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(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
file_id = osutils.safe_file_id(file_id)
118
return self.inventory.has_id(file_id)
120
__contains__ = has_id
122
def has_or_had_id(self, file_id):
123
file_id = osutils.safe_file_id(file_id)
124
if file_id == self.inventory.root.file_id:
126
return self.inventory.has_id(file_id)
129
return iter(self.inventory)
131
def id2path(self, file_id):
132
file_id = osutils.safe_file_id(file_id)
133
return self.inventory.id2path(file_id)
135
def is_control_filename(self, filename):
136
"""True if filename is the name of a control file in this tree.
138
:param filename: A filename within the tree. This is a relative path
139
from the root of this tree.
141
This is true IF and ONLY IF the filename is part of the meta data
142
that bzr controls in this tree. I.E. a random .bzr directory placed
143
on disk will not be a control file for this tree.
145
return self.bzrdir.is_control_filename(filename)
148
def iter_entries_by_dir(self, specific_file_ids=None):
149
"""Walk the tree in 'by_dir' order.
151
This will yield each entry in the tree as a (path, entry) tuple. The
152
order that they are yielded is: the contents of a directory are
153
preceeded by the parent of a directory, and all the contents of a
154
directory are grouped together.
156
return self.inventory.iter_entries_by_dir(
157
specific_file_ids=specific_file_ids)
159
def kind(self, file_id):
160
raise NotImplementedError("subclasses must implement kind")
162
def _comparison_data(self, entry, path):
163
"""Return a tuple of kind, executable, stat_value for a file.
165
entry may be None if there is no inventory entry for the file, but
166
path must always be supplied.
168
kind is None if there is no file present (even if an inventory id is
169
present). executable is False for non-file entries.
171
raise NotImplementedError(self._comparison_data)
173
def _file_size(self, entry, stat_value):
174
raise NotImplementedError(self._file_size)
176
def _get_inventory(self):
177
return self._inventory
179
def get_file(self, file_id):
180
"""Return a file object for the file file_id in the tree."""
181
raise NotImplementedError(self.get_file)
183
def get_file_by_path(self, path):
184
return self.get_file(self._inventory.path2id(path))
186
def get_symlink_target(self, file_id):
187
"""Get the target for a given file_id.
189
It is assumed that the caller already knows that file_id is referencing
191
:param file_id: Handle for the symlink entry.
192
:return: The path the symlink points to.
194
raise NotImplementedError(self.get_symlink_target)
196
def annotate_iter(self, file_id):
197
"""Return an iterator of revision_id, line tuples
199
For working trees (and mutable trees in general), the special
200
revision_id 'current:' will be used for lines that are new in this
201
tree, e.g. uncommitted changes.
202
:param file_id: The file to produce an annotated version from
204
raise NotImplementedError(self.annotate_iter)
206
inventory = property(_get_inventory,
207
doc="Inventory of this Tree")
209
def _check_retrieved(self, ie, f):
212
fp = fingerprint_file(f)
215
if ie.text_size is not None:
216
if ie.text_size != fp['size']:
217
raise BzrError("mismatched size for file %r in %r" % (ie.file_id, self._store),
218
["inventory expects %d bytes" % ie.text_size,
219
"file is actually %d bytes" % fp['size'],
220
"store is probably damaged/corrupt"])
222
if ie.text_sha1 != fp['sha1']:
223
raise BzrError("wrong SHA-1 for file %r in %r" % (ie.file_id, self._store),
224
["inventory expects %s" % ie.text_sha1,
225
"file is actually %s" % fp['sha1'],
226
"store is probably damaged/corrupt"])
228
def path2id(self, path):
229
"""Return the id for path in this tree."""
230
return self._inventory.path2id(path)
232
def paths2ids(self, paths, trees=[], require_versioned=True):
233
"""Return all the ids that can be reached by walking from paths.
235
Each path is looked up in each this tree and any extras provided in
236
trees, and this is repeated recursively: the children in an extra tree
237
of a directory that has been renamed under a provided path in this tree
238
are all returned, even if none exist until a provided path in this
239
tree, and vice versa.
