198
176
The yield order (ignoring root) would be::
199
177
a, f, a/b, a/d, a/b/c, a/d/e, f/g
201
return self.inventory.iter_entries_by_dir(
202
specific_file_ids=specific_file_ids)
179
:param yield_parents: If True, yield the parents from the root leading
180
down to specific_file_ids that have been requested. This has no
181
impact if specific_file_ids is None.
183
raise NotImplementedError(self.iter_entries_by_dir)
185
def list_files(self, include_root=False, from_dir=None, recursive=True):
186
"""List all files in this tree.
188
:param include_root: Whether to include the entry for the tree root
189
:param from_dir: Directory under which to list files
190
:param recursive: Whether to list files recursively
191
:return: iterator over tuples of (path, versioned, kind, file_id,
194
raise NotImplementedError(self.list_files)
204
196
def iter_references(self):
205
for path, entry in self.iter_entries_by_dir():
206
if entry.kind == 'tree-reference':
207
yield path, entry.file_id
197
if self.supports_tree_reference():
198
for path, entry in self.iter_entries_by_dir():
199
if entry.kind == 'tree-reference':
200
yield path, entry.file_id
209
202
def kind(self, file_id):
210
203
raise NotImplementedError("Tree subclass %s must implement kind"
344
370
raise NotImplementedError(self.get_symlink_target)
346
def get_canonical_inventory_paths(self, paths):
347
"""Like get_canonical_inventory_path() but works on multiple items.
349
:param paths: A sequence of paths relative to the root of the tree.
350
:return: A list of paths, with each item the corresponding input path
351
adjusted to account for existing elements that match case
354
return list(self._yield_canonical_inventory_paths(paths))
356
def get_canonical_inventory_path(self, path):
357
"""Returns the first inventory item that case-insensitively matches path.
359
If a path matches exactly, it is returned. If no path matches exactly
360
but more than one path matches case-insensitively, it is implementation
361
defined which is returned.
363
If no path matches case-insensitively, the input path is returned, but
364
with as many path entries that do exist changed to their canonical
367
If you need to resolve many names from the same tree, you should
368
use get_canonical_inventory_paths() to avoid O(N) behaviour.
370
:param path: A paths relative to the root of the tree.
371
:return: The input path adjusted to account for existing elements
372
that match case insensitively.
374
return self._yield_canonical_inventory_paths([path]).next()
376
def _yield_canonical_inventory_paths(self, paths):
378
# First, if the path as specified exists exactly, just use it.
379
if self.path2id(path) is not None:
383
cur_id = self.get_root_id()
385
bit_iter = iter(path.split("/"))
388
for child in self.iter_children(cur_id):
390
child_base = os.path.basename(self.id2path(child))
391
if child_base.lower() == lelt:
393
cur_path = osutils.pathjoin(cur_path, child_base)
396
# before a change is committed we can see this error...
399
# got to the end of this directory and no entries matched.
400
# Return what matched so far, plus the rest as specified.
401
cur_path = osutils.pathjoin(cur_path, elt, *list(bit_iter))
406
373
def get_root_id(self):
407
374
"""Return the file_id for the root of this tree."""
466
433
except errors.NoSuchRevisionInTree:
467
434
yield self.repository.revision_tree(revision_id)
470
def _file_revision(revision_tree, file_id):
471
"""Determine the revision associated with a file in a given tree."""
472
revision_tree.lock_read()
474
return revision_tree.inventory[file_id].revision
476
revision_tree.unlock()
478
436
def _get_file_revision(self, file_id, vf, tree_revision):
479
437
"""Ensure that file_id, tree_revision is in vf to plan the merge."""
