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
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', ]
122
def __init__(self, file_id, name, kind, parent_id, text_id=None):
123
"""Create an InventoryEntry
125
The filename must be a single component, relative to the
126
parent directory; it cannot be a whole path or relative name.
128
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
133
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
134
Traceback (most recent call last):
135
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
137
assert isinstance(name, basestring), name
138
if '/' in name or '\\' in name:
139
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
141
self.text_version = None
142
self.name_version = None
143
self.text_sha1 = None
144
self.text_size = None
145
self.file_id = file_id
148
self.text_id = text_id
149
self.parent_id = parent_id
150
if kind == 'directory':
155
raise BzrError("unhandled entry kind %r" % kind)
159
def sorted_children(self):
160
l = self.children.items()
166
other = InventoryEntry(self.file_id, self.name, self.kind,
168
other.text_id = self.text_id
169
other.text_sha1 = self.text_sha1
170
other.text_size = self.text_size
171
other.text_version = self.text_version
172
other.name_version = self.name_version
173
# note that children are *not* copied; they're pulled across when
179
return ("%s(%r, %r, kind=%r, parent_id=%r)"
180
% (self.__class__.__name__,
187
def __eq__(self, other):
188
if not isinstance(other, InventoryEntry):
189
return NotImplemented
191
return (self.file_id == other.file_id) \
192
and (self.name == other.name) \
193
and (self.text_sha1 == other.text_sha1) \
194
and (self.text_size == other.text_size) \
195
and (self.text_id == other.text_id) \
196
and (self.parent_id == other.parent_id) \
197
and (self.kind == other.kind) \
198
and (self.text_version == other.text_version) \
199
and (self.name_version == other.name_version)
202
def __ne__(self, other):
203
return not (self == other)
206
raise ValueError('not hashable')
210
class RootEntry(InventoryEntry):
211
def __init__(self, file_id):
212
self.file_id = file_id
214
self.kind = 'root_directory'
215
self.parent_id = None
218
def __eq__(self, other):
219
if not isinstance(other, RootEntry):
220
return NotImplemented
222
return (self.file_id == other.file_id) \
223
and (self.children == other.children)
227
class Inventory(object):
228
"""Inventory of versioned files in a tree.
230
This describes which file_id is present at each point in the tree,
231
and possibly the SHA-1 or other information about the file.
232
Entries can be looked up either by path or by file_id.
234
The inventory represents a typical unix file tree, with
235
directories containing files and subdirectories. We never store
236
the full path to a file, because renaming a directory implicitly
237
moves all of its contents. This class internally maintains a
238
lookup tree that allows the children under a directory to be
241
InventoryEntry objects must not be modified after they are
242
inserted, other than through the Inventory API.
244
>>> inv = Inventory()
245
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
246
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT')
247
>>> inv['123-123'].name
250
May be treated as an iterator or set to look up file ids:
252
>>> bool(inv.path2id('hello.c'))
257
May also look up by name:
259
>>> [x[0] for x in inv.iter_entries()]
261
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
262
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
263
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
265
def __init__(self, root_id=ROOT_ID):
266
"""Create or read an inventory.
268
If a working directory is specified, the inventory is read
269
from there. If the file is specified, read from that. If not,
270
the inventory is created empty.
272
The inventory is created with a default root directory, with
275
# We are letting Branch.initialize() create a unique inventory
276
# root id. Rather than generating a random one here.
278
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
279
self.root = RootEntry(root_id)
280
self._byid = {self.root.file_id: self.root}
284
other = Inventory(self.root.file_id)
285
# copy recursively so we know directories will be added before
286
# their children. There are more efficient ways than this...
287
for path, entry in self.iter_entries():
288
if entry == self.root:
290
other.add(entry.copy())
295
return iter(self._byid)
299
"""Returns number of entries."""
300
return len(self._byid)
303
def iter_entries(self, from_dir=None):
304
"""Return (path, entry) pairs, in order by name."""
308
elif isinstance(from_dir, basestring):
309
from_dir = self._byid[from_dir]
311
kids = from_dir.children.items()
313
for name, ie in kids:
315
if ie.kind == 'directory':
316
for cn, cie in self.iter_entries(from_dir=ie.file_id):
317
yield os.path.join(name, cn), cie
321
"""Return list of (path, ie) for all entries except the root.
323
This may be faster than iter_entries.
326
def descend(dir_ie, dir_path):
327
kids = dir_ie.children.items()
329
for name, ie in kids:
330
child_path = os.path.join(dir_path, name)
331
accum.append((child_path, ie))
332
if ie.kind == 'directory':
333
descend(ie, child_path)
335
descend(self.root, '')
339
def directories(self):
340
"""Return (path, entry) pairs for all directories, including the root.
343
def descend(parent_ie, parent_path):
344
accum.append((parent_path, parent_ie))
346
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
349
for name, child_ie in kids:
350
child_path = os.path.join(parent_path, name)
351
descend(child_ie, child_path)
352
descend(self.root, '')
357
def __contains__(self, file_id):
358
"""True if this entry contains a file with given id.
