« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 1
- compare with another revision
- download diff
- download tarball
- view history from revision 1
- Committer: mbp at sourcefrog
- Date: 2005-03-09 04:08:15 UTC
- Revision ID: mbp@sourcefrog.net-20050309040815-13242001617e4a06
import from baz patch-364
- files added:
- bzrlib/tests.py
- files removed:
- .bzrignore
- bzrlib/export
- bzrlib/plugins
- bzrlib/store
- bzrlib/tests
- bzrlib/tests/blackbox
- bzrlib/transport
- bzrlib/ui
- bzrlib/util
- bzrlib/util/configobj
- bzrlib/util/configobj/docs
- bzrlib/util/effbot
- bzrlib/util/effbot/org
- bzrlib/util/elementtree
- bzrlib/util/urlgrabber
- contrib
- contrib/bash
- contrib/emacs
- contrib/zsh
- doc
- notes
- patches
- tools
- files renamed:
- bzrlib/commands.py => bzr.py
- bzrlib/store/__init__.py => bzrlib/store.py
- files modified:
- README
added
removed
bzrlib/inventory.py
1
# (C) 2005 Canonical Ltd
2
#
1
#! /usr/bin/env python
2
# -*- coding: UTF-8 -*-
3
3
4
# This program is free software; you can redistribute it and/or modify
4
5
# it under the terms of the GNU General Public License as published by
5
6
# the Free Software Foundation; either version 2 of the License, or
6
7
# (at your option) any later version.
7
#
8
8
9
# This program is distributed in the hope that it will be useful,
9
10
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
11
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
12
# GNU General Public License for more details.
12
#
13
13
14
# You should have received a copy of the GNU General Public License
14
15
# along with this program; if not, write to the Free Software
15
16
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
16
17
17
# FIXME: This refactoring of the workingtree code doesn't seem to keep
18
# the WorkingTree's copy of the inventory in sync with the branch. The
19
# branch modifies its working inventory when it does a commit to make
20
# missing files permanently removed.
21
22
# TODO: Maybe also keep the full path of the entry, and the children?
23
# But those depend on its position within a particular inventory, and
24
# it would be nice not to need to hold the backpointer here.
25
26
# This should really be an id randomly assigned when the tree is
27
# created, but it's not for now.
28
ROOT_ID = "TREE_ROOT"
29
30
31
import os.path
32
import re
33
import sys
34
import tarfile
35
import types
36
37
import bzrlib
38
from bzrlib.osutils import (pumpfile, quotefn, splitpath, joinpath,
39
appendpath, sha_strings)
40
from bzrlib.trace import mutter
41
from bzrlib.errors import (NotVersionedError, InvalidEntryName,
42
BzrError, BzrCheckError)
43
44
45
class InventoryEntry(object):
18
"""Inventories map files to their name in a revision."""
19
20
21
__copyright__ = "Copyright (C) 2005 Canonical Ltd."
22
__author__ = "Martin Pool <mbp@canonical.com>"
23
24
import sys, os.path, types
25
from sets import Set
26
27
from xml import XMLMixin
28
from ElementTree import ElementTree, Element
29
from errors import bailout
30
from osutils import uuid, quotefn, splitpath, joinpath, appendpath
31
from trace import mutter
32
33
class InventoryEntry(XMLMixin):
46
34
"""Description of a versioned file.
47
35
48
36
An InventoryEntry has the following fields, which are also
49
37
present in the XML inventory-entry element:
50
38
51
file_id
52
53
name
54
(within the parent directory)
55
56
parent_id
57
file_id of the parent directory, or ROOT_ID
58
59
revision
60
the revision_id in which this variation of this file was
61
introduced.
62
63
executable
64
Indicates that this file should be executable on systems
65
that support it.
66
67
text_sha1
68
sha-1 of the text of the file
69
70
text_size
71
size in bytes of the text of the file
72
73
(reading a version 4 tree created a text_id field.)
39
* *file_id*
40
* *name*: (only the basename within the directory, must not
41
contain slashes)
42
* *kind*: "directory" or "file"
43
* *directory_id*: (if absent/null means the branch root directory)
44
* *text_sha1*: only for files
45
* *text_size*: in bytes, only for files
46
* *text_id*: identifier for the text version, only for files
47
48
InventoryEntries can also exist inside a WorkingTree
49
inventory, in which case they are not yet bound to a
50
particular revision of the file. In that case the text_sha1,
51
text_size and text_id are absent.
52
74
53
75
54
>>> i = Inventory()
76
55
>>> i.path2id('')
77
'TREE_ROOT'
78
>>> i.add(InventoryDirectory('123', 'src', ROOT_ID))
79
InventoryDirectory('123', 'src', parent_id='TREE_ROOT')
80
>>> i.add(InventoryFile('2323', 'hello.c', parent_id='123'))
81
InventoryFile('2323', 'hello.c', parent_id='123')
82
>>> shouldbe = {0: 'src', 1: os.path.join('src','hello.c')}
83
>>> for ix, j in enumerate(i.iter_entries()):
84
... print (j[0] == shouldbe[ix], j[1])
56
>>> i.add(InventoryEntry('123', 'src', kind='directory'))
57
>>> i.add(InventoryEntry('2323', 'hello.c', parent_id='123'))
58
>>> for j in i.iter_entries():
59
... print j
85
60
...