481
439
if getattr(self, '_repository', None) is None:
482
440
last_revision = tree_revision
483
parent_keys = [(file_id, self._file_revision(t, file_id)) for t in
441
parent_keys = [(file_id, t.get_file_revision(file_id)) for t in
484
442
self._iter_parent_trees()]
485
443
vf.add_lines((file_id, last_revision), parent_keys,
486
self.get_file(file_id).readlines())
444
self.get_file_lines(file_id))
487
445
repo = self.branch.repository
488
446
base_vf = repo.texts
490
last_revision = self._file_revision(self, file_id)
448
last_revision = self.get_file_revision(file_id)
491
449
base_vf = self._repository.texts
492
450
if base_vf not in vf.fallback_versionedfiles:
493
451
vf.fallback_versionedfiles.append(base_vf)
494
452
return last_revision
496
inventory = property(_get_inventory,
497
doc="Inventory of this Tree")
499
454
def _check_retrieved(self, ie, f):
500
455
if not __debug__:
502
fp = fingerprint_file(f)
457
fp = osutils.fingerprint_file(f)
505
460
if ie.text_size is not None:
506
461
if ie.text_size != fp['size']:
507
raise BzrError("mismatched size for file %r in %r" % (ie.file_id, self._store),
462
raise errors.BzrError(
463
"mismatched size for file %r in %r" %
464
(ie.file_id, self._store),
508
465
["inventory expects %d bytes" % ie.text_size,
509
466
"file is actually %d bytes" % fp['size'],
510
467
"store is probably damaged/corrupt"])
512
469
if ie.text_sha1 != fp['sha1']:
513
raise BzrError("wrong SHA-1 for file %r in %r" % (ie.file_id, self._store),
470
raise errors.BzrError("wrong SHA-1 for file %r in %r" %
471
(ie.file_id, self._store),
514
472
["inventory expects %s" % ie.text_sha1,
515
473
"file is actually %s" % fp['sha1'],
516
474
"store is probably damaged/corrupt"])
519
476
def path2id(self, path):
520
477
"""Return the id for path in this tree."""
521
return self._inventory.path2id(path)
478
raise NotImplementedError(self.path2id)
523
480
def paths2ids(self, paths, trees=[], require_versioned=True):
524
481
"""Return all the ids that can be reached by walking from paths.
676
635
for path in path_names:
677
636
yield searcher.get_items(path)
680
638
def _get_rules_searcher(self, default_searcher):
681
639
"""Get the RulesSearcher for this tree given the default one."""
682
640
searcher = default_searcher
644
class InventoryTree(Tree):
645
"""A tree that relies on an inventory for its metadata.
647
Trees contain an `Inventory` object, and also know how to retrieve
648
file texts mentioned in the inventory, either from a working
649
directory or from a store.
651
It is possible for trees to contain files that are not described
652
in their inventory or vice versa; for this use `filenames()`.
654
Subclasses should set the _inventory attribute, which is considered
655
private to external API users.
658
def get_canonical_inventory_paths(self, paths):
659
"""Like get_canonical_inventory_path() but works on multiple items.
661
:param paths: A sequence of paths relative to the root of the tree.
662
:return: A list of paths, with each item the corresponding input path
663
adjusted to account for existing elements that match case
666
return list(self._yield_canonical_inventory_paths(paths))
668
def get_canonical_inventory_path(self, path):
669
"""Returns the first inventory item that case-insensitively matches path.
671
If a path matches exactly, it is returned. If no path matches exactly
672
but more than one path matches case-insensitively, it is implementation
673
defined which is returned.
675
If no path matches case-insensitively, the input path is returned, but
676
with as many path entries that do exist changed to their canonical
679
If you need to resolve many names from the same tree, you should
680
use get_canonical_inventory_paths() to avoid O(N) behaviour.
682
:param path: A paths relative to the root of the tree.
683
:return: The input path adjusted to account for existing elements
684
that match case insensitively.
