1
# Copyright (C) 2005, 2006 Canonical Ltd
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.
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.
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
17
from copy import deepcopy
18
from cStringIO import StringIO
19
from unittest import TestSuite
21
import bzrlib.bzrdir as bzrdir
22
from bzrlib.decorators import needs_read_lock, needs_write_lock
23
import bzrlib.errors as errors
24
from bzrlib.errors import InvalidRevisionId
25
import bzrlib.gpg as gpg
26
from bzrlib.graph import Graph
27
from bzrlib.inter import InterObject
28
from bzrlib.knit import KnitVersionedFile, KnitPlainFactory
29
from bzrlib.lockable_files import LockableFiles, TransportLock
30
from bzrlib.lockdir import LockDir
31
from bzrlib.osutils import safe_unicode
32
from bzrlib.revision import NULL_REVISION
33
from bzrlib.store.versioned import VersionedFileStore, WeaveStore
34
from bzrlib.store.text import TextStore
35
from bzrlib.symbol_versioning import *
36
from bzrlib.trace import mutter, note
37
from bzrlib.tree import RevisionTree
38
from bzrlib.tsort import topo_sort
39
from bzrlib.testament import Testament
40
from bzrlib.tree import EmptyTree
42
from bzrlib.weave import WeaveFile
46
class Repository(object):
47
"""Repository holding history for one or more branches.
49
The repository holds and retrieves historical information including
50
revisions and file history. It's normally accessed only by the Branch,
51
which views a particular line of development through that history.
53
The Repository builds on top of Stores and a Transport, which respectively
54
describe the disk data format and the way of accessing the (possibly
59
def add_inventory(self, revid, inv, parents):
60
"""Add the inventory inv to the repository as revid.
62
:param parents: The revision ids of the parents that revid
63
is known to have and are in the repository already.
65
returns the sha1 of the serialized inventory.
67
inv_text = bzrlib.xml5.serializer_v5.write_inventory_to_string(inv)
68
inv_sha1 = bzrlib.osutils.sha_string(inv_text)
69
inv_vf = self.control_weaves.get_weave('inventory',
70
self.get_transaction())
71
inv_vf.add_lines(revid, parents, bzrlib.osutils.split_lines(inv_text))
75
def add_revision(self, rev_id, rev, inv=None, config=None):
76
"""Add rev to the revision store as rev_id.
78
:param rev_id: the revision id to use.
79
:param rev: The revision object.
80
:param inv: The inventory for the revision. if None, it will be looked
81
up in the inventory storer
82
:param config: If None no digital signature will be created.
83
If supplied its signature_needed method will be used
84
to determine if a signature should be made.
86
if config is not None and config.signature_needed():
88
inv = self.get_inventory(rev_id)
89
plaintext = Testament(rev, inv).as_short_text()
90
self.store_revision_signature(
91
gpg.GPGStrategy(config), plaintext, rev_id)
92
if not rev_id in self.get_inventory_weave():
94
raise errors.WeaveRevisionNotPresent(rev_id,
95
self.get_inventory_weave())
97
# yes, this is not suitable for adding with ghosts.
98
self.add_inventory(rev_id, inv, rev.parent_ids)
99
self._revision_store.add_revision(rev, self.get_transaction())
102
def _all_possible_ids(self):
103
"""Return all the possible revisions that we could find."""
104
return self.get_inventory_weave().versions()
107
def all_revision_ids(self):
108
"""Returns a list of all the revision ids in the repository.
110
These are in as much topological order as the underlying store can
111
present: for weaves ghosts may lead to a lack of correctness until
112
the reweave updates the parents list.
114
if self._revision_store.text_store.listable():
115
return self._revision_store.all_revision_ids(self.get_transaction())
116
result = self._all_possible_ids()
117
return self._eliminate_revisions_not_present(result)
120
def _eliminate_revisions_not_present(self, revision_ids):
121
"""Check every revision id in revision_ids to see if we have it.
123
Returns a set of the present revisions.
126
for id in revision_ids:
127
if self.has_revision(id):
132
def create(a_bzrdir):
133
"""Construct the current default format repository in a_bzrdir."""
134
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
136
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
137
"""instantiate a Repository.
139
:param _format: The format of the repository on disk.
140
:param a_bzrdir: The BzrDir of the repository.
142
In the future we will have a single api for all stores for
143
getting file texts, inventories and revisions, then
144
this construct will accept instances of those things.
146
super(Repository, self).__init__()
147
self._format = _format
148
# the following are part of the public API for Repository:
149
self.bzrdir = a_bzrdir
150
self.control_files = control_files
151
self._revision_store = _revision_store
152
self.text_store = text_store
153
# backwards compatability
154
self.weave_store = text_store
155
# not right yet - should be more semantically clear ?
157
self.control_store = control_store
158
self.control_weaves = control_store
159
# TODO: make sure to construct the right store classes, etc, depending
160
# on whether escaping is required.
163
return self.control_files.is_locked()
165
def lock_write(self):
166
self.control_files.lock_write()
169
self.control_files.lock_read()
171
def get_physical_lock_status(self):
172
return self.control_files.get_physical_lock_status()
175
def missing_revision_ids(self, other, revision_id=None):
176
"""Return the revision ids that other has that this does not.
178
These are returned in topological order.
180
revision_id: only return revision ids included by revision_id.
182
return InterRepository.get(other, self).missing_revision_ids(revision_id)
186
"""Open the repository rooted at base.
188
For instance, if the repository is at URL/.bzr/repository,
189
Repository.open(URL) -> a Repository instance.
191
control = bzrlib.bzrdir.BzrDir.open(base)
192
return control.open_repository()
194
def copy_content_into(self, destination, revision_id=None, basis=None):
195
"""Make a complete copy of the content in self into destination.
197
This is a destructive operation! Do not use it on existing
200
return InterRepository.get(self, destination).copy_content(revision_id, basis)
202
def fetch(self, source, revision_id=None, pb=None):
203
"""Fetch the content required to construct revision_id from source.
205
If revision_id is None all content is copied.
207
return InterRepository.get(source, self).fetch(revision_id=revision_id,
211
self.control_files.unlock()
214
def clone(self, a_bzrdir, revision_id=None, basis=None):
215
"""Clone this repository into a_bzrdir using the current format.
217
Currently no check is made that the format of this repository and
218
the bzrdir format are compatible. FIXME RBC 20060201.
220
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
221
# use target default format.
222
result = a_bzrdir.create_repository()
223
# FIXME RBC 20060209 split out the repository type to avoid this check ?
224
elif isinstance(a_bzrdir._format,
225
(bzrlib.bzrdir.BzrDirFormat4,
226
bzrlib.bzrdir.BzrDirFormat5,
227
bzrlib.bzrdir.BzrDirFormat6)):
228
result = a_bzrdir.open_repository()
230
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
231
self.copy_content_into(result, revision_id, basis)
235
def has_revision(self, revision_id):
236
"""True if this repository has a copy of the revision."""
