« back to all changes in this revision
Viewing changes to bzrlib/repository.py
- browse files at revision 1861.2.6
- compare with another revision
- download diff
- download tarball
- view history from revision 1861.2.6
- Committer: Alexander Belchenko
- Date: 2006-07-30 06:52:39 UTC
- mto: (1711.2.111 jam-integration)
- mto: This revision was merged to the branch mainline in revision 1906.
- Revision ID: bialix@ukr.net-20060730065239-03b4ac02d9f56202
branding: change Bazaar-NG to Bazaar
- files added:
- BRANCH.TODO
- bzrlib/benchmarks
- bzrlib/bundle
- bzrlib/bundle/old
- bzrlib/bundle/serializer
- bzrlib/doc
- bzrlib/doc/api
- bzrlib/export
- bzrlib/plugins
- bzrlib/plugins/launchpad
- bzrlib/store
- bzrlib/store/revision
- bzrlib/store/versioned
- bzrlib/tests
- bzrlib/tests/blackbox
- bzrlib/tests/branch_implementations
- bzrlib/tests/bzrdir_implementations
- bzrlib/tests/interrepository_implementations
- bzrlib/tests/interversionedfile_implementations
- bzrlib/tests/repository_implementations
- bzrlib/tests/revisionstore_implementations
- bzrlib/tests/test_patches_data
- bzrlib/tests/workingtree_implementations
- bzrlib/transport
- bzrlib/transport/http
- bzrlib/ui
- bzrlib/util
- bzrlib/util/configobj
- bzrlib/util/configobj/docs
- bzrlib/util/effbot
- bzrlib/util/effbot/org
- contrib/emacs
- tools
- tools/doc_generate
- tools/win32
- files removed:
- bzrlib/cache.py
- notes
- files renamed:
- bzrlib/store.py => bzrlib/store/__init__.py
- elementtree/ => bzrlib/util/elementtree/
- urlgrabber/ => bzrlib/util/urlgrabber/
- bzrlib/xml.py => bzrlib/xml_serializer.py
- contrib/bash/bzr => contrib/bash/bzr.simple
- bzrlib/newinventory.py => contrib/newinventory.py
- files modified:
- .bzrignore
- bzr *
- setup.py *
added
removed
bzrlib/repository.py
1
# Copyright (C) 2005, 2006 Canonical Ltd
2
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
7
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
12
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
16
17
from binascii import hexlify
18
from copy import deepcopy
19
from cStringIO import StringIO
20
import re
21
import time
22
from unittest import TestSuite
23
24
from bzrlib import bzrdir, check, delta, gpg, errors, xml5, ui, transactions, osutils
25
from bzrlib.decorators import needs_read_lock, needs_write_lock
26
from bzrlib.errors import InvalidRevisionId
27
from bzrlib.graph import Graph
28
from bzrlib.inter import InterObject
29
from bzrlib.inventory import Inventory
30
from bzrlib.knit import KnitVersionedFile, KnitPlainFactory
31
from bzrlib.lockable_files import LockableFiles, TransportLock
32
from bzrlib.lockdir import LockDir
33
from bzrlib.osutils import (safe_unicode, rand_bytes, compact_date,
34
local_time_offset)
35
from bzrlib.revision import NULL_REVISION, Revision
36
from bzrlib.store.versioned import VersionedFileStore, WeaveStore
37
from bzrlib.store.text import TextStore
38
from bzrlib.symbol_versioning import (deprecated_method,
39
zero_nine,
40
)
41
from bzrlib.trace import mutter, note
42
from bzrlib.tree import RevisionTree, EmptyTree
43
from bzrlib.tsort import topo_sort
44
from bzrlib.testament import Testament
45
from bzrlib.tree import EmptyTree
46
from bzrlib.weave import WeaveFile
47
48
49
class Repository(object):
50
"""Repository holding history for one or more branches.
51
52
The repository holds and retrieves historical information including
53
revisions and file history. It's normally accessed only by the Branch,
54
which views a particular line of development through that history.
55
56
The Repository builds on top of Stores and a Transport, which respectively
57
describe the disk data format and the way of accessing the (possibly
58
remote) disk.
59
"""
60
61
@needs_write_lock
62
def add_inventory(self, revid, inv, parents):
63
"""Add the inventory inv to the repository as revid.
64
65
:param parents: The revision ids of the parents that revid
66
is known to have and are in the repository already.
67
68
returns the sha1 of the serialized inventory.
69
"""
70
assert inv.revision_id is None or inv.revision_id == revid, \
71
"Mismatch between inventory revision" \
72
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
73
inv_text = xml5.serializer_v5.write_inventory_to_string(inv)
74
inv_sha1 = osutils.sha_string(inv_text)
75
inv_vf = self.control_weaves.get_weave('inventory',
76
self.get_transaction())
77
self._inventory_add_lines(inv_vf, revid, parents, osutils.split_lines(inv_text))
78
return inv_sha1
79
80
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
81
final_parents = []
82
for parent in parents:
83
if parent in inv_vf:
84
final_parents.append(parent)
85
86
inv_vf.add_lines(revid, final_parents, lines)
87
88
@needs_write_lock
89
def add_revision(self, rev_id, rev, inv=None, config=None):
90
"""Add rev to the revision store as rev_id.
91
92
:param rev_id: the revision id to use.
93
:param rev: The revision object.
94
:param inv: The inventory for the revision. if None, it will be looked
95
up in the inventory storer
96
:param config: If None no digital signature will be created.
97
If supplied its signature_needed method will be used
98
to determine if a signature should be made.
99
"""
100
if config is not None and config.signature_needed():
101
if inv is None:
102
inv = self.get_inventory(rev_id)
103
plaintext = Testament(rev, inv).as_short_text()
104
self.store_revision_signature(
105
gpg.GPGStrategy(config), plaintext, rev_id)
106
if not rev_id in self.get_inventory_weave():
107
if inv is None:
108
raise errors.WeaveRevisionNotPresent(rev_id,
109
self.get_inventory_weave())
110
else:
111
# yes, this is not suitable for adding with ghosts.
112
self.add_inventory(rev_id, inv, rev.parent_ids)
113
self._revision_store.add_revision(rev, self.get_transaction())
114
115
@needs_read_lock
116
def _all_possible_ids(self):
117
"""Return all the possible revisions that we could find."""
118
return self.get_inventory_weave().versions()
119
120
def all_revision_ids(self):
121
"""Returns a list of all the revision ids in the repository.
122
123
This is deprecated because code should generally work on the graph
124
reachable from a particular revision, and ignore any other revisions
125
that might be present. There is no direct replacement method.
126
"""
127
return self._all_revision_ids()
128
129
@needs_read_lock
130
def _all_revision_ids(self):
131
"""Returns a list of all the revision ids in the repository.
132
133
These are in as much topological order as the underlying store can
134
present: for weaves ghosts may lead to a lack of correctness until
135
the reweave updates the parents list.
136
"""
137
if self._revision_store.text_store.listable():
138
return self._revision_store.all_revision_ids(self.get_transaction())
139
result = self._all_possible_ids()
140
return self._eliminate_revisions_not_present(result)
141
142
def break_lock(self):
143
"""Break a lock if one is present from another instance.
144
145
Uses the ui factory to ask for confirmation if the lock may be from
146
an active process.
147
"""
148
self.control_files.break_lock()
149
150
@needs_read_lock
151
def _eliminate_revisions_not_present(self, revision_ids):
152
"""Check every revision id in revision_ids to see if we have it.
153
154
Returns a set of the present revisions.
155
"""
156
result = []
157
for id in revision_ids:
158
if self.has_revision(id):
159
result.append(id)
160
return result
161
162
@staticmethod
163
def create(a_bzrdir):
164
"""Construct the current default format repository in a_bzrdir."""
165
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
166
167
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
168
"""instantiate a Repository.
169
170
:param _format: The format of the repository on disk.
171
:param a_bzrdir: The BzrDir of the repository.
172
173
In the future we will have a single api for all stores for
174
getting file texts, inventories and revisions, then
175
this construct will accept instances of those things.
176
"""
177
super(Repository, self).__init__()
178
self._format = _format
179
# the following are part of the public API for Repository:
180
self.bzrdir = a_bzrdir
181
self.control_files = control_files
182
self._revision_store = _revision_store
183
self.text_store = text_store
184
# backwards compatibility
185
self.weave_store = text_store
186
# not right yet - should be more semantically clear ?
187
#
188
self.control_store = control_store
189
self.control_weaves = control_store
190
# TODO: make sure to construct the right store classes, etc, depending
191
# on whether escaping is required.
192
193
def __repr__(self):
194
return '%s(%r)' % (self.__class__.__name__,
195
self.bzrdir.transport.base)
196
197
def is_locked(self):
198
return self.control_files.is_locked()
199
200
def lock_write(self):
201
self.control_files.lock_write()
202
203
def lock_read(self):
204
self.control_files.lock_read()
205
206
def get_physical_lock_status(self):
207
return self.control_files.get_physical_lock_status()
208
209
@needs_read_lock
210
def missing_revision_ids(self, other, revision_id=None):
211
"""Return the revision ids that other has that this does not.
212
213
These are returned in topological order.
214
215
revision_id: only return revision ids included by revision_id.
216
"""
217
return InterRepository.get(other, self).missing_revision_ids(revision_id)
218
219
@staticmethod
220
def open(base):
221
"""Open the repository rooted at base.
222
223
For instance, if the repository is at URL/.bzr/repository,
224
Repository.open(URL) -> a Repository instance.
225
"""
226
control = bzrdir.BzrDir.open(base)
227
return control.open_repository()
228
229
def copy_content_into(self, destination, revision_id=None, basis=None):
230
"""Make a complete copy of the content in self into destination.
231
232
This is a destructive operation! Do not use it on existing
233
repositories.
