81
75
>>> i.add(InventoryDirectory('123', 'src', ROOT_ID))
82
InventoryDirectory('123', 'src', parent_id='TREE_ROOT', revision=None)
76
InventoryDirectory('123', 'src', parent_id='TREE_ROOT')
83
77
>>> i.add(InventoryFile('2323', 'hello.c', parent_id='123'))
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])
78
InventoryFile('2323', 'hello.c', parent_id='123')
79
>>> for j in i.iter_entries():
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))
82
('src', InventoryDirectory('123', 'src', parent_id='TREE_ROOT'))
83
('src/hello.c', InventoryFile('2323', 'hello.c', parent_id='123'))
92
84
>>> i.add(InventoryFile('2323', 'bye.c', '123'))
93
85
Traceback (most recent call last):
95
87
BzrError: inventory already contains entry with id {2323}
96
88
>>> i.add(InventoryFile('2324', 'bye.c', '123'))
97
InventoryFile('2324', 'bye.c', parent_id='123', sha1=None, len=None)
89
InventoryFile('2324', 'bye.c', parent_id='123')
98
90
>>> i.add(InventoryDirectory('2325', 'wibble', '123'))
99
InventoryDirectory('2325', 'wibble', parent_id='123', revision=None)
91
InventoryDirectory('2325', 'wibble', parent_id='123')
100
92
>>> i.path2id('src/wibble')
104
96
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
105
InventoryFile('2326', 'wibble.c', parent_id='2325', sha1=None, len=None)
97
InventoryFile('2326', 'wibble.c', parent_id='2325')
107
InventoryFile('2326', 'wibble.c', parent_id='2325', sha1=None, len=None)
99
InventoryFile('2326', 'wibble.c', parent_id='2325')
108
100
>>> for path, entry in i.iter_entries():
101
... print path.replace('\\\\', '/') # for win32 os.sep
110
102
... assert i.path2id(path)
117
108
src/wibble/wibble.c
118
>>> i.id2path('2326')
109
>>> i.id2path('2326').replace('\\\\', '/')
119
110
'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'
113
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
114
'text_id', 'parent_id', 'children', 'executable',
117
def _add_text_to_weave(self, new_lines, parents, weave_store, transaction):
118
weave_store.add_text(self.file_id, self.revision, new_lines, parents,
132
121
def detect_changes(self, old_entry):
133
122
"""Return a (text_modified, meta_modified) from this to old_entry.
158
147
output_to, reverse=False):
159
148
"""Perform a diff between two entries of the same kind."""
161
def find_previous_heads(self, previous_inventories,
162
versioned_file_store,
165
"""Return the revisions and entries that directly precede this.
150
def find_previous_heads(self, previous_inventories, entry_weave):
151
"""Return the revisions and entries that directly preceed this.
167
153
Returned as a map from revision to inventory entry.
169
155
This is a map containing the file revisions in all parents
170
156
for which the file exists, and its revision is not a parent of
171
157
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.
179
159
def get_ancestors(weave, entry):
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.
160
return set(map(weave.idx_to_name,
161
weave.inclusions([weave.lookup(entry.revision)])))
185
# revision: ancestor list for each head
186
163
head_ancestors = {}
187
# identify candidate head revision ids.
188
164
for inv in previous_inventories:
189
165
if self.file_id in inv:
190
166
ie = inv[self.file_id]
191
167
assert ie.file_id == self.file_id
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
168
if ie.revision in heads:
169
# fixup logic, there was a bug in revision updates.
170
# with x bit support.
198
if candidates[ie.revision].executable != ie.executable:
199
candidates[ie.revision].executable = False
172
if heads[ie.revision].executable != ie.executable:
173
heads[ie.revision].executable = False
200
174
ie.executable = False
201
175
except AttributeError:
203
# must now be the same.
204
assert candidates[ie.revision] == ie
177
assert 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
179
# may want to add it.
180
# may already be covered:
181
already_present = 0 != len(
182
[head for head in heads
183
if ie.revision in head_ancestors[head]])
185
# an ancestor of a known head.
188
ancestors = get_ancestors(entry_weave, ie)
189
# may knock something else out:
190
check_heads = list(heads.keys())
191
for head in check_heads:
192
if head in ancestors:
193
# this head is not really a head
195
head_ancestors[ie.revision] = ancestors
196
heads[ie.revision] = ie
247
199
def get_tar_item(self, root, dp, now, tree):
248
200
"""Get a tarfile item and a file stream for its content."""
