« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 856
- compare with another revision
- download diff
- download tarball
- view history from revision 856
- Committer: Martin Pool
- Date: 2005-07-07 10:31:36 UTC
- Revision ID: mbp@sourcefrog.net-20050707103135-9b4d911d8df6e880
- fix pwk help
- files added:
- bzrlib/meta_store.py
- plugins/changeset
- files removed:
- HACKING
- bzrlib/store
- bzrlib/transport
- bzrlib/util
- contrib/emacs
- files renamed:
- contrib/newinventory.py => bzrlib/newinventory.py
- bzrlib/store/__init__.py => bzrlib/store.py
- bzrlib/util/effbot/ => effbot/
- bzrlib/util/elementtree/ => elementtree/
- bzrlib/plugins/ => plugins/
- bzrlib/selftest/test_weave.py => tools/testweave.py
- bzrlib/util/urlgrabber/ => urlgrabber/
- files modified:
- .bzrignore
- bzr *
added
removed
bzrlib/inventory.py
1
1
# (C) 2005 Canonical Ltd
2
#
2
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
5
5
# the Free Software Foundation; either version 2 of the License, or
6
6
# (at your option) any later version.
7
#
7
8
8
# This program is distributed in the hope that it will be useful,
9
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
11
# GNU General Public License for more details.
12
#
12
13
13
# You should have received a copy of the GNU General Public License
14
14
# along with this program; if not, write to the Free Software
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
18
# This should really be an id randomly assigned when the tree is
23
19
# created, but it's not for now.
24
20
ROOT_ID = "TREE_ROOT"
25
21
26
22
27
import os.path
28
import re
29
import sys
30
import tarfile
31
import types
23
import sys, os.path, types, re
32
24
33
25
import bzrlib
34
26
from bzrlib.errors import BzrError, BzrCheckError
35
27
36
from bzrlib.osutils import (pumpfile, quotefn, splitpath, joinpath,
37
appendpath, sha_strings)
28
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
38
29
from bzrlib.trace import mutter
39
from bzrlib.errors import NotVersionedError
40
41
30
42
31
class InventoryEntry(object):
43
32
"""Description of a versioned file.
45
34
An InventoryEntry has the following fields, which are also
46
35
present in the XML inventory-entry element:
47
36
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.)
37
* *file_id*
38
* *name*: (only the basename within the directory, must not
39
contain slashes)
40
* *kind*: "directory" or "file"
41
* *directory_id*: (if absent/null means the branch root directory)
42
* *text_sha1*: only for files
43
* *text_size*: in bytes, only for files
44
* *text_id*: identifier for the text version, only for files
45
46
InventoryEntries can also exist inside a WorkingTree
47
inventory, in which case they are not yet bound to a
48
particular revision of the file. In that case the text_sha1,
49
text_size and text_id are absent.
50
71
51
72
52
>>> i = Inventory()
73
53
>>> i.path2id('')
74
54
'TREE_ROOT'
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')
55
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
56
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
79
57
>>> for j in i.iter_entries():
80
58
... print j
81
59
...
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'))
60
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
61
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
62
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
85
63
Traceback (most recent call last):
86
64
...
87
65
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')
66
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
67
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
92
68
>>> i.path2id('src/wibble')
93
69
'2325'
94
70
>>> '2325' in i
95
71
True
96
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
97
InventoryFile('2326', 'wibble.c', parent_id='2325')
72
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
98
73
>>> i['2326']
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)
74
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
75
>>> for j in i.iter_entries():
76
... print j[0]
77
... assert i.path2id(j[0])
103
78
...
104
79
src
105
80
src/bye.c
106
81
src/hello.c
107
82
src/wibble
108
83
src/wibble/wibble.c
109
>>> i.id2path('2326').replace('\\\\', '/')
84
>>> i.id2path('2326')
110
85
'src/wibble/wibble.c'
86
87
TODO: Maybe also keep the full path of the entry, and the children?
