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()]
282
"""Create or read an inventory.
284
If a working directory is specified, the inventory is read
285
from there. If the file is specified, read from that. If not,
286
the inventory is created empty.
288
The inventory is created with a default root directory, with
291
self.root = RootEntry(ROOT_ID)
292
self._byid = {self.root.file_id: self.root}
296
return iter(self._byid)
300
"""Returns number of entries."""
301
return len(self._byid)
304
def iter_entries(self, from_dir=None):
305
"""Return (path, entry) pairs, in order by name."""
309
elif isinstance(from_dir, basestring):
310
from_dir = self._byid[from_dir]
312
kids = from_dir.children.items()
314
for name, ie in kids:
316
if ie.kind == 'directory':
317
for cn, cie in self.iter_entries(from_dir=ie.file_id):
318
yield os.path.join(name, cn), cie
322
"""Return list of (path, ie) for all entries except the root.
324
This may be faster than iter_entries.
327
def descend(dir_ie, dir_path):
328
kids = dir_ie.children.items()
330
for name, ie in kids:
331
child_path = os.path.join(dir_path, name)
332
accum.append((child_path, ie))
333
if ie.kind == 'directory':
334
descend(ie, child_path)
336
descend(self.root, '')
340
def directories(self):
341
"""Return (path, entry) pairs for all directories, including the root.
344
def descend(parent_ie, parent_path):
345
accum.append((parent_path, parent_ie))
347
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
350
for name, child_ie in kids:
351
child_path = os.path.join(parent_path, name)
352
descend(child_ie, child_path)
353
descend(self.root, '')
358
def __contains__(self, file_id):
359
"""True if this entry contains a file with given id.
361
>>> inv = Inventory()
362
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
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
>>> inv['123123'].name
380
return self._byid[file_id]
383
raise BzrError("can't look up file_id None")
385
raise BzrError("file_id {%s} not in inventory" % file_id)
388
def get_file_kind(self, file_id):
389
return self._byid[file_id].kind
391
def get_child(self, parent_id, filename):
392
return self[parent_id].children.get(filename)
395
def add(self, entry):
396
"""Add entry to inventory.
398
To add a file to a branch ready to be committed, use Branch.add,
400
if entry.file_id in self._byid:
401
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
404
parent = self._byid[entry.parent_id]
406
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
408
if parent.children.has_key(entry.name):
409
raise BzrError("%s is already versioned" %
410
appendpath(self.id2path(parent.file_id), entry.name))
412
self._byid[entry.file_id] = entry
413
parent.children[entry.name] = entry
416
def add_path(self, relpath, kind, file_id=None):
417
"""Add entry from a path.
419
The immediate parent must already be versioned"""
420
from bzrlib.errors import NotVersionedError
422
parts = bzrlib.osutils.splitpath(relpath)
424
raise BzrError("cannot re-add root of inventory")
427
from bzrlib.branch import gen_file_id
428
file_id = gen_file_id(relpath)
430
parent_path = parts[:-1]
431
parent_id = self.path2id(parent_path)
432
if parent_id == None:
433
raise NotVersionedError(parent_path)
435
ie = InventoryEntry(file_id, parts[-1],
436
kind=kind, parent_id=parent_id)
440
def __delitem__(self, file_id):
441
"""Remove entry by id.
443
>>> inv = Inventory()
444
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
453
assert self[ie.parent_id].children[ie.name] == ie
455
# TODO: Test deleting all children; maybe hoist to a separate
457
if ie.kind == 'directory':
458
for cie in ie.children.values():
459
del self[cie.file_id]
462
del self._byid[file_id]
463
del self[ie.parent_id].children[ie.name]
466
def to_element(self):
467
"""Convert to XML Element"""
468
from bzrlib.xml import Element
470
e = Element('inventory')
472
for path, ie in self.iter_entries():
473
e.append(ie.to_element())
477
def from_element(cls, elt):
478
"""Construct from XML Element
480
>>> inv = Inventory()
481
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
482
>>> elt = inv.to_element()
483
>>> inv2 = Inventory.from_element(elt)
487
assert elt.tag == 'inventory'
490
o.add(InventoryEntry.from_element(e))
493
from_element = classmethod(from_element)
496
def __eq__(self, other):
497
"""Compare two sets by comparing their contents.
503
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
506
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
510
if not isinstance(other, Inventory):
511
return NotImplemented
513
if len(self._byid) != len(other._byid):
514
# shortcut: obviously not the same
517
return self._byid == other._byid
520
def __ne__(self, other):
521
return not (self == other)
525
raise ValueError('not hashable')
529
def get_idpath(self, file_id):
530
"""Return a list of file_ids for the path to an entry.
532
The list contains one element for each directory followed by
533
the id of the file itself. So the length of the returned list
534
is equal to the depth of the file in the tree, counting the
535
root directory as depth 1.
538
while file_id != None:
540
ie = self._byid[file_id]
542
raise BzrError("file_id {%s} not found in inventory" % file_id)
543
p.insert(0, ie.file_id)
544
file_id = ie.parent_id
548
def id2path(self, file_id):
549
"""Return as a list the path to file_id."""
551
# get all names, skipping root
552
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
553
return os.sep.join(p)
557
def path2id(self, name):
558
"""Walk down through directories to return entry of last component.
560
names may be either a list of path components, or a single
561
string, in which case it is automatically split.
563
This returns the entry of the last component in the path,
564
which may be either a file or a directory.
566
Returns None iff the path is not found.
568
if isinstance(name, types.StringTypes):
569
name = splitpath(name)
571
mutter("lookup path %r" % name)
576
cie = parent.children[f]
578
assert cie.parent_id == parent.file_id
584
return parent.file_id
587
def has_filename(self, names):
588
return bool(self.path2id(names))
591
def has_id(self, file_id):
592
return self._byid.has_key(file_id)
595
def rename(self, file_id, new_parent_id, new_name):
596
"""Move a file within the inventory.
598
This can change either the name, or the parent, or both.
600
This does not move the working file."""
601
if not is_valid_name(new_name):
602
raise BzrError("not an acceptable filename: %r" % new_name)
604
new_parent = self._byid[new_parent_id]
605
if new_name in new_parent.children:
606
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
608
new_parent_idpath = self.get_idpath(new_parent_id)
609
if file_id in new_parent_idpath:
610
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
611
% (self.id2path(file_id), self.id2path(new_parent_id)))
613
file_ie = self._byid[file_id]
614
old_parent = self._byid[file_ie.parent_id]
616
# TODO: Don't leave things messed up if this fails
618
del old_parent.children[file_ie.name]
619
new_parent.children[new_name] = file_ie
621
file_ie.name = new_name
622
file_ie.parent_id = new_parent_id
627
_NAME_RE = re.compile(r'^[^/\\]+$')
629
def is_valid_name(name):
630
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
