78
81
>>> i.add(InventoryDirectory('123', 'src', ROOT_ID))
79
InventoryDirectory('123', 'src', parent_id='TREE_ROOT')
82
InventoryDirectory('123', 'src', parent_id='TREE_ROOT', revision=None)
80
83
>>> i.add(InventoryFile('2323', 'hello.c', parent_id='123'))
81
InventoryFile('2323', 'hello.c', parent_id='123')
82
>>> for j in i.iter_entries():
84
InventoryFile('2323', 'hello.c', parent_id='123', sha1=None, len=None)
85
>>> shouldbe = {0: '', 1: 'src', 2: pathjoin('src','hello.c')}
86
>>> for ix, j in enumerate(i.iter_entries()):
87
... print (j[0] == shouldbe[ix], j[1])
85
('src', InventoryDirectory('123', 'src', parent_id='TREE_ROOT'))
86
('src/hello.c', InventoryFile('2323', 'hello.c', parent_id='123'))
89
(True, InventoryDirectory('TREE_ROOT', '', parent_id=None, revision=None))
90
(True, InventoryDirectory('123', 'src', parent_id='TREE_ROOT', revision=None))
91
(True, InventoryFile('2323', 'hello.c', parent_id='123', sha1=None, len=None))
87
92
>>> i.add(InventoryFile('2323', 'bye.c', '123'))
88
93
Traceback (most recent call last):
90
95
BzrError: inventory already contains entry with id {2323}
91
96
>>> i.add(InventoryFile('2324', 'bye.c', '123'))
92
InventoryFile('2324', 'bye.c', parent_id='123')
97
InventoryFile('2324', 'bye.c', parent_id='123', sha1=None, len=None)
93
98
>>> i.add(InventoryDirectory('2325', 'wibble', '123'))
94
InventoryDirectory('2325', 'wibble', parent_id='123')
99
InventoryDirectory('2325', 'wibble', parent_id='123', revision=None)
95
100
>>> i.path2id('src/wibble')
99
104
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
100
InventoryFile('2326', 'wibble.c', parent_id='2325')
105
InventoryFile('2326', 'wibble.c', parent_id='2325', sha1=None, len=None)
102
InventoryFile('2326', 'wibble.c', parent_id='2325')
107
InventoryFile('2326', 'wibble.c', parent_id='2325', sha1=None, len=None)
103
108
>>> for path, entry in i.iter_entries():
104
... print path.replace('\\\\', '/') # for win32 os.sep
105
110
... assert i.path2id(path)
111
117
src/wibble/wibble.c
112
>>> i.id2path('2326').replace('\\\\', '/')
118
>>> i.id2path('2326')
113
119
'src/wibble/wibble.c'
122
# Constants returned by describe_change()
124
# TODO: These should probably move to some kind of FileChangeDescription
125
# class; that's like what's inside a TreeDelta but we want to be able to
126
# generate them just for one file at a time.
128
MODIFIED_AND_RENAMED = 'modified and renamed'
116
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
117
'text_id', 'parent_id', 'children', 'executable',
120
def _add_text_to_weave(self, new_lines, parents, weave_store, transaction):
121
weave_store.add_text(self.file_id, self.revision, new_lines, parents,
124
132
def detect_changes(self, old_entry):
125
133
"""Return a (text_modified, meta_modified) from this to old_entry.
150
158
output_to, reverse=False):
151
159
"""Perform a diff between two entries of the same kind."""
153
def find_previous_heads(self, previous_inventories, entry_weave):
154
"""Return the revisions and entries that directly preceed this.
161
def find_previous_heads(self, previous_inventories,
162
versioned_file_store,
165
"""Return the revisions and entries that directly precede this.
156
167
Returned as a map from revision to inventory entry.
158
169
This is a map containing the file revisions in all parents
159
170
for which the file exists, and its revision is not a parent of
160
171
any other. If the file is new, the set will be empty.