86
(True, InventoryDirectory('123', 'src', parent_id='TREE_ROOT'))
87
(True, InventoryFile('2323', 'hello.c', parent_id='123'))
88
>>> i.add(InventoryFile('2323', 'bye.c', '123'))
61
('src', InventoryEntry('123', 'src', kind='directory', parent_id=None))
62
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
63
>>> i.add(InventoryEntry('2323', 'bye.c', parent_id='123'))
89
64
Traceback (most recent call last):
90
65
...
91
BzrError: inventory already contains entry with id {2323}
92
>>> i.add(InventoryFile('2324', 'bye.c', '123'))
93
InventoryFile('2324', 'bye.c', parent_id='123')
94
>>> i.add(InventoryDirectory('2325', 'wibble', '123'))
95
InventoryDirectory('2325', 'wibble', parent_id='123')
66
BzrError: ('inventory already contains entry with id {2323}', [])
67
>>> i.add(InventoryEntry('2324', 'bye.c', parent_id='123'))
68
>>> i.add(InventoryEntry('2325', 'wibble', parent_id='123', kind='directory'))
96
69
>>> i.path2id('src/wibble')
97
70
'2325'
98
71
>>> '2325' in i
99
72
True
100
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
101
InventoryFile('2326', 'wibble.c', parent_id='2325')
73
>>> i.add(InventoryEntry('2326', 'wibble.c', parent_id='2325'))
102
74
>>> i['2326']
103
InventoryFile('2326', 'wibble.c', parent_id='2325')
104
>>> for path, entry in i.iter_entries():
105
... print path.replace('\\\\', '/') # for win32 os.sep
106
... assert i.path2id(path)
75
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
76
>>> for j in i.iter_entries():
77
... print j[0]
78
... assert i.path2id(j[0])
107
79
...
108
80
src
109
81
src/bye.c
110
82
src/hello.c
111
83
src/wibble
112
84
src/wibble/wibble.c
113
>>> i.id2path('2326').replace('\\\\', '/')
85
>>> i.id2path('2326')
114
86
'src/wibble/wibble.c'
87
88
:todo: Maybe also keep the full path of the entry, and the children?
89
But those depend on its position within a particular inventory, and
90
it would be nice not to need to hold the backpointer here.
115
91
"""
116
117
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
118
'text_id', 'parent_id', 'children', 'executable',
119
'revision']
120
121
def _add_text_to_weave(self, new_lines, parents, weave_store, transaction):
122
weave_store.add_text(self.file_id, self.revision, new_lines, parents,
123
transaction)
124
125
def detect_changes(self, old_entry):
126
"""Return a (text_modified, meta_modified) from this to old_entry.
127
128
_read_tree_state must have been called on self and old_entry prior to
129
calling detect_changes.
130
"""
131
return False, False
132
133
def diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
134
output_to, reverse=False):
135
"""Perform a diff from this to to_entry.
136
137
text_diff will be used for textual difference calculation.
138
This is a template method, override _diff in child classes.
139
"""
140
self._read_tree_state(tree.id2path(self.file_id), tree)
141
if to_entry:
142
# cannot diff from one kind to another - you must do a removal
143
# and an addif they do not match.
144
assert self.kind == to_entry.kind
145
to_entry._read_tree_state(to_tree.id2path(to_entry.file_id),
146
to_tree)
147
self._diff(text_diff, from_label, tree, to_label, to_entry, to_tree,
148
output_to, reverse)
149
150
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
151
output_to, reverse=False):
152
"""Perform a diff between two entries of the same kind."""
153
154
def find_previous_heads(self, previous_inventories, entry_weave):
155
"""Return the revisions and entries that directly preceed this.
156
157
Returned as a map from revision to inventory entry.
158
159
This is a map containing the file revisions in all parents
160
for which the file exists, and its revision is not a parent of
161
any other. If the file is new, the set will be empty.
162
"""
163
def get_ancestors(weave, entry):
164
return set(map(weave.idx_to_name,
165
weave.inclusions([weave.lookup(entry.revision)])))
166
heads = {}
167
head_ancestors = {}
168
for inv in previous_inventories:
169
if self.file_id in inv:
170
ie = inv[self.file_id]
171
assert ie.file_id == self.file_id
172
if ie.revision in heads:
173
# fixup logic, there was a bug in revision updates.
174
# with x bit support.
175
try:
176
if heads[ie.revision].executable != ie.executable:
177
heads[ie.revision].executable = False
178
ie.executable = False
179
except AttributeError:
180
pass
181
assert heads[ie.revision] == ie
182
else:
183
# may want to add it.
184
# may already be covered:
185
already_present = 0 != len(
186
[head for head in heads
187
if ie.revision in head_ancestors[head]])
188
if already_present:
189
# an ancestor of a known head.
190
continue
191
# definately a head:
192
ancestors = get_ancestors(entry_weave, ie)
193
# may knock something else out:
194
check_heads = list(heads.keys())
195
for head in check_heads:
196
if head in ancestors:
197
# this head is not really a head
198
heads.pop(head)
199
head_ancestors[ie.revision] = ancestors
200
heads[ie.revision] = ie
201
return heads
202
203
def get_tar_item(self, root, dp, now, tree):
204
"""Get a tarfile item and a file stream for its content."""
205
item = tarfile.TarInfo(os.path.join(root, dp))
206
# TODO: would be cool to actually set it to the timestamp of the
207
# revision it was last changed
208
item.mtime = now
209
fileobj = self._put_in_tar(item, tree)
210
return item, fileobj
211
212
def has_text(self):
213
"""Return true if the object this entry represents has textual data.
214
215
Note that textual data includes binary content.
216
217
Also note that all entries get weave files created for them.
