« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 294
- compare with another revision
- download diff
- download tarball
- view history from revision 294
- Committer: Martin Pool
- Date: 2005-04-28 07:24:55 UTC
- Revision ID: mbp@sourcefrog.net-20050428072453-7b99afa993a1e549
todo
- files added:
- bzrlib/tests.py
- files removed:
- HACKING
- bzrlib/plugins
- bzrlib/selftest
- bzrlib/util
- bzrlib/util/effbot
- bzrlib/util/effbot/org
- contrib
- contrib/bash
- contrib/emacs
- contrib/zsh
- patches
- tools
- files renamed:
- contrib/newinventory.py => bzrlib/newinventory.py
- bzrlib/util/elementtree/ => elementtree/
- bzrlib/util/urlgrabber/ => urlgrabber/
- files modified:
- .bzrignore
added
removed
bzrlib/inventory.py
15
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
16
16
17
17
18
# TODO: Maybe also keep the full path of the entry, and the children?
19
# But those depend on its position within a particular inventory, and
20
# it would be nice not to need to hold the backpointer here.
21
22
# TODO: Perhaps split InventoryEntry into subclasses for files,
23
# directories, etc etc.
18
# TODO: Maybe store inventory_id in the file? Not really needed.
24
19
25
20
26
21
# This should really be an id randomly assigned when the tree is
29
24
30
25
31
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
32
36
33
37
import bzrlib
34
from bzrlib.errors import BzrError, BzrCheckError
35
36
38
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
37
39
from bzrlib.trace import mutter
38
from bzrlib.errors import NotVersionedError
39
40
40
41
class InventoryEntry(object):
41
class InventoryEntry(XMLMixin):
42
42
"""Description of a versioned file.
43
43
44
44
An InventoryEntry has the following fields, which are also
45
45
present in the XML inventory-entry element:
46
46
47
file_id
48
49
name
50
(within the parent directory)
51
52
kind
53
'directory' or 'file'
54
55
parent_id
56
file_id of the parent directory, or ROOT_ID
57
58
entry_version
59
the revision_id in which the name or parent of this file was
60
last changed
61
62
text_sha1
63
sha-1 of the text of the file
64
65
text_size
66
size in bytes of the text of the file
67
68
text_version
69
the revision_id in which the text of this file was introduced
70
71
(reading a version 4 tree created a text_id field.)
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
72
61
73
62
>>> i = Inventory()
74
63
>>> i.path2id('')
75
64
'TREE_ROOT'
76
65
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
77
InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT')
78
66
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
79
InventoryEntry('2323', 'hello.c', kind='file', parent_id='123')
80
67
>>> for j in i.iter_entries():
81
68
... print j
82
69
...
85
72
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
86
73
Traceback (most recent call last):
87
74
...
88
BzrError: inventory already contains entry with id {2323}
75
BzrError: ('inventory already contains entry with id {2323}', [])
89
76
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
90
InventoryEntry('2324', 'bye.c', kind='file', parent_id='123')
91
77
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
92
InventoryEntry('2325', 'wibble', kind='directory', parent_id='123')
93
78
>>> i.path2id('src/wibble')
94
79
'2325'
95
80
>>> '2325' in i
96
81
True
97
82
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
98
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
99
83
>>> i['2326']
100
84
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
101
85
>>> for j in i.iter_entries():
109
93
src/wibble/wibble.c
110
94
>>> i.id2path('2326')
111
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.
112
100
"""
101
102
# TODO: split InventoryEntry into subclasses for files,
103
# directories, etc etc.
113
104
114
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
115
'text_id', 'parent_id', 'children',
116
'text_version', 'entry_version', ]
117
118
119
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
120
106
"""Create an InventoryEntry
121
107
129
115
'123'
130
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
131
117
Traceback (most recent call last):
132
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
118
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
133
119
"""
134
assert isinstance(name, basestring), name
135
if '/' in name or '\\' in name:
136
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
137
138
self.text_version = None
139
self.entry_version = None
140
self.text_sha1 = None
141
self.text_size = None
120
121
if len(splitpath(name)) != 1:
122
bailout('InventoryEntry name is not a simple filename: %r'
123
% name)
124
142
125
self.file_id = file_id
143
126
self.name = name
144
127
self.kind = kind
145
128
self.text_id = text_id
146
129
self.parent_id = parent_id
130
self.text_sha1 = None
131
self.text_size = None
147
132
if kind == 'directory':
148
133
self.children = {}
149
134
elif kind == 'file':
161
146
162
147
def copy(self):
163
148
other = InventoryEntry(self.file_id, self.name, self.kind,
164
self.parent_id)
165
other.text_id = self.text_id
149
self.parent_id, text_id=self.text_id)
166
150
other.text_sha1 = self.text_sha1
167
151
other.text_size = self.text_size
168
other.text_version = self.text_version
169
# note that children are *not* copied; they're pulled across when
170
# others are added
171
152
return other
172
153
173
154
180
161
self.parent_id))
181
162
182
163
183
def __eq__(self, other):
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
184
219
if not isinstance(other, InventoryEntry):
185
220
return NotImplemented
186
221
187
return (self.file_id == other.file_id) \
188
and (self.name == other.name) \
189
and (self.text_sha1 == other.text_sha1) \
190
and (self.text_size == other.text_size) \
191
and (self.text_id == other.text_id) \
192
and (self.parent_id == other.parent_id) \
193
and (self.kind == other.kind) \
194
and (self.text_version == other.text_version) \
195
and (self.entry_version == other.entry_version)
196
197
198
def __ne__(self, other):
199
return not (self == other)
200
201
def __hash__(self):
202
raise ValueError('not hashable')
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)
203
229
204
230
205
231
211
237
self.parent_id = None
212
238
self.name = ''
213
239
214
def __eq__(self, other):
240
def __cmp__(self, other):
241
if self is other:
242
return 0
215
243
if not isinstance(other, RootEntry):
216
244
return NotImplemented
217
218
return (self.file_id == other.file_id) \
219
and (self.children == other.children)
220
221
222
223
class Inventory(object):
245
return cmp(self.file_id, other.file_id) \
246
or cmp(self.children, other.children)
247
248
249
250
class Inventory(XMLMixin):
224
251
"""Inventory of versioned files in a tree.