173
:param versioned_file_store: A store where ancestry data on this
174
file id can be queried.
175
:param transaction: The transaction that queries to the versioned
176
file store should be completed under.
177
:param entry_vf: The entry versioned file, if its already available.
162
179
def get_ancestors(weave, entry):
163
return set(map(weave.idx_to_name,
164
weave.inclusions([weave.lookup(entry.revision)])))
180
return set(weave.get_ancestry(entry.revision))
181
# revision:ie mapping for each ie found in previous_inventories.
183
# revision:ie mapping with one revision for each head.
185
# revision: ancestor list for each head
166
186
head_ancestors = {}
187
# identify candidate head revision ids.
167
188
for inv in previous_inventories:
168
189
if self.file_id in inv:
169
190
ie = inv[self.file_id]
170
191
assert ie.file_id == self.file_id
171
if ie.revision in heads:
172
# fixup logic, there was a bug in revision updates.
173
# with x bit support.
192
if ie.revision in candidates:
193
# same revision value in two different inventories:
194
# correct possible inconsistencies:
195
# * there was a bug in revision updates with 'x' bit
175
if heads[ie.revision].executable != ie.executable:
176
heads[ie.revision].executable = False
198
if candidates[ie.revision].executable != ie.executable:
199
candidates[ie.revision].executable = False
177
200
ie.executable = False
178
201
except AttributeError:
180
assert heads[ie.revision] == ie
203
# must now be the same.
204
assert candidates[ie.revision] == ie
182
# may want to add it.
183
# may already be covered:
184
already_present = 0 != len(
185
[head for head in heads
186
if ie.revision in head_ancestors[head]])
188
# an ancestor of a known head.
191
ancestors = get_ancestors(entry_weave, ie)
192
# may knock something else out:
193
check_heads = list(heads.keys())
194
for head in check_heads:
195
if head in ancestors:
196
# this head is not really a head
198
head_ancestors[ie.revision] = ancestors
199
heads[ie.revision] = ie
206
# add this revision as a candidate.
207
candidates[ie.revision] = ie
209
# common case optimisation
210
if len(candidates) == 1:
211
# if there is only one candidate revision found
212
# then we can opening the versioned file to access ancestry:
213
# there cannot be any ancestors to eliminate when there is
214
# only one revision available.
215
heads[ie.revision] = ie
218
# eliminate ancestors amongst the available candidates:
219
# heads are those that are not an ancestor of any other candidate
220
# - this provides convergence at a per-file level.
221
for ie in candidates.values():
222
# may be an ancestor of a known head:
223
already_present = 0 != len(
224
[head for head in heads
225
if ie.revision in head_ancestors[head]])
227
# an ancestor of an analyzed candidate.
229
# not an ancestor of a known head:
230
# load the versioned file for this file id if needed
232
entry_vf = versioned_file_store.get_weave_or_empty(
233
self.file_id, transaction)
234
ancestors = get_ancestors(entry_vf, ie)
235
# may knock something else out:
236
check_heads = list(heads.keys())
237
for head in check_heads:
238
if head in ancestors:
239
# this previously discovered 'head' is not
240
# really a head - its an ancestor of the newly
243
head_ancestors[ie.revision] = ancestors
244
heads[ie.revision] = ie
202
247
def get_tar_item(self, root, dp, now, tree):
203
248
"""Get a tarfile item and a file stream for its content."""
204
item = tarfile.TarInfo(os.path.join(root, dp))
249
item = tarfile.TarInfo(pathjoin(root, dp).encode('utf8'))
205
250
# TODO: would be cool to actually set it to the timestamp of the
206
251
# revision it was last changed
267
312
This is a template method - implement _put_on_disk in subclasses.
269
fullpath = appendpath(dest, dp)
314
fullpath = pathjoin(dest, dp)
270
315
self._put_on_disk(fullpath, tree)
271
mutter(" export {%s} kind %s to %s" % (self.file_id, self.kind, fullpath))
316
# mutter(" export {%s} kind %s to %s", self.file_id,
317
# self.kind, fullpath)
273
319
def _put_on_disk(self, fullpath, tree):
274
320
"""Put this entry onto disk at fullpath, from tree tree."""
