17
17
"""Tree classes, representing directory at point in time.
20
from __future__ import absolute_import
21
from collections import deque
24
from .lazy_import import lazy_import
25
lazy_import(globals(), """
25
29
conflicts as _mod_conflicts,
30
33
revision as _mod_revision,
33
from bzrlib.decorators import needs_read_lock
34
from bzrlib.errors import BzrError, NoSuchId
35
from bzrlib import errors
36
from bzrlib.inventory import InventoryFile
37
from bzrlib.inter import InterObject
38
from bzrlib.osutils import fingerprint_file
39
from bzrlib.symbol_versioning import deprecated_function, deprecated_in
40
from bzrlib.trace import note
37
from breezy.bzr import (
40
from breezy.i18n import gettext
48
from .inter import InterObject
54
class FileTimestampUnavailable(errors.BzrError):
56
_fmt = "The filestamp for %(path)s is not available."
60
def __init__(self, path):
43
64
class Tree(object):
50
71
* `RevisionTree` is a tree as recorded at some point in the past.
52
Trees contain an `Inventory` object, and also know how to retrieve
53
file texts mentioned in the inventory, either from a working
54
directory or from a store.
56
It is possible for trees to contain files that are not described
57
in their inventory or vice versa; for this use `filenames()`.
59
73
Trees can be compared, etc, regardless of whether they are working
60
74
trees or versioned trees.
77
def has_versioned_directories(self):
78
"""Whether this tree can contain explicitly versioned directories.
80
This defaults to True, but some implementations may want to override
63
85
def changes_from(self, other, want_unchanged=False, specific_files=None,
64
86
extra_trees=None, require_versioned=False, include_root=False,
65
87
want_unversioned=False):
147
return iter(self.inventory)
149
165
def all_file_ids(self):
150
166
"""Iterate through all file ids, including ids for missing files."""
151
return set(self.inventory)
167
raise NotImplementedError(self.all_file_ids)
169
def all_versioned_paths(self):
170
"""Iterate through all paths, including paths for missing files."""
171
raise NotImplementedError(self.all_versioned_paths)
153
173
def id2path(self, file_id):
154
174
"""Return the path for a file id.
156
176
:raises NoSuchId:
158
return self.inventory.id2path(file_id)
160
def is_control_filename(self, filename):
161
"""True if filename is the name of a control file in this tree.
163
:param filename: A filename within the tree. This is a relative path
164
from the root of this tree.
166
This is true IF and ONLY IF the filename is part of the meta data
167
that bzr controls in this tree. I.E. a random .bzr directory placed
168
on disk will not be a control file for this tree.
170
return self.bzrdir.is_control_filename(filename)
178
raise NotImplementedError(self.id2path)
173
180
def iter_entries_by_dir(self, specific_file_ids=None, yield_parents=False):
174
181
"""Walk the tree in 'by_dir' order.
194
201
The yield order (ignoring root) would be::
195
203
a, f, a/b, a/d, a/b/c, a/d/e, f/g
197
205
:param yield_parents: If True, yield the parents from the root leading
198
206
down to specific_file_ids that have been requested. This has no
199
207
impact if specific_file_ids is None.
201
return self.inventory.iter_entries_by_dir(
202
specific_file_ids=specific_file_ids, yield_parents=yield_parents)
209
raise NotImplementedError(self.iter_entries_by_dir)
211
def iter_child_entries(self, file_id, path=None):
212
"""Iterate over the children of a directory or tree reference.
214
:param file_id: File id of the directory/tree-reference
215
:param path: Optional path of the directory
216
:raise NoSuchId: When the file_id does not exist
217
:return: Iterator over entries in the directory
219
raise NotImplementedError(self.iter_child_entries)
221
def list_files(self, include_root=False, from_dir=None, recursive=True):
222
"""List all files in this tree.