234
"""
235
return InterRepository.get(self, destination).copy_content(revision_id, basis)
236
237
def fetch(self, source, revision_id=None, pb=None):
238
"""Fetch the content required to construct revision_id from source.
239
240
If revision_id is None all content is copied.
241
"""
242
return InterRepository.get(source, self).fetch(revision_id=revision_id,
243
pb=pb)
244
245
def get_commit_builder(self, branch, parents, config, timestamp=None,
246
timezone=None, committer=None, revprops=None,
247
revision_id=None):
248
"""Obtain a CommitBuilder for this repository.
249
250
:param branch: Branch to commit to.
251
:param parents: Revision ids of the parents of the new revision.
252
:param config: Configuration to use.
253
:param timestamp: Optional timestamp recorded for commit.
254
:param timezone: Optional timezone for timestamp.
255
:param committer: Optional committer to set for commit.
256
:param revprops: Optional dictionary of revision properties.
257
:param revision_id: Optional revision id.
258
"""
259
return CommitBuilder(self, parents, config, timestamp, timezone,
260
committer, revprops, revision_id)
261
262
def unlock(self):
263
self.control_files.unlock()
264
265
@needs_read_lock
266
def clone(self, a_bzrdir, revision_id=None, basis=None):
267
"""Clone this repository into a_bzrdir using the current format.
268
269
Currently no check is made that the format of this repository and
270
the bzrdir format are compatible. FIXME RBC 20060201.
271
"""
272
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
273
# use target default format.
274
result = a_bzrdir.create_repository()
275
# FIXME RBC 20060209 split out the repository type to avoid this check ?
276
elif isinstance(a_bzrdir._format,
277
(bzrdir.BzrDirFormat4,
278
bzrdir.BzrDirFormat5,
279
bzrdir.BzrDirFormat6)):
280
result = a_bzrdir.open_repository()
281
else:
282
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
283
self.copy_content_into(result, revision_id, basis)
284
return result
285
286
@needs_read_lock
287
def has_revision(self, revision_id):
288
"""True if this repository has a copy of the revision."""
289
return self._revision_store.has_revision_id(revision_id,
290
self.get_transaction())
291
292
@needs_read_lock
293
def get_revision_reconcile(self, revision_id):
294
"""'reconcile' helper routine that allows access to a revision always.
295
296
This variant of get_revision does not cross check the weave graph
297
against the revision one as get_revision does: but it should only
298
be used by reconcile, or reconcile-alike commands that are correcting
299
or testing the revision graph.
300
"""
301
if not revision_id or not isinstance(revision_id, basestring):
302
raise InvalidRevisionId(revision_id=revision_id, branch=self)
303
return self._revision_store.get_revisions([revision_id],
304
self.get_transaction())[0]
305
@needs_read_lock
306
def get_revisions(self, revision_ids):
307
return self._revision_store.get_revisions(revision_ids,
308
self.get_transaction())
309
310
@needs_read_lock
311
def get_revision_xml(self, revision_id):
312
rev = self.get_revision(revision_id)
313
rev_tmp = StringIO()
314
# the current serializer..
315
self._revision_store._serializer.write_revision(rev, rev_tmp)
316
rev_tmp.seek(0)
317
return rev_tmp.getvalue()
318
319
@needs_read_lock
320
def get_revision(self, revision_id):
321
"""Return the Revision object for a named revision"""
322
r = self.get_revision_reconcile(revision_id)
323
# weave corruption can lead to absent revision markers that should be
324
# present.
325
# the following test is reasonably cheap (it needs a single weave read)
326
# and the weave is cached in read transactions. In write transactions
327
# it is not cached but typically we only read a small number of
328
# revisions. For knits when they are introduced we will probably want
329
# to ensure that caching write transactions are in use.
330
inv = self.get_inventory_weave()
331
self._check_revision_parents(r, inv)
332
return r
333
334
@needs_read_lock
335
def get_deltas_for_revisions(self, revisions):
336
"""Produce a generator of revision deltas.
337
338
Note that the input is a sequence of REVISIONS, not revision_ids.
339
Trees will be held in memory until the generator exits.
340
Each delta is relative to the revision's lefthand predecessor.
341
"""
342
required_trees = set()
343
for revision in revisions:
344
required_trees.add(revision.revision_id)
345
required_trees.update(revision.parent_ids[:1])
346
trees = dict((t.get_revision_id(), t) for
347
t in self.revision_trees(required_trees))
348
for revision in revisions:
349
if not revision.parent_ids:
350
old_tree = EmptyTree()
351
else:
352
old_tree = trees[revision.parent_ids[0]]
353
yield delta.compare_trees(old_tree, trees[revision.revision_id])
354
355
@needs_read_lock
356
def get_revision_delta(self, revision_id):
357
"""Return the delta for one revision.
358
359
The delta is relative to the left-hand predecessor of the
360
revision.
361
"""
362
r = self.get_revision(revision_id)
363
return list(self.get_deltas_for_revisions([r]))[0]
364
365
def _check_revision_parents(self, revision, inventory):
366
"""Private to Repository and Fetch.
367
368
This checks the parentage of revision in an inventory weave for
369
consistency and is only applicable to inventory-weave-for-ancestry
370
using repository formats & fetchers.
371
"""
372
weave_parents = inventory.get_parents(revision.revision_id)
373
weave_names = inventory.versions()
374
for parent_id in revision.parent_ids:
375
if parent_id in weave_names:
376
# this parent must not be a ghost.
377
if not parent_id in weave_parents:
378
# but it is a ghost
379
raise errors.CorruptRepository(self)
380
381
@needs_write_lock
382
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
383
signature = gpg_strategy.sign(plaintext)
384
self._revision_store.add_revision_signature_text(revision_id,
385
signature,
386
self.get_transaction())
387
388
def fileids_altered_by_revision_ids(self, revision_ids):
389
"""Find the file ids and versions affected by revisions.
390
391
:param revisions: an iterable containing revision ids.
392
:return: a dictionary mapping altered file-ids to an iterable of
393
revision_ids. Each altered file-ids has the exact revision_ids that
394
altered it listed explicitly.
395
"""
396
assert isinstance(self._format, (RepositoryFormat5,
397
RepositoryFormat6,
398
RepositoryFormat7,
399
RepositoryFormatKnit1)), \
400
("fileids_altered_by_revision_ids only supported for branches "
401
"which store inventory as unnested xml, not on %r" % self)
402
selected_revision_ids = set(revision_ids)
403
w = self.get_inventory_weave()
404
result = {}
405
406
# this code needs to read every new line in every inventory for the
407
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
408
# not present in one of those inventories is unnecessary but not
409
# harmful because we are filtering by the revision id marker in the
410
# inventory lines : we only select file ids altered in one of those
411
# revisions. We don't need to see all lines in the inventory because
412
# only those added in an inventory in rev X can contain a revision=X
413
# line.
414
for line in w.iter_lines_added_or_present_in_versions(selected_revision_ids):
415
start = line.find('file_id="')+9
416
if start < 9: continue
417
end = line.find('"', start)
418
assert end>= 0
419
file_id = _unescape_xml(line[start:end])
420
421
start = line.find('revision="')+10
422
if start < 10: continue
423
end = line.find('"', start)
424
assert end>= 0
425
revision_id = _unescape_xml(line[start:end])
426
if revision_id in selected_revision_ids:
427
result.setdefault(file_id, set()).add(revision_id)
428
return result
429
430
@needs_read_lock
431
def get_inventory_weave(self):
432
return self.control_weaves.get_weave('inventory',
433
self.get_transaction())
434
435
@needs_read_lock
436
def get_inventory(self, revision_id):
437
"""Get Inventory object by hash."""
438
return self.deserialise_inventory(
439
revision_id, self.get_inventory_xml(revision_id))
440
441
def deserialise_inventory(self, revision_id, xml):
442
"""Transform the xml into an inventory object.
443
444
:param revision_id: The expected revision id of the inventory.
445
:param xml: A serialised inventory.
446
"""
447
return xml5.serializer_v5.read_inventory_from_string(xml)
448
449
@needs_read_lock
450
def get_inventory_xml(self, revision_id):
451
"""Get inventory XML as a file object."""
452
try:
453
assert isinstance(revision_id, basestring), type(revision_id)
454
iw = self.get_inventory_weave()
455
return iw.get_text(revision_id)
456
except IndexError:
457
raise errors.HistoryMissing(self, 'inventory', revision_id)
458
459
@needs_read_lock
460
def get_inventory_sha1(self, revision_id):
461
"""Return the sha1 hash of the inventory entry
462
"""
463
return self.get_revision(revision_id).inventory_sha1
464
465
@needs_read_lock
466
def get_revision_graph(self, revision_id=None):
467
"""Return a dictionary containing the revision graph.
468
469
:param revision_id: The revision_id to get a graph from. If None, then
470
the entire revision graph is returned. This is a deprecated mode of
471
operation and will be removed in the future.
472
:return: a dictionary of revision_id->revision_parents_list.
473
"""
474
# special case NULL_REVISION
475
if revision_id == NULL_REVISION:
476
return {}
477
weave = self.get_inventory_weave()
478
all_revisions = self._eliminate_revisions_not_present(weave.versions())
479
entire_graph = dict([(node, weave.get_parents(node)) for
480
node in all_revisions])
481
if revision_id is None:
482
return entire_graph
483
elif revision_id not in entire_graph:
484
raise errors.NoSuchRevision(self, revision_id)
485
else:
486
# add what can be reached from revision_id
487
result = {}
488
pending = set([revision_id])
489
while len(pending) > 0:
490
node = pending.pop()
491
result[node] = entire_graph[node]
492
for revision_id in result[node]:
493
if revision_id not in result:
494
pending.add(revision_id)
495
return result
496
497
@needs_read_lock
498
def get_revision_graph_with_ghosts(self, revision_ids=None):
499
"""Return a graph of the revisions with ghosts marked as applicable.
500
501
:param revision_ids: an iterable of revisions to graph or None for all.
