(root)/brz/remove-bazaar
» Revision
2116.4.1
(compared to revision 0.200.231)
: bzrlib/repository.py
To get this branch, use:
bzr branch
http://gegoxaren.bato24.eu/bzr/brz/remove-bazaar
« back to all changes in this revision
Viewing changes to bzrlib/repository.py
- browse files at revision 2116.4.1
- compare with another revision
- download diff
- download tarball
- view history from revision 2116.4.1
- Committer: John Arbash Meinel
- Date: 2006-11-02 22:48:49 UTC
- mto: This revision was merged to the branch mainline in revision 2132.
- Revision ID: john@arbash-meinel.com-20061102224849-9683d741ce3fbe38
Update file and revision id generators.
Move id generation to its own file, deprecate the old functions,
make gen_revision_id() not require an email address in the
username, and use rand_chars() instead of hexlify(rand_bytes())
Move id generation to its own file, deprecate the old functions,
make gen_revision_id() not require an email address in the
username, and use rand_chars() instead of hexlify(rand_bytes())
- files added:
- .bzrignore
- bzrlib
- bzrlib/benchmarks
- bzrlib/benchmarks/tree_creator
- 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/intertree_implementations
- bzrlib/tests/interversionedfile_implementations
- bzrlib/tests/repository_implementations
- bzrlib/tests/revisionstore_implementations
- bzrlib/tests/test_patches_data
- bzrlib/tests/tree_implementations
- 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
- bzrlib/util/elementtree
- bzrlib/version_info_formats
- contrib
- contrib/bash
- contrib/emacs
- contrib/zsh
- doc
- tools
- tools/doc_generate
- tools/win32
- files removed:
- .bzrignore
- foreign
- notes
- tests
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 cStringIO import StringIO
18
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
21
from binascii import hexlify
22
from copy import deepcopy
23
import re
24
import time
25
import unittest
26
27
from bzrlib import (
28
bzrdir,
29
check,
30
delta,
31
errors,
32
generate_ids,
33
gpg,
34
graph,
35
knit,
36
lockable_files,
37
lockdir,
38
osutils,
39
revision as _mod_revision,
40
symbol_versioning,
41
transactions,
42
ui,
43
weave,
44
weavefile,
45
xml5,
46
xml6,
47
)
48
from bzrlib.osutils import (
49
rand_bytes,
50
compact_date,
51
local_time_offset,
52
)
53
from bzrlib.revisiontree import RevisionTree
54
from bzrlib.store.versioned import VersionedFileStore
55
from bzrlib.store.text import TextStore
56
from bzrlib.testament import Testament
57
""")
58
59
from bzrlib.decorators import needs_read_lock, needs_write_lock
60
from bzrlib.inter import InterObject
61
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
62
from bzrlib.symbol_versioning import (
63
deprecated_method,
64
zero_nine,
65
)
66
from bzrlib.trace import mutter, note, warning
67
68
69
# Old formats display a warning, but only once
70
_deprecation_warning_done = False
71
72
73
class Repository(object):
74
"""Repository holding history for one or more branches.
75
76
The repository holds and retrieves historical information including
77
revisions and file history. It's normally accessed only by the Branch,
78
which views a particular line of development through that history.
79
80
The Repository builds on top of Stores and a Transport, which respectively
81
describe the disk data format and the way of accessing the (possibly
82
remote) disk.
83
"""
84
85
@needs_write_lock
86
def add_inventory(self, revid, inv, parents):
87
"""Add the inventory inv to the repository as revid.
88
89
:param parents: The revision ids of the parents that revid
90
is known to have and are in the repository already.
91
92
returns the sha1 of the serialized inventory.
93
"""
94
assert inv.revision_id is None or inv.revision_id == revid, \
95
"Mismatch between inventory revision" \
96
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
97
assert inv.root is not None
98
inv_text = self.serialise_inventory(inv)
99
inv_sha1 = osutils.sha_string(inv_text)
100
inv_vf = self.control_weaves.get_weave('inventory',
101
self.get_transaction())
102
self._inventory_add_lines(inv_vf, revid, parents, osutils.split_lines(inv_text))
103
return inv_sha1
104
105
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
106
final_parents = []
107
for parent in parents:
108
if parent in inv_vf:
109
final_parents.append(parent)
110
111
inv_vf.add_lines(revid, final_parents, lines)
112
113
@needs_write_lock
114
def add_revision(self, rev_id, rev, inv=None, config=None):
115
"""Add rev to the revision store as rev_id.
116
117
:param rev_id: the revision id to use.
118
:param rev: The revision object.
119
:param inv: The inventory for the revision. if None, it will be looked
120
up in the inventory storer
121
:param config: If None no digital signature will be created.
122
If supplied its signature_needed method will be used
123
to determine if a signature should be made.
124
"""
125
if config is not None and config.signature_needed():
126
if inv is None:
127
inv = self.get_inventory(rev_id)
128
plaintext = Testament(rev, inv).as_short_text()
129
self.store_revision_signature(
130
gpg.GPGStrategy(config), plaintext, rev_id)
131
if not rev_id in self.get_inventory_weave():
132
if inv is None:
133
raise errors.WeaveRevisionNotPresent(rev_id,
134
self.get_inventory_weave())
135
else:
136
# yes, this is not suitable for adding with ghosts.
137
self.add_inventory(rev_id, inv, rev.parent_ids)
138
self._revision_store.add_revision(rev, self.get_transaction())
139
140
@needs_read_lock
141
def _all_possible_ids(self):
142
"""Return all the possible revisions that we could find."""
143
return self.get_inventory_weave().versions()
144
145
def all_revision_ids(self):
146
"""Returns a list of all the revision ids in the repository.
147
148
This is deprecated because code should generally work on the graph
149
reachable from a particular revision, and ignore any other revisions
150
that might be present. There is no direct replacement method.
151
"""
152
return self._all_revision_ids()
153
154
@needs_read_lock
155
def _all_revision_ids(self):
156
"""Returns a list of all the revision ids in the repository.
157
158
These are in as much topological order as the underlying store can
159
present: for weaves ghosts may lead to a lack of correctness until
160
the reweave updates the parents list.
161
"""
162
if self._revision_store.text_store.listable():
163
return self._revision_store.all_revision_ids(self.get_transaction())
164
result = self._all_possible_ids()
165
return self._eliminate_revisions_not_present(result)
166
167
def break_lock(self):
168
"""Break a lock if one is present from another instance.
169
170
Uses the ui factory to ask for confirmation if the lock may be from
171
an active process.
172
"""
173
self.control_files.break_lock()
174
175
@needs_read_lock
176
def _eliminate_revisions_not_present(self, revision_ids):
177
"""Check every revision id in revision_ids to see if we have it.
178
179
Returns a set of the present revisions.
180
"""
181
result = []
182
for id in revision_ids:
183
if self.has_revision(id):
184
result.append(id)
185
return result
186
187
@staticmethod
188
def create(a_bzrdir):
189
"""Construct the current default format repository in a_bzrdir."""
190
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
191
192
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
193
"""instantiate a Repository.
194
195
:param _format: The format of the repository on disk.
196
:param a_bzrdir: The BzrDir of the repository.
197
198
In the future we will have a single api for all stores for
199
getting file texts, inventories and revisions, then
200
this construct will accept instances of those things.
201
"""
202
super(Repository, self).__init__()
203
self._format = _format
204
# the following are part of the public API for Repository:
205
self.bzrdir = a_bzrdir
206
self.control_files = control_files
207
self._revision_store = _revision_store
208
self.text_store = text_store
209
# backwards compatibility
210
self.weave_store = text_store
211
# not right yet - should be more semantically clear ?
212
#
213
self.control_store = control_store
214
self.control_weaves = control_store
215
# TODO: make sure to construct the right store classes, etc, depending
216
# on whether escaping is required.
217
self._warn_if_deprecated()
218
self._serializer = xml5.serializer_v5
219
220
def __repr__(self):
221
return '%s(%r)' % (self.__class__.__name__,
222
self.bzrdir.transport.base)
223
224
def is_locked(self):
225
return self.control_files.is_locked()
226
227
def lock_write(self):
228
self.control_files.lock_write()
229
230
def lock_read(self):
231
self.control_files.lock_read()
232
233
def get_physical_lock_status(self):
234
return self.control_files.get_physical_lock_status()
235
236
@needs_read_lock
237
def missing_revision_ids(self, other, revision_id=None):
238
"""Return the revision ids that other has that this does not.
239
240
These are returned in topological order.
241
242
revision_id: only return revision ids included by revision_id.
243
"""
244
return InterRepository.get(other, self).missing_revision_ids(revision_id)
245
246
@staticmethod
247
def open(base):
248
"""Open the repository rooted at base.
249
250
For instance, if the repository is at URL/.bzr/repository,
251
Repository.open(URL) -> a Repository instance.
252
"""
253
control = bzrdir.BzrDir.open(base)
254
return control.open_repository()
255
256
def copy_content_into(self, destination, revision_id=None, basis=None):
257
"""Make a complete copy of the content in self into destination.
258
259
This is a destructive operation! Do not use it on existing
260
repositories.
261
"""
262
return InterRepository.get(self, destination).copy_content(revision_id, basis)
263
264
def fetch(self, source, revision_id=None, pb=None):
265
"""Fetch the content required to construct revision_id from source.
266
267
If revision_id is None all content is copied.
268
"""
269
return InterRepository.get(source, self).fetch(revision_id=revision_id,
270
pb=pb)
271
272
def get_commit_builder(self, branch, parents, config, timestamp=None,
273
timezone=None, committer=None, revprops=None,
274
revision_id=None):
275
"""Obtain a CommitBuilder for this repository.
276
277
:param branch: Branch to commit to.
278
:param parents: Revision ids of the parents of the new revision.
279
:param config: Configuration to use.
280
:param timestamp: Optional timestamp recorded for commit.
281
:param timezone: Optional timezone for timestamp.
282
:param committer: Optional committer to set for commit.
283
:param revprops: Optional dictionary of revision properties.
284
:param revision_id: Optional revision id.
285
"""
286
return _CommitBuilder(self, parents, config, timestamp, timezone,
287
committer, revprops, revision_id)
288
289
def unlock(self):
290
self.control_files.unlock()
291
292
@needs_read_lock
293
def clone(self, a_bzrdir, revision_id=None, basis=None):
294
"""Clone this repository into a_bzrdir using the current format.
295
296
Currently no check is made that the format of this repository and
297
the bzrdir format are compatible. FIXME RBC 20060201.
298
"""
299
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
300
# use target default format.
301
result = a_bzrdir.create_repository()
302
# FIXME RBC 20060209 split out the repository type to avoid this check ?
303
elif isinstance(a_bzrdir._format,
304
(bzrdir.BzrDirFormat4,
305
bzrdir.BzrDirFormat5,
306
bzrdir.BzrDirFormat6)):
307
result = a_bzrdir.open_repository()
308
else:
309
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
310
self.copy_content_into(result, revision_id, basis)
311
return result
312
313
@needs_read_lock
314
def has_revision(self, revision_id):
315
"""True if this repository has a copy of the revision."""
316
return self._revision_store.has_revision_id(revision_id,
317
self.get_transaction())
318
319
@needs_read_lock
320
def get_revision_reconcile(self, revision_id):
321
"""'reconcile' helper routine that allows access to a revision always.