237
return self._revision_store.has_revision_id(revision_id,
238
self.get_transaction())
241
def get_revision_reconcile(self, revision_id):
242
"""'reconcile' helper routine that allows access to a revision always.
244
This variant of get_revision does not cross check the weave graph
245
against the revision one as get_revision does: but it should only
246
be used by reconcile, or reconcile-alike commands that are correcting
247
or testing the revision graph.
249
if not revision_id or not isinstance(revision_id, basestring):
250
raise InvalidRevisionId(revision_id=revision_id, branch=self)
251
return self._revision_store.get_revision(revision_id,
252
self.get_transaction())
255
def get_revision_xml(self, revision_id):
256
rev = self.get_revision(revision_id)
258
# the current serializer..
259
self._revision_store._serializer.write_revision(rev, rev_tmp)
261
return rev_tmp.getvalue()
264
def get_revision(self, revision_id):
265
"""Return the Revision object for a named revision"""
266
r = self.get_revision_reconcile(revision_id)
267
# weave corruption can lead to absent revision markers that should be
269
# the following test is reasonably cheap (it needs a single weave read)
270
# and the weave is cached in read transactions. In write transactions
271
# it is not cached but typically we only read a small number of
272
# revisions. For knits when they are introduced we will probably want
273
# to ensure that caching write transactions are in use.
274
inv = self.get_inventory_weave()
275
self._check_revision_parents(r, inv)
278
def _check_revision_parents(self, revision, inventory):
279
"""Private to Repository and Fetch.
281
This checks the parentage of revision in an inventory weave for
282
consistency and is only applicable to inventory-weave-for-ancestry
283
using repository formats & fetchers.
285
weave_parents = inventory.get_parents(revision.revision_id)
286
weave_names = inventory.versions()
287
for parent_id in revision.parent_ids:
288
if parent_id in weave_names:
289
# this parent must not be a ghost.
290
if not parent_id in weave_parents:
292
raise errors.CorruptRepository(self)
295
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
296
signature = gpg_strategy.sign(plaintext)
297
self._revision_store.add_revision_signature_text(revision_id,
299
self.get_transaction())
301
def fileids_altered_by_revision_ids(self, revision_ids):
302
"""Find the file ids and versions affected by revisions.
304
:param revisions: an iterable containing revision ids.
305
:return: a dictionary mapping altered file-ids to an iterable of
306
revision_ids. Each altered file-ids has the exact revision_ids that
307
altered it listed explicitly.
309
assert isinstance(self._format, (RepositoryFormat5,
312
RepositoryFormatKnit1)), \
313
"fileid_involved only supported for branches which store inventory as unnested xml"
314
selected_revision_ids = set(revision_ids)
315
w = self.get_inventory_weave()
318
# this code needs to read every new line in every inventory for the
319
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
320
# not pesent in one of those inventories is unnecessary but not
321
# harmful because we are filtering by the revision id marker in the
322
# inventory lines : we only select file ids altered in one of those
323
# revisions. We dont need to see all lines in the inventory because
324
# only those added in an inventory in rev X can contain a revision=X
326
for line in w.iter_lines_added_or_present_in_versions(selected_revision_ids):
327
start = line.find('file_id="')+9
328
if start < 9: continue
329
end = line.find('"', start)
331
file_id = _unescape_xml(line[start:end])
333
start = line.find('revision="')+10
334
if start < 10: continue
335
end = line.find('"', start)
337
revision_id = _unescape_xml(line[start:end])
338
if revision_id in selected_revision_ids:
339
result.setdefault(file_id, set()).add(revision_id)
343
def get_inventory_weave(self):
344
return self.control_weaves.get_weave('inventory',
345
self.get_transaction())
348
def get_inventory(self, revision_id):
349
"""Get Inventory object by hash."""
350
xml = self.get_inventory_xml(revision_id)
351
return bzrlib.xml5.serializer_v5.read_inventory_from_string(xml)
354
def get_inventory_xml(self, revision_id):
355
"""Get inventory XML as a file object."""
357
assert isinstance(revision_id, basestring), type(revision_id)
358
iw = self.get_inventory_weave()
359
return iw.get_text(revision_id)
361
raise bzrlib.errors.HistoryMissing(self, 'inventory', revision_id)
364
def get_inventory_sha1(self, revision_id):
365
"""Return the sha1 hash of the inventory entry
367
return self.get_revision(revision_id).inventory_sha1
370
def get_revision_graph(self, revision_id=None):
371
"""Return a dictionary containing the revision graph.
373
:return: a dictionary of revision_id->revision_parents_list.
375
weave = self.get_inventory_weave()
376
all_revisions = self._eliminate_revisions_not_present(weave.versions())
377
entire_graph = dict([(node, weave.get_parents(node)) for
378
node in all_revisions])
379
if revision_id is None:
381
elif revision_id not in entire_graph:
382
raise errors.NoSuchRevision(self, revision_id)
384
# add what can be reached from revision_id
386
pending = set([revision_id])
387
while len(pending) > 0:
389
result[node] = entire_graph[node]
390
for revision_id in result[node]:
391
if revision_id not in result:
392
pending.add(revision_id)
396
def get_revision_graph_with_ghosts(self, revision_ids=None):
397
"""Return a graph of the revisions with ghosts marked as applicable.
399
:param revision_ids: an iterable of revisions to graph or None for all.
400
:return: a Graph object with the graph reachable from revision_ids.
404
pending = set(self.all_revision_ids())
407
pending = set(revision_ids)
408
required = set(revision_ids)
411
revision_id = pending.pop()
413
rev = self.get_revision(revision_id)
414
except errors.NoSuchRevision:
415
if revision_id in required:
418
result.add_ghost(revision_id)
420
for parent_id in rev.parent_ids:
421
# is this queued or done ?
422
if (parent_id not in pending and
423
parent_id not in done):
425
pending.add(parent_id)
426
result.add_node(revision_id, rev.parent_ids)
427
done.add(revision_id)
431
def get_revision_inventory(self, revision_id):
432
"""Return inventory of a past revision."""
433
# TODO: Unify this with get_inventory()
434
# bzr 0.0.6 and later imposes the constraint that the inventory_id
435
# must be the same as its revision, so this is trivial.
436
if revision_id is None:
437
# This does not make sense: if there is no revision,
438
# then it is the current tree inventory surely ?!
439
# and thus get_root_id() is something that looks at the last
440
# commit on the branch, and the get_root_id is an inventory check.
441
raise NotImplementedError
442
# return Inventory(self.get_root_id())
444
return self.get_inventory(revision_id)
448
"""Return True if this repository is flagged as a shared repository."""