502
:return: a Graph object with the graph reachable from revision_ids.
503
"""
504
result = Graph()
505
if not revision_ids:
506
pending = set(self.all_revision_ids())
507
required = set([])
508
else:
509
pending = set(revision_ids)
510
# special case NULL_REVISION
511
if NULL_REVISION in pending:
512
pending.remove(NULL_REVISION)
513
required = set(pending)
514
done = set([])
515
while len(pending):
516
revision_id = pending.pop()
517
try:
518
rev = self.get_revision(revision_id)
519
except errors.NoSuchRevision:
520
if revision_id in required:
521
raise
522
# a ghost
523
result.add_ghost(revision_id)
524
continue
525
for parent_id in rev.parent_ids:
526
# is this queued or done ?
527
if (parent_id not in pending and
528
parent_id not in done):
529
# no, queue it.
530
pending.add(parent_id)
531
result.add_node(revision_id, rev.parent_ids)
532
done.add(revision_id)
533
return result
534
535
@needs_read_lock
536
def get_revision_inventory(self, revision_id):
537
"""Return inventory of a past revision."""
538
# TODO: Unify this with get_inventory()
539
# bzr 0.0.6 and later imposes the constraint that the inventory_id
540
# must be the same as its revision, so this is trivial.
541
if revision_id is None:
542
# This does not make sense: if there is no revision,
543
# then it is the current tree inventory surely ?!
544
# and thus get_root_id() is something that looks at the last
545
# commit on the branch, and the get_root_id is an inventory check.
546
raise NotImplementedError
547
# return Inventory(self.get_root_id())
548
else:
549
return self.get_inventory(revision_id)
550
551
@needs_read_lock
552
def is_shared(self):
553
"""Return True if this repository is flagged as a shared repository."""
554
raise NotImplementedError(self.is_shared)
555
556
@needs_write_lock
557
def reconcile(self, other=None, thorough=False):
558
"""Reconcile this repository."""
559
from bzrlib.reconcile import RepoReconciler
560
reconciler = RepoReconciler(self, thorough=thorough)
561
reconciler.reconcile()
562
return reconciler
563
564
@needs_read_lock
565
def revision_tree(self, revision_id):
566
"""Return Tree for a revision on this branch.
567
568
`revision_id` may be None for the null revision, in which case
569
an `EmptyTree` is returned."""
570
# TODO: refactor this to use an existing revision object
571
# so we don't need to read it in twice.
572
if revision_id is None or revision_id == NULL_REVISION:
573
return EmptyTree()
574
else:
575
inv = self.get_revision_inventory(revision_id)
576
return RevisionTree(self, inv, revision_id)
577
578
@needs_read_lock
579
def revision_trees(self, revision_ids):
580
"""Return Tree for a revision on this branch.
581
582
`revision_id` may not be None or 'null:'"""
583
assert None not in revision_ids
584
assert NULL_REVISION not in revision_ids
585
texts = self.get_inventory_weave().get_texts(revision_ids)
586
for text, revision_id in zip(texts, revision_ids):
587
inv = self.deserialise_inventory(revision_id, text)
588
yield RevisionTree(self, inv, revision_id)
589
590
@needs_read_lock
591
def get_ancestry(self, revision_id):
592
"""Return a list of revision-ids integrated by a revision.
593
594
The first element of the list is always None, indicating the origin
595
revision. This might change when we have history horizons, or
596
perhaps we should have a new API.
597
598
This is topologically sorted.
599
"""
600
if revision_id is None:
601
return [None]
602
if not self.has_revision(revision_id):
603
raise errors.NoSuchRevision(self, revision_id)
604
w = self.get_inventory_weave()
605
candidates = w.get_ancestry(revision_id)
606
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
607
608
@needs_read_lock
609
def print_file(self, file, revision_id):
610
"""Print `file` to stdout.
611
612
FIXME RBC 20060125 as John Meinel points out this is a bad api
613
- it writes to stdout, it assumes that that is valid etc. Fix
614
by creating a new more flexible convenience function.
615
"""
616
tree = self.revision_tree(revision_id)
617
# use inventory as it was in that revision
618
file_id = tree.inventory.path2id(file)
619
if not file_id:
620
# TODO: jam 20060427 Write a test for this code path
621
# it had a bug in it, and was raising the wrong
622
# exception.
623
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
624
tree.print_file(file_id)
625
626
def get_transaction(self):
627
return self.control_files.get_transaction()
628
629
def revision_parents(self, revid):
630
return self.get_inventory_weave().parent_names(revid)
631
632
@needs_write_lock
633
def set_make_working_trees(self, new_value):
634
"""Set the policy flag for making working trees when creating branches.
635
636
This only applies to branches that use this repository.
637
638
The default is 'True'.
639
:param new_value: True to restore the default, False to disable making
640
working trees.
641
"""
642
raise NotImplementedError(self.set_make_working_trees)
643
644
def make_working_trees(self):
645
"""Returns the policy for making working trees on new branches."""
646
raise NotImplementedError(self.make_working_trees)
647
648
@needs_write_lock
649
def sign_revision(self, revision_id, gpg_strategy):
650
plaintext = Testament.from_revision(self, revision_id).as_short_text()
651
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
652
653
@needs_read_lock
654
def has_signature_for_revision_id(self, revision_id):
655
"""Query for a revision signature for revision_id in the repository."""
656
return self._revision_store.has_signature(revision_id,
657
self.get_transaction())
658
659
@needs_read_lock
660
def get_signature_text(self, revision_id):
661
"""Return the text for a signature."""
662
return self._revision_store.get_signature_text(revision_id,
663
self.get_transaction())
664
665
@needs_read_lock
666
def check(self, revision_ids):
667
"""Check consistency of all history of given revision_ids.
668
669
Different repository implementations should override _check().
670
671
:param revision_ids: A non-empty list of revision_ids whose ancestry
672
will be checked. Typically the last revision_id of a branch.
673
"""
674
if not revision_ids:
675
raise ValueError("revision_ids must be non-empty in %s.check"
676
% (self,))
677
return self._check(revision_ids)
678
679
def _check(self, revision_ids):
680
result = check.Check(self)
681
result.check()
682
return result
683
684
685
class AllInOneRepository(Repository):
686
"""Legacy support - the repository behaviour for all-in-one branches."""
687
688
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
689
# we reuse one control files instance.
690
dir_mode = a_bzrdir._control_files._dir_mode
691
file_mode = a_bzrdir._control_files._file_mode
692
693
def get_store(name, compressed=True, prefixed=False):
694
# FIXME: This approach of assuming stores are all entirely compressed
695
# or entirely uncompressed is tidy, but breaks upgrade from
696
# some existing branches where there's a mixture; we probably
697
# still want the option to look for both.
698
relpath = a_bzrdir._control_files._escape(name)
699
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
700
prefixed=prefixed, compressed=compressed,
701
dir_mode=dir_mode,
702
file_mode=file_mode)
703
#if self._transport.should_cache():
704
# cache_path = os.path.join(self.cache_root, name)
705
# os.mkdir(cache_path)
706
# store = bzrlib.store.CachedStore(store, cache_path)
707
return store
708
709
# not broken out yet because the controlweaves|inventory_store
710
# and text_store | weave_store bits are still different.
711
if isinstance(_format, RepositoryFormat4):
712
# cannot remove these - there is still no consistent api
713
# which allows access to this old info.
714
self.inventory_store = get_store('inventory-store')
715
text_store = get_store('text-store')
716
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
717
718
@needs_read_lock
719
def is_shared(self):
720
"""AllInOne repositories cannot be shared."""
721
return False
722
723
@needs_write_lock
724
def set_make_working_trees(self, new_value):
725
"""Set the policy flag for making working trees when creating branches.
726
727
This only applies to branches that use this repository.
728
729
The default is 'True'.
730
:param new_value: True to restore the default, False to disable making
731
working trees.
732
"""
733
raise NotImplementedError(self.set_make_working_trees)
734
735
def make_working_trees(self):
736
"""Returns the policy for making working trees on new branches."""
737
return True
738
739
740
def install_revision(repository, rev, revision_tree):
741
"""Install all revision data into a repository."""
742
present_parents = []
743
parent_trees = {}
744
for p_id in rev.parent_ids:
745
if repository.has_revision(p_id):
746
present_parents.append(p_id)
747
parent_trees[p_id] = repository.revision_tree(p_id)
748
else:
749
parent_trees[p_id] = EmptyTree()
750
751
inv = revision_tree.inventory
752
753
# Add the texts that are not already present
754
for path, ie in inv.iter_entries():
755
w = repository.weave_store.get_weave_or_empty(ie.file_id,
756
repository.get_transaction())
757
if ie.revision not in w:
758
text_parents = []
759
# FIXME: TODO: The following loop *may* be overlapping/duplicate
760
# with InventoryEntry.find_previous_heads(). if it is, then there
761
# is a latent bug here where the parents may have ancestors of each
762
# other. RBC, AB
763
for revision, tree in parent_trees.iteritems():
764
if ie.file_id not in tree:
765
continue
766
parent_id = tree.inventory[ie.file_id].revision
767
if parent_id in text_parents:
768
continue
769
text_parents.append(parent_id)
770
771
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
772
repository.get_transaction())
773
lines = revision_tree.get_file(ie.file_id).readlines()
774
vfile.add_lines(rev.revision_id, text_parents, lines)
775
try:
776
# install the inventory
777
repository.add_inventory(rev.revision_id, inv, present_parents)
778
except errors.RevisionAlreadyPresent:
779
pass
780
repository.add_revision(rev.revision_id, rev, inv)
781
782
783
class MetaDirRepository(Repository):
784
"""Repositories in the new meta-dir layout."""
785
786
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
787
super(MetaDirRepository, self).__init__(_format,
788
a_bzrdir,
789
control_files,
790
_revision_store,
791
control_store,
792
text_store)
793
794
dir_mode = self.control_files._dir_mode
795
file_mode = self.control_files._file_mode
796
797
@needs_read_lock
798
def is_shared(self):
799
"""Return True if this repository is flagged as a shared repository."""
