1
# Copyright (C) 2005 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
"""Tree classes, representing directory at point in time.
21
from cStringIO import StringIO
29
from bzrlib.decorators import needs_read_lock
30
from bzrlib.errors import BzrError, BzrCheckError
31
from bzrlib import errors
32
from bzrlib.inventory import Inventory
33
from bzrlib.inter import InterObject
34
from bzrlib.osutils import fingerprint_file
35
import bzrlib.revision
36
from bzrlib.trace import mutter, note
40
"""Abstract file tree.
42
There are several subclasses:
44
* `WorkingTree` exists as files on disk editable by the user.
46
* `RevisionTree` is a tree as recorded at some point in the past.
48
Trees contain an `Inventory` object, and also know how to retrieve
49
file texts mentioned in the inventory, either from a working
50
directory or from a store.
52
It is possible for trees to contain files that are not described
53
in their inventory or vice versa; for this use `filenames()`.
55
Trees can be compared, etc, regardless of whether they are working
56
trees or versioned trees.
59
def changes_from(self, other, want_unchanged=False, specific_files=None,
60
extra_trees=None, require_versioned=False, include_root=False):
61
"""Return a TreeDelta of the changes from other to this tree.
63
:param other: A tree to compare with.
64
:param specific_files: An optional list of file paths to restrict the
65
comparison to. When mapping filenames to ids, all matches in all
66
trees (including optional extra_trees) are used, and all children of
67
matched directories are included.
68
:param want_unchanged: An optional boolean requesting the inclusion of
69
unchanged entries in the result.
70
:param extra_trees: An optional list of additional trees to use when
71
mapping the contents of specific_files (paths) to file_ids.
72
:param require_versioned: An optional boolean (defaults to False). When
73
supplied and True all the 'specific_files' must be versioned, or
74
a PathsNotVersionedError will be thrown.
76
The comparison will be performed by an InterTree object looked up on
79
# Martin observes that Tree.changes_from returns a TreeDelta and this
80
# may confuse people, because the class name of the returned object is
81
# a synonym of the object referenced in the method name.
82
return InterTree.get(other, self).compare(
83
want_unchanged=want_unchanged,
84
specific_files=specific_files,
85
extra_trees=extra_trees,
86
require_versioned=require_versioned,
87
include_root=include_root
90
def _iter_changes(self, from_tree, include_unchanged=False,
91
specific_file_ids=None, pb=None):
92
intertree = InterTree.get(from_tree, self)
93
return intertree._iter_changes(from_tree, self, include_unchanged,
94
specific_file_ids, pb)
97
"""Get a list of the conflicts in the tree.
99
Each conflict is an instance of bzrlib.conflicts.Conflict.
103
def get_parent_ids(self):
104
"""Get the parent ids for this tree.
106
:return: a list of parent ids. [] is returned to indicate
107
a tree with no parents.
108
:raises: BzrError if the parents are not known.
110
raise NotImplementedError(self.get_parent_ids)
112
def has_filename(self, filename):
113
"""True if the tree has given filename."""
114
raise NotImplementedError()
116
def has_id(self, file_id):
117
return self.inventory.has_id(file_id)
119
__contains__ = has_id
121
def has_or_had_id(self, file_id):
122
if file_id == self.inventory.root.file_id:
124
return self.inventory.has_id(file_id)
127
return iter(self.inventory)
129
def id2path(self, file_id):
130
return self.inventory.id2path(file_id)
132
def is_control_filename(self, filename):
133
"""True if filename is the name of a control file in this tree.
135
:param filename: A filename within the tree. This is a relative path
136
from the root of this tree.
138
This is true IF and ONLY IF the filename is part of the meta data
139
that bzr controls in this tree. I.E. a random .bzr directory placed
140
on disk will not be a control file for this tree.
142
return self.bzrdir.is_control_filename(filename)
145
def iter_entries_by_dir(self, specific_file_ids=None):
146
"""Walk the tree in 'by_dir' order.