88
But those depend on its position within a particular inventory, and
89
it would be nice not to need to hold the backpointer here.
111
90
"""
91
92
# TODO: split InventoryEntry into subclasses for files,
93
# directories, etc etc.
94
95
text_sha1 = None
96
text_size = None
112
97
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, transaction):
118
weave_store.add_text(self.file_id, self.revision, new_lines, parents,
119
transaction)
120
121
def detect_changes(self, old_entry):
122
"""Return a (text_modified, meta_modified) from this to old_entry.
123
124
_read_tree_state must have been called on self and old_entry prior to
125
calling detect_changes.
126
"""
127
return False, False
128
129
def diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
130
output_to, reverse=False):
131
"""Perform a diff from this to to_entry.
132
133
text_diff will be used for textual difference calculation.
134
This is a template method, override _diff in child classes.
135
"""
136
self._read_tree_state(tree.id2path(self.file_id), tree)
137
if to_entry:
138
# cannot diff from one kind to another - you must do a removal
139
# and an addif they do not match.
140
assert self.kind == to_entry.kind
141
to_entry._read_tree_state(to_tree.id2path(to_entry.file_id),
142
to_tree)
143
self._diff(text_diff, from_label, tree, to_label, to_entry, to_tree,
144
output_to, reverse)
145
146
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
147
output_to, reverse=False):
148
"""Perform a diff between two entries of the same kind."""
149
150
def find_previous_heads(self, previous_inventories, entry_weave):
151
"""Return the revisions and entries that directly preceed this.
152
153
Returned as a map from revision to inventory entry.
154
155
This is a map containing the file revisions in all parents
156
for which the file exists, and its revision is not a parent of
157
any other. If the file is new, the set will be empty.
158
"""
159
def get_ancestors(weave, entry):
160
return set(map(weave.idx_to_name,
161
weave.inclusions([weave.lookup(entry.revision)])))
162
heads = {}
163
head_ancestors = {}
164
for inv in previous_inventories:
165
if self.file_id in inv:
166
ie = inv[self.file_id]
167
assert ie.file_id == self.file_id
168
if ie.revision in heads:
169
# fixup logic, there was a bug in revision updates.
170
# with x bit support.
171
try:
172
if heads[ie.revision].executable != ie.executable:
173
heads[ie.revision].executable = False
174
ie.executable = False
175
except AttributeError:
176
pass
177
assert heads[ie.revision] == ie
178
else:
179
# may want to add it.
180
# may already be covered:
181
already_present = 0 != len(
182
[head for head in heads
183
if ie.revision in head_ancestors[head]])
184
if already_present:
185
# an ancestor of a known head.
186
continue
187
# definately a head:
188
ancestors = get_ancestors(entry_weave, ie)
189
# may knock something else out:
190
check_heads = list(heads.keys())
191
for head in check_heads:
192
if head in ancestors:
193
# this head is not really a head
194
heads.pop(head)
195
head_ancestors[ie.revision] = ancestors
196
heads[ie.revision] = ie
197
return heads
198
199
def get_tar_item(self, root, dp, now, tree):
200
"""Get a tarfile item and a file stream for its content."""
201
item = tarfile.TarInfo(os.path.join(root, dp))
202
# TODO: would be cool to actually set it to the timestamp of the
203
# revision it was last changed
204
item.mtime = now
205
fileobj = self._put_in_tar(item, tree)
206
return item, fileobj
207
208
def has_text(self):
209
"""Return true if the object this entry represents has textual data.
210
211
Note that textual data includes binary content.
212
213
Also note that all entries get weave files created for them.
214
This attribute is primarily used when upgrading from old trees that
215
did not have the weave index for all inventory entries.
216
"""
217
return False
218
219
def __init__(self, file_id, name, parent_id, text_id=None):
98
def __init__(self, file_id, name, kind, parent_id, text_id=None):
220
99
"""Create an InventoryEntry
221
100
222
101
The filename must be a single component, relative to the
223
102
parent directory; it cannot be a whole path or relative name.
