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 kind(self, file_id):
157
raise NotImplementedError("subclasses must implement kind")
159
def _comparison_data(self, entry, path):
160
"""Return a tuple of kind, executable, stat_value for a file.
162
entry may be None if there is no inventory entry for the file, but
163
path must always be supplied.
165
kind is None if there is no file present (even if an inventory id is
166
present). executable is False for non-file entries.
168
raise NotImplementedError(self._comparison_data)
170
def _file_size(self, entry, stat_value):
171
raise NotImplementedError(self._file_size)
173
def _get_inventory(self):
174
return self._inventory
176
def get_file(self, file_id):
177
"""Return a file object for the file file_id in the tree."""
178
raise NotImplementedError(self.get_file)
180
def get_file_by_path(self, path):
181
return self.get_file(self._inventory.path2id(path))
183
def annotate_iter(self, file_id):
184
"""Return an iterator of revision_id, line tuples
186
For working trees (and mutable trees in general), the special
187
revision_id 'current:' will be used for lines that are new in this
188
tree, e.g. uncommitted changes.
189
:param file_id: The file to produce an annotated version from
191
raise NotImplementedError(self.annotate_iter)
193
inventory = property(_get_inventory,
194
doc="Inventory of this Tree")
196
def _check_retrieved(self, ie, f):
199
fp = fingerprint_file(f)
202
if ie.text_size is not None:
203
if ie.text_size != fp['size']:
204
raise BzrError("mismatched size for file %r in %r" % (ie.file_id, self._store),
205
["inventory expects %d bytes" % ie.text_size,
206
"file is actually %d bytes" % fp['size'],
207
"store is probably damaged/corrupt"])
209
if ie.text_sha1 != fp['sha1']:
210
raise BzrError("wrong SHA-1 for file %r in %r" % (ie.file_id, self._store),
211
["inventory expects %s" % ie.text_sha1,
212
"file is actually %s" % fp['sha1'],
213
"store is probably damaged/corrupt"])
215
def path2id(self, path):
216
"""Return the id for path in this tree."""
217
return self._inventory.path2id(path)
219
def paths2ids(self, paths, trees=[], require_versioned=True):
220
"""Return all the ids that can be reached by walking from paths.
222
Each path is looked up in each this tree and any extras provided in
223
trees, and this is repeated recursively: the children in an extra tree
224
of a directory that has been renamed under a provided path in this tree
225
are all returned, even if none exist until a provided path in this
226
tree, and vice versa.
228
:param paths: An iterable of paths to start converting to ids from.
229
:param trees: Additional trees to consider.
230
:param require_versioned: If False, do not raise NotVersionedError if
231
an element of paths is not versioned in this tree and all of trees.
233
return find_ids_across_trees(paths, [self] + trees, require_versioned)
235
def print_file(self, file_id):
236
"""Print file with id `file_id` to stdout."""
238
sys.stdout.write(self.get_file_text(file_id))
243
def revision_tree(self, revision_id):
244
"""Obtain a revision tree for the revision revision_id.
246
The intention of this method is to allow access to possibly cached
247
tree data. Implementors of this method should raise NoSuchRevision if
248
the tree is not locally available, even if they could obtain the
249
tree via a repository or some other means. Callers are responsible
250
for finding the ultimate source for a revision tree.
252
:param revision_id: The revision_id of the requested tree.
254
:raises: NoSuchRevision if the tree cannot be obtained.
256
raise errors.NoSuchRevisionInTree(self, revision_id)
259
"""What files are present in this tree and unknown.
261
:return: an iterator over the unknown files.
268
def filter_unversioned_files(self, paths):
269
"""Filter out paths that are not versioned.
271
:return: set of paths.
273
# NB: we specifically *don't* call self.has_filename, because for
274
# WorkingTrees that can indicate files that exist on disk but that
276
pred = self.inventory.has_filename
277
return set((p for p in paths if not pred(p)))
279
def walkdirs(self, prefix=""):
280
"""Walk the contents of this tree from path down.
282
This yields all the data about the contents of a directory at a time.
283
After each directory has been yielded, if the caller has mutated the
284
list to exclude some directories, they are then not descended into.
286
The data yielded is of the form:
287
((directory-relpath, directory-path-from-root, directory-fileid),
288
[(relpath, basename, kind, lstat, path_from_tree_root, file_id,
289
versioned_kind), ...]),
290
- directory-relpath is the containing dirs relpath from prefix
291
- directory-path-from-root is the containing dirs path from /
292
- directory-fileid is the id of the directory if it is versioned.
293
- relpath is the relative path within the subtree being walked.
