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.errors import NotVersionedError
33
class InventoryEntry(object):
34
"""Description of a versioned file.
36
An InventoryEntry has the following fields, which are also
37
present in the XML inventory-entry element:
40
* *name*: (only the basename within the directory, must not
42
* *kind*: "directory" or "file"
43
* *directory_id*: (if absent/null means the branch root directory)
44
* *text_sha1*: only for files
45
* *text_size*: in bytes, only for files
46
* *text_id*: identifier for the text version, only for files
48
InventoryEntries can also exist inside a WorkingTree
49
inventory, in which case they are not yet bound to a
50
particular revision of the file. In that case the text_sha1,
51
text_size and text_id are absent.
57
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
58
InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT')
59
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
60
InventoryEntry('2323', 'hello.c', kind='file', parent_id='123')
61
>>> for j in i.iter_entries():
64
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
65
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
66
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
67
Traceback (most recent call last):
69
BzrError: inventory already contains entry with id {2323}
70
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
71
InventoryEntry('2324', 'bye.c', kind='file', parent_id='123')
72
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
73
InventoryEntry('2325', 'wibble', kind='directory', parent_id='123')
74
>>> i.path2id('src/wibble')
78
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
79
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
81
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
82
>>> for j in i.iter_entries():
84
... assert i.path2id(j[0])
94
TODO: Maybe also keep the full path of the entry, and the children?
95
But those depend on its position within a particular inventory, and
96
it would be nice not to need to hold the backpointer here.
99
# TODO: split InventoryEntry into subclasses for files,
100
# directories, etc etc.
102
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
103
'text_id', 'parent_id', 'children',
104
'text_version', 'entry_version', ]
106
def __init__(self, file_id, name, kind, parent_id, text_id=None):
107
"""Create an InventoryEntry
109
The filename must be a single component, relative to the
110
parent directory; it cannot be a whole path or relative name.
112
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
117
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
118
Traceback (most recent call last):
119
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
121
assert isinstance(name, basestring), name
122
if '/' in name or '\\' in name:
123
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
125
self.text_sha1 = None
126
self.text_size = None
128
self.file_id = file_id
131
self.text_id = text_id
132
self.parent_id = parent_id
133
if kind == 'directory':
138
raise BzrError("unhandled entry kind %r" % kind)
142
def sorted_children(self):
143
l = self.children.items()
149
other = InventoryEntry(self.file_id, self.name, self.kind,
150
self.parent_id, text_id=self.text_id)
151
other.text_sha1 = self.text_sha1
152
other.text_size = self.text_size
153
# note that children are *not* copied; they're pulled across when
159
return ("%s(%r, %r, kind=%r, parent_id=%r)"
160
% (self.__class__.__name__,
167
def __eq__(self, other):
168
if not isinstance(other, InventoryEntry):
169
return NotImplemented
171
return (self.file_id == other.file_id) \
172
and (self.name == other.name) \
173
and (self.text_sha1 == other.text_sha1) \
174
and (self.text_size == other.text_size) \
175
and (self.text_id == other.text_id) \
176
and (self.parent_id == other.parent_id) \
177
and (self.kind == other.kind)
180
def __ne__(self, other):
181
return not (self == other)
184
raise ValueError('not hashable')
188
class RootEntry(InventoryEntry):
189
def __init__(self, file_id):
190
self.file_id = file_id
192
self.kind = 'root_directory'
193
self.parent_id = None
196
def __eq__(self, other):
197
if not isinstance(other, RootEntry):
198
return NotImplemented
200
return (self.file_id == other.file_id) \
201
and (self.children == other.children)
205
class Inventory(object):
206
"""Inventory of versioned files in a tree.
208
This describes which file_id is present at each point in the tree,
209
and possibly the SHA-1 or other information about the file.
210
Entries can be looked up either by path or by file_id.
212
The inventory represents a typical unix file tree, with
213
directories containing files and subdirectories. We never store
214
the full path to a file, because renaming a directory implicitly
215
moves all of its contents. This class internally maintains a
216
lookup tree that allows the children under a directory to be
219
InventoryEntry objects must not be modified after they are
220
inserted, other than through the Inventory API.
222
>>> inv = Inventory()
223
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
224
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT')
225
>>> inv['123-123'].name
228
May be treated as an iterator or set to look up file ids:
230
>>> bool(inv.path2id('hello.c'))
235
May also look up by name:
237
>>> [x[0] for x in inv.iter_entries()]
239
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
240
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
241
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
243
def __init__(self, root_id=ROOT_ID):
244
"""Create or read an inventory.
246
If a working directory is specified, the inventory is read
247
from there. If the file is specified, read from that. If not,
248
the inventory is created empty.
250
The inventory is created with a default root directory, with
253
# We are letting Branch(init=True) create a unique inventory
254
# root id. Rather than generating a random one here.
256
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
257
self.root = RootEntry(root_id)
258
self._byid = {self.root.file_id: self.root}
262
return iter(self._byid)
266
"""Returns number of entries."""
267
return len(self._byid)
270
def iter_entries(self, from_dir=None):
271
"""Return (path, entry) pairs, in order by name."""
275
elif isinstance(from_dir, basestring):
276
from_dir = self._byid[from_dir]
278
kids = from_dir.children.items()
280
for name, ie in kids:
282
if ie.kind == 'directory':
283
for cn, cie in self.iter_entries(from_dir=ie.file_id):
284
yield os.path.join(name, cn), cie
288
"""Return list of (path, ie) for all entries except the root.
290
This may be faster than iter_entries.