224
103
225
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
104
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
226
105
>>> e.name
227
106
'hello.c'
228
107
>>> e.file_id
229
108
'123'
230
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
109
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
231
110
Traceback (most recent call last):
232
111
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
233
112
"""
234
assert isinstance(name, basestring), name
235
113
if '/' in name or '\\' in name:
236
114
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
237
115
238
self.executable = False
239
self.revision = None
240
self.text_sha1 = None
241
self.text_size = None
242
116
self.file_id = file_id
243
117
self.name = name
118
self.kind = kind
244
119
self.text_id = text_id
245
120
self.parent_id = parent_id
246
self.symlink_target = None
247
248
def kind_character(self):
249
"""Return a short kind indicator useful for appending to names."""
250
raise BzrError('unknown kind %r' % self.kind)
251
252
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
253
254
def _put_in_tar(self, item, tree):
255
"""populate item for stashing in a tar, and return the content stream.
256
257
If no content is available, return None.
258
"""
259
raise BzrError("don't know how to export {%s} of kind %r" %
260
(self.file_id, self.kind))
261
262
def put_on_disk(self, dest, dp, tree):
263
"""Create a representation of self on disk in the prefix dest.
264
265
This is a template method - implement _put_on_disk in subclasses.
266
"""
267
fullpath = appendpath(dest, dp)
268
self._put_on_disk(fullpath, tree)
269
mutter(" export {%s} kind %s to %s" % (self.file_id, self.kind, fullpath))
270
271
def _put_on_disk(self, fullpath, tree):
272
"""Put this entry onto disk at fullpath, from tree tree."""
273
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
121
if kind == 'directory':
122
self.children = {}
123
elif kind == 'file':
124
pass
125
else:
126
raise BzrError("unhandled entry kind %r" % kind)
127
128
274
129
275
130
def sorted_children(self):
276
131
l = self.children.items()
277
132
l.sort()
278
133
return l
279
134
280
@staticmethod
281
def versionable_kind(kind):
282
return kind in ('file', 'directory', 'symlink')
283
284
def check(self, checker, rev_id, inv, tree):
285
"""Check this inventory entry is intact.
286
287
This is a template method, override _check for kind specific
288
tests.
289
"""
290
if self.parent_id != None:
291
if not inv.has_id(self.parent_id):
292
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
293
% (self.parent_id, rev_id))
294
self._check(checker, rev_id, tree)
295
296
def _check(self, checker, rev_id, tree):
297
"""Check this inventory entry for kind specific errors."""
298
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
299
(self.kind, rev_id))
300
301
135
302
136
def copy(self):
303
"""Clone this inventory entry."""
304
raise NotImplementedError
137
other = InventoryEntry(self.file_id, self.name, self.kind,
138
self.parent_id, text_id=self.text_id)
139
other.text_sha1 = self.text_sha1
140
other.text_size = self.text_size
141
# note that children are *not* copied; they're pulled across when
142
# others are added
143
return other
305
144
306
def _get_snapshot_change(self, previous_entries):
307
if len(previous_entries) > 1:
308
return 'merged'
309
elif len(previous_entries) == 0:
310
return 'added'
311
else:
312
return 'modified/renamed/reparented'
313
145
314
146
def __repr__(self):
315
return ("%s(%r, %r, parent_id=%r)"
147
return ("%s(%r, %r, kind=%r, parent_id=%r)"
316
148
% (self.__class__.__name__,
317
149
self.file_id,
318
150
self.name,
151
self.kind,
319
152
self.parent_id))
320
153
321
def snapshot(self, revision, path, previous_entries,
322
work_tree, weave_store, transaction):
323
"""Make a snapshot of this entry which may or may not have changed.
324
325
This means that all its fields are populated, that it has its
326
text stored in the text store or weave.
