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
465
if extra_trees is not None:
466
for tree in extra_trees:
470
differ = format_cls.from_trees_options(old_tree, new_tree, to_file,
472
external_diff_options,
473
old_label, new_label, using,
474
context_lines=context)
475
return differ.show_diff(specific_files, extra_trees)
478
if extra_trees is not None:
479
for tree in extra_trees:
485
def _patch_header_date(tree, file_id, path):
486
"""Returns a timestamp suitable for use in a patch header."""
488
mtime = tree.get_file_mtime(file_id, path)
489
except FileTimestampUnavailable:
491
return timestamp.format_patch_date(mtime)
494
def get_executable_change(old_is_x, new_is_x):
495
descr = { True:"+x", False:"-x", None:"??" }
496
if old_is_x != new_is_x:
497
return ["%s to %s" % (descr[old_is_x], descr[new_is_x],)]
502
class DiffPath(object):
503
"""Base type for command object that compare files"""
505
# The type or contents of the file were unsuitable for diffing
506
CANNOT_DIFF = 'CANNOT_DIFF'
507
# The file has changed in a semantic way
509
# The file content may have changed, but there is no semantic change
510
UNCHANGED = 'UNCHANGED'
512
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8'):
515
:param old_tree: The tree to show as the old tree in the comparison
516
:param new_tree: The tree to show as new in the comparison
517
:param to_file: The file to write comparison data to
518
:param path_encoding: The character encoding to write paths in
520
self.old_tree = old_tree
521
self.new_tree = new_tree
522
self.to_file = to_file
523
self.path_encoding = path_encoding
529
def from_diff_tree(klass, diff_tree):
530
return klass(diff_tree.old_tree, diff_tree.new_tree,
531
diff_tree.to_file, diff_tree.path_encoding)
534
def _diff_many(differs, file_id, old_path, new_path, old_kind, new_kind):
535
for file_differ in differs:
536
result = file_differ.diff(file_id, old_path, new_path, old_kind,
538
if result is not DiffPath.CANNOT_DIFF:
541
return DiffPath.CANNOT_DIFF
544
class DiffKindChange(object):
545
"""Special differ for file kind changes.
547
Represents kind change as deletion + creation. Uses the other differs
550
def __init__(self, differs):
551
self.differs = differs
557
def from_diff_tree(klass, diff_tree):
558
return klass(diff_tree.differs)
560
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
561
"""Perform comparison
563
:param file_id: The file_id of the file to compare
564
:param old_path: Path of the file in the old tree
565
:param new_path: Path of the file in the new tree
566
:param old_kind: Old file-kind of the file
567
:param new_kind: New file-kind of the file
569
if None in (old_kind, new_kind):
570
return DiffPath.CANNOT_DIFF
571
result = DiffPath._diff_many(self.differs, file_id, old_path,
572
new_path, old_kind, None)
573
if result is DiffPath.CANNOT_DIFF:
575
return DiffPath._diff_many(self.differs, file_id, old_path, new_path,
579
class DiffDirectory(DiffPath):
581
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
582
"""Perform comparison between two directories. (dummy)
585
if 'directory' not in (old_kind, new_kind):
586
return self.CANNOT_DIFF
587
if old_kind not in ('directory', None):
588
return self.CANNOT_DIFF
589
if new_kind not in ('directory', None):
590
return self.CANNOT_DIFF
594
class DiffSymlink(DiffPath):
596
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
597
"""Perform comparison between two symlinks
599
:param file_id: The file_id of the file to compare
600
:param old_path: Path of the file in the old tree
601
:param new_path: Path of the file in the new tree
602
:param old_kind: Old file-kind of the file
603
:param new_kind: New file-kind of the file
605
if 'symlink' not in (old_kind, new_kind):
606
return self.CANNOT_DIFF
607
if old_kind == 'symlink':
608
old_target = self.old_tree.get_symlink_target(file_id)
609
elif old_kind is None:
612
return self.CANNOT_DIFF
613
if new_kind == 'symlink':
614
new_target = self.new_tree.get_symlink_target(file_id)
615
elif new_kind is None:
618
return self.CANNOT_DIFF
619
return self.diff_symlink(old_target, new_target)
621
def diff_symlink(self, old_target, new_target):
622
if old_target is None:
623
self.to_file.write('=== target is %r\n' % new_target)
624
elif new_target is None:
625
self.to_file.write('=== target was %r\n' % old_target)
627
self.to_file.write('=== target changed %r => %r\n' %
628
(old_target, new_target))
632
class DiffText(DiffPath):