225
252
226
253
This describes which file_id is present at each point in the tree,
238
265
inserted, other than through the Inventory API.
239
266
240
267
>>> inv = Inventory()
268
>>> inv.write_xml(sys.stdout)
269
<inventory>
270
</inventory>
241
271
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
242
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT')
243
272
>>> inv['123-123'].name
244
273
'hello.c'
245
274
254
283
255
284
>>> [x[0] for x in inv.iter_entries()]
256
285
['hello.c']
257
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
258
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
259
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
286
287
>>> inv.write_xml(sys.stdout)
288
<inventory>
289
<entry file_id="123-123" kind="file" name="hello.c" />
290
</inventory>
291
260
292
"""
261
def __init__(self, root_id=ROOT_ID):
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):
262
310
"""Create or read an inventory.
263
311
264
312
If a working directory is specified, the inventory is read
268
316
The inventory is created with a default root directory, with
269
317
an id of None.
270
318
"""
271
# We are letting Branch(init=True) create a unique inventory
272
# root id. Rather than generating a random one here.
273
#if root_id is None:
274
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
275
self.root = RootEntry(root_id)
319
self.root = RootEntry(ROOT_ID)
276
320
self._byid = {self.root.file_id: self.root}
277
321
278
322
279
def copy(self):
280
other = Inventory(self.root.file_id)
281
# copy recursively so we know directories will be added before
282
# their children. There are more efficient ways than this...
283
for path, entry in self.iter_entries():
284
if entry == self.root:
285
continue
286
other.add(entry.copy())
287
return other
288
289
290
323
def __iter__(self):
291
324
return iter(self._byid)
292
325
311
344
if ie.kind == 'directory':
312
345
for cn, cie in self.iter_entries(from_dir=ie.file_id):
313
346
yield os.path.join(name, cn), cie
314
315
316
def entries(self):
317
"""Return list of (path, ie) for all entries except the root.
318
319
This may be faster than iter_entries.
347
348
349
350
def directories(self):
351
"""Return (path, entry) pairs for all directories.
320
352
"""
321
accum = []
322
def descend(dir_ie, dir_path):
323
kids = dir_ie.children.items()
324
kids.sort()
325
for name, ie in kids:
326
child_path = os.path.join(dir_path, name)
327
accum.append((child_path, ie))
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():
328
360
if ie.kind == 'directory':
329
descend(ie, child_path)
330
331
descend(self.root, '')
332
return accum
333
334
335
def directories(self):
336
"""Return (path, entry) pairs for all directories, including the root.
337
"""
338
accum = []
339
def descend(parent_ie, parent_path):
340
accum.append((parent_path, parent_ie))
361
dn.append((ie.name, ie))
362
dn.sort()
341
363
342
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
343
kids.sort()
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
344
367
345
for name, child_ie in kids:
346
child_path = os.path.join(parent_path, name)
347
descend(child_ie, child_path)
348
descend(self.root, '')
349
return accum
368
for name, ie in descend(self.root):
369
yield name, ie
350
370
351
371
352
372
355
375
356
376
>>> inv = Inventory()
357
377
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
358
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
359
378
>>> '123' in inv
360
379
True
361
380
>>> '456' in inv
369
388
370
389
>>> inv = Inventory()
371
390
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
372
InventoryEntry('123123', 'hello.c', kind='file', parent_id='TREE_ROOT')
373
391
>>> inv['123123'].name
374
392
'hello.c'
375
393
"""
394
if file_id == None:
395
raise BzrError("can't look up file_id None")
396
376
397
try:
377
398
return self._byid[file_id]
378
399
except KeyError:
379
if file_id == None:
380
raise BzrError("can't look up file_id None")
381
else:
382
raise BzrError("file_id {%s} not in inventory" % file_id)
383
384
385
def get_file_kind(self, file_id):
386
return self._byid[file_id].kind
400
raise BzrError("file_id {%s} not in inventory" % file_id)
401
387
402
388
403
def get_child(self, parent_id, filename):
389
404
return self[parent_id].children.get(filename)
393
408
"""Add entry to inventory.
