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 _add_text_to_weave(self, new_lines, parents, weave_store):
119
weave_store.add_text(self.file_id, self.revision, new_lines, parents)
121
def __init__(self, file_id, name, kind, parent_id, text_id=None):
122
"""Create an InventoryEntry
124
The filename must be a single component, relative to the
125
parent directory; it cannot be a whole path or relative name.
127
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
132
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
133
Traceback (most recent call last):
134
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
136
assert isinstance(name, basestring), name
137
if '/' in name or '\\' in name:
138
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
141
self.text_sha1 = None
142
self.text_size = None
143
self.file_id = file_id
146
self.text_id = text_id
147
self.parent_id = parent_id
148
self.symlink_target = None
149
if kind == 'directory':
153
elif kind == 'symlink':
156
raise BzrError("unhandled entry kind %r" % kind)
158
def read_symlink_target(self, path):
159
if self.kind == 'symlink':
161
self.symlink_target = os.readlink(path)
163
raise BzrError("os.readlink error, %s" % e)
165
def sorted_children(self):
166
l = self.children.items()
170
def check(self, checker, rev_id, inv, tree):
171
if self.parent_id != None:
172
if not inv.has_id(self.parent_id):
173
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
174
% (self.parent_id, rev_id))
175
if self.kind == 'file':
176
revision = self.revision
177
t = (self.file_id, revision)
178
if t in checker.checked_texts:
179
prev_sha = checker.checked_texts[t]
180
if prev_sha != self.text_sha1:
181
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
182
(self.file_id, rev_id))
184
checker.repeated_text_cnt += 1
186
mutter('check version {%s} of {%s}', rev_id, self.file_id)
187
file_lines = tree.get_file_lines(self.file_id)
188
checker.checked_text_cnt += 1
189
if self.text_size != sum(map(len, file_lines)):
190
raise BzrCheckError('text {%s} wrong size' % self.text_id)
191
if self.text_sha1 != sha_strings(file_lines):
192
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
193
checker.checked_texts[t] = self.text_sha1
194
elif self.kind == 'directory':
195
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
196
raise BzrCheckError('directory {%s} has text in revision {%s}'
197
% (self.file_id, rev_id))
198
elif self.kind == 'root_directory':
200
elif self.kind == 'symlink':
201
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
202
raise BzrCheckError('symlink {%s} has text in revision {%s}'
203
% (self.file_id, rev_id))
204
if self.symlink_target == None:
205
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
206
% (self.file_id, rev_id))
208
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
213
other = InventoryEntry(self.file_id, self.name, self.kind,
215
other.text_id = self.text_id
216
other.text_sha1 = self.text_sha1
217
other.text_size = self.text_size
218
other.symlink_target = self.symlink_target
219
other.revision = self.revision
220
# note that children are *not* copied; they're pulled across when
224
def _get_snapshot_change(self, previous_entries):
225
if len(previous_entries) > 1:
227
elif len(previous_entries) == 0:
230
return 'modified/renamed/reparented'
233
return ("%s(%r, %r, kind=%r, parent_id=%r)"
234
% (self.__class__.__name__,
240
def snapshot(self, revision, path, previous_entries, work_tree,
242
"""Make a snapshot of this entry.
244
This means that all its fields are populated, that it has its
245
text stored in the text store or weave.
247
mutter('new parents of %s are %r', path, previous_entries)
248
self._read_tree_state(path, work_tree)
249
if len(previous_entries) == 1:
