« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 0.1.78
- compare with another revision
- download diff
- download tarball
- view history from revision 0.1.78
- Committer: Martin Pool
- Date: 2005-06-30 08:40:59 UTC
- mto: This revision was merged to the branch mainline in revision 852.
- Revision ID: mbp@sourcefrog.net-20050630084059-d6eb6cb46972365b
Rename Weave.get_included to inclusions and getiter to get_iter
Refactor annotate() code
Refactor annotate() code
- files added:
- .bzrignore
- files removed:
- .bzrignore
- bzrlib
- doc
- elementtree
- notes
- urlgrabber
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
18
# TODO: Maybe store inventory_id in the file? Not really needed.
19
20
21
# This should really be an id randomly assigned when the tree is
22
# created, but it's not for now.
23
ROOT_ID = "TREE_ROOT"
24
25
26
import sys, os.path, types, re
27
from sets import Set
28
29
try:
30
from cElementTree import Element, ElementTree, SubElement
31
except ImportError:
32
from elementtree.ElementTree import Element, ElementTree, SubElement
33
34
from xml import XMLMixin
35
from errors import bailout, BzrError
36
37
import bzrlib
38
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
39
from bzrlib.trace import mutter
40
41
class InventoryEntry(XMLMixin):
42
"""Description of a versioned file.
43
44
An InventoryEntry has the following fields, which are also
45
present in the XML inventory-entry element:
46
47
* *file_id*
48
* *name*: (only the basename within the directory, must not
49
contain slashes)
50
* *kind*: "directory" or "file"
51
* *directory_id*: (if absent/null means the branch root directory)
52
* *text_sha1*: only for files
53
* *text_size*: in bytes, only for files
54
* *text_id*: identifier for the text version, only for files
55
56
InventoryEntries can also exist inside a WorkingTree
57
inventory, in which case they are not yet bound to a
58
particular revision of the file. In that case the text_sha1,
59
text_size and text_id are absent.
60
61
62
>>> i = Inventory()
63
>>> i.path2id('')
64
'TREE_ROOT'
65
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
66
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
67
>>> for j in i.iter_entries():
68
... print j
69
...
70
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
71
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
72
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
73
Traceback (most recent call last):
74
...
75
BzrError: ('inventory already contains entry with id {2323}', [])
76
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
77
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
78
>>> i.path2id('src/wibble')
79
'2325'
80
>>> '2325' in i
81
True
82
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
83
>>> i['2326']
84
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
85
>>> for j in i.iter_entries():
86
... print j[0]
87
... assert i.path2id(j[0])
88
...
89
src
90
src/bye.c
91
src/hello.c
92
src/wibble
93
src/wibble/wibble.c
94
>>> i.id2path('2326')
95
'src/wibble/wibble.c'
96
97
TODO: Maybe also keep the full path of the entry, and the children?
98
But those depend on its position within a particular inventory, and
99
it would be nice not to need to hold the backpointer here.
100
"""
101
102
# TODO: split InventoryEntry into subclasses for files,
103
# directories, etc etc.
104
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
106
"""Create an InventoryEntry
107
108
The filename must be a single component, relative to the
109
parent directory; it cannot be a whole path or relative name.
110
111
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
112
>>> e.name
113
'hello.c'
114
>>> e.file_id
115
'123'
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
117
Traceback (most recent call last):
118
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
119
"""
120
121
if len(splitpath(name)) != 1:
122
bailout('InventoryEntry name is not a simple filename: %r'
123
% name)
124
125
self.file_id = file_id
126
self.name = name
127
self.kind = kind
128
self.text_id = text_id
129
self.parent_id = parent_id
130
self.text_sha1 = None
131
self.text_size = None
132
if kind == 'directory':
133
self.children = {}
134
elif kind == 'file':
135
pass
136
else:
137
raise BzrError("unhandled entry kind %r" % kind)
138
139
140
141
def sorted_children(self):
142
l = self.children.items()
143
l.sort()
144
return l
145
146
147
def copy(self):
148
other = InventoryEntry(self.file_id, self.name, self.kind,
149
self.parent_id, text_id=self.text_id)
150
other.text_sha1 = self.text_sha1
151
other.text_size = self.text_size
152
return other
153
154
155
def __repr__(self):
156
return ("%s(%r, %r, kind=%r, parent_id=%r)"
157
% (self.__class__.__name__,
158
self.file_id,
159
self.name,
160
self.kind,
161
self.parent_id))
162
163
164
def to_element(self):
165
"""Convert to XML element"""
166
e = Element('entry')
167
168
e.set('name', self.name)
169
e.set('file_id', self.file_id)
170
e.set('kind', self.kind)
171
172
if self.text_size != None:
173
e.set('text_size', '%d' % self.text_size)
174
175
for f in ['text_id', 'text_sha1']:
176
v = getattr(self, f)
177
if v != None:
178
e.set(f, v)