394
409
395
410
To add a file to a branch ready to be committed, use Branch.add,
396
which calls this.
397
398
Returns the new entry object.
399
"""
411
which calls this."""
400
412
if entry.file_id in self._byid:
401
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
402
403
if entry.parent_id == ROOT_ID or entry.parent_id is None:
404
entry.parent_id = self.root.file_id
413
bailout("inventory already contains entry with id {%s}" % entry.file_id)
405
414
406
415
try:
407
416
parent = self._byid[entry.parent_id]
408
417
except KeyError:
409
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
418
bailout("parent_id {%s} not in inventory" % entry.parent_id)
410
419
411
420
if parent.children.has_key(entry.name):
412
raise BzrError("%s is already versioned" %
421
bailout("%s is already versioned" %
413
422
appendpath(self.id2path(parent.file_id), entry.name))
414
423
415
424
self._byid[entry.file_id] = entry
416
425
parent.children[entry.name] = entry
417
return entry
418
426
419
427
420
428
def add_path(self, relpath, kind, file_id=None):
421
429
"""Add entry from a path.
422
430
423
The immediate parent must already be versioned.
424
425
Returns the new entry object."""
426
from bzrlib.branch import gen_file_id
427
431
The immediate parent must already be versioned"""
428
432
parts = bzrlib.osutils.splitpath(relpath)
429
433
if len(parts) == 0:
430
raise BzrError("cannot re-add root of inventory")
434
bailout("cannot re-add root of inventory")
431
435
432
436
if file_id == None:
433
file_id = gen_file_id(relpath)
434
435
parent_path = parts[:-1]
436
parent_id = self.path2id(parent_path)
437
if parent_id == None:
438
raise NotVersionedError(parent_path)
439
437
file_id = bzrlib.branch.gen_file_id(relpath)
438
439
parent_id = self.path2id(parts[:-1])
440
assert parent_id != None
440
441
ie = InventoryEntry(file_id, parts[-1],
441
442
kind=kind, parent_id=parent_id)
442
443
return self.add(ie)
447
448
448
449
>>> inv = Inventory()
449
450
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
450
InventoryEntry('123', 'foo.c', kind='file', parent_id='TREE_ROOT')
451
451
>>> '123' in inv
452
452
True
453
453
>>> del inv['123']
469
469
del self[ie.parent_id].children[ie.name]
470
470
471
471
472
def __eq__(self, other):
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):
473
505
"""Compare two sets by comparing their contents.
474
506
475
507
>>> i1 = Inventory()
477
509
>>> i1 == i2
478
510
True
479
511
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
480
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
481
512
>>> i1 == i2
482
513
False
483
514
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
484
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
485
515
>>> i1 == i2
486
516
True
487
517
"""
518
if self is other:
519
return 0
520
488
521
if not isinstance(other, Inventory):
489
522
return NotImplemented
490
523
491
if len(self._byid) != len(other._byid):
492
# shortcut: obviously not the same
493
return False
494
495
return self._byid == other._byid
496
497
498
def __ne__(self, other):
499
return not (self == other)
500
501
502
def __hash__(self):
503
raise ValueError('not hashable')
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
504
532
505
533
506
534
def get_idpath(self, file_id):
516
544
try:
517
545
ie = self._byid[file_id]
518
546
except KeyError:
519
raise BzrError("file_id {%s} not found in inventory" % file_id)
547
bailout("file_id {%s} not found in inventory" % file_id)
520
548
p.insert(0, ie.file_id)
521
549
file_id = ie.parent_id
522
550
return p
526
554
"""Return as a list the path to file_id."""
527
555
528
556
# get all names, skipping root
529
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
557
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
530
558
return os.sep.join(p)
531
559
532
560
576
604
577
605
This does not move the working file."""
578
606
if not is_valid_name(new_name):
579
raise BzrError("not an acceptable filename: %r" % new_name)
607
bailout("not an acceptable filename: %r" % new_name)
580
608
581
609
new_parent = self._byid[new_parent_id]
582
610
if new_name in new_parent.children:
583
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
611
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
584
612
585
613
new_parent_idpath = self.get_idpath(new_parent_id)
586
614
if file_id in new_parent_idpath:
587
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
615
bailout("cannot move directory %r into a subdirectory of itself, %r"
588
616
% (self.id2path(file_id), self.id2path(new_parent_id)))
589
617
590
618
file_ie = self._byid[file_id]
601
629
602
630
603
631
604
_NAME_RE = None
632
_NAME_RE = re.compile(r'^[^/\\]+$')
605
633
606
634
def is_valid_name(name):
607
global _NAME_RE
608
if _NAME_RE == None:
609
_NAME_RE = re.compile(r'^[^/\\]+$')
610
611
635
return bool(_NAME_RE.match(name))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1