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.xml import XMLMixin, Element
27
from bzrlib.errors import BzrError, BzrCheckError
29
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
30
from bzrlib.trace import mutter
32
class InventoryEntry(XMLMixin):
33
"""Description of a versioned file.
35
An InventoryEntry has the following fields, which are also
36
present in the XML inventory-entry element:
39
* *name*: (only the basename within the directory, must not
41
* *kind*: "directory" or "file"
42
* *directory_id*: (if absent/null means the branch root directory)
43
* *text_sha1*: only for files
44
* *text_size*: in bytes, only for files
45
* *text_id*: identifier for the text version, only for files
47
InventoryEntries can also exist inside a WorkingTree
48
inventory, in which case they are not yet bound to a
49
particular revision of the file. In that case the text_sha1,
50
text_size and text_id are absent.
56
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
57
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
58
>>> for j in i.iter_entries():
61
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
62
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
63
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
64
Traceback (most recent call last):
66
BzrError: inventory already contains entry with id {2323}
67
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
68
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
69
>>> i.path2id('src/wibble')
73
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
75
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
76
>>> for j in i.iter_entries():
78
... assert i.path2id(j[0])
88
TODO: Maybe also keep the full path of the entry, and the children?
89
But those depend on its position within a particular inventory, and
90
it would be nice not to need to hold the backpointer here.
93
# TODO: split InventoryEntry into subclasses for files,
94
# directories, etc etc.
99
def __init__(self, file_id, name, kind, parent_id, text_id=None):
100
"""Create an InventoryEntry
102
The filename must be a single component, relative to the
103
parent directory; it cannot be a whole path or relative name.
105
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
110
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
111
Traceback (most recent call last):
112
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
114
if '/' in name or '\\' in name:
115
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
117
self.file_id = file_id
120
self.text_id = text_id
121
self.parent_id = parent_id
122
if kind == 'directory':
127
raise BzrError("unhandled entry kind %r" % kind)
131
def sorted_children(self):
132
l = self.children.items()
138
other = InventoryEntry(self.file_id, self.name, self.kind,
139
self.parent_id, text_id=self.text_id)
140
other.text_sha1 = self.text_sha1
141
other.text_size = self.text_size
142
# note that children are *not* copied; they're pulled across when
148
return ("%s(%r, %r, kind=%r, parent_id=%r)"
149
% (self.__class__.__name__,
156
def to_element(self):
157
"""Convert to XML element"""
160
e.set('name', self.name)
161
e.set('file_id', self.file_id)
162
e.set('kind', self.kind)
164
if self.text_size != None:
165
e.set('text_size', '%d' % self.text_size)
167
for f in ['text_id', 'text_sha1']:
172
# to be conservative, we don't externalize the root pointers
173
# for now, leaving them as null in the xml form. in a future
174
# version it will be implied by nested elements.
175
if self.parent_id != ROOT_ID:
176
assert isinstance(self.parent_id, basestring)
177
e.set('parent_id', self.parent_id)
184
def from_element(cls, elt):
185
assert elt.tag == 'entry'
187
## original format inventories don't have a parent_id for
188
## nodes in the root directory, but it's cleaner to use one
190
parent_id = elt.get('parent_id')
191
if parent_id == None:
194
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
195
self.text_id = elt.get('text_id')
196
self.text_sha1 = elt.get('text_sha1')
198
## mutter("read inventoryentry: %r" % (elt.attrib))
200
v = elt.get('text_size')
201
self.text_size = v and int(v)
206
from_element = classmethod(from_element)
208
def __eq__(self, other):
209
if not isinstance(other, InventoryEntry):
210
return NotImplemented
212
return (self.file_id == other.file_id) \
213
and (self.name == other.name) \
214
and (self.text_sha1 == other.text_sha1) \
215
and (self.text_size == other.text_size) \
216
and (self.text_id == other.text_id) \
217
and (self.parent_id == other.parent_id) \
218
and (self.kind == other.kind)
221
def __ne__(self, other):
222
return not (self == other)
225
raise ValueError('not hashable')
229
class RootEntry(InventoryEntry):
230
def __init__(self, file_id):
231
self.file_id = file_id
233
self.kind = 'root_directory'
234
self.parent_id = None
237
def __eq__(self, other):
238
if not isinstance(other, RootEntry):
239
return NotImplemented
241
return (self.file_id == other.file_id) \
242
and (self.children == other.children)
246
class Inventory(XMLMixin):
247
"""Inventory of versioned files in a tree.