322
323
This variant of get_revision does not cross check the weave graph
324
against the revision one as get_revision does: but it should only
325
be used by reconcile, or reconcile-alike commands that are correcting
326
or testing the revision graph.
327
"""
328
if not revision_id or not isinstance(revision_id, basestring):
329
raise errors.InvalidRevisionId(revision_id=revision_id,
330
branch=self)
331
return self._revision_store.get_revisions([revision_id],
332
self.get_transaction())[0]
333
@needs_read_lock
334
def get_revisions(self, revision_ids):
335
return self._revision_store.get_revisions(revision_ids,
336
self.get_transaction())
337
338
@needs_read_lock
339
def get_revision_xml(self, revision_id):
340
rev = self.get_revision(revision_id)
341
rev_tmp = StringIO()
342
# the current serializer..
343
self._revision_store._serializer.write_revision(rev, rev_tmp)
344
rev_tmp.seek(0)
345
return rev_tmp.getvalue()
346
347
@needs_read_lock
348
def get_revision(self, revision_id):
349
"""Return the Revision object for a named revision"""
350
r = self.get_revision_reconcile(revision_id)
351
# weave corruption can lead to absent revision markers that should be
352
# present.
353
# the following test is reasonably cheap (it needs a single weave read)
354
# and the weave is cached in read transactions. In write transactions
355
# it is not cached but typically we only read a small number of
356
# revisions. For knits when they are introduced we will probably want
357
# to ensure that caching write transactions are in use.
358
inv = self.get_inventory_weave()
359
self._check_revision_parents(r, inv)
360
return r
361
362
@needs_read_lock
363
def get_deltas_for_revisions(self, revisions):
364
"""Produce a generator of revision deltas.
365
366
Note that the input is a sequence of REVISIONS, not revision_ids.
367
Trees will be held in memory until the generator exits.
368
Each delta is relative to the revision's lefthand predecessor.
369
"""
370
required_trees = set()
371
for revision in revisions:
372
required_trees.add(revision.revision_id)
373
required_trees.update(revision.parent_ids[:1])
374
trees = dict((t.get_revision_id(), t) for
375
t in self.revision_trees(required_trees))
376
for revision in revisions:
377
if not revision.parent_ids:
378
old_tree = self.revision_tree(None)
379
else:
380
old_tree = trees[revision.parent_ids[0]]
381
yield trees[revision.revision_id].changes_from(old_tree)
382
383
@needs_read_lock
384
def get_revision_delta(self, revision_id):
385
"""Return the delta for one revision.
386
387
The delta is relative to the left-hand predecessor of the
388
revision.
389
"""
390
r = self.get_revision(revision_id)
391
return list(self.get_deltas_for_revisions([r]))[0]
392
393
def _check_revision_parents(self, revision, inventory):
394
"""Private to Repository and Fetch.
395
396
This checks the parentage of revision in an inventory weave for
397
consistency and is only applicable to inventory-weave-for-ancestry
398
using repository formats & fetchers.
399
"""
400
weave_parents = inventory.get_parents(revision.revision_id)
401
weave_names = inventory.versions()
402
for parent_id in revision.parent_ids:
403
if parent_id in weave_names:
404
# this parent must not be a ghost.
405
if not parent_id in weave_parents:
406
# but it is a ghost
407
raise errors.CorruptRepository(self)
408
409
@needs_write_lock
410
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
411
signature = gpg_strategy.sign(plaintext)
412
self._revision_store.add_revision_signature_text(revision_id,
413
signature,
414
self.get_transaction())
415
416
def fileids_altered_by_revision_ids(self, revision_ids):
417
"""Find the file ids and versions affected by revisions.
418
419
:param revisions: an iterable containing revision ids.
420
:return: a dictionary mapping altered file-ids to an iterable of
421
revision_ids. Each altered file-ids has the exact revision_ids that
422
altered it listed explicitly.
423
"""
424
assert self._serializer.support_altered_by_hack, \
425
("fileids_altered_by_revision_ids only supported for branches "
426
"which store inventory as unnested xml, not on %r" % self)
427
selected_revision_ids = set(revision_ids)
428
w = self.get_inventory_weave()
429
result = {}
430
431
# this code needs to read every new line in every inventory for the
432
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
433
# not present in one of those inventories is unnecessary but not
434
# harmful because we are filtering by the revision id marker in the
435
# inventory lines : we only select file ids altered in one of those
436
# revisions. We don't need to see all lines in the inventory because
437
# only those added in an inventory in rev X can contain a revision=X
438
# line.
439
pb = ui.ui_factory.nested_progress_bar()
440
try:
441
for line in w.iter_lines_added_or_present_in_versions(
442
selected_revision_ids, pb=pb):
443
start = line.find('file_id="')+9
444
if start < 9: continue
445
end = line.find('"', start)
446
assert end>= 0
447
file_id = _unescape_xml(line[start:end])
448
449
start = line.find('revision="')+10
450
if start < 10: continue
451
end = line.find('"', start)
452
assert end>= 0
453
revision_id = _unescape_xml(line[start:end])
454
if revision_id in selected_revision_ids:
455
result.setdefault(file_id, set()).add(revision_id)
456
finally:
457
pb.finished()
458
return result
459
460
@needs_read_lock
461
def get_inventory_weave(self):
462
return self.control_weaves.get_weave('inventory',
463
self.get_transaction())
464
465
@needs_read_lock
466
def get_inventory(self, revision_id):
467
"""Get Inventory object by hash."""
468
return self.deserialise_inventory(
469
revision_id, self.get_inventory_xml(revision_id))
470
471
def deserialise_inventory(self, revision_id, xml):
472
"""Transform the xml into an inventory object.
473
474
:param revision_id: The expected revision id of the inventory.
475
:param xml: A serialised inventory.
476
"""
477
result = self._serializer.read_inventory_from_string(xml)
478
result.root.revision = revision_id
479
return result
480
481
def serialise_inventory(self, inv):
482
return self._serializer.write_inventory_to_string(inv)
483
484
@needs_read_lock
485
def get_inventory_xml(self, revision_id):
486
"""Get inventory XML as a file object."""
487
try:
488
assert isinstance(revision_id, basestring), type(revision_id)
489
iw = self.get_inventory_weave()
490
return iw.get_text(revision_id)
491
except IndexError:
492
raise errors.HistoryMissing(self, 'inventory', revision_id)
493
494
@needs_read_lock
495
def get_inventory_sha1(self, revision_id):
496
"""Return the sha1 hash of the inventory entry
497
"""
498
return self.get_revision(revision_id).inventory_sha1
499
500
@needs_read_lock
501
def get_revision_graph(self, revision_id=None):
502
"""Return a dictionary containing the revision graph.
503
504
:param revision_id: The revision_id to get a graph from. If None, then
505
the entire revision graph is returned. This is a deprecated mode of
506
operation and will be removed in the future.
507
:return: a dictionary of revision_id->revision_parents_list.
508
"""
509
# special case NULL_REVISION
510
if revision_id == _mod_revision.NULL_REVISION:
511
return {}
512
a_weave = self.get_inventory_weave()
513
all_revisions = self._eliminate_revisions_not_present(
514
a_weave.versions())
515
entire_graph = dict([(node, a_weave.get_parents(node)) for
516
node in all_revisions])
517
if revision_id is None:
518
return entire_graph
519
elif revision_id not in entire_graph:
520
raise errors.NoSuchRevision(self, revision_id)
521
else:
522
# add what can be reached from revision_id
523
result = {}
524
pending = set([revision_id])
525
while len(pending) > 0:
526
node = pending.pop()
527
result[node] = entire_graph[node]
528
for revision_id in result[node]:
529
if revision_id not in result:
530
pending.add(revision_id)
531
return result
532
533
@needs_read_lock
534
def get_revision_graph_with_ghosts(self, revision_ids=None):
535
"""Return a graph of the revisions with ghosts marked as applicable.
536
537
:param revision_ids: an iterable of revisions to graph or None for all.
538
:return: a Graph object with the graph reachable from revision_ids.
539
"""
540
result = graph.Graph()
541
if not revision_ids:
542
pending = set(self.all_revision_ids())
543
required = set([])
544
else:
545
pending = set(revision_ids)
546
# special case NULL_REVISION
547
if _mod_revision.NULL_REVISION in pending:
548
pending.remove(_mod_revision.NULL_REVISION)
549
required = set(pending)
550
done = set([])
551
while len(pending):
552
revision_id = pending.pop()
553
try:
554
rev = self.get_revision(revision_id)
555
except errors.NoSuchRevision:
556
if revision_id in required:
557
raise
558
# a ghost
559
result.add_ghost(revision_id)
560
continue
561
for parent_id in rev.parent_ids:
562
# is this queued or done ?
563
if (parent_id not in pending and
564
parent_id not in done):
565
# no, queue it.
566
pending.add(parent_id)
567
result.add_node(revision_id, rev.parent_ids)
568
done.add(revision_id)
569
return result
570
571
@needs_read_lock
572
def get_revision_inventory(self, revision_id):
573
"""Return inventory of a past revision."""
574
# TODO: Unify this with get_inventory()
575
# bzr 0.0.6 and later imposes the constraint that the inventory_id
576
# must be the same as its revision, so this is trivial.
577
if revision_id is None:
578
# This does not make sense: if there is no revision,
579
# then it is the current tree inventory surely ?!
580
# and thus get_root_id() is something that looks at the last
581
# commit on the branch, and the get_root_id is an inventory check.
582
raise NotImplementedError
583
# return Inventory(self.get_root_id())
584
else:
585
return self.get_inventory(revision_id)
586
587
@needs_read_lock
588
def is_shared(self):
589
"""Return True if this repository is flagged as a shared repository."""
590
raise NotImplementedError(self.is_shared)
591
592
@needs_write_lock
593
def reconcile(self, other=None, thorough=False):
594
"""Reconcile this repository."""
595
from bzrlib.reconcile import RepoReconciler
596
reconciler = RepoReconciler(self, thorough=thorough)
597
reconciler.reconcile()
598
return reconciler
599
600
@needs_read_lock
601
def revision_tree(self, revision_id):
602
"""Return Tree for a revision on this branch.
603
604
`revision_id` may be None for the empty tree revision.
605
"""
606
# TODO: refactor this to use an existing revision object
607
# so we don't need to read it in twice.
608
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
609
return RevisionTree(self, Inventory(root_id=None),
610
_mod_revision.NULL_REVISION)
611
else:
612
inv = self.get_revision_inventory(revision_id)
613
return RevisionTree(self, inv, revision_id)
614
615
@needs_read_lock
616
def revision_trees(self, revision_ids):
617
"""Return Tree for a revision on this branch.
618
619
`revision_id` may not be None or 'null:'"""
620
assert None not in revision_ids
621
assert _mod_revision.NULL_REVISION not in revision_ids
622
texts = self.get_inventory_weave().get_texts(revision_ids)
623
for text, revision_id in zip(texts, revision_ids):
624
inv = self.deserialise_inventory(revision_id, text)
625
yield RevisionTree(self, inv, revision_id)
626
627
@needs_read_lock
628
def get_ancestry(self, revision_id):
629
"""Return a list of revision-ids integrated by a revision.
630
631
The first element of the list is always None, indicating the origin
632
revision. This might change when we have history horizons, or
633
perhaps we should have a new API.
634
635
This is topologically sorted.
