1
# (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
# FIXME: This refactoring of the workingtree code doesn't seem to keep
18
# the WorkingTree's copy of the inventory in sync with the branch. The
19
# branch modifies its working inventory when it does a commit to make
20
# missing files permanently removed.
22
# TODO: Maybe also keep the full path of the entry, and the children?
23
# But those depend on its position within a particular inventory, and
24
# it would be nice not to need to hold the backpointer here.
26
# This should really be an id randomly assigned when the tree is
27
# created, but it's not for now.
38
from bzrlib.osutils import (pumpfile, quotefn, splitpath, joinpath,
39
pathjoin, sha_strings)
40
from bzrlib.errors import (NotVersionedError, InvalidEntryName,
41
BzrError, BzrCheckError, BinaryFile)
42
from bzrlib.trace import mutter
45
class InventoryEntry(object):
46
"""Description of a versioned file.
48
An InventoryEntry has the following fields, which are also
49
present in the XML inventory-entry element:
54
(within the parent directory)
57
file_id of the parent directory, or ROOT_ID
60
the revision_id in which this variation of this file was
64
Indicates that this file should be executable on systems
68
sha-1 of the text of the file
71
size in bytes of the text of the file
73
(reading a version 4 tree created a text_id field.)
78
>>> i.add(InventoryDirectory('123', 'src', ROOT_ID))
79
InventoryDirectory('123', 'src', parent_id='TREE_ROOT')
80
>>> i.add(InventoryFile('2323', 'hello.c', parent_id='123'))
81
InventoryFile('2323', 'hello.c', parent_id='123')
82
>>> shouldbe = {0: 'src', 1: pathjoin('src','hello.c')}
83
>>> for ix, j in enumerate(i.iter_entries()):
84
... print (j[0] == shouldbe[ix], j[1])
86
(True, InventoryDirectory('123', 'src', parent_id='TREE_ROOT'))
87
(True, InventoryFile('2323', 'hello.c', parent_id='123'))
88
>>> i.add(InventoryFile('2323', 'bye.c', '123'))
89
Traceback (most recent call last):
91
BzrError: inventory already contains entry with id {2323}
92
>>> i.add(InventoryFile('2324', 'bye.c', '123'))
93
InventoryFile('2324', 'bye.c', parent_id='123')
94
>>> i.add(InventoryDirectory('2325', 'wibble', '123'))
95
InventoryDirectory('2325', 'wibble', parent_id='123')
96
>>> i.path2id('src/wibble')
100
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
101
InventoryFile('2326', 'wibble.c', parent_id='2325')
103
InventoryFile('2326', 'wibble.c', parent_id='2325')
104
>>> for path, entry in i.iter_entries():
106
... assert i.path2id(path)
113
>>> i.id2path('2326')
114
'src/wibble/wibble.c'
117
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
118
'text_id', 'parent_id', 'children', 'executable',
121
def _add_text_to_weave(self, new_lines, parents, weave_store, transaction):
122
versionedfile = weave_store.get_weave_or_empty(self.file_id,
124
versionedfile.add_lines(self.revision, parents, new_lines)
126
def detect_changes(self, old_entry):
127
"""Return a (text_modified, meta_modified) from this to old_entry.
129
_read_tree_state must have been called on self and old_entry prior to
130
calling detect_changes.
134
def diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
135
output_to, reverse=False):
136
"""Perform a diff from this to to_entry.
138
text_diff will be used for textual difference calculation.
139
This is a template method, override _diff in child classes.
141
self._read_tree_state(tree.id2path(self.file_id), tree)
143
# cannot diff from one kind to another - you must do a removal
144
# and an addif they do not match.
145
assert self.kind == to_entry.kind
146
to_entry._read_tree_state(to_tree.id2path(to_entry.file_id),
148
self._diff(text_diff, from_label, tree, to_label, to_entry, to_tree,
151
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
152
output_to, reverse=False):
153
"""Perform a diff between two entries of the same kind."""
155
def find_previous_heads(self, previous_inventories,
156
versioned_file_store,
159
"""Return the revisions and entries that directly preceed this.
161
Returned as a map from revision to inventory entry.
163
This is a map containing the file revisions in all parents
164
for which the file exists, and its revision is not a parent of
165
any other. If the file is new, the set will be empty.