179
180
# to be conservative, we don't externalize the root pointers
181
# for now, leaving them as null in the xml form. in a future
182
# version it will be implied by nested elements.
183
if self.parent_id != ROOT_ID:
184
assert isinstance(self.parent_id, basestring)
185
e.set('parent_id', self.parent_id)
186
187
e.tail = '\n'
188
189
return e
190
191
192
def from_element(cls, elt):
193
assert elt.tag == 'entry'
194
195
## original format inventories don't have a parent_id for
196
## nodes in the root directory, but it's cleaner to use one
197
## internally.
198
parent_id = elt.get('parent_id')
199
if parent_id == None:
200
parent_id = ROOT_ID
201
202
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
203
self.text_id = elt.get('text_id')
204
self.text_sha1 = elt.get('text_sha1')
205
206
## mutter("read inventoryentry: %r" % (elt.attrib))
207
208
v = elt.get('text_size')
209
self.text_size = v and int(v)
210
211
return self
212
213
214
from_element = classmethod(from_element)
215
216
def __cmp__(self, other):
217
if self is other:
218
return 0
219
if not isinstance(other, InventoryEntry):
220
return NotImplemented
221
222
return cmp(self.file_id, other.file_id) \
223
or cmp(self.name, other.name) \
224
or cmp(self.text_sha1, other.text_sha1) \
225
or cmp(self.text_size, other.text_size) \
226
or cmp(self.text_id, other.text_id) \
227
or cmp(self.parent_id, other.parent_id) \
228
or cmp(self.kind, other.kind)
229
230
231
232
class RootEntry(InventoryEntry):
233
def __init__(self, file_id):
234
self.file_id = file_id
235
self.children = {}
236
self.kind = 'root_directory'
237
self.parent_id = None
238
self.name = ''
239
240
def __cmp__(self, other):
241
if self is other:
242
return 0
243
if not isinstance(other, RootEntry):
244
return NotImplemented
245
return cmp(self.file_id, other.file_id) \
246
or cmp(self.children, other.children)
247
248
249
250
class Inventory(XMLMixin):
251
"""Inventory of versioned files in a tree.
252
253
This describes which file_id is present at each point in the tree,
254
and possibly the SHA-1 or other information about the file.
255
Entries can be looked up either by path or by file_id.
256
257
The inventory represents a typical unix file tree, with
258
directories containing files and subdirectories. We never store
259
the full path to a file, because renaming a directory implicitly
260
moves all of its contents. This class internally maintains a
261
lookup tree that allows the children under a directory to be
262
returned quickly.
263
264
InventoryEntry objects must not be modified after they are
265
inserted, other than through the Inventory API.
266
267
>>> inv = Inventory()
268
>>> inv.write_xml(sys.stdout)
269
<inventory>
270
</inventory>
271
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
272
>>> inv['123-123'].name
273
'hello.c'
274
275
May be treated as an iterator or set to look up file ids:
276
277
>>> bool(inv.path2id('hello.c'))
278
True
279
>>> '123-123' in inv
280
True
281
282
May also look up by name:
283
284
>>> [x[0] for x in inv.iter_entries()]
285
['hello.c']
286
287
>>> inv.write_xml(sys.stdout)
288
<inventory>
289
<entry file_id="123-123" kind="file" name="hello.c" />
290
</inventory>
291
292
"""
293
294
## TODO: Make sure only canonical filenames are stored.
295
296
## TODO: Do something sensible about the possible collisions on
297
## case-losing filesystems. Perhaps we should just always forbid
298
## such collisions.
299
300
## TODO: No special cases for root, rather just give it a file id
301
## like everything else.
302
303
## TODO: Probably change XML serialization to use nesting rather
304
## than parent_id pointers.
305
306
## TODO: Perhaps hold the ElementTree in memory and work directly
307
## on that rather than converting into Python objects every time?
308
309
def __init__(self):
310
"""Create or read an inventory.
311
312
If a working directory is specified, the inventory is read
313
from there. If the file is specified, read from that. If not,
314
the inventory is created empty.
315
316
The inventory is created with a default root directory, with
317
an id of None.
318
"""
319
self.root = RootEntry(ROOT_ID)
320
self._byid = {self.root.file_id: self.root}
321
322
323
def __iter__(self):
324
return iter(self._byid)
325
326
327
def __len__(self):
328
"""Returns number of entries."""
329
return len(self._byid)
330
331
332
def iter_entries(self, from_dir=None):
333
"""Return (path, entry) pairs, in order by name."""
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()
341
kids.sort()
342
for name, ie in kids:
343
yield name, ie
344
if ie.kind == 'directory':
345
for cn, cie in self.iter_entries(from_dir=ie.file_id):
346
yield os.path.join(name, cn), cie
347
348
349
350
def directories(self):
351
"""Return (path, entry) pairs for all directories.
352
"""
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
370
371
372
373
def __contains__(self, file_id):
374
"""True if this entry contains a file with given id.
