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
18
# TODO: Maybe also keep the full path of the entry, and the children?
19
# But those depend on its position within a particular inventory, and
20
# it would be nice not to need to hold the backpointer here.
22
# TODO: Perhaps split InventoryEntry into subclasses for files,
23
# directories, etc etc.
26
# This should really be an id randomly assigned when the tree is
27
# created, but it's not for now.
37
from bzrlib.errors import BzrError, BzrCheckError
39
from bzrlib.osutils import quotefn, splitpath, joinpath, appendpath, sha_strings
40
from bzrlib.trace import mutter
41
from bzrlib.errors import NotVersionedError
44
class InventoryEntry(object):
45
"""Description of a versioned file.
47
An InventoryEntry has the following fields, which are also
48
present in the XML inventory-entry element:
53
(within the parent directory)
56
'directory' or 'file' or 'symlink'
59
file_id of the parent directory, or ROOT_ID
62
the revision_id in which this variation of this file was
66
Indicates that this file should be executable on systems
70
sha-1 of the text of the file
73
size in bytes of the text of the file
75
(reading a version 4 tree created a text_id field.)
80
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
81
InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT')
82
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
83
InventoryEntry('2323', 'hello.c', kind='file', parent_id='123')
84
>>> for j in i.iter_entries():
87
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
88
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
89
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
90
Traceback (most recent call last):
92
BzrError: inventory already contains entry with id {2323}
93
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
94
InventoryEntry('2324', 'bye.c', kind='file', parent_id='123')
95
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
96
InventoryEntry('2325', 'wibble', kind='directory', parent_id='123')
97
>>> i.path2id('src/wibble')
101
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
102
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
104
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
105
>>> for path, entry in i.iter_entries():
106
... print path.replace('\\\\', '/') # for win32 os.sep
107
... assert i.path2id(path)
114
>>> i.id2path('2326').replace('\\\\', '/')
115
'src/wibble/wibble.c'
118
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
119
'text_id', 'parent_id', 'children', 'executable',
120
'revision', 'symlink_target']
122
def _add_text_to_weave(self, new_lines, parents, weave_store):
123
weave_store.add_text(self.file_id, self.revision, new_lines, parents)
125
def detect_changes(self, old_entry):
126
"""Return a (text_modified, meta_modified) from this to old_entry.
128
_read_tree_state must have been called on self and old_entry prior to
129
calling detect_changes.
131
if self.kind == 'file':
132
assert self.text_sha1 != None
133
assert old_entry.text_sha1 != None
134
text_modified = (self.text_sha1 != old_entry.text_sha1)
135
meta_modified = (self.executable != old_entry.executable)
136
elif self.kind == 'symlink':
137
# FIXME: which _modified field should we use ? RBC 20051003
138
text_modified = (self.symlink_target != old_entry.symlink_target)
140
mutter(" symlink target changed")
141
meta_modified = False
143
text_modified = False
144
meta_modified = False
145
return text_modified, meta_modified
147
def __init__(self, file_id, name, kind, parent_id, text_id=None):
148
"""Create an InventoryEntry
150
The filename must be a single component, relative to the
151
parent directory; it cannot be a whole path or relative name.
153
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
158
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
159
Traceback (most recent call last):
160
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
162
assert isinstance(name, basestring), name
163
if '/' in name or '\\' in name:
164
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
166
self.executable = False
168
self.text_sha1 = None
169
self.text_size = None
170
self.file_id = file_id
173
self.text_id = text_id
174
self.parent_id = parent_id
175
self.symlink_target = None
176
if kind == 'directory':
180
elif kind == 'symlink':
183
raise BzrError("unhandled entry kind %r" % kind)
185
def kind_character(self):
186
"""Return a short kind indicator useful for appending to names."""
