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 __init__(self, file_id, name, kind, parent_id, text_id=None):
126
"""Create an InventoryEntry
128
The filename must be a single component, relative to the
129
parent directory; it cannot be a whole path or relative name.
131
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
136
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
137
Traceback (most recent call last):
138
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
140
assert isinstance(name, basestring), name
141
if '/' in name or '\\' in name:
142
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
144
self.executable = False
146
self.text_sha1 = None
147
self.text_size = None
148
self.file_id = file_id
151
self.text_id = text_id
152
self.parent_id = parent_id
153
self.symlink_target = None
154
if kind == 'directory':
158
elif kind == 'symlink':
161
raise BzrError("unhandled entry kind %r" % kind)
163
def read_symlink_target(self, path):
164
if self.kind == 'symlink':
166
self.symlink_target = os.readlink(path)
168
raise BzrError("os.readlink error, %s" % e)
170
def sorted_children(self):
171
l = self.children.items()
175
def check(self, checker, rev_id, inv, tree):
176
if self.parent_id != None:
177
if not inv.has_id(self.parent_id):
178
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
179
% (self.parent_id, rev_id))
180
if self.kind == 'file':
181
revision = self.revision
182
t = (self.file_id, revision)
183
if t in checker.checked_texts:
184
prev_sha = checker.checked_texts[t]
185
if prev_sha != self.text_sha1:
186
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
187
(self.file_id, rev_id))
189
checker.repeated_text_cnt += 1
191
mutter('check version {%s} of {%s}', rev_id, self.file_id)
192
file_lines = tree.get_file_lines(self.file_id)
193
checker.checked_text_cnt += 1
194
if self.text_size != sum(map(len, file_lines)):
195
raise BzrCheckError('text {%s} wrong size' % self.text_id)
196
if self.text_sha1 != sha_strings(file_lines):
197
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
198
checker.checked_texts[t] = self.text_sha1
199
elif self.kind == 'directory':
200
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
201
raise BzrCheckError('directory {%s} has text in revision {%s}'
202
% (self.file_id, rev_id))
203
elif self.kind == 'root_directory':
205
elif self.kind == 'symlink':
206
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
207
raise BzrCheckError('symlink {%s} has text in revision {%s}'
208
% (self.file_id, rev_id))
209
if self.symlink_target == None:
210
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
211
% (self.file_id, rev_id))
213
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
218
other = InventoryEntry(self.file_id, self.name, self.kind,
220
other.executable = self.executable
221
other.text_id = self.text_id
222
other.text_sha1 = self.text_sha1
223
other.text_size = self.text_size
224
other.symlink_target = self.symlink_target
225
other.revision = self.revision
226
# note that children are *not* copied; they're pulled across when
230
def _get_snapshot_change(self, previous_entries):
231
if len(previous_entries) > 1:
233
elif len(previous_entries) == 0:
236
return 'modified/renamed/reparented'
239
return ("%s(%r, %r, kind=%r, parent_id=%r)"
240
% (self.__class__.__name__,
246
def snapshot(self, revision, path, previous_entries, work_tree,
248
"""Make a snapshot of this entry.
250
This means that all its fields are populated, that it has its
251
text stored in the text store or weave.
253
mutter('new parents of %s are %r', path, previous_entries)
254
self._read_tree_state(path, work_tree)
255
if len(previous_entries) == 1:
