« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 1418
- compare with another revision
- download diff
- download tarball
- view history from revision 1418
- Committer: Robert Collins
- Date: 2005-10-06 12:14:01 UTC
- mfrom: (1393.1.67)
- Revision ID: robertc@robertcollins.net-20051006121401-ce87bcb93909bbdf
merge martins latest
- files added:
- HACKING
- bzrlib/plugins
- bzrlib/selftest
- bzrlib/store
- bzrlib/transport
- bzrlib/util
- bzrlib/util/effbot
- bzrlib/util/effbot/org
- contrib/emacs
- patches
- tools
- files removed:
- bzrlib/cache.py
- files renamed:
- bzrlib/store.py => bzrlib/store/__init__.py
- elementtree/ => bzrlib/util/elementtree/
- urlgrabber/ => bzrlib/util/urlgrabber/
- contrib/bash/bzr => contrib/bash/bzr.simple
- bzrlib/newinventory.py => contrib/newinventory.py
- 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
18
22
# This should really be an id randomly assigned when the tree is
19
23
# created, but it's not for now.
20
24
ROOT_ID = "TREE_ROOT"
21
25
22
26
23
import sys, os.path, types, re
24
from sets import Set
25
26
try:
27
from cElementTree import Element, ElementTree, SubElement
28
except ImportError:
29
from elementtree.ElementTree import Element, ElementTree, SubElement
30
31
from xml import XMLMixin
32
from errors import bailout, BzrError, BzrCheckError
27
import os.path
28
import re
29
import sys
30
import tarfile
31
import types
33
32
34
33
import bzrlib
35
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
34
from bzrlib.errors import BzrError, BzrCheckError
35
36
from bzrlib.osutils import (pumpfile, quotefn, splitpath, joinpath,
37
appendpath, sha_strings)
36
38
from bzrlib.trace import mutter
37
38
class InventoryEntry(XMLMixin):
39
from bzrlib.errors import NotVersionedError
40
41
42
class InventoryEntry(object):
39
43
"""Description of a versioned file.
40
44
41
45
An InventoryEntry has the following fields, which are also
42
46
present in the XML inventory-entry element:
43
47
44
* *file_id*
45
* *name*: (only the basename within the directory, must not
46
contain slashes)
47
* *kind*: "directory" or "file"
48
* *directory_id*: (if absent/null means the branch root directory)
49
* *text_sha1*: only for files
50
* *text_size*: in bytes, only for files
51
* *text_id*: identifier for the text version, only for files
52
53
InventoryEntries can also exist inside a WorkingTree
54
inventory, in which case they are not yet bound to a
55
particular revision of the file. In that case the text_sha1,
56
text_size and text_id are absent.
57
48
file_id
49
50
name
51
(within the parent directory)
52
53
parent_id
54
file_id of the parent directory, or ROOT_ID
55
56
revision
57
the revision_id in which this variation of this file was
58
introduced.
59
60
executable
61
Indicates that this file should be executable on systems
62
that support it.
63
64
text_sha1
65
sha-1 of the text of the file
66
67
text_size
68
size in bytes of the text of the file
69
70
(reading a version 4 tree created a text_id field.)
58
71
59
72
>>> i = Inventory()
60
73
>>> i.path2id('')
61
74
'TREE_ROOT'
62
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
63
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
75
>>> i.add(InventoryDirectory('123', 'src', ROOT_ID))
76
InventoryDirectory('123', 'src', parent_id='TREE_ROOT')
77
>>> i.add(InventoryFile('2323', 'hello.c', parent_id='123'))
78
InventoryFile('2323', 'hello.c', parent_id='123')
64
79
>>> for j in i.iter_entries():
65
80
... print j
66
81
...
67
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
68
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
69
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
82
('src', InventoryDirectory('123', 'src', parent_id='TREE_ROOT'))
83
('src/hello.c', InventoryFile('2323', 'hello.c', parent_id='123'))
84
>>> i.add(InventoryFile('2323', 'bye.c', '123'))
70
85
Traceback (most recent call last):
71
86
...
72
BzrError: ('inventory already contains entry with id {2323}', [])
73
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
74
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
87
BzrError: inventory already contains entry with id {2323}
88
>>> i.add(InventoryFile('2324', 'bye.c', '123'))
89
InventoryFile('2324', 'bye.c', parent_id='123')
90
>>> i.add(InventoryDirectory('2325', 'wibble', '123'))
91
InventoryDirectory('2325', 'wibble', parent_id='123')
75
92
>>> i.path2id('src/wibble')
76
93
'2325'
77
94
>>> '2325' in i
78
95
True
79
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
96
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
97
InventoryFile('2326', 'wibble.c', parent_id='2325')
80
98
>>> 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])
99
InventoryFile('2326', 'wibble.c', parent_id='2325')
100
>>> for path, entry in i.iter_entries():
101
... print path.replace('\\\\', '/') # for win32 os.sep
102
... assert i.path2id(path)
85
103
...
