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
30
from bzrlib.branch import gen_file_id
31
from bzrlib.errors import NotVersionedError
34
class InventoryEntry(object):
35
"""Description of a versioned file.
37
An InventoryEntry has the following fields, which are also
38
present in the XML inventory-entry element:
41
* *name*: (only the basename within the directory, must not
43
* *kind*: "directory" or "file"
44
* *directory_id*: (if absent/null means the branch root directory)
45
* *text_sha1*: only for files
46
* *text_size*: in bytes, only for files
47
* *text_id*: identifier for the text version, only for files
49
InventoryEntries can also exist inside a WorkingTree
50
inventory, in which case they are not yet bound to a
51
particular revision of the file. In that case the text_sha1,
52
text_size and text_id are absent.
58
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
59
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
60
>>> for j in i.iter_entries():
63
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
64
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
65
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
66
Traceback (most recent call last):
68
BzrError: inventory already contains entry with id {2323}
69
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
70
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
71
>>> i.path2id('src/wibble')
75
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
77
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
78
>>> for j in i.iter_entries():
80
... assert i.path2id(j[0])
90
TODO: Maybe also keep the full path of the entry, and the children?
91
But those depend on its position within a particular inventory, and
92
it would be nice not to need to hold the backpointer here.
95
# TODO: split InventoryEntry into subclasses for files,
96
# directories, etc etc.
98
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
99
'text_id', 'parent_id', 'children', ]
101
def __init__(self, file_id, name, kind, parent_id, text_id=None):
102
"""Create an InventoryEntry
104
The filename must be a single component, relative to the
105
parent directory; it cannot be a whole path or relative name.
107
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
112
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
113
Traceback (most recent call last):
114
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
116
if '/' in name or '\\' in name:
117
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
119
self.text_sha1 = None
120
self.text_size = None
122
self.file_id = file_id
125
self.text_id = text_id
126
self.parent_id = parent_id
127
if kind == 'directory':
132
raise BzrError("unhandled entry kind %r" % kind)
136
def sorted_children(self):
137
l = self.children.items()
143
other = InventoryEntry(self.file_id, self.name, self.kind,
144
self.parent_id, text_id=self.text_id)
145
other.text_sha1 = self.text_sha1
146
other.text_size = self.text_size
147
# note that children are *not* copied; they're pulled across when
153
return ("%s(%r, %r, kind=%r, parent_id=%r)"
154
% (self.__class__.__name__,
161
def to_element(self):
162
"""Convert to XML element"""
163
from bzrlib.xml import Element
167
e.set('name', self.name)
168
e.set('file_id', self.file_id)
169
e.set('kind', self.kind)
171
if self.text_size != None:
172
e.set('text_size', '%d' % self.text_size)
174
for f in ['text_id', 'text_sha1']:
179
# to be conservative, we don't externalize the root pointers
180
# for now, leaving them as null in the xml form. in a future
181
# version it will be implied by nested elements.
182
if self.parent_id != ROOT_ID:
183
assert isinstance(self.parent_id, basestring)
184
e.set('parent_id', self.parent_id)
191
def from_element(cls, elt):
192
assert elt.tag == 'entry'
194
## original format inventories don't have a parent_id for
195
## nodes in the root directory, but it's cleaner to use one
197
parent_id = elt.get('parent_id')
198
if parent_id == None:
201
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
202
self.text_id = elt.get('text_id')
203
self.text_sha1 = elt.get('text_sha1')
205
## mutter("read inventoryentry: %r" % (elt.attrib))
207
v = elt.get('text_size')
208
self.text_size = v and int(v)
213
from_element = classmethod(from_element)
215
def __eq__(self, other):
216
if not isinstance(other, InventoryEntry):
217
return NotImplemented
219
return (self.file_id == other.file_id) \
220
and (self.name == other.name) \
221
and (self.text_sha1 == other.text_sha1) \
222
and (self.text_size == other.text_size) \
223
and (self.text_id == other.text_id) \
224
and (self.parent_id == other.parent_id) \
225
and (self.kind == other.kind)
228
def __ne__(self, other):
229
return not (self == other)
232
raise ValueError('not hashable')
236
class RootEntry(InventoryEntry):
237
def __init__(self, file_id):
238
self.file_id = file_id
240
self.kind = 'root_directory'
241
self.parent_id = None
244
def __eq__(self, other):
245
if not isinstance(other, RootEntry):
246
return NotImplemented
248
return (self.file_id == other.file_id) \
249
and (self.children == other.children)
253
class Inventory(object):
254
"""Inventory of versioned files in a tree.
