~bzr-pqm/bzr/bzr.dev

« back to all changes in this revision

Viewing changes to bzrlib/diff.py

  • Committer: Jelmer Vernooij
  • Date: 2011-12-19 10:58:39 UTC
  • mfrom: (6383 +trunk)
  • mto: This revision was merged to the branch mainline in revision 6386.
  • Revision ID: jelmer@canonical.com-20111219105839-uji05ck4rkm1mj4j
Merge bzr.dev.

Show diffs side-by-side

added added

removed removed

Lines of Context:
1
 
# Copyright (C) 2004, 2005, 2006 Canonical Ltd.
 
1
# Copyright (C) 2005-2011 Canonical Ltd.
2
2
#
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
14
14
# along with this program; if not, write to the Free Software
15
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
16
16
 
 
17
from __future__ import absolute_import
 
18
 
17
19
import difflib
18
20
import os
19
21
import re
20
 
import shutil
 
22
import string
21
23
import sys
22
24
 
23
25
from bzrlib.lazy_import import lazy_import
25
27
import errno
26
28
import subprocess
27
29
import tempfile
28
 
import time
29
30
 
30
31
from bzrlib import (
31
 
    branch as _mod_branch,
32
 
    bzrdir,
33
 
    commands,
 
32
    cleanup,
 
33
    cmdline,
 
34
    controldir,
34
35
    errors,
35
36
    osutils,
36
37
    patiencediff,
38
39
    timestamp,
39
40
    views,
40
41
    )
 
42
 
 
43
from bzrlib.workingtree import WorkingTree
 
44
from bzrlib.i18n import gettext
41
45
""")
42
46
 
43
 
from bzrlib.symbol_versioning import (
44
 
    deprecated_function,
 
47
from bzrlib.registry import (
 
48
    Registry,
45
49
    )
46
50
from bzrlib.trace import mutter, note, warning
47
51
 
48
52
 
 
53
class AtTemplate(string.Template):
 
54
    """Templating class that uses @ instead of $."""
 
55
 
 
56
    delimiter = '@'
 
57
 
 
58
 
49
59
# TODO: Rather than building a changeset object, we should probably
50
60
# invoke callbacks on an object.  That object can either accumulate a
51
61
# list, write them out directly, etc etc.
86
96
    if sequence_matcher is None:
87
97
        sequence_matcher = patiencediff.PatienceSequenceMatcher
88
98
    ud = patiencediff.unified_diff(oldlines, newlines,
89
 
                      fromfile=old_filename.encode(path_encoding),
90
 
                      tofile=new_filename.encode(path_encoding),
 
99
                      fromfile=old_filename.encode(path_encoding, 'replace'),
 
100
                      tofile=new_filename.encode(path_encoding, 'replace'),
91
101
                      sequencematcher=sequence_matcher)
92
102
 
93
103
    ud = list(ud)
171
181
 
172
182
        if not diff_opts:
173
183
            diff_opts = []
 
184
        if sys.platform == 'win32':
 
185
            # Popen doesn't do the proper encoding for external commands
 
186
            # Since we are dealing with an ANSI api, use mbcs encoding
 
187
            old_filename = old_filename.encode('mbcs')
 
188
            new_filename = new_filename.encode('mbcs')
174
189
        diffcmd = ['diff',
175
190
                   '--label', old_filename,
176
191
                   old_abspath,
272
287
                        new_abspath, e)
273
288
 
274
289
 
275
 
def _get_trees_to_diff(path_list, revision_specs, old_url, new_url,
276
 
    apply_view=True):
 
290
def get_trees_and_branches_to_diff_locked(
 
291
    path_list, revision_specs, old_url, new_url, add_cleanup, apply_view=True):
277
292
    """Get the trees and specific files to diff given a list of paths.
278
293
 
279
294
    This method works out the trees to be diff'ed and the files of
290
305
    :param new_url:
291
306
        The url of the new branch or tree. If None, the tree to use is
292
307
        taken from the first path, if any, or the current working tree.
 