634
# GNU Patch uses the epoch date to detect files that are being added
635
# or removed in a diff.
636
EPOCH_DATE = '1970-01-01 00:00:00 +0000'
638
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
639
old_label='', new_label='', text_differ=internal_diff,
640
context_lines=DEFAULT_CONTEXT_AMOUNT):
641
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
642
self.text_differ = text_differ
643
self.old_label = old_label
644
self.new_label = new_label
645
self.path_encoding = path_encoding
646
self.context_lines = context_lines
648
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
649
"""Compare two files in unified diff format
651
:param file_id: The file_id of the file to compare
652
:param old_path: Path of the file in the old tree
653
:param new_path: Path of the file in the new tree
654
:param old_kind: Old file-kind of the file
655
:param new_kind: New file-kind of the file
657
if 'file' not in (old_kind, new_kind):
658
return self.CANNOT_DIFF
659
from_file_id = to_file_id = file_id
660
if old_kind == 'file':
661
old_date = _patch_header_date(self.old_tree, file_id, old_path)
662
elif old_kind is None:
663
old_date = self.EPOCH_DATE
666
return self.CANNOT_DIFF
667
if new_kind == 'file':
668
new_date = _patch_header_date(self.new_tree, file_id, new_path)
669
elif new_kind is None:
670
new_date = self.EPOCH_DATE
673
return self.CANNOT_DIFF
674
from_label = '%s%s\t%s' % (self.old_label, old_path, old_date)
675
to_label = '%s%s\t%s' % (self.new_label, new_path, new_date)
676
return self.diff_text(from_file_id, to_file_id, from_label, to_label,
679
def diff_text(self, from_file_id, to_file_id, from_label, to_label,
680
from_path=None, to_path=None):
681
"""Diff the content of given files in two trees
683
:param from_file_id: The id of the file in the from tree. If None,
684
the file is not present in the from tree.
685
:param to_file_id: The id of the file in the to tree. This may refer
686
to a different file from from_file_id. If None,
687
the file is not present in the to tree.
688
:param from_path: The path in the from tree or None if unknown.
689
:param to_path: The path in the to tree or None if unknown.