800
return self.control_files._transport.has('shared-storage')
801
802
@needs_write_lock
803
def set_make_working_trees(self, new_value):
804
"""Set the policy flag for making working trees when creating branches.
805
806
This only applies to branches that use this repository.
807
808
The default is 'True'.
809
:param new_value: True to restore the default, False to disable making
810
working trees.
811
"""
812
if new_value:
813
try:
814
self.control_files._transport.delete('no-working-trees')
815
except errors.NoSuchFile:
816
pass
817
else:
818
self.control_files.put_utf8('no-working-trees', '')
819
820
def make_working_trees(self):
821
"""Returns the policy for making working trees on new branches."""
822
return not self.control_files._transport.has('no-working-trees')
823
824
825
class KnitRepository(MetaDirRepository):
826
"""Knit format repository."""
827
828
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
829
inv_vf.add_lines_with_ghosts(revid, parents, lines)
830
831
@needs_read_lock
832
def _all_revision_ids(self):
833
"""See Repository.all_revision_ids()."""
834
# Knits get the revision graph from the index of the revision knit, so
835
# it's always possible even if they're on an unlistable transport.
836
return self._revision_store.all_revision_ids(self.get_transaction())
837
838
def fileid_involved_between_revs(self, from_revid, to_revid):
839
"""Find file_id(s) which are involved in the changes between revisions.
840
841
This determines the set of revisions which are involved, and then
842
finds all file ids affected by those revisions.
843
"""
844
vf = self._get_revision_vf()
845
from_set = set(vf.get_ancestry(from_revid))
846
to_set = set(vf.get_ancestry(to_revid))
847
changed = to_set.difference(from_set)
848
return self._fileid_involved_by_set(changed)
849
850
def fileid_involved(self, last_revid=None):
851
"""Find all file_ids modified in the ancestry of last_revid.
852
853
:param last_revid: If None, last_revision() will be used.
854
"""
855
if not last_revid:
856
changed = set(self.all_revision_ids())
857
else:
858
changed = set(self.get_ancestry(last_revid))
859
if None in changed:
860
changed.remove(None)
861
return self._fileid_involved_by_set(changed)
862
863
@needs_read_lock
864
def get_ancestry(self, revision_id):
865
"""Return a list of revision-ids integrated by a revision.
866
867
This is topologically sorted.
868
"""
869
if revision_id is None:
870
return [None]
871
vf = self._get_revision_vf()
872
try:
873
return [None] + vf.get_ancestry(revision_id)
874
except errors.RevisionNotPresent:
875
raise errors.NoSuchRevision(self, revision_id)
876
877
@needs_read_lock
878
def get_revision(self, revision_id):
879
"""Return the Revision object for a named revision"""
880
return self.get_revision_reconcile(revision_id)
881
882
@needs_read_lock
883
def get_revision_graph(self, revision_id=None):
884
"""Return a dictionary containing the revision graph.
885
886
:param revision_id: The revision_id to get a graph from. If None, then
887
the entire revision graph is returned. This is a deprecated mode of
888
operation and will be removed in the future.
889
:return: a dictionary of revision_id->revision_parents_list.
890
"""
891
# special case NULL_REVISION
892
if revision_id == NULL_REVISION:
893
return {}
894
weave = self._get_revision_vf()
895
entire_graph = weave.get_graph()
896
if revision_id is None:
897
return weave.get_graph()
898
elif revision_id not in weave:
899
raise errors.NoSuchRevision(self, revision_id)
900
else:
901
# add what can be reached from revision_id
902
result = {}
903
pending = set([revision_id])
904
while len(pending) > 0:
905
node = pending.pop()
906
result[node] = weave.get_parents(node)
907
for revision_id in result[node]:
908
if revision_id not in result:
909
pending.add(revision_id)
910
return result
911
912
@needs_read_lock
913
def get_revision_graph_with_ghosts(self, revision_ids=None):
914
"""Return a graph of the revisions with ghosts marked as applicable.
915
916
:param revision_ids: an iterable of revisions to graph or None for all.
917
:return: a Graph object with the graph reachable from revision_ids.
918
"""
919
result = Graph()
920
vf = self._get_revision_vf()
921
versions = set(vf.versions())
922
if not revision_ids:
923
pending = set(self.all_revision_ids())
924
required = set([])
925
else:
926
pending = set(revision_ids)
927
# special case NULL_REVISION
928
if NULL_REVISION in pending:
929
pending.remove(NULL_REVISION)
930
required = set(pending)
931
done = set([])
932
while len(pending):
933
revision_id = pending.pop()
934
if not revision_id in versions:
935
if revision_id in required:
936
raise errors.NoSuchRevision(self, revision_id)
937
# a ghost
938
result.add_ghost(revision_id)
939
# mark it as done so we don't try for it again.
940
done.add(revision_id)
941
continue
942
parent_ids = vf.get_parents_with_ghosts(revision_id)
943
for parent_id in parent_ids:
944
# is this queued or done ?
945
if (parent_id not in pending and
946
parent_id not in done):
947
# no, queue it.
948
pending.add(parent_id)
949
result.add_node(revision_id, parent_ids)
950
done.add(revision_id)
951
return result
952
953
def _get_revision_vf(self):
954
""":return: a versioned file containing the revisions."""
955
vf = self._revision_store.get_revision_file(self.get_transaction())
956
return vf
957
958
@needs_write_lock
959
def reconcile(self, other=None, thorough=False):
960
"""Reconcile this repository."""
961
from bzrlib.reconcile import KnitReconciler
962
reconciler = KnitReconciler(self, thorough=thorough)
963
reconciler.reconcile()
964
return reconciler
965
966
def revision_parents(self, revision_id):
967
return self._get_revision_vf().get_parents(revision_id)
968
969
970
class RepositoryFormat(object):
971
"""A repository format.
972
973
Formats provide three things:
974
* An initialization routine to construct repository data on disk.
975
* a format string which is used when the BzrDir supports versioned
976
children.
977
* an open routine which returns a Repository instance.
978
979
Formats are placed in an dict by their format string for reference
980
during opening. These should be subclasses of RepositoryFormat
981
for consistency.
982
983
Once a format is deprecated, just deprecate the initialize and open
984
methods on the format class. Do not deprecate the object, as the
985
object will be created every system load.
986
987
Common instance attributes:
988
_matchingbzrdir - the bzrdir format that the repository format was
989
originally written to work with. This can be used if manually
990
constructing a bzrdir and repository, or more commonly for test suite
991
parameterisation.
992
"""
993
994
_default_format = None
995
"""The default format used for new repositories."""
996
997
_formats = {}
998
"""The known formats."""
999
1000
@classmethod
1001
def find_format(klass, a_bzrdir):
1002
"""Return the format for the repository object in a_bzrdir."""
1003
try:
1004
transport = a_bzrdir.get_repository_transport(None)
1005
format_string = transport.get("format").read()
1006
return klass._formats[format_string]
1007
except errors.NoSuchFile:
1008
raise errors.NoRepositoryPresent(a_bzrdir)
1009
except KeyError:
1010
raise errors.UnknownFormatError(format=format_string)
1011
1012
def _get_control_store(self, repo_transport, control_files):
1013
"""Return the control store for this repository."""
1014
raise NotImplementedError(self._get_control_store)
1015
1016
@classmethod
1017
def get_default_format(klass):
1018
"""Return the current default format."""
1019
return klass._default_format
1020
1021
def get_format_string(self):
1022
"""Return the ASCII format string that identifies this format.
1023
1024
Note that in pre format ?? repositories the format string is
1025
not permitted nor written to disk.
1026
"""
1027
raise NotImplementedError(self.get_format_string)
1028
1029
def get_format_description(self):
1030
"""Return the short description for this format."""
1031
raise NotImplementedError(self.get_format_description)
1032
1033
def _get_revision_store(self, repo_transport, control_files):
1034
"""Return the revision store object for this a_bzrdir."""
1035
raise NotImplementedError(self._get_revision_store)
1036
1037
def _get_text_rev_store(self,
1038
transport,
1039
control_files,
1040
name,
1041
compressed=True,
1042
prefixed=False,
1043
serializer=None):
1044
"""Common logic for getting a revision store for a repository.
1045
1046
see self._get_revision_store for the subclass-overridable method to
1047
get the store for a repository.
1048
"""
1049
from bzrlib.store.revision.text import TextRevisionStore
1050
dir_mode = control_files._dir_mode
1051
file_mode = control_files._file_mode
1052
text_store =TextStore(transport.clone(name),
1053
prefixed=prefixed,
1054
compressed=compressed,
1055
dir_mode=dir_mode,
1056
file_mode=file_mode)
1057
_revision_store = TextRevisionStore(text_store, serializer)
1058
return _revision_store
1059
1060
def _get_versioned_file_store(self,
1061
name,
1062
transport,
1063
control_files,
1064
prefixed=True,
1065
versionedfile_class=WeaveFile,
1066
escaped=False):
1067
weave_transport = control_files._transport.clone(name)
1068
dir_mode = control_files._dir_mode
1069
file_mode = control_files._file_mode
1070
return VersionedFileStore(weave_transport, prefixed=prefixed,
1071
dir_mode=dir_mode,
1072
file_mode=file_mode,
1073
versionedfile_class=versionedfile_class,
1074
escaped=escaped)
1075
1076
def initialize(self, a_bzrdir, shared=False):
1077
"""Initialize a repository of this format in a_bzrdir.
1078
1079
:param a_bzrdir: The bzrdir to put the new repository in it.
1080
:param shared: The repository should be initialized as a sharable one.
1081
1082
This may raise UninitializableFormat if shared repository are not
1083
compatible the a_bzrdir.
1084
"""
1085
1086
def is_supported(self):
1087
"""Is this format supported?