686
return self._yield_canonical_inventory_paths([path]).next()
688
def _yield_canonical_inventory_paths(self, paths):
690
# First, if the path as specified exists exactly, just use it.
691
if self.path2id(path) is not None:
695
cur_id = self.get_root_id()
697
bit_iter = iter(path.split("/"))
701
for child in self.iter_children(cur_id):
703
# XXX: it seem like if the child is known to be in the
704
# tree, we shouldn't need to go from its id back to
705
# its path -- mbp 2010-02-11
707
# XXX: it seems like we could be more efficient
708
# by just directly looking up the original name and
709
# only then searching all children; also by not
710
# chopping paths so much. -- mbp 2010-02-11
711
child_base = os.path.basename(self.id2path(child))
712
if (child_base == elt):
713
# if we found an exact match, we can stop now; if
714
# we found an approximate match we need to keep
715
# searching because there might be an exact match
718
new_path = osutils.pathjoin(cur_path, child_base)
720
elif child_base.lower() == lelt:
722
new_path = osutils.pathjoin(cur_path, child_base)
723
except errors.NoSuchId:
724
# before a change is committed we can see this error...
729
# got to the end of this directory and no entries matched.
730
# Return what matched so far, plus the rest as specified.
731
cur_path = osutils.pathjoin(cur_path, elt, *list(bit_iter))
736
def _get_inventory(self):
737
return self._inventory
739
inventory = property(_get_inventory,
740
doc="Inventory of this Tree")
743
def path2id(self, path):
744
"""Return the id for path in this tree."""
745
return self._inventory.path2id(path)
747
def id2path(self, file_id):
748
"""Return the path for a file id.
752
return self.inventory.id2path(file_id)
754
def has_id(self, file_id):
755
return self.inventory.has_id(file_id)
757
def has_or_had_id(self, file_id):
758
return self.inventory.has_id(file_id)
761
return iter(self.inventory)
763
def filter_unversioned_files(self, paths):
764
"""Filter out paths that are versioned.
766
:return: set of paths.
768
# NB: we specifically *don't* call self.has_filename, because for
769
# WorkingTrees that can indicate files that exist on disk but that
771
pred = self.inventory.has_filename
772
return set((p for p in paths if not pred(p)))
775
def iter_entries_by_dir(self, specific_file_ids=None, yield_parents=False):
776
"""Walk the tree in 'by_dir' order.
778
This will yield each entry in the tree as a (path, entry) tuple.
779
The order that they are yielded is:
781
See Tree.iter_entries_by_dir for details.
783
:param yield_parents: If True, yield the parents from the root leading
784
down to specific_file_ids that have been requested. This has no
785
impact if specific_file_ids is None.
787
return self.inventory.iter_entries_by_dir(
788
specific_file_ids=specific_file_ids, yield_parents=yield_parents)
790
def get_file_by_path(self, path):
791
return self.get_file(self._inventory.path2id(path), path)
686
794
######################################################################
831
928
will pass through to InterTree as appropriate.
931
# Formats that will be used to test this InterTree. If both are
932
# None, this InterTree will not be tested (e.g. because a complex
934
_matching_from_tree_format = None
935
_matching_to_tree_format = None
939
def _changes_from_entries(self, source_entry, target_entry,
940
source_path=None, target_path=None):
941
"""Generate a iter_changes tuple between source_entry and target_entry.
943
:param source_entry: An inventory entry from self.source, or None.
944
:param target_entry: An inventory entry from self.target, or None.
945
:param source_path: The path of source_entry, if known. If not known
946
it will be looked up.
947
:param target_path: The path of target_entry, if known. If not known
948
it will be looked up.
949
:return: A tuple, item 0 of which is an iter_changes result tuple, and
950
item 1 is True if there are any changes in the result tuple.
952
if source_entry is None:
953
if target_entry is None:
955
file_id = target_entry.file_id
957
file_id = source_entry.file_id
958
if source_entry is not None:
959
source_versioned = True
960
source_name = source_entry.name
961
source_parent = source_entry.parent_id
962
if source_path is None:
963
source_path = self.source.id2path(file_id)
964
source_kind, source_executable, source_stat = \
965
self.source._comparison_data(source_entry, source_path)
967
source_versioned = False
971
source_executable = None
972
if target_entry is not None:
973
target_versioned = True
974
target_name = target_entry.name
975
target_parent = target_entry.parent_id
976
if target_path is None:
977
target_path = self.target.id2path(file_id)
978
target_kind, target_executable, target_stat = \
979
self.target._comparison_data(target_entry, target_path)
981
target_versioned = False
985
target_executable = None
986
versioned = (source_versioned, target_versioned)
987
kind = (source_kind, target_kind)
988
changed_content = False
989
if source_kind != target_kind:
990
changed_content = True
991
elif source_kind == 'file':
992
if (self.source.get_file_sha1(file_id, source_path, source_stat) !=
993
self.target.get_file_sha1(file_id, target_path, target_stat)):
994
changed_content = True
995
elif source_kind == 'symlink':
996
if (self.source.get_symlink_target(file_id) !=
997
self.target.get_symlink_target(file_id)):