691
def _get_text(tree, file_id, path):
692
if file_id is not None:
693
return tree.get_file_lines(file_id, path)
697
from_text = _get_text(self.old_tree, from_file_id, from_path)
698
to_text = _get_text(self.new_tree, to_file_id, to_path)
699
self.text_differ(from_label, from_text, to_label, to_text,
700
self.to_file, path_encoding=self.path_encoding,
701
context_lines=self.context_lines)
702
except errors.BinaryFile:
704
("Binary files %s and %s differ\n" %
705
(from_label, to_label)).encode(self.path_encoding,'replace'))
709
class DiffFromTool(DiffPath):
711
def __init__(self, command_template, old_tree, new_tree, to_file,
712
path_encoding='utf-8'):
713
DiffPath.__init__(self, old_tree, new_tree, to_file, path_encoding)
714
self.command_template = command_template
715
self._root = osutils.mkdtemp(prefix='brz-diff-')
718
def from_string(klass, command_string, old_tree, new_tree, to_file,
719
path_encoding='utf-8'):
720
command_template = cmdline.split(command_string)
721
if '@' not in command_string:
722
command_template.extend(['@old_path', '@new_path'])
723
return klass(command_template, old_tree, new_tree, to_file,
727
def make_from_diff_tree(klass, command_string, external_diff_options=None):
728
def from_diff_tree(diff_tree):
729
full_command_string = [command_string]
730
if external_diff_options is not None:
731
full_command_string += ' ' + external_diff_options
732
return klass.from_string(full_command_string, diff_tree.old_tree,
733
diff_tree.new_tree, diff_tree.to_file)
734
return from_diff_tree
736
def _get_command(self, old_path, new_path):
737
my_map = {'old_path': old_path, 'new_path': new_path}
738
command = [AtTemplate(t).substitute(my_map) for t in
739
self.command_template]
740
if sys.platform == 'win32': # Popen doesn't accept unicode on win32
743
if isinstance(c, unicode):
744
command_encoded.append(c.encode('mbcs'))
746
command_encoded.append(c)
747
return command_encoded
751
def _execute(self, old_path, new_path):
752
command = self._get_command(old_path, new_path)
754
proc = subprocess.Popen(command, stdout=subprocess.PIPE,
757
if e.errno == errno.ENOENT:
758
raise errors.ExecutableMissing(command[0])
761
self.to_file.write(proc.stdout.read())
764
def _try_symlink_root(self, tree, prefix):
765
if (getattr(tree, 'abspath', None) is None
766
or not osutils.host_os_dereferences_symlinks()):
769
os.symlink(tree.abspath(''), osutils.pathjoin(self._root, prefix))
771
if e.errno != errno.EEXIST:
777
"""Returns safe encoding for passing file path to diff tool"""
778
if sys.platform == 'win32':
781
# Don't fallback to 'utf-8' because subprocess may not be able to
782
# handle utf-8 correctly when locale is not utf-8.
783
return sys.getfilesystemencoding() or 'ascii'
785
def _is_safepath(self, path):
786
"""Return true if `path` may be able to pass to subprocess."""
789
return path == path.encode(fenc).decode(fenc)
793
def _safe_filename(self, prefix, relpath):
794
"""Replace unsafe character in `relpath` then join `self._root`,
795
`prefix` and `relpath`."""
797
# encoded_str.replace('?', '_') may break multibyte char.
798
# So we should encode, decode, then replace(u'?', u'_')
799
relpath_tmp = relpath.encode(fenc, 'replace').decode(fenc, 'replace')
800
relpath_tmp = relpath_tmp.replace(u'?', u'_')
801
return osutils.pathjoin(self._root, prefix, relpath_tmp)
803
def _write_file(self, file_id, tree, prefix, relpath, force_temp=False,
805
if not force_temp and isinstance(tree, WorkingTree):
806
full_path = tree.abspath(tree.id2path(file_id))
807
if self._is_safepath(full_path):
810
full_path = self._safe_filename(prefix, relpath)
811
if not force_temp and self._try_symlink_root(tree, prefix):
813
parent_dir = osutils.dirname(full_path)
815
os.makedirs(parent_dir)
817
if e.errno != errno.EEXIST:
819
source = tree.get_file(file_id, relpath)
821
target = open(full_path, 'wb')
823
osutils.pumpfile(source, target)
829
mtime = tree.get_file_mtime(file_id)
830
except FileTimestampUnavailable:
833
os.utime(full_path, (mtime, mtime))
835
osutils.make_readonly(full_path)
838
def _prepare_files(self, file_id, old_path, new_path, force_temp=False,
839
allow_write_new=False):
840
old_disk_path = self._write_file(file_id, self.old_tree, 'old',
841
old_path, force_temp)
842
new_disk_path = self._write_file(file_id, self.new_tree, 'new',
843
new_path, force_temp,
844
allow_write=allow_write_new)
845
return old_disk_path, new_disk_path
849
osutils.rmtree(self._root)
851
if e.errno != errno.ENOENT:
852
mutter("The temporary directory \"%s\" was not "
853
"cleanly removed: %s." % (self._root, e))
855
def diff(self, file_id, old_path, new_path, old_kind, new_kind):
856
if (old_kind, new_kind) != ('file', 'file'):
857
return DiffPath.CANNOT_DIFF
858
(old_disk_path, new_disk_path) = self._prepare_files(
859
file_id, old_path, new_path)
860
self._execute(old_disk_path, new_disk_path)
862
def edit_file(self, file_id):
863
"""Use this tool to edit a file.