327
"""
328
mutter('new parents of %s are %r', path, previous_entries)
329
self._read_tree_state(path, work_tree)
330
if len(previous_entries) == 1:
331
# cannot be unchanged unless there is only one parent file rev.
332
parent_ie = previous_entries.values()[0]
333
if self._unchanged(parent_ie):
334
mutter("found unchanged entry")
335
self.revision = parent_ie.revision
336
return "unchanged"
337
return self.snapshot_revision(revision, previous_entries,
338
work_tree, weave_store, transaction)
339
340
def snapshot_revision(self, revision, previous_entries, work_tree,
341
weave_store, transaction):
342
"""Record this revision unconditionally."""
343
mutter('new revision for {%s}', self.file_id)
344
self.revision = revision
345
change = self._get_snapshot_change(previous_entries)
346
self._snapshot_text(previous_entries, work_tree, weave_store,
347
transaction)
348
return change
349
350
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
351
"""Record the 'text' of this entry, whatever form that takes.
352
353
This default implementation simply adds an empty text.
354
"""
355
mutter('storing file {%s} in revision {%s}',
356
self.file_id, self.revision)
357
self._add_text_to_weave([], file_parents, weave_store, transaction)
154
155
def to_element(self):
156
"""Convert to XML element"""
157
from bzrlib.xml import Element
158
159
e = Element('entry')
160
161
e.set('name', self.name)
162
e.set('file_id', self.file_id)
163
e.set('kind', self.kind)
164
165
if self.text_size != None:
166
e.set('text_size', '%d' % self.text_size)
167
168
for f in ['text_id', 'text_sha1']:
169
v = getattr(self, f)
170
if v != None:
171
e.set(f, v)
172
173
# to be conservative, we don't externalize the root pointers
174
# for now, leaving them as null in the xml form. in a future
175
# version it will be implied by nested elements.
176
if self.parent_id != ROOT_ID:
177
assert isinstance(self.parent_id, basestring)
178
e.set('parent_id', self.parent_id)
179
180
e.tail = '\n'
181
182
return e
183
184
185
def from_element(cls, elt):
186
assert elt.tag == 'entry'
187
188
## original format inventories don't have a parent_id for
189
## nodes in the root directory, but it's cleaner to use one
190
## internally.
191
parent_id = elt.get('parent_id')
192
if parent_id == None:
193
parent_id = ROOT_ID
194
195
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
196
self.text_id = elt.get('text_id')
197
self.text_sha1 = elt.get('text_sha1')
198
199
## mutter("read inventoryentry: %r" % (elt.attrib))
200
201
v = elt.get('text_size')
202
self.text_size = v and int(v)
203
204
return self
205
206
207
from_element = classmethod(from_element)
358
208
359
209
def __eq__(self, other):
360
210
if not isinstance(other, InventoryEntry):
361
211
return NotImplemented
362
212
363
return ((self.file_id == other.file_id)
364
and (self.name == other.name)
365
and (other.symlink_target == self.symlink_target)
366
and (self.text_sha1 == other.text_sha1)
367
and (self.text_size == other.text_size)
368
and (self.text_id == other.text_id)
369
and (self.parent_id == other.parent_id)
370
and (self.kind == other.kind)
371
and (self.revision == other.revision)
372
and (self.executable == other.executable)
373
)
213
return (self.file_id == other.file_id) \
214
and (self.name == other.name) \
215
and (self.text_sha1 == other.text_sha1) \
216
and (self.text_size == other.text_size) \
217
and (self.text_id == other.text_id) \
218
and (self.parent_id == other.parent_id) \
219
and (self.kind == other.kind)
220
374
221
375
222
def __ne__(self, other):
376
223
return not (self == other)
378
225
def __hash__(self):
379
226
raise ValueError('not hashable')
380
227
381
def _unchanged(self, previous_ie):
382
"""Has this entry changed relative to previous_ie.
383
384
This method should be overriden in child classes.