224
:param include_root: Whether to include the entry for the tree root
225
:param from_dir: Directory under which to list files
226
:param recursive: Whether to list files recursively
227
:return: iterator over tuples of (path, versioned, kind, file_id,
230
raise NotImplementedError(self.list_files)
204
232
def iter_references(self):
205
233
if self.supports_tree_reference():
207
235
if entry.kind == 'tree-reference':
208
236
yield path, entry.file_id
210
def kind(self, file_id):
238
def kind(self, path, file_id=None):
211
239
raise NotImplementedError("Tree subclass %s must implement kind"
212
240
% self.__class__.__name__)
214
def stored_kind(self, file_id):
242
def stored_kind(self, path, file_id=None):
215
243
"""File kind stored for this file_id.
217
245
May not match kind on disk for working trees. Always available
218
246
for versioned files, even when the file itself is missing.
220
return self.kind(file_id)
248
return self.kind(path, file_id)
222
250
def path_content_summary(self, path):
223
251
"""Get a summary of the information about path.
268
293
raise NotImplementedError(self.get_file)
270
def get_file_with_stat(self, file_id, path=None):
295
def get_file_with_stat(self, path, file_id=None):
271
296
"""Get a file handle and stat object for file_id.
273
298
The default implementation returns (self.get_file, None) for backwards
276
:param file_id: The file id to read.
277
:param path: The path of the file, if it is known.
301
:param path: The path of the file.
302
:param file_id: The file id to read, if it is known.
278
303
:return: A tuple (file_handle, stat_value_or_None). If the tree has
279
304
no stat facility, or need for a stat cache feedback during commit,
280
305
it may return None for the second element of the tuple.
282
return (self.get_file(file_id, path), None)
307
return (self.get_file(path, file_id), None)
284
def get_file_text(self, file_id, path=None):
309
def get_file_text(self, path, file_id=None):
285
310
"""Return the byte content of a file.
312
:param path: The path of the file.
287
313
:param file_id: The file_id of the file.
288
:param path: The path of the file.
289
315
If both file_id and path are supplied, an implementation may use
318
:returns: A single byte string for the whole file.
292
my_file = self.get_file(file_id, path)
320
my_file = self.get_file(path, file_id)
294
322
return my_file.read()
298
def get_file_lines(self, file_id, path=None):
326
def get_file_lines(self, path, file_id=None):
299
327
"""Return the content of a file, as lines.
329
:param path: The path of the file.
301
330
:param file_id: The file_id of the file.
302
:param path: The path of the file.
303
332
If both file_id and path are supplied, an implementation may use
306
return osutils.split_lines(self.get_file_text(file_id, path))
308
def get_file_mtime(self, file_id, path=None):
335
return osutils.split_lines(self.get_file_text(path, file_id))
337
def get_file_verifier(self, path, file_id=None, stat_value=None):
338
"""Return a verifier for a file.
340
The default implementation returns a sha1.
342
:param file_id: The handle for this file.
343
:param path: The path that this file can be found at.
344
These must point to the same object.
345
:param stat_value: Optional stat value for the object
346
:return: Tuple with verifier name and verifier data
348
return ("SHA1", self.get_file_sha1(path, file_id,
349
stat_value=stat_value))
351
def get_file_sha1(self, path, file_id=None, stat_value=None):
352
"""Return the SHA1 file for a file.
354
:note: callers should use get_file_verifier instead
355
where possible, as the underlying repository implementation may
356
have quicker access to a non-sha1 verifier.
358
:param path: The path that this file can be found at.
359
:param file_id: The handle for this file.
360
These must point to the same object.
361
:param stat_value: Optional stat value for the object
363
raise NotImplementedError(self.get_file_sha1)
365
def get_file_mtime(self, path, file_id=None):
309
366
"""Return the modification time for a file.
368
:param path: The path that this file can be found at.
311
369
:param file_id: The handle for this file.
312
:param path: The path that this file can be found at.
313
370
These must point to the same object.
315
372
raise NotImplementedError(self.get_file_mtime)
317
def get_file_size(self, file_id):
374
def get_file_size(self, path, file_id=None):
318
375
"""Return the size of a file in bytes.