241
:param paths: An iterable of paths to start converting to ids from.
242
Alternatively, if paths is None, no ids should be calculated and None
243
will be returned. This is offered to make calling the api unconditional
244
for code that *might* take a list of files.
245
:param trees: Additional trees to consider.
246
:param require_versioned: If False, do not raise NotVersionedError if
247
an element of paths is not versioned in this tree and all of trees.
249
return find_ids_across_trees(paths, [self] + list(trees), require_versioned)
251
def print_file(self, file_id):
252
"""Print file with id `file_id` to stdout."""
253
file_id = osutils.safe_file_id(file_id)
255
sys.stdout.write(self.get_file_text(file_id))
260
def revision_tree(self, revision_id):
261
"""Obtain a revision tree for the revision revision_id.
263
The intention of this method is to allow access to possibly cached
264
tree data. Implementors of this method should raise NoSuchRevision if
265
the tree is not locally available, even if they could obtain the
266
tree via a repository or some other means. Callers are responsible
267
for finding the ultimate source for a revision tree.
269
:param revision_id: The revision_id of the requested tree.
271
:raises: NoSuchRevision if the tree cannot be obtained.
273
raise errors.NoSuchRevisionInTree(self, revision_id)
276
"""What files are present in this tree and unknown.
278
:return: an iterator over the unknown files.
285
def filter_unversioned_files(self, paths):
286
"""Filter out paths that are not versioned.
288
:return: set of paths.
290
# NB: we specifically *don't* call self.has_filename, because for
291
# WorkingTrees that can indicate files that exist on disk but that
293
pred = self.inventory.has_filename
294
return set((p for p in paths if not pred(p)))
296
def walkdirs(self, prefix=""):
297
"""Walk the contents of this tree from path down.
299
This yields all the data about the contents of a directory at a time.
300
After each directory has been yielded, if the caller has mutated the
301
list to exclude some directories, they are then not descended into.
303
The data yielded is of the form:
304
((directory-relpath, directory-path-from-root, directory-fileid),
305
[(relpath, basename, kind, lstat, path_from_tree_root, file_id,
306
versioned_kind), ...]),
307
- directory-relpath is the containing dirs relpath from prefix
308
- directory-path-from-root is the containing dirs path from /
309
- directory-fileid is the id of the directory if it is versioned.
310
- relpath is the relative path within the subtree being walked.
311
- basename is the basename
312
- kind is the kind of the file now. If unknonwn then the file is not
313
present within the tree - but it may be recorded as versioned. See
315
- lstat is the stat data *if* the file was statted.
316
- path_from_tree_root is the path from the root of the tree.
317
- file_id is the file_id is the entry is versioned.
318
- versioned_kind is the kind of the file as last recorded in the
319
versioning system. If 'unknown' the file is not versioned.
320
One of 'kind' and 'versioned_kind' must not be 'unknown'.
322
:param prefix: Start walking from prefix within the tree rather than
323
at the root. This allows one to walk a subtree but get paths that are
324
relative to a tree rooted higher up.
325
:return: an iterator over the directory data.
327
raise NotImplementedError(self.walkdirs)
330
class EmptyTree(Tree):
333
self._inventory = Inventory(root_id=None)
334
symbol_versioning.warn('EmptyTree is deprecated as of bzr 0.9 please'
335
' use repository.revision_tree instead.',
336
DeprecationWarning, stacklevel=2)
338
def get_parent_ids(self):
341
def get_symlink_target(self, file_id):
344
def has_filename(self, filename):
347
def kind(self, file_id):
348
file_id = osutils.safe_file_id(file_id)
349
assert self._inventory[file_id].kind == "directory"
352
def list_files(self, include_root=False):
355
def __contains__(self, file_id):
356
file_id = osutils.safe_file_id(file_id)
357
return (file_id in self._inventory)
359
def get_file_sha1(self, file_id, path=None, stat_value=None):
363
######################################################################
366
# TODO: Merge these two functions into a single one that can operate
367
# on either a whole tree or a set of files.
369
# TODO: Return the diff in order by filename, not by category or in
370
# random order. Can probably be done by lock-stepping through the
371
# filenames from both trees.
374
def file_status(filename, old_tree, new_tree):
375
"""Return single-letter status, old and new names for a file.