308
    :param add_cleanup:
 
309
        a callable like Command.add_cleanup.  get_trees_and_branches_to_diff
 
310
        will register cleanups that must be run to unlock the trees, etc.
293
311
    :param apply_view:
294
312
        if True and a view is set, apply the view or check that the paths
295
313
        are within it
296
314
    :returns:
297
 
        a tuple of (old_tree, new_tree, specific_files, extra_trees) where
298
 
        extra_trees is a sequence of additional trees to search in for
299
 
        file-ids.
 
315
        a tuple of (old_tree, new_tree, old_branch, new_branch,
 
316
        specific_files, extra_trees) where extra_trees is a sequence of
 
317
        additional trees to search in for file-ids.  The trees and branches
 
318
        will be read-locked until the cleanups registered via the add_cleanup
 
319
        param are run.
300
320
    """
301
321
    # Get the old and new revision specs
302
322
    old_revision_spec = None
325
345
        default_location = path_list[0]
326
346
        other_paths = path_list[1:]
327
347
 
 
348
    def lock_tree_or_branch(wt, br):
 
349
        if wt is not None:
 
350
            wt.lock_read()
 
351
            add_cleanup(wt.unlock)
 
352
        elif br is not None:
 
353
            br.lock_read()
 
354
            add_cleanup(br.unlock)
 
355
 
328
356
    # Get the old location
329
357
    specific_files = []
330
358
    if old_url is None:
331
359
        old_url = default_location
332
360
    working_tree, branch, relpath = \
333
 
        bzrdir.BzrDir.open_containing_tree_or_branch(old_url)
 
361
        controldir.ControlDir.open_containing_tree_or_branch(old_url)
 
362
    lock_tree_or_branch(working_tree, branch)
334
363
    if consider_relpath and relpath != '':
335
364
        if working_tree is not None and apply_view:
336
365
            views.check_path_in_view(working_tree, relpath)
337
366
        specific_files.append(relpath)
338
367
    old_tree = _get_tree_to_diff(old_revision_spec, working_tree, branch)
 
368
    old_branch = branch
339
369
 
340
370
    # Get the new location
341
371
    if new_url is None:
342
372
        new_url = default_location
343
373
    if new_url != old_url:
344
374
        working_tree, branch, relpath = \
345
 
            bzrdir.BzrDir.open_containing_tree_or_branch(new_url)
 
375
            controldir.ControlDir.open_containing_tree_or_branch(new_url)
 
376
        lock_tree_or_branch(working_tree, branch)
346
377
        if consider_relpath and relpath != '':
347
378
            if working_tree is not None and apply_view:
348
379
                views.check_path_in_view(working_tree, relpath)
349
380
            specific_files.append(relpath)
350
381
    new_tree = _get_tree_to_diff(new_revision_spec, working_tree, branch,
351
382
        basis_is_default=working_tree is None)
 
383
    new_branch = branch
352
384
 
353
385
    # Get the specific files (all files is None, no files is [])
354
386
    if make_paths_wt_relative and working_tree is not None:
355
 
        try:
356
 
            from bzrlib.builtins import safe_relpath_files
357
 
            other_paths = safe_relpath_files(working_tree, other_paths,
 
387
        other_paths = working_tree.safe_relpath_files(
 
388
            other_paths,
358
389
            apply_view=apply_view)
359
 
        except errors.FileInWrongBranch:
360
 
            raise errors.BzrCommandError("Files are in different branches")
361
390
    specific_files.extend(other_paths)
362
391
    if len(specific_files) == 0:
363
392
        specific_files = None
367
396
            if view_files:
368
397
                specific_files = view_files
369
398
                view_str = views.view_display_str(view_files)
370
 
                note("*** Ignoring files outside view. View is %s" % view_str)
 
399
                note(gettext("*** Ignoring files outside view. View is %s") % view_str)
371
400
 
372
401
    # Get extra trees that ought to be searched for file-ids
373
402
    extra_trees = None
374
403
    if working_tree is not None and working_tree not in (old_tree, new_tree):
375
404
        extra_trees = (working_tree,)
376
 
    return old_tree, new_tree, specific_files, extra_trees
 
405
    return (old_tree, new_tree, old_branch, new_branch,
 
406
            specific_files, extra_trees)
 