320
377
This applies only to regular files. If invoked on directories or
349
412
# of bytestrings. (Technically, a bytestring is also an iterable
350
413
# of bytestrings, but iterating through each character is not
352
cur_file = (self.get_file_text(file_id),)
415
# TODO(jelmer): Pass paths into iter_files_bytes
416
path = self.id2path(file_id)
417
cur_file = (self.get_file_text(path, file_id),)
353
418
yield identifier, cur_file
355
def get_symlink_target(self, file_id):
420
def get_symlink_target(self, path, file_id=None):
356
421
"""Get the target for a given file_id.
358
423
It is assumed that the caller already knows that file_id is referencing
360
425
:param file_id: Handle for the symlink entry.
426
:param path: The path of the file.
427
If both file_id and path are supplied, an implementation may use
361
429
:return: The path the symlink points to.
363
431
raise NotImplementedError(self.get_symlink_target)
365
def get_canonical_inventory_paths(self, paths):
366
"""Like get_canonical_inventory_path() but works on multiple items.
368
:param paths: A sequence of paths relative to the root of the tree.
369
:return: A list of paths, with each item the corresponding input path
370
adjusted to account for existing elements that match case
373
return list(self._yield_canonical_inventory_paths(paths))
375
def get_canonical_inventory_path(self, path):
376
"""Returns the first inventory item that case-insensitively matches path.
378
If a path matches exactly, it is returned. If no path matches exactly
379
but more than one path matches case-insensitively, it is implementation
380
defined which is returned.
382
If no path matches case-insensitively, the input path is returned, but
383
with as many path entries that do exist changed to their canonical
386
If you need to resolve many names from the same tree, you should
387
use get_canonical_inventory_paths() to avoid O(N) behaviour.
389
:param path: A paths relative to the root of the tree.
390
:return: The input path adjusted to account for existing elements
391
that match case insensitively.
393
return self._yield_canonical_inventory_paths([path]).next()
395
def _yield_canonical_inventory_paths(self, paths):
397
# First, if the path as specified exists exactly, just use it.
398
if self.path2id(path) is not None:
402
cur_id = self.get_root_id()
404
bit_iter = iter(path.split("/"))
408
for child in self.iter_children(cur_id):
410
# XXX: it seem like if the child is known to be in the
411
# tree, we shouldn't need to go from its id back to
412
# its path -- mbp 2010-02-11
414
# XXX: it seems like we could be more efficient
415
# by just directly looking up the original name and
416
# only then searching all children; also by not
417
# chopping paths so much. -- mbp 2010-02-11
418
child_base = os.path.basename(self.id2path(child))
419
if (child_base == elt):
420
# if we found an exact match, we can stop now; if
421
# we found an approximate match we need to keep
422
# searching because there might be an exact match
425
new_path = osutils.pathjoin(cur_path, child_base)
427
elif child_base.lower() == lelt:
429
new_path = osutils.pathjoin(cur_path, child_base)
431
# before a change is committed we can see this error...
436
# got to the end of this directory and no entries matched.
437
# Return what matched so far, plus the rest as specified.
438
cur_path = osutils.pathjoin(cur_path, elt, *list(bit_iter))
443
433
def get_root_id(self):
444
434
"""Return the file_id for the root of this tree."""
445
435
raise NotImplementedError(self.get_root_id)
447
def annotate_iter(self, file_id,
437
def annotate_iter(self, path, file_id=None,
448
438
default_revision=_mod_revision.CURRENT_REVISION):
449
439
"""Return an iterator of revision_id, line tuples.