865
A temporary copy will be edited, and the new contents will be
868
:param file_id: The id of the file to edit.
869
:return: The new contents of the file.
871
old_path = self.old_tree.id2path(file_id)
872
new_path = self.new_tree.id2path(file_id)
873
old_abs_path, new_abs_path = self._prepare_files(
874
file_id, old_path, new_path,
875
allow_write_new=True,
877
command = self._get_command(old_abs_path, new_abs_path)
878
subprocess.call(command, cwd=self._root)
879
new_file = open(new_abs_path, 'rb')
881
return new_file.read()
886
class DiffTree(object):
887
"""Provides textual representations of the difference between two trees.
889
A DiffTree examines two trees and where a file-id has altered
890
between them, generates a textual representation of the difference.
891
DiffTree uses a sequence of DiffPath objects which are each
892
given the opportunity to handle a given altered fileid. The list
893
of DiffPath objects can be extended globally by appending to
894
DiffTree.diff_factories, or for a specific diff operation by
895
supplying the extra_factories option to the appropriate method.
898
# list of factories that can provide instances of DiffPath objects
899
# may be extended by plugins.
900
diff_factories = [DiffSymlink.from_diff_tree,
901
DiffDirectory.from_diff_tree]
903
def __init__(self, old_tree, new_tree, to_file, path_encoding='utf-8',
904
diff_text=None, extra_factories=None):
907
:param old_tree: Tree to show as old in the comparison
908
:param new_tree: Tree to show as new in the comparison
909
:param to_file: File to write comparision to
910
:param path_encoding: Character encoding to write paths in
911
:param diff_text: DiffPath-type object to use as a last resort for
913
:param extra_factories: Factories of DiffPaths to try before any other
915
if diff_text is None:
916
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
917
'', '', internal_diff)
918
self.old_tree = old_tree
919
self.new_tree = new_tree
920
self.to_file = to_file
921
self.path_encoding = path_encoding
923
if extra_factories is not None:
924
self.differs.extend(f(self) for f in extra_factories)
925
self.differs.extend(f(self) for f in self.diff_factories)
926
self.differs.extend([diff_text, DiffKindChange.from_diff_tree(self)])
929
def from_trees_options(klass, old_tree, new_tree, to_file,
930
path_encoding, external_diff_options, old_label,
931
new_label, using, context_lines):
932
"""Factory for producing a DiffTree.
934
Designed to accept options used by show_diff_trees.
936
:param old_tree: The tree to show as old in the comparison
937
:param new_tree: The tree to show as new in the comparison
938
:param to_file: File to write comparisons to
939
:param path_encoding: Character encoding to use for writing paths
940
:param external_diff_options: If supplied, use the installed diff
941
binary to perform file comparison, using supplied options.
942
:param old_label: Prefix to use for old file labels
943
:param new_label: Prefix to use for new file labels
944
:param using: Commandline to use to invoke an external diff tool
946
if using is not None:
947
extra_factories = [DiffFromTool.make_from_diff_tree(using, external_diff_options)]
950
if external_diff_options:
951
opts = external_diff_options.split()
952
def diff_file(olab, olines, nlab, nlines, to_file, path_encoding=None, context_lines=None):
953
""":param path_encoding: not used but required
954
to match the signature of internal_diff.