998
changed_content = True
999
# XXX: Yes, the indentation below is wrong. But fixing it broke
1000
# test_merge.TestMergerEntriesLCAOnDisk.
1001
# test_nested_tree_subtree_renamed_and_modified. We'll wait for
1002
# the fix from bzr.dev -- vila 2009026
1003
elif source_kind == 'tree-reference':
1004
if (self.source.get_reference_revision(file_id, source_path)
1005
!= self.target.get_reference_revision(file_id, target_path)):
1006
changed_content = True
1007
parent = (source_parent, target_parent)
1008
name = (source_name, target_name)
1009
executable = (source_executable, target_executable)
1010
if (changed_content is not False or versioned[0] != versioned[1]
1011
or parent[0] != parent[1] or name[0] != name[1] or
1012
executable[0] != executable[1]):
1016
return (file_id, (source_path, target_path), changed_content,
1017
versioned, parent, name, kind, executable), changes
836
1019
@needs_read_lock
837
1020
def compare(self, want_unchanged=False, specific_files=None,
838
1021
extra_trees=None, require_versioned=False, include_root=False,
908
1095
lookup_trees = [self.source]
910
1097
lookup_trees.extend(extra_trees)
1098
# The ids of items we need to examine to insure delta consistency.
1099
precise_file_ids = set()
1100
changed_file_ids = []
911
1101
if specific_files == []:
912
1102
specific_file_ids = []
914
1104
specific_file_ids = self.target.paths2ids(specific_files,
915
1105
lookup_trees, require_versioned=require_versioned)
1106
if specific_files is not None:
1107
# reparented or added entries must have their parents included
1108
# so that valid deltas can be created. The seen_parents set
1109
# tracks the parents that we need to have.
1110
# The seen_dirs set tracks directory entries we've yielded.
1111
# After outputting version object in to_entries we set difference
1112
# the two seen sets and start checking parents.
1113
seen_parents = set()
916
1115
if want_unversioned:
917
1116
all_unversioned = sorted([(p.split('/'), p) for p in
918
1117
self.target.extras()
919
1118
if specific_files is None or
920
1119
osutils.is_inside_any(specific_files, p)])
921
all_unversioned = deque(all_unversioned)
1120
all_unversioned = collections.deque(all_unversioned)
923
all_unversioned = deque()
1122
all_unversioned = collections.deque()
925
1124
from_entries_by_dir = list(self.source.iter_entries_by_dir(
926
1125
specific_file_ids=specific_file_ids))