375
376
>>> inv = Inventory()
377
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
378
>>> '123' in inv
379
True
380
>>> '456' in inv
381
False
382
"""
383
return file_id in self._byid
384
385
386
def __getitem__(self, file_id):
387
"""Return the entry for given file_id.
388
389
>>> inv = Inventory()
390
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
391
>>> inv['123123'].name
392
'hello.c'
393
"""
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)
401
402
403
def get_child(self, parent_id, filename):
404
return self[parent_id].children.get(filename)
405
406
407
def add(self, entry):
408
"""Add entry to inventory.
409
410
To add a file to a branch ready to be committed, use Branch.add,
411
which calls this."""
412
if entry.file_id in self._byid:
413
bailout("inventory already contains entry with id {%s}" % entry.file_id)
414
415
try:
416
parent = self._byid[entry.parent_id]
417
except KeyError:
418
bailout("parent_id {%s} not in inventory" % entry.parent_id)
419
420
if parent.children.has_key(entry.name):
421
bailout("%s is already versioned" %
422
appendpath(self.id2path(parent.file_id), entry.name))
423
424
self._byid[entry.file_id] = entry
425
parent.children[entry.name] = entry
426
427
428
def add_path(self, relpath, kind, file_id=None):
429
"""Add entry from a path.
430
431
The immediate parent must already be versioned"""
432
parts = bzrlib.osutils.splitpath(relpath)
433
if len(parts) == 0:
434
bailout("cannot re-add root of inventory")
435
436
if file_id == None:
437
file_id = bzrlib.branch.gen_file_id(relpath)
438
439
parent_id = self.path2id(parts[:-1])
440
assert parent_id != None
441
ie = InventoryEntry(file_id, parts[-1],
442
kind=kind, parent_id=parent_id)
443
return self.add(ie)
444
445
446
def __delitem__(self, file_id):
447
"""Remove entry by id.
448
449
>>> inv = Inventory()
450
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
451
>>> '123' in inv
452
True
453
>>> del inv['123']
454
>>> '123' in inv
455
False
456
"""
457
ie = self[file_id]
458
459
assert self[ie.parent_id].children[ie.name] == ie
460
461
# TODO: Test deleting all children; maybe hoist to a separate
462
# deltree method?
463
if ie.kind == 'directory':
464
for cie in ie.children.values():
465
del self[cie.file_id]
466
del ie.children
467
468
del self._byid[file_id]
469
del self[ie.parent_id].children[ie.name]
470
471
472
def id_set(self):
473
return Set(self._byid)
474
475
476
def to_element(self):
477
"""Convert to XML Element"""
478
e = Element('inventory')
479
e.text = '\n'
480
for path, ie in self.iter_entries():
481
e.append(ie.to_element())
482
return e
483
484
485
def from_element(cls, elt):
486
"""Construct from XML Element
487
488
>>> inv = Inventory()
489
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
490
>>> elt = inv.to_element()
491
>>> inv2 = Inventory.from_element(elt)
492
>>> inv2 == inv
493
True
494
"""
495
assert elt.tag == 'inventory'
496
o = cls()
497
for e in elt:
498
o.add(InventoryEntry.from_element(e))
499
return o
500
501
from_element = classmethod(from_element)
502
503
504
def __cmp__(self, other):
505
"""Compare two sets by comparing their contents.
506
507
>>> i1 = Inventory()
508
>>> i2 = Inventory()
509
>>> i1 == i2
510
True
511
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
512
>>> i1 == i2
513
False
514
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
515
>>> i1 == i2
516
True
517
"""
518
if self is other:
519
return 0
520
521
if not isinstance(other, Inventory):
522
return NotImplemented
523
524
if self.id_set() ^ other.id_set():
525
return 1
526
527
for file_id in self._byid:
528
c = cmp(self[file_id], other[file_id])
529
if c: return c
530
531
return 0
532
533
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
"""
542
p = []
543
while file_id != None:
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)
549
file_id = ie.parent_id
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 os.sep.join(p)
559
560
561
562
def path2id(self, name):
563
"""Walk down through directories to return entry of last component.
564
565
names may be either a list of path components, or a single
566
string, in which case it is automatically split.
567
568
This returns the entry of the last component in the path,
569
which may be either a file or a directory.
570
571
Returns None iff the path is not found.
572
"""
573
if isinstance(name, types.StringTypes):
574
name = splitpath(name)
575
576
mutter("lookup path %r" % name)
577
578
parent = self.root
579
for f in name:
580
try:
581
cie = parent.children[f]
582
assert cie.name == f
583
assert cie.parent_id == parent.file_id
584
parent = cie
585
except KeyError:
586
# or raise an error?
587
return None
588
589
return parent.file_id
590
591
592
def has_filename(self, names):
593
return bool(self.path2id(names))
594
595
596
def has_id(self, file_id):
597
return self._byid.has_key(file_id)
598
599
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