636
"""
637
if revision_id is None:
638
return [None]
639
if not self.has_revision(revision_id):
640
raise errors.NoSuchRevision(self, revision_id)
641
w = self.get_inventory_weave()
642
candidates = w.get_ancestry(revision_id)
643
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
644
645
@needs_read_lock
646
def print_file(self, file, revision_id):
647
"""Print `file` to stdout.
648
649
FIXME RBC 20060125 as John Meinel points out this is a bad api
650
- it writes to stdout, it assumes that that is valid etc. Fix
651
by creating a new more flexible convenience function.
652
"""
653
tree = self.revision_tree(revision_id)
654
# use inventory as it was in that revision
655
file_id = tree.inventory.path2id(file)
656
if not file_id:
657
# TODO: jam 20060427 Write a test for this code path
658
# it had a bug in it, and was raising the wrong
659
# exception.
660
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
661
tree.print_file(file_id)
662
663
def get_transaction(self):
664
return self.control_files.get_transaction()
665
666
def revision_parents(self, revid):
667
return self.get_inventory_weave().parent_names(revid)
668
669
@needs_write_lock
670
def set_make_working_trees(self, new_value):
671
"""Set the policy flag for making working trees when creating branches.
672
673
This only applies to branches that use this repository.
674
675
The default is 'True'.
676
:param new_value: True to restore the default, False to disable making
677
working trees.
678
"""
679
raise NotImplementedError(self.set_make_working_trees)
680
681
def make_working_trees(self):
682
"""Returns the policy for making working trees on new branches."""
683
raise NotImplementedError(self.make_working_trees)
684
685
@needs_write_lock
686
def sign_revision(self, revision_id, gpg_strategy):
687
plaintext = Testament.from_revision(self, revision_id).as_short_text()
688
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
689
690
@needs_read_lock
691
def has_signature_for_revision_id(self, revision_id):
692
"""Query for a revision signature for revision_id in the repository."""
693
return self._revision_store.has_signature(revision_id,
694
self.get_transaction())
695
696
@needs_read_lock
697
def get_signature_text(self, revision_id):
698
"""Return the text for a signature."""
699
return self._revision_store.get_signature_text(revision_id,
700
self.get_transaction())
701
702
@needs_read_lock
703
def check(self, revision_ids):
704
"""Check consistency of all history of given revision_ids.
705
706
Different repository implementations should override _check().
707
708
:param revision_ids: A non-empty list of revision_ids whose ancestry
709
will be checked. Typically the last revision_id of a branch.
710
"""
711
if not revision_ids:
712
raise ValueError("revision_ids must be non-empty in %s.check"
713
% (self,))
714
return self._check(revision_ids)
715
716
def _check(self, revision_ids):
717
result = check.Check(self)
718
result.check()
719
return result
720
721
def _warn_if_deprecated(self):
722
global _deprecation_warning_done
723
if _deprecation_warning_done:
724
return
725
_deprecation_warning_done = True
726
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
727
% (self._format, self.bzrdir.transport.base))
728
729
def supports_rich_root(self):
730
return self._format.rich_root_data
731
732
733
class AllInOneRepository(Repository):
734
"""Legacy support - the repository behaviour for all-in-one branches."""
735
736
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
737
# we reuse one control files instance.
738
dir_mode = a_bzrdir._control_files._dir_mode
739
file_mode = a_bzrdir._control_files._file_mode
740
741
def get_store(name, compressed=True, prefixed=False):
742
# FIXME: This approach of assuming stores are all entirely compressed
743
# or entirely uncompressed is tidy, but breaks upgrade from
744
# some existing branches where there's a mixture; we probably
745
# still want the option to look for both.
746
relpath = a_bzrdir._control_files._escape(name)
747
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
748
prefixed=prefixed, compressed=compressed,
749
dir_mode=dir_mode,
750
file_mode=file_mode)
751
#if self._transport.should_cache():
752
# cache_path = os.path.join(self.cache_root, name)
753
# os.mkdir(cache_path)
754
# store = bzrlib.store.CachedStore(store, cache_path)
755
return store
756
757
# not broken out yet because the controlweaves|inventory_store
758
# and text_store | weave_store bits are still different.
759
if isinstance(_format, RepositoryFormat4):
760
# cannot remove these - there is still no consistent api
761
# which allows access to this old info.
762
self.inventory_store = get_store('inventory-store')
763
text_store = get_store('text-store')
764
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
765
766
@needs_read_lock
767
def is_shared(self):
768
"""AllInOne repositories cannot be shared."""
769
return False
770
771
@needs_write_lock
772
def set_make_working_trees(self, new_value):
773
"""Set the policy flag for making working trees when creating branches.
774
775
This only applies to branches that use this repository.
776
777
The default is 'True'.
778
:param new_value: True to restore the default, False to disable making
779
working trees.
780
"""
781
raise NotImplementedError(self.set_make_working_trees)
782
783
def make_working_trees(self):
784
"""Returns the policy for making working trees on new branches."""
785
return True
786
787
788
def install_revision(repository, rev, revision_tree):
789
"""Install all revision data into a repository."""
790
present_parents = []
791
parent_trees = {}
792
for p_id in rev.parent_ids:
793
if repository.has_revision(p_id):
794
present_parents.append(p_id)
795
parent_trees[p_id] = repository.revision_tree(p_id)
796
else:
797
parent_trees[p_id] = repository.revision_tree(None)
798
799
inv = revision_tree.inventory
800
entries = inv.iter_entries()
801
# backwards compatability hack: skip the root id.
802
if not repository.supports_rich_root():
803
path, root = entries.next()
804
if root.revision != rev.revision_id:
805
raise errors.IncompatibleRevision(repr(repository))
806
# Add the texts that are not already present
807
for path, ie in entries:
808
w = repository.weave_store.get_weave_or_empty(ie.file_id,
809
repository.get_transaction())
810
if ie.revision not in w:
811
text_parents = []
812
# FIXME: TODO: The following loop *may* be overlapping/duplicate
813
# with InventoryEntry.find_previous_heads(). if it is, then there
814
# is a latent bug here where the parents may have ancestors of each
815
# other. RBC, AB
816
for revision, tree in parent_trees.iteritems():
817
if ie.file_id not in tree:
818
continue
819
parent_id = tree.inventory[ie.file_id].revision
820
if parent_id in text_parents:
821
continue
822
text_parents.append(parent_id)
823
824
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
825
repository.get_transaction())
826
lines = revision_tree.get_file(ie.file_id).readlines()
827
vfile.add_lines(rev.revision_id, text_parents, lines)
828
try:
829
# install the inventory
830
repository.add_inventory(rev.revision_id, inv, present_parents)
831
except errors.RevisionAlreadyPresent:
832
pass
833
repository.add_revision(rev.revision_id, rev, inv)
834
835
836
class MetaDirRepository(Repository):
837
"""Repositories in the new meta-dir layout."""
838
839
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
840
super(MetaDirRepository, self).__init__(_format,
841
a_bzrdir,
842
control_files,
843
_revision_store,
844
control_store,
845
text_store)
846
dir_mode = self.control_files._dir_mode
847
file_mode = self.control_files._file_mode
848
849
@needs_read_lock
850
def is_shared(self):
851
"""Return True if this repository is flagged as a shared repository."""
852
return self.control_files._transport.has('shared-storage')
853
854
@needs_write_lock
855
def set_make_working_trees(self, new_value):
856
"""Set the policy flag for making working trees when creating branches.
857
858
This only applies to branches that use this repository.
859
860
The default is 'True'.
861
:param new_value: True to restore the default, False to disable making
862
working trees.
863
"""
864
if new_value:
865
try:
866
self.control_files._transport.delete('no-working-trees')
867
except errors.NoSuchFile:
868
pass
869
else:
870
self.control_files.put_utf8('no-working-trees', '')
871
872
def make_working_trees(self):
873
"""Returns the policy for making working trees on new branches."""
874
return not self.control_files._transport.has('no-working-trees')
875
876
877
class KnitRepository(MetaDirRepository):
878
"""Knit format repository."""
879
880
def _warn_if_deprecated(self):
881
# This class isn't deprecated
882
pass
883
884
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
885
inv_vf.add_lines_with_ghosts(revid, parents, lines)
886
887
@needs_read_lock
888
def _all_revision_ids(self):
889
"""See Repository.all_revision_ids()."""
890
# Knits get the revision graph from the index of the revision knit, so
891
# it's always possible even if they're on an unlistable transport.
892
return self._revision_store.all_revision_ids(self.get_transaction())
893
894
def fileid_involved_between_revs(self, from_revid, to_revid):
895
"""Find file_id(s) which are involved in the changes between revisions.
896
897
This determines the set of revisions which are involved, and then
898
finds all file ids affected by those revisions.
899
"""
900
vf = self._get_revision_vf()
901
from_set = set(vf.get_ancestry(from_revid))
902
to_set = set(vf.get_ancestry(to_revid))
903
changed = to_set.difference(from_set)
904
return self._fileid_involved_by_set(changed)
905
906
def fileid_involved(self, last_revid=None):
907
"""Find all file_ids modified in the ancestry of last_revid.
908
909
:param last_revid: If None, last_revision() will be used.
910
"""
911
if not last_revid:
912
changed = set(self.all_revision_ids())
913
else:
914
changed = set(self.get_ancestry(last_revid))
915
if None in changed:
916
changed.remove(None)
917
return self._fileid_involved_by_set(changed)
918
919
@needs_read_lock
920
def get_ancestry(self, revision_id):
921
"""Return a list of revision-ids integrated by a revision.
922
923
This is topologically sorted.
924
"""
925
if revision_id is None:
926
return [None]
927
vf = self._get_revision_vf()
928
try:
929
return [None] + vf.get_ancestry(revision_id)
930
except errors.RevisionNotPresent:
931
raise errors.NoSuchRevision(self, revision_id)
932
933
@needs_read_lock
934
def get_revision(self, revision_id):
935
"""Return the Revision object for a named revision"""
936
return self.get_revision_reconcile(revision_id)
937
938
@needs_read_lock
939
def get_revision_graph(self, revision_id=None):
940
"""Return a dictionary containing the revision graph.
941
942
:param revision_id: The revision_id to get a graph from. If None, then
943
the entire revision graph is returned. This is a deprecated mode of
944
operation and will be removed in the future.
945
:return: a dictionary of revision_id->revision_parents_list.
946
"""
947
# special case NULL_REVISION
948
if revision_id == _mod_revision.NULL_REVISION:
949
return {}
950
a_weave = self._get_revision_vf()
951
entire_graph = a_weave.get_graph()
952
if revision_id is None:
953
return a_weave.get_graph()
954
elif revision_id not in a_weave:
955
raise errors.NoSuchRevision(self, revision_id)
956
else:
957
# add what can be reached from revision_id
958
result = {}
959
pending = set([revision_id])
960
while len(pending) > 0:
961
node = pending.pop()
962
result[node] = a_weave.get_parents(node)
963
for revision_id in result[node]:
964
if revision_id not in result:
965
pending.add(revision_id)
966
return result
967
968
@needs_read_lock
969
def get_revision_graph_with_ghosts(self, revision_ids=None):
970
"""Return a graph of the revisions with ghosts marked as applicable.
971
972
:param revision_ids: an iterable of revisions to graph or None for all.
973
:return: a Graph object with the graph reachable from revision_ids.
