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 the name or parent of this file was
66
sha-1 of the text of the file
69
size in bytes of the text of the file
72
the revision_id in which the text of this file was introduced
74
(reading a version 4 tree created a text_id field.)
79
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
80
InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT')
81
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
82
InventoryEntry('2323', 'hello.c', kind='file', parent_id='123')
83
>>> for j in i.iter_entries():
86
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
87
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
88
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
89
Traceback (most recent call last):
91
BzrError: inventory already contains entry with id {2323}
92
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
93
InventoryEntry('2324', 'bye.c', kind='file', parent_id='123')
94
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
95
InventoryEntry('2325', 'wibble', kind='directory', parent_id='123')
96
>>> i.path2id('src/wibble')
100
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
101
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
103
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
104
>>> for path, entry in i.iter_entries():
105
... print path.replace('\\\\', '/') # for win32 os.sep
106
... assert i.path2id(path)
113
>>> i.id2path('2326').replace('\\\\', '/')
114
'src/wibble/wibble.c'
117
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
118
'text_id', 'parent_id', 'children',
119
'text_version', 'name_version', 'symlink_target']
121
def compatible_for_commit(self, previous_ie):
123
# different inv parent
124
if previous_ie.parent_id != self.parent_id:
127
elif previous_ie.name != self.name:
129
# changed link target
130
elif (hasattr(self,'symlink_target')
131
and self.symlink_target != previous_ie.symlink_target):
135
def __init__(self, file_id, name, kind, parent_id, text_id=None):
136
"""Create an InventoryEntry
138
The filename must be a single component, relative to the
139
parent directory; it cannot be a whole path or relative name.
141
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
146
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
147
Traceback (most recent call last):
148
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
150
assert isinstance(name, basestring), name
151
if '/' in name or '\\' in name:
152
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
154
self.text_version = None
155
self.name_version = None
156
self.text_sha1 = None
157
self.text_size = None
158
self.file_id = file_id
161
self.text_id = text_id
162
self.parent_id = parent_id
163
self.symlink_target = None
164
if kind == 'directory':
168
elif kind == 'symlink':
171
raise BzrError("unhandled entry kind %r" % kind)
173
def read_symlink_target(self, path):
174
if self.kind == 'symlink':
176
self.symlink_target = os.readlink(path)
178
raise BzrError("os.readlink error, %s" % e)
180
def sorted_children(self):
181
l = self.children.items()
185
def check(self, checker, rev_id, inv, tree):
186
if self.parent_id != None:
187
if not inv.has_id(self.parent_id):
188
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
189
% (self.parent_id, rev_id))
190
if self.kind == 'file':
191
text_version = self.text_version
192
t = (self.file_id, text_version)
193
if t in checker.checked_texts:
194
prev_sha = checker.checked_texts[t]
195
if prev_sha != self.text_sha1:
196
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
197
(self.file_id, rev_id))
199
checker.repeated_text_cnt += 1
201
mutter('check version {%s} of {%s}', rev_id, self.file_id)
202
file_lines = tree.get_file_lines(self.file_id)
203
checker.checked_text_cnt += 1
204
if self.text_size != sum(map(len, file_lines)):
205
raise BzrCheckError('text {%s} wrong size' % self.text_id)
206
if self.text_sha1 != sha_strings(file_lines):
207
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
208
checker.checked_texts[t] = self.text_sha1
209
elif self.kind == 'directory':
210
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
211
raise BzrCheckError('directory {%s} has text in revision {%s}'
212
% (self.file_id, rev_id))
213
elif self.kind == 'root_directory':
215
elif self.kind == 'symlink':
216
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
217
raise BzrCheckError('symlink {%s} has text in revision {%s}'
218
% (self.file_id, rev_id))
219
if self.symlink_target == None:
220
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
221
% (self.file_id, rev_id))
223
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
228
other = InventoryEntry(self.file_id, self.name, self.kind,
230
other.text_id = self.text_id
231
other.text_sha1 = self.text_sha1
232
other.text_size = self.text_size
233
other.symlink_target = self.symlink_target
234
other.text_version = self.text_version
235
other.name_version = self.name_version
236
# note that children are *not* copied; they're pulled across when
242
return ("%s(%r, %r, kind=%r, parent_id=%r)"
243
% (self.__class__.__name__,
249
def __eq__(self, other):
250
if not isinstance(other, InventoryEntry):
251
return NotImplemented
253
return (self.file_id == other.file_id) \
254
and (self.name == other.name) \
255
and (other.symlink_target == self.symlink_target) \
256
and (self.text_sha1 == other.text_sha1) \
257
and (self.text_size == other.text_size) \
258
and (self.text_id == other.text_id) \
259
and (self.parent_id == other.parent_id) \
260
and (self.kind == other.kind) \
261
and (self.text_version == other.text_version) \
262
and (self.name_version == other.name_version)
264
def __ne__(self, other):
265
return not (self == other)
268
raise ValueError('not hashable')
271
class RootEntry(InventoryEntry):
272
def __init__(self, file_id):
273
self.file_id = file_id
275
self.kind = 'root_directory'
276
self.parent_id = None
279
def __eq__(self, other):
280
if not isinstance(other, RootEntry):
281
return NotImplemented
283
return (self.file_id == other.file_id) \
284
and (self.children == other.children)
288
class Inventory(object):
289
"""Inventory of versioned files in a tree.