148
This will yield each entry in the tree as a (path, entry) tuple. The
149
order that they are yielded is: the contents of a directory are
150
preceeded by the parent of a directory, and all the contents of a
151
directory are grouped together.
153
return self.inventory.iter_entries_by_dir(
154
specific_file_ids=specific_file_ids)
156
def iter_reference_entries(self):
157
for path, entry in self.iter_entries_by_dir():
158
if entry.kind == 'tree-reference':
161
def kind(self, file_id):
162
raise NotImplementedError("subclasses must implement kind")
164
def get_reference_revision(self, entry, path=None):
165
raise NotImplementedError("subclasses must implement "
166
"get_reference_revision")
168
def _comparison_data(self, entry, path):
169
"""Return a tuple of kind, executable, stat_value for a file.
171
entry may be None if there is no inventory entry for the file, but
172
path must always be supplied.
174
kind is None if there is no file present (even if an inventory id is
175
present). executable is False for non-file entries.
177
raise NotImplementedError(self._comparison_data)
179
def _file_size(self, entry, stat_value):
180
raise NotImplementedError(self._file_size)
182
def _get_inventory(self):
183
return self._inventory
185
def get_file(self, file_id):
186
"""Return a file object for the file file_id in the tree."""
187
raise NotImplementedError(self.get_file)
189
def get_file_by_path(self, path):
190
return self.get_file(self._inventory.path2id(path))
192
def annotate_iter(self, file_id):
193
"""Return an iterator of revision_id, line tuples
195
For working trees (and mutable trees in general), the special
196
revision_id 'current:' will be used for lines that are new in this
197
tree, e.g. uncommitted changes.
198
:param file_id: The file to produce an annotated version from
200
raise NotImplementedError(self.annotate_iter)
202
inventory = property(_get_inventory,
203
doc="Inventory of this Tree")
205
def _check_retrieved(self, ie, f):
208
fp = fingerprint_file(f)
211
if ie.text_size is not None:
212
if ie.text_size != fp['size']:
213
raise BzrError("mismatched size for file %r in %r" % (ie.file_id, self._store),
214
["inventory expects %d bytes" % ie.text_size,
215
"file is actually %d bytes" % fp['size'],
216
"store is probably damaged/corrupt"])
218
if ie.text_sha1 != fp['sha1']:
219
raise BzrError("wrong SHA-1 for file %r in %r" % (ie.file_id, self._store),
220
["inventory expects %s" % ie.text_sha1,
221
"file is actually %s" % fp['sha1'],
222
"store is probably damaged/corrupt"])
224
def path2id(self, path):
225
"""Return the id for path in this tree."""
226
return self._inventory.path2id(path)
228
def paths2ids(self, paths, trees=[], require_versioned=True):
229
"""Return all the ids that can be reached by walking from paths.
231
Each path is looked up in each this tree and any extras provided in
232
trees, and this is repeated recursively: the children in an extra tree
233
of a directory that has been renamed under a provided path in this tree
234
are all returned, even if none exist until a provided path in this
235
tree, and vice versa.
237
:param paths: An iterable of paths to start converting to ids from.
238
Alternatively, if paths is None, no ids should be calculated and None
239
will be returned. This is offered to make calling the api unconditional
240
for code that *might* take a list of files.
241
:param trees: Additional trees to consider.
242
:param require_versioned: If False, do not raise NotVersionedError if
243
an element of paths is not versioned in this tree and all of trees.
245
return find_ids_across_trees(paths, [self] + trees, require_versioned)
247
def print_file(self, file_id):
248
"""Print file with id `file_id` to stdout."""
250
sys.stdout.write(self.get_file_text(file_id))
255
def revision_tree(self, revision_id):
256
"""Obtain a revision tree for the revision revision_id.
