~bzr-pqm/bzr/bzr.dev

« back to all changes in this revision

Viewing changes to bzrlib/diff.py

  • Committer: Canonical.com Patch Queue Manager
  • Date: 2009-10-13 06:08:53 UTC
  • mfrom: (4737.1.1 merge-2.0-into-devel)
  • Revision ID: pqm@pqm.ubuntu.com-20091013060853-erk2aaj80fnkrv25
(andrew) Merge lp:bzr/2.0 into lp:bzr, including fixes for #322807,
        #389413, #402623 and documentation improvements.

Show diffs side-by-side

added added

removed removed

Lines of Context:
1
 
# Copyright (C) 2005-2011 Canonical Ltd.
 
1
# Copyright (C) 2004, 2005, 2006 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
 
 
19
17
import difflib
20
18
import os
21
19
import re
22
 
import string
 
20
import shutil
23
21
import sys
24
22
 
25
23
from bzrlib.lazy_import import lazy_import
27
25
import errno
28
26
import subprocess
29
27
import tempfile
 
28
import time
30
29
 
31
30
from bzrlib import (
32
 
    cleanup,
33
 
    cmdline,
34
 
    controldir,
 
31
    branch as _mod_branch,
 
32
    bzrdir,
 
33
    commands,
35
34
    errors,
36
35
    osutils,
37
36
    patiencediff,
39
38
    timestamp,
40
39
    views,
41
40
    )
42
 
 
43
 
from bzrlib.workingtree import WorkingTree
44
 
from bzrlib.i18n import gettext
45
41
""")
46
42
 
47
 
from bzrlib.registry import (
48
 
    Registry,
 
43
from bzrlib.symbol_versioning import (
 
44
    deprecated_function,
49
45
    )
50
46
from bzrlib.trace import mutter, note, warning
51
47
 
52
48
 
53
 
class AtTemplate(string.Template):
54
 
    """Templating class that uses @ instead of $."""
55
 
 
56
 
    delimiter = '@'
57
 
 
58
 
 
59
49
# TODO: Rather than building a changeset object, we should probably
60
50
# invoke callbacks on an object.  That object can either accumulate a
61
51
# list, write them out directly, etc etc.
96
86
    if sequence_matcher is None:
97
87
        sequence_matcher = patiencediff.PatienceSequenceMatcher
98
88
    ud = patiencediff.unified_diff(oldlines, newlines,
99
 
                      fromfile=old_filename.encode(path_encoding, 'replace'),
100
 
                      tofile=new_filename.encode(path_encoding, 'replace'),
 
89
                      fromfile=old_filename.encode(path_encoding),
 
90
                      tofile=new_filename.encode(path_encoding),
101
91
                      sequencematcher=sequence_matcher)
102
92
 
103
93
    ud = list(ud)
287
277
                        new_abspath, e)
288
278
 
289
279
 
290
 
def get_trees_and_branches_to_diff_locked(
291
 
    path_list, revision_specs, old_url, new_url, add_cleanup, apply_view=True):
 
280
def get_trees_and_branches_to_diff(path_list, revision_specs, old_url, new_url,
 
281
                                   apply_view=True):
292
282
    """Get the trees and specific files to diff given a list of paths.
293
283
 
294
284
    This method works out the trees to be diff'ed and the files of
305
295
    :param new_url:
306
296
        The url of the new branch or tree. If None, the tree to use is
307
297
        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.
311
298
    :param apply_view:
312
299
        if True and a view is set, apply the view or check that the paths
313
300
        are within it
314
301
    :returns:
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.
 
302
        a tuple of (old_tree, new_tree, specific_files, extra_trees) where
 
303
        extra_trees is a sequence of additional trees to search in for
 
304
        file-ids.
320
305
    """
321
306
    # Get the old and new revision specs
322
307
    old_revision_spec = None
345
330
        default_location = path_list[0]
346
331
        other_paths = path_list[1:]
347
332
 
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
 
 
356
333
    # Get the old location
357
334
    specific_files = []
358
335
    if old_url is None:
359
336
        old_url = default_location
360
337
    working_tree, branch, relpath = \
361
 
        controldir.ControlDir.open_containing_tree_or_branch(old_url)
362
 
    lock_tree_or_branch(working_tree, branch)
 
338
        bzrdir.BzrDir.open_containing_tree_or_branch(old_url)
363
339
    if consider_relpath and relpath != '':
364
340
        if working_tree is not None and apply_view:
365
341
            views.check_path_in_view(working_tree, relpath)
372
348
        new_url = default_location
373
349
    if new_url != old_url:
374
350
        working_tree, branch, relpath = \
375
 
            controldir.ControlDir.open_containing_tree_or_branch(new_url)
376
 
        lock_tree_or_branch(working_tree, branch)
 