385
"""
386
compatible = True
387
# different inv parent
388
if previous_ie.parent_id != self.parent_id:
389
compatible = False
390
# renamed
391
elif previous_ie.name != self.name:
392
compatible = False
393
return compatible
394
395
def _read_tree_state(self, path, work_tree):
396
"""Populate fields in the inventory entry from the given tree.
397
398
Note that this should be modified to be a noop on virtual trees
399
as all entries created there are prepopulated.
400
"""
401
# TODO: Rather than running this manually, we should check the
402
# working sha1 and other expensive properties when they're
403
# first requested, or preload them if they're already known
404
pass # nothing to do by default
405
228
406
229
407
230
class RootEntry(InventoryEntry):
408
409
def _check(self, checker, rev_id, tree):
410
"""See InventoryEntry._check"""
411
412
231
def __init__(self, file_id):
413
232
self.file_id = file_id
414
233
self.children = {}
424
243
and (self.children == other.children)
425
244
426
245
427
class InventoryDirectory(InventoryEntry):
428
"""A directory in an inventory."""
429
430
def _check(self, checker, rev_id, tree):
431
"""See InventoryEntry._check"""
432
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
433
raise BzrCheckError('directory {%s} has text in revision {%s}'
434
% (self.file_id, rev_id))
435
436
def copy(self):
437
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
438
other.revision = self.revision
439
# note that children are *not* copied; they're pulled across when
440
# others are added
441
return other
442
443
def __init__(self, file_id, name, parent_id):
444
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
445
self.children = {}
446
self.kind = 'directory'
447
448
def kind_character(self):
449
"""See InventoryEntry.kind_character."""
450
return '/'
451
452
def _put_in_tar(self, item, tree):
453
"""See InventoryEntry._put_in_tar."""
454
item.type = tarfile.DIRTYPE
455
fileobj = None
456
item.name += '/'
457
item.size = 0
458
item.mode = 0755
459
return fileobj
460
461
def _put_on_disk(self, fullpath, tree):
462
"""See InventoryEntry._put_on_disk."""
463
os.mkdir(fullpath)
464
465
466
class InventoryFile(InventoryEntry):
467
"""A file in an inventory."""
468
469
def _check(self, checker, rev_id, tree):
470
"""See InventoryEntry._check"""
471
revision = self.revision
472
t = (self.file_id, revision)
473
if t in checker.checked_texts:
474
prev_sha = checker.checked_texts[t]
475
if prev_sha != self.text_sha1:
476
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
477
(self.file_id, rev_id))
478
else:
479
checker.repeated_text_cnt += 1
480
return
481
mutter('check version {%s} of {%s}', rev_id, self.file_id)
482
file_lines = tree.get_file_lines(self.file_id)
483
checker.checked_text_cnt += 1
484
if self.text_size != sum(map(len, file_lines)):
485
raise BzrCheckError('text {%s} wrong size' % self.text_id)
486
if self.text_sha1 != sha_strings(file_lines):
487
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
488
checker.checked_texts[t] = self.text_sha1
489
490
def copy(self):
491
other = InventoryFile(self.file_id, self.name, self.parent_id)
492
other.executable = self.executable
493
other.text_id = self.text_id
494
other.text_sha1 = self.text_sha1
495
other.text_size = self.text_size
496
other.revision = self.revision
497
return other
498
499
def detect_changes(self, old_entry):
500
"""See InventoryEntry.detect_changes."""
501
assert self.text_sha1 != None
502
assert old_entry.text_sha1 != None
503
text_modified = (self.text_sha1 != old_entry.text_sha1)
504
meta_modified = (self.executable != old_entry.executable)
505
return text_modified, meta_modified
506
507
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
508
output_to, reverse=False):
509
"""See InventoryEntry._diff."""
510
from_text = tree.get_file(self.file_id).readlines()
511
if to_entry:
512
to_text = to_tree.get_file(to_entry.file_id).readlines()
513
else:
514
to_text = []
515
if not reverse:
516
text_diff(from_label, from_text,
517
to_label, to_text, output_to)
518
else:
519
text_diff(to_label, to_text,
520
from_label, from_text, output_to)
521
522
def has_text(self):
523
"""See InventoryEntry.has_text."""
