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)
144
def iter_entries_by_dir(self, specific_file_ids=None):
145
"""Walk the tree in 'by_dir' order.
147
This will yield each entry in the tree as a (path, entry) tuple. The
148
order that they are yielded is: the contents of a directory are
149
preceeded by the parent of a directory, and all the contents of a
150
directory are grouped together.
152
return self.inventory.iter_entries_by_dir(
153
specific_file_ids=specific_file_ids)
155
def kind(self, file_id):
156
raise NotImplementedError("subclasses must implement kind")
158
def _comparison_data(self, entry, path):
159
"""Return a tuple of kind, executable, stat_value for a file.
161
entry may be None if there is no inventory entry for the file, but
162
path must always be supplied.
164
kind is None if there is no file present (even if an inventory id is
165
present). executable is False for non-file entries.
167
raise NotImplementedError(self._comparison_data)
169
def _file_size(entry, stat_value):
170
raise NotImplementedError(self._file_size)
172
def _get_inventory(self):
173
return self._inventory
175
def get_file(self, file_id):
176
"""Return a file object for the file file_id in the tree."""
177
raise NotImplementedError(self.get_file)
179
def get_file_by_path(self, path):
180
return self.get_file(self._inventory.path2id(path))
182
def annotate_iter(self, file_id):
183
"""Return an iterator of revision_id, line tuples
185
For working trees (and mutable trees in general), the special
186
revision_id 'current:' will be used for lines that are new in this
187
tree, e.g. uncommitted changes.
188
:param file_id: The file to produce an annotated version from
190
raise NotImplementedError(self.annotate_iter)
192
inventory = property(_get_inventory,
193
doc="Inventory of this Tree")
195
def _check_retrieved(self, ie, f):
198
fp = fingerprint_file(f)
201
if ie.text_size is not None:
202
if ie.text_size != fp['size']:
203
raise BzrError("mismatched size for file %r in %r" % (ie.file_id, self._store),
204
["inventory expects %d bytes" % ie.text_size,
205
"file is actually %d bytes" % fp['size'],
206
"store is probably damaged/corrupt"])
208
if ie.text_sha1 != fp['sha1']:
209
raise BzrError("wrong SHA-1 for file %r in %r" % (ie.file_id, self._store),
210
["inventory expects %s" % ie.text_sha1,
211
"file is actually %s" % fp['sha1'],
212
"store is probably damaged/corrupt"])
214
def path2id(self, path):
215
"""Return the id for path in this tree."""
216
return self._inventory.path2id(path)
218
def print_file(self, file_id):
219
"""Print file with id `file_id` to stdout."""
221
sys.stdout.write(self.get_file_text(file_id))
226
def revision_tree(self, revision_id):
227
"""Obtain a revision tree for the revision revision_id.
229
The intention of this method is to allow access to possibly cached
230
tree data. Implementors of this method should raise NoSuchRevision if
231
the tree is not locally available, even if they could obtain the
232
tree via a repository or some other means. Callers are responsible
233
for finding the ultimate source for a revision tree.
235
:param revision_id: The revision_id of the requested tree.
237
:raises: NoSuchRevision if the tree cannot be obtained.
239
raise errors.NoSuchRevisionInTree(self, revision_id)
242
"""What files are present in this tree and unknown.
244
:return: an iterator over the unknown files.
251
def filter_unversioned_files(self, paths):
252
"""Filter out paths that are not versioned.
254
:return: set of paths.
256
# NB: we specifically *don't* call self.has_filename, because for
257
# WorkingTrees that can indicate files that exist on disk but that
259
pred = self.inventory.has_filename
260
return set((p for p in paths if not pred(p)))
262
def walkdirs(self, prefix=""):
263
"""Walk the contents of this tree from path down.
265
This yields all the data about the contents of a directory at a time.
266
After each directory has been yielded, if the caller has mutated the
267
list to exclude some directories, they are then not descended into.