351
            bzrdir.BzrDir.open_containing_tree_or_branch(new_url)
377
352
        if consider_relpath and relpath != '':
378
353
            if working_tree is not None and apply_view:
379
354
                views.check_path_in_view(working_tree, relpath)
384
359
 
385
360
    # Get the specific files (all files is None, no files is [])
386
361
    if make_paths_wt_relative and working_tree is not None:
387
 
        other_paths = working_tree.safe_relpath_files(
388
 
            other_paths,
 
362
        try:
 
363
            from bzrlib.builtins import safe_relpath_files
 
364
            other_paths = safe_relpath_files(working_tree, other_paths,
389
365
            apply_view=apply_view)
 
366
        except errors.FileInWrongBranch:
 
367
            raise errors.BzrCommandError("Files are in different branches")
390
368
    specific_files.extend(other_paths)
391
369
    if len(specific_files) == 0:
392
370
        specific_files = None
396
374
            if view_files:
397
375
                specific_files = view_files
398
376
                view_str = views.view_display_str(view_files)
399
 
                note(gettext("*** Ignoring files outside view. View is %s") % view_str)
 
377
                note("*** Ignoring files outside view. View is %s" % view_str)
400
378
 
401
379
    # Get extra trees that ought to be searched for file-ids
402
380
    extra_trees = None
403
381
    if working_tree is not None and working_tree not in (old_tree, new_tree):
404
382
        extra_trees = (working_tree,)
405
 
    return (old_tree, new_tree, old_branch, new_branch,
406
 
            specific_files, extra_trees)
407
 
 
 
383
    return old_tree, new_tree, old_branch, new_branch, specific_files, extra_trees
408
384
 
409
385
def _get_tree_to_diff(spec, tree=None, branch=None, basis_is_default=True):
410
386
    if branch is None and tree is not None:
425
401
                    old_label='a/', new_label='b/',
426
402
                    extra_trees=None,
427
403
                    path_encoding='utf8',
428
 
                    using=None,
429
 
                    format_cls=None):
 
404
                    using=None):
430
405
    """Show in text form the changes from one tree to another.
431
406
 
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)
 
407
    to_file
 
408
        The output stream.
 
409
 
 
410
    specific_files
 
411
        Include only changes to these files - None for all changes.
 
412
 
 
413
    external_diff_options
 
414
        If set, use an external GNU diff and pass these options.
 
415
 
 
416
    extra_trees
 
417
        If set, more Trees to use for looking up file ids
 
418
 
 
419
    path_encoding
 
420
        If set, the path will be encoded as specified, otherwise is supposed
 
421
        to be utf8
441
422
    """
442
 
    if format_cls is None:
443
 
        format_cls = DiffTree
444
423
    old_tree.lock_read()
445
424
    try:
446
425
        if extra_trees is not None:
448
427
                tree.lock_read()
449
428
        new_tree.lock_read()
450
429
        try:
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)
 
430
            differ = DiffTree.from_trees_options(old_tree, new_tree, to_file,
 
431
                                                 path_encoding,
 
432
                                                 external_diff_options,
 
433
                                                 old_label, new_label, using)
455
434
            return differ.show_diff(specific_files, extra_trees)
456
435
        finally:
457
436
            new_tree.unlock()
464
443
 
465
444
def _patch_header_date(tree, file_id, path):
466
445
    """Returns a timestamp suitable for use in a patch header."""
467
 
    try:
468
 
        mtime = tree.get_file_mtime(file_id, path)
469
 
    except errors.FileTimestampUnavailable:
470
 
        mtime = 0
 
446
    mtime = tree.get_file_mtime(file_id, path)
471
447
    return timestamp.format_patch_date(mtime)
472
448
 
473
449
 
668
644
        """
669
645
        def _get_text(tree, file_id, path):
670
646
            if file_id is not None:
671
 
                return tree.get_file_lines(file_id, path)
 
647
                return tree.get_file(file_id, path).readlines()
672
648
            else:
673
649
                return []
674
650
        try:
675
651
            from_text = _get_text(self.old_tree, from_file_id, from_path)
676
652
            to_text = _get_text(self.new_tree, to_file_id, to_path)
677
653
            self.text_differ(from_label, from_text, to_label, to_text,
678
 
                             self.to_file, path_encoding=self.path_encoding)
 
654
                             self.to_file)
679
655
        except errors.BinaryFile:
680
656
            self.to_file.write(
681
657
                  ("Binary files %s and %s differ\n" %
682
 
                  (from_label, to_label)).encode(self.path_encoding,'replace'))
 
658
                  (from_label, to_label)).encode(self.path_encoding))
683
659
        return self.CHANGED
684
660
 
