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()
176
def versionable_kind(kind):
177
return kind in ('file', 'directory', 'symlink')
179
def check(self, checker, rev_id, inv, tree):
180
if self.parent_id != None:
181
if not inv.has_id(self.parent_id):
182
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
183
% (self.parent_id, rev_id))
184
if self.kind == 'file':
185
revision = self.revision
186
t = (self.file_id, revision)
187
if t in checker.checked_texts:
188
prev_sha = checker.checked_texts[t]
189
if prev_sha != self.text_sha1:
190
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
191
(self.file_id, rev_id))
193
checker.repeated_text_cnt += 1
195
mutter('check version {%s} of {%s}', rev_id, self.file_id)
196
file_lines = tree.get_file_lines(self.file_id)
197
checker.checked_text_cnt += 1
198
if self.text_size != sum(map(len, file_lines)):
199
raise BzrCheckError('text {%s} wrong size' % self.text_id)
200
if self.text_sha1 != sha_strings(file_lines):
201
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
202
checker.checked_texts[t] = self.text_sha1
203
elif self.kind == 'directory':
204
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
205
raise BzrCheckError('directory {%s} has text in revision {%s}'
206
% (self.file_id, rev_id))
207
elif self.kind == 'root_directory':
209
elif self.kind == 'symlink':
210
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
211
raise BzrCheckError('symlink {%s} has text in revision {%s}'
212
% (self.file_id, rev_id))
213
if self.symlink_target == None:
214
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
215
% (self.file_id, rev_id))
217
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
222
other = InventoryEntry(self.file_id, self.name, self.kind,
224
other.executable = self.executable
225
other.text_id = self.text_id
226
other.text_sha1 = self.text_sha1
227
other.text_size = self.text_size
228
other.symlink_target = self.symlink_target
229
other.revision = self.revision
230
# note that children are *not* copied; they're pulled across when
234
def _get_snapshot_change(self, previous_entries):
235
if len(previous_entries) > 1:
237
elif len(previous_entries) == 0:
240
return 'modified/renamed/reparented'
243
return ("%s(%r, %r, kind=%r, parent_id=%r)"
244
% (self.__class__.__name__,
250
def snapshot(self, revision, path, previous_entries, work_tree,
252
"""Make a snapshot of this entry.
254
This means that all its fields are populated, that it has its
255
text stored in the text store or weave.
257
mutter('new parents of %s are %r', path, previous_entries)
258
self._read_tree_state(path, work_tree)
259
if len(previous_entries) == 1:
260
# cannot be unchanged unless there is only one parent file rev.
261
parent_ie = previous_entries.values()[0]
262
if self._unchanged(path, parent_ie, work_tree):
263
mutter("found unchanged entry")
264
self.revision = parent_ie.revision
266
mutter('new revision for {%s}', self.file_id)
267
self.revision = revision
268
change = self._get_snapshot_change(previous_entries)
269
if self.kind != 'file':
271
self._snapshot_text(previous_entries, work_tree, weave_store)
274
def _snapshot_text(self, file_parents, work_tree, weave_store):
275
mutter('storing file {%s} in revision {%s}',
276
self.file_id, self.revision)
277
# special case to avoid diffing on renames or
279
if (len(file_parents) == 1
280
and self.text_sha1 == file_parents.values()[0].text_sha1
281
and self.text_size == file_parents.values()[0].text_size):
282
previous_ie = file_parents.values()[0]
283
weave_store.add_identical_text(
284
self.file_id, previous_ie.revision,
285
self.revision, file_parents)
287
new_lines = work_tree.get_file(self.file_id).readlines()
288
self._add_text_to_weave(new_lines, file_parents, weave_store)
289
self.text_sha1 = sha_strings(new_lines)
290
self.text_size = sum(map(len, new_lines))
292
def __eq__(self, other):
293
if not isinstance(other, InventoryEntry):
294
return NotImplemented
296
return ((self.file_id == other.file_id)
297
and (self.name == other.name)
298
and (other.symlink_target == self.symlink_target)
299
and (self.text_sha1 == other.text_sha1)
300
and (self.text_size == other.text_size)
301
and (self.text_id == other.text_id)
302
and (self.parent_id == other.parent_id)
303
and (self.kind == other.kind)
304
and (self.revision == other.revision)
305
and (self.executable == other.executable)
308
def __ne__(self, other):
309
return not (self == other)
312
raise ValueError('not hashable')
314
def _unchanged(self, path, previous_ie, work_tree):
316
# different inv parent
317
if previous_ie.parent_id != self.parent_id:
320
elif previous_ie.name != self.name:
322
if self.kind == 'symlink':
323
if self.symlink_target != previous_ie.symlink_target:
325
if self.kind == 'file':
326
if self.text_sha1 != previous_ie.text_sha1:
329
# FIXME: 20050930 probe for the text size when getting sha1
330
# in _read_tree_state
331
self.text_size = previous_ie.text_size
334
def _read_tree_state(self, path, work_tree):
335
if self.kind == 'symlink':
336
self.read_symlink_target(work_tree.abspath(path))
337
if self.kind == 'file':
338
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
339
self.executable = work_tree.is_executable(self.file_id)
342
class RootEntry(InventoryEntry):
343
def __init__(self, file_id):
344
self.file_id = file_id
346
self.kind = 'root_directory'
347
self.parent_id = None
350
def __eq__(self, other):
351
if not isinstance(other, RootEntry):
352
return NotImplemented
354
return (self.file_id == other.file_id) \
355
and (self.children == other.children)
359
class Inventory(object):
360
"""Inventory of versioned files in a tree.