167
:param versioned_file_store: A store where ancestry data on this
168
file id can be queried.
169
:param transaction: The transaction that queries to the versioned
170
file store should be completed under.
171
:param entry_vf: The entry versioned file, if its already available.
173
def get_ancestors(weave, entry):
174
return set(weave.get_ancestry(entry.revision))
175
# revision:ie mapping for each ie found in previous_inventories.
177
# revision:ie mapping with one revision for each head.
179
# revision: ancestor list for each head
181
# identify candidate head revision ids.
182
for inv in previous_inventories:
183
if self.file_id in inv:
184
ie = inv[self.file_id]
185
assert ie.file_id == self.file_id
186
if ie.revision in candidates:
187
# same revision value in two different inventories:
188
# correct possible inconsistencies:
189
# * there was a bug in revision updates with 'x' bit
192
if candidates[ie.revision].executable != ie.executable:
193
candidates[ie.revision].executable = False
194
ie.executable = False
195
except AttributeError:
197
# must now be the same.
198
assert candidates[ie.revision] == ie
200
# add this revision as a candidate.
201
candidates[ie.revision] = ie
203
# common case optimisation
204
if len(candidates) == 1:
205
# if there is only one candidate revision found
206
# then we can opening the versioned file to access ancestry:
207
# there cannot be any ancestors to eliminate when there is
208
# only one revision available.
209
heads[ie.revision] = ie
212
# eliminate ancestors amongst the available candidates:
213
# heads are those that are not an ancestor of any other candidate
214
# - this provides convergence at a per-file level.
215
for ie in candidates.values():
216
# may be an ancestor of a known head:
217
already_present = 0 != len(
218
[head for head in heads
219
if ie.revision in head_ancestors[head]])
221
# an ancestor of an analyzed candidate.
223
# not an ancestor of a known head:
224
# load the versioned file for this file id if needed
226
entry_vf = versioned_file_store.get_weave_or_empty(
227
self.file_id, transaction)
228
ancestors = get_ancestors(entry_vf, ie)
229
# may knock something else out:
230
check_heads = list(heads.keys())
231
for head in check_heads:
232
if head in ancestors:
233
# this previously discovered 'head' is not
234
# really a head - its an ancestor of the newly
237
head_ancestors[ie.revision] = ancestors
238
heads[ie.revision] = ie
241
def get_tar_item(self, root, dp, now, tree):
242
"""Get a tarfile item and a file stream for its content."""
243
item = tarfile.TarInfo(pathjoin(root, dp))
244
# TODO: would be cool to actually set it to the timestamp of the
245
# revision it was last changed
247
fileobj = self._put_in_tar(item, tree)
251
"""Return true if the object this entry represents has textual data.
253
Note that textual data includes binary content.
255
Also note that all entries get weave files created for them.
256
This attribute is primarily used when upgrading from old trees that
257
did not have the weave index for all inventory entries.
261
def __init__(self, file_id, name, parent_id, text_id=None):
262
"""Create an InventoryEntry
264
The filename must be a single component, relative to the
265
parent directory; it cannot be a whole path or relative name.
267
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
272
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
273
Traceback (most recent call last):
274
InvalidEntryName: Invalid entry name: src/hello.c
276
assert isinstance(name, basestring), name
277
if '/' in name or '\\' in name:
278
raise InvalidEntryName(name=name)
279
self.executable = False
281
self.text_sha1 = None
282
self.text_size = None
283
self.file_id = file_id
285
self.text_id = text_id
286
self.parent_id = parent_id
287
self.symlink_target = None
289
def kind_character(self):
290
"""Return a short kind indicator useful for appending to names."""
291
raise BzrError('unknown kind %r' % self.kind)
293
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
295
def _put_in_tar(self, item, tree):
296
"""populate item for stashing in a tar, and return the content stream.
298
If no content is available, return None.
300
raise BzrError("don't know how to export {%s} of kind %r" %
301
(self.file_id, self.kind))
303
def put_on_disk(self, dest, dp, tree):
304
"""Create a representation of self on disk in the prefix dest.
306
This is a template method - implement _put_on_disk in subclasses.
