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.
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.file_id = file_id
119
self.text_id = text_id
120
self.parent_id = parent_id
121
if kind == 'directory':
126
raise BzrError("unhandled entry kind %r" % kind)
130
def sorted_children(self):
131
l = self.children.items()
137
other = InventoryEntry(self.file_id, self.name, self.kind,
138
self.parent_id, text_id=self.text_id)
139
other.text_sha1 = self.text_sha1
140
other.text_size = self.text_size
141
# note that children are *not* copied; they're pulled across when
147
return ("%s(%r, %r, kind=%r, parent_id=%r)"
148
% (self.__class__.__name__,
155
def to_element(self):
156
"""Convert to XML element"""
157
from bzrlib.xml import Element
161
e.set('name', self.name)
162
e.set('file_id', self.file_id)
163
e.set('kind', self.kind)
165
if self.text_size != None:
166
e.set('text_size', '%d' % self.text_size)
168
for f in ['text_id', 'text_sha1']:
173
# to be conservative, we don't externalize the root pointers
174
# for now, leaving them as null in the xml form. in a future
175
# version it will be implied by nested elements.
176
if self.parent_id != ROOT_ID:
177
assert isinstance(self.parent_id, basestring)
178
e.set('parent_id', self.parent_id)
185
def from_element(cls, elt):
186
assert elt.tag == 'entry'
188
## original format inventories don't have a parent_id for
189
## nodes in the root directory, but it's cleaner to use one
191
parent_id = elt.get('parent_id')
192
if parent_id == None:
195
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
196
self.text_id = elt.get('text_id')
197
self.text_sha1 = elt.get('text_sha1')
199
## mutter("read inventoryentry: %r" % (elt.attrib))
201
v = elt.get('text_size')
202
self.text_size = v and int(v)
207
from_element = classmethod(from_element)
209
def __eq__(self, other):
210
if not isinstance(other, InventoryEntry):
211
return NotImplemented
213
return (self.file_id == other.file_id) \
214
and (self.name == other.name) \
215
and (self.text_sha1 == other.text_sha1) \
216
and (self.text_size == other.text_size) \
217
and (self.text_id == other.text_id) \
218
and (self.parent_id == other.parent_id) \
219
and (self.kind == other.kind)
222
def __ne__(self, other):
223
return not (self == other)
226
raise ValueError('not hashable')
230
class RootEntry(InventoryEntry):
231
def __init__(self, file_id):
232
self.file_id = file_id
234
self.kind = 'root_directory'
235
self.parent_id = None
238
def __eq__(self, other):
239
if not isinstance(other, RootEntry):
240
return NotImplemented
242
return (self.file_id == other.file_id) \
243
and (self.children == other.children)
247
class Inventory(object):
248
"""Inventory of versioned files in a tree.
250
This describes which file_id is present at each point in the tree,
251
and possibly the SHA-1 or other information about the file.
252
Entries can be looked up either by path or by file_id.
254
The inventory represents a typical unix file tree, with
255
directories containing files and subdirectories. We never store
256
the full path to a file, because renaming a directory implicitly
257
moves all of its contents. This class internally maintains a
258
lookup tree that allows the children under a directory to be
261
InventoryEntry objects must not be modified after they are
262
inserted, other than through the Inventory API.
264
>>> inv = Inventory()
265
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
266
>>> inv['123-123'].name
269
May be treated as an iterator or set to look up file ids:
271
>>> bool(inv.path2id('hello.c'))
276
May also look up by name:
278
>>> [x[0] for x in inv.iter_entries()]
280
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
281
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
283
def __init__(self, root_id=ROOT_ID):
284
"""Create or read an inventory.
286
If a working directory is specified, the inventory is read
287
from there. If the file is specified, read from that. If not,
288
the inventory is created empty.
290
The inventory is created with a default root directory, with
293
# We are letting Branch(init=True) create a unique inventory
294
# root id. Rather than generating a random one here.
296
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
297
self.root = RootEntry(root_id)
298
self._byid = {self.root.file_id: self.root}
302
return iter(self._byid)
306
"""Returns number of entries."""
307
return len(self._byid)
310
def iter_entries(self, from_dir=None):
311
"""Return (path, entry) pairs, in order by name."""
315
elif isinstance(from_dir, basestring):
316
from_dir = self._byid[from_dir]
318
kids = from_dir.children.items()
320
for name, ie in kids:
322
if ie.kind == 'directory':
323
for cn, cie in self.iter_entries(from_dir=ie.file_id):
324
yield os.path.join(name, cn), cie
328
"""Return list of (path, ie) for all entries except the root.
330
This may be faster than iter_entries.
333
def descend(dir_ie, dir_path):
334
kids = dir_ie.children.items()
336
for name, ie in kids:
337
child_path = os.path.join(dir_path, name)
338
accum.append((child_path, ie))
339
if ie.kind == 'directory':
340
descend(ie, child_path)
342
descend(self.root, '')
346
def directories(self):
347
"""Return (path, entry) pairs for all directories, including the root.
350
def descend(parent_ie, parent_path):
351
accum.append((parent_path, parent_ie))
353
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
356
for name, child_ie in kids:
357
child_path = os.path.join(parent_path, name)
358
descend(child_ie, child_path)
359
descend(self.root, '')
364
def __contains__(self, file_id):
365
"""True if this entry contains a file with given id.