294
- basename is the basename
295
- kind is the kind of the file now. If unknonwn then the file is not
296
present within the tree - but it may be recorded as versioned. See
298
- lstat is the stat data *if* the file was statted.
299
- path_from_tree_root is the path from the root of the tree.
300
- file_id is the file_id is the entry is versioned.
301
- versioned_kind is the kind of the file as last recorded in the
302
versioning system. If 'unknown' the file is not versioned.
303
One of 'kind' and 'versioned_kind' must not be 'unknown'.
305
:param prefix: Start walking from prefix within the tree rather than
306
at the root. This allows one to walk a subtree but get paths that are
307
relative to a tree rooted higher up.
308
:return: an iterator over the directory data.
310
raise NotImplementedError(self.walkdirs)
313
class EmptyTree(Tree):
316
self._inventory = Inventory(root_id=None)
317
symbol_versioning.warn('EmptyTree is deprecated as of bzr 0.9 please'
318
' use repository.revision_tree instead.',
319
DeprecationWarning, stacklevel=2)
321
def get_parent_ids(self):
324
def get_symlink_target(self, file_id):
327
def has_filename(self, filename):
330
def kind(self, file_id):
331
assert self._inventory[file_id].kind == "directory"
334
def list_files(self, include_root=False):
337
def __contains__(self, file_id):
338
return (file_id in self._inventory)
340
def get_file_sha1(self, file_id, path=None, stat_value=None):
344
######################################################################
347
# TODO: Merge these two functions into a single one that can operate
348
# on either a whole tree or a set of files.
350
# TODO: Return the diff in order by filename, not by category or in
351
# random order. Can probably be done by lock-stepping through the
352
# filenames from both trees.
355
def file_status(filename, old_tree, new_tree):
356
"""Return single-letter status, old and new names for a file.
358
The complexity here is in deciding how to represent renames;
359
many complex cases are possible.
361
old_inv = old_tree.inventory
362
new_inv = new_tree.inventory
363
new_id = new_inv.path2id(filename)
364
old_id = old_inv.path2id(filename)
366
if not new_id and not old_id:
367
# easy: doesn't exist in either; not versioned at all
368
if new_tree.is_ignored(filename):
369
return 'I', None, None
371
return '?', None, None
373
# There is now a file of this name, great.
376
# There is no longer a file of this name, but we can describe
377
# what happened to the file that used to have
378
# this name. There are two possibilities: either it was
379
# deleted entirely, or renamed.
381
if new_inv.has_id(old_id):
382
return 'X', old_inv.id2path(old_id), new_inv.id2path(old_id)
384
return 'D', old_inv.id2path(old_id), None
386
# if the file_id is new in this revision, it is added
387
if new_id and not old_inv.has_id(new_id):
390
# if there used to be a file of this name, but that ID has now
391
# disappeared, it is deleted
392
if old_id and not new_inv.has_id(old_id):
399
def find_renames(old_inv, new_inv):
400
for file_id in old_inv:
401
if file_id not in new_inv:
403
old_name = old_inv.id2path(file_id)
404
new_name = new_inv.id2path(file_id)
405
if old_name != new_name:
406
yield (old_name, new_name)
409
def find_ids_across_trees(filenames, trees, require_versioned=True):
410
"""Find the ids corresponding to specified filenames.
412
All matches in all trees will be used, and all children of matched
413
directories will be used.
415
:param filenames: The filenames to find file_ids for (if None, returns
417
:param trees: The trees to find file_ids within
418
:param require_versioned: if true, all specified filenames must occur in
420
:return: a set of file ids for the specified filenames and their children.
424
specified_path_ids = _find_ids_across_trees(filenames, trees,
426
return _find_children_across_trees(specified_path_ids, trees)
427
# specified_ids = [id for path, id in _find_path_ids_across_trees(filenames, trees, require_versioned)]
428
# return _find_children_across_trees(specified_ids, trees)
430
def find_path_ids_across_trees(filenames, trees, require_versioned=True):
431
"""Find the paths and ids corresponding to specified filenames.