407
 
377
408
 
378
409
def _get_tree_to_diff(spec, tree=None, branch=None, basis_is_default=True):
379
410
    if branch is None and tree is not None:
394
425
                    old_label='a/', new_label='b/',
395
426
                    extra_trees=None,
396
427
                    path_encoding='utf8',
397
 
                    using=None):
 
428
                    using=None,
 
429
                    format_cls=None):
398
430
    """Show in text form the changes from one tree to another.
399
431
 
400
 
    to_file
401
 
        The output stream.
402
 
 
403
 
    specific_files
404
 
        Include only changes to these files - None for all changes.
405
 
 
406
 
    external_diff_options
407
 
        If set, use an external GNU diff and pass these options.
408
 
 
409
 
    extra_trees
410
 
        If set, more Trees to use for looking up file ids
411
 
 
412
 
    path_encoding
413
 
        If set, the path will be encoded as specified, otherwise is supposed
414
 
        to be utf8
 
432
    :param to_file: The output stream.
 
433
    :param specific_files: Include only changes to these files - None for all
 
434
        changes.
 
435
    :param external_diff_options: If set, use an external GNU diff and pass 
 
436
        these options.
 
437
    :param extra_trees: If set, more Trees to use for looking up file ids
 
438
    :param path_encoding: If set, the path will be encoded as specified, 
 
439
        otherwise is supposed to be utf8
 
440
    :param format_cls: Formatter class (DiffTree subclass)
415
441
    """
 
442
    if format_cls is None:
 
443
        format_cls = DiffTree
416
444
    old_tree.lock_read()
417
445
    try:
418
446
        if extra_trees is not None:
420
448
                tree.lock_read()
421
449
        new_tree.lock_read()
422
450
        try:
423
 
            differ = DiffTree.from_trees_options(old_tree, new_tree, to_file,
424
 
                                                 path_encoding,
425
 
                                                 external_diff_options,
426
 
                                                 old_label, new_label, using)
 
451
            differ = format_cls.from_trees_options(old_tree, new_tree, to_file,
 
452
                                                   path_encoding,
 
453
                                                   external_diff_options,
 
454
                                                   old_label, new_label, using)
427
455
            return differ.show_diff(specific_files, extra_trees)
428
456
        finally:
429
457
            new_tree.unlock()
436
464
 
437
465
def _patch_header_date(tree, file_id, path):
438
466
    """Returns a timestamp suitable for use in a patch header."""
439
 
    mtime = tree.get_file_mtime(file_id, path)
 
467
    try:
 
468
        mtime = tree.get_file_mtime(file_id, path)
 
469
    except errors.FileTimestampUnavailable:
 
470
        mtime = 0
440
471
    return timestamp.format_patch_date(mtime)
441
472
 
442
473
 
637
668
        """
638
669
        def _get_text(tree, file_id, path):
639
670
            if file_id is not None:
640
 
                return tree.get_file(file_id, path).readlines()
 
671
                return tree.get_file_lines(file_id, path)
641
672
            else:
642
673
                return []
643
674
        try:
644
675
            from_text = _get_text(self.old_tree, from_file_id, from_path)
645
676
            to_text = _get_text(self.new_tree, to_file_id, to_path)
646
677
            self.text_differ(from_label, from_text, to_label, to_text,
647
 
                             self.to_file)
 
678
                             self.to_file, path_encoding=self.path_encoding)
648
679
        except errors.BinaryFile:
649
680
            self.to_file.write(
650
681
                  ("Binary files %s and %s differ\n" %
651
 
                  (from_label, to_label)).encode(self.path_encoding))
 
682
                  (from_label, to_label)).encode(self.path_encoding,'replace'))
652
683
        return self.CHANGED
653
684
 
654
685
 
663
694
    @classmethod
664
695
    def from_string(klass, command_string, old_tree, new_tree, to_file,
665
696
                    path_encoding='utf-8'):
666
 
        command_template = commands.shlex_split_unicode(command_string)
667
 
        command_template.extend(['%(old_path)s', '%(new_path)s'])
 
697
        command_template = cmdline.split(command_string)
 
698
        if '@' not in command_string:
 
699
            command_template.extend(['@old_path', '@new_path'])
668
700
        return klass(command_template, old_tree, new_tree, to_file,
669
701
                     path_encoding)
670
702
 
