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
# This should really be an id randomly assigned when the tree is
19
# created, but it's not for now.
23
import sys, os.path, types, re
26
from bzrlib.errors import BzrError, BzrCheckError
28
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
29
from bzrlib.trace import mutter
31
class InventoryEntry(object):
32
"""Description of a versioned file.
34
An InventoryEntry has the following fields, which are also
35
present in the XML inventory-entry element:
38
* *name*: (only the basename within the directory, must not
40
* *kind*: "directory" or "file"
41
* *directory_id*: (if absent/null means the branch root directory)
42
* *text_sha1*: only for files
43
* *text_size*: in bytes, only for files
44
* *text_id*: identifier for the text version, only for files
46
InventoryEntries can also exist inside a WorkingTree
47
inventory, in which case they are not yet bound to a
48
particular revision of the file. In that case the text_sha1,
49
text_size and text_id are absent.
55
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
56
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
57
>>> for j in i.iter_entries():
60
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
61
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
62
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
63
Traceback (most recent call last):
65
BzrError: inventory already contains entry with id {2323}
66
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
67
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
68
>>> i.path2id('src/wibble')
72
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
74
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
75
>>> for j in i.iter_entries():
77
... assert i.path2id(j[0])
87
TODO: Maybe also keep the full path of the entry, and the children?
88
But those depend on its position within a particular inventory, and
89
it would be nice not to need to hold the backpointer here.
92
# TODO: split InventoryEntry into subclasses for files,
93
# directories, etc etc.
95
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
96
'text_id', 'parent_id', 'children', ]
98
def __init__(self, file_id, name, kind, parent_id, text_id=None):
99
"""Create an InventoryEntry
101
The filename must be a single component, relative to the
102
parent directory; it cannot be a whole path or relative name.
104
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
109
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
110
Traceback (most recent call last):
111
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
113
if '/' in name or '\\' in name:
114
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
116
self.text_sha1 = None
117
self.text_size = None
119
self.file_id = file_id
122
self.text_id = text_id
123
self.parent_id = parent_id
124
if kind == 'directory':
129
raise BzrError("unhandled entry kind %r" % kind)
133
def sorted_children(self):
134
l = self.children.items()
140
other = InventoryEntry(self.file_id, self.name, self.kind,
141
self.parent_id, text_id=self.text_id)
142
other.text_sha1 = self.text_sha1
143
other.text_size = self.text_size
144
# note that children are *not* copied; they're pulled across when
150
return ("%s(%r, %r, kind=%r, parent_id=%r)"
151
% (self.__class__.__name__,
158
def to_element(self):
159
"""Convert to XML element"""
160
from bzrlib.xml import Element
164
e.set('name', self.name)
165
e.set('file_id', self.file_id)
166
e.set('kind', self.kind)
168
if self.text_size != None:
169
e.set('text_size', '%d' % self.text_size)
171
for f in ['text_id', 'text_sha1']:
176
# to be conservative, we don't externalize the root pointers
177
# for now, leaving them as null in the xml form. in a future
178
# version it will be implied by nested elements.
179
if self.parent_id != ROOT_ID:
180
assert isinstance(self.parent_id, basestring)
181
e.set('parent_id', self.parent_id)
188
def from_element(cls, elt):
189
assert elt.tag == 'entry'
191
## original format inventories don't have a parent_id for
192
## nodes in the root directory, but it's cleaner to use one
194
parent_id = elt.get('parent_id')
195
if parent_id == None:
198
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
199
self.text_id = elt.get('text_id')
200
self.text_sha1 = elt.get('text_sha1')
202
## mutter("read inventoryentry: %r" % (elt.attrib))
204
v = elt.get('text_size')
205
self.text_size = v and int(v)
210
from_element = classmethod(from_element)
212
def __eq__(self, other):
213
if not isinstance(other, InventoryEntry):
214
return NotImplemented
216
return (self.file_id == other.file_id) \
217
and (self.name == other.name) \
218
and (self.text_sha1 == other.text_sha1) \
219
and (self.text_size == other.text_size) \
220
and (self.text_id == other.text_id) \
221
and (self.parent_id == other.parent_id) \
222
and (self.kind == other.kind)
225
def __ne__(self, other):
226
return not (self == other)
229
raise ValueError('not hashable')
233
class RootEntry(InventoryEntry):
234
def __init__(self, file_id):
235
self.file_id = file_id
237
self.kind = 'root_directory'
238
self.parent_id = None
241
def __eq__(self, other):
242
if not isinstance(other, RootEntry):
243
return NotImplemented
245
return (self.file_id == other.file_id) \
246
and (self.children == other.children)
250
class Inventory(object):
251
"""Inventory of versioned files in a tree.