449
raise NotImplementedError(self.is_shared)
452
def reconcile(self, other=None, thorough=False):
453
"""Reconcile this repository."""
454
from bzrlib.reconcile import RepoReconciler
455
reconciler = RepoReconciler(self, thorough=thorough)
456
reconciler.reconcile()
460
def revision_tree(self, revision_id):
461
"""Return Tree for a revision on this branch.
463
`revision_id` may be None for the null revision, in which case
464
an `EmptyTree` is returned."""
465
# TODO: refactor this to use an existing revision object
466
# so we don't need to read it in twice.
467
if revision_id is None or revision_id == NULL_REVISION:
470
inv = self.get_revision_inventory(revision_id)
471
return RevisionTree(self, inv, revision_id)
474
def get_ancestry(self, revision_id):
475
"""Return a list of revision-ids integrated by a revision.
477
This is topologically sorted.
479
if revision_id is None:
481
if not self.has_revision(revision_id):
482
raise errors.NoSuchRevision(self, revision_id)
483
w = self.get_inventory_weave()
484
candidates = w.get_ancestry(revision_id)
485
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
488
def print_file(self, file, revision_id):
489
"""Print `file` to stdout.
491
FIXME RBC 20060125 as John Meinel points out this is a bad api
492
- it writes to stdout, it assumes that that is valid etc. Fix
493
by creating a new more flexible convenience function.
495
tree = self.revision_tree(revision_id)
496
# use inventory as it was in that revision
497
file_id = tree.inventory.path2id(file)
499
raise BzrError("%r is not present in revision %s" % (file, revno))
501
revno = self.revision_id_to_revno(revision_id)
502
except errors.NoSuchRevision:
503
# TODO: This should not be BzrError,
504
# but NoSuchFile doesn't fit either
505
raise BzrError('%r is not present in revision %s'
506
% (file, revision_id))
508
raise BzrError('%r is not present in revision %s'
510
tree.print_file(file_id)
512
def get_transaction(self):
513
return self.control_files.get_transaction()
515
def revision_parents(self, revid):
516
return self.get_inventory_weave().parent_names(revid)
519
def set_make_working_trees(self, new_value):
520
"""Set the policy flag for making working trees when creating branches.
522
This only applies to branches that use this repository.
524
The default is 'True'.
525
:param new_value: True to restore the default, False to disable making
528
raise NotImplementedError(self.set_make_working_trees)
530
def make_working_trees(self):
531
"""Returns the policy for making working trees on new branches."""
532
raise NotImplementedError(self.make_working_trees)
535
def sign_revision(self, revision_id, gpg_strategy):
536
plaintext = Testament.from_revision(self, revision_id).as_short_text()
537
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
540
def has_signature_for_revision_id(self, revision_id):
541
"""Query for a revision signature for revision_id in the repository."""
542
return self._revision_store.has_signature(revision_id,
543
self.get_transaction())
546
def get_signature_text(self, revision_id):
547
"""Return the text for a signature."""
548
return self._revision_store.get_signature_text(revision_id,
549
self.get_transaction())
552
class AllInOneRepository(Repository):
553
"""Legacy support - the repository behaviour for all-in-one branches."""
555
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
556
# we reuse one control files instance.
557
dir_mode = a_bzrdir._control_files._dir_mode
558
file_mode = a_bzrdir._control_files._file_mode
560
def get_store(name, compressed=True, prefixed=False):
561
# FIXME: This approach of assuming stores are all entirely compressed
562
# or entirely uncompressed is tidy, but breaks upgrade from
563
# some existing branches where there's a mixture; we probably
564
# still want the option to look for both.
565
relpath = a_bzrdir._control_files._escape(name)
566
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
567
prefixed=prefixed, compressed=compressed,
570
#if self._transport.should_cache():
571
# cache_path = os.path.join(self.cache_root, name)
572
# os.mkdir(cache_path)
573
# store = bzrlib.store.CachedStore(store, cache_path)
576
# not broken out yet because the controlweaves|inventory_store
577
# and text_store | weave_store bits are still different.
578
if isinstance(_format, RepositoryFormat4):
579
# cannot remove these - there is still no consistent api
580
# which allows access to this old info.
581
self.inventory_store = get_store('inventory-store')
582
text_store = get_store('text-store')
583
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
587
"""AllInOne repositories cannot be shared."""
591
def set_make_working_trees(self, new_value):
592
"""Set the policy flag for making working trees when creating branches.
594
This only applies to branches that use this repository.
596
The default is 'True'.
597
:param new_value: True to restore the default, False to disable making
600
raise NotImplementedError(self.set_make_working_trees)
602
def make_working_trees(self):
603
"""Returns the policy for making working trees on new branches."""
607
class MetaDirRepository(Repository):
608
"""Repositories in the new meta-dir layout."""
610
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
611
super(MetaDirRepository, self).__init__(_format,
618
dir_mode = self.control_files._dir_mode
619
file_mode = self.control_files._file_mode
623
"""Return True if this repository is flagged as a shared repository."""
624
return self.control_files._transport.has('shared-storage')
627
def set_make_working_trees(self, new_value):
628
"""Set the policy flag for making working trees when creating branches.
630
This only applies to branches that use this repository.
632
The default is 'True'.
633
:param new_value: True to restore the default, False to disable making
638
self.control_files._transport.delete('no-working-trees')
639
except errors.NoSuchFile:
642
self.control_files.put_utf8('no-working-trees', '')
644
def make_working_trees(self):
645
"""Returns the policy for making working trees on new branches."""
646
return not self.control_files._transport.has('no-working-trees')
649
class KnitRepository(MetaDirRepository):
650
"""Knit format repository."""
653
def all_revision_ids(self):
654
"""See Repository.all_revision_ids()."""
655
return self._revision_store.all_revision_ids(self.get_transaction())
657
def fileid_involved_between_revs(self, from_revid, to_revid):
658
"""Find file_id(s) which are involved in the changes between revisions.
660
This determines the set of revisions which are involved, and then
661
finds all file ids affected by those revisions.
663
vf = self._get_revision_vf()
664
from_set = set(vf.get_ancestry(from_revid))
665
to_set = set(vf.get_ancestry(to_revid))
666
changed = to_set.difference(from_set)
667
return self._fileid_involved_by_set(changed)
669
def fileid_involved(self, last_revid=None):
670
"""Find all file_ids modified in the ancestry of last_revid.
672
:param last_revid: If None, last_revision() will be used.
675
changed = set(self.all_revision_ids())
677
changed = set(self.get_ancestry(last_revid))
680
return self._fileid_involved_by_set(changed)
683
def get_ancestry(self, revision_id):
684
"""Return a list of revision-ids integrated by a revision.
686
This is topologically sorted.