258
The intention of this method is to allow access to possibly cached
259
tree data. Implementors of this method should raise NoSuchRevision if
260
the tree is not locally available, even if they could obtain the
261
tree via a repository or some other means. Callers are responsible
262
for finding the ultimate source for a revision tree.
264
:param revision_id: The revision_id of the requested tree.
266
:raises: NoSuchRevision if the tree cannot be obtained.
268
raise errors.NoSuchRevisionInTree(self, revision_id)
271
"""What files are present in this tree and unknown.
273
:return: an iterator over the unknown files.
280
def filter_unversioned_files(self, paths):
281
"""Filter out paths that are not versioned.
283
:return: set of paths.
285
# NB: we specifically *don't* call self.has_filename, because for
286
# WorkingTrees that can indicate files that exist on disk but that
288
pred = self.inventory.has_filename
289
return set((p for p in paths if not pred(p)))
291
def walkdirs(self, prefix=""):
292
"""Walk the contents of this tree from path down.
294
This yields all the data about the contents of a directory at a time.
295
After each directory has been yielded, if the caller has mutated the
296
list to exclude some directories, they are then not descended into.
298
The data yielded is of the form:
299
((directory-relpath, directory-path-from-root, directory-fileid),
300
[(relpath, basename, kind, lstat, path_from_tree_root, file_id,
301
versioned_kind), ...]),
302
- directory-relpath is the containing dirs relpath from prefix
303
- directory-path-from-root is the containing dirs path from /
304
- directory-fileid is the id of the directory if it is versioned.
305
- relpath is the relative path within the subtree being walked.
306
- basename is the basename
307
- kind is the kind of the file now. If unknonwn then the file is not
308
present within the tree - but it may be recorded as versioned. See
310
- lstat is the stat data *if* the file was statted.
311
- path_from_tree_root is the path from the root of the tree.
312
- file_id is the file_id is the entry is versioned.
313
- versioned_kind is the kind of the file as last recorded in the
314
versioning system. If 'unknown' the file is not versioned.
315
One of 'kind' and 'versioned_kind' must not be 'unknown'.
317
:param prefix: Start walking from prefix within the tree rather than
318
at the root. This allows one to walk a subtree but get paths that are
319
relative to a tree rooted higher up.
320
:return: an iterator over the directory data.
322
raise NotImplementedError(self.walkdirs)
325
class EmptyTree(Tree):
328
self._inventory = Inventory(root_id=None)
329
symbol_versioning.warn('EmptyTree is deprecated as of bzr 0.9 please'
330
' use repository.revision_tree instead.',
331
DeprecationWarning, stacklevel=2)
333
def get_parent_ids(self):
336
def get_symlink_target(self, file_id):
339
def has_filename(self, filename):
342
def kind(self, file_id):
343
assert self._inventory[file_id].kind == "directory"
346
def list_files(self, include_root=False):
349
def __contains__(self, file_id):
350
return (file_id in self._inventory)
352
def get_file_sha1(self, file_id, path=None, stat_value=None):
356
######################################################################
359
# TODO: Merge these two functions into a single one that can operate
360
# on either a whole tree or a set of files.
362
# TODO: Return the diff in order by filename, not by category or in
363
# random order. Can probably be done by lock-stepping through the
364
# filenames from both trees.
367
def file_status(filename, old_tree, new_tree):
368
"""Return single-letter status, old and new names for a file.
370
The complexity here is in deciding how to represent renames;
371
many complex cases are possible.