218
This attribute is primarily used when upgrading from old trees that
219
did not have the weave index for all inventory entries.
220
"""
221
return False
222
223
def __init__(self, file_id, name, parent_id, text_id=None):
92
def __init__(self, file_id, name, kind='file', text_id=None,
93
parent_id=None):
224
94
"""Create an InventoryEntry
225
95
226
96
The filename must be a single component, relative to the
227
97
parent directory; it cannot be a whole path or relative name.
228
98
229
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
99
>>> e = InventoryEntry('123', 'hello.c')
230
100
>>> e.name
231
101
'hello.c'
232
102
>>> e.file_id
233
103
'123'
234
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
104
>>> e = InventoryEntry('123', 'src/hello.c')
235
105
Traceback (most recent call last):
236
InvalidEntryName: Invalid entry name: src/hello.c
106
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
237
107
"""
238
assert isinstance(name, basestring), name
239
if '/' in name or '\\' in name:
240
raise InvalidEntryName(name=name)
241
self.executable = False
242
self.revision = None
243
self.text_sha1 = None
244
self.text_size = None
108
109
if len(splitpath(name)) != 1:
110
bailout('InventoryEntry name is not a simple filename: %r'
111
% name)
112
245
113
self.file_id = file_id
246
114
self.name = name
115
assert kind in ['file', 'directory']
116
self.kind = kind
247
117
self.text_id = text_id
248
118
self.parent_id = parent_id
249
self.symlink_target = None
250
251
def kind_character(self):
252
"""Return a short kind indicator useful for appending to names."""
253
raise BzrError('unknown kind %r' % self.kind)
254
255
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
256
257
def _put_in_tar(self, item, tree):
258
"""populate item for stashing in a tar, and return the content stream.
259
260
If no content is available, return None.
261
"""
262
raise BzrError("don't know how to export {%s} of kind %r" %
263
(self.file_id, self.kind))
264
265
def put_on_disk(self, dest, dp, tree):
266
"""Create a representation of self on disk in the prefix dest.
267
268
This is a template method - implement _put_on_disk in subclasses.
269
"""
270
fullpath = appendpath(dest, dp)
271
self._put_on_disk(fullpath, tree)
272
mutter(" export {%s} kind %s to %s", self.file_id,
273
self.kind, fullpath)
274
275
def _put_on_disk(self, fullpath, tree):
276
"""Put this entry onto disk at fullpath, from tree tree."""
277
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
278
279
def sorted_children(self):
280
l = self.children.items()
281
l.sort()
282
return l
283
284
@staticmethod
285
def versionable_kind(kind):
286
return kind in ('file', 'directory', 'symlink')
287
288
def check(self, checker, rev_id, inv, tree):
289
"""Check this inventory entry is intact.
290
291
This is a template method, override _check for kind specific
292
tests.
293
"""
294
if self.parent_id != None:
295
if not inv.has_id(self.parent_id):
296
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
297
% (self.parent_id, rev_id))
298
self._check(checker, rev_id, tree)
299
300
def _check(self, checker, rev_id, tree):
301
"""Check this inventory entry for kind specific errors."""
302
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
303
(self.kind, rev_id))
119
self.text_sha1 = None
120
self.text_size = None
304
121
305
122
306
123
def copy(self):
307
"""Clone this inventory entry."""
308
raise NotImplementedError
124
other = InventoryEntry(self.file_id, self.name, self.kind,
125
self.text_id, self.parent_id)
126
other.text_sha1 = self.text_sha1
127
other.text_size = self.text_size
128
return other
309
129
310
def _get_snapshot_change(self, previous_entries):
311
if len(previous_entries) > 1:
312
return 'merged'
313
elif len(previous_entries) == 0:
314
return 'added'
315
else:
316
return 'modified/renamed/reparented'
317
130
318
131
def __repr__(self):
319
return ("%s(%r, %r, parent_id=%r)"
132
return ("%s(%r, %r, kind=%r, parent_id=%r)"
320
133
% (self.__class__.__name__,
321
134
self.file_id,
322
135
self.name,
136
self.kind,
323
137
self.parent_id))
324
138
325
def snapshot(self, revision, path, previous_entries,
326
work_tree, weave_store, transaction):
327
"""Make a snapshot of this entry which may or may not have changed.
328
329
This means that all its fields are populated, that it has its
330
text stored in the text store or weave.
331
"""
332
mutter('new parents of %s are %r', path, previous_entries)
333
self._read_tree_state(path, work_tree)
334
if len(previous_entries) == 1:
335
# cannot be unchanged unless there is only one parent file rev.
336
parent_ie = previous_entries.values()[0]
337
if self._unchanged(parent_ie):
338
mutter("found unchanged entry")
339
self.revision = parent_ie.revision
340
return "unchanged"
341
return self.snapshot_revision(revision, previous_entries,
342
work_tree, weave_store, transaction)
343
344
def snapshot_revision(self, revision, previous_entries, work_tree,
345
weave_store, transaction):
346
"""Record this revision unconditionally."""
347
mutter('new revision for {%s}', self.file_id)
348
self.revision = revision
349
change = self._get_snapshot_change(previous_entries)
350
self._snapshot_text(previous_entries, work_tree, weave_store,
351
transaction)
352
return change
353
354
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
355
"""Record the 'text' of this entry, whatever form that takes.
356
357
This default implementation simply adds an empty text.