275
321
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
277
323
def sorted_children(self):
278
l = self.children.items()
324
return sorted(self.children.items())
283
327
def versionable_kind(kind):
284
return kind in ('file', 'directory', 'symlink')
328
return (kind in ('file', 'directory', 'symlink'))
286
330
def check(self, checker, rev_id, inv, tree):
287
331
"""Check this inventory entry is intact.
289
333
This is a template method, override _check for kind specific
336
:param checker: Check object providing context for the checks;
337
can be used to find out what parts of the repository have already
339
:param rev_id: Revision id from which this InventoryEntry was loaded.
340
Not necessarily the last-changed revision for this file.
341
:param inv: Inventory from which the entry was loaded.
342
:param tree: RevisionTree for this entry.
292
if self.parent_id != None:
344
if self.parent_id is not None:
293
345
if not inv.has_id(self.parent_id):
294
346
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
295
347
% (self.parent_id, rev_id))
300
352
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
301
353
(self.kind, rev_id))
305
356
"""Clone this inventory entry."""
306
357
raise NotImplementedError
308
def _get_snapshot_change(self, previous_entries):
309
if len(previous_entries) > 1:
311
elif len(previous_entries) == 0:
360
def describe_change(old_entry, new_entry):
361
"""Describe the change between old_entry and this.
363
This smells of being an InterInventoryEntry situation, but as its
364
the first one, we're making it a static method for now.
366
An entry with a different parent, or different name is considered
367
to be renamed. Reparenting is an internal detail.
368
Note that renaming the parent does not trigger a rename for the
371
# TODO: Perhaps return an object rather than just a string
372
if old_entry is new_entry:
373
# also the case of both being None
375
elif old_entry is None:
314
return 'modified/renamed/reparented'
377
elif new_entry is None:
379
text_modified, meta_modified = new_entry.detect_changes(old_entry)
380
if text_modified or meta_modified:
384
# TODO 20060511 (mbp, rbc) factor out 'detect_rename' here.
385
if old_entry.parent_id != new_entry.parent_id:
387
elif old_entry.name != new_entry.name:
391
if renamed and not modified:
392
return InventoryEntry.RENAMED
393
if modified and not renamed:
395
if modified and renamed:
396
return InventoryEntry.MODIFIED_AND_RENAMED
316
399
def __repr__(self):
317
return ("%s(%r, %r, parent_id=%r)"
400
return ("%s(%r, %r, parent_id=%r, revision=%r)"
318
401
% (self.__class__.__name__,
323
407
def snapshot(self, revision, path, previous_entries,
324
work_tree, weave_store, transaction):
408
work_tree, commit_builder):
325
409
"""Make a snapshot of this entry which may or may not have changed.
327
411
This means that all its fields are populated, that it has its
328
412
text stored in the text store or weave.
330
mutter('new parents of %s are %r', path, previous_entries)
414
# mutter('new parents of %s are %r', path, previous_entries)
331
415
self._read_tree_state(path, work_tree)
416
# TODO: Where should we determine whether to reuse a
417
# previous revision id or create a new revision? 20060606
332
418
if len(previous_entries) == 1:
333
419
# cannot be unchanged unless there is only one parent file rev.
334
420
parent_ie = previous_entries.values()[0]
335
421
if self._unchanged(parent_ie):
336
mutter("found unchanged entry")
422
# mutter("found unchanged entry")
337
423
self.revision = parent_ie.revision
338
424
return "unchanged"
339
return self.snapshot_revision(revision, previous_entries,
340
work_tree, weave_store, transaction)
342
def snapshot_revision(self, revision, previous_entries, work_tree,
343
weave_store, transaction):
344
"""Record this revision unconditionally."""