256
# cannot be unchanged unless there is only one parent file rev.
257
parent_ie = previous_entries.values()[0]
258
if self._unchanged(path, parent_ie, work_tree):
259
mutter("found unchanged entry")
260
self.revision = parent_ie.revision
262
mutter('new revision for {%s}', self.file_id)
263
self.revision = revision
264
change = self._get_snapshot_change(previous_entries)
265
if self.kind != 'file':
267
self._snapshot_text(previous_entries, work_tree, weave_store)
270
def _snapshot_text(self, file_parents, work_tree, weave_store):
271
mutter('storing file {%s} in revision {%s}',
272
self.file_id, self.revision)
273
# special case to avoid diffing on renames or
275
if (len(file_parents) == 1
276
and self.text_sha1 == file_parents.values()[0].text_sha1
277
and self.text_size == file_parents.values()[0].text_size):
278
previous_ie = file_parents.values()[0]
279
weave_store.add_identical_text(
280
self.file_id, previous_ie.revision,
281
self.revision, file_parents)
283
new_lines = work_tree.get_file(self.file_id).readlines()
284
self._add_text_to_weave(new_lines, file_parents, weave_store)
285
self.text_sha1 = sha_strings(new_lines)
286
self.text_size = sum(map(len, new_lines))
288
def __eq__(self, other):
289
if not isinstance(other, InventoryEntry):
290
return NotImplemented
292
return ((self.file_id == other.file_id)
293
and (self.name == other.name)
294
and (other.symlink_target == self.symlink_target)
295
and (self.text_sha1 == other.text_sha1)
296
and (self.text_size == other.text_size)
297
and (self.text_id == other.text_id)
298
and (self.parent_id == other.parent_id)
299
and (self.kind == other.kind)
300
and (self.revision == other.revision)
301
and (self.executable == other.executable)
304
def __ne__(self, other):
305
return not (self == other)
308
raise ValueError('not hashable')
310
def _unchanged(self, path, previous_ie, work_tree):
312
# different inv parent
313
if previous_ie.parent_id != self.parent_id:
316
elif previous_ie.name != self.name:
318
if self.kind == 'symlink':
319
if self.symlink_target != previous_ie.symlink_target:
321
if self.kind == 'file':
322
if self.text_sha1 != previous_ie.text_sha1:
325
# FIXME: 20050930 probe for the text size when getting sha1
326
# in _read_tree_state
327
self.text_size = previous_ie.text_size
330
def _read_tree_state(self, path, work_tree):
331
if self.kind == 'symlink':
332
self.read_symlink_target(work_tree.abspath(path))
333
if self.kind == 'file':
334
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
335
self.executable = work_tree.is_executable(self.file_id)
338
class RootEntry(InventoryEntry):
339
def __init__(self, file_id):
340
self.file_id = file_id
342
self.kind = 'root_directory'
343
self.parent_id = None
346
def __eq__(self, other):
347
if not isinstance(other, RootEntry):
348
return NotImplemented
350
return (self.file_id == other.file_id) \
351
and (self.children == other.children)
355
class Inventory(object):
356
"""Inventory of versioned files in a tree.
358
This describes which file_id is present at each point in the tree,
359
and possibly the SHA-1 or other information about the file.
360
Entries can be looked up either by path or by file_id.
362
The inventory represents a typical unix file tree, with
363
directories containing files and subdirectories. We never store
364
the full path to a file, because renaming a directory implicitly
365
moves all of its contents. This class internally maintains a
366
lookup tree that allows the children under a directory to be
369
InventoryEntry objects must not be modified after they are
370
inserted, other than through the Inventory API.
372
>>> inv = Inventory()
373
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
374
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT')
375
>>> inv['123-123'].name
378
May be treated as an iterator or set to look up file ids:
380
>>> bool(inv.path2id('hello.c'))
385
May also look up by name:
387
>>> [x[0] for x in inv.iter_entries()]
389
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
390
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
391
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
393
def __init__(self, root_id=ROOT_ID):
394
"""Create or read an inventory.
396
If a working directory is specified, the inventory is read
397
from there. If the file is specified, read from that. If not,
398
the inventory is created empty.
400
The inventory is created with a default root directory, with
403
# We are letting Branch.initialize() create a unique inventory
404
# root id. Rather than generating a random one here.
406
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
407
self.root = RootEntry(root_id)
408
self._byid = {self.root.file_id: self.root}
412
other = Inventory(self.root.file_id)
413
# copy recursively so we know directories will be added before
414
# their children. There are more efficient ways than this...
415
for path, entry in self.iter_entries():
416
if entry == self.root:
418
other.add(entry.copy())
423
return iter(self._byid)
427
"""Returns number of entries."""
428
return len(self._byid)
431
def iter_entries(self, from_dir=None):
432
"""Return (path, entry) pairs, in order by name."""
436
elif isinstance(from_dir, basestring):
437
from_dir = self._byid[from_dir]
439
kids = from_dir.children.items()
441
for name, ie in kids:
443
if ie.kind == 'directory':
444
for cn, cie in self.iter_entries(from_dir=ie.file_id):
445
yield os.path.join(name, cn), cie
449
"""Return list of (path, ie) for all entries except the root.
451
This may be faster than iter_entries.
454
def descend(dir_ie, dir_path):
455
kids = dir_ie.children.items()
457
for name, ie in kids:
458
child_path = os.path.join(dir_path, name)
459
accum.append((child_path, ie))
460
if ie.kind == 'directory':
461
descend(ie, child_path)
463
descend(self.root, '')
467
def directories(self):
468
"""Return (path, entry) pairs for all directories, including the root.
471
def descend(parent_ie, parent_path):
472
accum.append((parent_path, parent_ie))
474
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
477
for name, child_ie in kids:
478
child_path = os.path.join(parent_path, name)
479
descend(child_ie, child_path)
480
descend(self.root, '')
485
def __contains__(self, file_id):
486
"""True if this entry contains a file with given id.
488
>>> inv = Inventory()
489
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
490
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
496
return file_id in self._byid
499
def __getitem__(self, file_id):
500
"""Return the entry for given file_id.