253
This describes which file_id is present at each point in the tree,
254
and possibly the SHA-1 or other information about the file.
255
Entries can be looked up either by path or by file_id.
257
The inventory represents a typical unix file tree, with
258
directories containing files and subdirectories. We never store
259
the full path to a file, because renaming a directory implicitly
260
moves all of its contents. This class internally maintains a
261
lookup tree that allows the children under a directory to be
264
InventoryEntry objects must not be modified after they are
265
inserted, other than through the Inventory API.
267
>>> inv = Inventory()
268
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
269
>>> inv['123-123'].name
272
May be treated as an iterator or set to look up file ids:
274
>>> bool(inv.path2id('hello.c'))
279
May also look up by name:
281
>>> [x[0] for x in inv.iter_entries()]
283
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
284
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
286
def __init__(self, root_id=ROOT_ID):
287
"""Create or read an inventory.
289
If a working directory is specified, the inventory is read
290
from there. If the file is specified, read from that. If not,
291
the inventory is created empty.
293
The inventory is created with a default root directory, with
296
# We are letting Branch(init=True) create a unique inventory
297
# root id. Rather than generating a random one here.
299
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
300
self.root = RootEntry(root_id)
301
self._byid = {self.root.file_id: self.root}
305
return iter(self._byid)
309
"""Returns number of entries."""
310
return len(self._byid)
313
def iter_entries(self, from_dir=None):
314
"""Return (path, entry) pairs, in order by name."""
318
elif isinstance(from_dir, basestring):
319
from_dir = self._byid[from_dir]
321
kids = from_dir.children.items()
323
for name, ie in kids:
325
if ie.kind == 'directory':
326
for cn, cie in self.iter_entries(from_dir=ie.file_id):
327
yield os.path.join(name, cn), cie
331
"""Return list of (path, ie) for all entries except the root.
333
This may be faster than iter_entries.
336
def descend(dir_ie, dir_path):
337
kids = dir_ie.children.items()
339
for name, ie in kids:
340
child_path = os.path.join(dir_path, name)
341
accum.append((child_path, ie))
342
if ie.kind == 'directory':
343
descend(ie, child_path)
345
descend(self.root, '')
349
def directories(self):
350
"""Return (path, entry) pairs for all directories, including the root.
353
def descend(parent_ie, parent_path):
354
accum.append((parent_path, parent_ie))
356
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
359
for name, child_ie in kids:
360
child_path = os.path.join(parent_path, name)
361
descend(child_ie, child_path)
362
descend(self.root, '')
367
def __contains__(self, file_id):
368
"""True if this entry contains a file with given id.
370
>>> inv = Inventory()
371
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
377
return file_id in self._byid
380
def __getitem__(self, file_id):
381
"""Return the entry for given file_id.
383
>>> inv = Inventory()
384
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
385
>>> inv['123123'].name
389
return self._byid[file_id]
392
raise BzrError("can't look up file_id None")
394
raise BzrError("file_id {%s} not in inventory" % file_id)
397
def get_file_kind(self, file_id):
398
return self._byid[file_id].kind
400
def get_child(self, parent_id, filename):
401
return self[parent_id].children.get(filename)
404
def add(self, entry):
405
"""Add entry to inventory.
407
To add a file to a branch ready to be committed, use Branch.add,
409
if entry.file_id in self._byid:
410
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
412
if entry.parent_id == ROOT_ID or entry.parent_id is None:
413
entry.parent_id = self.root.file_id
416
parent = self._byid[entry.parent_id]
418
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
420
if parent.children.has_key(entry.name):
421
raise BzrError("%s is already versioned" %
422
appendpath(self.id2path(parent.file_id), entry.name))
424
self._byid[entry.file_id] = entry
425
parent.children[entry.name] = entry
428
def add_path(self, relpath, kind, file_id=None):
429
"""Add entry from a path.