249
item = tarfile.TarInfo(pathjoin(root, dp))
201
item = tarfile.TarInfo(os.path.join(root, dp))
250
202
# TODO: would be cool to actually set it to the timestamp of the
251
203
# revision it was last changed
312
265
This is a template method - implement _put_on_disk in subclasses.
314
fullpath = pathjoin(dest, dp)
267
fullpath = appendpath(dest, dp)
315
268
self._put_on_disk(fullpath, tree)
316
# mutter(" export {%s} kind %s to %s", self.file_id,
317
# self.kind, fullpath)
269
mutter(" export {%s} kind %s to %s" % (self.file_id, self.kind, fullpath))
319
271
def _put_on_disk(self, fullpath, tree):
320
272
"""Put this entry onto disk at fullpath, from tree tree."""
321
273
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
323
275
def sorted_children(self):
324
return sorted(self.children.items())
276
l = self.children.items()
327
281
def versionable_kind(kind):
328
return (kind in ('file', 'directory', 'symlink'))
282
return kind in ('file', 'directory', 'symlink')
330
284
def check(self, checker, rev_id, inv, tree):
331
285
"""Check this inventory entry is intact.
333
287
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.
344
if self.parent_id is not None:
290
if self.parent_id != None:
345
291
if not inv.has_id(self.parent_id):
346
292
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
347
293
% (self.parent_id, rev_id))
352
298
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
353
299
(self.kind, rev_id))
356
303
"""Clone this inventory entry."""
357
304
raise NotImplementedError
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:
306
def _get_snapshot_change(self, previous_entries):
307
if len(previous_entries) > 1:
309
elif len(previous_entries) == 0:
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
312
return 'modified/renamed/reparented'
399
314
def __repr__(self):
400
return ("%s(%r, %r, parent_id=%r, revision=%r)"
315
return ("%s(%r, %r, parent_id=%r)"
401
316
% (self.__class__.__name__,
407
321
def snapshot(self, revision, path, previous_entries,
408
work_tree, commit_builder):
322
work_tree, weave_store, transaction):
409
323
"""Make a snapshot of this entry which may or may not have changed.
411
325
This means that all its fields are populated, that it has its
412
326
text stored in the text store or weave.
414
# mutter('new parents of %s are %r', path, previous_entries)
328
mutter('new parents of %s are %r', path, previous_entries)
415
329
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
418
330
if len(previous_entries) == 1:
419
331
# cannot be unchanged unless there is only one parent file rev.
420
332
parent_ie = previous_entries.values()[0]
421
333
if self._unchanged(parent_ie):
422
# mutter("found unchanged entry")
334
mutter("found unchanged entry")
423
335
self.revision = parent_ie.revision
424
336
return "unchanged"
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)
337
return self.snapshot_revision(revision, previous_entries,
338
work_tree, weave_store, transaction)
340
def snapshot_revision(self, revision, previous_entries, work_tree,
341
weave_store, transaction):
342
"""Record this revision unconditionally."""
343
mutter('new revision for {%s}', self.file_id)
440
344
self.revision = revision
441
self._snapshot_text(previous_entries, work_tree, commit_builder)
345
change = self._get_snapshot_change(previous_entries)
346
self._snapshot_text(previous_entries, work_tree, weave_store,
443
def _snapshot_text(self, file_parents, work_tree, commit_builder):
350
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
444
351
"""Record the 'text' of this entry, whatever form that takes.
446
353
This default implementation simply adds an empty text.