688
if revision_id is None:
690
vf = self._get_revision_vf()
692
return [None] + vf.get_ancestry(revision_id)
693
except errors.RevisionNotPresent:
694
raise errors.NoSuchRevision(self, revision_id)
697
def get_revision(self, revision_id):
698
"""Return the Revision object for a named revision"""
699
return self.get_revision_reconcile(revision_id)
702
def get_revision_graph(self, revision_id=None):
703
"""Return a dictionary containing the revision graph.
705
:return: a dictionary of revision_id->revision_parents_list.
707
weave = self._get_revision_vf()
708
entire_graph = weave.get_graph()
709
if revision_id is None:
710
return weave.get_graph()
711
elif revision_id not in weave:
712
raise errors.NoSuchRevision(self, revision_id)
714
# add what can be reached from revision_id
716
pending = set([revision_id])
717
while len(pending) > 0:
719
result[node] = weave.get_parents(node)
720
for revision_id in result[node]:
721
if revision_id not in result:
722
pending.add(revision_id)
726
def get_revision_graph_with_ghosts(self, revision_ids=None):
727
"""Return a graph of the revisions with ghosts marked as applicable.
729
:param revision_ids: an iterable of revisions to graph or None for all.
730
:return: a Graph object with the graph reachable from revision_ids.
733
vf = self._get_revision_vf()
734
versions = set(vf.versions())
736
pending = set(self.all_revision_ids())
739
pending = set(revision_ids)
740
required = set(revision_ids)
743
revision_id = pending.pop()
744
if not revision_id in versions:
745
if revision_id in required:
746
raise errors.NoSuchRevision(self, revision_id)
748
result.add_ghost(revision_id)
749
# mark it as done so we dont try for it again.
750
done.add(revision_id)
752
parent_ids = vf.get_parents_with_ghosts(revision_id)
753
for parent_id in parent_ids:
754
# is this queued or done ?
755
if (parent_id not in pending and
756
parent_id not in done):
758
pending.add(parent_id)
759
result.add_node(revision_id, parent_ids)
760
done.add(revision_id)
763
def _get_revision_vf(self):
764
""":return: a versioned file containing the revisions."""
765
vf = self._revision_store.get_revision_file(self.get_transaction())
769
def reconcile(self, other=None, thorough=False):
770
"""Reconcile this repository."""
771
from bzrlib.reconcile import KnitReconciler
772
reconciler = KnitReconciler(self, thorough=thorough)
773
reconciler.reconcile()
776
def revision_parents(self, revid):
777
return self._get_revision_vf().get_parents(rev_id)
779
class RepositoryFormat(object):
780
"""A repository format.
782
Formats provide three things:
783
* An initialization routine to construct repository data on disk.
784
* a format string which is used when the BzrDir supports versioned
786
* an open routine which returns a Repository instance.
788
Formats are placed in an dict by their format string for reference
789
during opening. These should be subclasses of RepositoryFormat
792
Once a format is deprecated, just deprecate the initialize and open
793
methods on the format class. Do not deprecate the object, as the
794
object will be created every system load.
796
Common instance attributes:
797
_matchingbzrdir - the bzrdir format that the repository format was
798
originally written to work with. This can be used if manually
799
constructing a bzrdir and repository, or more commonly for test suite
803
_default_format = None
804
"""The default format used for new repositories."""
807
"""The known formats."""
810
def find_format(klass, a_bzrdir):
811
"""Return the format for the repository object in a_bzrdir."""
813
transport = a_bzrdir.get_repository_transport(None)
814
format_string = transport.get("format").read()
815
return klass._formats[format_string]
816
except errors.NoSuchFile:
817
raise errors.NoRepositoryPresent(a_bzrdir)
819
raise errors.UnknownFormatError(format_string)
821
def _get_control_store(self, repo_transport, control_files):
822
"""Return the control store for this repository."""
823
raise NotImplementedError(self._get_control_store)
826
def get_default_format(klass):
827
"""Return the current default format."""
828
return klass._default_format
830
def get_format_string(self):
831
"""Return the ASCII format string that identifies this format.
833
Note that in pre format ?? repositories the format string is
834
not permitted nor written to disk.
836
raise NotImplementedError(self.get_format_string)
838
def get_format_description(self):
839
"""Return the short desciption for this format."""
840
raise NotImplementedError(self.get_format_description)
842
def _get_revision_store(self, repo_transport, control_files):
843
"""Return the revision store object for this a_bzrdir."""
844
raise NotImplementedError(self._get_revision_store)
846
def _get_text_rev_store(self,
853
"""Common logic for getting a revision store for a repository.
855
see self._get_revision_store for the subclass-overridable method to
856
get the store for a repository.
858
from bzrlib.store.revision.text import TextRevisionStore
859
dir_mode = control_files._dir_mode
860
file_mode = control_files._file_mode
861
text_store =TextStore(transport.clone(name),
863
compressed=compressed,
866
_revision_store = TextRevisionStore(text_store, serializer)
867
return _revision_store
869
def _get_versioned_file_store(self,
874
versionedfile_class=WeaveFile,
876
weave_transport = control_files._transport.clone(name)
877
dir_mode = control_files._dir_mode
878
file_mode = control_files._file_mode
879
return VersionedFileStore(weave_transport, prefixed=prefixed,
882
versionedfile_class=versionedfile_class,
885
def initialize(self, a_bzrdir, shared=False):
886
"""Initialize a repository of this format in a_bzrdir.
888
:param a_bzrdir: The bzrdir to put the new repository in it.
889
:param shared: The repository should be initialized as a sharable one.
891
This may raise UninitializableFormat if shared repository are not
892
compatible the a_bzrdir.
895
def is_supported(self):
896
"""Is this format supported?
898
Supported formats must be initializable and openable.
899
Unsupported formats may not support initialization or committing or
900
some other features depending on the reason for not being supported.
904
def open(self, a_bzrdir, _found=False):
905
"""Return an instance of this format for the bzrdir a_bzrdir.
907
_found is a private parameter, do not use it.
909
raise NotImplementedError(self.open)
912
def register_format(klass, format):
913
klass._formats[format.get_format_string()] = format
916
def set_default_format(klass, format):
917
klass._default_format = format
920
def unregister_format(klass, format):
921
assert klass._formats[format.get_format_string()] is format
922
del klass._formats[format.get_format_string()]
925
class PreSplitOutRepositoryFormat(RepositoryFormat):
926
"""Base class for the pre split out repository formats."""
928
def initialize(self, a_bzrdir, shared=False, _internal=False):
929
"""Create a weave repository.
931
TODO: when creating split out bzr branch formats, move this to a common
932
base for Format5, Format6. or something like that.