377
The complexity here is in deciding how to represent renames;
378
many complex cases are possible.
380
old_inv = old_tree.inventory
381
new_inv = new_tree.inventory
382
new_id = new_inv.path2id(filename)
383
old_id = old_inv.path2id(filename)
385
if not new_id and not old_id:
386
# easy: doesn't exist in either; not versioned at all
387
if new_tree.is_ignored(filename):
388
return 'I', None, None
390
return '?', None, None
392
# There is now a file of this name, great.
395
# There is no longer a file of this name, but we can describe
396
# what happened to the file that used to have
397
# this name. There are two possibilities: either it was
398
# deleted entirely, or renamed.
400
if new_inv.has_id(old_id):
401
return 'X', old_inv.id2path(old_id), new_inv.id2path(old_id)
403
return 'D', old_inv.id2path(old_id), None
405
# if the file_id is new in this revision, it is added
406
if new_id and not old_inv.has_id(new_id):
409
# if there used to be a file of this name, but that ID has now
410
# disappeared, it is deleted
411
if old_id and not new_inv.has_id(old_id):
418
def find_renames(old_inv, new_inv):
419
for file_id in old_inv:
420
if file_id not in new_inv:
422
old_name = old_inv.id2path(file_id)
423
new_name = new_inv.id2path(file_id)
424
if old_name != new_name:
425
yield (old_name, new_name)
428
def find_ids_across_trees(filenames, trees, require_versioned=True):
429
"""Find the ids corresponding to specified filenames.
431
All matches in all trees will be used, and all children of matched
432
directories will be used.
434
:param filenames: The filenames to find file_ids for (if None, returns
436
:param trees: The trees to find file_ids within
437
:param require_versioned: if true, all specified filenames must occur in
439
:return: a set of file ids for the specified filenames and their children.
443
specified_path_ids = _find_ids_across_trees(filenames, trees,
445
return _find_children_across_trees(specified_path_ids, trees)
446
# specified_ids = [id for path, id in _find_path_ids_across_trees(filenames, trees, require_versioned)]
447
# return _find_children_across_trees(specified_ids, trees)
449
def find_path_ids_across_trees(filenames, trees, require_versioned=True):
450
"""Find the paths and ids corresponding to specified filenames.