256
This describes which file_id is present at each point in the tree,
257
and possibly the SHA-1 or other information about the file.
258
Entries can be looked up either by path or by file_id.
260
The inventory represents a typical unix file tree, with
261
directories containing files and subdirectories. We never store
262
the full path to a file, because renaming a directory implicitly
263
moves all of its contents. This class internally maintains a
264
lookup tree that allows the children under a directory to be
267
InventoryEntry objects must not be modified after they are
268
inserted, other than through the Inventory API.
270
>>> inv = Inventory()
271
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
272
>>> inv['123-123'].name
275
May be treated as an iterator or set to look up file ids:
277
>>> bool(inv.path2id('hello.c'))
282
May also look up by name:
284
>>> [x[0] for x in inv.iter_entries()]
286
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
287
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
289
def __init__(self, root_id=ROOT_ID):
290
"""Create or read an inventory.
292
If a working directory is specified, the inventory is read
293
from there. If the file is specified, read from that. If not,
294
the inventory is created empty.
296
The inventory is created with a default root directory, with
299
# We are letting Branch(init=True) create a unique inventory
300
# root id. Rather than generating a random one here.
302
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
303
self.root = RootEntry(root_id)
304
self._byid = {self.root.file_id: self.root}
308
return iter(self._byid)
312
"""Returns number of entries."""
313
return len(self._byid)
316
def iter_entries(self, from_dir=None):
317
"""Return (path, entry) pairs, in order by name."""
321
elif isinstance(from_dir, basestring):
322
from_dir = self._byid[from_dir]
324
kids = from_dir.children.items()
326
for name, ie in kids:
328
if ie.kind == 'directory':
329
for cn, cie in self.iter_entries(from_dir=ie.file_id):
330
yield os.path.join(name, cn), cie
334
"""Return list of (path, ie) for all entries except the root.
336
This may be faster than iter_entries.
339
def descend(dir_ie, dir_path):
340
kids = dir_ie.children.items()
342
for name, ie in kids:
343
child_path = os.path.join(dir_path, name)
344
accum.append((child_path, ie))
345
if ie.kind == 'directory':
346
descend(ie, child_path)
348
descend(self.root, '')
352
def directories(self):
353
"""Return (path, entry) pairs for all directories, including the root.
356
def descend(parent_ie, parent_path):
357
accum.append((parent_path, parent_ie))
359
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
362
for name, child_ie in kids:
363
child_path = os.path.join(parent_path, name)
364
descend(child_ie, child_path)
365
descend(self.root, '')
370
def __contains__(self, file_id):
371
"""True if this entry contains a file with given id.
373
>>> inv = Inventory()
374
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
380
return file_id in self._byid
383
def __getitem__(self, file_id):
384
"""Return the entry for given file_id.
386
>>> inv = Inventory()
387
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
388
>>> inv['123123'].name
392
return self._byid[file_id]
395
raise BzrError("can't look up file_id None")
397
raise BzrError("file_id {%s} not in inventory" % file_id)
400
def get_file_kind(self, file_id):
401
return self._byid[file_id].kind
403
def get_child(self, parent_id, filename):
404
return self[parent_id].children.get(filename)
407
def add(self, entry):
408
"""Add entry to inventory.
410
To add a file to a branch ready to be committed, use Branch.add,
412
if entry.file_id in self._byid:
413
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
415
if entry.parent_id == ROOT_ID or entry.parent_id is None:
416
entry.parent_id = self.root.file_id
419
parent = self._byid[entry.parent_id]
421
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
423
if parent.children.has_key(entry.name):
424
raise BzrError("%s is already versioned" %
425
appendpath(self.id2path(parent.file_id), entry.name))
427
self._byid[entry.file_id] = entry
428
parent.children[entry.name] = entry
431
def add_path(self, relpath, kind, file_id=None):
432
"""Add entry from a path.
434
The immediate parent must already be versioned"""
435
parts = bzrlib.osutils.splitpath(relpath)
437
raise BzrError("cannot re-add root of inventory")
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