269
The data yielded is of the form:
270
((directory-relpath, directory-path-from-root, directory-fileid),
271
[(relpath, basename, kind, lstat, path_from_tree_root, file_id,
272
versioned_kind), ...]),
273
- directory-relpath is the containing dirs relpath from prefix
274
- directory-path-from-root is the containing dirs path from /
275
- directory-fileid is the id of the directory if it is versioned.
276
- relpath is the relative path within the subtree being walked.
277
- basename is the basename
278
- kind is the kind of the file now. If unknonwn then the file is not
279
present within the tree - but it may be recorded as versioned. See
281
- lstat is the stat data *if* the file was statted.
282
- path_from_tree_root is the path from the root of the tree.
283
- file_id is the file_id is the entry is versioned.
284
- versioned_kind is the kind of the file as last recorded in the
285
versioning system. If 'unknown' the file is not versioned.
286
One of 'kind' and 'versioned_kind' must not be 'unknown'.
288
:param prefix: Start walking from prefix within the tree rather than
289
at the root. This allows one to walk a subtree but get paths that are
290
relative to a tree rooted higher up.
291
:return: an iterator over the directory data.
293
raise NotImplementedError(self.walkdirs)
296
class EmptyTree(Tree):
299
self._inventory = Inventory(root_id=None)
300
symbol_versioning.warn('EmptyTree is deprecated as of bzr 0.9 please'
301
' use repository.revision_tree instead.',
302
DeprecationWarning, stacklevel=2)
304
def get_parent_ids(self):
307
def get_symlink_target(self, file_id):
310
def has_filename(self, filename):
313
def kind(self, file_id):
314
assert self._inventory[file_id].kind == "directory"
317
def list_files(self, include_root=False):
320
def __contains__(self, file_id):
321
return (file_id in self._inventory)
323
def get_file_sha1(self, file_id, path=None, stat_value=None):
327
######################################################################
330
# TODO: Merge these two functions into a single one that can operate
331
# on either a whole tree or a set of files.
333
# TODO: Return the diff in order by filename, not by category or in
334
# random order. Can probably be done by lock-stepping through the
335
# filenames from both trees.
338
def file_status(filename, old_tree, new_tree):
339
"""Return single-letter status, old and new names for a file.
341
The complexity here is in deciding how to represent renames;
342
many complex cases are possible.
344
old_inv = old_tree.inventory
345
new_inv = new_tree.inventory
346
new_id = new_inv.path2id(filename)
347
old_id = old_inv.path2id(filename)
349
if not new_id and not old_id:
350
# easy: doesn't exist in either; not versioned at all
351
if new_tree.is_ignored(filename):
352
return 'I', None, None
354
return '?', None, None
356
# There is now a file of this name, great.
359
# There is no longer a file of this name, but we can describe
360
# what happened to the file that used to have
361
# this name. There are two possibilities: either it was
362
# deleted entirely, or renamed.
364
if new_inv.has_id(old_id):
365
return 'X', old_inv.id2path(old_id), new_inv.id2path(old_id)
367
return 'D', old_inv.id2path(old_id), None
369
# if the file_id is new in this revision, it is added
370
if new_id and not old_inv.has_id(new_id):
373
# if there used to be a file of this name, but that ID has now
374
# disappeared, it is deleted
375
if old_id and not new_inv.has_id(old_id):
382
def find_renames(old_inv, new_inv):
383
for file_id in old_inv:
384
if file_id not in new_inv:
386
old_name = old_inv.id2path(file_id)
387
new_name = new_inv.id2path(file_id)
388
if old_name != new_name:
389
yield (old_name, new_name)
392
def find_ids_across_trees(filenames, trees, require_versioned=True):
393
"""Find the ids corresponding to specified filenames.
395
All matches in all trees will be used, and all children of matched
396
directories will be used.