1088
1089
Supported formats must be initializable and openable.
1090
Unsupported formats may not support initialization or committing or
1091
some other features depending on the reason for not being supported.
1092
"""
1093
return True
1094
1095
def open(self, a_bzrdir, _found=False):
1096
"""Return an instance of this format for the bzrdir a_bzrdir.
1097
1098
_found is a private parameter, do not use it.
1099
"""
1100
raise NotImplementedError(self.open)
1101
1102
@classmethod
1103
def register_format(klass, format):
1104
klass._formats[format.get_format_string()] = format
1105
1106
@classmethod
1107
def set_default_format(klass, format):
1108
klass._default_format = format
1109
1110
@classmethod
1111
def unregister_format(klass, format):
1112
assert klass._formats[format.get_format_string()] is format
1113
del klass._formats[format.get_format_string()]
1114
1115
1116
class PreSplitOutRepositoryFormat(RepositoryFormat):
1117
"""Base class for the pre split out repository formats."""
1118
1119
def initialize(self, a_bzrdir, shared=False, _internal=False):
1120
"""Create a weave repository.
1121
1122
TODO: when creating split out bzr branch formats, move this to a common
1123
base for Format5, Format6. or something like that.
1124
"""
1125
from bzrlib.weavefile import write_weave_v5
1126
from bzrlib.weave import Weave
1127
1128
if shared:
1129
raise errors.IncompatibleFormat(self, a_bzrdir._format)
1130
1131
if not _internal:
1132
# always initialized when the bzrdir is.
1133
return self.open(a_bzrdir, _found=True)
1134
1135
# Create an empty weave
1136
sio = StringIO()
1137
write_weave_v5(Weave(), sio)
1138
empty_weave = sio.getvalue()
1139
1140
mutter('creating repository in %s.', a_bzrdir.transport.base)
1141
dirs = ['revision-store', 'weaves']
1142
files = [('inventory.weave', StringIO(empty_weave)),
1143
]
1144
1145
# FIXME: RBC 20060125 don't peek under the covers
1146
# NB: no need to escape relative paths that are url safe.
1147
control_files = LockableFiles(a_bzrdir.transport, 'branch-lock',
1148
TransportLock)
1149
control_files.create_lock()
1150
control_files.lock_write()
1151
control_files._transport.mkdir_multi(dirs,
1152
mode=control_files._dir_mode)
1153
try:
1154
for file, content in files:
1155
control_files.put(file, content)
1156
finally:
1157
control_files.unlock()
1158
return self.open(a_bzrdir, _found=True)
1159
1160
def _get_control_store(self, repo_transport, control_files):
1161
"""Return the control store for this repository."""
1162
return self._get_versioned_file_store('',
1163
repo_transport,
1164
control_files,
1165
prefixed=False)
1166
1167
def _get_text_store(self, transport, control_files):
1168
"""Get a store for file texts for this format."""
1169
raise NotImplementedError(self._get_text_store)
1170
1171
def open(self, a_bzrdir, _found=False):
1172
"""See RepositoryFormat.open()."""
1173
if not _found:
1174
# we are being called directly and must probe.
1175
raise NotImplementedError
1176
1177
repo_transport = a_bzrdir.get_repository_transport(None)
1178
control_files = a_bzrdir._control_files
1179
text_store = self._get_text_store(repo_transport, control_files)
1180
control_store = self._get_control_store(repo_transport, control_files)
1181
_revision_store = self._get_revision_store(repo_transport, control_files)
1182
return AllInOneRepository(_format=self,
1183
a_bzrdir=a_bzrdir,
1184
_revision_store=_revision_store,
1185
control_store=control_store,
1186
text_store=text_store)
1187
1188
1189
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1190
"""Bzr repository format 4.
1191
1192
This repository format has:
1193
- flat stores
1194
- TextStores for texts, inventories,revisions.
1195
1196
This format is deprecated: it indexes texts using a text id which is
1197
removed in format 5; initialization and write support for this format
1198
has been removed.
1199
"""
1200
1201
def __init__(self):
1202
super(RepositoryFormat4, self).__init__()
1203
self._matchingbzrdir = bzrdir.BzrDirFormat4()
1204
1205
def get_format_description(self):
1206
"""See RepositoryFormat.get_format_description()."""
1207
return "Repository format 4"
1208
1209
def initialize(self, url, shared=False, _internal=False):
1210
"""Format 4 branches cannot be created."""
1211
raise errors.UninitializableFormat(self)
1212
1213
def is_supported(self):
1214
"""Format 4 is not supported.
1215
1216
It is not supported because the model changed from 4 to 5 and the
1217
conversion logic is expensive - so doing it on the fly was not
1218
feasible.
1219
"""
1220
return False
1221
1222
def _get_control_store(self, repo_transport, control_files):
1223
"""Format 4 repositories have no formal control store at this point.
1224
1225
This will cause any control-file-needing apis to fail - this is desired.
1226
"""
1227
return None
1228
1229
def _get_revision_store(self, repo_transport, control_files):
1230
"""See RepositoryFormat._get_revision_store()."""
1231
from bzrlib.xml4 import serializer_v4
1232
return self._get_text_rev_store(repo_transport,
1233
control_files,
1234
'revision-store',
1235
serializer=serializer_v4)
1236
1237
def _get_text_store(self, transport, control_files):
1238
"""See RepositoryFormat._get_text_store()."""
1239
1240
1241
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1242
"""Bzr control format 5.
1243
1244
This repository format has:
1245
- weaves for file texts and inventory
1246
- flat stores
1247
- TextStores for revisions and signatures.
1248
"""
1249
1250
def __init__(self):
1251
super(RepositoryFormat5, self).__init__()
1252
self._matchingbzrdir = bzrdir.BzrDirFormat5()
1253
1254
def get_format_description(self):
1255
"""See RepositoryFormat.get_format_description()."""
1256
return "Weave repository format 5"
1257
1258
def _get_revision_store(self, repo_transport, control_files):
1259
"""See RepositoryFormat._get_revision_store()."""
1260
"""Return the revision store object for this a_bzrdir."""
1261
return self._get_text_rev_store(repo_transport,
1262
control_files,
1263
'revision-store',
1264
compressed=False)
1265
1266
def _get_text_store(self, transport, control_files):
1267
"""See RepositoryFormat._get_text_store()."""
1268
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1269
1270
1271
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1272
"""Bzr control format 6.
1273
1274
This repository format has:
1275
- weaves for file texts and inventory
1276
- hash subdirectory based stores.
1277
- TextStores for revisions and signatures.
1278
"""
1279
1280
def __init__(self):
1281
super(RepositoryFormat6, self).__init__()
1282
self._matchingbzrdir = bzrdir.BzrDirFormat6()
1283
1284
def get_format_description(self):
1285
"""See RepositoryFormat.get_format_description()."""
1286
return "Weave repository format 6"
1287
1288
def _get_revision_store(self, repo_transport, control_files):
1289
"""See RepositoryFormat._get_revision_store()."""
1290
return self._get_text_rev_store(repo_transport,
1291
control_files,
1292
'revision-store',
1293
compressed=False,
1294
prefixed=True)
1295
1296
def _get_text_store(self, transport, control_files):
1297
"""See RepositoryFormat._get_text_store()."""
1298
return self._get_versioned_file_store('weaves', transport, control_files)
1299
1300
1301
class MetaDirRepositoryFormat(RepositoryFormat):
1302
"""Common base class for the new repositories using the metadir layout."""
1303
1304
def __init__(self):
1305
super(MetaDirRepositoryFormat, self).__init__()
1306
self._matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1307
1308
def _create_control_files(self, a_bzrdir):
1309
"""Create the required files and the initial control_files object."""
1310
# FIXME: RBC 20060125 don't peek under the covers
1311
# NB: no need to escape relative paths that are url safe.
1312
repository_transport = a_bzrdir.get_repository_transport(self)
1313
control_files = LockableFiles(repository_transport, 'lock', LockDir)
1314
control_files.create_lock()
1315
return control_files
1316
1317
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1318
"""Upload the initial blank content."""
1319
control_files = self._create_control_files(a_bzrdir)
1320
control_files.lock_write()
1321
try:
1322
control_files._transport.mkdir_multi(dirs,
1323
mode=control_files._dir_mode)
1324
for file, content in files:
1325
control_files.put(file, content)
1326
for file, content in utf8_files:
1327
control_files.put_utf8(file, content)
1328
if shared == True:
1329
control_files.put_utf8('shared-storage', '')
1330
finally:
1331
control_files.unlock()
1332
1333
1334
class RepositoryFormat7(MetaDirRepositoryFormat):
1335
"""Bzr repository 7.
1336
1337
This repository format has:
1338
- weaves for file texts and inventory
1339
- hash subdirectory based stores.
1340
- TextStores for revisions and signatures.
1341
- a format marker of its own
1342
- an optional 'shared-storage' flag
1343
- an optional 'no-working-trees' flag
1344
"""
1345
1346
def _get_control_store(self, repo_transport, control_files):
1347
"""Return the control store for this repository."""
1348
return self._get_versioned_file_store('',
1349
repo_transport,
1350
control_files,
1351
prefixed=False)
1352
1353
def get_format_string(self):
1354
"""See RepositoryFormat.get_format_string()."""
1355
return "Bazaar-NG Repository format 7"
1356
1357
def get_format_description(self):
1358
"""See RepositoryFormat.get_format_description()."""
1359
return "Weave repository format 7"
1360
1361
def _get_revision_store(self, repo_transport, control_files):
1362
"""See RepositoryFormat._get_revision_store()."""
1363
return self._get_text_rev_store(repo_transport,
1364
control_files,
1365
'revision-store',
1366
compressed=False,
1367
prefixed=True,
1368
)
1369
1370
def _get_text_store(self, transport, control_files):
1371
"""See RepositoryFormat._get_text_store()."""