345
mutter('new revision for {%s}', self.file_id)
425
return self._snapshot_into_revision(revision, previous_entries,
426
work_tree, commit_builder)
428
def _snapshot_into_revision(self, revision, previous_entries, work_tree,
430
"""Record this revision unconditionally into a store.
432
The entry's last-changed revision property (`revision`) is updated to
433
that of the new revision.
435
:param revision: id of the new revision that is being recorded.
437
:returns: String description of the commit (e.g. "merged", "modified"), etc.
439
# mutter('new revision {%s} for {%s}', revision, self.file_id)
346
440
self.revision = revision
347
change = self._get_snapshot_change(previous_entries)
348
self._snapshot_text(previous_entries, work_tree, weave_store,
441
self._snapshot_text(previous_entries, work_tree, commit_builder)
352
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
443
def _snapshot_text(self, file_parents, work_tree, commit_builder):
353
444
"""Record the 'text' of this entry, whatever form that takes.
355
446
This default implementation simply adds an empty text.
357
mutter('storing file {%s} in revision {%s}',
358
self.file_id, self.revision)
359
self._add_text_to_weave([], file_parents, weave_store, transaction)
448
raise NotImplementedError(self._snapshot_text)
361
450
def __eq__(self, other):
362
451
if not isinstance(other, InventoryEntry):
405
494
# first requested, or preload them if they're already known
406
495
pass # nothing to do by default
497
def _forget_tree_state(self):
409
501
class RootEntry(InventoryEntry):
503
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
504
'text_id', 'parent_id', 'children', 'executable',
505
'revision', 'symlink_target']
411
507
def _check(self, checker, rev_id, tree):
412
508
"""See InventoryEntry._check"""
414
510
def __init__(self, file_id):
415
511
self.file_id = file_id
416
512
self.children = {}
417
self.kind = 'root_directory'
513
self.kind = 'directory'
418
514
self.parent_id = None
517
warn('RootEntry is deprecated as of bzr 0.10. Please use '
518
'InventoryDirectory instead.',
519
DeprecationWarning, stacklevel=2)
421
521
def __eq__(self, other):
422
522
if not isinstance(other, RootEntry):
464
568
"""See InventoryEntry._put_on_disk."""
465
569
os.mkdir(fullpath)
571
def _snapshot_text(self, file_parents, work_tree, commit_builder):
572
"""See InventoryEntry._snapshot_text."""
573
commit_builder.modified_directory(self.file_id, file_parents)
468
576
class InventoryFile(InventoryEntry):
469
577
"""A file in an inventory."""
471
def _check(self, checker, rev_id, tree):
579
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
580
'text_id', 'parent_id', 'children', 'executable',
581
'revision', 'symlink_target']
583
def _check(self, checker, tree_revision_id, tree):
472
584
"""See InventoryEntry._check"""
473
revision = self.revision
474
t = (self.file_id, revision)
585
t = (self.file_id, self.revision)
475
586
if t in checker.checked_texts:
476
prev_sha = checker.checked_texts[t]
587
prev_sha = checker.checked_texts[t]
477
588
if prev_sha != self.text_sha1:
478
589
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
479
(self.file_id, rev_id))
590
(self.file_id, tree_revision_id))
481
592
checker.repeated_text_cnt += 1
483
mutter('check version {%s} of {%s}', rev_id, self.file_id)
484
file_lines = tree.get_file_lines(self.file_id)
485
checker.checked_text_cnt += 1
486
if self.text_size != sum(map(len, file_lines)):
487
raise BzrCheckError('text {%s} wrong size' % self.text_id)
488
if self.text_sha1 != sha_strings(file_lines):
489
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
595
if self.file_id not in checker.checked_weaves:
596
mutter('check weave {%s}', self.file_id)
597
w = tree.get_weave(self.file_id)
598
# Not passing a progress bar, because it creates a new
599
# progress, which overwrites the current progress,
600
# and doesn't look nice
602
checker.checked_weaves[self.file_id] = True
604
w = tree.get_weave(self.file_id)
606
mutter('check version {%s} of {%s}', tree_revision_id, self.file_id)