86
104
src
87
105
src/bye.c
88
106
src/hello.c
89
107
src/wibble
90
108
src/wibble/wibble.c
91
>>> i.id2path('2326')
109
>>> i.id2path('2326').replace('\\\\', '/')
92
110
'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
111
"""
98
99
# TODO: split InventoryEntry into subclasses for files,
100
# directories, etc etc.
101
102
text_sha1 = None
103
text_size = None
104
112
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
113
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
114
'text_id', 'parent_id', 'children', 'executable',
115
'revision']
116
117
def _add_text_to_weave(self, new_lines, parents, weave_store):
118
weave_store.add_text(self.file_id, self.revision, new_lines, parents)
119
120
def detect_changes(self, old_entry):
121
"""Return a (text_modified, meta_modified) from this to old_entry.
122
123
_read_tree_state must have been called on self and old_entry prior to
124
calling detect_changes.
125
"""
126
return False, False
127
128
def diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
129
output_to, reverse=False):
130
"""Perform a diff from this to to_entry.
131
132
text_diff will be used for textual difference calculation.
133
This is a template method, override _diff in child classes.
134
"""
135
self._read_tree_state(tree.id2path(self.file_id), tree)
136
if to_entry:
137
# cannot diff from one kind to another - you must do a removal
138
# and an addif they do not match.
139
assert self.kind == to_entry.kind
140
to_entry._read_tree_state(to_tree.id2path(to_entry.file_id),
141
to_tree)
142
self._diff(text_diff, from_label, tree, to_label, to_entry, to_tree,
143
output_to, reverse)
144
145
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
146
output_to, reverse=False):
147
"""Perform a diff between two entries of the same kind."""
148
149
def find_previous_heads(self, previous_inventories, entry_weave):
150
"""Return the revisions and entries that directly preceed this.
151
152
Returned as a map from revision to inventory entry.
153
154
This is a map containing the file revisions in all parents
155
for which the file exists, and its revision is not a parent of
156
any other. If the file is new, the set will be empty.
157
"""
158
def get_ancestors(weave, entry):
159
return set(map(weave.idx_to_name,
160
weave.inclusions([weave.lookup(entry.revision)])))
161
heads = {}
162
head_ancestors = {}
163
for inv in previous_inventories:
164
if self.file_id in inv:
165
ie = inv[self.file_id]
166
assert ie.file_id == self.file_id
167
if ie.revision in heads:
168
assert heads[ie.revision] == ie
169
else:
170
# may want to add it.
171
# may already be covered:
172
already_present = 0 != len(
173
[head for head in heads
174
if ie.revision in head_ancestors[head]])
175
if already_present:
176
# an ancestor of a known head.
177
continue
178
# definately a head:
179
ancestors = get_ancestors(entry_weave, ie)
180
# may knock something else out:
181
check_heads = list(heads.keys())
182
for head in check_heads:
183
if head in ancestors:
184
# this head is not really a head
185
heads.pop(head)
186
head_ancestors[ie.revision] = ancestors
187
heads[ie.revision] = ie
188
return heads
189
190
def get_tar_item(self, root, dp, now, tree):
191
"""Get a tarfile item and a file stream for its content."""
192
item = tarfile.TarInfo(os.path.join(root, dp))
193
# TODO: would be cool to actually set it to the timestamp of the
194
# revision it was last changed
195
item.mtime = now
196
fileobj = self._put_in_tar(item, tree)
197
return item, fileobj
198
199
def has_text(self):
200
"""Return true if the object this entry represents has textual data.
201
202
Note that textual data includes binary content.
203
204
Also note that all entries get weave files created for them.
205
This attribute is primarily used when upgrading from old trees that
206
did not have the weave index for all inventory entries.
207
"""
208
return False
209
210
def __init__(self, file_id, name, parent_id, text_id=None):
106
211
"""Create an InventoryEntry
107
212
108
213
The filename must be a single component, relative to the
109
214
parent directory; it cannot be a whole path or relative name.