934
from bzrlib.weavefile import write_weave_v5
935
from bzrlib.weave import Weave
938
raise errors.IncompatibleFormat(self, a_bzrdir._format)
941
# always initialized when the bzrdir is.
942
return self.open(a_bzrdir, _found=True)
944
# Create an empty weave
946
bzrlib.weavefile.write_weave_v5(Weave(), sio)
947
empty_weave = sio.getvalue()
949
mutter('creating repository in %s.', a_bzrdir.transport.base)
950
dirs = ['revision-store', 'weaves']
951
files = [('inventory.weave', StringIO(empty_weave)),
954
# FIXME: RBC 20060125 dont peek under the covers
955
# NB: no need to escape relative paths that are url safe.
956
control_files = LockableFiles(a_bzrdir.transport, 'branch-lock',
958
control_files.create_lock()
959
control_files.lock_write()
960
control_files._transport.mkdir_multi(dirs,
961
mode=control_files._dir_mode)
963
for file, content in files:
964
control_files.put(file, content)
966
control_files.unlock()
967
return self.open(a_bzrdir, _found=True)
969
def _get_control_store(self, repo_transport, control_files):
970
"""Return the control store for this repository."""
971
return self._get_versioned_file_store('',
976
def _get_text_store(self, transport, control_files):
977
"""Get a store for file texts for this format."""
978
raise NotImplementedError(self._get_text_store)
980
def open(self, a_bzrdir, _found=False):
981
"""See RepositoryFormat.open()."""
983
# we are being called directly and must probe.
984
raise NotImplementedError
986
repo_transport = a_bzrdir.get_repository_transport(None)
987
control_files = a_bzrdir._control_files
988
text_store = self._get_text_store(repo_transport, control_files)
989
control_store = self._get_control_store(repo_transport, control_files)
990
_revision_store = self._get_revision_store(repo_transport, control_files)
991
return AllInOneRepository(_format=self,
993
_revision_store=_revision_store,
994
control_store=control_store,
995
text_store=text_store)
998
class RepositoryFormat4(PreSplitOutRepositoryFormat):
999
"""Bzr repository format 4.
1001
This repository format has:
1003
- TextStores for texts, inventories,revisions.
1005
This format is deprecated: it indexes texts using a text id which is
1006
removed in format 5; initializationa and write support for this format
1011
super(RepositoryFormat4, self).__init__()
1012
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat4()
1014
def get_format_description(self):
1015
"""See RepositoryFormat.get_format_description()."""
1016
return "Repository format 4"
1018
def initialize(self, url, shared=False, _internal=False):
1019
"""Format 4 branches cannot be created."""
1020
raise errors.UninitializableFormat(self)
1022
def is_supported(self):
1023
"""Format 4 is not supported.
1025
It is not supported because the model changed from 4 to 5 and the
1026
conversion logic is expensive - so doing it on the fly was not
1031
def _get_control_store(self, repo_transport, control_files):
1032
"""Format 4 repositories have no formal control store at this point.
1034
This will cause any control-file-needing apis to fail - this is desired.
1038
def _get_revision_store(self, repo_transport, control_files):
1039
"""See RepositoryFormat._get_revision_store()."""
1040
from bzrlib.xml4 import serializer_v4
1041
return self._get_text_rev_store(repo_transport,
1044
serializer=serializer_v4)
1046
def _get_text_store(self, transport, control_files):
1047
"""See RepositoryFormat._get_text_store()."""
1050
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1051
"""Bzr control format 5.
1053
This repository format has:
1054
- weaves for file texts and inventory
1056
- TextStores for revisions and signatures.
1060
super(RepositoryFormat5, self).__init__()
1061
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat5()
1063
def get_format_description(self):
1064
"""See RepositoryFormat.get_format_description()."""
1065
return "Weave repository format 5"
1067
def _get_revision_store(self, repo_transport, control_files):
1068
"""See RepositoryFormat._get_revision_store()."""
1069
"""Return the revision store object for this a_bzrdir."""
1070
return self._get_text_rev_store(repo_transport,
1075
def _get_text_store(self, transport, control_files):
1076
"""See RepositoryFormat._get_text_store()."""
1077
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1080
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1081
"""Bzr control format 6.
1083
This repository format has:
1084
- weaves for file texts and inventory
1085
- hash subdirectory based stores.
1086
- TextStores for revisions and signatures.
1090
super(RepositoryFormat6, self).__init__()
1091
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat6()
1093
def get_format_description(self):
1094
"""See RepositoryFormat.get_format_description()."""
1095
return "Weave repository format 6"
1097
def _get_revision_store(self, repo_transport, control_files):
1098
"""See RepositoryFormat._get_revision_store()."""
1099
return self._get_text_rev_store(repo_transport,
1105
def _get_text_store(self, transport, control_files):
1106
"""See RepositoryFormat._get_text_store()."""
1107
return self._get_versioned_file_store('weaves', transport, control_files)
1110
class MetaDirRepositoryFormat(RepositoryFormat):
1111
"""Common base class for the new repositories using the metadir layour."""
1114
super(MetaDirRepositoryFormat, self).__init__()
1115
self._matchingbzrdir = bzrlib.bzrdir.BzrDirMetaFormat1()
1117
def _create_control_files(self, a_bzrdir):
1118
"""Create the required files and the initial control_files object."""
1119
# FIXME: RBC 20060125 dont peek under the covers
1120
# NB: no need to escape relative paths that are url safe.
1121
repository_transport = a_bzrdir.get_repository_transport(self)
1122
control_files = LockableFiles(repository_transport, 'lock', LockDir)
1123
control_files.create_lock()
1124
return control_files
1126
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1127
"""Upload the initial blank content."""
1128
control_files = self._create_control_files(a_bzrdir)
1129
control_files.lock_write()
1131
control_files._transport.mkdir_multi(dirs,
1132
mode=control_files._dir_mode)
1133
for file, content in files:
1134
control_files.put(file, content)
1135
for file, content in utf8_files:
1136
control_files.put_utf8(file, content)
1138
control_files.put_utf8('shared-storage', '')
1140
control_files.unlock()
1143
class RepositoryFormat7(MetaDirRepositoryFormat):
1144
"""Bzr repository 7.
1146
This repository format has:
1147
- weaves for file texts and inventory
1148
- hash subdirectory based stores.
1149
- TextStores for revisions and signatures.
1150
- a format marker of its own
1151
- an optional 'shared-storage' flag
1152
- an optional 'no-working-trees' flag
1155
def _get_control_store(self, repo_transport, control_files):
1156
"""Return the control store for this repository."""
1157
return self._get_versioned_file_store('',
1162
def get_format_string(self):
1163
"""See RepositoryFormat.get_format_string()."""
1164
return "Bazaar-NG Repository format 7"
1166
def get_format_description(self):
1167
"""See RepositoryFormat.get_format_description()."""
1168
return "Weave repository format 7"
1170
def _get_revision_store(self, repo_transport, control_files):
1171
"""See RepositoryFormat._get_revision_store()."""