607
checker.checked_text_cnt += 1
608
# We can't check the length, because Weave doesn't store that
609
# information, and the whole point of looking at the weave's
610
# sha1sum is that we don't have to extract the text.
611
if self.text_sha1 != w.get_sha1(self.revision):
612
raise BzrCheckError('text {%s} version {%s} wrong sha1'
613
% (self.file_id, self.revision))
490
614
checker.checked_texts[t] = self.text_sha1
509
633
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
510
634
output_to, reverse=False):
511
635
"""See InventoryEntry._diff."""
512
from_text = tree.get_file(self.file_id).readlines()
514
to_text = to_tree.get_file(to_entry.file_id).readlines()
518
text_diff(from_label, from_text,
519
to_label, to_text, output_to)
521
text_diff(to_label, to_text,
522
from_label, from_text, output_to)
637
from_text = tree.get_file(self.file_id).readlines()
639
to_text = to_tree.get_file(to_entry.file_id).readlines()
643
text_diff(from_label, from_text,
644
to_label, to_text, output_to)
646
text_diff(to_label, to_text,
647
from_label, from_text, output_to)
650
label_pair = (to_label, from_label)
652
label_pair = (from_label, to_label)
653
print >> output_to, "Binary files %s and %s differ" % label_pair
524
655
def has_text(self):
525
656
"""See InventoryEntry.has_text."""
553
684
def _read_tree_state(self, path, work_tree):
554
685
"""See InventoryEntry._read_tree_state."""
555
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
556
self.executable = work_tree.is_executable(self.file_id)
558
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
686
self.text_sha1 = work_tree.get_file_sha1(self.file_id, path=path)
687
# FIXME: 20050930 probe for the text size when getting sha1
688
# in _read_tree_state
689
self.executable = work_tree.is_executable(self.file_id, path=path)
692
return ("%s(%r, %r, parent_id=%r, sha1=%r, len=%s)"
693
% (self.__class__.__name__,
700
def _forget_tree_state(self):
701
self.text_sha1 = None
703
def _snapshot_text(self, file_parents, work_tree, commit_builder):
559
704
"""See InventoryEntry._snapshot_text."""
560
mutter('storing file {%s} in revision {%s}',
561
self.file_id, self.revision)
562
# special case to avoid diffing on renames or
564
if (len(file_parents) == 1
565
and self.text_sha1 == file_parents.values()[0].text_sha1
566
and self.text_size == file_parents.values()[0].text_size):
567
previous_ie = file_parents.values()[0]
568
weave_store.add_identical_text(
569
self.file_id, previous_ie.revision,
570
self.revision, file_parents, transaction)
572
new_lines = work_tree.get_file(self.file_id).readlines()
573
self._add_text_to_weave(new_lines, file_parents, weave_store,
575
self.text_sha1 = sha_strings(new_lines)
576
self.text_size = sum(map(len, new_lines))
705
def get_content_byte_lines():
706
return work_tree.get_file(self.file_id).readlines()
707
self.text_sha1, self.text_size = commit_builder.modified_file_text(
708
self.file_id, file_parents, get_content_byte_lines, self.text_sha1, self.text_size)
579
710
def _unchanged(self, previous_ie):
580
711
"""See InventoryEntry._unchanged."""
593
724
class InventoryLink(InventoryEntry):
594
725
"""A file in an inventory."""