448
raise NotImplementedError(self._snapshot_text)
355
mutter('storing file {%s} in revision {%s}',
356
self.file_id, self.revision)
357
self._add_text_to_weave([], file_parents, weave_store, transaction)
450
359
def __eq__(self, other):
451
360
if not isinstance(other, InventoryEntry):
489
398
Note that this should be modified to be a noop on virtual trees
490
399
as all entries created there are prepopulated.
492
# TODO: Rather than running this manually, we should check the
493
# working sha1 and other expensive properties when they're
494
# first requested, or preload them if they're already known
495
pass # nothing to do by default
497
def _forget_tree_state(self):
501
403
class RootEntry(InventoryEntry):
503
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
504
'text_id', 'parent_id', 'children', 'executable',
505
'revision', 'symlink_target']
507
405
def _check(self, checker, rev_id, tree):
508
406
"""See InventoryEntry._check"""
510
408
def __init__(self, file_id):
511
409
self.file_id = file_id
512
410
self.children = {}
513
self.kind = 'directory'
411
self.kind = 'root_directory'
514
412
self.parent_id = None
517
warn('RootEntry is deprecated as of bzr 0.10. Please use '
518
'InventoryDirectory instead.',
519
DeprecationWarning, stacklevel=2)
521
415
def __eq__(self, other):
522
416
if not isinstance(other, RootEntry):
568
458
"""See InventoryEntry._put_on_disk."""
569
459
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)
576
462
class InventoryFile(InventoryEntry):
577
463
"""A file in an inventory."""
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):
465
def _check(self, checker, rev_id, tree):
584
466
"""See InventoryEntry._check"""
585
t = (self.file_id, self.revision)
467
revision = self.revision
468
t = (self.file_id, revision)
586
469
if t in checker.checked_texts:
587
prev_sha = checker.checked_texts[t]
470
prev_sha = checker.checked_texts[t]
588
471
if prev_sha != self.text_sha1:
589
472
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
590
(self.file_id, tree_revision_id))
473
(self.file_id, rev_id))
592
475
checker.repeated_text_cnt += 1
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))
477
mutter('check version {%s} of {%s}', rev_id, self.file_id)
478
file_lines = tree.get_file_lines(self.file_id)
479
checker.checked_text_cnt += 1
480
if self.text_size != sum(map(len, file_lines)):
481
raise BzrCheckError('text {%s} wrong size' % self.text_id)
482
if self.text_sha1 != sha_strings(file_lines):
483
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
614
484
checker.checked_texts[t] = self.text_sha1
633
503
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
634
504
output_to, reverse=False):
635
505
"""See InventoryEntry._diff."""
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
506
from_text = tree.get_file(self.file_id).readlines()
508
to_text = to_tree.get_file(to_entry.file_id).readlines()
512
text_diff(from_label, from_text,
513
to_label, to_text, output_to)
515
text_diff(to_label, to_text,
516
from_label, from_text, output_to)
655
518
def has_text(self):
656
519
"""See InventoryEntry.has_text."""
684
547
def _read_tree_state(self, path, work_tree):
685
548
"""See InventoryEntry._read_tree_state."""
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):
549
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
550
self.executable = work_tree.is_executable(self.file_id)
552
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
704
553
"""See InventoryEntry._snapshot_text."""
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)
554
mutter('storing file {%s} in revision {%s}',
555
self.file_id, self.revision)
556
# special case to avoid diffing on renames or
558
if (len(file_parents) == 1
559
and self.text_sha1 == file_parents.values()[0].text_sha1
560
and self.text_size == file_parents.values()[0].text_size):
561
previous_ie = file_parents.values()[0]
562
weave_store.add_identical_text(
563
self.file_id, previous_ie.revision,
564
self.revision, file_parents, transaction)
566
new_lines = work_tree.get_file(self.file_id).readlines()
567
self._add_text_to_weave(new_lines, file_parents, weave_store,
569
self.text_sha1 = sha_strings(new_lines)
570
self.text_size = sum(map(len, new_lines))
710
573
def _unchanged(self, previous_ie):
711
574
"""See InventoryEntry._unchanged."""
