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)
153
169
def id2path(self, file_id):
154
170
"""Return the path for a file id.
156
172
: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)
174
raise NotImplementedError(self.id2path)
173
176
def iter_entries_by_dir(self, specific_file_ids=None, yield_parents=False):
174
177
"""Walk the tree in 'by_dir' order.
194
197
The yield order (ignoring root) would be::
195
199
a, f, a/b, a/d, a/b/c, a/d/e, f/g
197
201
:param yield_parents: If True, yield the parents from the root leading
198
202
down to specific_file_ids that have been requested. This has no
199
203
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)
205
raise NotImplementedError(self.iter_entries_by_dir)
207
def iter_child_entries(self, file_id, path=None):
208
"""Iterate over the children of a directory or tree reference.
210
:param file_id: File id of the directory/tree-reference
211
:param path: Optional path of the directory
212
:raise NoSuchId: When the file_id does not exist
213
:return: Iterator over entries in the directory
215
raise NotImplementedError(self.iter_child_entries)
217
def list_files(self, include_root=False, from_dir=None, recursive=True):
218
"""List all files in this tree.
220
:param include_root: Whether to include the entry for the tree root
221
:param from_dir: Directory under which to list files
222
:param recursive: Whether to list files recursively
223
:return: iterator over tuples of (path, versioned, kind, file_id,
226
raise NotImplementedError(self.list_files)
204
228
def iter_references(self):
205
229
if self.supports_tree_reference():
207
231
if entry.kind == 'tree-reference':
208
232
yield path, entry.file_id
210
def kind(self, file_id):
234
def kind(self, path, file_id=None):
211
235
raise NotImplementedError("Tree subclass %s must implement kind"
212
236
% self.__class__.__name__)
214
def stored_kind(self, file_id):
238
def stored_kind(self, path, file_id=None):
215
239
"""File kind stored for this file_id.
217
241
May not match kind on disk for working trees. Always available
218
242
for versioned files, even when the file itself is missing.
220
return self.kind(file_id)
244
return self.kind(path, file_id)
222
246
def path_content_summary(self, path):
223
247
"""Get a summary of the information about path.
268
289
raise NotImplementedError(self.get_file)
270
def get_file_with_stat(self, file_id, path=None):
291
def get_file_with_stat(self, path, file_id=None):
271
292
"""Get a file handle and stat object for file_id.
273
294
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.
297
:param path: The path of the file.
298
:param file_id: The file id to read, if it is known.
278
299
:return: A tuple (file_handle, stat_value_or_None). If the tree has
279
300
no stat facility, or need for a stat cache feedback during commit,
280
301
it may return None for the second element of the tuple.
282
return (self.get_file(file_id, path), None)
303
return (self.get_file(path, file_id), None)
284
def get_file_text(self, file_id, path=None):
305
def get_file_text(self, path, file_id=None):
285
306
"""Return the byte content of a file.
308
:param path: The path of the file.
287
309
:param file_id: The file_id of the file.
288
:param path: The path of the file.
289
311
If both file_id and path are supplied, an implementation may use
314
:returns: A single byte string for the whole file.
292
my_file = self.get_file(file_id, path)
316
my_file = self.get_file(path, file_id)
294
318
return my_file.read()
298
def get_file_lines(self, file_id, path=None):
322
def get_file_lines(self, path, file_id=None):
299
323
"""Return the content of a file, as lines.
325
:param path: The path of the file.
301
326
:param file_id: The file_id of the file.
302
:param path: The path of the file.
303
328
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):
331
return osutils.split_lines(self.get_file_text(path, file_id))
333
def get_file_verifier(self, path, file_id=None, stat_value=None):
334
"""Return a verifier for a file.
336
The default implementation returns a sha1.
338
:param file_id: The handle for this file.
339
:param path: The path that this file can be found at.
340
These must point to the same object.
341
:param stat_value: Optional stat value for the object
342
:return: Tuple with verifier name and verifier data
344
return ("SHA1", self.get_file_sha1(path, file_id,
345
stat_value=stat_value))
347
def get_file_sha1(self, path, file_id=None, stat_value=None):
348
"""Return the SHA1 file for a file.
350
:note: callers should use get_file_verifier instead
351
where possible, as the underlying repository implementation may
352
have quicker access to a non-sha1 verifier.
354
:param path: The path that this file can be found at.
355
:param file_id: The handle for this file.
356
These must point to the same object.
357
:param stat_value: Optional stat value for the object
359
raise NotImplementedError(self.get_file_sha1)
361
def get_file_mtime(self, path, file_id=None):
309
362
"""Return the modification time for a file.
364
:param path: The path that this file can be found at.
311
365
:param file_id: The handle for this file.
312
:param path: The path that this file can be found at.
313
366
These must point to the same object.
