« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 233
- compare with another revision
- download diff
- download tarball
- view history from revision 233
- Committer: mbp at sourcefrog
- Date: 2005-04-11 01:58:58 UTC
- Revision ID: mbp@sourcefrog.net-20050411015857-f781b0d325632cf5cacd2481
- more output from test.sh
- write revison-history in a way that is hardlink-safe
- write revison-history in a way that is hardlink-safe
- files added:
- .rsyncexclude
- urlgrabber
- files modified:
- .bzrignore
added
removed
bzrlib/inventory.py
18
18
19
19
# TODO: Maybe store inventory_id in the file? Not really needed.
20
20
21
__copyright__ = "Copyright (C) 2005 Canonical Ltd."
22
21
__author__ = "Martin Pool <mbp@canonical.com>"
23
22
24
import sys, os.path, types
23
24
# This should really be an id randomly assigned when the tree is
25
# created, but it's not for now.
26
ROOT_ID = "TREE_ROOT"
27
28
29
import sys, os.path, types, re
25
30
from sets import Set
26
31
27
32
try:
30
35
from elementtree.ElementTree import Element, ElementTree, SubElement
31
36
32
37
from xml import XMLMixin
33
from errors import bailout
38
from errors import bailout, BzrError
34
39
35
40
import bzrlib
36
41
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
59
64
60
65
>>> i = Inventory()
61
66
>>> i.path2id('')
62
>>> i.add(InventoryEntry('123', 'src', kind='directory'))
63
>>> i.add(InventoryEntry('2323', 'hello.c', parent_id='123'))
67
'TREE_ROOT'
68
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
69
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
64
70
>>> for j in i.iter_entries():
65
71
... print j
66
72
...
67
('src', InventoryEntry('123', 'src', kind='directory', parent_id=None))
73
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
68
74
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
69
>>> i.add(InventoryEntry('2323', 'bye.c', parent_id='123'))
75
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
70
76
Traceback (most recent call last):
71
77
...
72
78
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'))
79
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
80
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
75
81
>>> i.path2id('src/wibble')
76
82
'2325'
77
83
>>> '2325' in i
78
84
True
79
>>> i.add(InventoryEntry('2326', 'wibble.c', parent_id='2325'))
85
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
80
86
>>> i['2326']
81
87
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
82
88
>>> for j in i.iter_entries():
95
101
But those depend on its position within a particular inventory, and
96
102
it would be nice not to need to hold the backpointer here.
97
103
"""
98
def __init__(self, file_id, name, kind='file', text_id=None,
99
parent_id=None):
104
105
# TODO: split InventoryEntry into subclasses for files,
106
# directories, etc etc.
107
108
def __init__(self, file_id, name, kind, parent_id, text_id=None):
100
109
"""Create an InventoryEntry
101
110
102
111
The filename must be a single component, relative to the
103
112
parent directory; it cannot be a whole path or relative name.
104
113
105
>>> e = InventoryEntry('123', 'hello.c')
114
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
106
115
>>> e.name
107
116
'hello.c'
108
117
>>> e.file_id
109
118
'123'
110
>>> e = InventoryEntry('123', 'src/hello.c')
119
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
111
120
Traceback (most recent call last):
112
121
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
113
122
"""
126
135
self.text_size = None
127
136
if kind == 'directory':
128
137
self.children = {}
138
else:
139
assert kind == 'file'
140
141
142
def sorted_children(self):
143
l = self.children.items()
144
l.sort()
145
return l
129
146
130
147
131
148
def copy(self):
132
149
other = InventoryEntry(self.file_id, self.name, self.kind,
133
self.text_id, self.parent_id)
150
self.parent_id, text_id=self.text_id)
134
151
other.text_sha1 = self.text_sha1
135
152
other.text_size = self.text_size
136
153
return other
153
170
e.set('file_id', self.file_id)
154
171
e.set('kind', self.kind)
155
172
156
if self.text_size is not None:
173
if self.text_size != None:
157
174
e.set('text_size', '%d' % self.text_size)
158
175
159
for f in ['text_id', 'text_sha1', 'parent_id']:
176
for f in ['text_id', 'text_sha1']:
160
177
v = getattr(self, f)
161
if v is not None:
178
if v != None:
162
179
e.set(f, v)
163
180
181
# to be conservative, we don't externalize the root pointers
182
# for now, leaving them as null in the xml form. in a future
183
# version it will be implied by nested elements.
184
if self.parent_id != ROOT_ID:
185
assert isinstance(self.parent_id, basestring)
186
e.set('parent_id', self.parent_id)
187
164
188
e.tail = '\n'
165
189
166
190
return e
168
192
169
193
def from_element(cls, elt):
170
194
assert elt.tag == 'entry'
171
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'))