358
"""
359
mutter('storing file {%s} in revision {%s}',
360
self.file_id, self.revision)
361
self._add_text_to_weave([], file_parents, weave_store, transaction)
362
363
def __eq__(self, other):
139
140
def to_element(self):
141
"""Convert to XML element"""
142
e = Element('entry')
143
144
e.set('name', self.name)
145
e.set('file_id', self.file_id)
146
e.set('kind', self.kind)
147
148
if self.text_size is not None:
149
e.set('text_size', '%d' % self.text_size)
150
151
for f in ['text_id', 'text_sha1', 'parent_id']:
152
v = getattr(self, f)
153
if v is not None:
154
e.set(f, v)
155
156
e.tail = '\n'
157
158
return e
159
160
161
def from_element(cls, elt):
162
assert elt.tag == 'entry'
163
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'))
164
self.text_id = elt.get('text_id')
165
self.text_sha1 = elt.get('text_sha1')
166
self.parent_id = elt.get('parent_id')
167
168
## mutter("read inventoryentry: %r" % (elt.attrib))
169
170
v = elt.get('text_size')
171
self.text_size = v and int(v)
172
173
return self
174
175
176
from_element = classmethod(from_element)
177
178
def __cmp__(self, other):
179
if self is other:
180
return 0
364
181
if not isinstance(other, InventoryEntry):
365
182
return NotImplemented
366
183
367
return ((self.file_id == other.file_id)
368
and (self.name == other.name)
369
and (other.symlink_target == self.symlink_target)
370
and (self.text_sha1 == other.text_sha1)
371
and (self.text_size == other.text_size)
372
and (self.text_id == other.text_id)
373
and (self.parent_id == other.parent_id)
374
and (self.kind == other.kind)
375
and (self.revision == other.revision)
376
and (self.executable == other.executable)
377
)
378
379
def __ne__(self, other):
380
return not (self == other)
381
382
def __hash__(self):
383
raise ValueError('not hashable')
384
385
def _unchanged(self, previous_ie):
386
"""Has this entry changed relative to previous_ie.
387
388
This method should be overriden in child classes.
389
"""
390
compatible = True
391
# different inv parent
392
if previous_ie.parent_id != self.parent_id:
393
compatible = False
394
# renamed
395
elif previous_ie.name != self.name:
396
compatible = False
397
return compatible
398
399
def _read_tree_state(self, path, work_tree):
400
"""Populate fields in the inventory entry from the given tree.
401
402
Note that this should be modified to be a noop on virtual trees
403
as all entries created there are prepopulated.
404
"""
405
# TODO: Rather than running this manually, we should check the
406
# working sha1 and other expensive properties when they're
407
# first requested, or preload them if they're already known
408
pass # nothing to do by default
409
410
411
class RootEntry(InventoryEntry):
412
413
def _check(self, checker, rev_id, tree):
414
"""See InventoryEntry._check"""
415
416
def __init__(self, file_id):
417
self.file_id = file_id
418
self.children = {}
419
self.kind = 'root_directory'
420
self.parent_id = None
421
self.name = u''
422
423
def __eq__(self, other):
424
if not isinstance(other, RootEntry):
425
return NotImplemented
426
427
return (self.file_id == other.file_id) \
428
and (self.children == other.children)
429
430
431
class InventoryDirectory(InventoryEntry):
432
"""A directory in an inventory."""
433
434
def _check(self, checker, rev_id, tree):
435
"""See InventoryEntry._check"""
436
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
437
raise BzrCheckError('directory {%s} has text in revision {%s}'
438
% (self.file_id, rev_id))
439
440
def copy(self):
441
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
442
other.revision = self.revision
443
# note that children are *not* copied; they're pulled across when
444
# others are added
445
return other
446
447
def __init__(self, file_id, name, parent_id):
448
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
449
self.children = {}
450
self.kind = 'directory'
451
452
def kind_character(self):
453
"""See InventoryEntry.kind_character."""
454
return '/'
455
456
def _put_in_tar(self, item, tree):
457
"""See InventoryEntry._put_in_tar."""
458
item.type = tarfile.DIRTYPE
459
fileobj = None
460
item.name += '/'
461
item.size = 0
462
item.mode = 0755
463
return fileobj
464
465
def _put_on_disk(self, fullpath, tree):
466
"""See InventoryEntry._put_on_disk."""
467
os.mkdir(fullpath)
468
469
470
class InventoryFile(InventoryEntry):
471
"""A file in an inventory."""
472
473
def _check(self, checker, rev_id, tree):
474
"""See InventoryEntry._check"""
475
revision = self.revision
476
t = (self.file_id, revision)
477
if t in checker.checked_texts:
478
prev_sha = checker.checked_texts[t]
479
if prev_sha != self.text_sha1:
480
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
481
(self.file_id, rev_id))
482
else:
483
checker.repeated_text_cnt += 1
484
return
485
mutter('check version {%s} of {%s}', rev_id, self.file_id)
486
file_lines = tree.get_file_lines(self.file_id)
487
checker.checked_text_cnt += 1
488
if self.text_size != sum(map(len, file_lines)):
489
raise BzrCheckError('text {%s} wrong size' % self.text_id)
490
if self.text_sha1 != sha_strings(file_lines):
491
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
492
checker.checked_texts[t] = self.text_sha1
493
494
def copy(self):
495
other = InventoryFile(self.file_id, self.name, self.parent_id)
496
other.executable = self.executable
497
other.text_id = self.text_id
498
other.text_sha1 = self.text_sha1
499
other.text_size = self.text_size
500
other.revision = self.revision
501
return other
502
503
def detect_changes(self, old_entry):
504
"""See InventoryEntry.detect_changes."""