315
368
raise NotImplementedError(self.get_file_mtime)
317
def get_file_size(self, file_id):
370
def get_file_size(self, path, file_id=None):
318
371
"""Return the size of a file in bytes.
320
373
This applies only to regular files. If invoked on directories or
349
408
# of bytestrings. (Technically, a bytestring is also an iterable
350
409
# of bytestrings, but iterating through each character is not
352
cur_file = (self.get_file_text(file_id),)
411
# TODO(jelmer): Pass paths into iter_files_bytes
412
path = self.id2path(file_id)
413
cur_file = (self.get_file_text(path, file_id),)
353
414
yield identifier, cur_file
355
def get_symlink_target(self, file_id):
416
def get_symlink_target(self, path, file_id=None):
356
417
"""Get the target for a given file_id.
358
419
It is assumed that the caller already knows that file_id is referencing
360
421
:param file_id: Handle for the symlink entry.
422
:param path: The path of the file.
423
If both file_id and path are supplied, an implementation may use
361
425
:return: The path the symlink points to.
363
427
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
429
def get_root_id(self):
444
430
"""Return the file_id for the root of this tree."""
445
431
raise NotImplementedError(self.get_root_id)
447
def annotate_iter(self, file_id,
433
def annotate_iter(self, path, file_id=None,
448
434
default_revision=_mod_revision.CURRENT_REVISION):
449
435
"""Return an iterator of revision_id, line tuples.
459
445
raise NotImplementedError(self.annotate_iter)
461
447
def _get_plan_merge_data(self, file_id, other, base):
462
from bzrlib import versionedfile
448
from .bzr import versionedfile
463
449
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:')
450
last_revision_a = self._get_file_revision(
451
self.id2path(file_id), file_id, vf, 'this:')
452
last_revision_b = other._get_file_revision(
453
other.id2path(file_id), file_id, vf, 'other:')
467
455
last_revision_base = None
469
last_revision_base = base._get_file_revision(file_id, vf, 'base:')
457
last_revision_base = base._get_file_revision(
458
base.id2path(file_id), file_id, vf, 'base:')
470
459
return vf, last_revision_a, last_revision_b, last_revision_base
472
461
def plan_file_merge(self, file_id, other, base=None):
503
492
except errors.NoSuchRevisionInTree:
504
493
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):
495
def _get_file_revision(self, path, file_id, vf, tree_revision):
516
496
"""Ensure that file_id, tree_revision is in vf to plan the merge."""
518
497
if getattr(self, '_repository', None) is None:
519
498
last_revision = tree_revision
520
parent_keys = [(file_id, self._file_revision(t, file_id)) for t in
499
parent_keys = [(file_id, t.get_file_revision(path, file_id)) for t in
521
500
self._iter_parent_trees()]
522
501
vf.add_lines((file_id, last_revision), parent_keys,
523
self.get_file(file_id).readlines())
502
self.get_file_lines(path, file_id))
524
503
repo = self.branch.repository
525
504
base_vf = repo.texts
527
last_revision = self._file_revision(self, file_id)
506
last_revision = self.get_file_revision(path, file_id)
528
507
base_vf = self._repository.texts
529
508
if base_vf not in vf.fallback_versionedfiles:
530
509
vf.fallback_versionedfiles.append(base_vf)
531
510
return last_revision
533
inventory = property(_get_inventory,
534
doc="Inventory of this Tree")
536
512
def _check_retrieved(self, ie, f):
537
513
if not __debug__:
539
fp = fingerprint_file(f)
515
fp = osutils.fingerprint_file(f)
542
518
if ie.text_size is not None:
543
519
if ie.text_size != fp['size']:
544
raise BzrError("mismatched size for file %r in %r" % (ie.file_id, self._store),
520
raise errors.BzrError(
521
"mismatched size for file %r in %r" %
522
(ie.file_id, self._store),
545
523
["inventory expects %d bytes" % ie.text_size,
546
524
"file is actually %d bytes" % fp['size'],
547
525
"store is probably damaged/corrupt"])
549
527
if ie.text_sha1 != fp['sha1']:
550
raise BzrError("wrong SHA-1 for file %r in %r" % (ie.file_id, self._store),
528
raise errors.BzrError("wrong SHA-1 for file %r in %r" %
529
(ie.file_id, self._store),
551
530
["inventory expects %s" % ie.text_sha1,
552
531
"file is actually %s" % fp['sha1'],
553
532
"store is probably damaged/corrupt"])
556
534
def path2id(self, path):
557
535
"""Return the id for path in this tree."""
558
return self._inventory.path2id(path)
536
raise NotImplementedError(self.path2id)
560
538
def paths2ids(self, paths, trees=[], require_versioned=True):
561
539
"""Return all the ids that can be reached by walking from paths.
577
555
return find_ids_across_trees(paths, [self] + list(trees), require_versioned)
579
557
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():
558
"""Iterate over the file ids of the children of an entry.
