,))
331
first_line = lang_c_out.split(b'\n', 1)[0]
332
# Starting with diffutils 2.8.4 the word "binary" was dropped.
333
m = re.match(b'^(binary )?files.*differ$', first_line, re.I)
335
raise errors.BzrError('external diff failed with exit code 2;'
336
' command: %r' % (diffcmd,))
338
# Binary files differ, just return
341
# If we got to here, we haven't written out the output of diff
345
# returns 1 if files differ; that's OK
347
msg = 'signal %d' % (-rc)
349
msg = 'exit code %d' % rc
351
raise errors.BzrError('external diff failed with %s; command: %r'
355
oldtmpf.close() # and delete
359
# Warn in case the file couldn't be deleted (in case windows still
360
# holds the file open, but not if the files have already been
365
if e.errno not in (errno.ENOENT,):
366
warning('Failed to delete temporary file: %s %s', path, e)
372
def get_trees_and_branches_to_diff_locked(
373
path_list, revision_specs, old_url, new_url, add_cleanup, apply_view=True):
374
"""Get the trees and specific files to diff given a list of paths.
376
This method works out the trees to be diff'ed and the files of
377
interest within those trees.
380
the list of arguments passed to the diff command
381
:param revision_specs:
382
Zero, one or two RevisionSpecs from the diff command line,
383
saying what revisions to compare.
385
The url of the old branch or tree. If None, the tree to use is
386
taken from the first path, if any, or the current working tree.
388
The url of the new branch or tree. If None, the tree to use is
389
taken from the first path, if any, or the current working tree.
391
a callable like Command.add_cleanup. get_trees_and_branches_to_diff
392
will register cleanups that must be run to unlock the trees, etc.
394
if True and a view is set, apply the view or check that the paths
397
a tuple of (old_tree, new_tree, old_branch, new_branch,
398
specific_files, extra_trees) where extra_trees is a sequence of
399
additional trees to search in for file-ids. The trees and branches
400
will be read-locked until the cleanups registered via the add_cleanup
403
# Get the old and new revision specs
404
old_revision_spec = None
405
new_revision_spec = None
406
if revision_specs is not None:
407
if len(revision_specs) > 0:
408
old_revision_spec = revision_specs[0]
410
old_url = old_revision_spec.get_branch()
411
if len(revision_specs) > 1:
412
new_revision_spec = revision_specs[1]
414
new_url = new_revision_spec.get_branch()
417
make_paths_wt_relative = True
418
consider_relpath = True
419
if path_list is None or len(path_list) == 0:
420
# If no path is given, the current working tree is used
421
default_location = u'.'
422
consider_relpath = False
423
elif old_url is not None and new_url is not None:
424
other_paths = path_list
425
make_paths_wt_relative = False
427
default_location = path_list[0]
428
other_paths = path_list[1:]
430
def lock_tree_or_branch(wt, br):
433
add_cleanup(wt.unlock)
436
add_cleanup(br.unlock)
438
# Get the old location
441
old_url = default_location
442
working_tree, branch, relpath = \
443
controldir.ControlDir.open_containing_tree_or_branch(old_url)
444
lock_tree_or_branch(working_tree, branch)
445
if consider_relpath and relpath != '':
446
if working_tree is not None and apply_view:
447
views.check_path_in_view(working_tree, relpath)
448
specific_files.append(relpath)
449
old_tree = _get_tree_to_diff(old_revision_spec, working_tree, branch)
452
# Get the new location
454
new_url = default_location
455
if new_url != old_url:
456
working_tree, branch, relpath = \
457
controldir.ControlDir.open_containing_tree_or_branch(new_url)
458
lock_tree_or_branch(working_tree, branch)
459
if consider_relpath and relpath != '':
460
if working_tree is not None and apply_view:
461
views.check_path_in_view(working_tree, relpath)
462
specific_files.append(relpath)
463
new_tree = _get_tree_to_diff(new_revision_spec, working_tree, branch,
464
basis_is_default=working_tree is None)
467
# Get the specific files (all files is None, no files is [])
468
if make_paths_wt_relative and working_tree is not None:
469
other_paths = working_tree.safe_relpath_files(
471
apply_view=apply_view)
472
specific_files.extend(other_paths)
473
if len(specific_files) == 0:
474
specific_files = None
475
if (working_tree is not None and working_tree.supports_views() and
477
view_files = working_tree.views.lookup_view()
479
specific_files = view_files
480
view_str = views.view_display_str(view_files)
481
note(gettext("*** Ignoring files outside view. View is %s") % view_str)
483
# Get extra trees that ought to be searched for file-ids
485
if working_tree is not None and working_tree not in (old_tree, new_tree):
486
extra_trees = (working_tree,)
487
return (old_tree, new_tree, old_branch, new_branch,
488
specific_files, extra_trees)
491
def _get_tree_to_diff(spec, tree=None, branch=None, basis_is_default=True):
492
if branch is None and tree is not None:
494
if spec is None or spec.spec is None:
497
return tree.basis_tree()
499
return branch.basis_tree()
502
return spec.as_tree(branch)
505
def show_diff_trees(old_tree, new_tree, to_file, specific_files=None,
506
external_diff_options=None,
507
old_label='a/', new_label='b/',
509
path_encoding='utf8',
512
context=DEFAULT_CONTEXT_AMOUNT):
513
"""Show in text form the changes from one tree to another.