1172
return self._get_text_rev_store(repo_transport,
1179
def _get_text_store(self, transport, control_files):
1180
"""See RepositoryFormat._get_text_store()."""
1181
return self._get_versioned_file_store('weaves',
1185
def initialize(self, a_bzrdir, shared=False):
1186
"""Create a weave repository.
1188
:param shared: If true the repository will be initialized as a shared
1191
from bzrlib.weavefile import write_weave_v5
1192
from bzrlib.weave import Weave
1194
# Create an empty weave
1196
bzrlib.weavefile.write_weave_v5(Weave(), sio)
1197
empty_weave = sio.getvalue()
1199
mutter('creating repository in %s.', a_bzrdir.transport.base)
1200
dirs = ['revision-store', 'weaves']
1201
files = [('inventory.weave', StringIO(empty_weave)),
1203
utf8_files = [('format', self.get_format_string())]
1205
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1206
return self.open(a_bzrdir=a_bzrdir, _found=True)
1208
def open(self, a_bzrdir, _found=False, _override_transport=None):
1209
"""See RepositoryFormat.open().
1211
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1212
repository at a slightly different url
1213
than normal. I.e. during 'upgrade'.
1216
format = RepositoryFormat.find_format(a_bzrdir)
1217
assert format.__class__ == self.__class__
1218
if _override_transport is not None:
1219
repo_transport = _override_transport
1221
repo_transport = a_bzrdir.get_repository_transport(None)
1222
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1223
text_store = self._get_text_store(repo_transport, control_files)
1224
control_store = self._get_control_store(repo_transport, control_files)
1225
_revision_store = self._get_revision_store(repo_transport, control_files)
1226
return MetaDirRepository(_format=self,
1228
control_files=control_files,
1229
_revision_store=_revision_store,
1230
control_store=control_store,
1231
text_store=text_store)
1234
class RepositoryFormatKnit1(MetaDirRepositoryFormat):
1235
"""Bzr repository knit format 1.
1237
This repository format has:
1238
- knits for file texts and inventory
1239
- hash subdirectory based stores.
1240
- knits for revisions and signatures
1241
- TextStores for revisions and signatures.
1242
- a format marker of its own
1243
- an optional 'shared-storage' flag
1244
- an optional 'no-working-trees' flag
1247
This format was introduced in bzr 0.8.
1250
def _get_control_store(self, repo_transport, control_files):
1251
"""Return the control store for this repository."""
1252
return VersionedFileStore(
1255
file_mode=control_files._file_mode,
1256
versionedfile_class=KnitVersionedFile,
1257
versionedfile_kwargs={'factory':KnitPlainFactory()},
1260
def get_format_string(self):
1261
"""See RepositoryFormat.get_format_string()."""
1262
return "Bazaar-NG Knit Repository Format 1"
1264
def get_format_description(self):
1265
"""See RepositoryFormat.get_format_description()."""
1266
return "Knit repository format 1"
1268
def _get_revision_store(self, repo_transport, control_files):
1269
"""See RepositoryFormat._get_revision_store()."""
1270
from bzrlib.store.revision.knit import KnitRevisionStore
1271
versioned_file_store = VersionedFileStore(
1273
file_mode=control_files._file_mode,
1276
versionedfile_class=KnitVersionedFile,
1277
versionedfile_kwargs={'delta':False, 'factory':KnitPlainFactory()},
1280
return KnitRevisionStore(versioned_file_store)
1282
def _get_text_store(self, transport, control_files):
1283
"""See RepositoryFormat._get_text_store()."""
1284
return self._get_versioned_file_store('knits',
1287
versionedfile_class=KnitVersionedFile,
1290
def initialize(self, a_bzrdir, shared=False):
1291
"""Create a knit format 1 repository.
1293
:param a_bzrdir: bzrdir to contain the new repository; must already
1295
:param shared: If true the repository will be initialized as a shared
1298
mutter('creating repository in %s.', a_bzrdir.transport.base)
1299
dirs = ['revision-store', 'knits']
1301
utf8_files = [('format', self.get_format_string())]
1303
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1304
repo_transport = a_bzrdir.get_repository_transport(None)
1305
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1306
control_store = self._get_control_store(repo_transport, control_files)
1307
transaction = bzrlib.transactions.WriteTransaction()
1308
# trigger a write of the inventory store.
1309
control_store.get_weave_or_empty('inventory', transaction)
1310
_revision_store = self._get_revision_store(repo_transport, control_files)
1311
_revision_store.has_revision_id('A', transaction)
1312
_revision_store.get_signature_file(transaction)
1313
return self.open(a_bzrdir=a_bzrdir, _found=True)
1315
def open(self, a_bzrdir, _found=False, _override_transport=None):
1316
"""See RepositoryFormat.open().
1318
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1319
repository at a slightly different url
1320
than normal. I.e. during 'upgrade'.
1323
format = RepositoryFormat.find_format(a_bzrdir)
1324
assert format.__class__ == self.__class__
1325
if _override_transport is not None:
1326
repo_transport = _override_transport
1328
repo_transport = a_bzrdir.get_repository_transport(None)
1329
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1330
text_store = self._get_text_store(repo_transport, control_files)
1331
control_store = self._get_control_store(repo_transport, control_files)
1332
_revision_store = self._get_revision_store(repo_transport, control_files)
1333
return KnitRepository(_format=self,
1335
control_files=control_files,
1336
_revision_store=_revision_store,
1337
control_store=control_store,
1338
text_store=text_store)
1341
# formats which have no format string are not discoverable
1342
# and not independently creatable, so are not registered.
1343
RepositoryFormat.register_format(RepositoryFormat7())
1344
_default_format = RepositoryFormatKnit1()
1345
RepositoryFormat.register_format(_default_format)
1346
RepositoryFormat.set_default_format(_default_format)
1347
_legacy_formats = [RepositoryFormat4(),
1348
RepositoryFormat5(),
1349
RepositoryFormat6()]
1352
class InterRepository(InterObject):
1353
"""This class represents operations taking place between two repositories.
1355
Its instances have methods like copy_content and fetch, and contain
1356
references to the source and target repositories these operations can be
1359
Often we will provide convenience methods on 'repository' which carry out
1360
operations with another repository - they will always forward to
1361
InterRepository.get(other).method_name(parameters).
1365
"""The available optimised InterRepository types."""
1368
def copy_content(self, revision_id=None, basis=None):
1369
"""Make a complete copy of the content in self into destination.
1371
This is a destructive operation! Do not use it on existing
1374
:param revision_id: Only copy the content needed to construct
1375
revision_id and its parents.
1376
:param basis: Copy the needed data preferentially from basis.