974
"""
975
result = graph.Graph()
976
vf = self._get_revision_vf()
977
versions = set(vf.versions())
978
if not revision_ids:
979
pending = set(self.all_revision_ids())
980
required = set([])
981
else:
982
pending = set(revision_ids)
983
# special case NULL_REVISION
984
if _mod_revision.NULL_REVISION in pending:
985
pending.remove(_mod_revision.NULL_REVISION)
986
required = set(pending)
987
done = set([])
988
while len(pending):
989
revision_id = pending.pop()
990
if not revision_id in versions:
991
if revision_id in required:
992
raise errors.NoSuchRevision(self, revision_id)
993
# a ghost
994
result.add_ghost(revision_id)
995
# mark it as done so we don't try for it again.
996
done.add(revision_id)
997
continue
998
parent_ids = vf.get_parents_with_ghosts(revision_id)
999
for parent_id in parent_ids:
1000
# is this queued or done ?
1001
if (parent_id not in pending and
1002
parent_id not in done):
1003
# no, queue it.
1004
pending.add(parent_id)
1005
result.add_node(revision_id, parent_ids)
1006
done.add(revision_id)
1007
return result
1008
1009
def _get_revision_vf(self):
1010
""":return: a versioned file containing the revisions."""
1011
vf = self._revision_store.get_revision_file(self.get_transaction())
1012
return vf
1013
1014
@needs_write_lock
1015
def reconcile(self, other=None, thorough=False):
1016
"""Reconcile this repository."""
1017
from bzrlib.reconcile import KnitReconciler
1018
reconciler = KnitReconciler(self, thorough=thorough)
1019
reconciler.reconcile()
1020
return reconciler
1021
1022
def revision_parents(self, revision_id):
1023
return self._get_revision_vf().get_parents(revision_id)
1024
1025
1026
class KnitRepository2(KnitRepository):
1027
""""""
1028
def __init__(self, _format, a_bzrdir, control_files, _revision_store,
1029
control_store, text_store):
1030
KnitRepository.__init__(self, _format, a_bzrdir, control_files,
1031
_revision_store, control_store, text_store)
1032
self._serializer = xml6.serializer_v6
1033
1034
def deserialise_inventory(self, revision_id, xml):
1035
"""Transform the xml into an inventory object.
1036
1037
:param revision_id: The expected revision id of the inventory.
1038
:param xml: A serialised inventory.
1039
"""
1040
result = self._serializer.read_inventory_from_string(xml)
1041
assert result.root.revision is not None
1042
return result
1043
1044
def serialise_inventory(self, inv):
1045
"""Transform the inventory object into XML text.
1046
1047
:param revision_id: The expected revision id of the inventory.
1048
:param xml: A serialised inventory.
1049
"""
1050
assert inv.revision_id is not None
1051
assert inv.root.revision is not None
1052
return KnitRepository.serialise_inventory(self, inv)
1053
1054
def get_commit_builder(self, branch, parents, config, timestamp=None,
1055
timezone=None, committer=None, revprops=None,
1056
revision_id=None):
1057
"""Obtain a CommitBuilder for this repository.
1058
1059
:param branch: Branch to commit to.
1060
:param parents: Revision ids of the parents of the new revision.
1061
:param config: Configuration to use.
1062
:param timestamp: Optional timestamp recorded for commit.
1063
:param timezone: Optional timezone for timestamp.
1064
:param committer: Optional committer to set for commit.
1065
:param revprops: Optional dictionary of revision properties.
1066
:param revision_id: Optional revision id.
1067
"""
1068
return RootCommitBuilder(self, parents, config, timestamp, timezone,
1069
committer, revprops, revision_id)
1070
1071
1072
class RepositoryFormat(object):
1073
"""A repository format.
1074
1075
Formats provide three things:
1076
* An initialization routine to construct repository data on disk.
1077
* a format string which is used when the BzrDir supports versioned
1078
children.
1079
* an open routine which returns a Repository instance.
1080
1081
Formats are placed in an dict by their format string for reference
1082
during opening. These should be subclasses of RepositoryFormat
1083
for consistency.
1084
1085
Once a format is deprecated, just deprecate the initialize and open
1086
methods on the format class. Do not deprecate the object, as the
1087
object will be created every system load.
1088
1089
Common instance attributes:
1090
_matchingbzrdir - the bzrdir format that the repository format was
1091
originally written to work with. This can be used if manually
1092
constructing a bzrdir and repository, or more commonly for test suite
1093
parameterisation.
1094
"""
1095
1096
_default_format = None
1097
"""The default format used for new repositories."""
1098
1099
_formats = {}
1100
"""The known formats."""
1101
1102
def __str__(self):
1103
return "<%s>" % self.__class__.__name__
1104
1105
@classmethod
1106
def find_format(klass, a_bzrdir):
1107
"""Return the format for the repository object in a_bzrdir."""
1108
try:
1109
transport = a_bzrdir.get_repository_transport(None)
1110
format_string = transport.get("format").read()
1111
return klass._formats[format_string]
1112
except errors.NoSuchFile:
1113
raise errors.NoRepositoryPresent(a_bzrdir)
1114
except KeyError:
1115
raise errors.UnknownFormatError(format=format_string)
1116
1117
def _get_control_store(self, repo_transport, control_files):
1118
"""Return the control store for this repository."""
1119
raise NotImplementedError(self._get_control_store)
1120
1121
@classmethod
1122
def get_default_format(klass):
1123
"""Return the current default format."""
1124
return klass._default_format
1125
1126
def get_format_string(self):
1127
"""Return the ASCII format string that identifies this format.
1128
1129
Note that in pre format ?? repositories the format string is
1130
not permitted nor written to disk.
1131
"""
1132
raise NotImplementedError(self.get_format_string)
1133
1134
def get_format_description(self):
1135
"""Return the short description for this format."""
1136
raise NotImplementedError(self.get_format_description)
1137
1138
def _get_revision_store(self, repo_transport, control_files):
1139
"""Return the revision store object for this a_bzrdir."""
1140
raise NotImplementedError(self._get_revision_store)
1141
1142
def _get_text_rev_store(self,
1143
transport,
1144
control_files,
1145
name,
1146
compressed=True,
1147
prefixed=False,
1148
serializer=None):
1149
"""Common logic for getting a revision store for a repository.
1150
1151
see self._get_revision_store for the subclass-overridable method to
1152
get the store for a repository.
1153
"""
1154
from bzrlib.store.revision.text import TextRevisionStore
1155
dir_mode = control_files._dir_mode
1156
file_mode = control_files._file_mode
1157
text_store =TextStore(transport.clone(name),
1158
prefixed=prefixed,
1159
compressed=compressed,
1160
dir_mode=dir_mode,
1161
file_mode=file_mode)
1162
_revision_store = TextRevisionStore(text_store, serializer)
1163
return _revision_store
1164
1165
def _get_versioned_file_store(self,
1166
name,
1167
transport,
1168
control_files,
1169
prefixed=True,
1170
versionedfile_class=weave.WeaveFile,
1171
versionedfile_kwargs={},
1172
escaped=False):
1173
weave_transport = control_files._transport.clone(name)
1174
dir_mode = control_files._dir_mode
1175
file_mode = control_files._file_mode
1176
return VersionedFileStore(weave_transport, prefixed=prefixed,
1177
dir_mode=dir_mode,
1178
file_mode=file_mode,
1179
versionedfile_class=versionedfile_class,
1180
versionedfile_kwargs=versionedfile_kwargs,
1181
escaped=escaped)
1182
1183
def initialize(self, a_bzrdir, shared=False):
1184
"""Initialize a repository of this format in a_bzrdir.
1185
1186
:param a_bzrdir: The bzrdir to put the new repository in it.
1187
:param shared: The repository should be initialized as a sharable one.
1188
1189
This may raise UninitializableFormat if shared repository are not
1190
compatible the a_bzrdir.
1191
"""
1192
1193
def is_supported(self):
1194
"""Is this format supported?
1195
1196
Supported formats must be initializable and openable.
1197
Unsupported formats may not support initialization or committing or
1198
some other features depending on the reason for not being supported.
1199
"""
1200
return True
1201
1202
def check_conversion_target(self, target_format):
1203
raise NotImplementedError(self.check_conversion_target)
1204
1205
def open(self, a_bzrdir, _found=False):
1206
"""Return an instance of this format for the bzrdir a_bzrdir.
1207
1208
_found is a private parameter, do not use it.
1209
"""
1210
raise NotImplementedError(self.open)
1211
1212
@classmethod
1213
def register_format(klass, format):
1214
klass._formats[format.get_format_string()] = format
1215
1216
@classmethod
1217
def set_default_format(klass, format):
1218
klass._default_format = format
1219
1220
@classmethod
1221
def unregister_format(klass, format):
1222
assert klass._formats[format.get_format_string()] is format
1223
del klass._formats[format.get_format_string()]
1224
1225
1226
class PreSplitOutRepositoryFormat(RepositoryFormat):
1227
"""Base class for the pre split out repository formats."""
1228
1229
rich_root_data = False
1230
1231
def initialize(self, a_bzrdir, shared=False, _internal=False):
1232
"""Create a weave repository.
1233
1234
TODO: when creating split out bzr branch formats, move this to a common
1235
base for Format5, Format6. or something like that.
1236
"""
1237
if shared:
1238
raise errors.IncompatibleFormat(self, a_bzrdir._format)
1239
1240
if not _internal:
1241
# always initialized when the bzrdir is.
1242
return self.open(a_bzrdir, _found=True)
1243
1244
# Create an empty weave
1245
sio = StringIO()
1246
weavefile.write_weave_v5(weave.Weave(), sio)
1247
empty_weave = sio.getvalue()
1248
1249
mutter('creating repository in %s.', a_bzrdir.transport.base)
1250
dirs = ['revision-store', 'weaves']
1251
files = [('inventory.weave', StringIO(empty_weave)),
1252
]
1253
1254
# FIXME: RBC 20060125 don't peek under the covers
1255
# NB: no need to escape relative paths that are url safe.
1256
control_files = lockable_files.LockableFiles(a_bzrdir.transport,
1257
'branch-lock', lockable_files.TransportLock)
1258
control_files.create_lock()
1259
control_files.lock_write()
1260
control_files._transport.mkdir_multi(dirs,
1261
mode=control_files._dir_mode)
1262
try:
1263
for file, content in files:
1264
control_files.put(file, content)
1265
finally:
1266
control_files.unlock()
1267
return self.open(a_bzrdir, _found=True)
1268
1269
def _get_control_store(self, repo_transport, control_files):
1270
"""Return the control store for this repository."""
1271
return self._get_versioned_file_store('',
1272
repo_transport,
1273
control_files,
1274
prefixed=False)
1275
1276
def _get_text_store(self, transport, control_files):
1277
"""Get a store for file texts for this format."""
1278
raise NotImplementedError(self._get_text_store)
1279
1280
def open(self, a_bzrdir, _found=False):
1281
"""See RepositoryFormat.open()."""
1282
if not _found:
1283
# we are being called directly and must probe.
1284
raise NotImplementedError
1285
1286
repo_transport = a_bzrdir.get_repository_transport(None)
1287
control_files = a_bzrdir._control_files
1288
text_store = self._get_text_store(repo_transport, control_files)
1289
control_store = self._get_control_store(repo_transport, control_files)
1290
_revision_store = self._get_revision_store(repo_transport, control_files)
1291
return AllInOneRepository(_format=self,
1292
a_bzrdir=a_bzrdir,
1293
_revision_store=_revision_store,
1294
control_store=control_store,
1295
text_store=text_store)
1296
1297
def check_conversion_target(self, target_format):
1298
pass
1299
1300
1301
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1302
"""Bzr repository format 4.