250
# cannot be unchanged unless there is only one parent file rev.
251
parent_ie = previous_entries.values()[0]
252
if self._unchanged(path, parent_ie, work_tree):
253
mutter("found unchanged entry")
254
self.revision = parent_ie.revision
256
mutter('new revision for {%s}', self.file_id)
257
self.revision = revision
258
change = self._get_snapshot_change(previous_entries)
259
if self.kind != 'file':
261
self._snapshot_text(previous_entries, work_tree, weave_store)
264
def _snapshot_text(self, file_parents, work_tree, weave_store):
265
mutter('storing file {%s} in revision {%s}',
266
self.file_id, self.revision)
267
# special case to avoid diffing on renames or
269
if (len(file_parents) == 1
270
and self.text_sha1 == file_parents.values()[0].text_sha1
271
and self.text_size == file_parents.values()[0].text_size):
272
previous_ie = file_parents.values()[0]
273
weave_store.add_identical_text(
274
self.file_id, previous_ie.revision,
275
self.revision, file_parents)
277
new_lines = work_tree.get_file(self.file_id).readlines()
278
self._add_text_to_weave(new_lines, file_parents, weave_store)
279
self.text_sha1 = sha_strings(new_lines)
280
self.text_size = sum(map(len, new_lines))
282
def __eq__(self, other):
283
if not isinstance(other, InventoryEntry):
284
return NotImplemented
286
return (self.file_id == other.file_id) \
287
and (self.name == other.name) \
288
and (other.symlink_target == self.symlink_target) \
289
and (self.text_sha1 == other.text_sha1) \
290
and (self.text_size == other.text_size) \
291
and (self.text_id == other.text_id) \
292
and (self.parent_id == other.parent_id) \
293
and (self.kind == other.kind) \
294
and (self.revision == other.revision)
296
def __ne__(self, other):
297
return not (self == other)
300
raise ValueError('not hashable')
302
def _unchanged(self, path, previous_ie, work_tree):
304
# different inv parent
305
if previous_ie.parent_id != self.parent_id:
308
elif previous_ie.name != self.name:
310
if self.kind == 'symlink':
311
if self.symlink_target != previous_ie.symlink_target:
313
if self.kind == 'file':
314
if self.text_sha1 != previous_ie.text_sha1:
317
# FIXME: 20050930 probe for the text size when getting sha1
318
# in _read_tree_state
319
self.text_size = previous_ie.text_size
322
def _read_tree_state(self, path, work_tree):
323
if self.kind == 'symlink':
324
self.read_symlink_target(work_tree.abspath(path))
325
if self.kind == 'file':
326
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
329
class RootEntry(InventoryEntry):
330
def __init__(self, file_id):
331
self.file_id = file_id
333
self.kind = 'root_directory'
334
self.parent_id = None
337
def __eq__(self, other):
338
if not isinstance(other, RootEntry):
339
return NotImplemented
341
return (self.file_id == other.file_id) \
342
and (self.children == other.children)
346
class Inventory(object):
347
"""Inventory of versioned files in a tree.
349
This describes which file_id is present at each point in the tree,
350
and possibly the SHA-1 or other information about the file.
351
Entries can be looked up either by path or by file_id.
353
The inventory represents a typical unix file tree, with
354
directories containing files and subdirectories. We never store
355
the full path to a file, because renaming a directory implicitly
356
moves all of its contents. This class internally maintains a
357
lookup tree that allows the children under a directory to be
360
InventoryEntry objects must not be modified after they are
361
inserted, other than through the Inventory API.
363
>>> inv = Inventory()
364
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
365
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT')
366
>>> inv['123-123'].name
369
May be treated as an iterator or set to look up file ids:
371
>>> bool(inv.path2id('hello.c'))
376
May also look up by name:
378
>>> [x[0] for x in inv.iter_entries()]
380
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
381
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
382
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
384
def __init__(self, root_id=ROOT_ID):
385
"""Create or read an inventory.
387
If a working directory is specified, the inventory is read
388
from there. If the file is specified, read from that. If not,
389
the inventory is created empty.
391
The inventory is created with a default root directory, with
394
# We are letting Branch.initialize() create a unique inventory
395
# root id. Rather than generating a random one here.
397
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
398
self.root = RootEntry(root_id)
399
self._byid = {self.root.file_id: self.root}
403
other = Inventory(self.root.file_id)
404
# copy recursively so we know directories will be added before
405
# their children. There are more efficient ways than this...
406
for path, entry in self.iter_entries():
407
if entry == self.root:
409
other.add(entry.copy())
414
return iter(self._byid)
418
"""Returns number of entries."""
419
return len(self._byid)
422
def iter_entries(self, from_dir=None):
423
"""Return (path, entry) pairs, in order by name."""
427
elif isinstance(from_dir, basestring):
428
from_dir = self._byid[from_dir]
430
kids = from_dir.children.items()
432
for name, ie in kids:
434
if ie.kind == 'directory':
435
for cn, cie in self.iter_entries(from_dir=ie.file_id):
436
yield os.path.join(name, cn), cie
440
"""Return list of (path, ie) for all entries except the root.
442
This may be faster than iter_entries.
445
def descend(dir_ie, dir_path):
446
kids = dir_ie.children.items()
448
for name, ie in kids:
449
child_path = os.path.join(dir_path, name)
450
accum.append((child_path, ie))
451
if ie.kind == 'directory':
452
descend(ie, child_path)
454
descend(self.root, '')
458
def directories(self):
459
"""Return (path, entry) pairs for all directories, including the root.
462
def descend(parent_ie, parent_path):
463
accum.append((parent_path, parent_ie))
465
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
468
for name, child_ie in kids:
469
child_path = os.path.join(parent_path, name)
470
descend(child_ie, child_path)
471
descend(self.root, '')
476
def __contains__(self, file_id):
477
"""True if this entry contains a file with given id.
479
>>> inv = Inventory()
480
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
481
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
487
return file_id in self._byid
490
def __getitem__(self, file_id):
491
"""Return the entry for given file_id.