195
196
## original format inventories don't have a parent_id for
197
## nodes in the root directory, but it's cleaner to use one
198
## internally.
199
parent_id = elt.get('parent_id')
200
if parent_id == None:
201
parent_id = ROOT_ID
202
203
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
172
204
self.text_id = elt.get('text_id')
173
205
self.text_sha1 = elt.get('text_sha1')
174
self.parent_id = elt.get('parent_id')
175
206
176
207
## mutter("read inventoryentry: %r" % (elt.attrib))
177
208
199
230
200
231
201
232
233
class RootEntry(InventoryEntry):
234
def __init__(self, file_id):
235
self.file_id = file_id
236
self.children = {}
237
self.kind = 'root_directory'
238
self.parent_id = None
239
self.name = ''
240
241
def __cmp__(self, other):
242
if self is other:
243
return 0
244
if not isinstance(other, RootEntry):
245
return NotImplemented
246
return cmp(self.file_id, other.file_id) \
247
or cmp(self.children, other.children)
248
249
250
202
251
class Inventory(XMLMixin):
203
252
"""Inventory of versioned files in a tree.
204
253
217
266
returned quickly.
218
267
219
268
InventoryEntry objects must not be modified after they are
220
inserted.
269
inserted, other than through the Inventory API.
221
270
222
271
>>> inv = Inventory()
223
272
>>> inv.write_xml(sys.stdout)
224
273
<inventory>
225
274
</inventory>
226
>>> inv.add(InventoryEntry('123-123', 'hello.c'))
275
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
227
276
>>> inv['123-123'].name
228
277
'hello.c'
229
278
263
312
If a working directory is specified, the inventory is read
264
313
from there. If the file is specified, read from that. If not,
265
314
the inventory is created empty.
315
316
The inventory is created with a default root directory, with
317
an id of None.
266
318
"""
267
self._root = InventoryEntry(None, '', kind='directory')
268
self._byid = {None: self._root}
319
self.root = RootEntry(ROOT_ID)
320
self._byid = {self.root.file_id: self.root}
269
321
270
322
271
323
def __iter__(self):
277
329
return len(self._byid)
278
330
279
331
280
def iter_entries(self, parent_id=None):
332
def iter_entries(self, from_dir=None):
281
333
"""Return (path, entry) pairs, in order by name."""
282
kids = self[parent_id].children.items()
334
if from_dir == None:
335
assert self.root
336
from_dir = self.root
337
elif isinstance(from_dir, basestring):
338
from_dir = self._byid[from_dir]
339
340
kids = from_dir.children.items()
283
341
kids.sort()
284
342
for name, ie in kids:
285
343
yield name, ie
286
344
if ie.kind == 'directory':
287
for cn, cie in self.iter_entries(parent_id=ie.file_id):
288
yield joinpath([name, cn]), cie
345
for cn, cie in self.iter_entries(from_dir=ie.file_id):
346
yield '/'.join((name, cn)), cie
347
289
348
290
349
291
350
def directories(self):
292
351
"""Return (path, entry) pairs for all directories.
293
352
"""
294
yield '', self._root
295
for path, entry in self.iter_entries():
296
if entry.kind == 'directory':
297
yield path, entry
353
def descend(parent_ie):
354
parent_name = parent_ie.name
355
yield parent_name, parent_ie
356
357
# directory children in sorted order
358
dn = []
359
for ie in parent_ie.children.itervalues():
360
if ie.kind == 'directory':
361
dn.append((ie.name, ie))
362
dn.sort()
363
364
for name, child_ie in dn:
365
for sub_name, sub_ie in descend(child_ie):
366
yield appendpath(parent_name, sub_name), sub_ie
367
368
for name, ie in descend(self.root):
369
yield name, ie
298
370
299
371
300
372
302
374
"""True if this entry contains a file with given id.
303
375
304
376
>>> inv = Inventory()
305
>>> inv.add(InventoryEntry('123', 'foo.c'))
377
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
306
378
>>> '123' in inv
307
379
True
308
380
>>> '456' in inv
315
387
"""Return the entry for given file_id.