308
fullpath = pathjoin(dest, dp)
309
self._put_on_disk(fullpath, tree)
310
mutter(" export {%s} kind %s to %s", self.file_id,
313
def _put_on_disk(self, fullpath, tree):
314
"""Put this entry onto disk at fullpath, from tree tree."""
315
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
317
def sorted_children(self):
318
l = self.children.items()
323
def versionable_kind(kind):
324
return kind in ('file', 'directory', 'symlink')
326
def check(self, checker, rev_id, inv, tree):
327
"""Check this inventory entry is intact.
329
This is a template method, override _check for kind specific
332
:param checker: Check object providing context for the checks;
333
can be used to find out what parts of the repository have already
335
:param rev_id: Revision id from which this InventoryEntry was loaded.
336
Not necessarily the last-changed revision for this file.
337
:param inv: Inventory from which the entry was loaded.
338
:param tree: RevisionTree for this entry.
340
if self.parent_id != None:
341
if not inv.has_id(self.parent_id):
342
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
343
% (self.parent_id, rev_id))
344
self._check(checker, rev_id, tree)
346
def _check(self, checker, rev_id, tree):
347
"""Check this inventory entry for kind specific errors."""
348
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
353
"""Clone this inventory entry."""
354
raise NotImplementedError
356
def _get_snapshot_change(self, previous_entries):
357
if len(previous_entries) > 1:
359
elif len(previous_entries) == 0:
362
return 'modified/renamed/reparented'
365
return ("%s(%r, %r, parent_id=%r)"
366
% (self.__class__.__name__,
371
def snapshot(self, revision, path, previous_entries,
372
work_tree, weave_store, transaction):
373
"""Make a snapshot of this entry which may or may not have changed.
375
This means that all its fields are populated, that it has its
376
text stored in the text store or weave.
378
mutter('new parents of %s are %r', path, previous_entries)
379
self._read_tree_state(path, work_tree)
380
if len(previous_entries) == 1:
381
# cannot be unchanged unless there is only one parent file rev.
382
parent_ie = previous_entries.values()[0]
383
if self._unchanged(parent_ie):
384
mutter("found unchanged entry")
385
self.revision = parent_ie.revision
387
return self.snapshot_revision(revision, previous_entries,
388
work_tree, weave_store, transaction)
390
def snapshot_revision(self, revision, previous_entries, work_tree,
391
weave_store, transaction):
392
"""Record this revision unconditionally."""
393
mutter('new revision for {%s}', self.file_id)
394
self.revision = revision
395
change = self._get_snapshot_change(previous_entries)
396
self._snapshot_text(previous_entries, work_tree, weave_store,
400
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
401
"""Record the 'text' of this entry, whatever form that takes.
403
This default implementation simply adds an empty text.
405
mutter('storing file {%s} in revision {%s}',
406
self.file_id, self.revision)
407
self._add_text_to_weave([], file_parents.keys(), weave_store, transaction)
409
def __eq__(self, other):
410
if not isinstance(other, InventoryEntry):
411
return NotImplemented
413
return ((self.file_id == other.file_id)
414
and (self.name == other.name)
415
and (other.symlink_target == self.symlink_target)
416
and (self.text_sha1 == other.text_sha1)
417
and (self.text_size == other.text_size)
418
and (self.text_id == other.text_id)
419
and (self.parent_id == other.parent_id)
420
and (self.kind == other.kind)
421
and (self.revision == other.revision)
422
and (self.executable == other.executable)
425
def __ne__(self, other):
426
return not (self == other)
429
raise ValueError('not hashable')
431
def _unchanged(self, previous_ie):
432
"""Has this entry changed relative to previous_ie.
434
This method should be overriden in child classes.
437
# different inv parent
438
if previous_ie.parent_id != self.parent_id:
441
elif previous_ie.name != self.name:
445
def _read_tree_state(self, path, work_tree):
446
"""Populate fields in the inventory entry from the given tree.
448
Note that this should be modified to be a noop on virtual trees
449
as all entries created there are prepopulated.