524
return True
525
526
def __init__(self, file_id, name, parent_id):
527
super(InventoryFile, self).__init__(file_id, name, parent_id)
528
self.kind = 'file'
529
530
def kind_character(self):
531
"""See InventoryEntry.kind_character."""
532
return ''
533
534
def _put_in_tar(self, item, tree):
535
"""See InventoryEntry._put_in_tar."""
536
item.type = tarfile.REGTYPE
537
fileobj = tree.get_file(self.file_id)
538
item.size = self.text_size
539
if tree.is_executable(self.file_id):
540
item.mode = 0755
541
else:
542
item.mode = 0644
543
return fileobj
544
545
def _put_on_disk(self, fullpath, tree):
546
"""See InventoryEntry._put_on_disk."""
547
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
548
if tree.is_executable(self.file_id):
549
os.chmod(fullpath, 0755)
550
551
def _read_tree_state(self, path, work_tree):
552
"""See InventoryEntry._read_tree_state."""
553
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
554
self.executable = work_tree.is_executable(self.file_id)
555
556
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
557
"""See InventoryEntry._snapshot_text."""
558
mutter('storing file {%s} in revision {%s}',
559
self.file_id, self.revision)
560
# special case to avoid diffing on renames or
561
# reparenting
562
if (len(file_parents) == 1
563
and self.text_sha1 == file_parents.values()[0].text_sha1
564
and self.text_size == file_parents.values()[0].text_size):
565
previous_ie = file_parents.values()[0]
566
weave_store.add_identical_text(
567
self.file_id, previous_ie.revision,
568
self.revision, file_parents, transaction)
569
else:
570
new_lines = work_tree.get_file(self.file_id).readlines()
571
self._add_text_to_weave(new_lines, file_parents, weave_store,
572
transaction)
573
self.text_sha1 = sha_strings(new_lines)
574
self.text_size = sum(map(len, new_lines))
575
576
577
def _unchanged(self, previous_ie):
578
"""See InventoryEntry._unchanged."""
579
compatible = super(InventoryFile, self)._unchanged(previous_ie)
580
if self.text_sha1 != previous_ie.text_sha1:
581
compatible = False
582
else:
583
# FIXME: 20050930 probe for the text size when getting sha1
584
# in _read_tree_state
585
self.text_size = previous_ie.text_size
586
if self.executable != previous_ie.executable:
587
compatible = False
588
return compatible
589
590
591
class InventoryLink(InventoryEntry):
592
"""A file in an inventory."""
593
594
__slots__ = ['symlink_target']
595
596
def _check(self, checker, rev_id, tree):
597
"""See InventoryEntry._check"""
598
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
599
raise BzrCheckError('symlink {%s} has text in revision {%s}'
600
% (self.file_id, rev_id))
601
if self.symlink_target == None:
602
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
603
% (self.file_id, rev_id))
604
605
def copy(self):
606
other = InventoryLink(self.file_id, self.name, self.parent_id)
607
other.symlink_target = self.symlink_target
608
other.revision = self.revision
609
return other
610
611
def detect_changes(self, old_entry):
612
"""See InventoryEntry.detect_changes."""
613
# FIXME: which _modified field should we use ? RBC 20051003
614
text_modified = (self.symlink_target != old_entry.symlink_target)
615
if text_modified:
616
mutter(" symlink target changed")
617
meta_modified = False
618
return text_modified, meta_modified
619
620
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
621
output_to, reverse=False):
622
"""See InventoryEntry._diff."""