493
>>> inv = Inventory()
494
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
495
InventoryEntry('123123', 'hello.c', kind='file', parent_id='TREE_ROOT')
496
>>> inv['123123'].name
500
return self._byid[file_id]
503
raise BzrError("can't look up file_id None")
505
raise BzrError("file_id {%s} not in inventory" % file_id)
508
def get_file_kind(self, file_id):
509
return self._byid[file_id].kind
511
def get_child(self, parent_id, filename):
512
return self[parent_id].children.get(filename)
515
def add(self, entry):
516
"""Add entry to inventory.
518
To add a file to a branch ready to be committed, use Branch.add,
521
Returns the new entry object.
523
if entry.file_id in self._byid:
524
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
526
if entry.parent_id == ROOT_ID or entry.parent_id is None:
527
entry.parent_id = self.root.file_id
530
parent = self._byid[entry.parent_id]
532
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
534
if parent.children.has_key(entry.name):
535
raise BzrError("%s is already versioned" %
536
appendpath(self.id2path(parent.file_id), entry.name))
538
self._byid[entry.file_id] = entry
539
parent.children[entry.name] = entry
543
def add_path(self, relpath, kind, file_id=None):
544
"""Add entry from a path.
546
The immediate parent must already be versioned.
548
Returns the new entry object."""
549
from bzrlib.branch import gen_file_id
551
parts = bzrlib.osutils.splitpath(relpath)
553
raise BzrError("cannot re-add root of inventory")
556
file_id = gen_file_id(relpath)
558
parent_path = parts[:-1]
559
parent_id = self.path2id(parent_path)
560
if parent_id == None:
561
raise NotVersionedError(parent_path)
563
ie = InventoryEntry(file_id, parts[-1],
564
kind=kind, parent_id=parent_id)
568
def __delitem__(self, file_id):
569
"""Remove entry by id.
571
>>> inv = Inventory()
572
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
573
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
582
assert self[ie.parent_id].children[ie.name] == ie
584
# TODO: Test deleting all children; maybe hoist to a separate
586
if ie.kind == 'directory':
587
for cie in ie.children.values():
588
del self[cie.file_id]
591
del self._byid[file_id]
592
del self[ie.parent_id].children[ie.name]
595
def __eq__(self, other):
596
"""Compare two sets by comparing their contents.
602
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
603
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
606
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
607
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
611
if not isinstance(other, Inventory):
612
return NotImplemented
614
if len(self._byid) != len(other._byid):
615
# shortcut: obviously not the same
618
return self._byid == other._byid
621
def __ne__(self, other):
622
return not self.__eq__(other)
626
raise ValueError('not hashable')
629
def get_idpath(self, file_id):
630
"""Return a list of file_ids for the path to an entry.
632
The list contains one element for each directory followed by
633
the id of the file itself. So the length of the returned list
634
is equal to the depth of the file in the tree, counting the
635
root directory as depth 1.
638
while file_id != None:
640
ie = self._byid[file_id]
642
raise BzrError("file_id {%s} not found in inventory" % file_id)
643
p.insert(0, ie.file_id)
644
file_id = ie.parent_id
648
def id2path(self, file_id):
649
"""Return as a list the path to file_id."""
651
# get all names, skipping root
652
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
653
return os.sep.join(p)
657
def path2id(self, name):
658
"""Walk down through directories to return entry of last component.
660
names may be either a list of path components, or a single
661
string, in which case it is automatically split.
663
This returns the entry of the last component in the path,
664
which may be either a file or a directory.
666
Returns None iff the path is not found.
668
if isinstance(name, types.StringTypes):
669
name = splitpath(name)
671
mutter("lookup path %r" % name)
676
cie = parent.children[f]
678
assert cie.parent_id == parent.file_id
684
return parent.file_id
687
def has_filename(self, names):
688
return bool(self.path2id(names))
691
def has_id(self, file_id):
692
return self._byid.has_key(file_id)
695
def rename(self, file_id, new_parent_id, new_name):
696
"""Move a file within the inventory.
698
This can change either the name, or the parent, or both.
700
This does not move the working file."""
701
if not is_valid_name(new_name):
702
raise BzrError("not an acceptable filename: %r" % new_name)
704
new_parent = self._byid[new_parent_id]
705
if new_name in new_parent.children:
706
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
708
new_parent_idpath = self.get_idpath(new_parent_id)
709
if file_id in new_parent_idpath:
710
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
711
% (self.id2path(file_id), self.id2path(new_parent_id)))
713
file_ie = self._byid[file_id]
714
old_parent = self._byid[file_ie.parent_id]
716
# TODO: Don't leave things messed up if this fails
718
del old_parent.children[file_ie.name]
719
new_parent.children[new_name] = file_ie
721
file_ie.name = new_name
722
file_ie.parent_id = new_parent_id
729
def is_valid_name(name):
732
_NAME_RE = re.compile(r'^[^/\\]+$')
734
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