1303
1304
This repository format has:
1305
- flat stores
1306
- TextStores for texts, inventories,revisions.
1307
1308
This format is deprecated: it indexes texts using a text id which is
1309
removed in format 5; initialization and write support for this format
1310
has been removed.
1311
"""
1312
1313
def __init__(self):
1314
super(RepositoryFormat4, self).__init__()
1315
self._matchingbzrdir = bzrdir.BzrDirFormat4()
1316
1317
def get_format_description(self):
1318
"""See RepositoryFormat.get_format_description()."""
1319
return "Repository format 4"
1320
1321
def initialize(self, url, shared=False, _internal=False):
1322
"""Format 4 branches cannot be created."""
1323
raise errors.UninitializableFormat(self)
1324
1325
def is_supported(self):
1326
"""Format 4 is not supported.
1327
1328
It is not supported because the model changed from 4 to 5 and the
1329
conversion logic is expensive - so doing it on the fly was not
1330
feasible.
1331
"""
1332
return False
1333
1334
def _get_control_store(self, repo_transport, control_files):
1335
"""Format 4 repositories have no formal control store at this point.
1336
1337
This will cause any control-file-needing apis to fail - this is desired.
1338
"""
1339
return None
1340
1341
def _get_revision_store(self, repo_transport, control_files):
1342
"""See RepositoryFormat._get_revision_store()."""
1343
from bzrlib.xml4 import serializer_v4
1344
return self._get_text_rev_store(repo_transport,
1345
control_files,
1346
'revision-store',
1347
serializer=serializer_v4)
1348
1349
def _get_text_store(self, transport, control_files):
1350
"""See RepositoryFormat._get_text_store()."""
1351
1352
1353
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1354
"""Bzr control format 5.
1355
1356
This repository format has:
1357
- weaves for file texts and inventory
1358
- flat stores
1359
- TextStores for revisions and signatures.
1360
"""
1361
1362
def __init__(self):
1363
super(RepositoryFormat5, self).__init__()
1364
self._matchingbzrdir = bzrdir.BzrDirFormat5()
1365
1366
def get_format_description(self):
1367
"""See RepositoryFormat.get_format_description()."""
1368
return "Weave repository format 5"
1369
1370
def _get_revision_store(self, repo_transport, control_files):
1371
"""See RepositoryFormat._get_revision_store()."""
1372
"""Return the revision store object for this a_bzrdir."""
1373
return self._get_text_rev_store(repo_transport,
1374
control_files,
1375
'revision-store',
1376
compressed=False)
1377
1378
def _get_text_store(self, transport, control_files):
1379
"""See RepositoryFormat._get_text_store()."""
1380
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1381
1382
1383
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1384
"""Bzr control format 6.
1385
1386
This repository format has:
1387
- weaves for file texts and inventory
1388
- hash subdirectory based stores.
1389
- TextStores for revisions and signatures.
1390
"""
1391
1392
def __init__(self):
1393
super(RepositoryFormat6, self).__init__()
1394
self._matchingbzrdir = bzrdir.BzrDirFormat6()
1395
1396
def get_format_description(self):
1397
"""See RepositoryFormat.get_format_description()."""
1398
return "Weave repository format 6"
1399
1400
def _get_revision_store(self, repo_transport, control_files):
1401
"""See RepositoryFormat._get_revision_store()."""
1402
return self._get_text_rev_store(repo_transport,
1403
control_files,
1404
'revision-store',
1405
compressed=False,
1406
prefixed=True)
1407
1408
def _get_text_store(self, transport, control_files):
1409
"""See RepositoryFormat._get_text_store()."""
1410
return self._get_versioned_file_store('weaves', transport, control_files)
1411
1412
1413
class MetaDirRepositoryFormat(RepositoryFormat):
1414
"""Common base class for the new repositories using the metadir layout."""
1415
1416
rich_root_data = False
1417
1418
def __init__(self):
1419
super(MetaDirRepositoryFormat, self).__init__()
1420
self._matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1421
1422
def _create_control_files(self, a_bzrdir):
1423
"""Create the required files and the initial control_files object."""
1424
# FIXME: RBC 20060125 don't peek under the covers
1425
# NB: no need to escape relative paths that are url safe.
1426
repository_transport = a_bzrdir.get_repository_transport(self)
1427
control_files = lockable_files.LockableFiles(repository_transport,
1428
'lock', lockdir.LockDir)
1429
control_files.create_lock()
1430
return control_files
1431
1432
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1433
"""Upload the initial blank content."""
1434
control_files = self._create_control_files(a_bzrdir)
1435
control_files.lock_write()
1436
try:
1437
control_files._transport.mkdir_multi(dirs,
1438
mode=control_files._dir_mode)
1439
for file, content in files:
1440
control_files.put(file, content)
1441
for file, content in utf8_files:
1442
control_files.put_utf8(file, content)
1443
if shared == True:
1444
control_files.put_utf8('shared-storage', '')
1445
finally:
1446
control_files.unlock()
1447
1448
1449
class RepositoryFormat7(MetaDirRepositoryFormat):
1450
"""Bzr repository 7.
1451
1452
This repository format has:
1453
- weaves for file texts and inventory
1454
- hash subdirectory based stores.
1455
- TextStores for revisions and signatures.
1456
- a format marker of its own
1457
- an optional 'shared-storage' flag
1458
- an optional 'no-working-trees' flag
1459
"""
1460
1461
def _get_control_store(self, repo_transport, control_files):
1462
"""Return the control store for this repository."""
1463
return self._get_versioned_file_store('',
1464
repo_transport,
1465
control_files,
1466
prefixed=False)
1467
1468
def get_format_string(self):
1469
"""See RepositoryFormat.get_format_string()."""
1470
return "Bazaar-NG Repository format 7"
1471
1472
def get_format_description(self):
1473
"""See RepositoryFormat.get_format_description()."""
1474
return "Weave repository format 7"
1475
1476
def check_conversion_target(self, target_format):
1477
pass
1478
1479
def _get_revision_store(self, repo_transport, control_files):
1480
"""See RepositoryFormat._get_revision_store()."""
1481
return self._get_text_rev_store(repo_transport,
1482
control_files,
1483
'revision-store',
1484
compressed=False,
1485
prefixed=True,
1486
)
1487
1488
def _get_text_store(self, transport, control_files):
1489
"""See RepositoryFormat._get_text_store()."""
1490
return self._get_versioned_file_store('weaves',
1491
transport,
1492
control_files)
1493
1494
def initialize(self, a_bzrdir, shared=False):
1495
"""Create a weave repository.
1496
1497
:param shared: If true the repository will be initialized as a shared
1498
repository.
1499
"""
1500
# Create an empty weave
1501
sio = StringIO()
1502
weavefile.write_weave_v5(weave.Weave(), sio)
1503
empty_weave = sio.getvalue()
1504
1505
mutter('creating repository in %s.', a_bzrdir.transport.base)
1506
dirs = ['revision-store', 'weaves']
1507
files = [('inventory.weave', StringIO(empty_weave)),
1508
]
1509
utf8_files = [('format', self.get_format_string())]
1510
1511
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1512
return self.open(a_bzrdir=a_bzrdir, _found=True)
1513
1514
def open(self, a_bzrdir, _found=False, _override_transport=None):
1515
"""See RepositoryFormat.open().
1516
1517
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1518
repository at a slightly different url
1519
than normal. I.e. during 'upgrade'.
1520
"""
1521
if not _found:
1522
format = RepositoryFormat.find_format(a_bzrdir)
1523
assert format.__class__ == self.__class__
1524
if _override_transport is not None:
1525
repo_transport = _override_transport
1526
else:
1527
repo_transport = a_bzrdir.get_repository_transport(None)
1528
control_files = lockable_files.LockableFiles(repo_transport,
1529
'lock', lockdir.LockDir)
1530
text_store = self._get_text_store(repo_transport, control_files)
1531
control_store = self._get_control_store(repo_transport, control_files)
1532
_revision_store = self._get_revision_store(repo_transport, control_files)
1533
return MetaDirRepository(_format=self,
1534
a_bzrdir=a_bzrdir,
1535
control_files=control_files,
1536
_revision_store=_revision_store,
1537
control_store=control_store,
1538
text_store=text_store)
1539
1540
1541
class RepositoryFormatKnit(MetaDirRepositoryFormat):
1542
"""Bzr repository knit format (generalized).
1543
1544
This repository format has:
1545
- knits for file texts and inventory
1546
- hash subdirectory based stores.
1547
- knits for revisions and signatures
1548
- TextStores for revisions and signatures.
1549
- a format marker of its own
1550
- an optional 'shared-storage' flag
1551
- an optional 'no-working-trees' flag
1552
- a LockDir lock
1553
"""
1554
1555
def _get_control_store(self, repo_transport, control_files):
1556
"""Return the control store for this repository."""
1557
return VersionedFileStore(
1558
repo_transport,
1559
prefixed=False,
1560
file_mode=control_files._file_mode,
1561
versionedfile_class=knit.KnitVersionedFile,
1562
versionedfile_kwargs={'factory':knit.KnitPlainFactory()},
1563
)
1564
1565
def _get_revision_store(self, repo_transport, control_files):
1566
"""See RepositoryFormat._get_revision_store()."""
1567
from bzrlib.store.revision.knit import KnitRevisionStore
1568
versioned_file_store = VersionedFileStore(
1569
repo_transport,
1570
file_mode=control_files._file_mode,
1571
prefixed=False,
1572
precious=True,
1573
versionedfile_class=knit.KnitVersionedFile,
1574
versionedfile_kwargs={'delta':False,
1575
'factory':knit.KnitPlainFactory(),
1576
},
1577
escaped=True,
1578
)
1579
return KnitRevisionStore(versioned_file_store)
1580
1581
def _get_text_store(self, transport, control_files):
1582
"""See RepositoryFormat._get_text_store()."""
1583
return self._get_versioned_file_store('knits',
1584
transport,
1585
control_files,
1586
versionedfile_class=knit.KnitVersionedFile,
1587
versionedfile_kwargs={
1588
'create_parent_dir':True,
1589
'delay_create':True,
1590
'dir_mode':control_files._dir_mode,
1591
},
1592
escaped=True)
1593
1594
def initialize(self, a_bzrdir, shared=False):
1595
"""Create a knit format 1 repository.
1596
1597
:param a_bzrdir: bzrdir to contain the new repository; must already
1598
be initialized.
1599
:param shared: If true the repository will be initialized as a shared
1600
repository.
1601
"""
1602
mutter('creating repository in %s.', a_bzrdir.transport.base)
1603
dirs = ['revision-store', 'knits']
1604
files = []
1605
utf8_files = [('format', self.get_format_string())]
1606
1607
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1608
repo_transport = a_bzrdir.get_repository_transport(None)
1609
control_files = lockable_files.LockableFiles(repo_transport,
1610
'lock', lockdir.LockDir)
1611
control_store = self._get_control_store(repo_transport, control_files)
1612
transaction = transactions.WriteTransaction()
1613
# trigger a write of the inventory store.
1614
control_store.get_weave_or_empty('inventory', transaction)
1615
_revision_store = self._get_revision_store(repo_transport, control_files)
1616
_revision_store.has_revision_id('A', transaction)
1617
_revision_store.get_signature_file(transaction)
1618
return self.open(a_bzrdir=a_bzrdir, _found=True)
1619
1620
def open(self, a_bzrdir, _found=False, _override_transport=None):
1621
"""See RepositoryFormat.open().