451
# TODO: Rather than running this manually, we should check the
452
# working sha1 and other expensive properties when they're
453
# first requested, or preload them if they're already known
454
pass # nothing to do by default
456
def _forget_tree_state(self):
460
class RootEntry(InventoryEntry):
462
def _check(self, checker, rev_id, tree):
463
"""See InventoryEntry._check"""
465
def __init__(self, file_id):
466
self.file_id = file_id
468
self.kind = 'root_directory'
469
self.parent_id = None
472
def __eq__(self, other):
473
if not isinstance(other, RootEntry):
474
return NotImplemented
476
return (self.file_id == other.file_id) \
477
and (self.children == other.children)
480
class InventoryDirectory(InventoryEntry):
481
"""A directory in an inventory."""
483
def _check(self, checker, rev_id, tree):
484
"""See InventoryEntry._check"""
485
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
486
raise BzrCheckError('directory {%s} has text in revision {%s}'
487
% (self.file_id, rev_id))
490
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
491
other.revision = self.revision
492
# note that children are *not* copied; they're pulled across when
496
def __init__(self, file_id, name, parent_id):
497
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
499
self.kind = 'directory'
501
def kind_character(self):
502
"""See InventoryEntry.kind_character."""
505
def _put_in_tar(self, item, tree):
506
"""See InventoryEntry._put_in_tar."""
507
item.type = tarfile.DIRTYPE
514
def _put_on_disk(self, fullpath, tree):
515
"""See InventoryEntry._put_on_disk."""
519
class InventoryFile(InventoryEntry):
520
"""A file in an inventory."""
522
def _check(self, checker, tree_revision_id, tree):
523
"""See InventoryEntry._check"""
524
t = (self.file_id, self.revision)
525
if t in checker.checked_texts:
526
prev_sha = checker.checked_texts[t]
527
if prev_sha != self.text_sha1:
528
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
529
(self.file_id, tree_revision_id))
531
checker.repeated_text_cnt += 1
534
if self.file_id not in checker.checked_weaves:
535
mutter('check weave {%s}', self.file_id)
536
w = tree.get_weave(self.file_id)
537
# Not passing a progress bar, because it creates a new
538
# progress, which overwrites the current progress,
539
# and doesn't look nice
541
checker.checked_weaves[self.file_id] = True
543
w = tree.get_weave(self.file_id)
545
mutter('check version {%s} of {%s}', tree_revision_id, self.file_id)
546
checker.checked_text_cnt += 1
547
# We can't check the length, because Weave doesn't store that
548
# information, and the whole point of looking at the weave's
549
# sha1sum is that we don't have to extract the text.
550
if self.text_sha1 != w.get_sha1(self.revision):
551
raise BzrCheckError('text {%s} version {%s} wrong sha1'
552
% (self.file_id, self.revision))
553
checker.checked_texts[t] = self.text_sha1
556
other = InventoryFile(self.file_id, self.name, self.parent_id)
557
other.executable = self.executable
558
other.text_id = self.text_id
559
other.text_sha1 = self.text_sha1
560
other.text_size = self.text_size
561
other.revision = self.revision
564
def detect_changes(self, old_entry):
565
"""See InventoryEntry.detect_changes."""
566
assert self.text_sha1 != None
567
assert old_entry.text_sha1 != None
568
text_modified = (self.text_sha1 != old_entry.text_sha1)
569
meta_modified = (self.executable != old_entry.executable)
570
return text_modified, meta_modified
572
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
573
output_to, reverse=False):
574
"""See InventoryEntry._diff."""
576
from_text = tree.get_file(self.file_id).readlines()
578
to_text = to_tree.get_file(to_entry.file_id).readlines()
582
text_diff(from_label, from_text,
583
to_label, to_text, output_to)
585
text_diff(to_label, to_text,
586
from_label, from_text, output_to)
588
print >> output_to, "Binary files differ"
591
"""See InventoryEntry.has_text."""
594
def __init__(self, file_id, name, parent_id):
595
super(InventoryFile, self).__init__(file_id, name, parent_id)
598
def kind_character(self):
599
"""See InventoryEntry.kind_character."""
602
def _put_in_tar(self, item, tree):
603
"""See InventoryEntry._put_in_tar."""
604
item.type = tarfile.REGTYPE
605
fileobj = tree.get_file(self.file_id)
606
item.size = self.text_size
607
if tree.is_executable(self.file_id):
613
def _put_on_disk(self, fullpath, tree):
614
"""See InventoryEntry._put_on_disk."""
615
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
616
if tree.is_executable(self.file_id):
617
os.chmod(fullpath, 0755)
619
def _read_tree_state(self, path, work_tree):
620
"""See InventoryEntry._read_tree_state."""
