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', ]
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
106
"""Create an InventoryEntry
108
The filename must be a single component, relative to the
109
parent directory; it cannot be a whole path or relative name.
111
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
117
Traceback (most recent call last):
118
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
120
if '/' in name or '\\' in name:
121
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
123
self.text_sha1 = None
124
self.text_size = None
126
self.file_id = file_id
129
self.text_id = text_id
130
self.parent_id = parent_id
131
if kind == 'directory':
136
raise BzrError("unhandled entry kind %r" % kind)
140
def sorted_children(self):
141
l = self.children.items()
147
other = InventoryEntry(self.file_id, self.name, self.kind,
148
self.parent_id, text_id=self.text_id)
149
other.text_sha1 = self.text_sha1
150
other.text_size = self.text_size
151
# note that children are *not* copied; they're pulled across when
157
return ("%s(%r, %r, kind=%r, parent_id=%r)"
158
% (self.__class__.__name__,
165
def __eq__(self, other):
166
if not isinstance(other, InventoryEntry):
167
return NotImplemented
169
return (self.file_id == other.file_id) \
170
and (self.name == other.name) \
171
and (self.text_sha1 == other.text_sha1) \
172
and (self.text_size == other.text_size) \
173
and (self.text_id == other.text_id) \
174
and (self.parent_id == other.parent_id) \
175
and (self.kind == other.kind)
178
def __ne__(self, other):
179
return not (self == other)
182
raise ValueError('not hashable')
186
class RootEntry(InventoryEntry):
187
def __init__(self, file_id):
188
self.file_id = file_id
190
self.kind = 'root_directory'
191
self.parent_id = None
194
def __eq__(self, other):
195
if not isinstance(other, RootEntry):
196
return NotImplemented
198
return (self.file_id == other.file_id) \
199
and (self.children == other.children)
203
class Inventory(object):
204
"""Inventory of versioned files in a tree.
206
This describes which file_id is present at each point in the tree,
207
and possibly the SHA-1 or other information about the file.
208
Entries can be looked up either by path or by file_id.
210
The inventory represents a typical unix file tree, with
211
directories containing files and subdirectories. We never store
212
the full path to a file, because renaming a directory implicitly
213
moves all of its contents. This class internally maintains a
214
lookup tree that allows the children under a directory to be
217
InventoryEntry objects must not be modified after they are
218
inserted, other than through the Inventory API.
220
>>> inv = Inventory()
221
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
222
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT')
223
>>> inv['123-123'].name
226
May be treated as an iterator or set to look up file ids:
228
>>> bool(inv.path2id('hello.c'))
233
May also look up by name:
235
>>> [x[0] for x in inv.iter_entries()]
237
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
238
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
239
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
241
def __init__(self, root_id=ROOT_ID):
242
"""Create or read an inventory.
244
If a working directory is specified, the inventory is read
245
from there. If the file is specified, read from that. If not,
246
the inventory is created empty.
248
The inventory is created with a default root directory, with
251
# We are letting Branch(init=True) create a unique inventory
252
# root id. Rather than generating a random one here.
254
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
255
self.root = RootEntry(root_id)
256
self._byid = {self.root.file_id: self.root}
260
return iter(self._byid)
264
"""Returns number of entries."""
265
return len(self._byid)
268
def iter_entries(self, from_dir=None):
269
"""Return (path, entry) pairs, in order by name."""
273
elif isinstance(from_dir, basestring):
274
from_dir = self._byid[from_dir]
276
kids = from_dir.children.items()
278
for name, ie in kids:
280
if ie.kind == 'directory':
281
for cn, cie in self.iter_entries(from_dir=ie.file_id):
282
yield os.path.join(name, cn), cie
286
"""Return list of (path, ie) for all entries except the root.
288
This may be faster than iter_entries.
291
def descend(dir_ie, dir_path):
292
kids = dir_ie.children.items()
294
for name, ie in kids:
295
child_path = os.path.join(dir_path, name)
296
accum.append((child_path, ie))
297
if ie.kind == 'directory':
298
descend(ie, child_path)
300
descend(self.root, '')
304
def directories(self):
305
"""Return (path, entry) pairs for all directories, including the root.
308
def descend(parent_ie, parent_path):
309
accum.append((parent_path, parent_ie))
311
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
314
for name, child_ie in kids:
315
child_path = os.path.join(parent_path, name)
316
descend(child_ie, child_path)
317
descend(self.root, '')
322
def __contains__(self, file_id):
323
"""True if this entry contains a file with given id.
325
>>> inv = Inventory()
326
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
327
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
333
return file_id in self._byid
336
def __getitem__(self, file_id):
337
"""Return the entry for given file_id.
