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_files=None, pb=None, extra_trees=None,
92
require_versioned=True):
93
intertree = InterTree.get(from_tree, self)
94
return intertree._iter_changes(include_unchanged, specific_files, pb,
95
extra_trees, require_versioned)
98
"""Get a list of the conflicts in the tree.
100
Each conflict is an instance of bzrlib.conflicts.Conflict.
104
def get_parent_ids(self):
105
"""Get the parent ids for this tree.
107
:return: a list of parent ids. [] is returned to indicate
108
a tree with no parents.
109
:raises: BzrError if the parents are not known.
111
raise NotImplementedError(self.get_parent_ids)
113
def has_filename(self, filename):
114
"""True if the tree has given filename."""
115
raise NotImplementedError()
117
def has_id(self, file_id):
118
file_id = osutils.safe_file_id(file_id)
119
return self.inventory.has_id(file_id)
121
__contains__ = has_id
123
def has_or_had_id(self, file_id):
124
file_id = osutils.safe_file_id(file_id)
125
if file_id == self.inventory.root.file_id:
127
return self.inventory.has_id(file_id)
130
return iter(self.inventory)
132
def id2path(self, file_id):
133
file_id = osutils.safe_file_id(file_id)
134
return self.inventory.id2path(file_id)
136
def is_control_filename(self, filename):
137
"""True if filename is the name of a control file in this tree.
139
:param filename: A filename within the tree. This is a relative path
140
from the root of this tree.
142
This is true IF and ONLY IF the filename is part of the meta data
143
that bzr controls in this tree. I.E. a random .bzr directory placed
144
on disk will not be a control file for this tree.
146
return self.bzrdir.is_control_filename(filename)
149
def iter_entries_by_dir(self, specific_file_ids=None):
150
"""Walk the tree in 'by_dir' order.
152
This will yield each entry in the tree as a (path, entry) tuple. The
153
order that they are yielded is: the contents of a directory are
154
preceeded by the parent of a directory, and all the contents of a
155
directory are grouped together.
157
return self.inventory.iter_entries_by_dir(
158
specific_file_ids=specific_file_ids)
160
def iter_reference_entries(self):
161
for path, entry in self.iter_entries_by_dir():
162
if entry.kind == 'tree-reference':
165
def kind(self, file_id):
166
raise NotImplementedError("Tree subclass %s must implement kind"
167
% self.__class__.__name__)
169
def get_reference_revision(self, entry, path=None):
170
raise NotImplementedError("Tree subclass %s must implement "
171
"get_reference_revision"
172
% self.__class__.__name__)
174
def _comparison_data(self, entry, path):
175
"""Return a tuple of kind, executable, stat_value for a file.
177
entry may be None if there is no inventory entry for the file, but
178
path must always be supplied.
180
kind is None if there is no file present (even if an inventory id is
181
present). executable is False for non-file entries.
183
raise NotImplementedError(self._comparison_data)
185
def _file_size(self, entry, stat_value):
186
raise NotImplementedError(self._file_size)
188
def _get_inventory(self):
189
return self._inventory
191
def get_file(self, file_id):
192
"""Return a file object for the file file_id in the tree."""
193
raise NotImplementedError(self.get_file)
195
def get_file_mtime(self, file_id, path=None):
196
"""Return the modification time for a file.
198
:param file_id: The handle for this file.
199
:param path: The path that this file can be found at.
200
These must point to the same object.
202
raise NotImplementedError(self.get_file_mtime)
204
def get_file_by_path(self, path):
205
return self.get_file(self._inventory.path2id(path))
207
def get_symlink_target(self, file_id):
208
"""Get the target for a given file_id.
210
It is assumed that the caller already knows that file_id is referencing
212
:param file_id: Handle for the symlink entry.
213
:return: The path the symlink points to.
215
raise NotImplementedError(self.get_symlink_target)
217
def annotate_iter(self, file_id):
218
"""Return an iterator of revision_id, line tuples
220
For working trees (and mutable trees in general), the special
221
revision_id 'current:' will be used for lines that are new in this
222
tree, e.g. uncommitted changes.
223
:param file_id: The file to produce an annotated version from
225
raise NotImplementedError(self.annotate_iter)
227
inventory = property(_get_inventory,
228
doc="Inventory of this Tree")
230
def _check_retrieved(self, ie, f):
233
fp = fingerprint_file(f)
236
if ie.text_size is not None:
237
if ie.text_size != fp['size']:
238
raise BzrError("mismatched size for file %r in %r" % (ie.file_id, self._store),
239
["inventory expects %d bytes" % ie.text_size,
240
"file is actually %d bytes" % fp['size'],
241
"store is probably damaged/corrupt"])
243
if ie.text_sha1 != fp['sha1']:
244
raise BzrError("wrong SHA-1 for file %r in %r" % (ie.file_id, self._store),
245
["inventory expects %s" % ie.text_sha1,
246
"file is actually %s" % fp['sha1'],
247
"store is probably damaged/corrupt"])
249
def path2id(self, path):
250
"""Return the id for path in this tree."""