187
if self.kind == 'directory':
189
if self.kind == 'file':
191
if self.kind == 'symlink':
193
raise RuntimeError('unreachable code')
195
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
197
def sorted_children(self):
198
l = self.children.items()
203
def versionable_kind(kind):
204
return kind in ('file', 'directory', 'symlink')
206
def check(self, checker, rev_id, inv, tree):
207
if self.parent_id != None:
208
if not inv.has_id(self.parent_id):
209
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
210
% (self.parent_id, rev_id))
211
if self.kind == 'file':
212
revision = self.revision
213
t = (self.file_id, revision)
214
if t in checker.checked_texts:
215
prev_sha = checker.checked_texts[t]
216
if prev_sha != self.text_sha1:
217
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
218
(self.file_id, rev_id))
220
checker.repeated_text_cnt += 1
222
mutter('check version {%s} of {%s}', rev_id, self.file_id)
223
file_lines = tree.get_file_lines(self.file_id)
224
checker.checked_text_cnt += 1
225
if self.text_size != sum(map(len, file_lines)):
226
raise BzrCheckError('text {%s} wrong size' % self.text_id)
227
if self.text_sha1 != sha_strings(file_lines):
228
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
229
checker.checked_texts[t] = self.text_sha1
230
elif self.kind == 'directory':
231
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
232
raise BzrCheckError('directory {%s} has text in revision {%s}'
233
% (self.file_id, rev_id))
234
elif self.kind == 'root_directory':
236
elif self.kind == 'symlink':
237
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
238
raise BzrCheckError('symlink {%s} has text in revision {%s}'
239
% (self.file_id, rev_id))
240
if self.symlink_target == None:
241
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
242
% (self.file_id, rev_id))
244
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
249
other = InventoryEntry(self.file_id, self.name, self.kind,
251
other.executable = self.executable
252
other.text_id = self.text_id
253
other.text_sha1 = self.text_sha1
254
other.text_size = self.text_size
255
other.symlink_target = self.symlink_target
256
other.revision = self.revision
257
# note that children are *not* copied; they're pulled across when
261
def _get_snapshot_change(self, previous_entries):
262
if len(previous_entries) > 1:
264
elif len(previous_entries) == 0:
267
return 'modified/renamed/reparented'
270
return ("%s(%r, %r, kind=%r, parent_id=%r)"
271
% (self.__class__.__name__,
277
def snapshot(self, revision, path, previous_entries, work_tree,
279
"""Make a snapshot of this entry.
281
This means that all its fields are populated, that it has its
282
text stored in the text store or weave.
284
mutter('new parents of %s are %r', path, previous_entries)
285
self._read_tree_state(path, work_tree)
286
if len(previous_entries) == 1:
287
# cannot be unchanged unless there is only one parent file rev.
288
parent_ie = previous_entries.values()[0]
289
if self._unchanged(path, parent_ie, work_tree):
290
mutter("found unchanged entry")
291
self.revision = parent_ie.revision
293
mutter('new revision for {%s}', self.file_id)
294
self.revision = revision
295
change = self._get_snapshot_change(previous_entries)
296
if self.kind != 'file':
298
self._snapshot_text(previous_entries, work_tree, weave_store)
301
def _snapshot_text(self, file_parents, work_tree, weave_store):
302
mutter('storing file {%s} in revision {%s}',
303
self.file_id, self.revision)
304
# special case to avoid diffing on renames or
306
if (len(file_parents) == 1
307
and self.text_sha1 == file_parents.values()[0].text_sha1
308
and self.text_size == file_parents.values()[0].text_size):
309
previous_ie = file_parents.values()[0]
310
weave_store.add_identical_text(
311
self.file_id, previous_ie.revision,
312
self.revision, file_parents)
314
new_lines = work_tree.get_file(self.file_id).readlines()
315
self._add_text_to_weave(new_lines, file_parents, weave_store)
316
self.text_sha1 = sha_strings(new_lines)
317
self.text_size = sum(map(len, new_lines))
319
def __eq__(self, other):
320
if not isinstance(other, InventoryEntry):
321
return NotImplemented
323
return ((self.file_id == other.file_id)
324
and (self.name == other.name)
325
and (other.symlink_target == self.symlink_target)
326
and (self.text_sha1 == other.text_sha1)
327
and (self.text_size == other.text_size)
328
and (self.text_id == other.text_id)
329
and (self.parent_id == other.parent_id)
330
and (self.kind == other.kind)
331
and (self.revision == other.revision)
332
and (self.executable == other.executable)
335
def __ne__(self, other):
336
return not (self == other)
339
raise ValueError('not hashable')
341
def _unchanged(self, path, previous_ie, work_tree):
343
# different inv parent
344
if previous_ie.parent_id != self.parent_id:
347
elif previous_ie.name != self.name:
349
if self.kind == 'symlink':
350
if self.symlink_target != previous_ie.symlink_target:
352
if self.kind == 'file':
353
if self.text_sha1 != previous_ie.text_sha1:
356
# FIXME: 20050930 probe for the text size when getting sha1
357
# in _read_tree_state
358
self.text_size = previous_ie.text_size
361
def _read_tree_state(self, path, work_tree):
362
if self.kind == 'symlink':
363
self.symlink_target = work_tree.get_symlink_target(self.file_id)
364
if self.kind == 'file':
365
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
366
self.executable = work_tree.is_executable(self.file_id)
369
class RootEntry(InventoryEntry):
370
def __init__(self, file_id):
371
self.file_id = file_id
373
self.kind = 'root_directory'
374
self.parent_id = None
377
def __eq__(self, other):
378
if not isinstance(other, RootEntry):
379
return NotImplemented
381
return (self.file_id == other.file_id) \
382
and (self.children == other.children)
386
class Inventory(object):
387
"""Inventory of versioned files in a tree.