1622
1623
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1624
repository at a slightly different url
1625
than normal. I.e. during 'upgrade'.
1626
"""
1627
if not _found:
1628
format = RepositoryFormat.find_format(a_bzrdir)
1629
assert format.__class__ == self.__class__
1630
if _override_transport is not None:
1631
repo_transport = _override_transport
1632
else:
1633
repo_transport = a_bzrdir.get_repository_transport(None)
1634
control_files = lockable_files.LockableFiles(repo_transport,
1635
'lock', lockdir.LockDir)
1636
text_store = self._get_text_store(repo_transport, control_files)
1637
control_store = self._get_control_store(repo_transport, control_files)
1638
_revision_store = self._get_revision_store(repo_transport, control_files)
1639
return KnitRepository(_format=self,
1640
a_bzrdir=a_bzrdir,
1641
control_files=control_files,
1642
_revision_store=_revision_store,
1643
control_store=control_store,
1644
text_store=text_store)
1645
1646
1647
class RepositoryFormatKnit1(RepositoryFormatKnit):
1648
"""Bzr repository knit format 1.
1649
1650
This repository format has:
1651
- knits for file texts and inventory
1652
- hash subdirectory based stores.
1653
- knits for revisions and signatures
1654
- TextStores for revisions and signatures.
1655
- a format marker of its own
1656
- an optional 'shared-storage' flag
1657
- an optional 'no-working-trees' flag
1658
- a LockDir lock
1659
1660
This format was introduced in bzr 0.8.
1661
"""
1662
def get_format_string(self):
1663
"""See RepositoryFormat.get_format_string()."""
1664
return "Bazaar-NG Knit Repository Format 1"
1665
1666
def get_format_description(self):
1667
"""See RepositoryFormat.get_format_description()."""
1668
return "Knit repository format 1"
1669
1670
def check_conversion_target(self, target_format):
1671
pass
1672
1673
1674
class RepositoryFormatKnit2(RepositoryFormatKnit):
1675
"""Bzr repository knit format 2.
1676
1677
THIS FORMAT IS EXPERIMENTAL
1678
This repository format has:
1679
- knits for file texts and inventory
1680
- hash subdirectory based stores.
1681
- knits for revisions and signatures
1682
- TextStores for revisions and signatures.
1683
- a format marker of its own
1684
- an optional 'shared-storage' flag
1685
- an optional 'no-working-trees' flag
1686
- a LockDir lock
1687
- Support for recording full info about the tree root
1688
1689
"""
1690
1691
rich_root_data = True
1692
1693
def get_format_string(self):
1694
"""See RepositoryFormat.get_format_string()."""
1695
return "Bazaar Knit Repository Format 2\n"
1696
1697
def get_format_description(self):
1698
"""See RepositoryFormat.get_format_description()."""
1699
return "Knit repository format 2"
1700
1701
def check_conversion_target(self, target_format):
1702
if not target_format.rich_root_data:
1703
raise errors.BadConversionTarget(
1704
'Does not support rich root data.', target_format)
1705
1706
def open(self, a_bzrdir, _found=False, _override_transport=None):
1707
"""See RepositoryFormat.open().
1708
1709
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1710
repository at a slightly different url
1711
than normal. I.e. during 'upgrade'.
1712
"""
1713
if not _found:
1714
format = RepositoryFormat.find_format(a_bzrdir)
1715
assert format.__class__ == self.__class__
1716
if _override_transport is not None:
1717
repo_transport = _override_transport
1718
else:
1719
repo_transport = a_bzrdir.get_repository_transport(None)
1720
control_files = lockable_files.LockableFiles(repo_transport, 'lock',
1721
lockdir.LockDir)
1722
text_store = self._get_text_store(repo_transport, control_files)
1723
control_store = self._get_control_store(repo_transport, control_files)
1724
_revision_store = self._get_revision_store(repo_transport, control_files)
1725
return KnitRepository2(_format=self,
1726
a_bzrdir=a_bzrdir,
1727
control_files=control_files,
1728
_revision_store=_revision_store,
1729
control_store=control_store,
1730
text_store=text_store)
1731
1732
1733
1734
# formats which have no format string are not discoverable
1735
# and not independently creatable, so are not registered.
1736
RepositoryFormat.register_format(RepositoryFormat7())
1737
_default_format = RepositoryFormatKnit1()
1738
RepositoryFormat.register_format(_default_format)
1739
RepositoryFormat.register_format(RepositoryFormatKnit2())
1740
RepositoryFormat.set_default_format(_default_format)
1741
_legacy_formats = [RepositoryFormat4(),
1742
RepositoryFormat5(),
1743
RepositoryFormat6()]
1744
1745
1746
class InterRepository(InterObject):
1747
"""This class represents operations taking place between two repositories.
1748
1749
Its instances have methods like copy_content and fetch, and contain
1750
references to the source and target repositories these operations can be
1751
carried out on.
1752
1753
Often we will provide convenience methods on 'repository' which carry out
1754
operations with another repository - they will always forward to
1755
InterRepository.get(other).method_name(parameters).
1756
"""
1757
1758
_optimisers = []
1759
"""The available optimised InterRepository types."""
1760
1761
def copy_content(self, revision_id=None, basis=None):
1762
raise NotImplementedError(self.copy_content)
1763
1764
def fetch(self, revision_id=None, pb=None):
1765
"""Fetch the content required to construct revision_id.
1766
1767
The content is copied from self.source to self.target.
1768
1769
:param revision_id: if None all content is copied, if NULL_REVISION no
1770
content is copied.
1771
:param pb: optional progress bar to use for progress reports. If not
1772
provided a default one will be created.
1773
1774
Returns the copied revision count and the failed revisions in a tuple:
1775
(copied, failures).
1776
"""
1777
raise NotImplementedError(self.fetch)
1778
1779
@needs_read_lock
1780
def missing_revision_ids(self, revision_id=None):
1781
"""Return the revision ids that source has that target does not.
1782
1783
These are returned in topological order.
1784
1785
:param revision_id: only return revision ids included by this
1786
revision_id.
1787
"""
1788
# generic, possibly worst case, slow code path.
1789
target_ids = set(self.target.all_revision_ids())
1790
if revision_id is not None:
1791
source_ids = self.source.get_ancestry(revision_id)
1792
assert source_ids[0] is None
1793
source_ids.pop(0)
1794
else:
1795
source_ids = self.source.all_revision_ids()
1796
result_set = set(source_ids).difference(target_ids)
1797
# this may look like a no-op: its not. It preserves the ordering
1798
# other_ids had while only returning the members from other_ids
1799
# that we've decided we need.
1800
return [rev_id for rev_id in source_ids if rev_id in result_set]
1801
1802
1803
class InterSameDataRepository(InterRepository):
1804
"""Code for converting between repositories that represent the same data.
1805
1806
Data format and model must match for this to work.
1807
"""
1808
1809
_matching_repo_format = RepositoryFormat4()
1810
"""Repository format for testing with."""
1811
1812
@staticmethod
1813
def is_compatible(source, target):
1814
if not isinstance(source, Repository):
1815
return False
1816
if not isinstance(target, Repository):
1817
return False
1818
if source._format.rich_root_data == target._format.rich_root_data:
1819
return True
1820
else:
1821
return False
1822
1823
@needs_write_lock
1824
def copy_content(self, revision_id=None, basis=None):
1825
"""Make a complete copy of the content in self into destination.
1826
1827
This is a destructive operation! Do not use it on existing
1828
repositories.
1829
1830
:param revision_id: Only copy the content needed to construct
1831
revision_id and its parents.
1832
:param basis: Copy the needed data preferentially from basis.
1833
"""
1834
try:
1835
self.target.set_make_working_trees(self.source.make_working_trees())
1836
except NotImplementedError:
1837
pass
1838
# grab the basis available data
1839
if basis is not None:
1840
self.target.fetch(basis, revision_id=revision_id)
1841
# but don't bother fetching if we have the needed data now.
1842
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1843
self.target.has_revision(revision_id)):
1844
return
1845
self.target.fetch(self.source, revision_id=revision_id)
1846
1847
@needs_write_lock
1848
def fetch(self, revision_id=None, pb=None):
1849
"""See InterRepository.fetch()."""
1850
from bzrlib.fetch import GenericRepoFetcher
1851
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1852
self.source, self.source._format, self.target,
1853
self.target._format)
1854
f = GenericRepoFetcher(to_repository=self.target,
1855
from_repository=self.source,
1856
last_revision=revision_id,
1857
pb=pb)
1858
return f.count_copied, f.failed_revisions
1859
1860
1861
class InterWeaveRepo(InterSameDataRepository):
1862
"""Optimised code paths between Weave based repositories."""
1863
1864
_matching_repo_format = RepositoryFormat7()
1865
"""Repository format for testing with."""
1866
1867
@staticmethod
1868
def is_compatible(source, target):
1869
"""Be compatible with known Weave formats.
1870
1871
We don't test for the stores being of specific types because that
1872
could lead to confusing results, and there is no need to be
1873
overly general.
1874
"""
1875
try:
1876
return (isinstance(source._format, (RepositoryFormat5,
1877
RepositoryFormat6,
1878
RepositoryFormat7)) and
1879
isinstance(target._format, (RepositoryFormat5,
1880
RepositoryFormat6,
1881
RepositoryFormat7)))
1882
except AttributeError:
1883
return False
1884
1885
@needs_write_lock
1886
def copy_content(self, revision_id=None, basis=None):
1887
"""See InterRepository.copy_content()."""
1888
# weave specific optimised path:
1889
if basis is not None:
1890
# copy the basis in, then fetch remaining data.
1891
basis.copy_content_into(self.target, revision_id)
1892
# the basis copy_content_into could miss-set this.
1893
try:
1894
self.target.set_make_working_trees(self.source.make_working_trees())
1895
except NotImplementedError:
1896
pass
1897
self.target.fetch(self.source, revision_id=revision_id)
1898
else:
1899
try:
1900
self.target.set_make_working_trees(self.source.make_working_trees())
1901
except NotImplementedError:
1902
pass
1903
# FIXME do not peek!
1904
if self.source.control_files._transport.listable():
1905
pb = ui.ui_factory.nested_progress_bar()
1906
try:
1907
self.target.weave_store.copy_all_ids(
1908
self.source.weave_store,
1909
pb=pb,
1910
from_transaction=self.source.get_transaction(),
1911
to_transaction=self.target.get_transaction())
1912
pb.update('copying inventory', 0, 1)
1913
self.target.control_weaves.copy_multi(
1914
self.source.control_weaves, ['inventory'],
1915
from_transaction=self.source.get_transaction(),
1916
to_transaction=self.target.get_transaction())
1917
self.target._revision_store.text_store.copy_all_ids(
1918
self.source._revision_store.text_store,
1919
pb=pb)
1920
finally:
1921
pb.finished()
1922
else:
1923
self.target.fetch(self.source, revision_id=revision_id)
1924
1925
@needs_write_lock
1926
def fetch(self, revision_id=None, pb=None):
1927
"""See InterRepository.fetch()."""