956
external_diff(olab, olines, nlab, nlines, to_file, opts)
958
diff_file = internal_diff
959
diff_text = DiffText(old_tree, new_tree, to_file, path_encoding,
960
old_label, new_label, diff_file, context_lines=context_lines)
961
return klass(old_tree, new_tree, to_file, path_encoding, diff_text,
964
def show_diff(self, specific_files, extra_trees=None):
965
"""Write tree diff to self.to_file
967
:param specific_files: the specific files to compare (recursive)
968
:param extra_trees: extra trees to use for mapping paths to file_ids
971
return self._show_diff(specific_files, extra_trees)
973
for differ in self.differs:
976
def _show_diff(self, specific_files, extra_trees):
977
# TODO: Generation of pseudo-diffs for added/deleted files could
978
# be usefully made into a much faster special case.
979
iterator = self.new_tree.iter_changes(self.old_tree,
980
specific_files=specific_files,
981
extra_trees=extra_trees,
982
require_versioned=True)
984
def changes_key(change):
985
old_path, new_path = change[1]
990
def get_encoded_path(path):
992
return path.encode(self.path_encoding, "replace")
993
for (file_id, paths, changed_content, versioned, parent, name, kind,
994
executable) in sorted(iterator, key=changes_key):
995
# The root does not get diffed, and items with no known kind (that
996
# is, missing) in both trees are skipped as well.
997
if parent == (None, None) or kind == (None, None):
999
oldpath, newpath = paths
1000
oldpath_encoded = get_encoded_path(paths[0])
1001
newpath_encoded = get_encoded_path(paths[1])
1002
old_present = (kind[0] is not None and versioned[0])
1003
new_present = (kind[1] is not None and versioned[1])
1004
renamed = (parent[0], name[0]) != (parent[1], name[1])
1006
properties_changed = []
1007
properties_changed.extend(get_executable_change(executable[0], executable[1]))
1009
if properties_changed:
1010
prop_str = " (properties changed: %s)" % (", ".join(properties_changed),)
1014
if (old_present, new_present) == (True, False):
1015
self.to_file.write("=== removed %s '%s'\n" %
1016
(kind[0], oldpath_encoded))
1018
elif (old_present, new_present) == (False, True):
1019
self.to_file.write("=== added %s '%s'\n" %
1020
(kind[1], newpath_encoded))
1023
self.to_file.write("=== renamed %s '%s' => '%s'%s\n" %
1024
(kind[0], oldpath_encoded, newpath_encoded, prop_str))
1026
# if it was produced by iter_changes, it must be
1027
# modified *somehow*, either content or execute bit.
1028
self.to_file.write("=== modified %s '%s'%s\n" % (kind[0],
1029
newpath_encoded, prop_str))
1031
self._diff(file_id, oldpath, newpath, kind[0], kind[1])
1037
def diff(self, file_id, old_path, new_path):
1038
"""Perform a diff of a single file
1040
:param file_id: file-id of the file
1041
:param old_path: The path of the file in the old tree
1042
:param new_path: The path of the file in the new tree
1045
old_kind = self.old_tree.kind(file_id)
1046
except (errors.NoSuchId, errors.NoSuchFile):
1049
new_kind = self.new_tree.kind(file_id)
1050
except (errors.NoSuchId, errors.NoSuchFile):
1052
self._diff(file_id, old_path, new_path, old_kind, new_kind)
1055
def _diff(self, file_id, old_path, new_path, old_kind, new_kind):
1056
result = DiffPath._diff_many(self.differs, file_id, old_path,
1057
new_path, old_kind, new_kind)
1058
if result is DiffPath.CANNOT_DIFF:
1059
error_path = new_path
1060
if error_path is None:
1061
error_path = old_path
1062
raise errors.NoDiffFound(error_path)
1065
format_registry = Registry()
1066
format_registry.register('default', DiffTree)
added
removed
breezy/diff.py
