« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 162
- compare with another revision
- download diff
- download tarball
- view history from revision 162
- Committer: mbp at sourcefrog
- Date: 2005-04-04 09:50:24 UTC
- Revision ID: mbp@sourcefrog.net-20050404095024-4646dbcc42eada9e
workaround for python2.3 difflib bug
- files added:
- .bzrignore
- bzrlib
- doc
- elementtree
- notes
- files removed:
- .bzrignore
added
removed
bzrlib/inventory.py
1
# (C) 2005 Canonical Ltd
2
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.
7
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.
12
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
16
17
"""Inventories map files to their name in a revision."""
18
19
# TODO: Maybe store inventory_id in the file? Not really needed.
20
21
__copyright__ = "Copyright (C) 2005 Canonical Ltd."
22
__author__ = "Martin Pool <mbp@canonical.com>"
23
24
import sys, os.path, types, re
25
from sets import Set
26
27
try:
28
from cElementTree import Element, ElementTree, SubElement
29
except ImportError:
30
from elementtree.ElementTree import Element, ElementTree, SubElement
31
32
from xml import XMLMixin
33
from errors import bailout
34
35
import bzrlib
36
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
37
from bzrlib.trace import mutter
38
39
class InventoryEntry(XMLMixin):
40
"""Description of a versioned file.
41
42
An InventoryEntry has the following fields, which are also
43
present in the XML inventory-entry element:
44
45
* *file_id*
46
* *name*: (only the basename within the directory, must not
47
contain slashes)
48
* *kind*: "directory" or "file"
49
* *directory_id*: (if absent/null means the branch root directory)
50
* *text_sha1*: only for files
51
* *text_size*: in bytes, only for files
52
* *text_id*: identifier for the text version, only for files
53
54
InventoryEntries can also exist inside a WorkingTree
55
inventory, in which case they are not yet bound to a
56
particular revision of the file. In that case the text_sha1,
57
text_size and text_id are absent.
58
59
60
>>> i = Inventory()
61
>>> i.path2id('')
62
>>> i.add(InventoryEntry('123', 'src', kind='directory'))
63
>>> i.add(InventoryEntry('2323', 'hello.c', parent_id='123'))
64
>>> for j in i.iter_entries():
65
... print j
66
...
67
('src', InventoryEntry('123', 'src', kind='directory', parent_id=None))
68
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
69
>>> i.add(InventoryEntry('2323', 'bye.c', parent_id='123'))
70
Traceback (most recent call last):
71
...
72
BzrError: ('inventory already contains entry with id {2323}', [])
73
>>> i.add(InventoryEntry('2324', 'bye.c', parent_id='123'))
74
>>> i.add(InventoryEntry('2325', 'wibble', parent_id='123', kind='directory'))
75
>>> i.path2id('src/wibble')
76
'2325'
77
>>> '2325' in i
78
True
79
>>> i.add(InventoryEntry('2326', 'wibble.c', parent_id='2325'))
80
>>> i['2326']
81
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
82
>>> for j in i.iter_entries():
83
... print j[0]
84
... assert i.path2id(j[0])
85
...
86
src
87
src/bye.c
88
src/hello.c
89
src/wibble
90
src/wibble/wibble.c
91
>>> i.id2path('2326')
92
'src/wibble/wibble.c'
93
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.
97
"""
98
def __init__(self, file_id, name, kind='file', text_id=None,
99
parent_id=None):
100
"""Create an InventoryEntry
101
102
The filename must be a single component, relative to the
103
parent directory; it cannot be a whole path or relative name.