596
__slots__ = ['symlink_target']
727
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
728
'text_id', 'parent_id', 'children', 'executable',
729
'revision', 'symlink_target']
598
731
def _check(self, checker, rev_id, tree):
599
732
"""See InventoryEntry._check"""
600
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
733
if self.text_sha1 is not None or self.text_size is not None or self.text_id is not None:
601
734
raise BzrCheckError('symlink {%s} has text in revision {%s}'
602
735
% (self.file_id, rev_id))
603
if self.symlink_target == None:
736
if self.symlink_target is None:
604
737
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
605
738
% (self.file_id, rev_id))
720
861
The inventory is created with a default root directory, with
723
# We are letting Branch.initialize() create a unique inventory
864
# We are letting Branch.create() create a unique inventory
724
865
# root id. Rather than generating a random one here.
725
866
#if root_id is None:
726
867
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
727
self.root = RootEntry(root_id)
868
if root_id is not None:
869
self._set_root(InventoryDirectory(root_id, '', None))
873
# FIXME: this isn't ever used, changing it to self.revision may break
874
# things. TODO make everything use self.revision_id
875
self.revision_id = revision_id
877
def _set_root(self, ie):
728
879
self._byid = {self.root.file_id: self.root}
732
other = Inventory(self.root.file_id)
882
# TODO: jam 20051218 Should copy also copy the revision_id?
883
entries = self.iter_entries()
884
other = Inventory(entries.next()[1].file_id)
733
885
# copy recursively so we know directories will be added before
734
886
# their children. There are more efficient ways than this...
735
for path, entry in self.iter_entries():
736
if entry == self.root:
887
for path, entry in entries():
738
888
other.add(entry.copy())
742
891
def __iter__(self):
743
892
return iter(self._byid)
746
894
def __len__(self):
747
895
"""Returns number of entries."""
748
896
return len(self._byid)
751
898
def iter_entries(self, from_dir=None):
752
899
"""Return (path, entry) pairs, in order by name."""
756
elif isinstance(from_dir, basestring):
757
from_dir = self._byid[from_dir]
759
kids = from_dir.children.items()
761
for name, ie in kids:
763
if ie.kind == 'directory':
764
for cn, cie in self.iter_entries(from_dir=ie.file_id):
765
yield os.path.join(name, cn), cie
904
elif isinstance(from_dir, basestring):
905
from_dir = self._byid[from_dir]
907
# unrolling the recursive called changed the time from
908
# 440ms/663ms (inline/total) to 116ms/116ms
909
children = from_dir.children.items()
911
children = collections.deque(children)
912
stack = [(u'', children)]
914
from_dir_relpath, children = stack[-1]
917
name, ie = children.popleft()
919
# we know that from_dir_relpath never ends in a slash
920
# and 'f' doesn't begin with one, we can do a string op, rather
921
# than the checks of pathjoin(), though this means that all paths
923
path = from_dir_relpath + '/' + name
927
if ie.kind != 'directory':
930
# But do this child first
931
new_children = ie.children.items()
933
new_children = collections.deque(new_children)
934
stack.append((path, new_children))
935
# Break out of inner loop, so that we start outer loop with child
938
# if we finished all children, pop it off the stack
941
def iter_entries_by_dir(self, from_dir=None):
942
"""Iterate over the entries in a directory first order.
944
This returns all entries for a directory before returning
945
the entries for children of a directory. This is not
946
lexicographically sorted order, and is a hybrid between
947
depth-first and breadth-first.
949
:return: This yields (path, entry) pairs
951
# TODO? Perhaps this should return the from_dir so that the root is
952
# yielded? or maybe an option?
957
elif isinstance(from_dir, basestring):
958
from_dir = self._byid[from_dir]
960
stack = [(u'', from_dir)]
962
cur_relpath, cur_dir = stack.pop()
965
for child_name, child_ie in sorted(cur_dir.children.iteritems()):
967
child_relpath = cur_relpath + child_name
969
yield child_relpath, child_ie
971
if child_ie.kind == 'directory':
972
child_dirs.append((child_relpath+'/', child_ie))
973
stack.extend(reversed(child_dirs))
768
975
def entries(self):
769
976
"""Return list of (path, ie) for all entries except the root.