249
This describes which file_id is present at each point in the tree,
250
and possibly the SHA-1 or other information about the file.
251
Entries can be looked up either by path or by file_id.
253
The inventory represents a typical unix file tree, with
254
directories containing files and subdirectories. We never store
255
the full path to a file, because renaming a directory implicitly
256
moves all of its contents. This class internally maintains a
257
lookup tree that allows the children under a directory to be
260
InventoryEntry objects must not be modified after they are
261
inserted, other than through the Inventory API.
263
>>> inv = Inventory()
264
>>> inv.write_xml(sys.stdout)
267
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
268
>>> inv['123-123'].name
271
May be treated as an iterator or set to look up file ids:
273
>>> bool(inv.path2id('hello.c'))
278
May also look up by name:
280
>>> [x[0] for x in inv.iter_entries()]
283
>>> inv.write_xml(sys.stdout)
285
<entry file_id="123-123" kind="file" name="hello.c" />
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
self.root = RootEntry(ROOT_ID)
300
self._byid = {self.root.file_id: self.root}
304
return iter(self._byid)
308
"""Returns number of entries."""
309
return len(self._byid)
312
def iter_entries(self, from_dir=None):
313
"""Return (path, entry) pairs, in order by name."""
317
elif isinstance(from_dir, basestring):
318
from_dir = self._byid[from_dir]
320
kids = from_dir.children.items()
322
for name, ie in kids:
324
if ie.kind == 'directory':
325
for cn, cie in self.iter_entries(from_dir=ie.file_id):
326
yield os.path.join(name, cn), cie
330
"""Return list of (path, ie) for all entries except the root.
332
This may be faster than iter_entries.
335
def descend(dir_ie, dir_path):
336
kids = dir_ie.children.items()
338
for name, ie in kids:
339
child_path = os.path.join(dir_path, name)
340
accum.append((child_path, ie))
341
if ie.kind == 'directory':
342
descend(ie, child_path)
344
descend(self.root, '')
348
def directories(self):
349
"""Return (path, entry) pairs for all directories, including the root.
352
def descend(parent_ie, parent_path):
353
accum.append((parent_path, parent_ie))
355
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
358
for name, child_ie in kids:
359
child_path = os.path.join(parent_path, name)
360
descend(child_ie, child_path)
361
descend(self.root, '')
366
def __contains__(self, file_id):
367
"""True if this entry contains a file with given id.
369
>>> inv = Inventory()
370
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
376
return file_id in self._byid
379
def __getitem__(self, file_id):
380
"""Return the entry for given file_id.
382
>>> inv = Inventory()
383
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
384
>>> inv['123123'].name
388
return self._byid[file_id]
391
raise BzrError("can't look up file_id None")
393
raise BzrError("file_id {%s} not in inventory" % file_id)
396
def get_file_kind(self, file_id):
397
return self._byid[file_id].kind
399
def get_child(self, parent_id, filename):
400
return self[parent_id].children.get(filename)
403
def add(self, entry):
404
"""Add entry to inventory.
406
To add a file to a branch ready to be committed, use Branch.add,
408
if entry.file_id in self._byid:
409
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
412
parent = self._byid[entry.parent_id]
414
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
416
if parent.children.has_key(entry.name):
417
raise BzrError("%s is already versioned" %
418
appendpath(self.id2path(parent.file_id), entry.name))
420
self._byid[entry.file_id] = entry
421
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"""
428
from bzrlib.errors import NotVersionedError
430
parts = bzrlib.osutils.splitpath(relpath)
432
raise BzrError("cannot re-add root of inventory")
435
from bzrlib.branch import gen_file_id
436
file_id = gen_file_id(relpath)
438
parent_path = parts[:-1]
439
parent_id = self.path2id(parent_path)
440
if parent_id == None:
441
raise NotVersionedError(parent_path)
443
ie = InventoryEntry(file_id, parts[-1],
444
kind=kind, parent_id=parent_id)
448
def __delitem__(self, file_id):
449
"""Remove entry by id.