459
449
raise NotImplementedError(self.annotate_iter)
461
451
def _get_plan_merge_data(self, file_id, other, base):
462
from bzrlib import versionedfile
452
from .bzr import versionedfile
463
453
vf = versionedfile._PlanMergeVersionedFile(file_id)
464
last_revision_a = self._get_file_revision(file_id, vf, 'this:')
465
last_revision_b = other._get_file_revision(file_id, vf, 'other:')
454
last_revision_a = self._get_file_revision(
455
self.id2path(file_id), file_id, vf, 'this:')
456
last_revision_b = other._get_file_revision(
457
other.id2path(file_id), file_id, vf, 'other:')
467
459
last_revision_base = None
469
last_revision_base = base._get_file_revision(file_id, vf, 'base:')
461
last_revision_base = base._get_file_revision(
462
base.id2path(file_id), file_id, vf, 'base:')
470
463
return vf, last_revision_a, last_revision_b, last_revision_base
472
465
def plan_file_merge(self, file_id, other, base=None):
503
496
except errors.NoSuchRevisionInTree:
504
497
yield self.repository.revision_tree(revision_id)
507
def _file_revision(revision_tree, file_id):
508
"""Determine the revision associated with a file in a given tree."""
509
revision_tree.lock_read()
511
return revision_tree.inventory[file_id].revision
513
revision_tree.unlock()
515
def _get_file_revision(self, file_id, vf, tree_revision):
499
def _get_file_revision(self, path, file_id, vf, tree_revision):
516
500
"""Ensure that file_id, tree_revision is in vf to plan the merge."""
518
501
if getattr(self, '_repository', None) is None:
519
502
last_revision = tree_revision
520
parent_keys = [(file_id, self._file_revision(t, file_id)) for t in
503
parent_keys = [(file_id, t.get_file_revision(path, file_id)) for t in
521
504
self._iter_parent_trees()]
522
505
vf.add_lines((file_id, last_revision), parent_keys,
523
self.get_file(file_id).readlines())
506
self.get_file_lines(path, file_id))
524
507
repo = self.branch.repository
525
508
base_vf = repo.texts
527
last_revision = self._file_revision(self, file_id)
510
last_revision = self.get_file_revision(path, file_id)
528
511
base_vf = self._repository.texts
529
512
if base_vf not in vf.fallback_versionedfiles:
530
513
vf.fallback_versionedfiles.append(base_vf)
531
514
return last_revision
533
inventory = property(_get_inventory,
534
doc="Inventory of this Tree")
536
516
def _check_retrieved(self, ie, f):
537
517
if not __debug__:
539
fp = fingerprint_file(f)
519
fp = osutils.fingerprint_file(f)
542
522
if ie.text_size is not None:
543
523
if ie.text_size != fp['size']:
544
raise BzrError("mismatched size for file %r in %r" % (ie.file_id, self._store),
524
raise errors.BzrError(
525
"mismatched size for file %r in %r" %
526
(ie.file_id, self._store),
545
527
["inventory expects %d bytes" % ie.text_size,
546
528
"file is actually %d bytes" % fp['size'],
547
529
"store is probably damaged/corrupt"])
549
531
if ie.text_sha1 != fp['sha1']:
550
raise BzrError("wrong SHA-1 for file %r in %r" % (ie.file_id, self._store),
532
raise errors.BzrError("wrong SHA-1 for file %r in %r" %
533
(ie.file_id, self._store),
551
534
["inventory expects %s" % ie.text_sha1,
552
535
"file is actually %s" % fp['sha1'],
553
536
"store is probably damaged/corrupt"])
556
538
def path2id(self, path):
557
539
"""Return the id for path in this tree."""
558
return self._inventory.path2id(path)
540
raise NotImplementedError(self.path2id)
560
542
def paths2ids(self, paths, trees=[], require_versioned=True):
561
543
"""Return all the ids that can be reached by walking from paths.
577
559
return find_ids_across_trees(paths, [self] + list(trees), require_versioned)
579
561
def iter_children(self, file_id):
580
entry = self.iter_entries_by_dir([file_id]).next()[1]
581
for child in getattr(entry, 'children', {}).itervalues():
562
"""Iterate over the file ids of the children of an entry.
564
:param file_id: File id of the entry
565
:return: Iterator over child file ids.
567
raise NotImplementedError(self.iter_children)
584
569
def lock_read(self):
570
"""Lock this tree for multiple read only operations.