362
This describes which file_id is present at each point in the tree,
363
and possibly the SHA-1 or other information about the file.
364
Entries can be looked up either by path or by file_id.
366
The inventory represents a typical unix file tree, with
367
directories containing files and subdirectories. We never store
368
the full path to a file, because renaming a directory implicitly
369
moves all of its contents. This class internally maintains a
370
lookup tree that allows the children under a directory to be
373
InventoryEntry objects must not be modified after they are
374
inserted, other than through the Inventory API.
376
>>> inv = Inventory()
377
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
378
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT')
379
>>> inv['123-123'].name
382
May be treated as an iterator or set to look up file ids:
384
>>> bool(inv.path2id('hello.c'))
389
May also look up by name:
391
>>> [x[0] for x in inv.iter_entries()]
393
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
394
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
395
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
397
def __init__(self, root_id=ROOT_ID):
398
"""Create or read an inventory.
400
If a working directory is specified, the inventory is read
401
from there. If the file is specified, read from that. If not,
402
the inventory is created empty.
404
The inventory is created with a default root directory, with
407
# We are letting Branch.initialize() create a unique inventory
408
# root id. Rather than generating a random one here.
410
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
411
self.root = RootEntry(root_id)
412
self._byid = {self.root.file_id: self.root}
416
other = Inventory(self.root.file_id)
417
# copy recursively so we know directories will be added before
418
# their children. There are more efficient ways than this...
419
for path, entry in self.iter_entries():
420
if entry == self.root:
422
other.add(entry.copy())
427
return iter(self._byid)
431
"""Returns number of entries."""
432
return len(self._byid)
435
def iter_entries(self, from_dir=None):
436
"""Return (path, entry) pairs, in order by name."""
440
elif isinstance(from_dir, basestring):
441
from_dir = self._byid[from_dir]
443
kids = from_dir.children.items()
445
for name, ie in kids:
447
if ie.kind == 'directory':
448
for cn, cie in self.iter_entries(from_dir=ie.file_id):
449
yield os.path.join(name, cn), cie
453
"""Return list of (path, ie) for all entries except the root.
455
This may be faster than iter_entries.
458
def descend(dir_ie, dir_path):
459
kids = dir_ie.children.items()
461
for name, ie in kids:
462
child_path = os.path.join(dir_path, name)
463
accum.append((child_path, ie))
464
if ie.kind == 'directory':
465
descend(ie, child_path)
467
descend(self.root, '')
471
def directories(self):
472
"""Return (path, entry) pairs for all directories, including the root.
475
def descend(parent_ie, parent_path):
476
accum.append((parent_path, parent_ie))
478
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
481
for name, child_ie in kids:
482
child_path = os.path.join(parent_path, name)
483
descend(child_ie, child_path)
484
descend(self.root, '')
489
def __contains__(self, file_id):
490
"""True if this entry contains a file with given id.
492
>>> inv = Inventory()
493
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
494
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
500
return file_id in self._byid
503
def __getitem__(self, file_id):
504
"""Return the entry for given file_id.
506
>>> inv = Inventory()
507
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
508
InventoryEntry('123123', 'hello.c', kind='file', parent_id='TREE_ROOT')
509
>>> inv['123123'].name
513
return self._byid[file_id]
516
raise BzrError("can't look up file_id None")
518
raise BzrError("file_id {%s} not in inventory" % file_id)
521
def get_file_kind(self, file_id):
522
return self._byid[file_id].kind
524
def get_child(self, parent_id, filename):
525
return self[parent_id].children.get(filename)
528
def add(self, entry):
529
"""Add entry to inventory.