724
587
class InventoryLink(InventoryEntry):
725
588
"""A file in an inventory."""
727
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
728
'text_id', 'parent_id', 'children', 'executable',
729
'revision', 'symlink_target']
590
__slots__ = ['symlink_target']
731
592
def _check(self, checker, rev_id, tree):
732
593
"""See InventoryEntry._check"""
733
if self.text_sha1 is not None or self.text_size is not None or self.text_id is not None:
594
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
734
595
raise BzrCheckError('symlink {%s} has text in revision {%s}'
735
596
% (self.file_id, rev_id))
736
if self.symlink_target is None:
597
if self.symlink_target == None:
737
598
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
738
599
% (self.file_id, rev_id))
861
714
The inventory is created with a default root directory, with
864
# We are letting Branch.create() create a unique inventory
717
# We are letting Branch.initialize() create a unique inventory
865
718
# root id. Rather than generating a random one here.
866
719
#if root_id is None:
867
720
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
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):
721
self.root = RootEntry(root_id)
879
722
self._byid = {self.root.file_id: self.root}
882
# TODO: jam 20051218 Should copy also copy the revision_id?
883
entries = self.iter_entries()
884
other = Inventory(entries.next()[1].file_id)
726
other = Inventory(self.root.file_id)
885
727
# copy recursively so we know directories will be added before
886
728
# their children. There are more efficient ways than this...
887
for path, entry in entries():
729
for path, entry in self.iter_entries():
730
if entry == self.root:
888
732
other.add(entry.copy())
891
736
def __iter__(self):
892
737
return iter(self._byid)
894
740
def __len__(self):
895
741
"""Returns number of entries."""
896
742
return len(self._byid)
898
745
def iter_entries(self, from_dir=None):
899
746
"""Return (path, entry) pairs, in order by name."""
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))
750
elif isinstance(from_dir, basestring):
751
from_dir = self._byid[from_dir]
753
kids = from_dir.children.items()
755
for name, ie in kids:
757
if ie.kind == 'directory':
758
for cn, cie in self.iter_entries(from_dir=ie.file_id):
759
yield os.path.join(name, cn), cie
975
762
def entries(self):
976
763
"""Return list of (path, ie) for all entries except the root.
1003
791
for name, child_ie in kids:
1004
child_path = pathjoin(parent_path, name)
792
child_path = os.path.join(parent_path, name)
1005
793
descend(child_ie, child_path)
1006
descend(self.root, u'')
794
descend(self.root, '')
1009
799
def __contains__(self, file_id):
1010
800
"""True if this entry contains a file with given id.
1012
802
>>> inv = Inventory()
1013
803
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
1014
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT', sha1=None, len=None)
804
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
1015
805
>>> '123' in inv
1017
807
>>> '456' in inv
1020
return (file_id in self._byid)
810
return file_id in self._byid
1022
813
def __getitem__(self, file_id):
1023
814
"""Return the entry for given file_id.
1025
816
>>> inv = Inventory()
1026
817
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
1027
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT', sha1=None, len=None)
818
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
1028
819
>>> inv['123123'].name
1032
823
return self._byid[file_id]
1033
824
except KeyError:
1035
826
raise BzrError("can't look up file_id None")
1037
828
raise BzrError("file_id {%s} not in inventory" % file_id)
1039
831
def get_file_kind(self, file_id):
1040
832
return self._byid[file_id].kind
1042
834
def get_child(self, parent_id, filename):
1043
835
return self[parent_id].children.get(filename)
1045
838
def add(self, entry):
1046
839
"""Add entry to inventory.
1066
854
except KeyError:
1067
855
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
1069
if entry.name in parent.children:
857
if parent.children.has_key(entry.name):
1070
858
raise BzrError("%s is already versioned" %
1071
pathjoin(self.id2path(parent.file_id), entry.name))
859
appendpath(self.id2path(parent.file_id), entry.name))
1073
861
self._byid[entry.file_id] = entry
1074
862
parent.children[entry.name] = entry
1077
def add_path(self, relpath, kind, file_id=None, parent_id=None):
866
def add_path(self, relpath, kind, file_id=None):
1078
867
"""Add entry from a path.