367
>>> inv = Inventory()
368
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
374
return file_id in self._byid
377
def __getitem__(self, file_id):
378
"""Return the entry for given file_id.
380
>>> inv = Inventory()
381
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
382
>>> inv['123123'].name
386
return self._byid[file_id]
389
raise BzrError("can't look up file_id None")
391
raise BzrError("file_id {%s} not in inventory" % file_id)
394
def get_file_kind(self, file_id):
395
return self._byid[file_id].kind
397
def get_child(self, parent_id, filename):
398
return self[parent_id].children.get(filename)
401
def add(self, entry):
402
"""Add entry to inventory.
404
To add a file to a branch ready to be committed, use Branch.add,
406
if entry.file_id in self._byid:
407
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
409
if entry.parent_id == ROOT_ID or entry.parent_id is None:
410
entry.parent_id = self.root.file_id
413
parent = self._byid[entry.parent_id]
415
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
417
if parent.children.has_key(entry.name):
418
raise BzrError("%s is already versioned" %
419
appendpath(self.id2path(parent.file_id), entry.name))
421
self._byid[entry.file_id] = entry
422
parent.children[entry.name] = entry
425
def add_path(self, relpath, kind, file_id=None):
426
"""Add entry from a path.
428
The immediate parent must already be versioned"""
429
from bzrlib.errors import NotVersionedError
431
parts = bzrlib.osutils.splitpath(relpath)
433
raise BzrError("cannot re-add root of inventory")
436
from bzrlib.branch import gen_file_id
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))
462
assert self[ie.parent_id].children[ie.name] == ie
464
# TODO: Test deleting all children; maybe hoist to a separate
466
if ie.kind == 'directory':
467
for cie in ie.children.values():
468
del self[cie.file_id]
471
del self._byid[file_id]
472
del self[ie.parent_id].children[ie.name]
475
def to_element(self):
476
"""Convert to XML Element"""
477
from bzrlib.xml import Element
479
e = Element('inventory')
481
if self.root.file_id not in (None, ROOT_ID):
482
e.set('file_id', self.root.file_id)
483
for path, ie in self.iter_entries():
484
e.append(ie.to_element())
488
def from_element(cls, elt):
489
"""Construct from XML Element
491
>>> inv = Inventory()
492
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
493
>>> elt = inv.to_element()
494
>>> inv2 = Inventory.from_element(elt)
498
# XXXX: doctest doesn't run this properly under python2.3
499
assert elt.tag == 'inventory'
500
root_id = elt.get('file_id') or ROOT_ID
503
ie = InventoryEntry.from_element(e)
504
if ie.parent_id == ROOT_ID:
505
ie.parent_id = root_id
509
from_element = classmethod(from_element)
512
def __eq__(self, other):
513
"""Compare two sets by comparing their contents.
519
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
522
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
526
if not isinstance(other, Inventory):
527
return NotImplemented
529
if len(self._byid) != len(other._byid):
530
# shortcut: obviously not the same
533
return self._byid == other._byid
536
def __ne__(self, other):
537
return not (self == other)
541
raise ValueError('not hashable')
545
def get_idpath(self, file_id):
546
"""Return a list of file_ids for the path to an entry.
548
The list contains one element for each directory followed by
549
the id of the file itself. So the length of the returned list
550
is equal to the depth of the file in the tree, counting the
551
root directory as depth 1.
554
while file_id != None:
556
ie = self._byid[file_id]
558
raise BzrError("file_id {%s} not found in inventory" % file_id)
559
p.insert(0, ie.file_id)
560
file_id = ie.parent_id
564
def id2path(self, file_id):
565
"""Return as a list the path to file_id."""
567
# get all names, skipping root
568
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
569
return os.sep.join(p)
573
def path2id(self, name):
574
"""Walk down through directories to return entry of last component.
576
names may be either a list of path components, or a single
577
string, in which case it is automatically split.
579
This returns the entry of the last component in the path,
580
which may be either a file or a directory.
582
Returns None iff the path is not found.
584
if isinstance(name, types.StringTypes):
585
name = splitpath(name)
587
mutter("lookup path %r" % name)
592
cie = parent.children[f]
594
assert cie.parent_id == parent.file_id
600
return parent.file_id
603
def has_filename(self, names):
604
return bool(self.path2id(names))
607
def has_id(self, file_id):
608
return self._byid.has_key(file_id)
611
def rename(self, file_id, new_parent_id, new_name):
612
"""Move a file within the inventory.
614
This can change either the name, or the parent, or both.
616
This does not move the working file."""
617
if not is_valid_name(new_name):
618
raise BzrError("not an acceptable filename: %r" % new_name)
620
new_parent = self._byid[new_parent_id]
621
if new_name in new_parent.children:
622
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
624
new_parent_idpath = self.get_idpath(new_parent_id)
625
if file_id in new_parent_idpath:
626
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
627
% (self.id2path(file_id), self.id2path(new_parent_id)))
629
file_ie = self._byid[file_id]
630
old_parent = self._byid[file_ie.parent_id]
632
# TODO: Don't leave things messed up if this fails
634
del old_parent.children[file_ie.name]
635
new_parent.children[new_name] = file_ie
637
file_ie.name = new_name
638
file_ie.parent_id = new_parent_id
643
_NAME_RE = re.compile(r'^[^/\\]+$')
645
def is_valid_name(name):
646
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