505
assert self.text_sha1 != None
506
assert old_entry.text_sha1 != None
507
text_modified = (self.text_sha1 != old_entry.text_sha1)
508
meta_modified = (self.executable != old_entry.executable)
509
return text_modified, meta_modified
510
511
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
512
output_to, reverse=False):
513
"""See InventoryEntry._diff."""
514
from_text = tree.get_file(self.file_id).readlines()
515
if to_entry:
516
to_text = to_tree.get_file(to_entry.file_id).readlines()
517
else:
518
to_text = []
519
if not reverse:
520
text_diff(from_label, from_text,
521
to_label, to_text, output_to)
522
else:
523
text_diff(to_label, to_text,
524
from_label, from_text, output_to)
525
526
def has_text(self):
527
"""See InventoryEntry.has_text."""
528
return True
529
530
def __init__(self, file_id, name, parent_id):
531
super(InventoryFile, self).__init__(file_id, name, parent_id)
532
self.kind = 'file'
533
534
def kind_character(self):
535
"""See InventoryEntry.kind_character."""
536
return ''
537
538
def _put_in_tar(self, item, tree):
539
"""See InventoryEntry._put_in_tar."""
540
item.type = tarfile.REGTYPE
541
fileobj = tree.get_file(self.file_id)
542
item.size = self.text_size
543
if tree.is_executable(self.file_id):
544
item.mode = 0755
545
else:
546
item.mode = 0644
547
return fileobj
548
549
def _put_on_disk(self, fullpath, tree):
550
"""See InventoryEntry._put_on_disk."""
551
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
552
if tree.is_executable(self.file_id):
553
os.chmod(fullpath, 0755)
554
555
def _read_tree_state(self, path, work_tree):
556
"""See InventoryEntry._read_tree_state."""
557
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
558
self.executable = work_tree.is_executable(self.file_id)
559
560
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
561
"""See InventoryEntry._snapshot_text."""
562
mutter('storing file {%s} in revision {%s}',
563
self.file_id, self.revision)
564
# special case to avoid diffing on renames or
565
# reparenting
566
if (len(file_parents) == 1
567
and self.text_sha1 == file_parents.values()[0].text_sha1
568
and self.text_size == file_parents.values()[0].text_size):
569
previous_ie = file_parents.values()[0]
570
weave_store.add_identical_text(
571
self.file_id, previous_ie.revision,
572
self.revision, file_parents, transaction)
573
else:
574
new_lines = work_tree.get_file(self.file_id).readlines()
575
self._add_text_to_weave(new_lines, file_parents, weave_store,
576
transaction)
577
self.text_sha1 = sha_strings(new_lines)
578
self.text_size = sum(map(len, new_lines))
579
580
581
def _unchanged(self, previous_ie):
582
"""See InventoryEntry._unchanged."""
583
compatible = super(InventoryFile, self)._unchanged(previous_ie)
584
if self.text_sha1 != previous_ie.text_sha1:
585
compatible = False
586
else:
587
# FIXME: 20050930 probe for the text size when getting sha1
588
# in _read_tree_state
589
self.text_size = previous_ie.text_size
590
if self.executable != previous_ie.executable:
591
compatible = False
592
return compatible
593
594
595
class InventoryLink(InventoryEntry):
596
"""A file in an inventory."""
597
598
__slots__ = ['symlink_target']
599
600
def _check(self, checker, rev_id, tree):
601
"""See InventoryEntry._check"""
602
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
603
raise BzrCheckError('symlink {%s} has text in revision {%s}'
604
% (self.file_id, rev_id))
605
if self.symlink_target == None:
606
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
607
% (self.file_id, rev_id))
608
609
def copy(self):
610
other = InventoryLink(self.file_id, self.name, self.parent_id)
611
other.symlink_target = self.symlink_target
612
other.revision = self.revision
613
return other
614
615
def detect_changes(self, old_entry):
616
"""See InventoryEntry.detect_changes."""
617
# FIXME: which _modified field should we use ? RBC 20051003
618
text_modified = (self.symlink_target != old_entry.symlink_target)
619
if text_modified:
620
mutter(" symlink target changed")
621
meta_modified = False
622
return text_modified, meta_modified
623
624
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
625
output_to, reverse=False):
626
"""See InventoryEntry._diff."""
627
from_text = self.symlink_target
628
if to_entry is not None:
629
to_text = to_entry.symlink_target
630
if reverse:
631
temp = from_text
632
from_text = to_text
633
to_text = temp
634
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
635
else:
636
if not reverse:
637
print >>output_to, '=== target was %r' % self.symlink_target
638
else:
639
print >>output_to, '=== target is %r' % self.symlink_target
640
641
def __init__(self, file_id, name, parent_id):
642
super(InventoryLink, self).__init__(file_id, name, parent_id)
643
self.kind = 'symlink'
644
645
def kind_character(self):
646
"""See InventoryEntry.kind_character."""
647
return ''
648
649
def _put_in_tar(self, item, tree):
650
"""See InventoryEntry._put_in_tar."""
651
item.type = tarfile.SYMTYPE
652
fileobj = None
653
item.size = 0
654
item.mode = 0755
655
item.linkname = self.symlink_target
656
return fileobj
657
658
def _put_on_disk(self, fullpath, tree):
659
"""See InventoryEntry._put_on_disk."""
660
try:
661
os.symlink(self.symlink_target, fullpath)
662
except OSError,e:
663
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
664
665
def _read_tree_state(self, path, work_tree):
666
"""See InventoryEntry._read_tree_state."""