685
661
 
694
670
    @classmethod
695
671
    def from_string(klass, command_string, old_tree, new_tree, to_file,
696
672
                    path_encoding='utf-8'):
697
 
        command_template = cmdline.split(command_string)
698
 
        if '@' not in command_string:
699
 
            command_template.extend(['@old_path', '@new_path'])
 
673
        command_template = commands.shlex_split_unicode(command_string)
 
674
        command_template.extend(['%(old_path)s', '%(new_path)s'])
700
675
        return klass(command_template, old_tree, new_tree, to_file,
701
676
                     path_encoding)
702
677
 
703
678
    @classmethod
704
 
    def make_from_diff_tree(klass, command_string, external_diff_options=None):
 
679
    def make_from_diff_tree(klass, command_string):
705
680
        def from_diff_tree(diff_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,
 
681
            return klass.from_string(command_string, diff_tree.old_tree,
710
682
                                     diff_tree.new_tree, diff_tree.to_file)
711
683
        return from_diff_tree
712
684
 
713
685
    def _get_command(self, old_path, new_path):
714
686
        my_map = {'old_path': old_path, 'new_path': new_path}
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
 
687
        return [t % my_map for t in self.command_template]
727
688
 
728
689
    def _execute(self, old_path, new_path):
729
690
        command = self._get_command(old_path, new_path)
749
710
                raise
750
711
        return True
751
712
 
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):
 
713
    def _write_file(self, file_id, tree, prefix, relpath):
 
714
        full_path = osutils.pathjoin(self._root, prefix, relpath)
 
715
        if self._try_symlink_root(tree, prefix):
789
716
            return full_path
790
717
        parent_dir = osutils.dirname(full_path)
791
718
        try:
802
729
                target.close()
803
730
        finally:
804
731
            source.close()
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)
 
732
        osutils.make_readonly(full_path)
 
733
        mtime = tree.get_file_mtime(file_id)
 
734
        os.utime(full_path, (mtime, mtime))
813
735
        return full_path
814
736
 
815
 
    def _prepare_files(self, file_id, old_path, new_path, force_temp=False,
816
 
                       allow_write_new=False):
 
737
    def _prepare_files(self, file_id, old_path, new_path):
817
738
        old_disk_path = self._write_file(file_id, self.old_tree, 'old',
818
 
                                         old_path, force_temp)
 
739
                                         old_path)
819
740
        new_disk_path = self._write_file(file_id, self.new_tree, 'new',
820
 
                                         new_path, force_temp,
821
 
                                         allow_write=allow_write_new)
 
741
                                         new_path)
822
742
        return old_disk_path, new_disk_path
823
743
 
824
744
    def finish(self):
832
752
    def diff(self, file_id, old_path, new_path, old_kind, new_kind):
833
753
        if (old_kind, new_kind) != ('file', 'file'):
834
754
            return DiffPath.CANNOT_DIFF
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()
 
755
        self._prepare_files(file_id, old_path, new_path)
 
756
        self._execute(osutils.pathjoin('old', old_path),
 
757
                      osutils.pathjoin('new', new_path))
861
758
 
862
759
 
863
760
class DiffTree(object):
909
806
        """Factory for producing a DiffTree.
910
807
 
911
808
        Designed to accept options used by show_diff_trees.
912
 
 
913
809
        :param old_tree: The tree to show as old in the comparison
914
810
        :param new_tree: The tree to show as new in the comparison
915
811
        :param to_file: File to write comparisons to
921
817
        :param using: Commandline to use to invoke an external diff tool
922
818
        """
923
819
        if using is not None:
924
 
            extra_factories = [DiffFromTool.make_from_diff_tree(using, external_diff_options)]
 
820
            extra_factories = [DiffFromTool.make_from_diff_tree(using)]
925
821
        else:
926
822
            extra_factories = []
927
823
        if external_diff_options:
928
824
            opts = external_diff_options.split()
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
 
                """
 
825
            def diff_file(olab, olines, nlab, nlines, to_file):
933
826
                external_diff(olab, olines, nlab, nlines, to_file, opts)
934
827
        else:
935
828
            diff_file = internal_diff
941
834
    def show_diff(self, specific_files, extra_trees=None):
942
835
        """Write tree diff to self.to_file
943
836
 
944
 
        :param specific_files: the specific files to compare (recursive)
 
837
        :param sepecific_files: the specific files to compare (recursive)
945
838
        :param extra_trees: extra trees to use for mapping paths to file_ids
946
839
        """
947
840
        try:
1037
930
            if error_path is None:
1038
931
                error_path = old_path
1039
932
            raise errors.NoDiffFound(error_path)
1040
 
 
1041
 
 
1042
 
format_registry = Registry()
1043
 
format_registry.register('default', DiffTree)