373
old_inv = old_tree.inventory
374
new_inv = new_tree.inventory
375
new_id = new_inv.path2id(filename)
376
old_id = old_inv.path2id(filename)
378
if not new_id and not old_id:
379
# easy: doesn't exist in either; not versioned at all
380
if new_tree.is_ignored(filename):
381
return 'I', None, None
383
return '?', None, None
385
# There is now a file of this name, great.
388
# There is no longer a file of this name, but we can describe
389
# what happened to the file that used to have
390
# this name. There are two possibilities: either it was
391
# deleted entirely, or renamed.
393
if new_inv.has_id(old_id):
394
return 'X', old_inv.id2path(old_id), new_inv.id2path(old_id)
396
return 'D', old_inv.id2path(old_id), None
398
# if the file_id is new in this revision, it is added
399
if new_id and not old_inv.has_id(new_id):
402
# if there used to be a file of this name, but that ID has now
403
# disappeared, it is deleted
404
if old_id and not new_inv.has_id(old_id):
411
def find_renames(old_inv, new_inv):
412
for file_id in old_inv:
413
if file_id not in new_inv:
415
old_name = old_inv.id2path(file_id)
416
new_name = new_inv.id2path(file_id)
417
if old_name != new_name:
418
yield (old_name, new_name)
421
def find_ids_across_trees(filenames, trees, require_versioned=True):
422
"""Find the ids corresponding to specified filenames.
424
All matches in all trees will be used, and all children of matched
425
directories will be used.
427
:param filenames: The filenames to find file_ids for (if None, returns
429
:param trees: The trees to find file_ids within
430
:param require_versioned: if true, all specified filenames must occur in
432
:return: a set of file ids for the specified filenames and their children.
436
specified_path_ids = _find_ids_across_trees(filenames, trees,
438
return _find_children_across_trees(specified_path_ids, trees)
439
# specified_ids = [id for path, id in _find_path_ids_across_trees(filenames, trees, require_versioned)]
440
# return _find_children_across_trees(specified_ids, trees)
442
def find_path_ids_across_trees(filenames, trees, require_versioned=True):
443
"""Find the paths and ids corresponding to specified filenames.
445
All matches in all trees will be used, and all children of matched
446
directories will be included
448
:param filenames: The filenames to find file_ids for
449
:param trees: The trees to find file_ids within
450
:param require_versioned: if true, all specified filenames must occur in
452
:return: a set of (path, file ids) for the specified filenames and their
453
children. The returned path is the path of the id in the first tree
454
that contains it. This matters when files have been moved
458
# This function needs to know the ids for filenames in all trees, then
459
# search for those same files and children in all the other trees.
460
# it is complicated by the same path in two trees being able to have
461
# different ids, which might both be present in both trees.
462
# consider two trees, which have had 'mv foo bar' and 'mv baz foo' done
463
# in this case, a diff of 'foo' should should changes to both the current
464
# 'bar' and the current 'foo' which was baz. Its arguable that if
465
# the situation is 'mv parent/foo bar' and 'mv baz parent/foo', that
466
# we should return the current bar and the current parent/foo' - at the
467
# moment we do, but we loop around all ids and all trees: I*T checks.
469
# Updating this algorithm to be fast in the common case:
470
# nothing has moved, all files have the same id in parent, child and there
471
# are only two trees (or one is working tree and the others are parents).
472
# walk the dirstate. as we find each path, gather the paths of that
473
# id in all trees. add a mapping from the id to the path in those trees.
474
# now lookup children by id, again in all trees; for these trees that
475
# nothing has moved in, the id->path mapping will allow us to find the
476
# parent trivially. To answer 'has anything been moved' in one of the
477
# dirstate parent trees though, we will need to stare harder at it.
479
# Now, given a path index, that is trivial for any one tree, and given
480
# that we can ask for additional data from a dirstate tree, its a single
481
# pass, though it will require scanning the entire tree to find paths
482
# that were at the current location.
483
# ideal results?: There are three things: tree, path, id. Pathologically
484
# we can have completely disjoint ids for each tree; but we cannot have
485
# disjoin paths for each tree, except if we scan each tree for the
486
# different ids from other trees.
488
specified_path_ids = _find_ids_across_trees(filenames, trees,
490
return _find_path_id_children_across_trees(specified_path_ids, trees)
493
def _find_ids_across_trees(filenames, trees, require_versioned):
494
"""Find the ids corresponding to specified filenames.