621
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
622
self.executable = work_tree.is_executable(self.file_id)
624
def _forget_tree_state(self):
625
self.text_sha1 = None
626
self.executable = None
628
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
629
"""See InventoryEntry._snapshot_text."""
630
mutter('storing file {%s} in revision {%s}',
631
self.file_id, self.revision)
632
# special case to avoid diffing on renames or
634
if (len(file_parents) == 1
635
and self.text_sha1 == file_parents.values()[0].text_sha1
636
and self.text_size == file_parents.values()[0].text_size):
637
previous_ie = file_parents.values()[0]
638
versionedfile = weave_store.get_weave(self.file_id, transaction)
639
versionedfile.clone_text(self.revision, previous_ie.revision, file_parents.keys())
641
new_lines = work_tree.get_file(self.file_id).readlines()
642
self._add_text_to_weave(new_lines, file_parents.keys(), weave_store,
644
self.text_sha1 = sha_strings(new_lines)
645
self.text_size = sum(map(len, new_lines))
648
def _unchanged(self, previous_ie):
649
"""See InventoryEntry._unchanged."""
650
compatible = super(InventoryFile, self)._unchanged(previous_ie)
651
if self.text_sha1 != previous_ie.text_sha1:
654
# FIXME: 20050930 probe for the text size when getting sha1
655
# in _read_tree_state
656
self.text_size = previous_ie.text_size
657
if self.executable != previous_ie.executable:
662
class InventoryLink(InventoryEntry):
663
"""A file in an inventory."""
665
__slots__ = ['symlink_target']
667
def _check(self, checker, rev_id, tree):
668
"""See InventoryEntry._check"""
669
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
670
raise BzrCheckError('symlink {%s} has text in revision {%s}'
671
% (self.file_id, rev_id))
672
if self.symlink_target == None:
673
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
674
% (self.file_id, rev_id))
677
other = InventoryLink(self.file_id, self.name, self.parent_id)
678
other.symlink_target = self.symlink_target
679
other.revision = self.revision
682
def detect_changes(self, old_entry):
683
"""See InventoryEntry.detect_changes."""
684
# FIXME: which _modified field should we use ? RBC 20051003
685
text_modified = (self.symlink_target != old_entry.symlink_target)
687
mutter(" symlink target changed")
688
meta_modified = False
689
return text_modified, meta_modified
691
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
692
output_to, reverse=False):
693
"""See InventoryEntry._diff."""
694
from_text = self.symlink_target
695
if to_entry is not None:
696
to_text = to_entry.symlink_target
701
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
704
print >>output_to, '=== target was %r' % self.symlink_target
706
print >>output_to, '=== target is %r' % self.symlink_target
708
def __init__(self, file_id, name, parent_id):
709
super(InventoryLink, self).__init__(file_id, name, parent_id)
710
self.kind = 'symlink'
712
def kind_character(self):
713
"""See InventoryEntry.kind_character."""
716
def _put_in_tar(self, item, tree):
717
"""See InventoryEntry._put_in_tar."""
718
item.type = tarfile.SYMTYPE
722
item.linkname = self.symlink_target
725
def _put_on_disk(self, fullpath, tree):
726
"""See InventoryEntry._put_on_disk."""
728
os.symlink(self.symlink_target, fullpath)
730
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
732
def _read_tree_state(self, path, work_tree):
733
"""See InventoryEntry._read_tree_state."""
734
self.symlink_target = work_tree.get_symlink_target(self.file_id)
736
def _forget_tree_state(self):
737
self.symlink_target = None
739
def _unchanged(self, previous_ie):
740
"""See InventoryEntry._unchanged."""
741
compatible = super(InventoryLink, self)._unchanged(previous_ie)
742
if self.symlink_target != previous_ie.symlink_target:
747
class Inventory(object):
748
"""Inventory of versioned files in a tree.
750
This describes which file_id is present at each point in the tree,
751
and possibly the SHA-1 or other information about the file.
752
Entries can be looked up either by path or by file_id.
754
The inventory represents a typical unix file tree, with
755
directories containing files and subdirectories. We never store
756
the full path to a file, because renaming a directory implicitly
757
moves all of its contents. This class internally maintains a
758
lookup tree that allows the children under a directory to be
761
InventoryEntry objects must not be modified after they are
762
inserted, other than through the Inventory API.