431
The immediate parent must already be versioned"""
432
from bzrlib.errors import NotVersionedError
434
parts = bzrlib.osutils.splitpath(relpath)
436
raise BzrError("cannot re-add root of inventory")
439
from bzrlib.branch import gen_file_id
440
file_id = gen_file_id(relpath)
442
parent_path = parts[:-1]
443
parent_id = self.path2id(parent_path)
444
if parent_id == None:
445
raise NotVersionedError(parent_path)
447
ie = InventoryEntry(file_id, parts[-1],
448
kind=kind, parent_id=parent_id)
452
def __delitem__(self, file_id):
453
"""Remove entry by id.
455
>>> inv = Inventory()
456
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
465
assert self[ie.parent_id].children[ie.name] == ie
467
# TODO: Test deleting all children; maybe hoist to a separate
469
if ie.kind == 'directory':
470
for cie in ie.children.values():
471
del self[cie.file_id]
474
del self._byid[file_id]
475
del self[ie.parent_id].children[ie.name]
478
def to_element(self):
479
"""Convert to XML Element"""
480
from bzrlib.xml import Element
482
e = Element('inventory')
484
if self.root.file_id not in (None, ROOT_ID):
485
e.set('file_id', self.root.file_id)
486
for path, ie in self.iter_entries():
487
e.append(ie.to_element())
491
def from_element(cls, elt):
492
"""Construct from XML Element
494
>>> inv = Inventory()
495
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
496
>>> elt = inv.to_element()
497
>>> inv2 = Inventory.from_element(elt)
501
# XXXX: doctest doesn't run this properly under python2.3
502
assert elt.tag == 'inventory'
503
root_id = elt.get('file_id') or ROOT_ID
506
ie = InventoryEntry.from_element(e)
507
if ie.parent_id == ROOT_ID:
508
ie.parent_id = root_id
512
from_element = classmethod(from_element)
515
def __eq__(self, other):
516
"""Compare two sets by comparing their contents.
522
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
525
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
529
if not isinstance(other, Inventory):
530
return NotImplemented
532
if len(self._byid) != len(other._byid):
533
# shortcut: obviously not the same
536
return self._byid == other._byid
539
def __ne__(self, other):
540
return not (self == other)
544
raise ValueError('not hashable')
548
def get_idpath(self, file_id):
549
"""Return a list of file_ids for the path to an entry.
551
The list contains one element for each directory followed by
552
the id of the file itself. So the length of the returned list
553
is equal to the depth of the file in the tree, counting the
554
root directory as depth 1.
557
while file_id != None:
559
ie = self._byid[file_id]
561
raise BzrError("file_id {%s} not found in inventory" % file_id)
562
p.insert(0, ie.file_id)
563
file_id = ie.parent_id
567
def id2path(self, file_id):
568
"""Return as a list the path to file_id."""
570
# get all names, skipping root
571
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
572
return os.sep.join(p)
576
def path2id(self, name):
577
"""Walk down through directories to return entry of last component.
579
names may be either a list of path components, or a single
580
string, in which case it is automatically split.
582
This returns the entry of the last component in the path,
583
which may be either a file or a directory.
585
Returns None iff the path is not found.
587
if isinstance(name, types.StringTypes):
588
name = splitpath(name)
590
mutter("lookup path %r" % name)
595
cie = parent.children[f]
597
assert cie.parent_id == parent.file_id
603
return parent.file_id
606
def has_filename(self, names):
607
return bool(self.path2id(names))
610
def has_id(self, file_id):
611
return self._byid.has_key(file_id)
614
def rename(self, file_id, new_parent_id, new_name):
615
"""Move a file within the inventory.
617
This can change either the name, or the parent, or both.
619
This does not move the working file."""
620
if not is_valid_name(new_name):
621
raise BzrError("not an acceptable filename: %r" % new_name)
623
new_parent = self._byid[new_parent_id]
624
if new_name in new_parent.children:
625
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
627
new_parent_idpath = self.get_idpath(new_parent_id)
628
if file_id in new_parent_idpath:
629
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
630
% (self.id2path(file_id), self.id2path(new_parent_id)))
632
file_ie = self._byid[file_id]
633
old_parent = self._byid[file_ie.parent_id]
635
# TODO: Don't leave things messed up if this fails
637
del old_parent.children[file_ie.name]
638
new_parent.children[new_name] = file_ie
640
file_ie.name = new_name
641
file_ie.parent_id = new_parent_id
646
_NAME_RE = re.compile(r'^[^/\\]+$')
648
def is_valid_name(name):
649
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