389
This describes which file_id is present at each point in the tree,
390
and possibly the SHA-1 or other information about the file.
391
Entries can be looked up either by path or by file_id.
393
The inventory represents a typical unix file tree, with
394
directories containing files and subdirectories. We never store
395
the full path to a file, because renaming a directory implicitly
396
moves all of its contents. This class internally maintains a
397
lookup tree that allows the children under a directory to be
400
InventoryEntry objects must not be modified after they are
401
inserted, other than through the Inventory API.
403
>>> inv = Inventory()
404
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
405
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT')
406
>>> inv['123-123'].name
409
May be treated as an iterator or set to look up file ids:
411
>>> bool(inv.path2id('hello.c'))
416
May also look up by name:
418
>>> [x[0] for x in inv.iter_entries()]
420
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
421
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
422
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
424
def __init__(self, root_id=ROOT_ID):
425
"""Create or read an inventory.
427
If a working directory is specified, the inventory is read
428
from there. If the file is specified, read from that. If not,
429
the inventory is created empty.
431
The inventory is created with a default root directory, with
434
# We are letting Branch.initialize() create a unique inventory
435
# root id. Rather than generating a random one here.
437
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
438
self.root = RootEntry(root_id)
439
self._byid = {self.root.file_id: self.root}
443
other = Inventory(self.root.file_id)
444
# copy recursively so we know directories will be added before
445
# their children. There are more efficient ways than this...
446
for path, entry in self.iter_entries():
447
if entry == self.root:
449
other.add(entry.copy())
454
return iter(self._byid)
458
"""Returns number of entries."""
459
return len(self._byid)
462
def iter_entries(self, from_dir=None):
463
"""Return (path, entry) pairs, in order by name."""
467
elif isinstance(from_dir, basestring):
468
from_dir = self._byid[from_dir]
470
kids = from_dir.children.items()
472
for name, ie in kids:
474
if ie.kind == 'directory':
475
for cn, cie in self.iter_entries(from_dir=ie.file_id):
476
yield os.path.join(name, cn), cie
480
"""Return list of (path, ie) for all entries except the root.
482
This may be faster than iter_entries.
485
def descend(dir_ie, dir_path):
486
kids = dir_ie.children.items()
488
for name, ie in kids:
489
child_path = os.path.join(dir_path, name)
490
accum.append((child_path, ie))
491
if ie.kind == 'directory':
492
descend(ie, child_path)
494
descend(self.root, '')
498
def directories(self):
499
"""Return (path, entry) pairs for all directories, including the root.
502
def descend(parent_ie, parent_path):
503
accum.append((parent_path, parent_ie))
505
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
508
for name, child_ie in kids:
509
child_path = os.path.join(parent_path, name)
510
descend(child_ie, child_path)
511
descend(self.root, '')
516
def __contains__(self, file_id):
517
"""True if this entry contains a file with given id.
519
>>> inv = Inventory()
520
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
521
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
527
return file_id in self._byid
530
def __getitem__(self, file_id):
531
"""Return the entry for given file_id.
533
>>> inv = Inventory()
534
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
535
InventoryEntry('123123', 'hello.c', kind='file', parent_id='TREE_ROOT')
536
>>> inv['123123'].name
540
return self._byid[file_id]
543
raise BzrError("can't look up file_id None")
545
raise BzrError("file_id {%s} not in inventory" % file_id)
548
def get_file_kind(self, file_id):
549
return self._byid[file_id].kind
551
def get_child(self, parent_id, filename):
552
return self[parent_id].children.get(filename)
555
def add(self, entry):
556
"""Add entry to inventory.