515
:param to_file: The output stream.
516
:param specific_files: Include only changes to these files - None for all
518
:param external_diff_options: If set, use an external GNU diff and pass
520
:param extra_trees: If set, more Trees to use for looking up file ids
521
:param path_encoding: If set, the path will be encoded as specified,
522
otherwise is supposed to be utf8
523
:param format_cls: Formatter class (DiffTree subclass)
526
context = DEFAULT_CONTEXT_AMOUNT
527
if format_cls is None:
528
format_cls = DiffTree
529
with old_tree.lock_read():
530
if extra_trees is not None:
531
for tree in extra_trees:
535
differ = format_cls.from_trees_options(old_tree, new_tree, to_file,
537
external_diff_options,
538
old_label, new_label, using,
539
context_lines=context)
540
return differ.show_diff(specific_files, extra_trees)
543
if extra_trees is not None:
544
for tree in extra_trees:
548
def _patch_header_date(tree, path):
549
"""Returns a timestamp suitable for use in a patch header."""
551
mtime = tree.get_file_mtime(path)
552
except FileTimestampUnavailable:
554
return timestamp.format_patch_date(mtime)
557
def get_executable_change(old_is_x, new_is_x):
558
descr = {True: b"+x", False: b"-x", None: b"??"}
559
if old_is_x != new_is_x:
560
return [b"%s to %s" % (descr[old_is_x], descr[new_is_x],)]
565
class DiffPath(object):
566
"""Base type for command object that compare files"""
568
# The type or contents of the file were unsuitable for diffing
569
CANNOT_DIFF = 'CANNOT_DIFF'
570
# The file has changed in a semantic way
572
# The file content may have changed, but there is no semantic change
573
UNCHANGED = 'UNCHANGED'
575
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8'):
578
:param old_tree: The tree to show as the old tree in the comparison
579
:param new_tree: The tree to show as new in the comparison
580
:param to_file: The file to write comparison data to
581
:param path_encoding: The character encoding to write paths in
583
self.old_tree = old_tree
584
self.new_tree = new_tree
585
self.to_file = to_file
586
self.path_encoding = path_encoding
592
def from_diff_tree(klass, diff_tree):
593
return klass(diff_tree.old_tree, diff_tree.new_tree,
594
diff_tree.to_file, diff_tree.path_encoding)
597
def _diff_many(differs, old_path, new_path, old_kind, new_kind):
598
for file_differ in differs:
599
result = file_differ.diff(old_path, new_path, old_kind, new_kind)
600
if result is not DiffPath.CANNOT_DIFF:
603
return DiffPath.CANNOT_DIFF
606
class DiffKindChange(object):
607
"""Special differ for file kind changes.