398
:param filenames: The filenames to find file_ids for
399
:param trees: The trees to find file_ids within
400
:param require_versioned: if true, all specified filenames must occur in
402
:return: a set of file ids for the specified filenames and their children.
406
specified_ids = _find_filename_ids_across_trees(filenames, trees,
408
return _find_children_across_trees(specified_ids, trees)
411
def _find_filename_ids_across_trees(filenames, trees, require_versioned):
412
"""Find the ids corresponding to specified filenames.
414
All matches in all trees will be used.
416
:param filenames: The filenames to find file_ids for
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
423
interesting_ids = set()
424
for tree_path in filenames:
427
file_id = tree.inventory.path2id(tree_path)
428
if file_id is not None:
429
interesting_ids.add(file_id)
432
not_versioned.append(tree_path)
433
if len(not_versioned) > 0 and require_versioned:
434
raise errors.PathsNotVersionedError(not_versioned)
435
return interesting_ids
438
def _find_children_across_trees(specified_ids, trees):
439
"""Return a set including specified ids and their children
441
All matches in all trees will be used.
443
:param trees: The trees to find file_ids within
444
:return: a set containing all specified ids and their children
446
interesting_ids = set(specified_ids)
447
pending = interesting_ids
448
# now handle children of interesting ids
449
# we loop so that we handle all children of each id in both trees
450
while len(pending) > 0:
452
for file_id in pending:
454
if file_id not in tree:
456
entry = tree.inventory[file_id]
457
for child in getattr(entry, 'children', {}).itervalues():
458
if child.file_id not in interesting_ids:
459
new_pending.add(child.file_id)
460
interesting_ids.update(new_pending)
461
pending = new_pending
462
return interesting_ids
465
class InterTree(InterObject):
466
"""This class represents operations taking place between two Trees.
468
Its instances have methods like 'compare' and contain references to the
469
source and target trees these operations are to be carried out on.
471
clients of bzrlib should not need to use InterTree directly, rather they
472
should use the convenience methods on Tree such as 'Tree.compare()' which
473
will pass through to InterTree as appropriate.
479
def compare(self, want_unchanged=False, specific_files=None,
480
extra_trees=None, require_versioned=False, include_root=False):
481
"""Return the changes from source to target.
483
:return: A TreeDelta.
484
:param specific_files: An optional list of file paths to restrict the
485
comparison to. When mapping filenames to ids, all matches in all
486
trees (including optional extra_trees) are used, and all children of
487
matched directories are included.
488
:param want_unchanged: An optional boolean requesting the inclusion of
489
unchanged entries in the result.
490
:param extra_trees: An optional list of additional trees to use when
491
mapping the contents of specific_files (paths) to file_ids.
492
:param require_versioned: An optional boolean (defaults to False). When
493
supplied and True all the 'specific_files' must be versioned, or
494
a PathsNotVersionedError will be thrown.
496
# NB: show_status depends on being able to pass in non-versioned files and
497
# report them as unknown
498
trees = (self.source, self.target)
499
if extra_trees is not None:
500
trees = trees + tuple(extra_trees)
501
specific_file_ids = find_ids_across_trees(specific_files,
502
trees, require_versioned=require_versioned)
503
if specific_files and not specific_file_ids:
504
# All files are unversioned, so just return an empty delta
505
# _compare_trees would think we want a complete delta
506
return delta.TreeDelta()
507
return delta._compare_trees(self.source, self.target, want_unchanged,
508
specific_file_ids, include_root)
510
def _iter_changes(self, from_tree, to_tree, include_unchanged,
511
specific_file_ids, pb):
512
"""Generate an iterator of changes between trees.
515
(file_id, path, changed_content, versioned, parent, name, kind,
518
Path is relative to the to_tree. changed_content is True if the file's
519
content has changed. This includes changes to its kind, and to
522
versioned, parent, name, kind, executable are tuples of (from, to).
523
If a file is missing in a tree, its kind is None.