496
All matches in all trees will be used, but subdirectories are not scanned.
498
:param filenames: The filenames to find file_ids for
499
:param trees: The trees to find file_ids within
500
:param require_versioned: if true, all specified filenames must occur in
502
:return: a set of (path, file ids) for the specified filenames
505
interesting_ids = set()
506
for tree_path in filenames:
509
file_id = tree.path2id(tree_path)
510
if file_id is not None:
511
interesting_ids.add(file_id)
514
not_versioned.append(tree_path)
515
if len(not_versioned) > 0 and require_versioned:
516
raise errors.PathsNotVersionedError(not_versioned)
517
return interesting_ids
520
def _find_children_across_trees(specified_ids, trees):
521
"""Return a set including specified ids and their children
523
All matches in all trees will be used.
525
:param trees: The trees to find file_ids within
526
:return: a set containing all specified ids and their children
528
interesting_ids = set(specified_ids)
529
pending = interesting_ids
530
# now handle children of interesting ids
531
# we loop so that we handle all children of each id in both trees
532
while len(pending) > 0:
534
for file_id in pending:
536
if not tree.has_id(file_id):
538
entry = tree.inventory[file_id]
539
for child in getattr(entry, 'children', {}).itervalues():
540
if child.file_id not in interesting_ids:
541
new_pending.add(child.file_id)
542
interesting_ids.update(new_pending)
543
pending = new_pending
544
return interesting_ids
547
class InterTree(InterObject):
548
"""This class represents operations taking place between two Trees.
550
Its instances have methods like 'compare' and contain references to the
551
source and target trees these operations are to be carried out on.
553
clients of bzrlib should not need to use InterTree directly, rather they
554
should use the convenience methods on Tree such as 'Tree.compare()' which
555
will pass through to InterTree as appropriate.
561
def compare(self, want_unchanged=False, specific_files=None,
562
extra_trees=None, require_versioned=False, include_root=False):
563
"""Return the changes from source to target.
565
:return: A TreeDelta.
566
:param specific_files: An optional list of file paths to restrict the
567
comparison to. When mapping filenames to ids, all matches in all
568
trees (including optional extra_trees) are used, and all children of
569
matched directories are included.
570
:param want_unchanged: An optional boolean requesting the inclusion of
571
unchanged entries in the result.
572
:param extra_trees: An optional list of additional trees to use when
573
mapping the contents of specific_files (paths) to file_ids.
574
:param require_versioned: An optional boolean (defaults to False). When
575
supplied and True all the 'specific_files' must be versioned, or
576
a PathsNotVersionedError will be thrown.
578
# NB: show_status depends on being able to pass in non-versioned files
579
# and report them as unknown
580
trees = (self.source, self.target)
581
if extra_trees is not None:
582
trees = trees + tuple(extra_trees)
583
# target is usually the newer tree:
584
specific_file_ids = self.target.paths2ids(specific_files, [self.source],
585
require_versioned=require_versioned)
586
if specific_files and not specific_file_ids:
587
# All files are unversioned, so just return an empty delta
588
# _compare_trees would think we want a complete delta
589
return delta.TreeDelta()
590
return delta._compare_trees(self.source, self.target, want_unchanged,
591
specific_file_ids, include_root)
593
def _iter_changes(self, from_tree, to_tree, include_unchanged,
594
specific_file_ids, pb):
595
"""Generate an iterator of changes between trees.
598
(file_id, path, changed_content, versioned, parent, name, kind,
601
Path is relative to the to_tree. changed_content is True if the file's
602
content has changed. This includes changes to its kind, and to
605
versioned, parent, name, kind, executable are tuples of (from, to).
606
If a file is missing in a tree, its kind is None.
608
Iteration is done in parent-to-child order, relative to the to_tree.