609
Represents kind change as deletion + creation. Uses the other differs
613
def __init__(self, differs):
614
self.differs = differs
620
def from_diff_tree(klass, diff_tree):
621
return klass(diff_tree.differs)
623
def diff(self, old_path, new_path, old_kind, new_kind):
624
"""Perform comparison
626
:param old_path: Path of the file in the old tree
627
:param new_path: Path of the file in the new tree
628
:param old_kind: Old file-kind of the file
629
:param new_kind: New file-kind of the file
631
if None in (old_kind, new_kind):
632
return DiffPath.CANNOT_DIFF
633
result = DiffPath._diff_many(
634
self.differs, old_path, new_path, old_kind, None)
635
if result is DiffPath.CANNOT_DIFF:
637
return DiffPath._diff_many(
638
self.differs, old_path, new_path, None, new_kind)
641
class DiffTreeReference(DiffPath):
643
def diff(self, old_path, new_path, old_kind, new_kind):
644
"""Perform comparison between two tree references. (dummy)
647
if 'tree-reference' not in (old_kind, new_kind):
648
return self.CANNOT_DIFF
649
if old_kind not in ('tree-reference', None):
650
return self.CANNOT_DIFF
651
if new_kind not in ('tree-reference', None):
652
return self.CANNOT_DIFF
656
class DiffDirectory(DiffPath):
658
def diff(self, old_path, new_path, old_kind, new_kind):
659
"""Perform comparison between two directories. (dummy)
662
if 'directory' not in (old_kind, new_kind):
663
return self.CANNOT_DIFF
664
if old_kind not in ('directory', None):
665
return self.CANNOT_DIFF
666
if new_kind not in ('directory', None):
667
return self.CANNOT_DIFF
671
class DiffSymlink(DiffPath):
673
def diff(self, old_path, new_path, old_kind, new_kind):
674
"""Perform comparison between two symlinks
676
:param old_path: Path of the file in the old tree
677
:param new_path: Path of the file in the new tree
678
:param old_kind: Old file-kind of the file
679
:param new_kind: New file-kind of the file
681
if 'symlink' not in (old_kind, new_kind):
682
return self.CANNOT_DIFF
683
if old_kind == 'symlink':
684
old_target = self.old_tree.get_symlink_target(old_path)
685
elif old_kind is None:
688
return self.CANNOT_DIFF
689
if new_kind == 'symlink':
690
new_target = self.new_tree.get_symlink_target(new_path)
691
elif new_kind is None:
694
return self.CANNOT_DIFF
695
return self.diff_symlink(old_target, new_target)
697
def diff_symlink(self, old_target, new_target):
698
if old_target is None:
699
self.to_file.write(b'=== target is \'%s\'\n' %
700
new_target.encode(self.path_encoding, 'replace'))
701
elif new_target is None:
702
self.to_file.write(b'=== target was \'%s\'\n' %
703
old_target.encode(self.path_encoding, 'replace'))
705
self.to_file.write(b'=== target changed \'%s\' => \'%s\'\n' %
706
(old_target.encode(self.path_encoding, 'replace'),
707
new_target.encode(self.path_encoding, 'replace')))
711
class DiffText(DiffPath):
713
# GNU Patch uses the epoch date to detect files that are being added
714
# or removed in a diff.
715
EPOCH_DATE = '1970-01-01 00:00:00 +0000'
717
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
718
old_label='', new_label='', text_differ=internal_diff,
719
context_lines=DEFAULT_CONTEXT_AMOUNT):
720
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
721
self.text_differ = text_differ
722
self.old_label = old_label
723
self.new_label = new_label
724
self.path_encoding = path_encoding
725
self.context_lines = context_lines
727
def diff(self, old_path, new_path, old_kind, new_kind):
728
"""Compare two files in unified diff format
730
:param old_path: Path of the file in the old tree
731
:param new_path: Path of the file in the new tree
732
:param old_kind: Old file-kind of the file
733
:param new_kind: New file-kind of the file
735
if 'file' not in (old_kind, new_kind):
736
return self.CANNOT_DIFF
737
if old_kind == 'file':
738
old_date = _patch_header_date(self.old_tree, old_path)
739
elif old_kind is None:
740
old_date = self.EPOCH_DATE
742
return self.CANNOT_DIFF
743
if new_kind == 'file':
744
new_date = _patch_header_date(self.new_tree, new_path)
745
elif new_kind is None:
746
new_date = self.EPOCH_DATE
748
return self.CANNOT_DIFF
749
from_label = '%s%s\t%s' % (self.old_label, old_path,
751
to_label = '%s%s\t%s' % (self.new_label, new_path,
753
return self.diff_text(old_path, new_path, from_label, to_label)
755
def diff_text(self, from_path, to_path, from_label, to_label):
756
"""Diff the content of given files in two trees
758
:param from_path: The path in the from tree. If None,
759
the file is not present in the from tree.