932
1131
# the unversioned path lookup only occurs on real trees - where there
933
1132
# can be extras. So the fake_entry is solely used to look up
934
1133
# executable it values when execute is not supported.
935
fake_entry = InventoryFile('unused', 'unused', 'unused')
936
for to_path, to_entry in to_entries_by_dir:
937
while all_unversioned and all_unversioned[0][0] < to_path.split('/'):
1134
fake_entry = inventory.InventoryFile('unused', 'unused', 'unused')
1135
for target_path, target_entry in to_entries_by_dir:
1136
while (all_unversioned and
1137
all_unversioned[0][0] < target_path.split('/')):
938
1138
unversioned_path = all_unversioned.popleft()
939
to_kind, to_executable, to_stat = \
1139
target_kind, target_executable, target_stat = \
940
1140
self.target._comparison_data(fake_entry, unversioned_path[1])
941
1141
yield (None, (None, unversioned_path[1]), True, (False, False),
943
1143
(None, unversioned_path[0][-1]),
945
(None, to_executable))
946
file_id = to_entry.file_id
947
to_paths[file_id] = to_path
1144
(None, target_kind),
1145
(None, target_executable))
1146
source_path, source_entry = from_data.get(target_entry.file_id,
1148
result, changes = self._changes_from_entries(source_entry,
1149
target_entry, source_path=source_path, target_path=target_path)
1150
to_paths[result[0]] = result[1][1]
948
1151
entry_count += 1
949
changed_content = False
950
from_path, from_entry = from_data.get(file_id, (None, None))
951
from_versioned = (from_entry is not None)
952
if from_entry is not None:
953
from_versioned = True
954
from_name = from_entry.name
955
from_parent = from_entry.parent_id
956
from_kind, from_executable, from_stat = \
957
self.source._comparison_data(from_entry, from_path)
958
1153
entry_count += 1
960
from_versioned = False
964
from_executable = None
965
versioned = (from_versioned, True)
966
to_kind, to_executable, to_stat = \
967
self.target._comparison_data(to_entry, to_path)
968
kind = (from_kind, to_kind)
969
if kind[0] != kind[1]:
970
changed_content = True
971
elif from_kind == 'file':
972
if (self.source.get_file_sha1(file_id, from_path, from_stat) !=
973
self.target.get_file_sha1(file_id, to_path, to_stat)):
974
changed_content = True
975
elif from_kind == 'symlink':
976
if (self.source.get_symlink_target(file_id) !=
977
self.target.get_symlink_target(file_id)):
978
changed_content = True
979
# XXX: Yes, the indentation below is wrong. But fixing it broke
980
# test_merge.TestMergerEntriesLCAOnDisk.
981
# test_nested_tree_subtree_renamed_and_modified. We'll wait for
982
# the fix from bzr.dev -- vila 2009026
983
elif from_kind == 'tree-reference':
984
if (self.source.get_reference_revision(file_id, from_path)
985
!= self.target.get_reference_revision(file_id, to_path)):
986
changed_content = True
987
parent = (from_parent, to_entry.parent_id)
988
name = (from_name, to_entry.name)
989
executable = (from_executable, to_executable)
990
1154
if pb is not None:
991
1155
pb.update('comparing files', entry_count, num_entries)
992
if (changed_content is not False or versioned[0] != versioned[1]
993
or parent[0] != parent[1] or name[0] != name[1] or
994
executable[0] != executable[1] or include_unchanged):
995
yield (file_id, (from_path, to_path), changed_content,
996
versioned, parent, name, kind, executable)
1156
if changes or include_unchanged:
1157
if specific_file_ids is not None:
1158
new_parent_id = result[4][1]
1159
precise_file_ids.add(new_parent_id)
1160
changed_file_ids.append(result[0])
1162
# Ensure correct behaviour for reparented/added specific files.
1163
if specific_files is not None:
1164
# Record output dirs
1165
if result[6][1] == 'directory':
1166
seen_dirs.add(result[0])
1167
# Record parents of reparented/added entries.
1168
versioned = result[3]
1170
if not versioned[0] or parents[0] != parents[1]:
1171
seen_parents.add(parents[1])
998
1172
while all_unversioned:
999
1173
# yield any trailing unversioned paths
1000
1174
unversioned_path = all_unversioned.popleft()
1005
1179
(None, unversioned_path[0][-1]),
1006
1180
(None, to_kind),
1007
1181
(None, to_executable))
1009
def get_to_path(to_entry):
1010
if to_entry.parent_id is None:
1011
to_path = '' # the root
1013
if to_entry.parent_id not in to_paths:
1015
return get_to_path(self.target.inventory[to_entry.parent_id])
1016
to_path = osutils.pathjoin(to_paths[to_entry.parent_id],
1018
to_paths[to_entry.file_id] = to_path
1182
# Yield all remaining source paths
1021
1183
for path, from_entry in from_entries_by_dir:
1022
1184
file_id = from_entry.file_id
1023
1185
if file_id in to_paths:
1024
1186
# already returned
1026
if not file_id in self.target.all_file_ids():
1188
if not self.target.has_id(file_id):
1027
1189
# common case - paths we have not emitted are not present in
1031
to_path = get_to_path(self.target.inventory[file_id])
1193
to_path = self.target.id2path(file_id)
1032
1194
entry_count += 1
1033
1195
if pb is not None:
1034
1196
pb.update('comparing files', entry_count, num_entries)
1041
1203
executable = (from_executable, None)