1372
return self._get_versioned_file_store('weaves',
1373
transport,
1374
control_files)
1375
1376
def initialize(self, a_bzrdir, shared=False):
1377
"""Create a weave repository.
1378
1379
:param shared: If true the repository will be initialized as a shared
1380
repository.
1381
"""
1382
from bzrlib.weavefile import write_weave_v5
1383
from bzrlib.weave import Weave
1384
1385
# Create an empty weave
1386
sio = StringIO()
1387
write_weave_v5(Weave(), sio)
1388
empty_weave = sio.getvalue()
1389
1390
mutter('creating repository in %s.', a_bzrdir.transport.base)
1391
dirs = ['revision-store', 'weaves']
1392
files = [('inventory.weave', StringIO(empty_weave)),
1393
]
1394
utf8_files = [('format', self.get_format_string())]
1395
1396
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1397
return self.open(a_bzrdir=a_bzrdir, _found=True)
1398
1399
def open(self, a_bzrdir, _found=False, _override_transport=None):
1400
"""See RepositoryFormat.open().
1401
1402
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1403
repository at a slightly different url
1404
than normal. I.e. during 'upgrade'.
1405
"""
1406
if not _found:
1407
format = RepositoryFormat.find_format(a_bzrdir)
1408
assert format.__class__ == self.__class__
1409
if _override_transport is not None:
1410
repo_transport = _override_transport
1411
else:
1412
repo_transport = a_bzrdir.get_repository_transport(None)
1413
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1414
text_store = self._get_text_store(repo_transport, control_files)
1415
control_store = self._get_control_store(repo_transport, control_files)
1416
_revision_store = self._get_revision_store(repo_transport, control_files)
1417
return MetaDirRepository(_format=self,
1418
a_bzrdir=a_bzrdir,
1419
control_files=control_files,
1420
_revision_store=_revision_store,
1421
control_store=control_store,
1422
text_store=text_store)
1423
1424
1425
class RepositoryFormatKnit1(MetaDirRepositoryFormat):
1426
"""Bzr repository knit format 1.
1427
1428
This repository format has:
1429
- knits for file texts and inventory
1430
- hash subdirectory based stores.
1431
- knits for revisions and signatures
1432
- TextStores for revisions and signatures.
1433
- a format marker of its own
1434
- an optional 'shared-storage' flag
1435
- an optional 'no-working-trees' flag
1436
- a LockDir lock
1437
1438
This format was introduced in bzr 0.8.
1439
"""
1440
1441
def _get_control_store(self, repo_transport, control_files):
1442
"""Return the control store for this repository."""
1443
return VersionedFileStore(
1444
repo_transport,
1445
prefixed=False,
1446
file_mode=control_files._file_mode,
1447
versionedfile_class=KnitVersionedFile,
1448
versionedfile_kwargs={'factory':KnitPlainFactory()},
1449
)
1450
1451
def get_format_string(self):
1452
"""See RepositoryFormat.get_format_string()."""
1453
return "Bazaar-NG Knit Repository Format 1"
1454
1455
def get_format_description(self):
1456
"""See RepositoryFormat.get_format_description()."""
1457
return "Knit repository format 1"
1458
1459
def _get_revision_store(self, repo_transport, control_files):
1460
"""See RepositoryFormat._get_revision_store()."""
1461
from bzrlib.store.revision.knit import KnitRevisionStore
1462
versioned_file_store = VersionedFileStore(
1463
repo_transport,
1464
file_mode=control_files._file_mode,
1465
prefixed=False,
1466
precious=True,
1467
versionedfile_class=KnitVersionedFile,
1468
versionedfile_kwargs={'delta':False, 'factory':KnitPlainFactory()},
1469
escaped=True,
1470
)
1471
return KnitRevisionStore(versioned_file_store)
1472
1473
def _get_text_store(self, transport, control_files):
1474
"""See RepositoryFormat._get_text_store()."""
1475
return self._get_versioned_file_store('knits',
1476
transport,
1477
control_files,
1478
versionedfile_class=KnitVersionedFile,
1479
escaped=True)
1480
1481
def initialize(self, a_bzrdir, shared=False):
1482
"""Create a knit format 1 repository.
1483
1484
:param a_bzrdir: bzrdir to contain the new repository; must already
1485
be initialized.
1486
:param shared: If true the repository will be initialized as a shared
1487
repository.
1488
"""
1489
mutter('creating repository in %s.', a_bzrdir.transport.base)
1490
dirs = ['revision-store', 'knits']
1491
files = []
1492
utf8_files = [('format', self.get_format_string())]
1493
1494
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1495
repo_transport = a_bzrdir.get_repository_transport(None)
1496
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1497
control_store = self._get_control_store(repo_transport, control_files)
1498
transaction = transactions.WriteTransaction()
1499
# trigger a write of the inventory store.
1500
control_store.get_weave_or_empty('inventory', transaction)
1501
_revision_store = self._get_revision_store(repo_transport, control_files)
1502
_revision_store.has_revision_id('A', transaction)
1503
_revision_store.get_signature_file(transaction)
1504
return self.open(a_bzrdir=a_bzrdir, _found=True)
1505
1506
def open(self, a_bzrdir, _found=False, _override_transport=None):
1507
"""See RepositoryFormat.open().
1508
1509
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1510
repository at a slightly different url
1511
than normal. I.e. during 'upgrade'.
1512
"""
1513
if not _found:
1514
format = RepositoryFormat.find_format(a_bzrdir)
1515
assert format.__class__ == self.__class__
1516
if _override_transport is not None:
1517
repo_transport = _override_transport
1518
else:
1519
repo_transport = a_bzrdir.get_repository_transport(None)
1520
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1521
text_store = self._get_text_store(repo_transport, control_files)
1522
control_store = self._get_control_store(repo_transport, control_files)
1523
_revision_store = self._get_revision_store(repo_transport, control_files)
1524
return KnitRepository(_format=self,
1525
a_bzrdir=a_bzrdir,
1526
control_files=control_files,
1527
_revision_store=_revision_store,
1528
control_store=control_store,
1529
text_store=text_store)
1530
1531
1532
# formats which have no format string are not discoverable
1533
# and not independently creatable, so are not registered.
1534
RepositoryFormat.register_format(RepositoryFormat7())
1535
_default_format = RepositoryFormatKnit1()
1536
RepositoryFormat.register_format(_default_format)
1537
RepositoryFormat.set_default_format(_default_format)
1538
_legacy_formats = [RepositoryFormat4(),
1539
RepositoryFormat5(),
1540
RepositoryFormat6()]
1541
1542
1543
class InterRepository(InterObject):
1544
"""This class represents operations taking place between two repositories.
1545
1546
Its instances have methods like copy_content and fetch, and contain
1547
references to the source and target repositories these operations can be
1548
carried out on.
1549
1550
Often we will provide convenience methods on 'repository' which carry out
1551
operations with another repository - they will always forward to
1552
InterRepository.get(other).method_name(parameters).
1553
"""
1554
1555
_optimisers = set()
1556
"""The available optimised InterRepository types."""
1557
1558
@needs_write_lock
1559
def copy_content(self, revision_id=None, basis=None):
1560
"""Make a complete copy of the content in self into destination.
1561
1562
This is a destructive operation! Do not use it on existing
1563
repositories.
1564
1565
:param revision_id: Only copy the content needed to construct
1566
revision_id and its parents.
1567
:param basis: Copy the needed data preferentially from basis.
1568
"""
1569
try:
1570
self.target.set_make_working_trees(self.source.make_working_trees())
1571
except NotImplementedError:
1572
pass
1573
# grab the basis available data
1574
if basis is not None:
1575
self.target.fetch(basis, revision_id=revision_id)
1576
# but don't bother fetching if we have the needed data now.
1577
if (revision_id not in (None, NULL_REVISION) and
1578
self.target.has_revision(revision_id)):
1579
return
1580
self.target.fetch(self.source, revision_id=revision_id)
1581
1582
def _double_lock(self, lock_source, lock_target):
1583
"""Take out too locks, rolling back the first if the second throws."""
1584
lock_source()
1585
try:
1586
lock_target()
1587
except Exception:
1588
# we want to ensure that we don't leave source locked by mistake.
1589
# and any error on target should not confuse source.
1590
self.source.unlock()
1591
raise
1592
1593
@needs_write_lock
1594
def fetch(self, revision_id=None, pb=None):
1595
"""Fetch the content required to construct revision_id.
1596
1597
The content is copied from source to target.
1598
1599
:param revision_id: if None all content is copied, if NULL_REVISION no
1600
content is copied.
1601
:param pb: optional progress bar to use for progress reports. If not
1602
provided a default one will be created.
1603
1604
Returns the copied revision count and the failed revisions in a tuple:
1605
(copied, failures).
1606
"""
1607
from bzrlib.fetch import GenericRepoFetcher
1608
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1609
self.source, self.source._format, self.target, self.target._format)
1610
f = GenericRepoFetcher(to_repository=self.target,
1611
from_repository=self.source,
1612
last_revision=revision_id,
1613
pb=pb)
1614
return f.count_copied, f.failed_revisions
1615
1616
def lock_read(self):
1617
"""Take out a logical read lock.
1618
1619
This will lock the source branch and the target branch. The source gets
1620
a read lock and the target a read lock.
1621
"""
1622
self._double_lock(self.source.lock_read, self.target.lock_read)
1623
1624
def lock_write(self):
1625
"""Take out a logical write lock.
1626
1627
This will lock the source branch and the target branch. The source gets
1628
a read lock and the target a write lock.
1629
"""
1630
self._double_lock(self.source.lock_read, self.target.lock_write)
1631
1632
@needs_read_lock
1633
def missing_revision_ids(self, revision_id=None):
1634
"""Return the revision ids that source has that target does not.
1635
1636
These are returned in topological order.
1637
1638
:param revision_id: only return revision ids included by this
1639
revision_id.