433
All matches in all trees will be used, and all children of matched
434
directories will be included
436
:param filenames: The filenames to find file_ids for
437
:param trees: The trees to find file_ids within
438
:param require_versioned: if true, all specified filenames must occur in
440
:return: a set of (path, file ids) for the specified filenames and their
441
children. The returned path is the path of the id in the first tree
442
that contains it. This matters when files have been moved
446
# This function needs to know the ids for filenames in all trees, then
447
# search for those same files and children in all the other trees.
448
# it is complicated by the same path in two trees being able to have
449
# different ids, which might both be present in both trees.
450
# consider two trees, which have had 'mv foo bar' and 'mv baz foo' done
451
# in this case, a diff of 'foo' should should changes to both the current
452
# 'bar' and the current 'foo' which was baz. Its arguable that if
453
# the situation is 'mv parent/foo bar' and 'mv baz parent/foo', that
454
# we should return the current bar and the current parent/foo' - at the
455
# moment we do, but we loop around all ids and all trees: I*T checks.
457
# Updating this algorithm to be fast in the common case:
458
# nothing has moved, all files have the same id in parent, child and there
459
# are only two trees (or one is working tree and the others are parents).
460
# walk the dirstate. as we find each path, gather the paths of that
461
# id in all trees. add a mapping from the id to the path in those trees.
462
# now lookup children by id, again in all trees; for these trees that
463
# nothing has moved in, the id->path mapping will allow us to find the
464
# parent trivially. To answer 'has anything been moved' in one of the
465
# dirstate parent trees though, we will need to stare harder at it.
467
# Now, given a path index, that is trivial for any one tree, and given
468
# that we can ask for additional data from a dirstate tree, its a single
469
# pass, though it will require scanning the entire tree to find paths
470
# that were at the current location.
471
# ideal results?: There are three things: tree, path, id. Pathologically
472
# we can have completely disjoint ids for each tree; but we cannot have
473
# disjoin paths for each tree, except if we scan each tree for the
474
# different ids from other trees.
476
specified_path_ids = _find_ids_across_trees(filenames, trees,
478
return _find_path_id_children_across_trees(specified_path_ids, trees)
481
def _find_ids_across_trees(filenames, trees, require_versioned):
482
"""Find the ids corresponding to specified filenames.
484
All matches in all trees will be used, but subdirectories are not scanned.
486
:param filenames: The filenames to find file_ids for
487
:param trees: The trees to find file_ids within
488
:param require_versioned: if true, all specified filenames must occur in
490
:return: a set of (path, file ids) for the specified filenames
493
interesting_ids = set()
494
for tree_path in filenames:
497
file_id = tree.path2id(tree_path)
498
if file_id is not None:
499
interesting_ids.add(file_id)
502
not_versioned.append(tree_path)
503
if len(not_versioned) > 0 and require_versioned:
504
raise errors.PathsNotVersionedError(not_versioned)
505
return interesting_ids
508
def _find_children_across_trees(specified_ids, trees):
509
"""Return a set including specified ids and their children
511
All matches in all trees will be used.
513
:param trees: The trees to find file_ids within
514
:return: a set containing all specified ids and their children
516
interesting_ids = set(specified_ids)
517
pending = interesting_ids
518
# now handle children of interesting ids
519
# we loop so that we handle all children of each id in both trees
520
while len(pending) > 0:
522
for file_id in pending:
524
if not tree.has_id(file_id):
526
entry = tree.inventory[file_id]
527
for child in getattr(entry, 'children', {}).itervalues():
528
if child.file_id not in interesting_ids:
529
new_pending.add(child.file_id)
530
interesting_ids.update(new_pending)
531
pending = new_pending
532
return interesting_ids
535
class InterTree(InterObject):
536
"""This class represents operations taking place between two Trees.
538
Its instances have methods like 'compare' and contain references to the
539
source and target trees these operations are to be carried out on.
541
clients of bzrlib should not need to use InterTree directly, rather they
542
should use the convenience methods on Tree such as 'Tree.compare()' which
543
will pass through to InterTree as appropriate.
549
def compare(self, want_unchanged=False, specific_files=None,
550
extra_trees=None, require_versioned=False, include_root=False):
551
"""Return the changes from source to target.
553
:return: A TreeDelta.
554
:param specific_files: An optional list of file paths to restrict the
555
comparison to. When mapping filenames to ids, all matches in all
556
trees (including optional extra_trees) are used, and all children of
557
matched directories are included.
558
:param want_unchanged: An optional boolean requesting the inclusion of
559
unchanged entries in the result.
560
:param extra_trees: An optional list of additional trees to use when
561
mapping the contents of specific_files (paths) to file_ids.
562
:param require_versioned: An optional boolean (defaults to False). When
563
supplied and True all the 'specific_files' must be versioned, or
564
a PathsNotVersionedError will be thrown.