760
:param to_path: The path in the to tree. This may refer
761
to a different file from from_path. If None,
762
the file is not present in the to tree.
764
def _get_text(tree, path):
768
return tree.get_file_lines(path)
769
except errors.NoSuchFile:
772
from_text = _get_text(self.old_tree, from_path)
773
to_text = _get_text(self.new_tree, to_path)
774
self.text_differ(from_label, from_text, to_label, to_text,
775
self.to_file, path_encoding=self.path_encoding,
776
context_lines=self.context_lines)
777
except errors.BinaryFile:
779
("Binary files %s and %s differ\n" %
780
(from_label, to_label)).encode(self.path_encoding, 'replace'))
784
class DiffFromTool(DiffPath):
786
def __init__(self, command_template, old_tree, new_tree, to_file,
787
path_encoding='utf-8'):
788
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
789
self.command_template = command_template
790
self._root = osutils.mkdtemp(prefix='brz-diff-')
793
def from_string(klass, command_string, old_tree, new_tree, to_file,
794
path_encoding='utf-8'):
795
command_template = cmdline.split(command_string)
796
if '@' not in command_string:
797
command_template.extend(['@old_path', '@new_path'])
798
return klass(command_template, old_tree, new_tree, to_file,
802
def make_from_diff_tree(klass, command_string, external_diff_options=None):
803
def from_diff_tree(diff_tree):
804
full_command_string = [command_string]
805
if external_diff_options is not None:
806
full_command_string += ' ' + external_diff_options
807
return klass.from_string(full_command_string, diff_tree.old_tree,
808
diff_tree.new_tree, diff_tree.to_file)
809
return from_diff_tree
811
def _get_command(self, old_path, new_path):
812
my_map = {'old_path': old_path, 'new_path': new_path}
813
command = [AtTemplate(t).substitute(my_map) for t in
814
self.command_template]
815
if sys.platform == 'win32': # Popen doesn't accept unicode on win32
818
if isinstance(c, text_type):
819
command_encoded.append(c.encode('mbcs'))
821
command_encoded.append(c)
822
return command_encoded
826
def _execute(self, old_path, new_path):
827
command = self._get_command(old_path, new_path)
829
proc = subprocess.Popen(command, stdout=subprocess.PIPE,
832
if e.errno == errno.ENOENT:
833
raise errors.ExecutableMissing(command[0])
836
self.to_file.write(proc.stdout.read())
840
def _try_symlink_root(self, tree, prefix):
841
if (getattr(tree, 'abspath', None) is None or
842
not osutils.host_os_dereferences_symlinks()):
845
os.symlink(tree.abspath(''), osutils.pathjoin(self._root, prefix))
847
if e.errno != errno.EEXIST:
853
"""Returns safe encoding for passing file path to diff tool"""
854
if sys.platform == 'win32':
857
# Don't fallback to 'utf-8' because subprocess may not be able to
858
# handle utf-8 correctly when locale is not utf-8.
859
return sys.getfilesystemencoding() or 'ascii'
861
def _is_safepath(self, path):
862
"""Return true if `path` may be able to pass to subprocess."""
865
return path == path.encode(fenc).decode(fenc)
869
def _safe_filename(self, prefix, relpath):
870
"""Replace unsafe character in `relpath` then join `self._root`,
871
`prefix` and `relpath`."""