623
from_text = self.symlink_target
624
if to_entry is not None:
625
to_text = to_entry.symlink_target
626
if reverse:
627
temp = from_text
628
from_text = to_text
629
to_text = temp
630
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
631
else:
632
if not reverse:
633
print >>output_to, '=== target was %r' % self.symlink_target
634
else:
635
print >>output_to, '=== target is %r' % self.symlink_target
636
637
def __init__(self, file_id, name, parent_id):
638
super(InventoryLink, self).__init__(file_id, name, parent_id)
639
self.kind = 'symlink'
640
641
def kind_character(self):
642
"""See InventoryEntry.kind_character."""
643
return ''
644
645
def _put_in_tar(self, item, tree):
646
"""See InventoryEntry._put_in_tar."""
647
iterm.type = tarfile.SYMTYPE
648
fileobj = None
649
item.size = 0
650
item.mode = 0755
651
item.linkname = self.symlink_target
652
return fileobj
653
654
def _put_on_disk(self, fullpath, tree):
655
"""See InventoryEntry._put_on_disk."""
656
try:
657
os.symlink(self.symlink_target, fullpath)
658
except OSError,e:
659
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
660
661
def _read_tree_state(self, path, work_tree):
662
"""See InventoryEntry._read_tree_state."""
663
self.symlink_target = work_tree.get_symlink_target(self.file_id)
664
665
def _unchanged(self, previous_ie):
666
"""See InventoryEntry._unchanged."""
667
compatible = super(InventoryLink, self)._unchanged(previous_ie)
668
if self.symlink_target != previous_ie.symlink_target:
669
compatible = False
670
return compatible
671
672
246
673
247
class Inventory(object):
674
248
"""Inventory of versioned files in a tree.
688
262
inserted, other than through the Inventory API.
689
263
690
264
>>> inv = Inventory()
691
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
692
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT')
265
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
693
266
>>> inv['123-123'].name
694
267
'hello.c'
695
268
704
277
705
278
>>> [x[0] for x in inv.iter_entries()]
706
279
['hello.c']
707
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
708
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
709
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
710
280
"""
711
def __init__(self, root_id=ROOT_ID):
281
def __init__(self):
712
282
"""Create or read an inventory.
713
283
714
284
If a working directory is specified, the inventory is read
718
288
The inventory is created with a default root directory, with
719
289
an id of None.
720
290
"""
721
# We are letting Branch.initialize() create a unique inventory
722
# root id. Rather than generating a random one here.
723
#if root_id is None:
724
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
725
self.root = RootEntry(root_id)
291
self.root = RootEntry(ROOT_ID)
726
292
self._byid = {self.root.file_id: self.root}
727
293
728
294
729
def copy(self):
730
other = Inventory(self.root.file_id)
731
# copy recursively so we know directories will be added before
732
# their children. There are more efficient ways than this...
733
for path, entry in self.iter_entries():
734
if entry == self.root:
735
continue
736
other.add(entry.copy())
737
return other
738
739
740
295
def __iter__(self):
741
296
return iter(self._byid)
742
297
804
359
"""True if this entry contains a file with given id.
805
360
806
361
>>> inv = Inventory()
807
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
808
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
362
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
809
363
>>> '123' in inv
810
364
True
811
365
>>> '456' in inv
818
372
"""Return the entry for given file_id.
819
373
820
374
>>> inv = Inventory()
821
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
822
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
375
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
823
376
>>> inv['123123'].name
824
377
'hello.c'
825
378
"""
843
396
"""Add entry to inventory.
844
397
845
398
To add a file to a branch ready to be committed, use Branch.add,
846
which calls this.
847
848
Returns the new entry object.
849
"""
399
which calls this."""
850
400
if entry.file_id in self._byid:
851
401
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
852
402
853
if entry.parent_id == ROOT_ID or entry.parent_id is None:
854
entry.parent_id = self.root.file_id
855
856
403
try:
857
404
parent = self._byid[entry.parent_id]
858
405
except KeyError:
864
411
865
412
self._byid[entry.file_id] = entry
866
413
parent.children[entry.name] = entry
867
return entry
868
414
869
415
870
416
def add_path(self, relpath, kind, file_id=None):
871
417
"""Add entry from a path.