251
return self._inventory.path2id(path)
253
def paths2ids(self, paths, trees=[], require_versioned=True):
254
"""Return all the ids that can be reached by walking from paths.
256
Each path is looked up in each this tree and any extras provided in
257
trees, and this is repeated recursively: the children in an extra tree
258
of a directory that has been renamed under a provided path in this tree
259
are all returned, even if none exist until a provided path in this
260
tree, and vice versa.
262
:param paths: An iterable of paths to start converting to ids from.
263
Alternatively, if paths is None, no ids should be calculated and None
264
will be returned. This is offered to make calling the api unconditional
265
for code that *might* take a list of files.
266
:param trees: Additional trees to consider.
267
:param require_versioned: If False, do not raise NotVersionedError if
268
an element of paths is not versioned in this tree and all of trees.
270
return find_ids_across_trees(paths, [self] + list(trees), require_versioned)
272
def print_file(self, file_id):
273
"""Print file with id `file_id` to stdout."""
274
file_id = osutils.safe_file_id(file_id)
276
sys.stdout.write(self.get_file_text(file_id))
281
def revision_tree(self, revision_id):
282
"""Obtain a revision tree for the revision revision_id.
284
The intention of this method is to allow access to possibly cached
285
tree data. Implementors of this method should raise NoSuchRevision if
286
the tree is not locally available, even if they could obtain the
287
tree via a repository or some other means. Callers are responsible
288
for finding the ultimate source for a revision tree.
290
:param revision_id: The revision_id of the requested tree.
292
:raises: NoSuchRevision if the tree cannot be obtained.
294
raise errors.NoSuchRevisionInTree(self, revision_id)
297
"""What files are present in this tree and unknown.
299
:return: an iterator over the unknown files.
306
def filter_unversioned_files(self, paths):
307
"""Filter out paths that are versioned.
309
:return: set of paths.
311
# NB: we specifically *don't* call self.has_filename, because for
312
# WorkingTrees that can indicate files that exist on disk but that
314
pred = self.inventory.has_filename
315
return set((p for p in paths if not pred(p)))
317
def walkdirs(self, prefix=""):
318
"""Walk the contents of this tree from path down.
320
This yields all the data about the contents of a directory at a time.
321
After each directory has been yielded, if the caller has mutated the
322
list to exclude some directories, they are then not descended into.
324
The data yielded is of the form:
325
((directory-relpath, directory-path-from-root, directory-fileid),
326
[(relpath, basename, kind, lstat, path_from_tree_root, file_id,
327
versioned_kind), ...]),
328
- directory-relpath is the containing dirs relpath from prefix
329
- directory-path-from-root is the containing dirs path from /
330
- directory-fileid is the id of the directory if it is versioned.
331
- relpath is the relative path within the subtree being walked.
332
- basename is the basename
333
- kind is the kind of the file now. If unknonwn then the file is not
334
present within the tree - but it may be recorded as versioned. See
336
- lstat is the stat data *if* the file was statted.
337
- path_from_tree_root is the path from the root of the tree.
338
- file_id is the file_id is the entry is versioned.
339
- versioned_kind is the kind of the file as last recorded in the
340
versioning system. If 'unknown' the file is not versioned.
341
One of 'kind' and 'versioned_kind' must not be 'unknown'.
343
:param prefix: Start walking from prefix within the tree rather than
344
at the root. This allows one to walk a subtree but get paths that are
345
relative to a tree rooted higher up.
346
:return: an iterator over the directory data.
348
raise NotImplementedError(self.walkdirs)
351
class EmptyTree(Tree):
354
self._inventory = Inventory(root_id=None)
355
symbol_versioning.warn('EmptyTree is deprecated as of bzr 0.9 please'
356
' use repository.revision_tree instead.',
357
DeprecationWarning, stacklevel=2)
359
def get_parent_ids(self):
362
def get_symlink_target(self, file_id):
365
def has_filename(self, filename):
368
def kind(self, file_id):
369
file_id = osutils.safe_file_id(file_id)
370
assert self._inventory[file_id].kind == "directory"
373
def list_files(self, include_root=False):
376
def __contains__(self, file_id):
377
file_id = osutils.safe_file_id(file_id)
378
return (file_id in self._inventory)
380
def get_file_sha1(self, file_id, path=None, stat_value=None):
384
######################################################################
387
# TODO: Merge these two functions into a single one that can operate
388
# on either a whole tree or a set of files.
390
# TODO: Return the diff in order by filename, not by category or in
391
# random order. Can probably be done by lock-stepping through the
392
# filenames from both trees.
395
def file_status(filename, old_tree, new_tree):
396
"""Return single-letter status, old and new names for a file.