611
from_entries_by_dir = list(from_tree.inventory.iter_entries_by_dir(
612
specific_file_ids=specific_file_ids))
613
from_data = dict((e.file_id, (p, e)) for p, e in from_entries_by_dir)
614
to_entries_by_dir = list(to_tree.inventory.iter_entries_by_dir(
615
specific_file_ids=specific_file_ids))
616
num_entries = len(from_entries_by_dir) + len(to_entries_by_dir)
618
for to_path, to_entry in to_entries_by_dir:
619
file_id = to_entry.file_id
620
to_paths[file_id] = to_path
622
changed_content = False
623
from_path, from_entry = from_data.get(file_id, (None, None))
624
from_versioned = (from_entry is not None)
625
if from_entry is not None:
626
from_versioned = True
627
from_name = from_entry.name
628
from_parent = from_entry.parent_id
629
from_kind, from_executable, from_stat = \
630
from_tree._comparison_data(from_entry, from_path)
633
from_versioned = False
637
from_executable = None
638
versioned = (from_versioned, True)
639
to_kind, to_executable, to_stat = \
640
to_tree._comparison_data(to_entry, to_path)
641
kind = (from_kind, to_kind)
642
if kind[0] != kind[1]:
643
changed_content = True
644
elif from_kind == 'file':
645
from_size = from_tree._file_size(from_entry, from_stat)
646
to_size = to_tree._file_size(to_entry, to_stat)
647
if from_size != to_size:
648
changed_content = True
649
elif (from_tree.get_file_sha1(file_id, from_path, from_stat) !=
650
to_tree.get_file_sha1(file_id, to_path, to_stat)):
651
changed_content = True
652
elif from_kind == 'symlink':
653
if (from_tree.get_symlink_target(file_id) !=
654
to_tree.get_symlink_target(file_id)):
655
changed_content = True
656
elif from_kind == 'tree-reference':
657
if (from_tree.get_reference_revision(from_entry, from_path) !=
658
to_tree.get_reference_revision(to_entry, to_path)):
659
changed_content = True
660
parent = (from_parent, to_entry.parent_id)
661
name = (from_name, to_entry.name)
662
executable = (from_executable, to_executable)
664
pb.update('comparing files', entry_count, num_entries)
665
if (changed_content is not False or versioned[0] != versioned[1]
666
or parent[0] != parent[1] or name[0] != name[1] or
667
executable[0] != executable[1] or include_unchanged):
668
yield (file_id, to_path, changed_content, versioned, parent,
669
name, kind, executable)
671
def get_to_path(from_entry):
672
if from_entry.parent_id is None:
675
if from_entry.parent_id not in to_paths:
676
get_to_path(from_tree.inventory[from_entry.parent_id])
677
to_path = osutils.pathjoin(to_paths[from_entry.parent_id],
679
to_paths[from_entry.file_id] = to_path
682
for path, from_entry in from_entries_by_dir:
683
file_id = from_entry.file_id
684
if file_id in to_paths:
686
to_path = get_to_path(from_entry)
689
pb.update('comparing files', entry_count, num_entries)
690
versioned = (True, False)
691
parent = (from_entry.parent_id, None)
692
name = (from_entry.name, None)
693
from_kind, from_executable, stat_value = \
694
from_tree._comparison_data(from_entry, path)
695
kind = (from_kind, None)
696
executable = (from_executable, None)
697
changed_content = True
698
# the parent's path is necessarily known at this point.
699
yield(file_id, to_path, changed_content, versioned, parent,
700
name, kind, executable)
703
# This was deprecated before 0.12, but did not have an official warning
704
@symbol_versioning.deprecated_function(symbol_versioning.zero_twelve)
705
def RevisionTree(*args, **kwargs):
706
"""RevisionTree has moved to bzrlib.revisiontree.RevisionTree()
708
Accessing it as bzrlib.tree.RevisionTree has been deprecated as of
711
from bzrlib.revisiontree import RevisionTree as _RevisionTree
712
return _RevisionTree(*args, **kwargs)
added
removed
bzrlib/tree.py