502
>>> inv = Inventory()
503
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
504
InventoryEntry('123123', 'hello.c', kind='file', parent_id='TREE_ROOT')
505
>>> inv['123123'].name
509
return self._byid[file_id]
512
raise BzrError("can't look up file_id None")
514
raise BzrError("file_id {%s} not in inventory" % file_id)
517
def get_file_kind(self, file_id):
518
return self._byid[file_id].kind
520
def get_child(self, parent_id, filename):
521
return self[parent_id].children.get(filename)
524
def add(self, entry):
525
"""Add entry to inventory.
527
To add a file to a branch ready to be committed, use Branch.add,
530
Returns the new entry object.
532
if entry.file_id in self._byid:
533
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
535
if entry.parent_id == ROOT_ID or entry.parent_id is None:
536
entry.parent_id = self.root.file_id
539
parent = self._byid[entry.parent_id]
541
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
543
if parent.children.has_key(entry.name):
544
raise BzrError("%s is already versioned" %
545
appendpath(self.id2path(parent.file_id), entry.name))
547
self._byid[entry.file_id] = entry
548
parent.children[entry.name] = entry
552
def add_path(self, relpath, kind, file_id=None):
553
"""Add entry from a path.
555
The immediate parent must already be versioned.
557
Returns the new entry object."""
558
from bzrlib.branch import gen_file_id
560
parts = bzrlib.osutils.splitpath(relpath)
562
raise BzrError("cannot re-add root of inventory")
565
file_id = gen_file_id(relpath)
567
parent_path = parts[:-1]
568
parent_id = self.path2id(parent_path)
569
if parent_id == None:
570
raise NotVersionedError(parent_path)
572
ie = InventoryEntry(file_id, parts[-1],
573
kind=kind, parent_id=parent_id)
577
def __delitem__(self, file_id):
578
"""Remove entry by id.
580
>>> inv = Inventory()
581
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
582
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
591
assert self[ie.parent_id].children[ie.name] == ie
593
# TODO: Test deleting all children; maybe hoist to a separate
595
if ie.kind == 'directory':
596
for cie in ie.children.values():
597
del self[cie.file_id]
600
del self._byid[file_id]
601
del self[ie.parent_id].children[ie.name]
604
def __eq__(self, other):
605
"""Compare two sets by comparing their contents.
611
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
612
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
615
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
616
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
620
if not isinstance(other, Inventory):
621
return NotImplemented
623
if len(self._byid) != len(other._byid):
624
# shortcut: obviously not the same
627
return self._byid == other._byid
630
def __ne__(self, other):
631
return not self.__eq__(other)
635
raise ValueError('not hashable')
638
def get_idpath(self, file_id):
639
"""Return a list of file_ids for the path to an entry.
641
The list contains one element for each directory followed by
642
the id of the file itself. So the length of the returned list
643
is equal to the depth of the file in the tree, counting the
644
root directory as depth 1.
647
while file_id != None:
649
ie = self._byid[file_id]
651
raise BzrError("file_id {%s} not found in inventory" % file_id)
652
p.insert(0, ie.file_id)
653
file_id = ie.parent_id
657
def id2path(self, file_id):
658
"""Return as a list the path to file_id."""
660
# get all names, skipping root
661
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
662
return os.sep.join(p)
666
def path2id(self, name):
667
"""Walk down through directories to return entry of last component.
669
names may be either a list of path components, or a single
670
string, in which case it is automatically split.
672
This returns the entry of the last component in the path,
673
which may be either a file or a directory.
675
Returns None iff the path is not found.
677
if isinstance(name, types.StringTypes):
678
name = splitpath(name)
680
mutter("lookup path %r" % name)
685
cie = parent.children[f]
687
assert cie.parent_id == parent.file_id
693
return parent.file_id
696
def has_filename(self, names):
697
return bool(self.path2id(names))
700
def has_id(self, file_id):
701
return self._byid.has_key(file_id)
704
def rename(self, file_id, new_parent_id, new_name):
705
"""Move a file within the inventory.
707
This can change either the name, or the parent, or both.
709
This does not move the working file."""
710
if not is_valid_name(new_name):
711
raise BzrError("not an acceptable filename: %r" % new_name)
713
new_parent = self._byid[new_parent_id]
714
if new_name in new_parent.children:
715
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
717
new_parent_idpath = self.get_idpath(new_parent_id)
718
if file_id in new_parent_idpath:
719
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
720
% (self.id2path(file_id), self.id2path(new_parent_id)))
722
file_ie = self._byid[file_id]
723
old_parent = self._byid[file_ie.parent_id]
725
# TODO: Don't leave things messed up if this fails
727
del old_parent.children[file_ie.name]
728
new_parent.children[new_name] = file_ie
730
file_ie.name = new_name
731
file_ie.parent_id = new_parent_id
738
def is_valid_name(name):
741
_NAME_RE = re.compile(r'^[^/\\]+$')
743
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