1640
"""
1641
# generic, possibly worst case, slow code path.
1642
target_ids = set(self.target.all_revision_ids())
1643
if revision_id is not None:
1644
source_ids = self.source.get_ancestry(revision_id)
1645
assert source_ids[0] == None
1646
source_ids.pop(0)
1647
else:
1648
source_ids = self.source.all_revision_ids()
1649
result_set = set(source_ids).difference(target_ids)
1650
# this may look like a no-op: its not. It preserves the ordering
1651
# other_ids had while only returning the members from other_ids
1652
# that we've decided we need.
1653
return [rev_id for rev_id in source_ids if rev_id in result_set]
1654
1655
def unlock(self):
1656
"""Release the locks on source and target."""
1657
try:
1658
self.target.unlock()
1659
finally:
1660
self.source.unlock()
1661
1662
1663
class InterWeaveRepo(InterRepository):
1664
"""Optimised code paths between Weave based repositories."""
1665
1666
_matching_repo_format = RepositoryFormat7()
1667
"""Repository format for testing with."""
1668
1669
@staticmethod
1670
def is_compatible(source, target):
1671
"""Be compatible with known Weave formats.
1672
1673
We don't test for the stores being of specific types because that
1674
could lead to confusing results, and there is no need to be
1675
overly general.
1676
"""
1677
try:
1678
return (isinstance(source._format, (RepositoryFormat5,
1679
RepositoryFormat6,
1680
RepositoryFormat7)) and
1681
isinstance(target._format, (RepositoryFormat5,
1682
RepositoryFormat6,
1683
RepositoryFormat7)))
1684
except AttributeError:
1685
return False
1686
1687
@needs_write_lock
1688
def copy_content(self, revision_id=None, basis=None):
1689
"""See InterRepository.copy_content()."""
1690
# weave specific optimised path:
1691
if basis is not None:
1692
# copy the basis in, then fetch remaining data.
1693
basis.copy_content_into(self.target, revision_id)
1694
# the basis copy_content_into could miss-set this.
1695
try:
1696
self.target.set_make_working_trees(self.source.make_working_trees())
1697
except NotImplementedError:
1698
pass
1699
self.target.fetch(self.source, revision_id=revision_id)
1700
else:
1701
try:
1702
self.target.set_make_working_trees(self.source.make_working_trees())
1703
except NotImplementedError:
1704
pass
1705
# FIXME do not peek!
1706
if self.source.control_files._transport.listable():
1707
pb = ui.ui_factory.nested_progress_bar()
1708
try:
1709
self.target.weave_store.copy_all_ids(
1710
self.source.weave_store,
1711
pb=pb,
1712
from_transaction=self.source.get_transaction(),
1713
to_transaction=self.target.get_transaction())
1714
pb.update('copying inventory', 0, 1)
1715
self.target.control_weaves.copy_multi(
1716
self.source.control_weaves, ['inventory'],
1717
from_transaction=self.source.get_transaction(),
1718
to_transaction=self.target.get_transaction())
1719
self.target._revision_store.text_store.copy_all_ids(
1720
self.source._revision_store.text_store,
1721
pb=pb)
1722
finally:
1723
pb.finished()
1724
else:
1725
self.target.fetch(self.source, revision_id=revision_id)
1726
1727
@needs_write_lock
1728
def fetch(self, revision_id=None, pb=None):
1729
"""See InterRepository.fetch()."""
1730
from bzrlib.fetch import GenericRepoFetcher
1731
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1732
self.source, self.source._format, self.target, self.target._format)
1733
f = GenericRepoFetcher(to_repository=self.target,
1734
from_repository=self.source,
1735
last_revision=revision_id,
1736
pb=pb)
1737
return f.count_copied, f.failed_revisions
1738
1739
@needs_read_lock
1740
def missing_revision_ids(self, revision_id=None):
1741
"""See InterRepository.missing_revision_ids()."""
1742
# we want all revisions to satisfy revision_id in source.
1743
# but we don't want to stat every file here and there.
1744
# we want then, all revisions other needs to satisfy revision_id
1745
# checked, but not those that we have locally.
1746
# so the first thing is to get a subset of the revisions to
1747
# satisfy revision_id in source, and then eliminate those that
1748
# we do already have.
1749
# this is slow on high latency connection to self, but as as this
1750
# disk format scales terribly for push anyway due to rewriting
1751
# inventory.weave, this is considered acceptable.
1752
# - RBC 20060209
1753
if revision_id is not None:
1754
source_ids = self.source.get_ancestry(revision_id)
1755
assert source_ids[0] == None
1756
source_ids.pop(0)
1757
else:
1758
source_ids = self.source._all_possible_ids()
1759
source_ids_set = set(source_ids)
1760
# source_ids is the worst possible case we may need to pull.
1761
# now we want to filter source_ids against what we actually
1762
# have in target, but don't try to check for existence where we know
1763
# we do not have a revision as that would be pointless.
1764
target_ids = set(self.target._all_possible_ids())
1765
possibly_present_revisions = target_ids.intersection(source_ids_set)
1766
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1767
required_revisions = source_ids_set.difference(actually_present_revisions)
1768
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1769
if revision_id is not None:
1770
# we used get_ancestry to determine source_ids then we are assured all
1771
# revisions referenced are present as they are installed in topological order.
1772
# and the tip revision was validated by get_ancestry.
1773
return required_topo_revisions
1774
else:
1775
# if we just grabbed the possibly available ids, then
1776
# we only have an estimate of whats available and need to validate
1777
# that against the revision records.
1778
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1779
1780
1781
class InterKnitRepo(InterRepository):
1782
"""Optimised code paths between Knit based repositories."""
1783
1784
_matching_repo_format = RepositoryFormatKnit1()
1785
"""Repository format for testing with."""
1786
1787
@staticmethod
1788
def is_compatible(source, target):
1789
"""Be compatible with known Knit formats.
1790
1791
We don't test for the stores being of specific types because that
1792
could lead to confusing results, and there is no need to be
1793
overly general.
1794
"""
1795
try:
1796
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1797
isinstance(target._format, (RepositoryFormatKnit1)))
1798
except AttributeError:
1799
return False
1800
1801
@needs_write_lock
1802
def fetch(self, revision_id=None, pb=None):
1803
"""See InterRepository.fetch()."""
1804
from bzrlib.fetch import KnitRepoFetcher
1805
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1806
self.source, self.source._format, self.target, self.target._format)
1807
f = KnitRepoFetcher(to_repository=self.target,
1808
from_repository=self.source,
1809
last_revision=revision_id,
1810
pb=pb)
1811
return f.count_copied, f.failed_revisions
1812
1813
@needs_read_lock
1814
def missing_revision_ids(self, revision_id=None):
1815
"""See InterRepository.missing_revision_ids()."""
1816
if revision_id is not None:
1817
source_ids = self.source.get_ancestry(revision_id)
1818
assert source_ids[0] == None
1819
source_ids.pop(0)
1820
else:
1821
source_ids = self.source._all_possible_ids()
1822
source_ids_set = set(source_ids)
1823
# source_ids is the worst possible case we may need to pull.
1824
# now we want to filter source_ids against what we actually
1825
# have in target, but don't try to check for existence where we know
1826
# we do not have a revision as that would be pointless.
1827
target_ids = set(self.target._all_possible_ids())
1828
possibly_present_revisions = target_ids.intersection(source_ids_set)
1829
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1830
required_revisions = source_ids_set.difference(actually_present_revisions)
1831
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1832
if revision_id is not None:
1833
# we used get_ancestry to determine source_ids then we are assured all
1834
# revisions referenced are present as they are installed in topological order.
1835
# and the tip revision was validated by get_ancestry.
1836
return required_topo_revisions
1837
else:
1838
# if we just grabbed the possibly available ids, then
1839
# we only have an estimate of whats available and need to validate
1840
# that against the revision records.
1841
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1842
1843
InterRepository.register_optimiser(InterWeaveRepo)
1844
InterRepository.register_optimiser(InterKnitRepo)
1845
1846
1847
class RepositoryTestProviderAdapter(object):
1848
"""A tool to generate a suite testing multiple repository formats at once.
1849
1850
This is done by copying the test once for each transport and injecting
1851
the transport_server, transport_readonly_server, and bzrdir_format and
1852
repository_format classes into each copy. Each copy is also given a new id()
1853
to make it easy to identify.
1854
"""
1855
1856
def __init__(self, transport_server, transport_readonly_server, formats):
1857
self._transport_server = transport_server
1858
self._transport_readonly_server = transport_readonly_server
1859
self._formats = formats
1860
1861
def adapt(self, test):
1862
result = TestSuite()
1863
for repository_format, bzrdir_format in self._formats:
1864
new_test = deepcopy(test)
1865
new_test.transport_server = self._transport_server
1866
new_test.transport_readonly_server = self._transport_readonly_server
1867
new_test.bzrdir_format = bzrdir_format
1868
new_test.repository_format = repository_format
1869
def make_new_test_id():
1870
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1871
return lambda: new_id
1872
new_test.id = make_new_test_id()
1873
result.addTest(new_test)
1874
return result
1875
1876
1877
class InterRepositoryTestProviderAdapter(object):
1878
"""A tool to generate a suite testing multiple inter repository formats.
1879
1880
This is done by copying the test once for each interrepo provider and injecting
1881
the transport_server, transport_readonly_server, repository_format and
1882
repository_to_format classes into each copy.
1883
Each copy is also given a new id() to make it easy to identify.
1884
"""
1885
1886
def __init__(self, transport_server, transport_readonly_server, formats):
1887
self._transport_server = transport_server
1888
self._transport_readonly_server = transport_readonly_server
1889
self._formats = formats
1890
1891
def adapt(self, test):
1892
result = TestSuite()
1893
for interrepo_class, repository_format, repository_format_to in self._formats:
1894
new_test = deepcopy(test)
1895
new_test.transport_server = self._transport_server
1896
new_test.transport_readonly_server = self._transport_readonly_server
1897
new_test.interrepo_class = interrepo_class
1898
new_test.repository_format = repository_format
1899
new_test.repository_format_to = repository_format_to
1900
def make_new_test_id():
1901
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1902
return lambda: new_id
1903
new_test.id = make_new_test_id()
1904
result.addTest(new_test)
1905
return result
1906
1907
@staticmethod
1908
def default_test_list():
1909
"""Generate the default list of interrepo permutations to test."""