293
def descend(dir_ie, dir_path):
294
kids = dir_ie.children.items()
296
for name, ie in kids:
297
child_path = os.path.join(dir_path, name)
298
accum.append((child_path, ie))
299
if ie.kind == 'directory':
300
descend(ie, child_path)
302
descend(self.root, '')
306
def directories(self):
307
"""Return (path, entry) pairs for all directories, including the root.
310
def descend(parent_ie, parent_path):
311
accum.append((parent_path, parent_ie))
313
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
316
for name, child_ie in kids:
317
child_path = os.path.join(parent_path, name)
318
descend(child_ie, child_path)
319
descend(self.root, '')
324
def __contains__(self, file_id):
325
"""True if this entry contains a file with given id.
327
>>> inv = Inventory()
328
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
329
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
335
return file_id in self._byid
338
def __getitem__(self, file_id):
339
"""Return the entry for given file_id.
341
>>> inv = Inventory()
342
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
343
InventoryEntry('123123', 'hello.c', kind='file', parent_id='TREE_ROOT')
344
>>> inv['123123'].name
348
return self._byid[file_id]
351
raise BzrError("can't look up file_id None")
353
raise BzrError("file_id {%s} not in inventory" % file_id)
356
def get_file_kind(self, file_id):
357
return self._byid[file_id].kind
359
def get_child(self, parent_id, filename):
360
return self[parent_id].children.get(filename)
363
def add(self, entry):
364
"""Add entry to inventory.
366
To add a file to a branch ready to be committed, use Branch.add,
369
Returns the new entry object.
371
if entry.file_id in self._byid:
372
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
374
if entry.parent_id == ROOT_ID or entry.parent_id is None:
375
entry.parent_id = self.root.file_id
378
parent = self._byid[entry.parent_id]
380
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
382
if parent.children.has_key(entry.name):
383
raise BzrError("%s is already versioned" %
384
appendpath(self.id2path(parent.file_id), entry.name))
386
self._byid[entry.file_id] = entry
387
parent.children[entry.name] = entry
391
def add_path(self, relpath, kind, file_id=None):
392
"""Add entry from a path.
394
The immediate parent must already be versioned.
396
Returns the new entry object."""
397
from bzrlib.branch import gen_file_id
399
parts = bzrlib.osutils.splitpath(relpath)
401
raise BzrError("cannot re-add root of inventory")
404
file_id = gen_file_id(relpath)
406
parent_path = parts[:-1]
407
parent_id = self.path2id(parent_path)
408
if parent_id == None:
409
raise NotVersionedError(parent_path)
411
ie = InventoryEntry(file_id, parts[-1],
412
kind=kind, parent_id=parent_id)
416
def __delitem__(self, file_id):
417
"""Remove entry by id.
419
>>> inv = Inventory()
420
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
421
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
430
assert self[ie.parent_id].children[ie.name] == ie
432
# TODO: Test deleting all children; maybe hoist to a separate
434
if ie.kind == 'directory':
435
for cie in ie.children.values():
436
del self[cie.file_id]
439
del self._byid[file_id]
440
del self[ie.parent_id].children[ie.name]
443
def __eq__(self, other):
444
"""Compare two sets by comparing their contents.
450
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
451
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
454
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
455
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
459
if not isinstance(other, Inventory):
460
return NotImplemented
462
if len(self._byid) != len(other._byid):
463
# shortcut: obviously not the same
466
return self._byid == other._byid
469
def __ne__(self, other):
470
return not (self == other)
474
raise ValueError('not hashable')
477
def get_idpath(self, file_id):
478
"""Return a list of file_ids for the path to an entry.
480
The list contains one element for each directory followed by
481
the id of the file itself. So the length of the returned list
482
is equal to the depth of the file in the tree, counting the
483
root directory as depth 1.
486
while file_id != None:
488
ie = self._byid[file_id]
490
raise BzrError("file_id {%s} not found in inventory" % file_id)
491
p.insert(0, ie.file_id)
492
file_id = ie.parent_id
496
def id2path(self, file_id):
497
"""Return as a list the path to file_id."""
499
# get all names, skipping root
500
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
501
return os.sep.join(p)
505
def path2id(self, name):
506
"""Walk down through directories to return entry of last component.
508
names may be either a list of path components, or a single
509
string, in which case it is automatically split.
511
This returns the entry of the last component in the path,
512
which may be either a file or a directory.
514
Returns None iff the path is not found.
516
if isinstance(name, types.StringTypes):
517
name = splitpath(name)
519
mutter("lookup path %r" % name)
524
cie = parent.children[f]
526
assert cie.parent_id == parent.file_id
532
return parent.file_id
535
def has_filename(self, names):
536
return bool(self.path2id(names))
539
def has_id(self, file_id):
540
return self._byid.has_key(file_id)
543
def rename(self, file_id, new_parent_id, new_name):
544
"""Move a file within the inventory.
546
This can change either the name, or the parent, or both.
548
This does not move the working file."""
549
if not is_valid_name(new_name):
550
raise BzrError("not an acceptable filename: %r" % new_name)
552
new_parent = self._byid[new_parent_id]
553
if new_name in new_parent.children:
554
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
556
new_parent_idpath = self.get_idpath(new_parent_id)
557
if file_id in new_parent_idpath:
558
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
559
% (self.id2path(file_id), self.id2path(new_parent_id)))
561
file_ie = self._byid[file_id]
562
old_parent = self._byid[file_ie.parent_id]
564
# TODO: Don't leave things messed up if this fails
566
del old_parent.children[file_ie.name]
567
new_parent.children[new_name] = file_ie
569
file_ie.name = new_name
570
file_ie.parent_id = new_parent_id
577
def is_valid_name(name):
580
_NAME_RE = re.compile(r'^[^/\\]+$')
582
return bool(_NAME_RE.match(name))