1928
from bzrlib.fetch import GenericRepoFetcher
1929
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1930
self.source, self.source._format, self.target, self.target._format)
1931
f = GenericRepoFetcher(to_repository=self.target,
1932
from_repository=self.source,
1933
last_revision=revision_id,
1934
pb=pb)
1935
return f.count_copied, f.failed_revisions
1936
1937
@needs_read_lock
1938
def missing_revision_ids(self, revision_id=None):
1939
"""See InterRepository.missing_revision_ids()."""
1940
# we want all revisions to satisfy revision_id in source.
1941
# but we don't want to stat every file here and there.
1942
# we want then, all revisions other needs to satisfy revision_id
1943
# checked, but not those that we have locally.
1944
# so the first thing is to get a subset of the revisions to
1945
# satisfy revision_id in source, and then eliminate those that
1946
# we do already have.
1947
# this is slow on high latency connection to self, but as as this
1948
# disk format scales terribly for push anyway due to rewriting
1949
# inventory.weave, this is considered acceptable.
1950
# - RBC 20060209
1951
if revision_id is not None:
1952
source_ids = self.source.get_ancestry(revision_id)
1953
assert source_ids[0] is None
1954
source_ids.pop(0)
1955
else:
1956
source_ids = self.source._all_possible_ids()
1957
source_ids_set = set(source_ids)
1958
# source_ids is the worst possible case we may need to pull.
1959
# now we want to filter source_ids against what we actually
1960
# have in target, but don't try to check for existence where we know
1961
# we do not have a revision as that would be pointless.
1962
target_ids = set(self.target._all_possible_ids())
1963
possibly_present_revisions = target_ids.intersection(source_ids_set)
1964
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1965
required_revisions = source_ids_set.difference(actually_present_revisions)
1966
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1967
if revision_id is not None:
1968
# we used get_ancestry to determine source_ids then we are assured all
1969
# revisions referenced are present as they are installed in topological order.
1970
# and the tip revision was validated by get_ancestry.
1971
return required_topo_revisions
1972
else:
1973
# if we just grabbed the possibly available ids, then
1974
# we only have an estimate of whats available and need to validate
1975
# that against the revision records.
1976
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1977
1978
1979
class InterKnitRepo(InterSameDataRepository):
1980
"""Optimised code paths between Knit based repositories."""
1981
1982
_matching_repo_format = RepositoryFormatKnit1()
1983
"""Repository format for testing with."""
1984
1985
@staticmethod
1986
def is_compatible(source, target):
1987
"""Be compatible with known Knit formats.
1988
1989
We don't test for the stores being of specific types because that
1990
could lead to confusing results, and there is no need to be
1991
overly general.
1992
"""
1993
try:
1994
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1995
isinstance(target._format, (RepositoryFormatKnit1)))
1996
except AttributeError:
1997
return False
1998
1999
@needs_write_lock
2000
def fetch(self, revision_id=None, pb=None):
2001
"""See InterRepository.fetch()."""
2002
from bzrlib.fetch import KnitRepoFetcher
2003
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2004
self.source, self.source._format, self.target, self.target._format)
2005
f = KnitRepoFetcher(to_repository=self.target,
2006
from_repository=self.source,
2007
last_revision=revision_id,
2008
pb=pb)
2009
return f.count_copied, f.failed_revisions
2010
2011
@needs_read_lock
2012
def missing_revision_ids(self, revision_id=None):
2013
"""See InterRepository.missing_revision_ids()."""
2014
if revision_id is not None:
2015
source_ids = self.source.get_ancestry(revision_id)
2016
assert source_ids[0] is None
2017
source_ids.pop(0)
2018
else:
2019
source_ids = self.source._all_possible_ids()
2020
source_ids_set = set(source_ids)
2021
# source_ids is the worst possible case we may need to pull.
2022
# now we want to filter source_ids against what we actually
2023
# have in target, but don't try to check for existence where we know
2024
# we do not have a revision as that would be pointless.
2025
target_ids = set(self.target._all_possible_ids())
2026
possibly_present_revisions = target_ids.intersection(source_ids_set)
2027
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2028
required_revisions = source_ids_set.difference(actually_present_revisions)
2029
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2030
if revision_id is not None:
2031
# we used get_ancestry to determine source_ids then we are assured all
2032
# revisions referenced are present as they are installed in topological order.
2033
# and the tip revision was validated by get_ancestry.
2034
return required_topo_revisions
2035
else:
2036
# if we just grabbed the possibly available ids, then
2037
# we only have an estimate of whats available and need to validate
2038
# that against the revision records.
2039
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2040
2041
2042
class InterModel1and2(InterRepository):
2043
2044
_matching_repo_format = None
2045
2046
@staticmethod
2047
def is_compatible(source, target):
2048
if not isinstance(source, Repository):
2049
return False
2050
if not isinstance(target, Repository):
2051
return False
2052
if not source._format.rich_root_data and target._format.rich_root_data:
2053
return True
2054
else:
2055
return False
2056
2057
@needs_write_lock
2058
def fetch(self, revision_id=None, pb=None):
2059
"""See InterRepository.fetch()."""
2060
from bzrlib.fetch import Model1toKnit2Fetcher
2061
f = Model1toKnit2Fetcher(to_repository=self.target,
2062
from_repository=self.source,
2063
last_revision=revision_id,
2064
pb=pb)
2065
return f.count_copied, f.failed_revisions
2066
2067
@needs_write_lock
2068
def copy_content(self, revision_id=None, basis=None):
2069
"""Make a complete copy of the content in self into destination.
2070
2071
This is a destructive operation! Do not use it on existing
2072
repositories.
2073
2074
:param revision_id: Only copy the content needed to construct
2075
revision_id and its parents.
2076
:param basis: Copy the needed data preferentially from basis.
2077
"""
2078
try:
2079
self.target.set_make_working_trees(self.source.make_working_trees())
2080
except NotImplementedError:
2081
pass
2082
# grab the basis available data
2083
if basis is not None:
2084
self.target.fetch(basis, revision_id=revision_id)
2085
# but don't bother fetching if we have the needed data now.
2086
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2087
self.target.has_revision(revision_id)):
2088
return
2089
self.target.fetch(self.source, revision_id=revision_id)
2090
2091
2092
class InterKnit1and2(InterKnitRepo):
2093
2094
_matching_repo_format = None
2095
2096
@staticmethod
2097
def is_compatible(source, target):
2098
"""Be compatible with Knit1 source and Knit2 target"""
2099
try:
2100
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2101
isinstance(target._format, (RepositoryFormatKnit2)))
2102
except AttributeError:
2103
return False
2104
2105
@needs_write_lock
2106
def fetch(self, revision_id=None, pb=None):
2107
"""See InterRepository.fetch()."""
2108
from bzrlib.fetch import Knit1to2Fetcher
2109
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2110
self.source, self.source._format, self.target,
2111
self.target._format)
2112
f = Knit1to2Fetcher(to_repository=self.target,
2113
from_repository=self.source,
2114
last_revision=revision_id,
2115
pb=pb)
2116
return f.count_copied, f.failed_revisions
2117
2118
2119
InterRepository.register_optimiser(InterSameDataRepository)
2120
InterRepository.register_optimiser(InterWeaveRepo)
2121
InterRepository.register_optimiser(InterKnitRepo)
2122
InterRepository.register_optimiser(InterModel1and2)
2123
InterRepository.register_optimiser(InterKnit1and2)
2124
2125
2126
class RepositoryTestProviderAdapter(object):
2127
"""A tool to generate a suite testing multiple repository formats at once.
2128
2129
This is done by copying the test once for each transport and injecting
2130
the transport_server, transport_readonly_server, and bzrdir_format and
2131
repository_format classes into each copy. Each copy is also given a new id()
2132
to make it easy to identify.
2133
"""
2134
2135
def __init__(self, transport_server, transport_readonly_server, formats):
2136
self._transport_server = transport_server
2137
self._transport_readonly_server = transport_readonly_server
2138
self._formats = formats
2139
2140
def adapt(self, test):
2141
result = unittest.TestSuite()
2142
for repository_format, bzrdir_format in self._formats:
2143
new_test = deepcopy(test)
2144
new_test.transport_server = self._transport_server
2145
new_test.transport_readonly_server = self._transport_readonly_server
2146
new_test.bzrdir_format = bzrdir_format
2147
new_test.repository_format = repository_format
2148
def make_new_test_id():
2149
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
2150
return lambda: new_id
2151
new_test.id = make_new_test_id()
2152
result.addTest(new_test)
2153
return result
2154
2155
2156
class InterRepositoryTestProviderAdapter(object):
2157
"""A tool to generate a suite testing multiple inter repository formats.
2158
2159
This is done by copying the test once for each interrepo provider and injecting
2160
the transport_server, transport_readonly_server, repository_format and
2161
repository_to_format classes into each copy.
2162
Each copy is also given a new id() to make it easy to identify.
2163
"""
2164
2165
def __init__(self, transport_server, transport_readonly_server, formats):
2166
self._transport_server = transport_server
2167
self._transport_readonly_server = transport_readonly_server
2168
self._formats = formats
2169
2170
def adapt(self, test):
2171
result = unittest.TestSuite()
2172
for interrepo_class, repository_format, repository_format_to in self._formats:
2173
new_test = deepcopy(test)
2174
new_test.transport_server = self._transport_server
2175
new_test.transport_readonly_server = self._transport_readonly_server
2176
new_test.interrepo_class = interrepo_class
2177
new_test.repository_format = repository_format
2178
new_test.repository_format_to = repository_format_to
2179
def make_new_test_id():
2180
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
2181
return lambda: new_id
2182
new_test.id = make_new_test_id()
2183
result.addTest(new_test)
2184
return result
2185
2186
@staticmethod
2187
def default_test_list():
2188
"""Generate the default list of interrepo permutations to test."""
2189
result = []
2190
# test the default InterRepository between format 6 and the current
2191
# default format.
2192
# XXX: robertc 20060220 reinstate this when there are two supported
2193
# formats which do not have an optimal code path between them.
2194
#result.append((InterRepository,
2195
# RepositoryFormat6(),
2196
# RepositoryFormatKnit1()))
2197
for optimiser in InterRepository._optimisers:
2198
if optimiser._matching_repo_format is not None:
2199
result.append((optimiser,
2200
optimiser._matching_repo_format,
2201
optimiser._matching_repo_format
2202
))
2203
# if there are specific combinations we want to use, we can add them
2204
# here.
2205
result.append((InterModel1and2, RepositoryFormat5(),
2206
RepositoryFormatKnit2()))
2207
result.append((InterKnit1and2, RepositoryFormatKnit1(),
2208
RepositoryFormatKnit2()))
2209
return result
2210
2211
2212
class CopyConverter(object):
2213
"""A repository conversion tool which just performs a copy of the content.
2214
2215
This is slow but quite reliable.
2216
"""
2217
2218
def __init__(self, target_format):
2219
"""Create a CopyConverter.
2220
2221
:param target_format: The format the resulting repository should be.
2222
"""
2223
self.target_format = target_format
2224
2225
def convert(self, repo, pb):
2226
"""Perform the conversion of to_convert, giving feedback via pb.