452
All matches in all trees will be used, and all children of matched
453
directories will be included
455
:param filenames: The filenames to find file_ids for
456
:param trees: The trees to find file_ids within
457
:param require_versioned: if true, all specified filenames must occur in
459
:return: a set of (path, file ids) for the specified filenames and their
460
children. The returned path is the path of the id in the first tree
461
that contains it. This matters when files have been moved
465
# This function needs to know the ids for filenames in all trees, then
466
# search for those same files and children in all the other trees.
467
# it is complicated by the same path in two trees being able to have
468
# different ids, which might both be present in both trees.
469
# consider two trees, which have had 'mv foo bar' and 'mv baz foo' done
470
# in this case, a diff of 'foo' should should changes to both the current
471
# 'bar' and the current 'foo' which was baz. Its arguable that if
472
# the situation is 'mv parent/foo bar' and 'mv baz parent/foo', that
473
# we should return the current bar and the current parent/foo' - at the
474
# moment we do, but we loop around all ids and all trees: I*T checks.
476
# Updating this algorithm to be fast in the common case:
477
# nothing has moved, all files have the same id in parent, child and there
478
# are only two trees (or one is working tree and the others are parents).
479
# walk the dirstate. as we find each path, gather the paths of that
480
# id in all trees. add a mapping from the id to the path in those trees.
481
# now lookup children by id, again in all trees; for these trees that
482
# nothing has moved in, the id->path mapping will allow us to find the
483
# parent trivially. To answer 'has anything been moved' in one of the
484
# dirstate parent trees though, we will need to stare harder at it.
486
# Now, given a path index, that is trivial for any one tree, and given
487
# that we can ask for additional data from a dirstate tree, its a single
488
# pass, though it will require scanning the entire tree to find paths
489
# that were at the current location.
490
# ideal results?: There are three things: tree, path, id. Pathologically
491
# we can have completely disjoint ids for each tree; but we cannot have
492
# disjoin paths for each tree, except if we scan each tree for the
493
# different ids from other trees.
495
specified_path_ids = _find_ids_across_trees(filenames, trees,
497
return _find_path_id_children_across_trees(specified_path_ids, trees)
500
def _find_ids_across_trees(filenames, trees, require_versioned):
501
"""Find the ids corresponding to specified filenames.
503
All matches in all trees will be used, but subdirectories are not scanned.
505
:param filenames: The filenames to find file_ids for
506
:param trees: The trees to find file_ids within
507
:param require_versioned: if true, all specified filenames must occur in
509
:return: a set of (path, file ids) for the specified filenames
512
interesting_ids = set()
513
for tree_path in filenames:
516
file_id = tree.path2id(tree_path)
517
if file_id is not None:
518
interesting_ids.add(file_id)
521
not_versioned.append(tree_path)
522
if len(not_versioned) > 0 and require_versioned:
523
raise errors.PathsNotVersionedError(not_versioned)
524
return interesting_ids
527
def _find_children_across_trees(specified_ids, trees):
528
"""Return a set including specified ids and their children
530
All matches in all trees will be used.
532
:param trees: The trees to find file_ids within
533
:return: a set containing all specified ids and their children
535
interesting_ids = set(specified_ids)
536
pending = interesting_ids
537
# now handle children of interesting ids
538
# we loop so that we handle all children of each id in both trees
539
while len(pending) > 0:
541
for file_id in pending:
543
if not tree.has_id(file_id):
545
entry = tree.inventory[file_id]
546
for child in getattr(entry, 'children', {}).itervalues():
547
if child.file_id not in interesting_ids:
548
new_pending.add(child.file_id)
549
interesting_ids.update(new_pending)
550
pending = new_pending
551
return interesting_ids
554
class InterTree(InterObject):
555
"""This class represents operations taking place between two Trees.
557
Its instances have methods like 'compare' and contain references to the
558
source and target trees these operations are to be carried out on.
560
clients of bzrlib should not need to use InterTree directly, rather they
561
should use the convenience methods on Tree such as 'Tree.compare()' which
562
will pass through to InterTree as appropriate.
568
def compare(self, want_unchanged=False, specific_files=None,
569
extra_trees=None, require_versioned=False, include_root=False):
570
"""Return the changes from source to target.
572
:return: A TreeDelta.
573
:param specific_files: An optional list of file paths to restrict the
574
comparison to. When mapping filenames to ids, all matches in all
575
trees (including optional extra_trees) are used, and all children of
576
matched directories are included.
577
:param want_unchanged: An optional boolean requesting the inclusion of
578
unchanged entries in the result.
579
:param extra_trees: An optional list of additional trees to use when
580
mapping the contents of specific_files (paths) to file_ids.
581
:param require_versioned: An optional boolean (defaults to False). When
582
supplied and True all the 'specific_files' must be versioned, or
583
a PathsNotVersionedError will be thrown.
585
# NB: show_status depends on being able to pass in non-versioned files
586
# and report them as unknown
587
trees = (self.source,)
588
if extra_trees is not None:
589
trees = trees + tuple(extra_trees)
590
# target is usually the newer tree:
591
specific_file_ids = self.target.paths2ids(specific_files, trees,
592
require_versioned=require_versioned)
593
if specific_files and not specific_file_ids:
594
# All files are unversioned, so just return an empty delta
595
# _compare_trees would think we want a complete delta
596
return delta.TreeDelta()
597
return delta._compare_trees(self.source, self.target, want_unchanged,
598
specific_file_ids, include_root)
600
def _iter_changes(self, include_unchanged=False,
601
specific_file_ids=None, pb=None):
602
"""Generate an iterator of changes between trees.