110
215
111
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
216
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
112
217
>>> e.name
113
218
'hello.c'
114
219
>>> e.file_id
115
220
'123'
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
221
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
117
222
Traceback (most recent call last):
118
223
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
119
224
"""
225
assert isinstance(name, basestring), name
120
226
if '/' in name or '\\' in name:
121
227
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
122
228
229
self.executable = False
230
self.revision = None
231
self.text_sha1 = None
232
self.text_size = None
123
233
self.file_id = file_id
124
234
self.name = name
125
self.kind = kind
126
235
self.text_id = text_id
127
236
self.parent_id = parent_id
128
if kind == 'directory':
129
self.children = {}
130
elif kind == 'file':
131
pass
132
else:
133
raise BzrError("unhandled entry kind %r" % kind)
134
135
237
self.symlink_target = None
238
239
def kind_character(self):
240
"""Return a short kind indicator useful for appending to names."""
241
raise BzrError('unknown kind %r' % self.kind)
242
243
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
244
245
def _put_in_tar(self, item, tree):
246
"""populate item for stashing in a tar, and return the content stream.
247
248
If no content is available, return None.
249
"""
250
raise BzrError("don't know how to export {%s} of kind %r" %
251
(self.file_id, self.kind))
252
253
def put_on_disk(self, dest, dp, tree):
254
"""Create a representation of self on disk in the prefix dest.
255
256
This is a template method - implement _put_on_disk in subclasses.
257
"""
258
fullpath = appendpath(dest, dp)
259
self._put_on_disk(fullpath, tree)
260
mutter(" export {%s} kind %s to %s" % (self.file_id, self.kind, fullpath))
261
262
def _put_on_disk(self, fullpath, tree):
263
"""Put this entry onto disk at fullpath, from tree tree."""
264
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
136
265
137
266
def sorted_children(self):
138
267
l = self.children.items()
139
268
l.sort()
140
269
return l
141
270
271
@staticmethod
272
def versionable_kind(kind):
273
return kind in ('file', 'directory', 'symlink')
274
275
def check(self, checker, rev_id, inv, tree):
276
"""Check this inventory entry is intact.
277
278
This is a template method, override _check for kind specific
279
tests.
280
"""
281
if self.parent_id != None:
282
if not inv.has_id(self.parent_id):
283
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
284
% (self.parent_id, rev_id))
285
self._check(checker, rev_id, tree)
286
287
def _check(self, checker, rev_id, tree):
288
"""Check this inventory entry for kind specific errors."""
289
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
290
(self.kind, rev_id))
291
142
292
143
293
def copy(self):
144
other = InventoryEntry(self.file_id, self.name, self.kind,
145
self.parent_id, text_id=self.text_id)
146
other.text_sha1 = self.text_sha1
147
other.text_size = self.text_size
148
return other
294
"""Clone this inventory entry."""
295
raise NotImplementedError
149
296
297
def _get_snapshot_change(self, previous_entries):
298
if len(previous_entries) > 1:
299
return 'merged'
300
elif len(previous_entries) == 0:
301
return 'added'
302
else:
303
return 'modified/renamed/reparented'
150
304
151
305
def __repr__(self):
152
return ("%s(%r, %r, kind=%r, parent_id=%r)"
306
return ("%s(%r, %r, parent_id=%r)"
153
307
% (self.__class__.__name__,
154
308
self.file_id,
155
309
self.name,
156
self.kind,
157
310
self.parent_id))
158
311
159
160
def to_element(self):
161
"""Convert to XML element"""
162
e = Element('entry')
163
164
e.set('name', self.name)
165
e.set('file_id', self.file_id)
166
e.set('kind', self.kind)
167
168
if self.text_size != None:
169
e.set('text_size', '%d' % self.text_size)
170
171
for f in ['text_id', 'text_sha1']:
172
v = getattr(self, f)
173
if v != None:
174
e.set(f, v)
175
176
# to be conservative, we don't externalize the root pointers
177
# for now, leaving them as null in the xml form. in a future
178
# version it will be implied by nested elements.
179
if self.parent_id != ROOT_ID:
180
assert isinstance(self.parent_id, basestring)
181
e.set('parent_id', self.parent_id)
182
183
e.tail = '\n'
184
185
return e
186
187
188
def from_element(cls, elt):
189
assert elt.tag == 'entry'
190
191
## original format inventories don't have a parent_id for
192
## nodes in the root directory, but it's cleaner to use one
193
## internally.
194
parent_id = elt.get('parent_id')
195
if parent_id == None:
196
parent_id = ROOT_ID
197
198
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
199
self.text_id = elt.get('text_id')
200
self.text_sha1 = elt.get('text_sha1')
201
202
## mutter("read inventoryentry: %r" % (elt.attrib))
203
204
v = elt.get('text_size')
205
self.text_size = v and int(v)
206
207
return self
208
209
210
from_element = classmethod(from_element)
211
212
def __cmp__(self, other):
213
if self is other:
214
return 0
312
def snapshot(self, revision, path, previous_entries,
313
work_tree, weave_store):
314
"""Make a snapshot of this entry which may or may not have changed.