764
>>> inv = Inventory()
765
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
766
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT')
767
>>> inv['123-123'].name
770
May be treated as an iterator or set to look up file ids:
772
>>> bool(inv.path2id('hello.c'))
777
May also look up by name:
779
>>> [x[0] for x in inv.iter_entries()]
781
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
782
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
783
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
785
def __init__(self, root_id=ROOT_ID, revision_id=None):
786
"""Create or read an inventory.
788
If a working directory is specified, the inventory is read
789
from there. If the file is specified, read from that. If not,
790
the inventory is created empty.
792
The inventory is created with a default root directory, with
795
# We are letting Branch.create() create a unique inventory
796
# root id. Rather than generating a random one here.
798
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
799
self.root = RootEntry(root_id)
800
self.revision_id = revision_id
801
self._byid = {self.root.file_id: self.root}
805
# TODO: jam 20051218 Should copy also copy the revision_id?
806
other = Inventory(self.root.file_id)
807
# copy recursively so we know directories will be added before
808
# their children. There are more efficient ways than this...
809
for path, entry in self.iter_entries():
810
if entry == self.root:
812
other.add(entry.copy())
817
return iter(self._byid)
821
"""Returns number of entries."""
822
return len(self._byid)
825
def iter_entries(self, from_dir=None):
826
"""Return (path, entry) pairs, in order by name."""
830
elif isinstance(from_dir, basestring):
831
from_dir = self._byid[from_dir]
833
kids = from_dir.children.items()
835
for name, ie in kids:
837
if ie.kind == 'directory':
838
for cn, cie in self.iter_entries(from_dir=ie.file_id):
839
yield pathjoin(name, cn), cie
843
"""Return list of (path, ie) for all entries except the root.
845
This may be faster than iter_entries.
848
def descend(dir_ie, dir_path):
849
kids = dir_ie.children.items()
851
for name, ie in kids:
852
child_path = pathjoin(dir_path, name)
853
accum.append((child_path, ie))
854
if ie.kind == 'directory':
855
descend(ie, child_path)
857
descend(self.root, u'')
861
def directories(self):
862
"""Return (path, entry) pairs for all directories, including the root.
865
def descend(parent_ie, parent_path):
866
accum.append((parent_path, parent_ie))
868
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
871
for name, child_ie in kids:
872
child_path = pathjoin(parent_path, name)
873
descend(child_ie, child_path)
874
descend(self.root, u'')
879
def __contains__(self, file_id):
880
"""True if this entry contains a file with given id.
882
>>> inv = Inventory()
883
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
884
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
890
return file_id in self._byid
893
def __getitem__(self, file_id):
894
"""Return the entry for given file_id.
896
>>> inv = Inventory()
897
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
898
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
899
>>> inv['123123'].name
903
return self._byid[file_id]
906
raise BzrError("can't look up file_id None")
908
raise BzrError("file_id {%s} not in inventory" % file_id)
911
def get_file_kind(self, file_id):
912
return self._byid[file_id].kind
914
def get_child(self, parent_id, filename):
915
return self[parent_id].children.get(filename)
918
def add(self, entry):
919
"""Add entry to inventory.
921
To add a file to a branch ready to be committed, use Branch.add,
924
Returns the new entry object.
926
if entry.file_id in self._byid:
927
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
929
if entry.parent_id == ROOT_ID or entry.parent_id is None:
930
entry.parent_id = self.root.file_id
933
parent = self._byid[entry.parent_id]
935
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
937
if parent.children.has_key(entry.name):
938
raise BzrError("%s is already versioned" %
939
pathjoin(self.id2path(parent.file_id), entry.name))
941
self._byid[entry.file_id] = entry
942
parent.children[entry.name] = entry
946
def add_path(self, relpath, kind, file_id=None):
947
"""Add entry from a path.
949
The immediate parent must already be versioned.
951
Returns the new entry object."""