671
703
    @classmethod
672
 
    def make_from_diff_tree(klass, command_string):
 
704
    def make_from_diff_tree(klass, command_string, external_diff_options=None):
673
705
        def from_diff_tree(diff_tree):
674
 
            return klass.from_string(command_string, diff_tree.old_tree,
 
706
            full_command_string = [command_string]
 
707
            if external_diff_options is not None:
 
708
                full_command_string += ' ' + external_diff_options
 
709
            return klass.from_string(full_command_string, diff_tree.old_tree,
675
710
                                     diff_tree.new_tree, diff_tree.to_file)
676
711
        return from_diff_tree
677
712
 
678
713
    def _get_command(self, old_path, new_path):
679
714
        my_map = {'old_path': old_path, 'new_path': new_path}
680
 
        return [t % my_map for t in self.command_template]
 
715
        command = [AtTemplate(t).substitute(my_map) for t in
 
716
                   self.command_template]
 
717
        if sys.platform == 'win32': # Popen doesn't accept unicode on win32
 
718
            command_encoded = []
 
719
            for c in command:
 
720
                if isinstance(c, unicode):
 
721
                    command_encoded.append(c.encode('mbcs'))
 
722
                else:
 
723
                    command_encoded.append(c)
 
724
            return command_encoded
 
725
        else:
 
726
            return command
681
727
 
682
728
    def _execute(self, old_path, new_path):
683
729
        command = self._get_command(old_path, new_path)
703
749
                raise
704
750
        return True
705
751
 
706
 
    def _write_file(self, file_id, tree, prefix, relpath):
707
 
        full_path = osutils.pathjoin(self._root, prefix, relpath)
708
 
        if self._try_symlink_root(tree, prefix):
 
752
    @staticmethod
 
753
    def _fenc():
 
754
        """Returns safe encoding for passing file path to diff tool"""
 
755
        if sys.platform == 'win32':
 
756
            return 'mbcs'
 
757
        else:
 
758
            # Don't fallback to 'utf-8' because subprocess may not be able to
 
759
            # handle utf-8 correctly when locale is not utf-8.
 
760
            return sys.getfilesystemencoding() or 'ascii'
 
761
 
 
762
    def _is_safepath(self, path):
 
763
        """Return true if `path` may be able to pass to subprocess."""
 
764
        fenc = self._fenc()
 
765
        try:
 
766
            return path == path.encode(fenc).decode(fenc)
 
767
        except UnicodeError:
 
768
            return False
 
769
 
 
770
    def _safe_filename(self, prefix, relpath):
 
771
        """Replace unsafe character in `relpath` then join `self._root`,
 
772
        `prefix` and `relpath`."""
 
773
        fenc = self._fenc()
 
774
        # encoded_str.replace('?', '_') may break multibyte char.
 
775
        # So we should encode, decode, then replace(u'?', u'_')
 
776
        relpath_tmp = relpath.encode(fenc, 'replace').decode(fenc, 'replace')
 
777
        relpath_tmp = relpath_tmp.replace(u'?', u'_')
 
778
        return osutils.pathjoin(self._root, prefix, relpath_tmp)
 
779
 
 
780
    def _write_file(self, file_id, tree, prefix, relpath, force_temp=False,
 
781
                    allow_write=False):
 
782
        if not force_temp and isinstance(tree, WorkingTree):
 
783
            full_path = tree.abspath(tree.id2path(file_id))
 
784
            if self._is_safepath(full_path):
 
785
                return full_path
 
786
 
 
787
        full_path = self._safe_filename(prefix, relpath)
 
788
        if not force_temp and self._try_symlink_root(tree, prefix):
709
789
            return full_path
710
790
        parent_dir = osutils.dirname(full_path)
711
791
        try:
722
802
                target.close()
723
803
        finally:
724
804
            source.close()
725
 
        osutils.make_readonly(full_path)
726
 
        mtime = tree.get_file_mtime(file_id)
727
 
        os.utime(full_path, (mtime, mtime))
 
805
        try:
 
806
            mtime = tree.get_file_mtime(file_id)
 
807
        except errors.FileTimestampUnavailable:
 
808
            pass
 
809
        else:
 
810
            os.utime(full_path, (mtime, mtime))
 