398
The complexity here is in deciding how to represent renames;
399
many complex cases are possible.
401
old_inv = old_tree.inventory
402
new_inv = new_tree.inventory
403
new_id = new_inv.path2id(filename)
404
old_id = old_inv.path2id(filename)
406
if not new_id and not old_id:
407
# easy: doesn't exist in either; not versioned at all
408
if new_tree.is_ignored(filename):
409
return 'I', None, None
411
return '?', None, None
413
# There is now a file of this name, great.
416
# There is no longer a file of this name, but we can describe
417
# what happened to the file that used to have
418
# this name. There are two possibilities: either it was
419
# deleted entirely, or renamed.
421
if new_inv.has_id(old_id):
422
return 'X', old_inv.id2path(old_id), new_inv.id2path(old_id)
424
return 'D', old_inv.id2path(old_id), None
426
# if the file_id is new in this revision, it is added
427
if new_id and not old_inv.has_id(new_id):
430
# if there used to be a file of this name, but that ID has now
431
# disappeared, it is deleted
432
if old_id and not new_inv.has_id(old_id):
439
def find_renames(old_inv, new_inv):
440
for file_id in old_inv:
441
if file_id not in new_inv:
443
old_name = old_inv.id2path(file_id)
444
new_name = new_inv.id2path(file_id)
445
if old_name != new_name:
446
yield (old_name, new_name)
449
def find_ids_across_trees(filenames, trees, require_versioned=True):
450
"""Find the ids corresponding to specified filenames.
452
All matches in all trees will be used, and all children of matched
453
directories will be used.
455
:param filenames: The filenames to find file_ids for (if None, returns
457
:param trees: The trees to find file_ids within
458
:param require_versioned: if true, all specified filenames must occur in
460
:return: a set of file ids for the specified filenames and their children.
464
specified_path_ids = _find_ids_across_trees(filenames, trees,
466
return _find_children_across_trees(specified_path_ids, trees)
469
def _find_ids_across_trees(filenames, trees, require_versioned):
470
"""Find the ids corresponding to specified filenames.
472
All matches in all trees will be used, but subdirectories are not scanned.
474
:param filenames: The filenames to find file_ids for
475
:param trees: The trees to find file_ids within
476
:param require_versioned: if true, all specified filenames must occur in
478
:return: a set of (path, file ids) for the specified filenames
481
interesting_ids = set()
482
for tree_path in filenames:
485
file_id = tree.path2id(tree_path)
486
if file_id is not None:
487
interesting_ids.add(file_id)
490
not_versioned.append(tree_path)
491
if len(not_versioned) > 0 and require_versioned:
492
raise errors.PathsNotVersionedError(not_versioned)
493
return interesting_ids
496
def _find_children_across_trees(specified_ids, trees):
497
"""Return a set including specified ids and their children
499
All matches in all trees will be used.
501
:param trees: The trees to find file_ids within
502
:return: a set containing all specified ids and their children
504
interesting_ids = set(specified_ids)
505
pending = interesting_ids
506
# now handle children of interesting ids
507
# we loop so that we handle all children of each id in both trees
508
while len(pending) > 0:
510
for file_id in pending:
512
if not tree.has_id(file_id):
514
entry = tree.inventory[file_id]
515
for child in getattr(entry, 'children', {}).itervalues():
516
if child.file_id not in interesting_ids:
517
new_pending.add(child.file_id)
518
interesting_ids.update(new_pending)
519
pending = new_pending
520
return interesting_ids
523
class InterTree(InterObject):
524
"""This class represents operations taking place between two Trees.
526
Its instances have methods like 'compare' and contain references to the
527
source and target trees these operations are to be carried out on.
529
clients of bzrlib should not need to use InterTree directly, rather they
530
should use the convenience methods on Tree such as 'Tree.compare()' which
531
will pass through to InterTree as appropriate.
537
def compare(self, want_unchanged=False, specific_files=None,
538
extra_trees=None, require_versioned=False, include_root=False):
539
"""Return the changes from source to target.
541
:return: A TreeDelta.
542
:param specific_files: An optional list of file paths to restrict the
543
comparison to. When mapping filenames to ids, all matches in all
544
trees (including optional extra_trees) are used, and all children of
545
matched directories are included.
546
:param want_unchanged: An optional boolean requesting the inclusion of
547
unchanged entries in the result.