104
105
>>> e = InventoryEntry('123', 'hello.c')
106
>>> e.name
107
'hello.c'
108
>>> e.file_id
109
'123'
110
>>> e = InventoryEntry('123', 'src/hello.c')
111
Traceback (most recent call last):
112
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
113
"""
114
115
if len(splitpath(name)) != 1:
116
bailout('InventoryEntry name is not a simple filename: %r'
117
% name)
118
119
self.file_id = file_id
120
self.name = name
121
assert kind in ['file', 'directory']
122
self.kind = kind
123
self.text_id = text_id
124
self.parent_id = parent_id
125
self.text_sha1 = None
126
self.text_size = None
127
if kind == 'directory':
128
self.children = {}
129
130
131
def sorted_children(self):
132
l = self.children.items()
133
l.sort()
134
return l
135
136
137
def copy(self):
138
other = InventoryEntry(self.file_id, self.name, self.kind,
139
self.text_id, self.parent_id)
140
other.text_sha1 = self.text_sha1
141
other.text_size = self.text_size
142
return other
143
144
145
def __repr__(self):
146
return ("%s(%r, %r, kind=%r, parent_id=%r)"
147
% (self.__class__.__name__,
148
self.file_id,
149
self.name,
150
self.kind,
151
self.parent_id))
152
153
154
def to_element(self):
155
"""Convert to XML element"""
156
e = Element('entry')
157
158
e.set('name', self.name)
159
e.set('file_id', self.file_id)
160
e.set('kind', self.kind)
161
162
if self.text_size is not None:
163
e.set('text_size', '%d' % self.text_size)
164
165
for f in ['text_id', 'text_sha1', 'parent_id']:
166
v = getattr(self, f)
167
if v is not None:
168
e.set(f, v)
169
170
e.tail = '\n'
171
172
return e
173
174
175
def from_element(cls, elt):
176
assert elt.tag == 'entry'
177
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'))
178
self.text_id = elt.get('text_id')
179
self.text_sha1 = elt.get('text_sha1')
180
self.parent_id = elt.get('parent_id')
181
182
## mutter("read inventoryentry: %r" % (elt.attrib))
183
184
v = elt.get('text_size')
185
self.text_size = v and int(v)
186
187
return self
188
189
190
from_element = classmethod(from_element)
191
192
def __cmp__(self, other):
193
if self is other:
194
return 0
195
if not isinstance(other, InventoryEntry):
196
return NotImplemented
197
198
return cmp(self.file_id, other.file_id) \
199
or cmp(self.name, other.name) \
200
or cmp(self.text_sha1, other.text_sha1) \
201
or cmp(self.text_size, other.text_size) \
202
or cmp(self.text_id, other.text_id) \
203
or cmp(self.parent_id, other.parent_id) \
204
or cmp(self.kind, other.kind)
205
206
207
208
class RootEntry(InventoryEntry):
209
def __init__(self, file_id):
210
self.file_id = file_id
211
self.children = {}
212
self.kind = 'root_directory'
213
self.parent_id = None
214
self.name = ''
215
216
def __cmp__(self, other):
217
if self is other:
218
return 0
219
if not isinstance(other, RootEntry):
220
return NotImplemented
221
return cmp(self.file_id, other.file_id) \
222
or cmp(self.children, other.children)
223
224
225
226
class Inventory(XMLMixin):
227
"""Inventory of versioned files in a tree.
228
229
An Inventory acts like a set of InventoryEntry items. You can
230
also look files up by their file_id or name.
231
232
May be read from and written to a metadata file in a tree. To
233
manipulate the inventory (for example to add a file), it is read
234
in, modified, and then written back out.
235
236
The inventory represents a typical unix file tree, with
237
directories containing files and subdirectories. We never store
238
the full path to a file, because renaming a directory implicitly
239
moves all of its contents. This class internally maintains a
240
lookup tree that allows the children under a directory to be
241
returned quickly.
242
243
InventoryEntry objects must not be modified after they are
244
inserted, other than through the Inventory API.
245
246
>>> inv = Inventory()
247
>>> inv.write_xml(sys.stdout)
248
<inventory>
249
</inventory>
250
>>> inv.add(InventoryEntry('123-123', 'hello.c'))
251
>>> inv['123-123'].name
252
'hello.c'
253
254
May be treated as an iterator or set to look up file ids:
255
256
>>> bool(inv.path2id('hello.c'))
257
True
258
>>> '123-123' in inv
259
True
260
261
May also look up by name:
262
263
>>> [x[0] for x in inv.iter_entries()]
264
['hello.c']
265
266
>>> inv.write_xml(sys.stdout)
267
<inventory>
268
<entry file_id="123-123" kind="file" name="hello.c" />
269
</inventory>
270
271
"""
272
273
## TODO: Make sure only canonical filenames are stored.
274
275
## TODO: Do something sensible about the possible collisions on
276
## case-losing filesystems. Perhaps we should just always forbid
277
## such collisions.
278
279
## TODO: No special cases for root, rather just give it a file id
280
## like everything else.
281
282
## TODO: Probably change XML serialization to use nesting
283
284
def __init__(self):
285
"""Create or read an inventory.
