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)
59
file_id of the parent directory, or ROOT_ID
62
the revision_id in which this variationo f this file was
66
sha-1 of the text of the file
69
size in bytes of the text of the file
71
(reading a version 4 tree created a text_id field.)
76
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
77
InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT')
78
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
79
InventoryEntry('2323', 'hello.c', kind='file', parent_id='123')
80
>>> for j in i.iter_entries():
83
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
84
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
85
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
86
Traceback (most recent call last):
88
BzrError: inventory already contains entry with id {2323}
89
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
90
InventoryEntry('2324', 'bye.c', kind='file', parent_id='123')
91
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
92
InventoryEntry('2325', 'wibble', kind='directory', parent_id='123')
93
>>> i.path2id('src/wibble')
97
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
98
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
100
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
101
>>> for path, entry in i.iter_entries():
102
... print path.replace('\\\\', '/') # for win32 os.sep
103
... assert i.path2id(path)
110
>>> i.id2path('2326').replace('\\\\', '/')
111
'src/wibble/wibble.c'
114
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
115
'text_id', 'parent_id', 'children',
116
'revision', 'symlink_target']
118
def __init__(self, file_id, name, kind, parent_id, text_id=None):
119
"""Create an InventoryEntry
121
The filename must be a single component, relative to the
122
parent directory; it cannot be a whole path or relative name.
124
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
129
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
130
Traceback (most recent call last):
131
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
133
assert isinstance(name, basestring), name
134
if '/' in name or '\\' in name:
135
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
138
self.text_sha1 = None
139
self.text_size = None
140
self.file_id = file_id
143
self.text_id = text_id
144
self.parent_id = parent_id
145
self.symlink_target = None
146
if kind == 'directory':
150
elif kind == 'symlink':
153
raise BzrError("unhandled entry kind %r" % kind)
155
def read_symlink_target(self, path):
156
if self.kind == 'symlink':
158
self.symlink_target = os.readlink(path)
160
raise BzrError("os.readlink error, %s" % e)
162
def sorted_children(self):
163
l = self.children.items()
167
def check(self, checker, rev_id, inv, tree):
168
if self.parent_id != None:
169
if not inv.has_id(self.parent_id):
170
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
171
% (self.parent_id, rev_id))
172
if self.kind == 'file':
173
revision = self.revision
174
t = (self.file_id, revision)
175
if t in checker.checked_texts:
176
prev_sha = checker.checked_texts[t]
177
if prev_sha != self.text_sha1:
178
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
179
(self.file_id, rev_id))
181
checker.repeated_text_cnt += 1
183
mutter('check version {%s} of {%s}', rev_id, self.file_id)
184
file_lines = tree.get_file_lines(self.file_id)
185
checker.checked_text_cnt += 1
186
if self.text_size != sum(map(len, file_lines)):
187
raise BzrCheckError('text {%s} wrong size' % self.text_id)
188
if self.text_sha1 != sha_strings(file_lines):
189
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
190
checker.checked_texts[t] = self.text_sha1
191
elif self.kind == 'directory':
192
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
193
raise BzrCheckError('directory {%s} has text in revision {%s}'
194
% (self.file_id, rev_id))
195
elif self.kind == 'root_directory':
197
elif self.kind == 'symlink':
198
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
199
raise BzrCheckError('symlink {%s} has text in revision {%s}'
200
% (self.file_id, rev_id))
201
if self.symlink_target == None:
202
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
203
% (self.file_id, rev_id))
205
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
210
other = InventoryEntry(self.file_id, self.name, self.kind,
212
other.text_id = self.text_id
213
other.text_sha1 = self.text_sha1
214
other.text_size = self.text_size
215
other.symlink_target = self.symlink_target
216
other.revision = self.revision
217
# note that children are *not* copied; they're pulled across when
223
return ("%s(%r, %r, kind=%r, parent_id=%r)"
224
% (self.__class__.__name__,
230
def __eq__(self, other):
231
if not isinstance(other, InventoryEntry):
232
return NotImplemented
234
return (self.file_id == other.file_id) \
235
and (self.name == other.name) \
236
and (other.symlink_target == self.symlink_target) \
237
and (self.text_sha1 == other.text_sha1) \
238
and (self.text_size == other.text_size) \
239
and (self.text_id == other.text_id) \
240
and (self.parent_id == other.parent_id) \
241
and (self.kind == other.kind) \
242
and (self.revision == other.revision)
244
def __ne__(self, other):
245
return not (self == other)
248
raise ValueError('not hashable')
250
def unchanged(self, previous_ie, work_tree):
252
# different inv parent
253
if previous_ie.parent_id != self.parent_id:
256
elif previous_ie.name != self.name:
258
# changed link target
259
elif (hasattr(self,'symlink_target')
260
and self.symlink_target != previous_ie.symlink_target):
262
if self.kind != 'file':
264
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
265
if self.text_sha1 != previous_ie.text_sha1:
268
# FIXME: 20050930 probe for the text size when getting sha1
269
self.text_size = previous_ie.text_size
273
class RootEntry(InventoryEntry):
274
def __init__(self, file_id):
275
self.file_id = file_id
277
self.kind = 'root_directory'
278
self.parent_id = None
281
def __eq__(self, other):
282
if not isinstance(other, RootEntry):
283
return NotImplemented
285
return (self.file_id == other.file_id) \
286
and (self.children == other.children)
290
class Inventory(object):
291
"""Inventory of versioned files in a tree.