872
418
873
The immediate parent must already be versioned.
874
875
Returns the new entry object."""
876
from bzrlib.branch import gen_file_id
419
The immediate parent must already be versioned"""
420
from bzrlib.errors import NotVersionedError
877
421
878
422
parts = bzrlib.osutils.splitpath(relpath)
879
423
if len(parts) == 0:
880
424
raise BzrError("cannot re-add root of inventory")
881
425
882
426
if file_id == None:
427
from bzrlib.branch import gen_file_id
883
428
file_id = gen_file_id(relpath)
884
429
885
430
parent_path = parts[:-1]
887
432
if parent_id == None:
888
433
raise NotVersionedError(parent_path)
889
434
890
if kind == 'directory':
891
ie = InventoryDirectory(file_id, parts[-1], parent_id)
892
elif kind == 'file':
893
ie = InventoryFile(file_id, parts[-1], parent_id)
894
elif kind == 'symlink':
895
ie = InventoryLink(file_id, parts[-1], parent_id)
896
else:
897
raise BzrError("unknown kind %r" % kind)
435
ie = InventoryEntry(file_id, parts[-1],
436
kind=kind, parent_id=parent_id)
898
437
return self.add(ie)
899
438
900
439
902
441
"""Remove entry by id.
903
442
904
443
>>> inv = Inventory()
905
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
906
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
444
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
907
445
>>> '123' in inv
908
446
True
909
447
>>> del inv['123']
925
463
del self[ie.parent_id].children[ie.name]
926
464
927
465
466
def to_element(self):
467
"""Convert to XML Element"""
468
from bzrlib.xml import Element
469
470
e = Element('inventory')
471
e.text = '\n'
472
for path, ie in self.iter_entries():
473
e.append(ie.to_element())
474
return e
475
476
477
def from_element(cls, elt):
478
"""Construct from XML Element
479
480
>>> inv = Inventory()
481
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
482
>>> elt = inv.to_element()
483
>>> inv2 = Inventory.from_element(elt)
484
>>> inv2 == inv
485
True
486
"""
487
# XXXX: doctest doesn't run this properly under python2.3
488
assert elt.tag == 'inventory'
489
o = cls()
490
for e in elt:
491
o.add(InventoryEntry.from_element(e))
492
return o
493
494
from_element = classmethod(from_element)
495
496
928
497
def __eq__(self, other):
929
498
"""Compare two sets by comparing their contents.
930
499
932
501
>>> i2 = Inventory()
933
502
>>> i1 == i2
934
503
True
935
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
936
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
504
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
937
505
>>> i1 == i2
938
506
False
939
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
940
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
507
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
941
508
>>> i1 == i2
942
509
True
943
510
"""
952
519
953
520
954
521
def __ne__(self, other):
955
return not self.__eq__(other)
522
return not (self == other)
956
523
957
524
958
525
def __hash__(self):
959
526
raise ValueError('not hashable')
960
527
961
528
529
962
530
def get_idpath(self, file_id):
963
531
"""Return a list of file_ids for the path to an entry.
964
532
979
547
980
548
981
549
def id2path(self, file_id):
982
"""Return as a list the path to file_id.
983
984
>>> i = Inventory()
985
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
986
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
987
>>> print i.id2path('foo-id').replace(os.sep, '/')
988
src/foo.c
989
"""
550
"""Return as a list the path to file_id."""
551
990
552
# get all names, skipping root
991
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
553
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
992
554
return os.sep.join(p)
993
555
994
556
1063
625
1064
626
1065
627
1066
_NAME_RE = None
628
_NAME_RE = re.compile(r'^[^/\\]+$')
1067
629
1068
630
def is_valid_name(name):
1069
global _NAME_RE
1070
if _NAME_RE == None:
1071
_NAME_RE = re.compile(r'^[^/\\]+$')
1072
1073
631
return bool(_NAME_RE.match(name))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1