548
:param extra_trees: An optional list of additional trees to use when
549
mapping the contents of specific_files (paths) to file_ids.
550
:param require_versioned: An optional boolean (defaults to False). When
551
supplied and True all the 'specific_files' must be versioned, or
552
a PathsNotVersionedError will be thrown.
554
# NB: show_status depends on being able to pass in non-versioned files
555
# and report them as unknown
556
trees = (self.source,)
557
if extra_trees is not None:
558
trees = trees + tuple(extra_trees)
559
# target is usually the newer tree:
560
specific_file_ids = self.target.paths2ids(specific_files, trees,
561
require_versioned=require_versioned)
562
if specific_files and not specific_file_ids:
563
# All files are unversioned, so just return an empty delta
564
# _compare_trees would think we want a complete delta
565
return delta.TreeDelta()
566
return delta._compare_trees(self.source, self.target, want_unchanged,
567
specific_files, include_root, extra_trees=extra_trees)
569
def _iter_changes(self, include_unchanged=False,
570
specific_files=None, pb=None, extra_trees=[],
571
require_versioned=True):
572
"""Generate an iterator of changes between trees.
575
(file_id, path, changed_content, versioned, parent, name, kind,
578
Path is relative to the target tree. changed_content is True if the
579
file's content has changed. This includes changes to its kind, and to
582
versioned, parent, name, kind, executable are tuples of (from, to).
583
If a file is missing in a tree, its kind is None.
585
Iteration is done in parent-to-child order, relative to the target
588
There is no guarantee that all paths are in sorted order: the
589
requirement to expand the search due to renames may result in children
590
that should be found early being found late in the search, after
591
lexically later results have been returned.
592
:param require_versioned: Raise errors.PathsNotVersionedError if a
593
path in the specific_files list is not versioned in one of
594
source, target or extra_trees.
596
# this must return a sequence rather than a list so that it can hold a
597
# read-lock for the whole time.
599
# TODO: this really only needs to lock the trees not the branches, so
600
# could do with lock_tree_read() -- mbp 20070227
602
self.source.lock_read()
603
self.target.lock_read()
605
lookup_trees = [self.source]
607
lookup_trees.extend(extra_trees)
608
specific_file_ids = self.target.paths2ids(specific_files,
609
lookup_trees, require_versioned=require_versioned)
611
from_entries_by_dir = list(self.source.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(self.target.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
self.source._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
self.target._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 = self.source._file_size(from_entry, from_stat)
646
to_size = self.target._file_size(to_entry, to_stat)
647
if from_size != to_size:
648
changed_content = True
649
elif (self.source.get_file_sha1(file_id, from_path, from_stat) !=
650
self.target.get_file_sha1(file_id, to_path, to_stat)):
651
changed_content = True
652
elif from_kind == 'symlink':
653
if (self.source.get_symlink_target(file_id) !=
654
self.target.get_symlink_target(file_id)):
655
changed_content = True
656
elif from_kind == 'tree-reference':
657
if (self.source.get_reference_revision(from_entry, from_path)
658
!= self.target.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
result.append((file_id, to_path, changed_content, versioned, parent,
669
name, kind, executable))
670
def get_to_path(from_entry):
671
if from_entry.parent_id is None:
674
if from_entry.parent_id not in to_paths:
675
get_to_path(self.source.inventory[from_entry.parent_id])
676
to_path = osutils.pathjoin(to_paths[from_entry.parent_id],
678
to_paths[from_entry.file_id] = to_path
681
for path, from_entry in from_entries_by_dir:
682
file_id = from_entry.file_id
683
if file_id in to_paths:
685
to_path = get_to_path(from_entry)
688
pb.update('comparing files', entry_count, num_entries)
689
versioned = (True, False)
690
parent = (from_entry.parent_id, None)
691
name = (from_entry.name, None)
692
from_kind, from_executable, stat_value = \
693
self.source._comparison_data(from_entry, path)
694
kind = (from_kind, None)
695
executable = (from_executable, None)
696
changed_content = True
697
# the parent's path is necessarily known at this point.
698
result.append((file_id, to_path, changed_content, versioned, parent,
699
name, kind, executable))
706
# This was deprecated before 0.12, but did not have an official warning
707
@symbol_versioning.deprecated_function(symbol_versioning.zero_twelve)
708
def RevisionTree(*args, **kwargs):
709
"""RevisionTree has moved to bzrlib.revisiontree.RevisionTree()
711
Accessing it as bzrlib.tree.RevisionTree has been deprecated as of
714
from bzrlib.revisiontree import RevisionTree as _RevisionTree
715
return _RevisionTree(*args, **kwargs)
added
removed
bzrlib/tree.py