667
self.symlink_target = work_tree.get_symlink_target(self.file_id)
668
669
def _unchanged(self, previous_ie):
670
"""See InventoryEntry._unchanged."""
671
compatible = super(InventoryLink, self)._unchanged(previous_ie)
672
if self.symlink_target != previous_ie.symlink_target:
673
compatible = False
674
return compatible
675
676
677
class Inventory(object):
184
return cmp(self.file_id, other.file_id) \
185
or cmp(self.name, other.name) \
186
or cmp(self.text_sha1, other.text_sha1) \
187
or cmp(self.text_size, other.text_size) \
188
or cmp(self.text_id, other.text_id) \
189
or cmp(self.parent_id, other.parent_id) \
190
or cmp(self.kind, other.kind)
191
192
193
194
class Inventory(XMLMixin):
678
195
"""Inventory of versioned files in a tree.
679
196
680
This describes which file_id is present at each point in the tree,
681
and possibly the SHA-1 or other information about the file.
682
Entries can be looked up either by path or by file_id.
197
An Inventory acts like a set of InventoryEntry items. You can
198
also look files up by their file_id or name.
199
200
May be read from and written to a metadata file in a tree. To
201
manipulate the inventory (for example to add a file), it is read
202
in, modified, and then written back out.
683
203
684
204
The inventory represents a typical unix file tree, with
685
205
directories containing files and subdirectories. We never store
689
209
returned quickly.
690
210
691
211
InventoryEntry objects must not be modified after they are
692
inserted, other than through the Inventory API.
212
inserted.
693
213
694
214
>>> inv = Inventory()
695
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
696
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT')
215
>>> inv.write_xml(sys.stdout)
216
<inventory>
217
</inventory>
218
>>> inv.add(InventoryEntry('123-123', 'hello.c'))
697
219
>>> inv['123-123'].name
698
220
'hello.c'
221
>>> for file_id in inv: print file_id
222
...
223
123-123
699
224
700
225
May be treated as an iterator or set to look up file ids:
701
226
708
233
709
234
>>> [x[0] for x in inv.iter_entries()]
710
235
['hello.c']
711
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
712
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
713
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
236
237
>>> inv.write_xml(sys.stdout)
238
<inventory>
239
<entry file_id="123-123" kind="file" name="hello.c" />
240
</inventory>
241
714
242
"""
715
def __init__(self, root_id=ROOT_ID):
243
244
## TODO: Clear up handling of files in subdirectories; we probably
245
## do want to be able to just look them up by name but this
246
## probably means gradually walking down the path, looking up as we go.
247
248
## TODO: Make sure only canonical filenames are stored.
249
250
## TODO: Do something sensible about the possible collisions on
251
## case-losing filesystems. Perhaps we should just always forbid
252
## such collisions.
253
254
## _tree should probably just be stored as
255
## InventoryEntry._children on each directory.
256
257
def __init__(self):
716
258
"""Create or read an inventory.
717
259
718
260
If a working directory is specified, the inventory is read
719
261
from there. If the file is specified, read from that. If not,
720
262
the inventory is created empty.
721
722
The inventory is created with a default root directory, with
723
an id of None.
724
263
"""
725
# We are letting Branch.initialize() create a unique inventory
726
# root id. Rather than generating a random one here.
727
#if root_id is None:
728
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
729
self.root = RootEntry(root_id)
730
self._byid = {self.root.file_id: self.root}
731
732
733
def copy(self):
734
other = Inventory(self.root.file_id)
735
# copy recursively so we know directories will be added before
736
# their children. There are more efficient ways than this...
737
for path, entry in self.iter_entries():
738
if entry == self.root:
739
continue
740
other.add(entry.copy())
741
return other
264
self._byid = dict()
265
266
# _tree is indexed by parent_id; at each level a map from name
267
# to ie. The None entry is the root.
268
self._tree = {None: {}}
742
269
743
270
744
271
def __iter__(self):
750
277
return len(self._byid)
751
278
752
279
753
def iter_entries(self, from_dir=None):
280
def iter_entries(self, parent_id=None):
754
281
"""Return (path, entry) pairs, in order by name."""
755
if from_dir == None:
756
assert self.root
757
from_dir = self.root
758
elif isinstance(from_dir, basestring):
759
from_dir = self._byid[from_dir]
760
761
kids = from_dir.children.items()
282
kids = self._tree[parent_id].items()
762
283
kids.sort()
763
284
for name, ie in kids:
764
285
yield name, ie
765
286
if ie.kind == 'directory':
766
for cn, cie in self.iter_entries(from_dir=ie.file_id):
767
yield os.path.join(name, cn), cie
768
769
770
def entries(self):
771
"""Return list of (path, ie) for all entries except the root.
772
773
This may be faster than iter_entries.
774
"""
775
accum = []
776
def descend(dir_ie, dir_path):
777
kids = dir_ie.children.items()
778
kids.sort()
779
for name, ie in kids:
780
child_path = os.path.join(dir_path, name)
781
accum.append((child_path, ie))
782
if ie.kind == 'directory':
783
descend(ie, child_path)
784
785
descend(self.root, u'')
786
return accum
787
788
789
def directories(self):
790
"""Return (path, entry) pairs for all directories, including the root.