797
1003
for name, child_ie in kids:
798
child_path = os.path.join(parent_path, name)
1004
child_path = pathjoin(parent_path, name)
799
1005
descend(child_ie, child_path)
800
descend(self.root, '')
1006
descend(self.root, u'')
805
1009
def __contains__(self, file_id):
806
1010
"""True if this entry contains a file with given id.
808
1012
>>> inv = Inventory()
809
1013
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
810
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
1014
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT', sha1=None, len=None)
811
1015
>>> '123' in inv
813
1017
>>> '456' in inv
816
return file_id in self._byid
1020
return (file_id in self._byid)
819
1022
def __getitem__(self, file_id):
820
1023
"""Return the entry for given file_id.
822
1025
>>> inv = Inventory()
823
1026
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
824
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
1027
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT', sha1=None, len=None)
825
1028
>>> inv['123123'].name
829
1032
return self._byid[file_id]
830
1033
except KeyError:
832
1035
raise BzrError("can't look up file_id None")
834
1037
raise BzrError("file_id {%s} not in inventory" % file_id)
837
1039
def get_file_kind(self, file_id):
838
1040
return self._byid[file_id].kind
840
1042
def get_child(self, parent_id, filename):
841
1043
return self[parent_id].children.get(filename)
844
1045
def add(self, entry):
845
1046
"""Add entry to inventory.
860
1066
except KeyError:
861
1067
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
863
if parent.children.has_key(entry.name):
1069
if entry.name in parent.children:
864
1070
raise BzrError("%s is already versioned" %
865
appendpath(self.id2path(parent.file_id), entry.name))
1071
pathjoin(self.id2path(parent.file_id), entry.name))
867
1073
self._byid[entry.file_id] = entry
868
1074
parent.children[entry.name] = entry
872
def add_path(self, relpath, kind, file_id=None):
1077
def add_path(self, relpath, kind, file_id=None, parent_id=None):
873
1078
"""Add entry from a path.
875
1080
The immediate parent must already be versioned.
877
1082
Returns the new entry object."""
878
from bzrlib.branch import gen_file_id
880
parts = bzrlib.osutils.splitpath(relpath)
1084
parts = osutils.splitpath(relpath)
881
1086
if len(parts) == 0:
882
raise BzrError("cannot re-add root of inventory")
885
file_id = gen_file_id(relpath)
887
parent_path = parts[:-1]
888
parent_id = self.path2id(parent_path)
889
if parent_id == None:
890
raise NotVersionedError(path=parent_path)
891
if kind == 'directory':
892
ie = InventoryDirectory(file_id, parts[-1], parent_id)
894
ie = InventoryFile(file_id, parts[-1], parent_id)
895
elif kind == 'symlink':
896
ie = InventoryLink(file_id, parts[-1], parent_id)
1088
file_id = bzrlib.workingtree.gen_root_id()
1089
self.root = InventoryDirectory(file_id, '', None)
1090
self._byid = {self.root.file_id: self.root}
898
raise BzrError("unknown kind %r" % kind)
1093
parent_path = parts[:-1]
1094
parent_id = self.path2id(parent_path)
1095
if parent_id is None:
1096
raise NotVersionedError(path=parent_path)
1097
ie = make_entry(kind, parts[-1], parent_id, file_id)
899
1098
return self.add(ie)
902
1100
def __delitem__(self, file_id):
903
1101
"""Remove entry by id.
905
1103
>>> inv = Inventory()
906
1104
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
907
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
1105
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT', sha1=None, len=None)
908
1106
>>> '123' in inv
910
1108
>>> del inv['123']
936
1128
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
937
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
1129
InventoryFile('123', 'foo', parent_id='TREE_ROOT', sha1=None, len=None)
940
1132
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
941
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
1133
InventoryFile('123', 'foo', parent_id='TREE_ROOT', sha1=None, len=None)
945
1137
if not isinstance(other, Inventory):
946
1138
return NotImplemented
948
if len(self._byid) != len(other._byid):
949
# shortcut: obviously not the same
952
1140
return self._byid == other._byid
955
1142
def __ne__(self, other):
956
1143
return not self.__eq__(other)
959
1145
def __hash__(self):
960
1146
raise ValueError('not hashable')
1148
def _iter_file_id_parents(self, file_id):
1149
"""Yield the parents of file_id up to the root."""