293
This describes which file_id is present at each point in the tree,
294
and possibly the SHA-1 or other information about the file.
295
Entries can be looked up either by path or by file_id.
297
The inventory represents a typical unix file tree, with
298
directories containing files and subdirectories. We never store
299
the full path to a file, because renaming a directory implicitly
300
moves all of its contents. This class internally maintains a
301
lookup tree that allows the children under a directory to be
304
InventoryEntry objects must not be modified after they are
305
inserted, other than through the Inventory API.
307
>>> inv = Inventory()
308
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
309
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT')
310
>>> inv['123-123'].name
313
May be treated as an iterator or set to look up file ids:
315
>>> bool(inv.path2id('hello.c'))
320
May also look up by name:
322
>>> [x[0] for x in inv.iter_entries()]
324
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
325
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
326
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
328
def __init__(self, root_id=ROOT_ID):
329
"""Create or read an inventory.
331
If a working directory is specified, the inventory is read
332
from there. If the file is specified, read from that. If not,
333
the inventory is created empty.
335
The inventory is created with a default root directory, with
338
# We are letting Branch.initialize() create a unique inventory
339
# root id. Rather than generating a random one here.
341
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
342
self.root = RootEntry(root_id)
343
self._byid = {self.root.file_id: self.root}
347
other = Inventory(self.root.file_id)
348
# copy recursively so we know directories will be added before
349
# their children. There are more efficient ways than this...
350
for path, entry in self.iter_entries():
351
if entry == self.root:
353
other.add(entry.copy())
358
return iter(self._byid)
362
"""Returns number of entries."""
363
return len(self._byid)
366
def iter_entries(self, from_dir=None):
367
"""Return (path, entry) pairs, in order by name."""
371
elif isinstance(from_dir, basestring):
372
from_dir = self._byid[from_dir]
374
kids = from_dir.children.items()
376
for name, ie in kids:
378
if ie.kind == 'directory':
379
for cn, cie in self.iter_entries(from_dir=ie.file_id):
380
yield os.path.join(name, cn), cie
384
"""Return list of (path, ie) for all entries except the root.
386
This may be faster than iter_entries.
389
def descend(dir_ie, dir_path):
390
kids = dir_ie.children.items()
392
for name, ie in kids:
393
child_path = os.path.join(dir_path, name)
394
accum.append((child_path, ie))
395
if ie.kind == 'directory':
396
descend(ie, child_path)
398
descend(self.root, '')
402
def directories(self):
403
"""Return (path, entry) pairs for all directories, including the root.
406
def descend(parent_ie, parent_path):
407
accum.append((parent_path, parent_ie))
409
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
412
for name, child_ie in kids:
413
child_path = os.path.join(parent_path, name)
414
descend(child_ie, child_path)
415
descend(self.root, '')
420
def __contains__(self, file_id):
421
"""True if this entry contains a file with given id.
423
>>> inv = Inventory()
424
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
425
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
431
return file_id in self._byid
434
def __getitem__(self, file_id):
435
"""Return the entry for given file_id.
437
>>> inv = Inventory()
438
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
439
InventoryEntry('123123', 'hello.c', kind='file', parent_id='TREE_ROOT')
440
>>> inv['123123'].name
444
return self._byid[file_id]
447
raise BzrError("can't look up file_id None")
449
raise BzrError("file_id {%s} not in inventory" % file_id)
452
def get_file_kind(self, file_id):
453
return self._byid[file_id].kind
455
def get_child(self, parent_id, filename):
456
return self[parent_id].children.get(filename)
459
def add(self, entry):
460
"""Add entry to inventory.