315
316
This means that all its fields are populated, that it has its
317
text stored in the text store or weave.
318
"""
319
mutter('new parents of %s are %r', path, previous_entries)
320
self._read_tree_state(path, work_tree)
321
if len(previous_entries) == 1:
322
# cannot be unchanged unless there is only one parent file rev.
323
parent_ie = previous_entries.values()[0]
324
if self._unchanged(parent_ie):
325
mutter("found unchanged entry")
326
self.revision = parent_ie.revision
327
return "unchanged"
328
return self.snapshot_revision(revision, previous_entries,
329
work_tree, weave_store)
330
331
def snapshot_revision(self, revision, previous_entries, work_tree,
332
weave_store):
333
"""Record this revision unconditionally."""
334
mutter('new revision for {%s}', self.file_id)
335
self.revision = revision
336
change = self._get_snapshot_change(previous_entries)
337
self._snapshot_text(previous_entries, work_tree, weave_store)
338
return change
339
340
def _snapshot_text(self, file_parents, work_tree, weave_store):
341
"""Record the 'text' of this entry, whatever form that takes.
342
343
This default implementation simply adds an empty text.
344
"""
345
mutter('storing file {%s} in revision {%s}',
346
self.file_id, self.revision)
347
self._add_text_to_weave([], file_parents, weave_store)
348
349
def __eq__(self, other):
215
350
if not isinstance(other, InventoryEntry):
216
351
return NotImplemented
217
352
218
return cmp(self.file_id, other.file_id) \
219
or cmp(self.name, other.name) \
220
or cmp(self.text_sha1, other.text_sha1) \
221
or cmp(self.text_size, other.text_size) \
222
or cmp(self.text_id, other.text_id) \
223
or cmp(self.parent_id, other.parent_id) \
224
or cmp(self.kind, other.kind)
225
353
return ((self.file_id == other.file_id)
354
and (self.name == other.name)
355
and (other.symlink_target == self.symlink_target)
356
and (self.text_sha1 == other.text_sha1)
357
and (self.text_size == other.text_size)
358
and (self.text_id == other.text_id)
359
and (self.parent_id == other.parent_id)
360
and (self.kind == other.kind)
361
and (self.revision == other.revision)
362
and (self.executable == other.executable)
363
)
364
365
def __ne__(self, other):
366
return not (self == other)
367
368
def __hash__(self):
369
raise ValueError('not hashable')
370
371
def _unchanged(self, previous_ie):
372
"""Has this entry changed relative to previous_ie.
373
374
This method should be overriden in child classes.
375
"""
376
compatible = True
377
# different inv parent
378
if previous_ie.parent_id != self.parent_id:
379
compatible = False
380
# renamed
381
elif previous_ie.name != self.name:
382
compatible = False
383
return compatible
384
385
def _read_tree_state(self, path, work_tree):
386
"""Populate fields in the inventory entry from the given tree.
387
388
Note that this should be modified to be a noop on virtual trees
389
as all entries created there are prepopulated.
390
"""
226
391
227
392
228
393
class RootEntry(InventoryEntry):
394
395
def _check(self, checker, rev_id, tree):
396
"""See InventoryEntry._check"""
397
229
398
def __init__(self, file_id):
230
399
self.file_id = file_id
231
400
self.children = {}
233
402
self.parent_id = None
234
403
self.name = ''
235
404
236
def __cmp__(self, other):
237
if self is other:
238
return 0
405
def __eq__(self, other):
239
406
if not isinstance(other, RootEntry):
240
407
return NotImplemented
241
return cmp(self.file_id, other.file_id) \
242
or cmp(self.children, other.children)
243
244
245
246
class Inventory(XMLMixin):
408
409
return (self.file_id == other.file_id) \
410
and (self.children == other.children)
411
412
413
class InventoryDirectory(InventoryEntry):
414
"""A directory in an inventory."""
415
416
def _check(self, checker, rev_id, tree):
417
"""See InventoryEntry._check"""
418
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
419
raise BzrCheckError('directory {%s} has text in revision {%s}'
420
% (self.file_id, rev_id))
421
422
def copy(self):
423
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
424
other.revision = self.revision
425
# note that children are *not* copied; they're pulled across when
426
# others are added
427
return other
428
429
def __init__(self, file_id, name, parent_id):
430
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
431
self.children = {}
432
self.kind = 'directory'
433
434
def kind_character(self):
435
"""See InventoryEntry.kind_character."""