1379
self.target.set_make_working_trees(self.source.make_working_trees())
1380
except NotImplementedError:
1382
# grab the basis available data
1383
if basis is not None:
1384
self.target.fetch(basis, revision_id=revision_id)
1385
# but dont bother fetching if we have the needed data now.
1386
if (revision_id not in (None, NULL_REVISION) and
1387
self.target.has_revision(revision_id)):
1389
self.target.fetch(self.source, revision_id=revision_id)
1391
def _double_lock(self, lock_source, lock_target):
1392
"""Take out too locks, rolling back the first if the second throws."""
1397
# we want to ensure that we don't leave source locked by mistake.
1398
# and any error on target should not confuse source.
1399
self.source.unlock()
1403
def fetch(self, revision_id=None, pb=None):
1404
"""Fetch the content required to construct revision_id.
1406
The content is copied from source to target.
1408
:param revision_id: if None all content is copied, if NULL_REVISION no
1410
:param pb: optional progress bar to use for progress reports. If not
1411
provided a default one will be created.
1413
Returns the copied revision count and the failed revisions in a tuple:
1416
from bzrlib.fetch import GenericRepoFetcher
1417
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1418
self.source, self.source._format, self.target, self.target._format)
1419
f = GenericRepoFetcher(to_repository=self.target,
1420
from_repository=self.source,
1421
last_revision=revision_id,
1423
return f.count_copied, f.failed_revisions
1425
def lock_read(self):
1426
"""Take out a logical read lock.
1428
This will lock the source branch and the target branch. The source gets
1429
a read lock and the target a read lock.
1431
self._double_lock(self.source.lock_read, self.target.lock_read)
1433
def lock_write(self):
1434
"""Take out a logical write lock.
1436
This will lock the source branch and the target branch. The source gets
1437
a read lock and the target a write lock.
1439
self._double_lock(self.source.lock_read, self.target.lock_write)
1442
def missing_revision_ids(self, revision_id=None):
1443
"""Return the revision ids that source has that target does not.
1445
These are returned in topological order.
1447
:param revision_id: only return revision ids included by this
1450
# generic, possibly worst case, slow code path.
1451
target_ids = set(self.target.all_revision_ids())
1452
if revision_id is not None:
1453
source_ids = self.source.get_ancestry(revision_id)
1454
assert source_ids.pop(0) == None
1456
source_ids = self.source.all_revision_ids()
1457
result_set = set(source_ids).difference(target_ids)
1458
# this may look like a no-op: its not. It preserves the ordering
1459
# other_ids had while only returning the members from other_ids
1460
# that we've decided we need.
1461
return [rev_id for rev_id in source_ids if rev_id in result_set]
1464
"""Release the locks on source and target."""
1466
self.target.unlock()
1468
self.source.unlock()
1471
class InterWeaveRepo(InterRepository):
1472
"""Optimised code paths between Weave based repositories."""
1474
_matching_repo_format = RepositoryFormat7()
1475
"""Repository format for testing with."""
1478
def is_compatible(source, target):
1479
"""Be compatible with known Weave formats.
1481
We dont test for the stores being of specific types becase that
1482
could lead to confusing results, and there is no need to be
1486
return (isinstance(source._format, (RepositoryFormat5,
1488
RepositoryFormat7)) and
1489
isinstance(target._format, (RepositoryFormat5,
1491
RepositoryFormat7)))
1492
except AttributeError:
1496
def copy_content(self, revision_id=None, basis=None):
1497
"""See InterRepository.copy_content()."""
1498
# weave specific optimised path:
1499
if basis is not None:
1500
# copy the basis in, then fetch remaining data.
1501
basis.copy_content_into(self.target, revision_id)
1502
# the basis copy_content_into could misset this.
1504
self.target.set_make_working_trees(self.source.make_working_trees())
1505
except NotImplementedError:
1507
self.target.fetch(self.source, revision_id=revision_id)
1510
self.target.set_make_working_trees(self.source.make_working_trees())
1511
except NotImplementedError:
1513
# FIXME do not peek!
1514
if self.source.control_files._transport.listable():
1515
pb = bzrlib.ui.ui_factory.nested_progress_bar()
1517
self.target.weave_store.copy_all_ids(
1518
self.source.weave_store,
1520
from_transaction=self.source.get_transaction(),
1521
to_transaction=self.target.get_transaction())
1522
pb.update('copying inventory', 0, 1)
1523
self.target.control_weaves.copy_multi(
1524
self.source.control_weaves, ['inventory'],
1525
from_transaction=self.source.get_transaction(),
1526
to_transaction=self.target.get_transaction())
1527
self.target._revision_store.text_store.copy_all_ids(
1528
self.source._revision_store.text_store,
1533
self.target.fetch(self.source, revision_id=revision_id)
1536
def fetch(self, revision_id=None, pb=None):
1537
"""See InterRepository.fetch()."""
1538
from bzrlib.fetch import GenericRepoFetcher
1539
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1540
self.source, self.source._format, self.target, self.target._format)
1541
f = GenericRepoFetcher(to_repository=self.target,
1542
from_repository=self.source,
1543
last_revision=revision_id,
1545
return f.count_copied, f.failed_revisions
1548
def missing_revision_ids(self, revision_id=None):
1549
"""See InterRepository.missing_revision_ids()."""
1550
# we want all revisions to satisfy revision_id in source.
1551
# but we dont want to stat every file here and there.
1552
# we want then, all revisions other needs to satisfy revision_id
1553
# checked, but not those that we have locally.
1554
# so the first thing is to get a subset of the revisions to
1555
# satisfy revision_id in source, and then eliminate those that
1556
# we do already have.
1557
# this is slow on high latency connection to self, but as as this
1558
# disk format scales terribly for push anyway due to rewriting
1559
# inventory.weave, this is considered acceptable.
1561
if revision_id is not None:
1562
source_ids = self.source.get_ancestry(revision_id)
1563
assert source_ids.pop(0) == None
1565
source_ids = self.source._all_possible_ids()
1566
source_ids_set = set(source_ids)
1567
# source_ids is the worst possible case we may need to pull.
1568
# now we want to filter source_ids against what we actually
1569
# have in target, but dont try to check for existence where we know
1570
# we do not have a revision as that would be pointless.
1571
target_ids = set(self.target._all_possible_ids())
1572
possibly_present_revisions = target_ids.intersection(source_ids_set)
1573
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1574
required_revisions = source_ids_set.difference(actually_present_revisions)
1575
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1576
if revision_id is not None:
1577
# we used get_ancestry to determine source_ids then we are assured all
1578
# revisions referenced are present as they are installed in topological order.
1579
# and the tip revision was validated by get_ancestry.
1580
return required_topo_revisions
1582
# if we just grabbed the possibly available ids, then
1583
# we only have an estimate of whats available and need to validate
1584
# that against the revision records.
1585
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1588
class InterKnitRepo(InterRepository):
1589
"""Optimised code paths between Knit based repositories."""