1042
1204
changed_content = from_kind is not None
1043
1205
# the parent's path is necessarily known at this point.
1206
changed_file_ids.append(file_id)
1044
1207
yield(file_id, (path, to_path), changed_content, versioned, parent,
1045
1208
name, kind, executable)
1209
changed_file_ids = set(changed_file_ids)
1210
if specific_file_ids is not None:
1211
for result in self._handle_precise_ids(precise_file_ids,
1215
def _get_entry(self, tree, file_id):
1216
"""Get an inventory entry from a tree, with missing entries as None.
1218
If the tree raises NotImplementedError on accessing .inventory, then
1219
this is worked around using iter_entries_by_dir on just the file id
1222
:param tree: The tree to lookup the entry in.
1223
:param file_id: The file_id to lookup.
1226
inventory = tree.inventory
1227
except NotImplementedError:
1228
# No inventory available.
1230
iterator = tree.iter_entries_by_dir(specific_file_ids=[file_id])
1231
return iterator.next()[1]
1232
except StopIteration:
1236
return inventory[file_id]
1237
except errors.NoSuchId:
1240
def _handle_precise_ids(self, precise_file_ids, changed_file_ids,
1241
discarded_changes=None):
1242
"""Fill out a partial iter_changes to be consistent.
1244
:param precise_file_ids: The file ids of parents that were seen during
1246
:param changed_file_ids: The file ids of already emitted items.
1247
:param discarded_changes: An optional dict of precalculated
1248
iter_changes items which the partial iter_changes had not output
1250
:return: A generator of iter_changes items to output.
1252
# process parents of things that had changed under the users
1253
# requested paths to prevent incorrect paths or parent ids which
1254
# aren't in the tree.
1255
while precise_file_ids:
1256
precise_file_ids.discard(None)
1257
# Don't emit file_ids twice
1258
precise_file_ids.difference_update(changed_file_ids)
1259
if not precise_file_ids:
1261
# If the there was something at a given output path in source, we
1262
# have to include the entry from source in the delta, or we would
1263
# be putting this entry into a used path.
1265
for parent_id in precise_file_ids:
1267
paths.append(self.target.id2path(parent_id))
1268
except errors.NoSuchId:
1269
# This id has been dragged in from the source by delta
1270
# expansion and isn't present in target at all: we don't
1271
# need to check for path collisions on it.
1274
old_id = self.source.path2id(path)
1275
precise_file_ids.add(old_id)
1276
precise_file_ids.discard(None)
1277
current_ids = precise_file_ids
1278
precise_file_ids = set()
1279
# We have to emit all of precise_file_ids that have been altered.
1280
# We may have to output the children of some of those ids if any
1281
# directories have stopped being directories.
1282
for file_id in current_ids:
1284
if discarded_changes:
1285
result = discarded_changes.get(file_id)
1290
old_entry = self._get_entry(self.source, file_id)
1291
new_entry = self._get_entry(self.target, file_id)
1292
result, changes = self._changes_from_entries(
1293
old_entry, new_entry)
1296
# Get this parents parent to examine.
1297
new_parent_id = result[4][1]
1298
precise_file_ids.add(new_parent_id)
1300
if (result[6][0] == 'directory' and
1301
result[6][1] != 'directory'):
1302
# This stopped being a directory, the old children have
1304
if old_entry is None:
1305
# Reusing a discarded change.
1306
old_entry = self._get_entry(self.source, file_id)
1307
for child in old_entry.children.values():
1308
precise_file_ids.add(child.file_id)
1309
changed_file_ids.add(result[0])
1048
1313
class MultiWalker(object):
added
removed
bzrlib/tree.py