286
287
If a working directory is specified, the inventory is read
288
from there. If the file is specified, read from that. If not,
289
the inventory is created empty.
290
291
The inventory is created with a default root directory, with
292
an id of None.
293
"""
294
self.root = RootEntry(None)
295
self._byid = {None: self.root}
296
297
298
def __iter__(self):
299
return iter(self._byid)
300
301
302
def __len__(self):
303
"""Returns number of entries."""
304
return len(self._byid)
305
306
307
def iter_entries(self, from_dir=None):
308
"""Return (path, entry) pairs, in order by name."""
309
if from_dir == None:
310
assert self.root
311
from_dir = self.root
312
elif isinstance(from_dir, basestring):
313
from_dir = self._byid[from_dir]
314
315
kids = from_dir.children.items()
316
kids.sort()
317
for name, ie in kids:
318
yield name, ie
319
if ie.kind == 'directory':
320
for cn, cie in self.iter_entries(from_dir=ie.file_id):
321
yield '/'.join((name, cn)), cie
322
323
324
325
def directories(self, from_dir=None):
326
"""Return (path, entry) pairs for all directories.
327
"""
328
def descend(parent_ie):
329
parent_name = parent_ie.name
330
yield parent_name, parent_ie
331
332
# directory children in sorted order
333
dn = []
334
for ie in parent_ie.children.itervalues():
335
if ie.kind == 'directory':
336
dn.append((ie.name, ie))
337
dn.sort()
338
339
for name, child_ie in dn:
340
for sub_name, sub_ie in descend(child_ie):
341
yield appendpath(parent_name, sub_name), sub_ie
342
343
for name, ie in descend(self.root):
344
yield name, ie
345
346
347
348
def __contains__(self, file_id):
349
"""True if this entry contains a file with given id.
350
351
>>> inv = Inventory()
352
>>> inv.add(InventoryEntry('123', 'foo.c'))
353
>>> '123' in inv
354
True
355
>>> '456' in inv
356
False
357
"""
358
return file_id in self._byid
359
360
361
def __getitem__(self, file_id):
362
"""Return the entry for given file_id.
363
364
>>> inv = Inventory()
365
>>> inv.add(InventoryEntry('123123', 'hello.c'))
366
>>> inv['123123'].name
367
'hello.c'
368
"""
369
return self._byid[file_id]
370
371
372
def get_child(self, parent_id, filename):
373
return self[parent_id].children.get(filename)
374
375
376
def add(self, entry):
377
"""Add entry to inventory.
378
379
To add a file to a branch ready to be committed, use Branch.add,
380
which calls this."""
381
if entry.file_id in self._byid:
382
bailout("inventory already contains entry with id {%s}" % entry.file_id)
383
384
try:
385
parent = self._byid[entry.parent_id]
386
except KeyError:
387
bailout("parent_id %r not in inventory" % entry.parent_id)
388
389
if parent.children.has_key(entry.name):
390
bailout("%s is already versioned" %
391
appendpath(self.id2path(parent.file_id), entry.name))
392
393
self._byid[entry.file_id] = entry
394
parent.children[entry.name] = entry
395
396
397
def add_path(self, relpath, kind, file_id=None):
398
"""Add entry from a path.
399
400
The immediate parent must already be versioned"""
401
parts = bzrlib.osutils.splitpath(relpath)
402
if len(parts) == 0:
403
bailout("cannot re-add root of inventory")
404
405
if file_id is None:
406
file_id = bzrlib.branch.gen_file_id(relpath)
407
408
parent_id = self.path2id(parts[:-1])
409
ie = InventoryEntry(file_id, parts[-1],
410
kind=kind, parent_id=parent_id)
411
return self.add(ie)
412
413
414
def __delitem__(self, file_id):
415
"""Remove entry by id.