316
388
317
389
>>> inv = Inventory()
318
>>> inv.add(InventoryEntry('123123', 'hello.c'))
390
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
319
391
>>> inv['123123'].name
320
392
'hello.c'
321
393
"""
322
return self._byid[file_id]
394
if file_id == None:
395
raise BzrError("can't look up file_id None")
396
397
try:
398
return self._byid[file_id]
399
except KeyError:
400
raise BzrError("file_id {%s} not in inventory" % file_id)
323
401
324
402
325
403
def get_child(self, parent_id, filename):
326
if parent_id == None:
327
return self._root.children.get(filename)
328
else:
329
return self[parent_id].children.get(filename)
404
return self[parent_id].children.get(filename)
330
405
331
406
332
407
def add(self, entry):
337
412
if entry.file_id in self._byid:
338
413
bailout("inventory already contains entry with id {%s}" % entry.file_id)
339
414
340
parent = self._byid[entry.parent_id]
341
if parent.kind != 'directory':
342
bailout("attempt to add under non-directory {%s}" % parent.file_id)
415
try:
416
parent = self._byid[entry.parent_id]
417
except KeyError:
418
bailout("parent_id {%s} not in inventory" % entry.parent_id)
343
419
344
420
if parent.children.has_key(entry.name):
345
421
bailout("%s is already versioned" %
357
433
if len(parts) == 0:
358
434
bailout("cannot re-add root of inventory")
359
435
360
if file_id is None:
436
if file_id == None:
361
437
file_id = bzrlib.branch.gen_file_id(relpath)
362
438
363
439
parent_id = self.path2id(parts[:-1])
440
assert parent_id != None
364
441
ie = InventoryEntry(file_id, parts[-1],
365
442
kind=kind, parent_id=parent_id)
366
443
return self.add(ie)
370
447
"""Remove entry by id.
371
448
372
449
>>> inv = Inventory()
373
>>> inv.add(InventoryEntry('123', 'foo.c'))
450
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
374
451
>>> '123' in inv
375
452
True
376
453
>>> del inv['123']
409
486
"""Construct from XML Element
410
487
411
488
>>> inv = Inventory()
412
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c'))
489
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
413
490
>>> elt = inv.to_element()
414
491
>>> inv2 = Inventory.from_element(elt)
415
492
>>> inv2 == inv
431
508
>>> i2 = Inventory()
432
509
>>> i1 == i2
433
510
True
434
>>> i1.add(InventoryEntry('123', 'foo'))
511
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
435
512
>>> i1 == i2
436
513
False
437
>>> i2.add(InventoryEntry('123', 'foo'))
514
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
438
515
>>> i1 == i2
439
516
True
440
517
"""
454
531
return 0
455
532
456
533
457
def id2path(self, file_id):
458
"""Return as a list the path to file_id."""
534
def get_idpath(self, file_id):
535
"""Return a list of file_ids for the path to an entry.
536
537
The list contains one element for each directory followed by
538
the id of the file itself. So the length of the returned list
539
is equal to the depth of the file in the tree, counting the
540
root directory as depth 1.
541
"""
459
542
p = []
460
543
while file_id != None:
461
ie = self[file_id]
462
p = [ie.name] + p
544
try:
545
ie = self._byid[file_id]
546
except KeyError:
547
bailout("file_id {%s} not found in inventory" % file_id)
548
p.insert(0, ie.file_id)
463
549
file_id = ie.parent_id
464
return joinpath(p)
550
return p
551
552
553
def id2path(self, file_id):
554
"""Return as a list the path to file_id."""
555
556
# get all names, skipping root
557
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
558
return '/'.join(p)
465
559
466
560
467
561
473
567
474
568
This returns the entry of the last component in the path,
475
569
which may be either a file or a directory.
570
571
Returns None iff the path is not found.
476
572
"""
477
573
if isinstance(name, types.StringTypes):
478
574
name = splitpath(name)
479
575
480
parent = self[None]
576
mutter("lookup path %r" % name)
577
578
parent = self.root
481
579
for f in name:
482
580
try:
483
581
cie = parent.children[f]
484
582
assert cie.name == f
583
assert cie.parent_id == parent.file_id
485
584
parent = cie
486
585
except KeyError:
487
586
# or raise an error?
495
594
496
595
497
596
def has_id(self, file_id):
498
assert isinstance(file_id, str)
499
597
return self._byid.has_key(file_id)
500
598
501
599
502
503
504
505
if __name__ == '__main__':
506
import doctest, inventory
507
doctest.testmod(inventory)
600
def rename(self, file_id, new_parent_id, new_name):
601
"""Move a file within the inventory.
602
603
This can change either the name, or the parent, or both.
604
605
This does not move the working file."""
606
if not is_valid_name(new_name):
607
bailout("not an acceptable filename: %r" % new_name)
608
609
new_parent = self._byid[new_parent_id]
610
if new_name in new_parent.children:
611
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
612
613
new_parent_idpath = self.get_idpath(new_parent_id)
614
if file_id in new_parent_idpath:
615
bailout("cannot move directory %r into a subdirectory of itself, %r"
616
% (self.id2path(file_id), self.id2path(new_parent_id)))
617
618
file_ie = self._byid[file_id]
619
old_parent = self._byid[file_ie.parent_id]
620
621
# TODO: Don't leave things messed up if this fails
622
623
del old_parent.children[file_ie.name]
624
new_parent.children[new_name] = file_ie
625
626
file_ie.name = new_name
627
file_ie.parent_id = new_parent_id
628
629
630
631
632
_NAME_RE = re.compile(r'^[^/\\]+$')
633
634
def is_valid_name(name):
635
return bool(_NAME_RE.match(name))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1