1080
869
The immediate parent must already be versioned.
1082
871
Returns the new entry object."""
872
from bzrlib.branch import gen_file_id
1084
parts = osutils.splitpath(relpath)
874
parts = bzrlib.osutils.splitpath(relpath)
1086
875
if len(parts) == 0:
1088
file_id = bzrlib.workingtree.gen_root_id()
1089
self.root = InventoryDirectory(file_id, '', None)
1090
self._byid = {self.root.file_id: self.root}
876
raise BzrError("cannot re-add root of inventory")
879
file_id = gen_file_id(relpath)
881
parent_path = parts[:-1]
882
parent_id = self.path2id(parent_path)
883
if parent_id == None:
884
raise NotVersionedError(parent_path)
886
if kind == 'directory':
887
ie = InventoryDirectory(file_id, parts[-1], parent_id)
889
ie = InventoryFile(file_id, parts[-1], parent_id)
890
elif kind == 'symlink':
891
ie = InventoryLink(file_id, parts[-1], parent_id)
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)
893
raise BzrError("unknown kind %r" % kind)
1098
894
return self.add(ie)
1100
897
def __delitem__(self, file_id):
1101
898
"""Remove entry by id.
1103
900
>>> inv = Inventory()
1104
901
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
1105
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT', sha1=None, len=None)
902
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
1106
903
>>> '123' in inv
1108
905
>>> del inv['123']
1128
931
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
1129
InventoryFile('123', 'foo', parent_id='TREE_ROOT', sha1=None, len=None)
932
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
1132
935
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
1133
InventoryFile('123', 'foo', parent_id='TREE_ROOT', sha1=None, len=None)
936
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
1137
940
if not isinstance(other, Inventory):
1138
941
return NotImplemented
943
if len(self._byid) != len(other._byid):
944
# shortcut: obviously not the same
1140
947
return self._byid == other._byid
1142
950
def __ne__(self, other):
1143
951
return not self.__eq__(other)
1145
954
def __hash__(self):
1146
955
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
1158
958
def get_idpath(self, file_id):
1159
959
"""Return a list of file_ids for the path to an entry.
1164
964
root directory as depth 1.
1167
for parent in self._iter_file_id_parents(file_id):
1168
p.insert(0, parent.file_id)
967
while file_id != None:
969
ie = self._byid[file_id]
971
raise BzrError("file_id {%s} not found in inventory" % file_id)
972
p.insert(0, ie.file_id)
973
file_id = ie.parent_id
1171
977
def id2path(self, file_id):
1172
"""Return as a string the path to file_id.
978
"""Return as a list the path to file_id.
1174
980
>>> i = Inventory()
1175
981
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
1176
982
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
1177
>>> print i.id2path('foo-id')
983
>>> print i.id2path('foo-id').replace(os.sep, '/')
1180
986
# get all names, skipping root
1181
return '/'.join(reversed(
1182
[parent.name for parent in
1183
self._iter_file_id_parents(file_id)][:-1]))
987
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
988
return os.sep.join(p)
1185
992
def path2id(self, name):
1186
993
"""Walk down through directories to return entry of last component.
1212
1019
return parent.file_id
1214
1022
def has_filename(self, names):
1215
1023
return bool(self.path2id(names))
1217
1026
def has_id(self, file_id):
1218
return (file_id in self._byid)
1027
return self._byid.has_key(file_id)
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]
1238
1030
def rename(self, file_id, new_parent_id, new_name):
1239
1031
"""Move a file within the inventory.
1265
1057
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)
1298
1062
_NAME_RE = None
1300
1064
def is_valid_name(name):
1301
1065
global _NAME_RE
1302
if _NAME_RE is None:
1066
if _NAME_RE == None:
1303
1067
_NAME_RE = re.compile(r'^[^/\\]+$')
1305
1069
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