791
"""
792
accum = []
793
def descend(parent_ie, parent_path):
794
accum.append((parent_path, parent_ie))
795
796
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
797
kids.sort()
798
799
for name, child_ie in kids:
800
child_path = os.path.join(parent_path, name)
801
descend(child_ie, child_path)
802
descend(self.root, u'')
803
return accum
287
for cn, cie in self.iter_entries(parent_id=ie.file_id):
288
yield joinpath([name, cn]), cie
289
290
291
def directories(self, include_root=True):
292
"""Return (path, entry) pairs for all directories.
293
"""
294
if include_root:
295
yield '', None
296
for path, entry in self.iter_entries():
297
if entry.kind == 'directory':
298
yield path, entry
804
299
805
300
806
301
302
def children(self, parent_id):
303
"""Return entries that are direct children of parent_id."""
304
return self._tree[parent_id]
305
306
307
308
# TODO: return all paths and entries
309
310
807
311
def __contains__(self, file_id):
808
312
"""True if this entry contains a file with given id.
809
313
810
314
>>> inv = Inventory()
811
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
812
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
315
>>> inv.add(InventoryEntry('123', 'foo.c'))
813
316
>>> '123' in inv
814
317
True
815
318
>>> '456' in inv
822
325
"""Return the entry for given file_id.
823
326
824
327
>>> inv = Inventory()
825
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
826
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
328
>>> inv.add(InventoryEntry('123123', 'hello.c'))
827
329
>>> inv['123123'].name
828
330
'hello.c'
829
331
"""
830
try:
831
return self._byid[file_id]
832
except KeyError:
833
if file_id == None:
834
raise BzrError("can't look up file_id None")
835
else:
836
raise BzrError("file_id {%s} not in inventory" % file_id)
837
838
839
def get_file_kind(self, file_id):
840
return self._byid[file_id].kind
841
842
def get_child(self, parent_id, filename):
843
return self[parent_id].children.get(filename)
332
return self._byid[file_id]
844
333
845
334
846
335
def add(self, entry):
847
336
"""Add entry to inventory.
848
337
849
338
To add a file to a branch ready to be committed, use Branch.add,
850
which calls this.
851
852
Returns the new entry object.
853
"""
854
if entry.file_id in self._byid:
855
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
856
857
if entry.parent_id == ROOT_ID or entry.parent_id is None:
858
entry.parent_id = self.root.file_id
859
860
try:
861
parent = self._byid[entry.parent_id]
862
except KeyError:
863
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
864
865
if parent.children.has_key(entry.name):
866
raise BzrError("%s is already versioned" %
867
appendpath(self.id2path(parent.file_id), entry.name))
339
which calls this."""
340
if entry.file_id in self:
341
bailout("inventory already contains entry with id {%s}" % entry.file_id)
342
343
if entry.parent_id != None:
344
if entry.parent_id not in self:
345
bailout("parent_id %s of new entry not found in inventory"
346
% entry.parent_id)
347
348
if self._tree[entry.parent_id].has_key(entry.name):
349
bailout("%s is already versioned"
350
% appendpath(self.id2path(entry.parent_id), entry.name))
868
351
869
352
self._byid[entry.file_id] = entry
870
parent.children[entry.name] = entry
871
return entry
872
873
874
def add_path(self, relpath, kind, file_id=None):
875
"""Add entry from a path.
876
877
The immediate parent must already be versioned.
878
879
Returns the new entry object."""
880
from bzrlib.workingtree import gen_file_id
881
882
parts = bzrlib.osutils.splitpath(relpath)
883
if len(parts) == 0:
884
raise BzrError("cannot re-add root of inventory")
885
886
if file_id == None:
887
file_id = gen_file_id(relpath)
888
889
parent_path = parts[:-1]
890
parent_id = self.path2id(parent_path)
891
if parent_id == None:
892
raise NotVersionedError(path=parent_path)
893
if kind == 'directory':
894
ie = InventoryDirectory(file_id, parts[-1], parent_id)
895
elif kind == 'file':
896
ie = InventoryFile(file_id, parts[-1], parent_id)
897
elif kind == 'symlink':
898
ie = InventoryLink(file_id, parts[-1], parent_id)
899
else:
900
raise BzrError("unknown kind %r" % kind)
901
return self.add(ie)
353
self._tree[entry.parent_id][entry.name] = entry
354
355
if entry.kind == 'directory':
356
self._tree[entry.file_id] = {}
902
357
903
358
904
359
def __delitem__(self, file_id):
905
360
"""Remove entry by id.
906
361
907
362
>>> inv = Inventory()
908
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
909
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
363
>>> inv.add(InventoryEntry('123', 'foo.c'))
910
364
>>> '123' in inv
911
365
True
912
366
>>> del inv['123']
915
369
"""
916
370
ie = self[file_id]
917
371
918
assert self[ie.parent_id].children[ie.name] == ie
372
assert self._tree[ie.parent_id][ie.name] == ie
919
373
920
374
# TODO: Test deleting all children; maybe hoist to a separate
921
375
# deltree method?
922
376
if ie.kind == 'directory':
923
for cie in ie.children.values():
377
for cie in self._tree[file_id].values():
924
378
del self[cie.file_id]
925
del ie.children
379
del self._tree[file_id]
926
380
927
381
del self._byid[file_id]
928
del self[ie.parent_id].children[ie.name]
929
930
931
def __eq__(self, other):
382
del self._tree[ie.parent_id][ie.name]
383
384
385
def id_set(self):
386
return Set(self._byid)
387
388
389
def to_element(self):
390
"""Convert to XML Element"""
391
e = Element('inventory')
392
e.text = '\n'
393
for path, ie in self.iter_entries():
394
e.append(ie.to_element())
395
return e
396
397
398
def from_element(cls, elt):
399
"""Construct from XML Element
400
401
>>> inv = Inventory()
402
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c'))
403
>>> elt = inv.to_element()
404
>>> inv2 = Inventory.from_element(elt)
405
>>> inv2 == inv
406
True
407
"""
408
assert elt.tag == 'inventory'
409
o = cls()
410
for e in elt:
411
o.add(InventoryEntry.from_element(e))
412
return o
413
414
from_element = classmethod(from_element)
415
416
417
def __cmp__(self, other):
932
418
"""Compare two sets by comparing their contents.