291
This describes which file_id is present at each point in the tree,
292
and possibly the SHA-1 or other information about the file.
293
Entries can be looked up either by path or by file_id.
295
The inventory represents a typical unix file tree, with
296
directories containing files and subdirectories. We never store
297
the full path to a file, because renaming a directory implicitly
298
moves all of its contents. This class internally maintains a
299
lookup tree that allows the children under a directory to be
302
InventoryEntry objects must not be modified after they are
303
inserted, other than through the Inventory API.
305
>>> inv = Inventory()
306
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
307
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT')
308
>>> inv['123-123'].name
311
May be treated as an iterator or set to look up file ids:
313
>>> bool(inv.path2id('hello.c'))
318
May also look up by name:
320
>>> [x[0] for x in inv.iter_entries()]
322
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
323
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
324
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
326
def __init__(self, root_id=ROOT_ID):
327
"""Create or read an inventory.
329
If a working directory is specified, the inventory is read
330
from there. If the file is specified, read from that. If not,
331
the inventory is created empty.
333
The inventory is created with a default root directory, with
336
# We are letting Branch.initialize() create a unique inventory
337
# root id. Rather than generating a random one here.
339
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
340
self.root = RootEntry(root_id)
341
self._byid = {self.root.file_id: self.root}
345
other = Inventory(self.root.file_id)
346
# copy recursively so we know directories will be added before
347
# their children. There are more efficient ways than this...
348
for path, entry in self.iter_entries():
349
if entry == self.root:
351
other.add(entry.copy())
356
return iter(self._byid)
360
"""Returns number of entries."""
361
return len(self._byid)
364
def iter_entries(self, from_dir=None):
365
"""Return (path, entry) pairs, in order by name."""
369
elif isinstance(from_dir, basestring):
370
from_dir = self._byid[from_dir]
372
kids = from_dir.children.items()
374
for name, ie in kids:
376
if ie.kind == 'directory':
377
for cn, cie in self.iter_entries(from_dir=ie.file_id):
378
yield os.path.join(name, cn), cie
382
"""Return list of (path, ie) for all entries except the root.
384
This may be faster than iter_entries.
387
def descend(dir_ie, dir_path):
388
kids = dir_ie.children.items()
390
for name, ie in kids:
391
child_path = os.path.join(dir_path, name)
392
accum.append((child_path, ie))
393
if ie.kind == 'directory':
394
descend(ie, child_path)
396
descend(self.root, '')
400
def directories(self):
401
"""Return (path, entry) pairs for all directories, including the root.
404
def descend(parent_ie, parent_path):
405
accum.append((parent_path, parent_ie))
407
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
410
for name, child_ie in kids:
411
child_path = os.path.join(parent_path, name)
412
descend(child_ie, child_path)
413
descend(self.root, '')
418
def __contains__(self, file_id):
419
"""True if this entry contains a file with given id.
421
>>> inv = Inventory()
422
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
423
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
429
return file_id in self._byid
432
def __getitem__(self, file_id):
433
"""Return the entry for given file_id.
435
>>> inv = Inventory()
436
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
437
InventoryEntry('123123', 'hello.c', kind='file', parent_id='TREE_ROOT')
438
>>> inv['123123'].name
442
return self._byid[file_id]
445
raise BzrError("can't look up file_id None")
447
raise BzrError("file_id {%s} not in inventory" % file_id)
450
def get_file_kind(self, file_id):
451
return self._byid[file_id].kind
453
def get_child(self, parent_id, filename):
454
return self[parent_id].children.get(filename)
457
def add(self, entry):
458
"""Add entry to inventory.