873
# encoded_str.replace('?', '_') may break multibyte char.
874
# So we should encode, decode, then replace(u'?', u'_')
875
relpath_tmp = relpath.encode(fenc, 'replace').decode(fenc, 'replace')
876
relpath_tmp = relpath_tmp.replace(u'?', u'_')
877
return osutils.pathjoin(self._root, prefix, relpath_tmp)
879
def _write_file(self, relpath, tree, prefix, force_temp=False,
881
if not force_temp and isinstance(tree, WorkingTree):
882
full_path = tree.abspath(relpath)
883
if self._is_safepath(full_path):
886
full_path = self._safe_filename(prefix, relpath)
887
if not force_temp and self._try_symlink_root(tree, prefix):
889
parent_dir = osutils.dirname(full_path)
891
os.makedirs(parent_dir)
893
if e.errno != errno.EEXIST:
895
source = tree.get_file(relpath)
897
with open(full_path, 'wb') as target:
898
osutils.pumpfile(source, target)
902
mtime = tree.get_file_mtime(relpath)
903
except FileTimestampUnavailable:
906
os.utime(full_path, (mtime, mtime))
908
osutils.make_readonly(full_path)
911
def _prepare_files(self, old_path, new_path, force_temp=False,
912
allow_write_new=False):
913
old_disk_path = self._write_file(
914
old_path, self.old_tree, 'old', force_temp)
915
new_disk_path = self._write_file(
916
new_path, self.new_tree, 'new', force_temp,
917
allow_write=allow_write_new)
918
return old_disk_path, new_disk_path
922
osutils.rmtree(self._root)
924
if e.errno != errno.ENOENT:
925
mutter("The temporary directory \"%s\" was not "
926
"cleanly removed: %s." % (self._root, e))
928
def diff(self, old_path, new_path, old_kind, new_kind):
929
if (old_kind, new_kind) != ('file', 'file'):
930
return DiffPath.CANNOT_DIFF
931
(old_disk_path, new_disk_path) = self._prepare_files(
933
self._execute(old_disk_path, new_disk_path)
935
def edit_file(self, old_path, new_path):
936
"""Use this tool to edit a file.
938
A temporary copy will be edited, and the new contents will be
941
:return: The new contents of the file.
943
old_abs_path, new_abs_path = self._prepare_files(
944
old_path, new_path, allow_write_new=True, force_temp=True)
945
command = self._get_command(old_abs_path, new_abs_path)
946
subprocess.call(command, cwd=self._root)
947
with open(new_abs_path, 'rb') as new_file:
948
return new_file.read()
951
class DiffTree(object):
952
"""Provides textual representations of the difference between two trees.
954
A DiffTree examines two trees and where a file-id has altered
955
between them, generates a textual representation of the difference.
956
DiffTree uses a sequence of DiffPath objects which are each
957
given the opportunity to handle a given altered fileid. The list
958
of DiffPath objects can be extended globally by appending to
959
DiffTree.diff_factories, or for a specific diff operation by
960
supplying the extra_factories option to the appropriate method.
963
# list of factories that can provide instances of DiffPath objects
964
# may be extended by plugins.
965
diff_factories = [DiffSymlink.from_diff_tree,
966
DiffDirectory.from_diff_tree,
967
DiffTreeReference.from_diff_tree]
969
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
970
diff_text=None, extra_factories=None):
973
:param old_tree: Tree to show as old in the comparison
974
:param new_tree: Tree to show as new in the comparison
975
:param to_file: File to write comparision to
976
:param path_encoding: Character encoding to write paths in
977
:param diff_text: DiffPath-type object to use as a last resort for
979
:param extra_factories: Factories of DiffPaths to try before any other
981
if diff_text is None:
982
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
983
'', '', internal_diff)
984
self.old_tree = old_tree
985
self.new_tree = new_tree
986
self.to_file = to_file
987
self.path_encoding = path_encoding
989
if extra_factories is not None:
990
self.differs.extend(f(self) for f in extra_factories)
991
self.differs.extend(f(self) for f in self.diff_factories)
992
self.differs.extend([diff_text, DiffKindChange.from_diff_tree(self)])
995
def from_trees_options(klass, old_tree, new_tree, to_file,
996
path_encoding, external_diff_options, old_label,
997
new_label, using, context_lines):
998
"""Factory for producing a DiffTree.
1000
Designed to accept options used by show_diff_trees.
1002
:param old_tree: The tree to show as old in the comparison
1003
:param new_tree: The tree to show as new in the comparison
1004
:param to_file: File to write comparisons to
1005
:param path_encoding: Character encoding to use for writing paths
1006
:param external_diff_options: If supplied, use the installed diff
1007
binary to perform file comparison, using supplied options.
1008
:param old_label: Prefix to use for old file labels
1009
:param new_label: Prefix to use for new file labels
1010
:param using: Commandline to use to invoke an external diff tool
1012
if using is not None:
1013
extra_factories = [DiffFromTool.make_from_diff_tree(
1014
using, external_diff_options)]
1016
extra_factories = []
1017
if external_diff_options:
1018
opts = external_diff_options.split()
1020
def diff_file(olab, olines, nlab, nlines, to_file, path_encoding=None, context_lines=None):
1021
""":param path_encoding: not used but required
1022
to match the signature of internal_diff.