1910
result = []
1911
# test the default InterRepository between format 6 and the current
1912
# default format.
1913
# XXX: robertc 20060220 reinstate this when there are two supported
1914
# formats which do not have an optimal code path between them.
1915
result.append((InterRepository,
1916
RepositoryFormat6(),
1917
RepositoryFormatKnit1()))
1918
for optimiser in InterRepository._optimisers:
1919
result.append((optimiser,
1920
optimiser._matching_repo_format,
1921
optimiser._matching_repo_format
1922
))
1923
# if there are specific combinations we want to use, we can add them
1924
# here.
1925
return result
1926
1927
1928
class CopyConverter(object):
1929
"""A repository conversion tool which just performs a copy of the content.
1930
1931
This is slow but quite reliable.
1932
"""
1933
1934
def __init__(self, target_format):
1935
"""Create a CopyConverter.
1936
1937
:param target_format: The format the resulting repository should be.
1938
"""
1939
self.target_format = target_format
1940
1941
def convert(self, repo, pb):
1942
"""Perform the conversion of to_convert, giving feedback via pb.
1943
1944
:param to_convert: The disk object to convert.
1945
:param pb: a progress bar to use for progress information.
1946
"""
1947
self.pb = pb
1948
self.count = 0
1949
self.total = 4
1950
# this is only useful with metadir layouts - separated repo content.
1951
# trigger an assertion if not such
1952
repo._format.get_format_string()
1953
self.repo_dir = repo.bzrdir
1954
self.step('Moving repository to repository.backup')
1955
self.repo_dir.transport.move('repository', 'repository.backup')
1956
backup_transport = self.repo_dir.transport.clone('repository.backup')
1957
self.source_repo = repo._format.open(self.repo_dir,
1958
_found=True,
1959
_override_transport=backup_transport)
1960
self.step('Creating new repository')
1961
converted = self.target_format.initialize(self.repo_dir,
1962
self.source_repo.is_shared())
1963
converted.lock_write()
1964
try:
1965
self.step('Copying content into repository.')
1966
self.source_repo.copy_content_into(converted)
1967
finally:
1968
converted.unlock()
1969
self.step('Deleting old repository content.')
1970
self.repo_dir.transport.delete_tree('repository.backup')
1971
self.pb.note('repository converted')
1972
1973
def step(self, message):
1974
"""Update the pb by a step."""
1975
self.count +=1
1976
self.pb.update(message, self.count, self.total)
1977
1978
1979
class CommitBuilder(object):
1980
"""Provides an interface to build up a commit.
1981
1982
This allows describing a tree to be committed without needing to
1983
know the internals of the format of the repository.
1984
"""
1985
def __init__(self, repository, parents, config, timestamp=None,
1986
timezone=None, committer=None, revprops=None,
1987
revision_id=None):
1988
"""Initiate a CommitBuilder.
1989
1990
:param repository: Repository to commit to.
1991
:param parents: Revision ids of the parents of the new revision.
1992
:param config: Configuration to use.
1993
:param timestamp: Optional timestamp recorded for commit.
1994
:param timezone: Optional timezone for timestamp.
1995
:param committer: Optional committer to set for commit.
1996
:param revprops: Optional dictionary of revision properties.
1997
:param revision_id: Optional revision id.
1998
"""
1999
self._config = config
2000
2001
if committer is None:
2002
self._committer = self._config.username()
2003
else:
2004
assert isinstance(committer, basestring), type(committer)
2005
self._committer = committer
2006
2007
self.new_inventory = Inventory()
2008
self._new_revision_id = revision_id
2009
self.parents = parents
2010
self.repository = repository
2011
2012
self._revprops = {}
2013
if revprops is not None:
2014
self._revprops.update(revprops)
2015
2016
if timestamp is None:
2017
self._timestamp = time.time()
2018
else:
2019
self._timestamp = long(timestamp)
2020
2021
if timezone is None:
2022
self._timezone = local_time_offset()
2023
else:
2024
self._timezone = int(timezone)
2025
2026
self._generate_revision_if_needed()
2027
2028
def commit(self, message):
2029
"""Make the actual commit.
2030
2031
:return: The revision id of the recorded revision.
2032
"""
2033
rev = Revision(timestamp=self._timestamp,
2034
timezone=self._timezone,
2035
committer=self._committer,
2036
message=message,
2037
inventory_sha1=self.inv_sha1,
2038
revision_id=self._new_revision_id,
2039
properties=self._revprops)
2040
rev.parent_ids = self.parents
2041
self.repository.add_revision(self._new_revision_id, rev,
2042
self.new_inventory, self._config)
2043
return self._new_revision_id
2044
2045
def finish_inventory(self):
2046
"""Tell the builder that the inventory is finished."""
2047
self.new_inventory.revision_id = self._new_revision_id
2048
self.inv_sha1 = self.repository.add_inventory(
2049
self._new_revision_id,
2050
self.new_inventory,
2051
self.parents
2052
)
2053
2054
def _gen_revision_id(self):
2055
"""Return new revision-id."""
2056
s = '%s-%s-' % (self._config.user_email(),
2057
compact_date(self._timestamp))
2058
s += hexlify(rand_bytes(8))
2059
return s
2060
2061
def _generate_revision_if_needed(self):
2062
"""Create a revision id if None was supplied.
2063
2064
If the repository can not support user-specified revision ids
2065
they should override this function and raise UnsupportedOperation
2066
if _new_revision_id is not None.
2067
2068
:raises: UnsupportedOperation
2069
"""
2070
if self._new_revision_id is None:
2071
self._new_revision_id = self._gen_revision_id()
2072
2073
def record_entry_contents(self, ie, parent_invs, path, tree):
2074
"""Record the content of ie from tree into the commit if needed.
2075
2076
:param ie: An inventory entry present in the commit.
2077
:param parent_invs: The inventories of the parent revisions of the
2078
commit.
2079
:param path: The path the entry is at in the tree.
2080
:param tree: The tree which contains this entry and should be used to
2081
obtain content.
2082
"""
2083
self.new_inventory.add(ie)
2084
2085
# ie.revision is always None if the InventoryEntry is considered
2086
# for committing. ie.snapshot will record the correct revision
2087
# which may be the sole parent if it is untouched.
2088
if ie.revision is not None:
2089
return
2090
previous_entries = ie.find_previous_heads(
2091
parent_invs,
2092
self.repository.weave_store,
2093
self.repository.get_transaction())
2094
# we are creating a new revision for ie in the history store
2095
# and inventory.
2096
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2097
2098
def modified_directory(self, file_id, file_parents):
2099
"""Record the presence of a symbolic link.
2100
2101
:param file_id: The file_id of the link to record.
2102
:param file_parents: The per-file parent revision ids.
2103
"""
2104
self._add_text_to_weave(file_id, [], file_parents.keys())
2105
2106
def modified_file_text(self, file_id, file_parents,
2107
get_content_byte_lines, text_sha1=None,
2108
text_size=None):
2109
"""Record the text of file file_id
2110
2111
:param file_id: The file_id of the file to record the text of.
2112
:param file_parents: The per-file parent revision ids.
2113
:param get_content_byte_lines: A callable which will return the byte
2114
lines for the file.
2115
:param text_sha1: Optional SHA1 of the file contents.
2116
:param text_size: Optional size of the file contents.
2117
"""
2118
mutter('storing text of file {%s} in revision {%s} into %r',
2119
file_id, self._new_revision_id, self.repository.weave_store)
2120
# special case to avoid diffing on renames or
2121
# reparenting
2122
if (len(file_parents) == 1
2123
and text_sha1 == file_parents.values()[0].text_sha1
2124
and text_size == file_parents.values()[0].text_size):
2125
previous_ie = file_parents.values()[0]
2126
versionedfile = self.repository.weave_store.get_weave(file_id,
2127
self.repository.get_transaction())
2128
versionedfile.clone_text(self._new_revision_id,
2129
previous_ie.revision, file_parents.keys())
2130
return text_sha1, text_size
2131
else:
2132
new_lines = get_content_byte_lines()
2133
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2134
# should return the SHA1 and size
2135
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2136
return osutils.sha_strings(new_lines), \
2137
sum(map(len, new_lines))
2138
2139
def modified_link(self, file_id, file_parents, link_target):
2140
"""Record the presence of a symbolic link.
2141
2142
:param file_id: The file_id of the link to record.
2143
:param file_parents: The per-file parent revision ids.
2144
:param link_target: Target location of this link.
2145
"""
2146
self._add_text_to_weave(file_id, [], file_parents.keys())
2147
2148
def _add_text_to_weave(self, file_id, new_lines, parents):
2149
versionedfile = self.repository.weave_store.get_weave_or_empty(
2150
file_id, self.repository.get_transaction())
2151
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2152
versionedfile.clear_cache()
2153
2154
2155
_unescape_map = {
2156
'apos':"'",
2157
'quot':'"',
2158
'amp':'&',
2159
'lt':'<',
2160
'gt':'>'
2161
}
2162
2163
2164
def _unescaper(match, _map=_unescape_map):
2165
return _map[match.group(1)]
2166
2167
2168
_unescape_re = None
2169
2170
2171
def _unescape_xml(data):
2172
"""Unescape predefined XML entities in a string of data."""
2173
global _unescape_re
2174
if _unescape_re is None:
2175
_unescape_re = re.compile('\&([^;]*);')
2176
return _unescape_re.sub(_unescaper, data)
Loggerhead is a web-based interface for Breezy
Version: 2.0.1