416
417
>>> inv = Inventory()
418
>>> inv.add(InventoryEntry('123', 'foo.c'))
419
>>> '123' in inv
420
True
421
>>> del inv['123']
422
>>> '123' in inv
423
False
424
"""
425
ie = self[file_id]
426
427
assert self[ie.parent_id].children[ie.name] == ie
428
429
# TODO: Test deleting all children; maybe hoist to a separate
430
# deltree method?
431
if ie.kind == 'directory':
432
for cie in ie.children.values():
433
del self[cie.file_id]
434
del ie.children
435
436
del self._byid[file_id]
437
del self[ie.parent_id].children[ie.name]
438
439
440
def id_set(self):
441
return Set(self._byid)
442
443
444
def to_element(self):
445
"""Convert to XML Element"""
446
e = Element('inventory')
447
e.text = '\n'
448
for path, ie in self.iter_entries():
449
e.append(ie.to_element())
450
return e
451
452
453
def from_element(cls, elt):
454
"""Construct from XML Element
455
456
>>> inv = Inventory()
457
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c'))
458
>>> elt = inv.to_element()
459
>>> inv2 = Inventory.from_element(elt)
460
>>> inv2 == inv
461
True
462
"""
463
assert elt.tag == 'inventory'
464
o = cls()
465
for e in elt:
466
o.add(InventoryEntry.from_element(e))
467
return o
468
469
from_element = classmethod(from_element)
470
471
472
def __cmp__(self, other):
473
"""Compare two sets by comparing their contents.
474
475
>>> i1 = Inventory()
476
>>> i2 = Inventory()
477
>>> i1 == i2
478
True
479
>>> i1.add(InventoryEntry('123', 'foo'))
480
>>> i1 == i2
481
False
482
>>> i2.add(InventoryEntry('123', 'foo'))
483
>>> i1 == i2
484
True
485
"""
486
if self is other:
487
return 0
488
489
if not isinstance(other, Inventory):
490
return NotImplemented
491
492
if self.id_set() ^ other.id_set():
493
return 1
494
495
for file_id in self._byid:
496
c = cmp(self[file_id], other[file_id])
497
if c: return c
498
499
return 0
500
501
502
def id2path(self, file_id):
503
"""Return as a list the path to file_id."""
504
p = []
505
while file_id != None:
506
ie = self._byid[file_id]
507
p.insert(0, ie.name)
508
file_id = ie.parent_id
509
return '/'.join(p)
510
511
512
513
def path2id(self, name):
514
"""Walk down through directories to return entry of last component.
515
516
names may be either a list of path components, or a single
517
string, in which case it is automatically split.
518
519
This returns the entry of the last component in the path,
520
which may be either a file or a directory.
521
"""
522
if isinstance(name, types.StringTypes):
523
name = splitpath(name)
524
525
parent = self[None]
526
for f in name:
527
try:
528
cie = parent.children[f]
529
assert cie.name == f
530
parent = cie
531
except KeyError:
532
# or raise an error?
533
return None
534
535
return parent.file_id
536
537
538
def has_filename(self, names):
539
return bool(self.path2id(names))
540
541
542
def has_id(self, file_id):
543
return self._byid.has_key(file_id)
544
545
546
def rename(self, file_id, new_parent_id, new_name):
547
"""Move a file within the inventory.
548
549
This can change either the name, or the parent, or both.
550
551
This does not move the working file."""
552
if not is_valid_name(new_name):
553
bailout("not an acceptable filename: %r" % new_name)
554
555
new_parent = self._byid[new_parent_id]
556
if new_name in new_parent.children:
557
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
558
559
file_ie = self._byid[file_id]
560
old_parent = self._byid[file_ie.parent_id]
561
562
# TODO: Don't leave things messed up if this fails
563
564
del old_parent.children[file_ie.name]
565
new_parent.children[new_name] = file_ie
566
567
file_ie.name = new_name
568
file_ie.parent_id = new_parent_id
569
570
571
572
573
_NAME_RE = re.compile(r'^[^/\\]+$')
574
575
def is_valid_name(name):
576
return bool(_NAME_RE.match(name))
577
578
579
580
if __name__ == '__main__':
581
import doctest, inventory
582
doctest.testmod(inventory)
Loggerhead is a web-based interface for Breezy
Version: 2.0.1