436
return '/'
437
438
def _put_in_tar(self, item, tree):
439
"""See InventoryEntry._put_in_tar."""
440
item.type = tarfile.DIRTYPE
441
fileobj = None
442
item.name += '/'
443
item.size = 0
444
item.mode = 0755
445
return fileobj
446
447
def _put_on_disk(self, fullpath, tree):
448
"""See InventoryEntry._put_on_disk."""
449
os.mkdir(fullpath)
450
451
452
class InventoryFile(InventoryEntry):
453
"""A file in an inventory."""
454
455
def _check(self, checker, rev_id, tree):
456
"""See InventoryEntry._check"""
457
revision = self.revision
458
t = (self.file_id, revision)
459
if t in checker.checked_texts:
460
prev_sha = checker.checked_texts[t]
461
if prev_sha != self.text_sha1:
462
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
463
(self.file_id, rev_id))
464
else:
465
checker.repeated_text_cnt += 1
466
return
467
mutter('check version {%s} of {%s}', rev_id, self.file_id)
468
file_lines = tree.get_file_lines(self.file_id)
469
checker.checked_text_cnt += 1
470
if self.text_size != sum(map(len, file_lines)):
471
raise BzrCheckError('text {%s} wrong size' % self.text_id)
472
if self.text_sha1 != sha_strings(file_lines):
473
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
474
checker.checked_texts[t] = self.text_sha1
475
476
def copy(self):
477
other = InventoryFile(self.file_id, self.name, self.parent_id)
478
other.executable = self.executable
479
other.text_id = self.text_id
480
other.text_sha1 = self.text_sha1
481
other.text_size = self.text_size
482
other.revision = self.revision
483
return other
484
485
def detect_changes(self, old_entry):
486
"""See InventoryEntry.detect_changes."""
487
assert self.text_sha1 != None
488
assert old_entry.text_sha1 != None
489
text_modified = (self.text_sha1 != old_entry.text_sha1)
490
meta_modified = (self.executable != old_entry.executable)
491
return text_modified, meta_modified
492
493
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
494
output_to, reverse=False):
495
"""See InventoryEntry._diff."""
496
from_text = tree.get_file(self.file_id).readlines()
497
if to_entry:
498
to_text = to_tree.get_file(to_entry.file_id).readlines()
499
else:
500
to_text = []
501
if not reverse:
502
text_diff(from_label, from_text,
503
to_label, to_text, output_to)
504
else:
505
text_diff(to_label, to_text,
506
from_label, from_text, output_to)
507
508
def has_text(self):
509
"""See InventoryEntry.has_text."""
510
return True
511
512
def __init__(self, file_id, name, parent_id):
513
super(InventoryFile, self).__init__(file_id, name, parent_id)
514
self.kind = 'file'
515
516
def kind_character(self):
517
"""See InventoryEntry.kind_character."""
518
return ''
519
520
def _put_in_tar(self, item, tree):
521
"""See InventoryEntry._put_in_tar."""
522
item.type = tarfile.REGTYPE
523
fileobj = tree.get_file(self.file_id)
524
item.size = self.text_size
525
if tree.is_executable(self.file_id):
526
item.mode = 0755
527
else:
528
item.mode = 0644
529
return fileobj
530
531
def _put_on_disk(self, fullpath, tree):
532
"""See InventoryEntry._put_on_disk."""
533
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
534
if tree.is_executable(self.file_id):
535
os.chmod(fullpath, 0755)
536
537
def _read_tree_state(self, path, work_tree):
538
"""See InventoryEntry._read_tree_state."""
539
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
540
self.executable = work_tree.is_executable(self.file_id)
541
542
def _snapshot_text(self, file_parents, work_tree, weave_store):
543
"""See InventoryEntry._snapshot_text."""
544
mutter('storing file {%s} in revision {%s}',
545
self.file_id, self.revision)
546
# special case to avoid diffing on renames or
547
# reparenting
548
if (len(file_parents) == 1
549
and self.text_sha1 == file_parents.values()[0].text_sha1
550
and self.text_size == file_parents.values()[0].text_size):
551
previous_ie = file_parents.values()[0]
552
weave_store.add_identical_text(
553
self.file_id, previous_ie.revision,
554
self.revision, file_parents)
555
else:
556
new_lines = work_tree.get_file(self.file_id).readlines()
557
self._add_text_to_weave(new_lines, file_parents, weave_store)
558
self.text_sha1 = sha_strings(new_lines)
559
self.text_size = sum(map(len, new_lines))
560
561
562
def _unchanged(self, previous_ie):
563
"""See InventoryEntry._unchanged."""