933
419
934
420
>>> i1 = Inventory()
935
421
>>> i2 = Inventory()
936
422
>>> i1 == i2
937
423
True
938
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
939
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
424
>>> i1.add(InventoryEntry('123', 'foo'))
940
425
>>> i1 == i2
941
426
False
942
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
943
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
427
>>> i2.add(InventoryEntry('123', 'foo'))
944
428
>>> i1 == i2
945
429
True
946
430
"""
431
if self is other:
432
return 0
433
947
434
if not isinstance(other, Inventory):
948
435
return NotImplemented
949
436
950
if len(self._byid) != len(other._byid):
951
# shortcut: obviously not the same
952
return False
953
954
return self._byid == other._byid
955
956
957
def __ne__(self, other):
958
return not self.__eq__(other)
959
960
961
def __hash__(self):
962
raise ValueError('not hashable')
963
964
965
def get_idpath(self, file_id):
966
"""Return a list of file_ids for the path to an entry.
967
968
The list contains one element for each directory followed by
969
the id of the file itself. So the length of the returned list
970
is equal to the depth of the file in the tree, counting the
971
root directory as depth 1.
972
"""
437
if self.id_set() ^ other.id_set():
438
return 1
439
440
for file_id in self._byid:
441
c = cmp(self[file_id], other[file_id])
442
if c: return c
443
444
return 0
445
446
447
def id2path(self, file_id):
448
"""Return as a list the path to file_id."""
973
449
p = []
974
450
while file_id != None:
975
try:
976
ie = self._byid[file_id]
977
except KeyError:
978
raise BzrError("file_id {%s} not found in inventory" % file_id)
979
p.insert(0, ie.file_id)
451
ie = self[file_id]
452
p = [ie.name] + p
980
453
file_id = ie.parent_id
981
return p
982
983
984
def id2path(self, file_id):
985
"""Return as a list the path to file_id.
986
987
>>> i = Inventory()
988
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
989
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
990
>>> print i.id2path('foo-id').replace(os.sep, '/')
991
src/foo.c
992
"""
993
# get all names, skipping root
994
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
995
return os.sep.join(p)
454
return joinpath(p)
996
455
997
456
998
457
1004
463
1005
464
This returns the entry of the last component in the path,
1006
465
which may be either a file or a directory.
1007
1008
Returns None iff the path is not found.
1009
466
"""
1010
if isinstance(name, types.StringTypes):
1011
name = splitpath(name)
1012
1013
mutter("lookup path %r" % name)
1014
1015
parent = self.root
1016
for f in name:
467
assert isinstance(name, types.StringTypes)
468
469
parent_id = None
470
for f in splitpath(name):
1017
471
try:
1018
cie = parent.children[f]
472
cie = self._tree[parent_id][f]
1019
473
assert cie.name == f
1020
assert cie.parent_id == parent.file_id
1021
parent = cie
474
parent_id = cie.file_id
1022
475
except KeyError:
1023
476
# or raise an error?
1024
477
return None
1025
478
1026
return parent.file_id
479
return parent_id
480
481
482
def get_child(self, parent_id, child_name):
483
return self._tree[parent_id].get(child_name)
1027
484
1028
485
1029
486
def has_filename(self, names):
1031
488
1032
489
1033
490
def has_id(self, file_id):
491
assert isinstance(file_id, str)
1034
492
return self._byid.has_key(file_id)
1035
493
1036
494
1037
def rename(self, file_id, new_parent_id, new_name):
1038
"""Move a file within the inventory.
1039
1040
This can change either the name, or the parent, or both.
1041
1042
This does not move the working file."""
1043
if not is_valid_name(new_name):
1044
raise BzrError("not an acceptable filename: %r" % new_name)
1045
1046
new_parent = self._byid[new_parent_id]
1047
if new_name in new_parent.children:
1048
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
1049
1050
new_parent_idpath = self.get_idpath(new_parent_id)
1051
if file_id in new_parent_idpath:
1052
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
1053
% (self.id2path(file_id), self.id2path(new_parent_id)))
1054
1055
file_ie = self._byid[file_id]
1056
old_parent = self._byid[file_ie.parent_id]
1057
1058
# TODO: Don't leave things messed up if this fails
1059
1060
del old_parent.children[file_ie.name]
1061
new_parent.children[new_name] = file_ie
1062
1063
file_ie.name = new_name
1064
file_ie.parent_id = new_parent_id
1065
1066
1067
1068
1069
_NAME_RE = None
1070
1071
def is_valid_name(name):
1072
global _NAME_RE
1073
if _NAME_RE == None:
1074
_NAME_RE = re.compile(r'^[^/\\]+$')
1075
1076
return bool(_NAME_RE.match(name))
495
496
if __name__ == '__main__':
497
import doctest, inventory
498
doctest.testmod(inventory)
Loggerhead is a web-based interface for Breezy
Version: 2.0.1