360
>>> inv = Inventory()
361
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
362
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
368
return file_id in self._byid
371
def __getitem__(self, file_id):
372
"""Return the entry for given file_id.
374
>>> inv = Inventory()
375
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
376
InventoryEntry('123123', 'hello.c', kind='file', parent_id='TREE_ROOT')
377
>>> inv['123123'].name
381
return self._byid[file_id]
384
raise BzrError("can't look up file_id None")
386
raise BzrError("file_id {%s} not in inventory" % file_id)
389
def get_file_kind(self, file_id):
390
return self._byid[file_id].kind
392
def get_child(self, parent_id, filename):
393
return self[parent_id].children.get(filename)
396
def add(self, entry):
397
"""Add entry to inventory.
399
To add a file to a branch ready to be committed, use Branch.add,
402
Returns the new entry object.
404
if entry.file_id in self._byid:
405
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
407
if entry.parent_id == ROOT_ID or entry.parent_id is None:
408
entry.parent_id = self.root.file_id
411
parent = self._byid[entry.parent_id]
413
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
415
if parent.children.has_key(entry.name):
416
raise BzrError("%s is already versioned" %
417
appendpath(self.id2path(parent.file_id), entry.name))
419
self._byid[entry.file_id] = entry
420
parent.children[entry.name] = entry
424
def add_path(self, relpath, kind, file_id=None):
425
"""Add entry from a path.
427
The immediate parent must already be versioned.
429
Returns the new entry object."""
430
from bzrlib.branch import gen_file_id
432
parts = bzrlib.osutils.splitpath(relpath)
434
raise BzrError("cannot re-add root of inventory")
437
file_id = gen_file_id(relpath)
439
parent_path = parts[:-1]
440
parent_id = self.path2id(parent_path)
441
if parent_id == None:
442
raise NotVersionedError(parent_path)
444
ie = InventoryEntry(file_id, parts[-1],
445
kind=kind, parent_id=parent_id)
449
def __delitem__(self, file_id):
450
"""Remove entry by id.
452
>>> inv = Inventory()
453
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
454
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
463
assert self[ie.parent_id].children[ie.name] == ie
465
# TODO: Test deleting all children; maybe hoist to a separate
467
if ie.kind == 'directory':
468
for cie in ie.children.values():
469
del self[cie.file_id]
472
del self._byid[file_id]
473
del self[ie.parent_id].children[ie.name]
476
def __eq__(self, other):
477
"""Compare two sets by comparing their contents.
483
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
484
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
487
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
488
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
492
if not isinstance(other, Inventory):
493
return NotImplemented
495
if len(self._byid) != len(other._byid):
496
# shortcut: obviously not the same
499
return self._byid == other._byid
502
def __ne__(self, other):
503
return not self.__eq__(other)
507
raise ValueError('not hashable')
510
def get_idpath(self, file_id):
511
"""Return a list of file_ids for the path to an entry.
513
The list contains one element for each directory followed by
514
the id of the file itself. So the length of the returned list
515
is equal to the depth of the file in the tree, counting the
516
root directory as depth 1.
519
while file_id != None:
521
ie = self._byid[file_id]
523
raise BzrError("file_id {%s} not found in inventory" % file_id)
524
p.insert(0, ie.file_id)
525
file_id = ie.parent_id
529
def id2path(self, file_id):
530
"""Return as a list the path to file_id."""
532
# get all names, skipping root
533
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
534
return os.sep.join(p)
538
def path2id(self, name):
539
"""Walk down through directories to return entry of last component.
541
names may be either a list of path components, or a single
542
string, in which case it is automatically split.
544
This returns the entry of the last component in the path,
545
which may be either a file or a directory.
547
Returns None iff the path is not found.
549
if isinstance(name, types.StringTypes):
550
name = splitpath(name)
552
mutter("lookup path %r" % name)
557
cie = parent.children[f]
559
assert cie.parent_id == parent.file_id
565
return parent.file_id
568
def has_filename(self, names):
569
return bool(self.path2id(names))
572
def has_id(self, file_id):
573
return self._byid.has_key(file_id)
576
def rename(self, file_id, new_parent_id, new_name):
577
"""Move a file within the inventory.
579
This can change either the name, or the parent, or both.
581
This does not move the working file."""
582
if not is_valid_name(new_name):
583
raise BzrError("not an acceptable filename: %r" % new_name)
585
new_parent = self._byid[new_parent_id]
586
if new_name in new_parent.children:
587
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
589
new_parent_idpath = self.get_idpath(new_parent_id)
590
if file_id in new_parent_idpath:
591
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
592
% (self.id2path(file_id), self.id2path(new_parent_id)))
594
file_ie = self._byid[file_id]
595
old_parent = self._byid[file_ie.parent_id]
597
# TODO: Don't leave things messed up if this fails
599
del old_parent.children[file_ie.name]
600
new_parent.children[new_name] = file_ie
602
file_ie.name = new_name
603
file_ie.parent_id = new_parent_id
610
def is_valid_name(name):
613
_NAME_RE = re.compile(r'^[^/\\]+$')
615
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