564
compatible = super(InventoryFile, self)._unchanged(previous_ie)
565
if self.text_sha1 != previous_ie.text_sha1:
566
compatible = False
567
else:
568
# FIXME: 20050930 probe for the text size when getting sha1
569
# in _read_tree_state
570
self.text_size = previous_ie.text_size
571
return compatible
572
573
574
class InventoryLink(InventoryEntry):
575
"""A file in an inventory."""
576
577
__slots__ = ['symlink_target']
578
579
def _check(self, checker, rev_id, tree):
580
"""See InventoryEntry._check"""
581
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
582
raise BzrCheckError('symlink {%s} has text in revision {%s}'
583
% (self.file_id, rev_id))
584
if self.symlink_target == None:
585
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
586
% (self.file_id, rev_id))
587
588
def copy(self):
589
other = InventoryLink(self.file_id, self.name, self.parent_id)
590
other.symlink_target = self.symlink_target
591
other.revision = self.revision
592
return other
593
594
def detect_changes(self, old_entry):
595
"""See InventoryEntry.detect_changes."""
596
# FIXME: which _modified field should we use ? RBC 20051003
597
text_modified = (self.symlink_target != old_entry.symlink_target)
598
if text_modified:
599
mutter(" symlink target changed")
600
meta_modified = False
601
return text_modified, meta_modified
602
603
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
604
output_to, reverse=False):
605
"""See InventoryEntry._diff."""
606
from_text = self.symlink_target
607
if to_entry is not None:
608
to_text = to_entry.symlink_target
609
if reverse:
610
temp = from_text
611
from_text = to_text
612
to_text = temp
613
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
614
else:
615
if not reverse:
616
print >>output_to, '=== target was %r' % self.symlink_target
617
else:
618
print >>output_to, '=== target is %r' % self.symlink_target
619
620
def __init__(self, file_id, name, parent_id):
621
super(InventoryLink, self).__init__(file_id, name, parent_id)
622
self.kind = 'symlink'
623
624
def kind_character(self):
625
"""See InventoryEntry.kind_character."""
626
return ''
627
628
def _put_in_tar(self, item, tree):
629
"""See InventoryEntry._put_in_tar."""
630
iterm.type = tarfile.SYMTYPE
631
fileobj = None
632
item.size = 0
633
item.mode = 0755
634
item.linkname = self.symlink_target
635
return fileobj
636
637
def _put_on_disk(self, fullpath, tree):
638
"""See InventoryEntry._put_on_disk."""
639
try:
640
os.symlink(self.symlink_target, fullpath)
641
except OSError,e:
642
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
643
644
def _read_tree_state(self, path, work_tree):
645
"""See InventoryEntry._read_tree_state."""
646
self.symlink_target = work_tree.get_symlink_target(self.file_id)
647
648
def _unchanged(self, previous_ie):
649
"""See InventoryEntry._unchanged."""
650
compatible = super(InventoryLink, self)._unchanged(previous_ie)
651
if self.symlink_target != previous_ie.symlink_target:
652
compatible = False
653
return compatible
654
655
656
class Inventory(object):
247
657
"""Inventory of versioned files in a tree.
248
658
249
659
This describes which file_id is present at each point in the tree,
261
671
inserted, other than through the Inventory API.
262
672
263
673
>>> inv = Inventory()
264
>>> inv.write_xml(sys.stdout)
265
<inventory>
266
</inventory>
267
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
674
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
675
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT')
268
676
>>> inv['123-123'].name
269
677
'hello.c'
270
678
279
687
280
688
>>> [x[0] for x in inv.iter_entries()]
281
689
['hello.c']
282
283
>>> inv.write_xml(sys.stdout)
284
<inventory>
285
<entry file_id="123-123" kind="file" name="hello.c" />
286
</inventory>
287
690
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
691
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
692
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
288
693
"""
289
def __init__(self):
694
def __init__(self, root_id=ROOT_ID):
290
695
"""Create or read an inventory.
291
696
292
697
If a working directory is specified, the inventory is read
296
701
The inventory is created with a default root directory, with
297
702
an id of None.
298
703
"""
299
self.root = RootEntry(ROOT_ID)
704
# We are letting Branch.initialize() create a unique inventory
705
# root id. Rather than generating a random one here.
706
#if root_id is None:
707
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
708
self.root = RootEntry(root_id)
300
709
self._byid = {self.root.file_id: self.root}
301
710
302
711
712
def copy(self):
713
other = Inventory(self.root.file_id)
714
# copy recursively so we know directories will be added before
715
# their children. There are more efficient ways than this...
716
for path, entry in self.iter_entries():
717
if entry == self.root:
718
continue
719
other.add(entry.copy())
720
return other
721
722
303
723
def __iter__(self):
304
724
return iter(self._byid)
305
725
324
744
if ie.kind == 'directory':
325
745
for cn, cie in self.iter_entries(from_dir=ie.file_id):
326
746
yield os.path.join(name, cn), cie
327
747
748
749
def entries(self):
750
"""Return list of (path, ie) for all entries except the root.