531
To add a file to a branch ready to be committed, use Branch.add,
534
Returns the new entry object.
536
if entry.file_id in self._byid:
537
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
539
if entry.parent_id == ROOT_ID or entry.parent_id is None:
540
entry.parent_id = self.root.file_id
543
parent = self._byid[entry.parent_id]
545
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
547
if parent.children.has_key(entry.name):
548
raise BzrError("%s is already versioned" %
549
appendpath(self.id2path(parent.file_id), entry.name))
551
self._byid[entry.file_id] = entry
552
parent.children[entry.name] = entry
556
def add_path(self, relpath, kind, file_id=None):
557
"""Add entry from a path.
559
The immediate parent must already be versioned.
561
Returns the new entry object."""
562
from bzrlib.branch import gen_file_id
564
parts = bzrlib.osutils.splitpath(relpath)
566
raise BzrError("cannot re-add root of inventory")
569
file_id = gen_file_id(relpath)
571
parent_path = parts[:-1]
572
parent_id = self.path2id(parent_path)
573
if parent_id == None:
574
raise NotVersionedError(parent_path)
576
ie = InventoryEntry(file_id, parts[-1],
577
kind=kind, parent_id=parent_id)
581
def __delitem__(self, file_id):
582
"""Remove entry by id.
584
>>> inv = Inventory()
585
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
586
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
595
assert self[ie.parent_id].children[ie.name] == ie
597
# TODO: Test deleting all children; maybe hoist to a separate
599
if ie.kind == 'directory':
600
for cie in ie.children.values():
601
del self[cie.file_id]
604
del self._byid[file_id]
605
del self[ie.parent_id].children[ie.name]
608
def __eq__(self, other):
609
"""Compare two sets by comparing their contents.
615
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
616
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
619
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
620
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
624
if not isinstance(other, Inventory):
625
return NotImplemented
627
if len(self._byid) != len(other._byid):
628
# shortcut: obviously not the same
631
return self._byid == other._byid
634
def __ne__(self, other):
635
return not self.__eq__(other)
639
raise ValueError('not hashable')
642
def get_idpath(self, file_id):
643
"""Return a list of file_ids for the path to an entry.
645
The list contains one element for each directory followed by
646
the id of the file itself. So the length of the returned list
647
is equal to the depth of the file in the tree, counting the
648
root directory as depth 1.
651
while file_id != None:
653
ie = self._byid[file_id]
655
raise BzrError("file_id {%s} not found in inventory" % file_id)
656
p.insert(0, ie.file_id)
657
file_id = ie.parent_id
661
def id2path(self, file_id):
662
"""Return as a list the path to file_id."""
664
# get all names, skipping root
665
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
666
return os.sep.join(p)
670
def path2id(self, name):
671
"""Walk down through directories to return entry of last component.
673
names may be either a list of path components, or a single
674
string, in which case it is automatically split.
676
This returns the entry of the last component in the path,
677
which may be either a file or a directory.
679
Returns None iff the path is not found.
681
if isinstance(name, types.StringTypes):
682
name = splitpath(name)
684
mutter("lookup path %r" % name)
689
cie = parent.children[f]
691
assert cie.parent_id == parent.file_id
697
return parent.file_id
700
def has_filename(self, names):
701
return bool(self.path2id(names))
704
def has_id(self, file_id):
705
return self._byid.has_key(file_id)
708
def rename(self, file_id, new_parent_id, new_name):
709
"""Move a file within the inventory.
711
This can change either the name, or the parent, or both.
713
This does not move the working file."""
714
if not is_valid_name(new_name):
715
raise BzrError("not an acceptable filename: %r" % new_name)
717
new_parent = self._byid[new_parent_id]
718
if new_name in new_parent.children:
719
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
721
new_parent_idpath = self.get_idpath(new_parent_id)
722
if file_id in new_parent_idpath:
723
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
724
% (self.id2path(file_id), self.id2path(new_parent_id)))
726
file_ie = self._byid[file_id]
727
old_parent = self._byid[file_ie.parent_id]
729
# TODO: Don't leave things messed up if this fails
731
del old_parent.children[file_ie.name]
732
new_parent.children[new_name] = file_ie
734
file_ie.name = new_name
735
file_ie.parent_id = new_parent_id
742
def is_valid_name(name):
745
_NAME_RE = re.compile(r'^[^/\\]+$')
747
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