460
To add a file to a branch ready to be committed, use Branch.add,
463
Returns the new entry object.
465
if entry.file_id in self._byid:
466
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
468
if entry.parent_id == ROOT_ID or entry.parent_id is None:
469
entry.parent_id = self.root.file_id
472
parent = self._byid[entry.parent_id]
474
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
476
if parent.children.has_key(entry.name):
477
raise BzrError("%s is already versioned" %
478
appendpath(self.id2path(parent.file_id), entry.name))
480
self._byid[entry.file_id] = entry
481
parent.children[entry.name] = entry
485
def add_path(self, relpath, kind, file_id=None):
486
"""Add entry from a path.
488
The immediate parent must already be versioned.
490
Returns the new entry object."""
491
from bzrlib.branch import gen_file_id
493
parts = bzrlib.osutils.splitpath(relpath)
495
raise BzrError("cannot re-add root of inventory")
498
file_id = gen_file_id(relpath)
500
parent_path = parts[:-1]
501
parent_id = self.path2id(parent_path)
502
if parent_id == None:
503
raise NotVersionedError(parent_path)
505
ie = InventoryEntry(file_id, parts[-1],
506
kind=kind, parent_id=parent_id)
510
def __delitem__(self, file_id):
511
"""Remove entry by id.
513
>>> inv = Inventory()
514
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
515
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
524
assert self[ie.parent_id].children[ie.name] == ie
526
# TODO: Test deleting all children; maybe hoist to a separate
528
if ie.kind == 'directory':
529
for cie in ie.children.values():
530
del self[cie.file_id]
533
del self._byid[file_id]
534
del self[ie.parent_id].children[ie.name]
537
def __eq__(self, other):
538
"""Compare two sets by comparing their contents.
544
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
545
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
548
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
549
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
553
if not isinstance(other, Inventory):
554
return NotImplemented
556
if len(self._byid) != len(other._byid):
557
# shortcut: obviously not the same
560
return self._byid == other._byid
563
def __ne__(self, other):
564
return not self.__eq__(other)
568
raise ValueError('not hashable')
571
def get_idpath(self, file_id):
572
"""Return a list of file_ids for the path to an entry.
574
The list contains one element for each directory followed by
575
the id of the file itself. So the length of the returned list
576
is equal to the depth of the file in the tree, counting the
577
root directory as depth 1.
580
while file_id != None:
582
ie = self._byid[file_id]
584
raise BzrError("file_id {%s} not found in inventory" % file_id)
585
p.insert(0, ie.file_id)
586
file_id = ie.parent_id
590
def id2path(self, file_id):
591
"""Return as a list the path to file_id."""
593
# get all names, skipping root
594
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
595
return os.sep.join(p)
599
def path2id(self, name):
600
"""Walk down through directories to return entry of last component.
602
names may be either a list of path components, or a single
603
string, in which case it is automatically split.
605
This returns the entry of the last component in the path,
606
which may be either a file or a directory.
608
Returns None iff the path is not found.
610
if isinstance(name, types.StringTypes):
611
name = splitpath(name)
613
mutter("lookup path %r" % name)
618
cie = parent.children[f]
620
assert cie.parent_id == parent.file_id
626
return parent.file_id
629
def has_filename(self, names):
630
return bool(self.path2id(names))
633
def has_id(self, file_id):
634
return self._byid.has_key(file_id)
637
def rename(self, file_id, new_parent_id, new_name):
638
"""Move a file within the inventory.
640
This can change either the name, or the parent, or both.
642
This does not move the working file."""
643
if not is_valid_name(new_name):
644
raise BzrError("not an acceptable filename: %r" % new_name)
646
new_parent = self._byid[new_parent_id]
647
if new_name in new_parent.children:
648
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
650
new_parent_idpath = self.get_idpath(new_parent_id)
651
if file_id in new_parent_idpath:
652
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
653
% (self.id2path(file_id), self.id2path(new_parent_id)))
655
file_ie = self._byid[file_id]
656
old_parent = self._byid[file_ie.parent_id]
658
# TODO: Don't leave things messed up if this fails
660
del old_parent.children[file_ie.name]
661
new_parent.children[new_name] = file_ie
663
file_ie.name = new_name
664
file_ie.parent_id = new_parent_id
671
def is_valid_name(name):
674
_NAME_RE = re.compile(r'^[^/\\]+$')
676
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