462
To add a file to a branch ready to be committed, use Branch.add,
465
Returns the new entry object.
467
if entry.file_id in self._byid:
468
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
470
if entry.parent_id == ROOT_ID or entry.parent_id is None:
471
entry.parent_id = self.root.file_id
474
parent = self._byid[entry.parent_id]
476
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
478
if parent.children.has_key(entry.name):
479
raise BzrError("%s is already versioned" %
480
appendpath(self.id2path(parent.file_id), entry.name))
482
self._byid[entry.file_id] = entry
483
parent.children[entry.name] = entry
487
def add_path(self, relpath, kind, file_id=None):
488
"""Add entry from a path.
490
The immediate parent must already be versioned.
492
Returns the new entry object."""
493
from bzrlib.branch import gen_file_id
495
parts = bzrlib.osutils.splitpath(relpath)
497
raise BzrError("cannot re-add root of inventory")
500
file_id = gen_file_id(relpath)
502
parent_path = parts[:-1]
503
parent_id = self.path2id(parent_path)
504
if parent_id == None:
505
raise NotVersionedError(parent_path)
507
ie = InventoryEntry(file_id, parts[-1],
508
kind=kind, parent_id=parent_id)
512
def __delitem__(self, file_id):
513
"""Remove entry by id.
515
>>> inv = Inventory()
516
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
517
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
526
assert self[ie.parent_id].children[ie.name] == ie
528
# TODO: Test deleting all children; maybe hoist to a separate
530
if ie.kind == 'directory':
531
for cie in ie.children.values():
532
del self[cie.file_id]
535
del self._byid[file_id]
536
del self[ie.parent_id].children[ie.name]
539
def __eq__(self, other):
540
"""Compare two sets by comparing their contents.
546
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
547
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
550
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
551
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
555
if not isinstance(other, Inventory):
556
return NotImplemented
558
if len(self._byid) != len(other._byid):
559
# shortcut: obviously not the same
562
return self._byid == other._byid
565
def __ne__(self, other):
566
return not self.__eq__(other)
570
raise ValueError('not hashable')
573
def get_idpath(self, file_id):
574
"""Return a list of file_ids for the path to an entry.
576
The list contains one element for each directory followed by
577
the id of the file itself. So the length of the returned list
578
is equal to the depth of the file in the tree, counting the
579
root directory as depth 1.
582
while file_id != None:
584
ie = self._byid[file_id]
586
raise BzrError("file_id {%s} not found in inventory" % file_id)
587
p.insert(0, ie.file_id)
588
file_id = ie.parent_id
592
def id2path(self, file_id):
593
"""Return as a list the path to file_id."""
595
# get all names, skipping root
596
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
597
return os.sep.join(p)
601
def path2id(self, name):
602
"""Walk down through directories to return entry of last component.
604
names may be either a list of path components, or a single
605
string, in which case it is automatically split.
607
This returns the entry of the last component in the path,
608
which may be either a file or a directory.
610
Returns None iff the path is not found.
612
if isinstance(name, types.StringTypes):
613
name = splitpath(name)
615
mutter("lookup path %r" % name)
620
cie = parent.children[f]
622
assert cie.parent_id == parent.file_id
628
return parent.file_id
631
def has_filename(self, names):
632
return bool(self.path2id(names))
635
def has_id(self, file_id):
636
return self._byid.has_key(file_id)
639
def rename(self, file_id, new_parent_id, new_name):
640
"""Move a file within the inventory.
642
This can change either the name, or the parent, or both.
644
This does not move the working file."""
645
if not is_valid_name(new_name):
646
raise BzrError("not an acceptable filename: %r" % new_name)
648
new_parent = self._byid[new_parent_id]
649
if new_name in new_parent.children:
650
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
652
new_parent_idpath = self.get_idpath(new_parent_id)
653
if file_id in new_parent_idpath:
654
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
655
% (self.id2path(file_id), self.id2path(new_parent_id)))
657
file_ie = self._byid[file_id]
658
old_parent = self._byid[file_ie.parent_id]
660
# TODO: Don't leave things messed up if this fails
662
del old_parent.children[file_ie.name]
663
new_parent.children[new_name] = file_ie
665
file_ie.name = new_name
666
file_ie.parent_id = new_parent_id
673
def is_valid_name(name):
676
_NAME_RE = re.compile(r'^[^/\\]+$')
678
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