1150
while file_id is not None:
1152
ie = self._byid[file_id]
1154
raise BzrError("file_id {%s} not found in inventory" % file_id)
1156
file_id = ie.parent_id
963
1158
def get_idpath(self, file_id):
964
1159
"""Return a list of file_ids for the path to an entry.
969
1164
root directory as depth 1.
972
while file_id != None:
974
ie = self._byid[file_id]
976
raise BzrError("file_id {%s} not found in inventory" % file_id)
977
p.insert(0, ie.file_id)
978
file_id = ie.parent_id
1167
for parent in self._iter_file_id_parents(file_id):
1168
p.insert(0, parent.file_id)
982
1171
def id2path(self, file_id):
983
"""Return as a list the path to file_id.
1172
"""Return as a string the path to file_id.
985
1174
>>> i = Inventory()
986
1175
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
987
1176
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
988
>>> print i.id2path('foo-id').replace(os.sep, '/')
1177
>>> print i.id2path('foo-id')
991
1180
# get all names, skipping root
992
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
993
return os.sep.join(p)
1181
return '/'.join(reversed(
1182
[parent.name for parent in
1183
self._iter_file_id_parents(file_id)][:-1]))
997
1185
def path2id(self, name):
998
1186
"""Walk down through directories to return entry of last component.
1024
1212
return parent.file_id
1027
1214
def has_filename(self, names):
1028
1215
return bool(self.path2id(names))
1031
1217
def has_id(self, file_id):
1032
return self._byid.has_key(file_id)
1218
return (file_id in self._byid)
1220
def remove_recursive_id(self, file_id):
1221
"""Remove file_id, and children, from the inventory.
1223
:param file_id: A file_id to remove.
1225
to_find_delete = [self._byid[file_id]]
1227
while to_find_delete:
1228
ie = to_find_delete.pop()
1229
to_delete.append(ie.file_id)
1230
if ie.kind == 'directory':
1231
to_find_delete.extend(ie.children.values())
1232
for file_id in reversed(to_delete):
1234
del self._byid[file_id]
1235
if ie.parent_id is not None:
1236
del self[ie.parent_id].children[ie.name]
1035
1238
def rename(self, file_id, new_parent_id, new_name):
1036
1239
"""Move a file within the inventory.
1062
1265
file_ie.parent_id = new_parent_id
1268
def make_entry(kind, name, parent_id, file_id=None):
1269
"""Create an inventory entry.
1271
:param kind: the type of inventory entry to create.
1272
:param name: the basename of the entry.
1273
:param parent_id: the parent_id of the entry.
1274
:param file_id: the file_id to use. if None, one will be created.
1277
file_id = bzrlib.workingtree.gen_file_id(name)
1279
norm_name, can_access = osutils.normalized_filename(name)
1280
if norm_name != name:
1284
# TODO: jam 20060701 This would probably be more useful
1285
# if the error was raised with the full path
1286
raise errors.InvalidNormalization(name)
1288
if kind == 'directory':
1289
return InventoryDirectory(file_id, name, parent_id)
1290
elif kind == 'file':
1291
return InventoryFile(file_id, name, parent_id)
1292
elif kind == 'symlink':
1293
return InventoryLink(file_id, name, parent_id)
1295
raise BzrError("unknown kind %r" % kind)
1067
1298
_NAME_RE = None
1069
1300
def is_valid_name(name):
1070
1301
global _NAME_RE
1071
if _NAME_RE == None:
1302
if _NAME_RE is None:
1072
1303
_NAME_RE = re.compile(r'^[^/\\]+$')
1074
1305
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