558
To add a file to a branch ready to be committed, use Branch.add,
561
Returns the new entry object.
563
if entry.file_id in self._byid:
564
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
566
if entry.parent_id == ROOT_ID or entry.parent_id is None:
567
entry.parent_id = self.root.file_id
570
parent = self._byid[entry.parent_id]
572
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
574
if parent.children.has_key(entry.name):
575
raise BzrError("%s is already versioned" %
576
appendpath(self.id2path(parent.file_id), entry.name))
578
self._byid[entry.file_id] = entry
579
parent.children[entry.name] = entry
583
def add_path(self, relpath, kind, file_id=None):
584
"""Add entry from a path.
586
The immediate parent must already be versioned.
588
Returns the new entry object."""
589
from bzrlib.branch import gen_file_id
591
parts = bzrlib.osutils.splitpath(relpath)
593
raise BzrError("cannot re-add root of inventory")
596
file_id = gen_file_id(relpath)
598
parent_path = parts[:-1]
599
parent_id = self.path2id(parent_path)
600
if parent_id == None:
601
raise NotVersionedError(parent_path)
603
ie = InventoryEntry(file_id, parts[-1],
604
kind=kind, parent_id=parent_id)
608
def __delitem__(self, file_id):
609
"""Remove entry by id.
611
>>> inv = Inventory()
612
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
613
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
622
assert self[ie.parent_id].children[ie.name] == ie
624
# TODO: Test deleting all children; maybe hoist to a separate
626
if ie.kind == 'directory':
627
for cie in ie.children.values():
628
del self[cie.file_id]
631
del self._byid[file_id]
632
del self[ie.parent_id].children[ie.name]
635
def __eq__(self, other):
636
"""Compare two sets by comparing their contents.
642
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
643
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
646
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
647
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
651
if not isinstance(other, Inventory):
652
return NotImplemented
654
if len(self._byid) != len(other._byid):
655
# shortcut: obviously not the same
658
return self._byid == other._byid
661
def __ne__(self, other):
662
return not self.__eq__(other)
666
raise ValueError('not hashable')
669
def get_idpath(self, file_id):
670
"""Return a list of file_ids for the path to an entry.
672
The list contains one element for each directory followed by
673
the id of the file itself. So the length of the returned list
674
is equal to the depth of the file in the tree, counting the
675
root directory as depth 1.
678
while file_id != None:
680
ie = self._byid[file_id]
682
raise BzrError("file_id {%s} not found in inventory" % file_id)
683
p.insert(0, ie.file_id)
684
file_id = ie.parent_id
688
def id2path(self, file_id):
689
"""Return as a list the path to file_id."""
691
# get all names, skipping root
692
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
693
return os.sep.join(p)
697
def path2id(self, name):
698
"""Walk down through directories to return entry of last component.
700
names may be either a list of path components, or a single
701
string, in which case it is automatically split.
703
This returns the entry of the last component in the path,
704
which may be either a file or a directory.
706
Returns None iff the path is not found.
708
if isinstance(name, types.StringTypes):
709
name = splitpath(name)
711
mutter("lookup path %r" % name)
716
cie = parent.children[f]
718
assert cie.parent_id == parent.file_id
724
return parent.file_id
727
def has_filename(self, names):
728
return bool(self.path2id(names))
731
def has_id(self, file_id):
732
return self._byid.has_key(file_id)
735
def rename(self, file_id, new_parent_id, new_name):
736
"""Move a file within the inventory.
738
This can change either the name, or the parent, or both.
740
This does not move the working file."""
741
if not is_valid_name(new_name):
742
raise BzrError("not an acceptable filename: %r" % new_name)
744
new_parent = self._byid[new_parent_id]
745
if new_name in new_parent.children:
746
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
748
new_parent_idpath = self.get_idpath(new_parent_id)
749
if file_id in new_parent_idpath:
750
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
751
% (self.id2path(file_id), self.id2path(new_parent_id)))
753
file_ie = self._byid[file_id]
754
old_parent = self._byid[file_ie.parent_id]
756
# TODO: Don't leave things messed up if this fails
758
del old_parent.children[file_ie.name]
759
new_parent.children[new_name] = file_ie
761
file_ie.name = new_name
762
file_ie.parent_id = new_parent_id
769
def is_valid_name(name):
772
_NAME_RE = re.compile(r'^[^/\\]+$')
774
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