1591
_matching_repo_format = RepositoryFormatKnit1()
1592
"""Repository format for testing with."""
1595
def is_compatible(source, target):
1596
"""Be compatible with known Knit formats.
1598
We dont test for the stores being of specific types becase that
1599
could lead to confusing results, and there is no need to be
1603
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1604
isinstance(target._format, (RepositoryFormatKnit1)))
1605
except AttributeError:
1609
def fetch(self, revision_id=None, pb=None):
1610
"""See InterRepository.fetch()."""
1611
from bzrlib.fetch import KnitRepoFetcher
1612
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1613
self.source, self.source._format, self.target, self.target._format)
1614
f = KnitRepoFetcher(to_repository=self.target,
1615
from_repository=self.source,
1616
last_revision=revision_id,
1618
return f.count_copied, f.failed_revisions
1621
def missing_revision_ids(self, revision_id=None):
1622
"""See InterRepository.missing_revision_ids()."""
1623
if revision_id is not None:
1624
source_ids = self.source.get_ancestry(revision_id)
1625
assert source_ids.pop(0) == None
1627
source_ids = self.source._all_possible_ids()
1628
source_ids_set = set(source_ids)
1629
# source_ids is the worst possible case we may need to pull.
1630
# now we want to filter source_ids against what we actually
1631
# have in target, but dont try to check for existence where we know
1632
# we do not have a revision as that would be pointless.
1633
target_ids = set(self.target._all_possible_ids())
1634
possibly_present_revisions = target_ids.intersection(source_ids_set)
1635
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1636
required_revisions = source_ids_set.difference(actually_present_revisions)
1637
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1638
if revision_id is not None:
1639
# we used get_ancestry to determine source_ids then we are assured all
1640
# revisions referenced are present as they are installed in topological order.
1641
# and the tip revision was validated by get_ancestry.
1642
return required_topo_revisions
1644
# if we just grabbed the possibly available ids, then
1645
# we only have an estimate of whats available and need to validate
1646
# that against the revision records.
1647
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1649
InterRepository.register_optimiser(InterWeaveRepo)
1650
InterRepository.register_optimiser(InterKnitRepo)
1653
class RepositoryTestProviderAdapter(object):
1654
"""A tool to generate a suite testing multiple repository formats at once.
1656
This is done by copying the test once for each transport and injecting
1657
the transport_server, transport_readonly_server, and bzrdir_format and
1658
repository_format classes into each copy. Each copy is also given a new id()
1659
to make it easy to identify.
1662
def __init__(self, transport_server, transport_readonly_server, formats):
1663
self._transport_server = transport_server
1664
self._transport_readonly_server = transport_readonly_server
1665
self._formats = formats
1667
def adapt(self, test):
1668
result = TestSuite()
1669
for repository_format, bzrdir_format in self._formats:
1670
new_test = deepcopy(test)
1671
new_test.transport_server = self._transport_server
1672
new_test.transport_readonly_server = self._transport_readonly_server
1673
new_test.bzrdir_format = bzrdir_format
1674
new_test.repository_format = repository_format
1675
def make_new_test_id():
1676
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1677
return lambda: new_id
1678
new_test.id = make_new_test_id()
1679
result.addTest(new_test)
1683
class InterRepositoryTestProviderAdapter(object):
1684
"""A tool to generate a suite testing multiple inter repository formats.
1686
This is done by copying the test once for each interrepo provider and injecting
1687
the transport_server, transport_readonly_server, repository_format and
1688
repository_to_format classes into each copy.
1689
Each copy is also given a new id() to make it easy to identify.
1692
def __init__(self, transport_server, transport_readonly_server, formats):
1693
self._transport_server = transport_server
1694
self._transport_readonly_server = transport_readonly_server
1695
self._formats = formats
1697
def adapt(self, test):
1698
result = TestSuite()
1699
for interrepo_class, repository_format, repository_format_to in self._formats:
1700
new_test = deepcopy(test)
1701
new_test.transport_server = self._transport_server
1702
new_test.transport_readonly_server = self._transport_readonly_server
1703
new_test.interrepo_class = interrepo_class
1704
new_test.repository_format = repository_format
1705
new_test.repository_format_to = repository_format_to
1706
def make_new_test_id():
1707
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1708
return lambda: new_id
1709
new_test.id = make_new_test_id()
1710
result.addTest(new_test)
1714
def default_test_list():
1715
"""Generate the default list of interrepo permutations to test."""
1717
# test the default InterRepository between format 6 and the current
1719
# XXX: robertc 20060220 reinstate this when there are two supported
1720
# formats which do not have an optimal code path between them.
1721
result.append((InterRepository,
1722
RepositoryFormat6(),
1723
RepositoryFormatKnit1()))
1724
for optimiser in InterRepository._optimisers:
1725
result.append((optimiser,
1726
optimiser._matching_repo_format,
1727
optimiser._matching_repo_format
1729
# if there are specific combinations we want to use, we can add them
1734
class CopyConverter(object):
1735
"""A repository conversion tool which just performs a copy of the content.
1737
This is slow but quite reliable.
1740
def __init__(self, target_format):
1741
"""Create a CopyConverter.
1743
:param target_format: The format the resulting repository should be.
1745
self.target_format = target_format
1747
def convert(self, repo, pb):
1748
"""Perform the conversion of to_convert, giving feedback via pb.
1750
:param to_convert: The disk object to convert.
1751
:param pb: a progress bar to use for progress information.
1756
# this is only useful with metadir layouts - separated repo content.
1757
# trigger an assertion if not such
1758
repo._format.get_format_string()
1759
self.repo_dir = repo.bzrdir
1760
self.step('Moving repository to repository.backup')
1761
self.repo_dir.transport.move('repository', 'repository.backup')
1762
backup_transport = self.repo_dir.transport.clone('repository.backup')
1763
self.source_repo = repo._format.open(self.repo_dir,
1765
_override_transport=backup_transport)
1766
self.step('Creating new repository')
1767
converted = self.target_format.initialize(self.repo_dir,
1768
self.source_repo.is_shared())
1769
converted.lock_write()
1771
self.step('Copying content into repository.')
1772
self.source_repo.copy_content_into(converted)
1775
self.step('Deleting old repository content.')
1776
self.repo_dir.transport.delete_tree('repository.backup')
1777
self.pb.note('repository converted')
1779
def step(self, message):
1780
"""Update the pb by a step."""
1782
self.pb.update(message, self.count, self.total)
1785
# Copied from xml.sax.saxutils
1786
def _unescape_xml(data):
1787
"""Unescape &, <, and > in a string of data.
1789
data = data.replace("<", "<")
1790
data = data.replace(">", ">")
1791
# must do ampersand last
1792
return data.replace("&", "&")
added
removed
bzrlib/repository.py