2227
2228
:param to_convert: The disk object to convert.
2229
:param pb: a progress bar to use for progress information.
2230
"""
2231
self.pb = pb
2232
self.count = 0
2233
self.total = 4
2234
# this is only useful with metadir layouts - separated repo content.
2235
# trigger an assertion if not such
2236
repo._format.get_format_string()
2237
self.repo_dir = repo.bzrdir
2238
self.step('Moving repository to repository.backup')
2239
self.repo_dir.transport.move('repository', 'repository.backup')
2240
backup_transport = self.repo_dir.transport.clone('repository.backup')
2241
repo._format.check_conversion_target(self.target_format)
2242
self.source_repo = repo._format.open(self.repo_dir,
2243
_found=True,
2244
_override_transport=backup_transport)
2245
self.step('Creating new repository')
2246
converted = self.target_format.initialize(self.repo_dir,
2247
self.source_repo.is_shared())
2248
converted.lock_write()
2249
try:
2250
self.step('Copying content into repository.')
2251
self.source_repo.copy_content_into(converted)
2252
finally:
2253
converted.unlock()
2254
self.step('Deleting old repository content.')
2255
self.repo_dir.transport.delete_tree('repository.backup')
2256
self.pb.note('repository converted')
2257
2258
def step(self, message):
2259
"""Update the pb by a step."""
2260
self.count +=1
2261
self.pb.update(message, self.count, self.total)
2262
2263
2264
class CommitBuilder(object):
2265
"""Provides an interface to build up a commit.
2266
2267
This allows describing a tree to be committed without needing to
2268
know the internals of the format of the repository.
2269
"""
2270
2271
record_root_entry = False
2272
def __init__(self, repository, parents, config, timestamp=None,
2273
timezone=None, committer=None, revprops=None,
2274
revision_id=None):
2275
"""Initiate a CommitBuilder.
2276
2277
:param repository: Repository to commit to.
2278
:param parents: Revision ids of the parents of the new revision.
2279
:param config: Configuration to use.
2280
:param timestamp: Optional timestamp recorded for commit.
2281
:param timezone: Optional timezone for timestamp.
2282
:param committer: Optional committer to set for commit.
2283
:param revprops: Optional dictionary of revision properties.
2284
:param revision_id: Optional revision id.
2285
"""
2286
self._config = config
2287
2288
if committer is None:
2289
self._committer = self._config.username()
2290
else:
2291
assert isinstance(committer, basestring), type(committer)
2292
self._committer = committer
2293
2294
self.new_inventory = Inventory(None)
2295
self._new_revision_id = revision_id
2296
self.parents = parents
2297
self.repository = repository
2298
2299
self._revprops = {}
2300
if revprops is not None:
2301
self._revprops.update(revprops)
2302
2303
if timestamp is None:
2304
timestamp = time.time()
2305
# Restrict resolution to 1ms
2306
self._timestamp = round(timestamp, 3)
2307
2308
if timezone is None:
2309
self._timezone = local_time_offset()
2310
else:
2311
self._timezone = int(timezone)
2312
2313
self._generate_revision_if_needed()
2314
2315
def commit(self, message):
2316
"""Make the actual commit.
2317
2318
:return: The revision id of the recorded revision.
2319
"""
2320
rev = _mod_revision.Revision(
2321
timestamp=self._timestamp,
2322
timezone=self._timezone,
2323
committer=self._committer,
2324
message=message,
2325
inventory_sha1=self.inv_sha1,
2326
revision_id=self._new_revision_id,
2327
properties=self._revprops)
2328
rev.parent_ids = self.parents
2329
self.repository.add_revision(self._new_revision_id, rev,
2330
self.new_inventory, self._config)
2331
return self._new_revision_id
2332
2333
def revision_tree(self):
2334
"""Return the tree that was just committed.
2335
2336
After calling commit() this can be called to get a RevisionTree
2337
representing the newly committed tree. This is preferred to
2338
calling Repository.revision_tree() because that may require
2339
deserializing the inventory, while we already have a copy in
2340
memory.
2341
"""
2342
return RevisionTree(self.repository, self.new_inventory,
2343
self._new_revision_id)
2344
2345
def finish_inventory(self):
2346
"""Tell the builder that the inventory is finished."""
2347
if self.new_inventory.root is None:
2348
symbol_versioning.warn('Root entry should be supplied to'
2349
' record_entry_contents, as of bzr 0.10.',
2350
DeprecationWarning, stacklevel=2)
2351
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
2352
self.new_inventory.revision_id = self._new_revision_id
2353
self.inv_sha1 = self.repository.add_inventory(
2354
self._new_revision_id,
2355
self.new_inventory,
2356
self.parents
2357
)
2358
2359
@staticmethod
2360
def _sanitize_for_revision_id(text):
2361
"""This just removes whitespace, etc to make friendlier ids.
2362
"""
2363
2364
def _gen_revision_id(self):
2365
"""Return new revision-id."""
2366
return generate_ids.gen_revision_id(self._config.username(),
2367
self._timestamp)
2368
2369
def _generate_revision_if_needed(self):
2370
"""Create a revision id if None was supplied.
2371
2372
If the repository can not support user-specified revision ids
2373
they should override this function and raise UnsupportedOperation
2374
if _new_revision_id is not None.
2375
2376
:raises: UnsupportedOperation
2377
"""
2378
if self._new_revision_id is None:
2379
self._new_revision_id = self._gen_revision_id()
2380
2381
def record_entry_contents(self, ie, parent_invs, path, tree):
2382
"""Record the content of ie from tree into the commit if needed.
2383
2384
Side effect: sets ie.revision when unchanged
2385
2386
:param ie: An inventory entry present in the commit.
2387
:param parent_invs: The inventories of the parent revisions of the
2388
commit.
2389
:param path: The path the entry is at in the tree.
2390
:param tree: The tree which contains this entry and should be used to
2391
obtain content.
2392
"""
2393
if self.new_inventory.root is None and ie.parent_id is not None:
2394
symbol_versioning.warn('Root entry should be supplied to'
2395
' record_entry_contents, as of bzr 0.10.',
2396
DeprecationWarning, stacklevel=2)
2397
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
2398
'', tree)
2399
self.new_inventory.add(ie)
2400
2401
# ie.revision is always None if the InventoryEntry is considered
2402
# for committing. ie.snapshot will record the correct revision
2403
# which may be the sole parent if it is untouched.
2404
if ie.revision is not None:
2405
return
2406
2407
# In this revision format, root entries have no knit or weave
2408
if ie is self.new_inventory.root:
2409
# When serializing out to disk and back in
2410
# root.revision is always _new_revision_id
2411
ie.revision = self._new_revision_id
2412
return
2413
previous_entries = ie.find_previous_heads(
2414
parent_invs,
2415
self.repository.weave_store,
2416
self.repository.get_transaction())
2417
# we are creating a new revision for ie in the history store
2418
# and inventory.
2419
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2420
2421
def modified_directory(self, file_id, file_parents):
2422
"""Record the presence of a symbolic link.
2423
2424
:param file_id: The file_id of the link to record.
2425
:param file_parents: The per-file parent revision ids.
2426
"""
2427
self._add_text_to_weave(file_id, [], file_parents.keys())
2428
2429
def modified_file_text(self, file_id, file_parents,
2430
get_content_byte_lines, text_sha1=None,
2431
text_size=None):
2432
"""Record the text of file file_id
2433
2434
:param file_id: The file_id of the file to record the text of.
2435
:param file_parents: The per-file parent revision ids.
2436
:param get_content_byte_lines: A callable which will return the byte
2437
lines for the file.
2438
:param text_sha1: Optional SHA1 of the file contents.
2439
:param text_size: Optional size of the file contents.
2440
"""
2441
# mutter('storing text of file {%s} in revision {%s} into %r',
2442
# file_id, self._new_revision_id, self.repository.weave_store)
2443
# special case to avoid diffing on renames or
2444
# reparenting
2445
if (len(file_parents) == 1
2446
and text_sha1 == file_parents.values()[0].text_sha1
2447
and text_size == file_parents.values()[0].text_size):
2448
previous_ie = file_parents.values()[0]
2449
versionedfile = self.repository.weave_store.get_weave(file_id,
2450
self.repository.get_transaction())
2451
versionedfile.clone_text(self._new_revision_id,
2452
previous_ie.revision, file_parents.keys())
2453
return text_sha1, text_size
2454
else:
2455
new_lines = get_content_byte_lines()
2456
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2457
# should return the SHA1 and size
2458
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2459
return osutils.sha_strings(new_lines), \
2460
sum(map(len, new_lines))
2461
2462
def modified_link(self, file_id, file_parents, link_target):
2463
"""Record the presence of a symbolic link.
2464
2465
:param file_id: The file_id of the link to record.
2466
:param file_parents: The per-file parent revision ids.
2467
:param link_target: Target location of this link.
2468
"""
2469
self._add_text_to_weave(file_id, [], file_parents.keys())
2470
2471
def _add_text_to_weave(self, file_id, new_lines, parents):
2472
versionedfile = self.repository.weave_store.get_weave_or_empty(
2473
file_id, self.repository.get_transaction())
2474
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2475
versionedfile.clear_cache()
2476
2477
2478
class _CommitBuilder(CommitBuilder):
2479
"""Temporary class so old CommitBuilders are detected properly
2480
2481
Note: CommitBuilder works whether or not root entry is recorded.
2482
"""
2483
2484
record_root_entry = True
2485
2486
2487
class RootCommitBuilder(CommitBuilder):
2488
"""This commitbuilder actually records the root id"""
2489
2490
record_root_entry = True
2491
2492
def record_entry_contents(self, ie, parent_invs, path, tree):
2493
"""Record the content of ie from tree into the commit if needed.
2494
2495
Side effect: sets ie.revision when unchanged
2496
2497
:param ie: An inventory entry present in the commit.
2498
:param parent_invs: The inventories of the parent revisions of the
2499
commit.
2500
:param path: The path the entry is at in the tree.
2501
:param tree: The tree which contains this entry and should be used to
2502
obtain content.
2503
"""
2504
assert self.new_inventory.root is not None or ie.parent_id is None
2505
self.new_inventory.add(ie)
2506
2507
# ie.revision is always None if the InventoryEntry is considered
2508
# for committing. ie.snapshot will record the correct revision
2509
# which may be the sole parent if it is untouched.
2510
if ie.revision is not None:
2511
return
2512
2513
previous_entries = ie.find_previous_heads(
2514
parent_invs,
2515
self.repository.weave_store,
2516
self.repository.get_transaction())
2517
# we are creating a new revision for ie in the history store
2518
# and inventory.
2519
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2520
2521
2522
_unescape_map = {
2523
'apos':"'",
2524
'quot':'"',
2525
'amp':'&',
2526
'lt':'<',
2527
'gt':'>'
2528
}
2529
2530
2531
def _unescaper(match, _map=_unescape_map):
2532
return _map[match.group(1)]
2533
2534
2535
_unescape_re = None
2536
2537
2538
def _unescape_xml(data):
2539
"""Unescape predefined XML entities in a string of data."""
2540
global _unescape_re
2541
if _unescape_re is None:
2542
_unescape_re = re.compile('\&([^;]*);')
2543
return _unescape_re.sub(_unescaper, data)
Loggerhead is a web-based interface for Breezy
Version: 2.0.2.dev0