525
Iteration is done in parent-to-child order, relative to the to_tree.
528
from_entries_by_dir = list(from_tree.inventory.iter_entries_by_dir(
529
specific_file_ids=specific_file_ids))
530
from_data = dict((e.file_id, (p, e)) for p, e in from_entries_by_dir)
531
to_entries_by_dir = list(to_tree.inventory.iter_entries_by_dir(
532
specific_file_ids=specific_file_ids))
533
num_entries = len(from_entries_by_dir) + len(to_entries_by_dir)
535
for to_path, to_entry in to_entries_by_dir:
536
file_id = to_entry.file_id
537
to_paths[file_id] = to_path
539
changed_content = False
540
from_path, from_entry = from_data.get(file_id, (None, None))
541
from_versioned = (from_entry is not None)
542
if from_entry is not None:
543
from_versioned = True
544
from_name = from_entry.name
545
from_parent = from_entry.parent_id
546
from_kind, from_executable, from_stat = \
547
from_tree._comparison_data(from_entry, from_path)
550
from_versioned = False
554
from_executable = None
555
versioned = (from_versioned, True)
556
to_kind, to_executable, to_stat = \
557
to_tree._comparison_data(to_entry, to_path)
558
kind = (from_kind, to_kind)
559
if kind[0] != kind[1]:
560
changed_content = True
561
elif from_kind == 'file':
562
from_size = from_tree._file_size(from_entry, from_stat)
563
to_size = to_tree._file_size(to_entry, to_stat)
564
if from_size != to_size:
565
changed_content = True
566
elif (from_tree.get_file_sha1(file_id, from_path, from_stat) !=
567
to_tree.get_file_sha1(file_id, to_path, to_stat)):
568
changed_content = True
569
elif from_kind == 'symlink':
570
if (from_tree.get_symlink_target(file_id) !=
571
to_tree.get_symlink_target(file_id)):
572
changed_content = True
573
parent = (from_parent, to_entry.parent_id)
574
name = (from_name, to_entry.name)
575
executable = (from_executable, to_executable)
577
pb.update('comparing files', entry_count, num_entries)
578
if (changed_content is not False or versioned[0] != versioned[1]
579
or parent[0] != parent[1] or name[0] != name[1] or
580
executable[0] != executable[1] or include_unchanged):
581
yield (file_id, to_path, changed_content, versioned, parent,
582
name, kind, executable)
584
def get_to_path(from_entry):
585
if from_entry.parent_id is None:
588
if from_entry.parent_id not in to_paths:
589
get_to_path(from_tree.inventory[from_entry.parent_id])
590
to_path = osutils.pathjoin(to_paths[from_entry.parent_id],
592
to_paths[from_entry.file_id] = to_path
595
for path, from_entry in from_entries_by_dir:
596
file_id = from_entry.file_id
597
if file_id in to_paths:
599
to_path = get_to_path(from_entry)
602
pb.update('comparing files', entry_count, num_entries)
603
versioned = (True, False)
604
parent = (from_entry.parent_id, None)
605
name = (from_entry.name, None)
606
from_kind, from_executable, stat_value = \
607
from_tree._comparison_data(from_entry, path)
608
kind = (from_kind, None)
609
executable = (from_executable, None)
610
changed_content = True
611
# the parent's path is necessarily known at this point.
612
yield(file_id, to_path, changed_content, versioned, parent,
613
name, kind, executable)
616
# This was deprecated before 0.12, but did not have an official warning
617
@symbol_versioning.deprecated_function(symbol_versioning.zero_twelve)
618
def RevisionTree(*args, **kwargs):
619
"""RevisionTree has moved to bzrlib.revisiontree.RevisionTree()
621
Accessing it as bzrlib.tree.RevisionTree has been deprecated as of
624
from bzrlib.revisiontree import RevisionTree as _RevisionTree
625
return _RevisionTree(*args, **kwargs)
added
removed
bzrlib/tree.py