566
# NB: show_status depends on being able to pass in non-versioned files
567
# and report them as unknown
568
trees = (self.source, self.target)
569
if extra_trees is not None:
570
trees = trees + tuple(extra_trees)
571
# target is usually the newer tree:
572
specific_file_ids = self.target.paths2ids(specific_files, [self.source],
573
require_versioned=require_versioned)
574
if specific_files and not specific_file_ids:
575
# All files are unversioned, so just return an empty delta
576
# _compare_trees would think we want a complete delta
577
return delta.TreeDelta()
578
return delta._compare_trees(self.source, self.target, want_unchanged,
579
specific_file_ids, include_root)
581
def _iter_changes(self, from_tree, to_tree, include_unchanged,
582
specific_file_ids, pb):
583
"""Generate an iterator of changes between trees.
586
(file_id, path, changed_content, versioned, parent, name, kind,
589
Path is relative to the to_tree. changed_content is True if the file's
590
content has changed. This includes changes to its kind, and to
593
versioned, parent, name, kind, executable are tuples of (from, to).
594
If a file is missing in a tree, its kind is None.
596
Iteration is done in parent-to-child order, relative to the to_tree.
599
from_entries_by_dir = list(from_tree.inventory.iter_entries_by_dir(
600
specific_file_ids=specific_file_ids))
601
from_data = dict((e.file_id, (p, e)) for p, e in from_entries_by_dir)
602
to_entries_by_dir = list(to_tree.inventory.iter_entries_by_dir(
603
specific_file_ids=specific_file_ids))
604
num_entries = len(from_entries_by_dir) + len(to_entries_by_dir)
606
for to_path, to_entry in to_entries_by_dir:
607
file_id = to_entry.file_id
608
to_paths[file_id] = to_path
610
changed_content = False
611
from_path, from_entry = from_data.get(file_id, (None, None))
612
from_versioned = (from_entry is not None)
613
if from_entry is not None:
614
from_versioned = True
615
from_name = from_entry.name
616
from_parent = from_entry.parent_id
617
from_kind, from_executable, from_stat = \
618
from_tree._comparison_data(from_entry, from_path)
621
from_versioned = False
625
from_executable = None
626
versioned = (from_versioned, True)
627
to_kind, to_executable, to_stat = \
628
to_tree._comparison_data(to_entry, to_path)
629
kind = (from_kind, to_kind)
630
if kind[0] != kind[1]:
631
changed_content = True
632
elif from_kind == 'file':
633
from_size = from_tree._file_size(from_entry, from_stat)
634
to_size = to_tree._file_size(to_entry, to_stat)
635
if from_size != to_size:
636
changed_content = True
637
elif (from_tree.get_file_sha1(file_id, from_path, from_stat) !=
638
to_tree.get_file_sha1(file_id, to_path, to_stat)):
639
changed_content = True
640
elif from_kind == 'symlink':
641
if (from_tree.get_symlink_target(file_id) !=
642
to_tree.get_symlink_target(file_id)):
643
changed_content = True
644
parent = (from_parent, to_entry.parent_id)
645
name = (from_name, to_entry.name)
646
executable = (from_executable, to_executable)
648
pb.update('comparing files', entry_count, num_entries)
649
if (changed_content is not False or versioned[0] != versioned[1]
650
or parent[0] != parent[1] or name[0] != name[1] or
651
executable[0] != executable[1] or include_unchanged):
652
yield (file_id, to_path, changed_content, versioned, parent,
653
name, kind, executable)
655
def get_to_path(from_entry):
656
if from_entry.parent_id is None:
659
if from_entry.parent_id not in to_paths:
660
get_to_path(from_tree.inventory[from_entry.parent_id])
661
to_path = osutils.pathjoin(to_paths[from_entry.parent_id],
663
to_paths[from_entry.file_id] = to_path
666
for path, from_entry in from_entries_by_dir:
667
file_id = from_entry.file_id
668
if file_id in to_paths:
670
to_path = get_to_path(from_entry)
673
pb.update('comparing files', entry_count, num_entries)
674
versioned = (True, False)
675
parent = (from_entry.parent_id, None)
676
name = (from_entry.name, None)
677
from_kind, from_executable, stat_value = \
678
from_tree._comparison_data(from_entry, path)
679
kind = (from_kind, None)
680
executable = (from_executable, None)
681
changed_content = True
682
# the parent's path is necessarily known at this point.
683
yield(file_id, to_path, changed_content, versioned, parent,
684
name, kind, executable)
687
# This was deprecated before 0.12, but did not have an official warning
688
@symbol_versioning.deprecated_function(symbol_versioning.zero_twelve)
689
def RevisionTree(*args, **kwargs):
690
"""RevisionTree has moved to bzrlib.revisiontree.RevisionTree()
692
Accessing it as bzrlib.tree.RevisionTree has been deprecated as of
695
from bzrlib.revisiontree import RevisionTree as _RevisionTree
696
return _RevisionTree(*args, **kwargs)
added
removed
bzrlib/tree.py