451
>>> inv = Inventory()
452
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
461
assert self[ie.parent_id].children[ie.name] == ie
463
# TODO: Test deleting all children; maybe hoist to a separate
465
if ie.kind == 'directory':
466
for cie in ie.children.values():
467
del self[cie.file_id]
470
del self._byid[file_id]
471
del self[ie.parent_id].children[ie.name]
474
def to_element(self):
475
"""Convert to XML Element"""
476
e = Element('inventory')
478
for path, ie in self.iter_entries():
479
e.append(ie.to_element())
483
def from_element(cls, elt):
484
"""Construct from XML Element
486
>>> inv = Inventory()
487
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
488
>>> elt = inv.to_element()
489
>>> inv2 = Inventory.from_element(elt)
493
assert elt.tag == 'inventory'
496
o.add(InventoryEntry.from_element(e))
499
from_element = classmethod(from_element)
502
def __eq__(self, other):
503
"""Compare two sets by comparing their contents.
509
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
512
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
516
if not isinstance(other, Inventory):
517
return NotImplemented
519
if len(self._byid) != len(other._byid):
520
# shortcut: obviously not the same
523
return self._byid == other._byid
526
def __ne__(self, other):
527
return not (self == other)
531
raise ValueError('not hashable')
535
def get_idpath(self, file_id):
536
"""Return a list of file_ids for the path to an entry.
538
The list contains one element for each directory followed by
539
the id of the file itself. So the length of the returned list
540
is equal to the depth of the file in the tree, counting the
541
root directory as depth 1.
544
while file_id != None:
546
ie = self._byid[file_id]
548
raise BzrError("file_id {%s} not found in inventory" % file_id)
549
p.insert(0, ie.file_id)
550
file_id = ie.parent_id
554
def id2path(self, file_id):
555
"""Return as a list the path to file_id."""
557
# get all names, skipping root
558
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
559
return os.sep.join(p)
563
def path2id(self, name):
564
"""Walk down through directories to return entry of last component.
566
names may be either a list of path components, or a single
567
string, in which case it is automatically split.
569
This returns the entry of the last component in the path,
570
which may be either a file or a directory.
572
Returns None iff the path is not found.
574
if isinstance(name, types.StringTypes):
575
name = splitpath(name)
577
mutter("lookup path %r" % name)
582
cie = parent.children[f]
584
assert cie.parent_id == parent.file_id
590
return parent.file_id
593
def has_filename(self, names):
594
return bool(self.path2id(names))
597
def has_id(self, file_id):
598
return self._byid.has_key(file_id)
601
def rename(self, file_id, new_parent_id, new_name):
602
"""Move a file within the inventory.
604
This can change either the name, or the parent, or both.
606
This does not move the working file."""
607
if not is_valid_name(new_name):
608
raise BzrError("not an acceptable filename: %r" % new_name)
610
new_parent = self._byid[new_parent_id]
611
if new_name in new_parent.children:
612
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
614
new_parent_idpath = self.get_idpath(new_parent_id)
615
if file_id in new_parent_idpath:
616
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
617
% (self.id2path(file_id), self.id2path(new_parent_id)))
619
file_ie = self._byid[file_id]
620
old_parent = self._byid[file_ie.parent_id]
622
# TODO: Don't leave things messed up if this fails
624
del old_parent.children[file_ie.name]
625
new_parent.children[new_name] = file_ie
627
file_ie.name = new_name
628
file_ie.parent_id = new_parent_id
633
_NAME_RE = re.compile(r'^[^/\\]+$')
635
def is_valid_name(name):
636
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
