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 cElementTree import Element, ElementTree, SubElement
28
from elementtree.ElementTree import Element, ElementTree, SubElement
30
from bzrlib.xml import XMLMixin
31
from bzrlib.errors import BzrError, BzrCheckError
34
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
35
from bzrlib.trace import mutter
37
class InventoryEntry(XMLMixin):
38
"""Description of a versioned file.
40
An InventoryEntry has the following fields, which are also
41
present in the XML inventory-entry element:
44
* *name*: (only the basename within the directory, must not
46
* *kind*: "directory" or "file"
47
* *directory_id*: (if absent/null means the branch root directory)
48
* *text_sha1*: only for files
49
* *text_size*: in bytes, only for files
50
* *text_id*: identifier for the text version, only for files
52
InventoryEntries can also exist inside a WorkingTree
53
inventory, in which case they are not yet bound to a
54
particular revision of the file. In that case the text_sha1,
55
text_size and text_id are absent.
61
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
62
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
63
>>> for j in i.iter_entries():
66
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
67
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
68
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
69
Traceback (most recent call last):
71
BzrError: inventory already contains entry with id {2323}
72
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
73
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
74
>>> i.path2id('src/wibble')
78
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
80
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
81
>>> for j in i.iter_entries():
83
... assert i.path2id(j[0])
93
TODO: Maybe also keep the full path of the entry, and the children?
94
But those depend on its position within a particular inventory, and
95
it would be nice not to need to hold the backpointer here.
98
# TODO: split InventoryEntry into subclasses for files,
99
# directories, etc etc.
104
def __init__(self, file_id, name, kind, parent_id, text_id=None):
105
"""Create an InventoryEntry
107
The filename must be a single component, relative to the
108
parent directory; it cannot be a whole path or relative name.
110
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
115
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
116
Traceback (most recent call last):
117
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
119
if '/' in name or '\\' in name:
120
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
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"""
165
e.set('name', self.name)
166
e.set('file_id', self.file_id)
167
e.set('kind', self.kind)
169
if self.text_size != None:
170
e.set('text_size', '%d' % self.text_size)
172
for f in ['text_id', 'text_sha1']:
177
# to be conservative, we don't externalize the root pointers
178
# for now, leaving them as null in the xml form. in a future
179
# version it will be implied by nested elements.
180
if self.parent_id != ROOT_ID:
181
assert isinstance(self.parent_id, basestring)
182
e.set('parent_id', self.parent_id)
189
def from_element(cls, elt):
190
assert elt.tag == 'entry'
192
## original format inventories don't have a parent_id for
193
## nodes in the root directory, but it's cleaner to use one
195
parent_id = elt.get('parent_id')
196
if parent_id == None:
199
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
200
self.text_id = elt.get('text_id')
201
self.text_sha1 = elt.get('text_sha1')
203
## mutter("read inventoryentry: %r" % (elt.attrib))
205
v = elt.get('text_size')
206
self.text_size = v and int(v)
211
from_element = classmethod(from_element)
213
def __eq__(self, other):
214
if not isinstance(other, InventoryEntry):
215
return NotImplemented
217
return (self.file_id == other.file_id) \
218
and (self.name == other.name) \
219
and (self.text_sha1 == other.text_sha1) \
220
and (self.text_size == other.text_size) \
221
and (self.text_id == other.text_id) \
222
and (self.parent_id == other.parent_id) \
223
and (self.kind == other.kind)
226
def __ne__(self, other):
227
return not (self == other)
230
raise ValueError('not hashable')
234
class RootEntry(InventoryEntry):
235
def __init__(self, file_id):
236
self.file_id = file_id
238
self.kind = 'root_directory'
239
self.parent_id = None
242
def __eq__(self, other):
243
if not isinstance(other, RootEntry):
244
return NotImplemented
246
return (self.file_id == other.file_id) \
247
and (self.children == other.children)
251
class Inventory(XMLMixin):
252
"""Inventory of versioned files in a tree.
254
This describes which file_id is present at each point in the tree,
255
and possibly the SHA-1 or other information about the file.
256
Entries can be looked up either by path or by file_id.
258
The inventory represents a typical unix file tree, with
259
directories containing files and subdirectories. We never store
260
the full path to a file, because renaming a directory implicitly
261
moves all of its contents. This class internally maintains a
262
lookup tree that allows the children under a directory to be
265
InventoryEntry objects must not be modified after they are
266
inserted, other than through the Inventory API.
268
>>> inv = Inventory()
269
>>> inv.write_xml(sys.stdout)
272
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
273
>>> inv['123-123'].name
276
May be treated as an iterator or set to look up file ids:
278
>>> bool(inv.path2id('hello.c'))
283
May also look up by name:
285
>>> [x[0] for x in inv.iter_entries()]
288
>>> inv.write_xml(sys.stdout)
290
<entry file_id="123-123" kind="file" name="hello.c" />
295
"""Create or read an inventory.
297
If a working directory is specified, the inventory is read
298
from there. If the file is specified, read from that. If not,
299
the inventory is created empty.
301
The inventory is created with a default root directory, with
304
self.root = RootEntry(ROOT_ID)
305
self._byid = {self.root.file_id: self.root}
309
return iter(self._byid)
313
"""Returns number of entries."""
314
return len(self._byid)
317
def iter_entries(self, from_dir=None):
318
"""Return (path, entry) pairs, in order by name."""