952
from bzrlib.workingtree import gen_file_id
954
parts = bzrlib.osutils.splitpath(relpath)
957
file_id = gen_file_id(relpath)
960
self.root = RootEntry(file_id)
961
self._byid = {self.root.file_id: self.root}
964
parent_path = parts[:-1]
965
parent_id = self.path2id(parent_path)
966
if parent_id == None:
967
raise NotVersionedError(path=parent_path)
968
if kind == 'directory':
969
ie = InventoryDirectory(file_id, parts[-1], parent_id)
971
ie = InventoryFile(file_id, parts[-1], parent_id)
972
elif kind == 'symlink':
973
ie = InventoryLink(file_id, parts[-1], parent_id)
975
raise BzrError("unknown kind %r" % kind)
979
def __delitem__(self, file_id):
980
"""Remove entry by id.
982
>>> inv = Inventory()
983
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
984
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
993
assert ie.parent_id is None or \
994
self[ie.parent_id].children[ie.name] == ie
996
del self._byid[file_id]
997
if ie.parent_id is not None:
998
del self[ie.parent_id].children[ie.name]
1001
def __eq__(self, other):
1002
"""Compare two sets by comparing their contents.
1004
>>> i1 = Inventory()
1005
>>> i2 = Inventory()
1008
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
1009
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
1012
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
1013
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
1017
if not isinstance(other, Inventory):
1018
return NotImplemented
1020
if len(self._byid) != len(other._byid):
1021
# shortcut: obviously not the same
1024
return self._byid == other._byid
1027
def __ne__(self, other):
1028
return not self.__eq__(other)
1032
raise ValueError('not hashable')
1034
def _iter_file_id_parents(self, file_id):
1035
"""Yield the parents of file_id up to the root."""
1036
while file_id != None:
1038
ie = self._byid[file_id]
1040
raise BzrError("file_id {%s} not found in inventory" % file_id)
1042
file_id = ie.parent_id
1044
def get_idpath(self, file_id):
1045
"""Return a list of file_ids for the path to an entry.
1047
The list contains one element for each directory followed by
1048
the id of the file itself. So the length of the returned list
1049
is equal to the depth of the file in the tree, counting the
1050
root directory as depth 1.
1053
for parent in self._iter_file_id_parents(file_id):
1054
p.insert(0, parent.file_id)
1057
def id2path(self, file_id):
1058
"""Return as a string the path to file_id.
1061
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
1062
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
1063
>>> print i.id2path('foo-id')
1066
# get all names, skipping root
1067
return '/'.join(reversed(
1068
[parent.name for parent in
1069
self._iter_file_id_parents(file_id)][:-1]))
1071
def path2id(self, name):
1072
"""Walk down through directories to return entry of last component.
1074
names may be either a list of path components, or a single
1075
string, in which case it is automatically split.
1077
This returns the entry of the last component in the path,
1078
which may be either a file or a directory.
1080
Returns None iff the path is not found.
1082
if isinstance(name, types.StringTypes):
1083
name = splitpath(name)
1085
mutter("lookup path %r" % name)
1090
cie = parent.children[f]
1091
assert cie.name == f
1092
assert cie.parent_id == parent.file_id
1095
# or raise an error?
1098
return parent.file_id
1101
def has_filename(self, names):
1102
return bool(self.path2id(names))
1105
def has_id(self, file_id):
1106
return self._byid.has_key(file_id)
1109
def rename(self, file_id, new_parent_id, new_name):
1110
"""Move a file within the inventory.
1112
This can change either the name, or the parent, or both.
1114
This does not move the working file."""
1115
if not is_valid_name(new_name):
1116
raise BzrError("not an acceptable filename: %r" % new_name)
1118
new_parent = self._byid[new_parent_id]
1119
if new_name in new_parent.children:
1120
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
1122
new_parent_idpath = self.get_idpath(new_parent_id)
1123
if file_id in new_parent_idpath:
1124
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
1125
% (self.id2path(file_id), self.id2path(new_parent_id)))
1127
file_ie = self._byid[file_id]
1128
old_parent = self._byid[file_ie.parent_id]
1130
# TODO: Don't leave things messed up if this fails
1132
del old_parent.children[file_ie.name]
1133
new_parent.children[new_name] = file_ie
1135
file_ie.name = new_name
1136
file_ie.parent_id = new_parent_id
1143
def is_valid_name(name):
1145
if _NAME_RE == None:
1146
_NAME_RE = re.compile(r'^[^/\\]+$')
1148
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