751
752
This may be faster than iter_entries.
753
"""
754
accum = []
755
def descend(dir_ie, dir_path):
756
kids = dir_ie.children.items()
757
kids.sort()
758
for name, ie in kids:
759
child_path = os.path.join(dir_path, name)
760
accum.append((child_path, ie))
761
if ie.kind == 'directory':
762
descend(ie, child_path)
763
764
descend(self.root, '')
765
return accum
328
766
329
767
330
768
def directories(self):
331
"""Return (path, entry) pairs for all directories.
769
"""Return (path, entry) pairs for all directories, including the root.
332
770
"""
333
def descend(parent_ie):
334
parent_name = parent_ie.name
335
yield parent_name, parent_ie
336
337
# directory children in sorted order
338
dn = []
339
for ie in parent_ie.children.itervalues():
340
if ie.kind == 'directory':
341
dn.append((ie.name, ie))
342
dn.sort()
771
accum = []
772
def descend(parent_ie, parent_path):
773
accum.append((parent_path, parent_ie))
343
774
344
for name, child_ie in dn:
345
for sub_name, sub_ie in descend(child_ie):
346
yield appendpath(parent_name, sub_name), sub_ie
775
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
776
kids.sort()
347
777
348
for name, ie in descend(self.root):
349
yield name, ie
778
for name, child_ie in kids:
779
child_path = os.path.join(parent_path, name)
780
descend(child_ie, child_path)
781
descend(self.root, '')
782
return accum
350
783
351
784
352
785
354
787
"""True if this entry contains a file with given id.
355
788
356
789
>>> inv = Inventory()
357
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
790
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
791
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
358
792
>>> '123' in inv
359
793
True
360
794
>>> '456' in inv
367
801
"""Return the entry for given file_id.
368
802
369
803
>>> inv = Inventory()
370
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
804
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
805
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
371
806
>>> inv['123123'].name
372
807
'hello.c'
373
808
"""
380
815
raise BzrError("file_id {%s} not in inventory" % file_id)
381
816
382
817
818
def get_file_kind(self, file_id):
819
return self._byid[file_id].kind
820
383
821
def get_child(self, parent_id, filename):
384
822
return self[parent_id].children.get(filename)
385
823
388
826
"""Add entry to inventory.
389
827
390
828
To add a file to a branch ready to be committed, use Branch.add,
391
which calls this."""
829
which calls this.
830
831
Returns the new entry object.
832
"""
392
833
if entry.file_id in self._byid:
393
bailout("inventory already contains entry with id {%s}" % entry.file_id)
834
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
835
836
if entry.parent_id == ROOT_ID or entry.parent_id is None:
837
entry.parent_id = self.root.file_id
394
838
395
839
try:
396
840
parent = self._byid[entry.parent_id]
397
841
except KeyError:
398
bailout("parent_id {%s} not in inventory" % entry.parent_id)
842
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
399
843
400
844
if parent.children.has_key(entry.name):
401
bailout("%s is already versioned" %
845
raise BzrError("%s is already versioned" %
402
846
appendpath(self.id2path(parent.file_id), entry.name))
403
847
404
848
self._byid[entry.file_id] = entry
405
849
parent.children[entry.name] = entry
850
return entry
406
851
407
852
408
853
def add_path(self, relpath, kind, file_id=None):
409
854
"""Add entry from a path.
410
855
411
The immediate parent must already be versioned"""
856
The immediate parent must already be versioned.
857
858
Returns the new entry object."""
859
from bzrlib.branch import gen_file_id
860
412
861
parts = bzrlib.osutils.splitpath(relpath)
413
862
if len(parts) == 0:
414
bailout("cannot re-add root of inventory")
863
raise BzrError("cannot re-add root of inventory")
415
864
416
865
if file_id == None:
417
file_id = bzrlib.branch.gen_file_id(relpath)
418
419
parent_id = self.path2id(parts[:-1])
420
assert parent_id != None
421
ie = InventoryEntry(file_id, parts[-1],
422
kind=kind, parent_id=parent_id)
866
file_id = gen_file_id(relpath)
867
868
parent_path = parts[:-1]
869
parent_id = self.path2id(parent_path)
870
if parent_id == None:
871
raise NotVersionedError(parent_path)
872
873
if kind == 'directory':
874
ie = InventoryDirectory(file_id, parts[-1], parent_id)
875
elif kind == 'file':
876
ie = InventoryFile(file_id, parts[-1], parent_id)
877
elif kind == 'symlink':
878
ie = InventoryLink(file_id, parts[-1], parent_id)
879
else:
880
raise BzrError("unknown kind %r" % kind)
423
881
return self.add(ie)
424
882
425
883
427
885
"""Remove entry by id.