322
elif isinstance(from_dir, basestring):
323
from_dir = self._byid[from_dir]
325
kids = from_dir.children.items()
327
for name, ie in kids:
329
if ie.kind == 'directory':
330
for cn, cie in self.iter_entries(from_dir=ie.file_id):
331
yield os.path.join(name, cn), cie
335
"""Return list of (path, ie) for all entries except the root.
337
This may be faster than iter_entries.
340
def descend(dir_ie, dir_path):
341
kids = dir_ie.children.items()
343
for name, ie in kids:
344
child_path = os.path.join(dir_path, name)
345
accum.append((child_path, ie))
346
if ie.kind == 'directory':
347
descend(ie, child_path)
349
descend(self.root, '')
353
def directories(self):
354
"""Return (path, entry) pairs for all directories, including the root.
357
def descend(parent_ie, parent_path):
358
accum.append((parent_path, parent_ie))
360
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
363
for name, child_ie in kids:
364
child_path = os.path.join(parent_path, name)
365
descend(child_ie, child_path)
366
descend(self.root, '')
371
def __contains__(self, file_id):
372
"""True if this entry contains a file with given id.
374
>>> inv = Inventory()
375
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
381
return file_id in self._byid
384
def __getitem__(self, file_id):
385
"""Return the entry for given file_id.
387
>>> inv = Inventory()
388
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
389
>>> inv['123123'].name
393
return self._byid[file_id]
396
raise BzrError("can't look up file_id None")
398
raise BzrError("file_id {%s} not in inventory" % file_id)
401
def get_file_kind(self, file_id):
402
return self._byid[file_id].kind
404
def get_child(self, parent_id, filename):
405
return self[parent_id].children.get(filename)
408
def add(self, entry):
409
"""Add entry to inventory.
411
To add a file to a branch ready to be committed, use Branch.add,
413
if entry.file_id in self._byid:
414
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
417
parent = self._byid[entry.parent_id]
419
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
421
if parent.children.has_key(entry.name):
422
raise BzrError("%s is already versioned" %
423
appendpath(self.id2path(parent.file_id), entry.name))
425
self._byid[entry.file_id] = entry
426
parent.children[entry.name] = entry
429
def add_path(self, relpath, kind, file_id=None):
430
"""Add entry from a path.
432
The immediate parent must already be versioned"""
433
from bzrlib.errors import NotVersionedError
435
parts = bzrlib.osutils.splitpath(relpath)
437
raise BzrError("cannot re-add root of inventory")
440
file_id = bzrlib.branch.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
e = Element('inventory')
482
for path, ie in self.iter_entries():
483
e.append(ie.to_element())
487
def from_element(cls, elt):
488
"""Construct from XML Element
490
>>> inv = Inventory()
491
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
492
>>> elt = inv.to_element()
493
>>> inv2 = Inventory.from_element(elt)
497
assert elt.tag == 'inventory'
500
o.add(InventoryEntry.from_element(e))
503
from_element = classmethod(from_element)
506
def __eq__(self, other):
507
"""Compare two sets by comparing their contents.
513
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
516
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
520
if not isinstance(other, Inventory):
521
return NotImplemented
523
if len(self._byid) != len(other._byid):
524
# shortcut: obviously not the same
527
return self._byid == other._byid
530
def __ne__(self, other):
531
return not (self == other)
535
raise ValueError('not hashable')
539
def get_idpath(self, file_id):
540
"""Return a list of file_ids for the path to an entry.
542
The list contains one element for each directory followed by
543
the id of the file itself. So the length of the returned list
544
is equal to the depth of the file in the tree, counting the
545
root directory as depth 1.
548
while file_id != None:
550
ie = self._byid[file_id]
552
raise BzrError("file_id {%s} not found in inventory" % file_id)
553
p.insert(0, ie.file_id)
554
file_id = ie.parent_id
558
def id2path(self, file_id):
559
"""Return as a list the path to file_id."""
561
# get all names, skipping root
562
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
563
return os.sep.join(p)
567
def path2id(self, name):
568
"""Walk down through directories to return entry of last component.
570
names may be either a list of path components, or a single
571
string, in which case it is automatically split.
573
This returns the entry of the last component in the path,
574
which may be either a file or a directory.
576
Returns None iff the path is not found.
578
if isinstance(name, types.StringTypes):
579
name = splitpath(name)
581
mutter("lookup path %r" % name)
586
cie = parent.children[f]
588
assert cie.parent_id == parent.file_id
594
return parent.file_id
597
def has_filename(self, names):
598
return bool(self.path2id(names))
601
def has_id(self, file_id):
602
return self._byid.has_key(file_id)
605
def rename(self, file_id, new_parent_id, new_name):
606
"""Move a file within the inventory.
608
This can change either the name, or the parent, or both.
610
This does not move the working file."""
611
if not is_valid_name(new_name):
612
raise BzrError("not an acceptable filename: %r" % new_name)
614
new_parent = self._byid[new_parent_id]
615
if new_name in new_parent.children:
616
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
618
new_parent_idpath = self.get_idpath(new_parent_id)
619
if file_id in new_parent_idpath:
620
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
621
% (self.id2path(file_id), self.id2path(new_parent_id)))
623
file_ie = self._byid[file_id]
624
old_parent = self._byid[file_ie.parent_id]
626
# TODO: Don't leave things messed up if this fails
628
del old_parent.children[file_ie.name]
629
new_parent.children[new_name] = file_ie
631
file_ie.name = new_name
632
file_ie.parent_id = new_parent_id
637
_NAME_RE = re.compile(r'^[^/\\]+$')
639
def is_valid_name(name):
640
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