811
        if not allow_write:
 
812
            osutils.make_readonly(full_path)
728
813
        return full_path
729
814
 
730
 
    def _prepare_files(self, file_id, old_path, new_path):
 
815
    def _prepare_files(self, file_id, old_path, new_path, force_temp=False,
 
816
                       allow_write_new=False):
731
817
        old_disk_path = self._write_file(file_id, self.old_tree, 'old',
732
 
                                         old_path)
 
818
                                         old_path, force_temp)
733
819
        new_disk_path = self._write_file(file_id, self.new_tree, 'new',
734
 
                                         new_path)
 
820
                                         new_path, force_temp,
 
821
                                         allow_write=allow_write_new)
735
822
        return old_disk_path, new_disk_path
736
823
 
737
824
    def finish(self):
745
832
    def diff(self, file_id, old_path, new_path, old_kind, new_kind):
746
833
        if (old_kind, new_kind) != ('file', 'file'):
747
834
            return DiffPath.CANNOT_DIFF
748
 
        self._prepare_files(file_id, old_path, new_path)
749
 
        self._execute(osutils.pathjoin('old', old_path),
750
 
                      osutils.pathjoin('new', new_path))
 
835
        (old_disk_path, new_disk_path) = self._prepare_files(
 
836
                                                file_id, old_path, new_path)
 
837
        self._execute(old_disk_path, new_disk_path)
 
838
 
 
839
    def edit_file(self, file_id):
 
840
        """Use this tool to edit a file.
 
841
 
 
842
        A temporary copy will be edited, and the new contents will be
 
843
        returned.
 
844
 
 
845
        :param file_id: The id of the file to edit.
 
846
        :return: The new contents of the file.
 
847
        """
 
848
        old_path = self.old_tree.id2path(file_id)
 
849
        new_path = self.new_tree.id2path(file_id)
 
850
        old_abs_path, new_abs_path = self._prepare_files(
 
851
                                            file_id, old_path, new_path,
 
852
                                            allow_write_new=True,
 
853
                                            force_temp=True)
 
854
        command = self._get_command(old_abs_path, new_abs_path)
 
855
        subprocess.call(command, cwd=self._root)
 
856
        new_file = open(new_abs_path, 'rb')
 
857
        try:
 
858
            return new_file.read()
 
859
        finally:
 
860
            new_file.close()
751
861
 
752
862
 
753
863
class DiffTree(object):
799
909
        """Factory for producing a DiffTree.
800
910
 
801
911
        Designed to accept options used by show_diff_trees.
 
912
 
802
913
        :param old_tree: The tree to show as old in the comparison
803
914
        :param new_tree: The tree to show as new in the comparison
804
915
        :param to_file: File to write comparisons to
810
921
        :param using: Commandline to use to invoke an external diff tool
811
922
        """
812
923
        if using is not None:
813
 
            extra_factories = [DiffFromTool.make_from_diff_tree(using)]
 
924
            extra_factories = [DiffFromTool.make_from_diff_tree(using, external_diff_options)]
814
925
        else:
815
926
            extra_factories = []
816
927
        if external_diff_options:
817
928
            opts = external_diff_options.split()
818
 
            def diff_file(olab, olines, nlab, nlines, to_file):
 
929
            def diff_file(olab, olines, nlab, nlines, to_file, path_encoding=None):
 
930
                """:param path_encoding: not used but required
 
931
                        to match the signature of internal_diff.
 
932
                """
819
933
                external_diff(olab, olines, nlab, nlines, to_file, opts)
820
934
        else:
821
935
            diff_file = internal_diff
827
941
    def show_diff(self, specific_files, extra_trees=None):
828
942
        """Write tree diff to self.to_file
829
943
 
830
 
        :param sepecific_files: the specific files to compare (recursive)
 
944
        :param specific_files: the specific files to compare (recursive)
831
945
        :param extra_trees: extra trees to use for mapping paths to file_ids
832
946
        """
833
947
        try:
923
1037
            if error_path is None:
924
1038
                error_path = old_path
925
1039
            raise errors.NoDiffFound(error_path)
 
1040
 
 
1041
 
 
1042
format_registry = Registry()
 
1043
format_registry.register('default', DiffTree)