572
:return: A breezy.lock.LogicalLockResult.
574
return lock.LogicalLockResult(self.unlock)
587
576
def revision_tree(self, revision_id):
588
577
"""Obtain a revision tree for the revision revision_id.
724
######################################################################
727
# TODO: Merge these two functions into a single one that can operate
728
# on either a whole tree or a set of files.
730
# TODO: Return the diff in order by filename, not by category or in
731
# random order. Can probably be done by lock-stepping through the
732
# filenames from both trees.
735
def file_status(filename, old_tree, new_tree):
736
"""Return single-letter status, old and new names for a file.
738
The complexity here is in deciding how to represent renames;
739
many complex cases are possible.
741
old_inv = old_tree.inventory
742
new_inv = new_tree.inventory
743
new_id = new_inv.path2id(filename)
744
old_id = old_inv.path2id(filename)
746
if not new_id and not old_id:
747
# easy: doesn't exist in either; not versioned at all
748
if new_tree.is_ignored(filename):
749
return 'I', None, None
751
return '?', None, None
753
# There is now a file of this name, great.
756
# There is no longer a file of this name, but we can describe
757
# what happened to the file that used to have
758
# this name. There are two possibilities: either it was
759
# deleted entirely, or renamed.
760
if new_inv.has_id(old_id):
761
return 'X', old_inv.id2path(old_id), new_inv.id2path(old_id)
763
return 'D', old_inv.id2path(old_id), None
765
# if the file_id is new in this revision, it is added
766
if new_id and not old_inv.has_id(new_id):
769
# if there used to be a file of this name, but that ID has now
770
# disappeared, it is deleted
771
if old_id and not new_inv.has_id(old_id):
777
@deprecated_function(deprecated_in((1, 9, 0)))
778
def find_renames(old_inv, new_inv):
779
for file_id in old_inv:
780
if file_id not in new_inv:
782
old_name = old_inv.id2path(file_id)
783
new_name = new_inv.id2path(file_id)
784
if old_name != new_name:
785
yield (old_name, new_name)
788
709
def find_ids_across_trees(filenames, trees, require_versioned=True):
789
710
"""Find the ids corresponding to specified filenames.
930
857
if source_kind != target_kind:
931
858
changed_content = True
932
859
elif source_kind == 'file':
933
if (self.source.get_file_sha1(file_id, source_path, source_stat) !=
934
self.target.get_file_sha1(file_id, target_path, target_stat)):
860
if not self.file_content_matches(file_id, file_id, source_path,
861
target_path, source_stat, target_stat):
935
862
changed_content = True
936
863
elif source_kind == 'symlink':
937
if (self.source.get_symlink_target(file_id) !=
938
self.target.get_symlink_target(file_id)):
864
if (self.source.get_symlink_target(source_path, file_id) !=
865
self.target.get_symlink_target(target_path, file_id)):