428
886
429
887
>>> inv = Inventory()
430
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
888
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
889
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
431
890
>>> '123' in inv
432
891
True
433
892
>>> del inv['123']
449
908
del self[ie.parent_id].children[ie.name]
450
909
451
910
452
def id_set(self):
453
return Set(self._byid)
454
455
456
def to_element(self):
457
"""Convert to XML Element"""
458
e = Element('inventory')
459
e.text = '\n'
460
for path, ie in self.iter_entries():
461
e.append(ie.to_element())
462
return e
463
464
465
def from_element(cls, elt):
466
"""Construct from XML Element
467
468
>>> inv = Inventory()
469
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
470
>>> elt = inv.to_element()
471
>>> inv2 = Inventory.from_element(elt)
472
>>> inv2 == inv
473
True
474
"""
475
assert elt.tag == 'inventory'
476
o = cls()
477
for e in elt:
478
o.add(InventoryEntry.from_element(e))
479
return o
480
481
from_element = classmethod(from_element)
482
483
484
def __cmp__(self, other):
911
def __eq__(self, other):
485
912
"""Compare two sets by comparing their contents.
486
913
487
914
>>> i1 = Inventory()
488
915
>>> i2 = Inventory()
489
916
>>> i1 == i2
490
917
True
491
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
918
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
919
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
492
920
>>> i1 == i2
493
921
False
494
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
922
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
923
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
495
924
>>> i1 == i2
496
925
True
497
926
"""
498
if self is other:
499
return 0
500
501
927
if not isinstance(other, Inventory):
502
928
return NotImplemented
503
929
504
if self.id_set() ^ other.id_set():
505
return 1
506
507
for file_id in self._byid:
508
c = cmp(self[file_id], other[file_id])
509
if c: return c
510
511
return 0
930
if len(self._byid) != len(other._byid):
931
# shortcut: obviously not the same
932
return False
933
934
return self._byid == other._byid
935
936
937
def __ne__(self, other):
938
return not self.__eq__(other)
939
940
941
def __hash__(self):
942
raise ValueError('not hashable')
512
943
513
944
514
945
def get_idpath(self, file_id):
524
955
try:
525
956
ie = self._byid[file_id]
526
957
except KeyError:
527
bailout("file_id {%s} not found in inventory" % file_id)
958
raise BzrError("file_id {%s} not found in inventory" % file_id)
528
959
p.insert(0, ie.file_id)
529
960
file_id = ie.parent_id
530
961
return p
531
962
532
963
533
964
def id2path(self, file_id):
534
"""Return as a list the path to file_id."""
535
965
"""Return as a list the path to file_id.
966
967
>>> i = Inventory()
968
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
969
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
970
>>> print i.id2path('foo-id').replace(os.sep, '/')
971
src/foo.c
972
"""
536
973
# get all names, skipping root
537
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
974
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
538
975
return os.sep.join(p)
539
976
540
977
584
1021
585
1022
This does not move the working file."""
586
1023
if not is_valid_name(new_name):
587
bailout("not an acceptable filename: %r" % new_name)
1024
raise BzrError("not an acceptable filename: %r" % new_name)
588
1025
589
1026
new_parent = self._byid[new_parent_id]
590
1027
if new_name in new_parent.children:
591
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
1028
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
592
1029
593
1030
new_parent_idpath = self.get_idpath(new_parent_id)
594
1031
if file_id in new_parent_idpath:
595
bailout("cannot move directory %r into a subdirectory of itself, %r"
1032
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
596
1033
% (self.id2path(file_id), self.id2path(new_parent_id)))
597
1034
598
1035
file_ie = self._byid[file_id]
609
1046
610
1047
611
1048
612
_NAME_RE = re.compile(r'^[^/\\]+$')
1049
_NAME_RE = None
613
1050
614
1051
def is_valid_name(name):
1052
global _NAME_RE
1053
if _NAME_RE == None:
1054
_NAME_RE = re.compile(r'^[^/\\]+$')
1055
615
1056
return bool(_NAME_RE.match(name))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1