605
(file_id, path, changed_content, versioned, parent, name, kind,
608
Path is relative to the target tree. changed_content is True if the
609
file's content has changed. This includes changes to its kind, and to
612
versioned, parent, name, kind, executable are tuples of (from, to).
613
If a file is missing in a tree, its kind is None.
615
Iteration is done in parent-to-child order, relative to the target
619
from_entries_by_dir = list(self.source.inventory.iter_entries_by_dir(
620
specific_file_ids=specific_file_ids))
621
from_data = dict((e.file_id, (p, e)) for p, e in from_entries_by_dir)
622
to_entries_by_dir = list(self.target.inventory.iter_entries_by_dir(
623
specific_file_ids=specific_file_ids))
624
num_entries = len(from_entries_by_dir) + len(to_entries_by_dir)
626
for to_path, to_entry in to_entries_by_dir:
627
file_id = to_entry.file_id
628
to_paths[file_id] = to_path
630
changed_content = False
631
from_path, from_entry = from_data.get(file_id, (None, None))
632
from_versioned = (from_entry is not None)
633
if from_entry is not None:
634
from_versioned = True
635
from_name = from_entry.name
636
from_parent = from_entry.parent_id
637
from_kind, from_executable, from_stat = \
638
self.source._comparison_data(from_entry, from_path)
641
from_versioned = False
645
from_executable = None
646
versioned = (from_versioned, True)
647
to_kind, to_executable, to_stat = \
648
self.target._comparison_data(to_entry, to_path)
649
kind = (from_kind, to_kind)
650
if kind[0] != kind[1]:
651
changed_content = True
652
elif from_kind == 'file':
653
from_size = self.source._file_size(from_entry, from_stat)
654
to_size = self.target._file_size(to_entry, to_stat)
655
if from_size != to_size:
656
changed_content = True
657
elif (self.source.get_file_sha1(file_id, from_path, from_stat) !=
658
self.target.get_file_sha1(file_id, to_path, to_stat)):
659
changed_content = True
660
elif from_kind == 'symlink':
661
if (self.source.get_symlink_target(file_id) !=
662
self.target.get_symlink_target(file_id)):
663
changed_content = True
664
parent = (from_parent, to_entry.parent_id)
665
name = (from_name, to_entry.name)
666
executable = (from_executable, to_executable)
668
pb.update('comparing files', entry_count, num_entries)
669
if (changed_content is not False or versioned[0] != versioned[1]
670
or parent[0] != parent[1] or name[0] != name[1] or
671
executable[0] != executable[1] or include_unchanged):
672
yield (file_id, to_path, changed_content, versioned, parent,
673
name, kind, executable)
675
def get_to_path(from_entry):
676
if from_entry.parent_id is None:
679
if from_entry.parent_id not in to_paths:
680
get_to_path(self.source.inventory[from_entry.parent_id])
681
to_path = osutils.pathjoin(to_paths[from_entry.parent_id],
683
to_paths[from_entry.file_id] = to_path
686
for path, from_entry in from_entries_by_dir:
687
file_id = from_entry.file_id
688
if file_id in to_paths:
690
to_path = get_to_path(from_entry)
693
pb.update('comparing files', entry_count, num_entries)
694
versioned = (True, False)
695
parent = (from_entry.parent_id, None)
696
name = (from_entry.name, None)
697
from_kind, from_executable, stat_value = \
698
self.source._comparison_data(from_entry, path)
699
kind = (from_kind, None)
700
executable = (from_executable, None)
701
changed_content = True
702
# the parent's path is necessarily known at this point.
703
yield(file_id, to_path, changed_content, versioned, parent,
704
name, kind, executable)
707
# This was deprecated before 0.12, but did not have an official warning
708
@symbol_versioning.deprecated_function(symbol_versioning.zero_twelve)
709
def RevisionTree(*args, **kwargs):
710
"""RevisionTree has moved to bzrlib.revisiontree.RevisionTree()
712
Accessing it as bzrlib.tree.RevisionTree has been deprecated as of
715
from bzrlib.revisiontree import RevisionTree as _RevisionTree
716
return _RevisionTree(*args, **kwargs)
added
removed
bzrlib/tree.py