560
:param file_id: File id of the entry
561
:return: Iterator over child file ids.
563
raise NotImplementedError(self.iter_children)
584
565
def lock_read(self):
566
"""Lock this tree for multiple read only operations.
568
:return: A breezy.lock.LogicalLockResult.
570
return lock.LogicalLockResult(self.unlock)
587
572
def revision_tree(self, revision_id):
588
573
"""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
705
def find_ids_across_trees(filenames, trees, require_versioned=True):
789
706
"""Find the ids corresponding to specified filenames.
930
853
if source_kind != target_kind:
931
854
changed_content = True
932
855
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)):
856
if not self.file_content_matches(file_id, file_id, source_path,
857
target_path, source_stat, target_stat):
935
858
changed_content = True
936
859
elif source_kind == 'symlink':
937
if (self.source.get_symlink_target(file_id) !=
938
self.target.get_symlink_target(file_id)):
860
if (self.source.get_symlink_target(source_path, file_id) !=
861
self.target.get_symlink_target(target_path, file_id)):
939
862
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)):
863
elif source_kind == 'tree-reference':
864
if (self.source.get_reference_revision(source_path, file_id)
865
!= self.target.get_reference_revision(target_path, file_id)):
947
866
changed_content = True
948
867
parent = (source_parent, target_parent)
949
868
name = (source_name, target_name)
980
898
trees = (self.source,)
981
899
if extra_trees is not None:
982
900
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)
901
with self.lock_read():
902
# target is usually the newer tree:
903
specific_file_ids = self.target.paths2ids(specific_files, trees,
904
require_versioned=require_versioned)
905
if specific_files and not specific_file_ids:
906
# All files are unversioned, so just return an empty delta
907
# _compare_trees would think we want a complete delta
908
result = delta.TreeDelta()
909
fake_entry = inventory.InventoryFile('unused', 'unused', 'unused')
910
result.unversioned = [(path, None,
911
self.target._comparison_data(fake_entry, path)[0]) for path in
914
return delta._compare_trees(self.source, self.target, want_unchanged,
915
specific_files, include_root, extra_trees=extra_trees,
916
require_versioned=require_versioned,
917
want_unversioned=want_unversioned)
1000
919
def iter_changes(self, include_unchanged=False,
1001
920
specific_files=None, pb=None, extra_trees=[],
1239
1158
precise_file_ids.add(new_parent_id)
1241
1160
if (result[6][0] == 'directory' and
1242
result[6][1] != 'directory'):
1161
result[6][1] != 'directory'):
1243
1162
# This stopped being a directory, the old children have
1244
1163
# to be included.
1245
1164
if old_entry is None:
1246
1165
# Reusing a discarded change.
1247
1166
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)
1167
precise_file_ids.update(
1168
self.source.iter_children(file_id))
1250
1169
changed_file_ids.add(result[0])
1172
def file_content_matches(
1173
self, source_file_id, target_file_id, source_path=None,
1174
target_path=None, source_stat=None, target_stat=None):
1175
"""Check if two files are the same in the source and target trees.
1177
This only checks that the contents of the files are the same,
1178
it does not touch anything else.
1180
:param source_file_id: File id of the file in the source tree
1181
:param target_file_id: File id of the file in the target tree
1182
:param source_path: Path of the file in the source tree
1183
:param target_path: Path of the file in the target tree
1184
:param source_stat: Optional stat value of the file in the source tree
1185
:param target_stat: Optional stat value of the file in the target tree
1186
:return: Boolean indicating whether the files have the same contents
1188
with self.lock_read():
1189
if source_path is None:
1190
source_path = self.source.id2path(source_file_id)
1191
if target_path is None:
1192
target_path = self.target.id2path(target_file_id)
1193
source_verifier_kind, source_verifier_data = (
1194
self.source.get_file_verifier(
1195
source_path, source_file_id, source_stat))
1196
target_verifier_kind, target_verifier_data = (
1197
self.target.get_file_verifier(
1198
target_path, target_file_id, target_stat))
1199
if source_verifier_kind == target_verifier_kind:
1200
return (source_verifier_data == target_verifier_data)
1201
# Fall back to SHA1 for now
1202
if source_verifier_kind != "SHA1":
1203
source_sha1 = self.source.get_file_sha1(
1204
source_path, source_file_id, source_stat)
1206
source_sha1 = source_verifier_data
1207
if target_verifier_kind != "SHA1":
1208
target_sha1 = self.target.get_file_sha1(
1209
target_path, target_file_id, target_stat)
1211
target_sha1 = target_verifier_data
1212
return (source_sha1 == target_sha1)
1214
InterTree.register_optimiser(InterTree)
1254
1217
class MultiWalker(object):
1255
1218
"""Walk multiple trees simultaneously, getting combined results."""
added
removed
breezy/tree.py