339
>>> inv = Inventory()
340
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
341
InventoryEntry('123123', 'hello.c', kind='file', parent_id='TREE_ROOT')
342
>>> inv['123123'].name
346
return self._byid[file_id]
349
raise BzrError("can't look up file_id None")
351
raise BzrError("file_id {%s} not in inventory" % file_id)
354
def get_file_kind(self, file_id):
355
return self._byid[file_id].kind
357
def get_child(self, parent_id, filename):
358
return self[parent_id].children.get(filename)
361
def add(self, entry):
362
"""Add entry to inventory.
364
To add a file to a branch ready to be committed, use Branch.add,
367
Returns the new entry object.
369
if entry.file_id in self._byid:
370
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
372
if entry.parent_id == ROOT_ID or entry.parent_id is None:
373
entry.parent_id = self.root.file_id
376
parent = self._byid[entry.parent_id]
378
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
380
if parent.children.has_key(entry.name):
381
raise BzrError("%s is already versioned" %
382
appendpath(self.id2path(parent.file_id), entry.name))
384
self._byid[entry.file_id] = entry
385
parent.children[entry.name] = entry
389
def add_path(self, relpath, kind, file_id=None):
390
"""Add entry from a path.
392
The immediate parent must already be versioned.
394
Returns the new entry object."""
395
from bzrlib.branch import gen_file_id
397
parts = bzrlib.osutils.splitpath(relpath)
399
raise BzrError("cannot re-add root of inventory")
402
file_id = gen_file_id(relpath)
404
parent_path = parts[:-1]
405
parent_id = self.path2id(parent_path)
406
if parent_id == None:
407
raise NotVersionedError(parent_path)
409
ie = InventoryEntry(file_id, parts[-1],
410
kind=kind, parent_id=parent_id)
414
def __delitem__(self, file_id):
415
"""Remove entry by id.
417
>>> inv = Inventory()
418
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
419
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
428
assert self[ie.parent_id].children[ie.name] == ie
430
# TODO: Test deleting all children; maybe hoist to a separate
432
if ie.kind == 'directory':
433
for cie in ie.children.values():
434
del self[cie.file_id]
437
del self._byid[file_id]
438
del self[ie.parent_id].children[ie.name]
441
def __eq__(self, other):
442
"""Compare two sets by comparing their contents.
448
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
449
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
452
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
453
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
457
if not isinstance(other, Inventory):
458
return NotImplemented
460
if len(self._byid) != len(other._byid):
461
# shortcut: obviously not the same
464
return self._byid == other._byid
467
def __ne__(self, other):
468
return not (self == other)
472
raise ValueError('not hashable')
475
def get_idpath(self, file_id):
476
"""Return a list of file_ids for the path to an entry.
478
The list contains one element for each directory followed by
479
the id of the file itself. So the length of the returned list
480
is equal to the depth of the file in the tree, counting the
481
root directory as depth 1.
484
while file_id != None:
486
ie = self._byid[file_id]
488
raise BzrError("file_id {%s} not found in inventory" % file_id)
489
p.insert(0, ie.file_id)
490
file_id = ie.parent_id
494
def id2path(self, file_id):
495
"""Return as a list the path to file_id."""
497
# get all names, skipping root
498
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
499
return os.sep.join(p)
503
def path2id(self, name):
504
"""Walk down through directories to return entry of last component.
506
names may be either a list of path components, or a single
507
string, in which case it is automatically split.
509
This returns the entry of the last component in the path,
510
which may be either a file or a directory.
512
Returns None iff the path is not found.
514
if isinstance(name, types.StringTypes):
515
name = splitpath(name)
517
mutter("lookup path %r" % name)
522
cie = parent.children[f]
524
assert cie.parent_id == parent.file_id
530
return parent.file_id
533
def has_filename(self, names):
534
return bool(self.path2id(names))
537
def has_id(self, file_id):
538
return self._byid.has_key(file_id)
541
def rename(self, file_id, new_parent_id, new_name):
542
"""Move a file within the inventory.
544
This can change either the name, or the parent, or both.
546
This does not move the working file."""
547
if not is_valid_name(new_name):
548
raise BzrError("not an acceptable filename: %r" % new_name)
550
new_parent = self._byid[new_parent_id]
551
if new_name in new_parent.children:
552
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
554
new_parent_idpath = self.get_idpath(new_parent_id)
555
if file_id in new_parent_idpath:
556
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
557
% (self.id2path(file_id), self.id2path(new_parent_id)))
559
file_ie = self._byid[file_id]
560
old_parent = self._byid[file_ie.parent_id]
562
# TODO: Don't leave things messed up if this fails
564
del old_parent.children[file_ie.name]
565
new_parent.children[new_name] = file_ie
567
file_ie.name = new_name
568
file_ie.parent_id = new_parent_id
575
def is_valid_name(name):
578
_NAME_RE = re.compile(r'^[^/\\]+$')
580
return bool(_NAME_RE.match(name))