1024
external_diff(olab, olines, nlab, nlines, to_file, opts)
1026
diff_file = internal_diff
1027
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
1028
old_label, new_label, diff_file, context_lines=context_lines)
1029
return klass(old_tree, new_tree, to_file, path_encoding, diff_text,
1032
def show_diff(self, specific_files, extra_trees=None):
1033
"""Write tree diff to self.to_file
1035
:param specific_files: the specific files to compare (recursive)
1036
:param extra_trees: extra trees to use for mapping paths to file_ids
1039
return self._show_diff(specific_files, extra_trees)
1041
for differ in self.differs:
1044
def _show_diff(self, specific_files, extra_trees):
1045
# TODO: Generation of pseudo-diffs for added/deleted files could
1046
# be usefully made into a much faster special case.
1047
iterator = self.new_tree.iter_changes(self.old_tree,
1048
specific_files=specific_files,
1049
extra_trees=extra_trees,
1050
require_versioned=True)
1053
def changes_key(change):
1054
old_path, new_path = change[1]
1060
def get_encoded_path(path):
1061
if path is not None:
1062
return path.encode(self.path_encoding, "replace")
1063
for change in sorted(iterator, key=changes_key):
1064
# The root does not get diffed, and items with no known kind (that
1065
# is, missing) in both trees are skipped as well.
1066
if change.parent_id == (None, None) or change.kind == (None, None):
1068
if change.kind[0] == 'symlink' and not self.new_tree.supports_symlinks():
1070
'Ignoring "%s" as symlinks are not '
1071
'supported on this filesystem.' % (change.path[0],))
1073
oldpath, newpath = change.path
1074
oldpath_encoded = get_encoded_path(change.path[0])
1075
newpath_encoded = get_encoded_path(change.path[1])
1076
old_present = (change.kind[0] is not None and change.versioned[0])
1077
new_present = (change.kind[1] is not None and change.versioned[1])
1078
executable = change.executable
1080
renamed = (change.parent_id[0], change.name[0]) != (change.parent_id[1], change.name[1])
1082
properties_changed = []
1083
properties_changed.extend(
1084
get_executable_change(executable[0], executable[1]))
1086
if properties_changed:
1087
prop_str = b" (properties changed: %s)" % (
1088
b", ".join(properties_changed),)
1092
if (old_present, new_present) == (True, False):
1093
self.to_file.write(b"=== removed %s '%s'\n" %
1094
(kind[0].encode('ascii'), oldpath_encoded))
1096
elif (old_present, new_present) == (False, True):
1097
self.to_file.write(b"=== added %s '%s'\n" %
1098
(kind[1].encode('ascii'), newpath_encoded))
1101
self.to_file.write(b"=== renamed %s '%s' => '%s'%s\n" %
1102
(kind[0].encode('ascii'), oldpath_encoded, newpath_encoded, prop_str))
1104
# if it was produced by iter_changes, it must be
1105
# modified *somehow*, either content or execute bit.
1106
self.to_file.write(b"=== modified %s '%s'%s\n" % (kind[0].encode('ascii'),
1107
newpath_encoded, prop_str))
1108
if change.changed_content:
1109
self._diff(oldpath, newpath, kind[0], kind[1])
1115
def diff(self, old_path, new_path):
1116
"""Perform a diff of a single file
1118
:param old_path: The path of the file in the old tree
1119
:param new_path: The path of the file in the new tree
1121
if old_path is None:
1124
old_kind = self.old_tree.kind(old_path)
1125
if new_path is None:
1128
new_kind = self.new_tree.kind(new_path)
1129
self._diff(old_path, new_path, old_kind, new_kind)
1131
def _diff(self, old_path, new_path, old_kind, new_kind):
1132
result = DiffPath._diff_many(
1133
self.differs, old_path, new_path, old_kind, new_kind)
1134
if result is DiffPath.CANNOT_DIFF:
1135
error_path = new_path
1136
if error_path is None:
1137
error_path = old_path
1138
raise errors.NoDiffFound(error_path)
1141
format_registry = Registry()
1142
format_registry.register('default', DiffTree)
added
removed
breezy/diff.py