939
866
changed_content = True
940
# XXX: Yes, the indentation below is wrong. But fixing it broke
941
# test_merge.TestMergerEntriesLCAOnDisk.
942
# test_nested_tree_subtree_renamed_and_modified. We'll wait for
943
# the fix from bzr.dev -- vila 2009026
944
elif source_kind == 'tree-reference':
945
if (self.source.get_reference_revision(file_id, source_path)
946
!= self.target.get_reference_revision(file_id, target_path)):
867
elif source_kind == 'tree-reference':
868
if (self.source.get_reference_revision(source_path, file_id)
869
!= self.target.get_reference_revision(target_path, file_id)):
947
870
changed_content = True
948
871
parent = (source_parent, target_parent)
949
872
name = (source_name, target_name)
980
902
trees = (self.source,)
981
903
if extra_trees is not None:
982
904
trees = trees + tuple(extra_trees)
983
# target is usually the newer tree:
984
specific_file_ids = self.target.paths2ids(specific_files, trees,
985
require_versioned=require_versioned)
986
if specific_files and not specific_file_ids:
987
# All files are unversioned, so just return an empty delta
988
# _compare_trees would think we want a complete delta
989
result = delta.TreeDelta()
990
fake_entry = InventoryFile('unused', 'unused', 'unused')
991
result.unversioned = [(path, None,
992
self.target._comparison_data(fake_entry, path)[0]) for path in
995
return delta._compare_trees(self.source, self.target, want_unchanged,
996
specific_files, include_root, extra_trees=extra_trees,
997
require_versioned=require_versioned,
998
want_unversioned=want_unversioned)
905
with self.lock_read():
906
# target is usually the newer tree:
907
specific_file_ids = self.target.paths2ids(specific_files, trees,
908
require_versioned=require_versioned)
909
if specific_files and not specific_file_ids:
910
# All files are unversioned, so just return an empty delta
911
# _compare_trees would think we want a complete delta
912
result = delta.TreeDelta()
913
fake_entry = inventory.InventoryFile('unused', 'unused', 'unused')
914
result.unversioned = [(path, None,
915
self.target._comparison_data(fake_entry, path)[0]) for path in
918
return delta._compare_trees(self.source, self.target, want_unchanged,
919
specific_files, include_root, extra_trees=extra_trees,
920
require_versioned=require_versioned,
921
want_unversioned=want_unversioned)
1000
923
def iter_changes(self, include_unchanged=False,
1001
924
specific_files=None, pb=None, extra_trees=[],
1239
1162
precise_file_ids.add(new_parent_id)
1241
1164
if (result[6][0] == 'directory' and
1242
result[6][1] != 'directory'):
1165
result[6][1] != 'directory'):
1243
1166
# This stopped being a directory, the old children have
1244
1167
# to be included.
1245
1168
if old_entry is None:
1246
1169
# Reusing a discarded change.
1247
1170
old_entry = self._get_entry(self.source, file_id)
1248
for child in old_entry.children.values():
1249
precise_file_ids.add(child.file_id)
1171
precise_file_ids.update(
1172
self.source.iter_children(file_id))
1250
1173
changed_file_ids.add(result[0])
1176
def file_content_matches(
1177
self, source_file_id, target_file_id, source_path=None,
1178
target_path=None, source_stat=None, target_stat=None):
1179
"""Check if two files are the same in the source and target trees.
1181
This only checks that the contents of the files are the same,
1182
it does not touch anything else.
1184
:param source_file_id: File id of the file in the source tree
1185
:param target_file_id: File id of the file in the target tree
1186
:param source_path: Path of the file in the source tree
1187
:param target_path: Path of the file in the target tree
1188
:param source_stat: Optional stat value of the file in the source tree
1189
:param target_stat: Optional stat value of the file in the target tree
1190
:return: Boolean indicating whether the files have the same contents
1192
with self.lock_read():
1193
if source_path is None:
1194
source_path = self.source.id2path(source_file_id)
1195
if target_path is None:
1196
target_path = self.target.id2path(target_file_id)
1197
source_verifier_kind, source_verifier_data = (
1198
self.source.get_file_verifier(
1199
source_path, source_file_id, source_stat))
1200
target_verifier_kind, target_verifier_data = (
1201
self.target.get_file_verifier(
1202
target_path, target_file_id, target_stat))
1203
if source_verifier_kind == target_verifier_kind:
1204
return (source_verifier_data == target_verifier_data)
1205
# Fall back to SHA1 for now
1206
if source_verifier_kind != "SHA1":
1207
source_sha1 = self.source.get_file_sha1(
1208
source_path, source_file_id, source_stat)
1210
source_sha1 = source_verifier_data
1211
if target_verifier_kind != "SHA1":
1212
target_sha1 = self.target.get_file_sha1(
1213
target_path, target_file_id, target_stat)
1215
target_sha1 = target_verifier_data
1216
return (source_sha1 == target_sha1)
1218
InterTree.register_optimiser(InterTree)
1254
1221
class MultiWalker(object):
1255
1222
"""Walk multiple trees simultaneously, getting combined results."""