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
assert inv.revision_id is None or inv.revision_id == revid, \
68
"Mismatch between inventory revision" \
69
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
70
inv_text = bzrlib.xml5.serializer_v5.write_inventory_to_string(inv)
71
inv_sha1 = bzrlib.osutils.sha_string(inv_text)
72
inv_vf = self.control_weaves.get_weave('inventory',
73
self.get_transaction())
74
inv_vf.add_lines(revid, parents, bzrlib.osutils.split_lines(inv_text))
78
def add_revision(self, rev_id, rev, inv=None, config=None):
79
"""Add rev to the revision store as rev_id.
81
:param rev_id: the revision id to use.
82
:param rev: The revision object.
83
:param inv: The inventory for the revision. if None, it will be looked
84
up in the inventory storer
85
:param config: If None no digital signature will be created.
86
If supplied its signature_needed method will be used
87
to determine if a signature should be made.
89
if config is not None and config.signature_needed():
91
inv = self.get_inventory(rev_id)
92
plaintext = Testament(rev, inv).as_short_text()
93
self.store_revision_signature(
94
gpg.GPGStrategy(config), plaintext, rev_id)
95
if not rev_id in self.get_inventory_weave():
97
raise errors.WeaveRevisionNotPresent(rev_id,
98
self.get_inventory_weave())
100
# yes, this is not suitable for adding with ghosts.
101
self.add_inventory(rev_id, inv, rev.parent_ids)
102
self._revision_store.add_revision(rev, self.get_transaction())
105
def _all_possible_ids(self):
106
"""Return all the possible revisions that we could find."""
107
return self.get_inventory_weave().versions()
110
def all_revision_ids(self):
111
"""Returns a list of all the revision ids in the repository.
113
These are in as much topological order as the underlying store can
114
present: for weaves ghosts may lead to a lack of correctness until
115
the reweave updates the parents list.
117
if self._revision_store.text_store.listable():
118
return self._revision_store.all_revision_ids(self.get_transaction())
119
result = self._all_possible_ids()
120
return self._eliminate_revisions_not_present(result)
122
def break_lock(self):
123
"""Break a lock if one is present from another instance.
125
Uses the ui factory to ask for confirmation if the lock may be from
128
self.control_files.break_lock()
131
def _eliminate_revisions_not_present(self, revision_ids):
132
"""Check every revision id in revision_ids to see if we have it.
134
Returns a set of the present revisions.
137
for id in revision_ids:
138
if self.has_revision(id):
143
def create(a_bzrdir):
144
"""Construct the current default format repository in a_bzrdir."""
145
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
147
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
148
"""instantiate a Repository.
150
:param _format: The format of the repository on disk.
151
:param a_bzrdir: The BzrDir of the repository.
153
In the future we will have a single api for all stores for
154
getting file texts, inventories and revisions, then
155
this construct will accept instances of those things.
157
super(Repository, self).__init__()
158
self._format = _format
159
# the following are part of the public API for Repository:
160
self.bzrdir = a_bzrdir
161
self.control_files = control_files
162
self._revision_store = _revision_store
163
self.text_store = text_store
164
# backwards compatability
165
self.weave_store = text_store
166
# not right yet - should be more semantically clear ?
168
self.control_store = control_store
169
self.control_weaves = control_store
170
# TODO: make sure to construct the right store classes, etc, depending
171
# on whether escaping is required.
174
return '%s(%r)' % (self.__class__.__name__,
175
self.bzrdir.transport.base)
178
return self.control_files.is_locked()
180
def lock_write(self):
181
self.control_files.lock_write()
184
self.control_files.lock_read()
186
def get_physical_lock_status(self):
187
return self.control_files.get_physical_lock_status()
190
def missing_revision_ids(self, other, revision_id=None):
191
"""Return the revision ids that other has that this does not.
193
These are returned in topological order.
195
revision_id: only return revision ids included by revision_id.
197
return InterRepository.get(other, self).missing_revision_ids(revision_id)
201
"""Open the repository rooted at base.
203
For instance, if the repository is at URL/.bzr/repository,
204
Repository.open(URL) -> a Repository instance.
206
control = bzrlib.bzrdir.BzrDir.open(base)
207
return control.open_repository()
209
def copy_content_into(self, destination, revision_id=None, basis=None):
210
"""Make a complete copy of the content in self into destination.
212
This is a destructive operation! Do not use it on existing
215
return InterRepository.get(self, destination).copy_content(revision_id, basis)
217
def fetch(self, source, revision_id=None, pb=None):
218
"""Fetch the content required to construct revision_id from source.
220
If revision_id is None all content is copied.
222
return InterRepository.get(source, self).fetch(revision_id=revision_id,
226
self.control_files.unlock()
229
def clone(self, a_bzrdir, revision_id=None, basis=None):
230
"""Clone this repository into a_bzrdir using the current format.
232
Currently no check is made that the format of this repository and
233
the bzrdir format are compatible. FIXME RBC 20060201.
235
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
236
# use target default format.
237
result = a_bzrdir.create_repository()
238
# FIXME RBC 20060209 split out the repository type to avoid this check ?
239
elif isinstance(a_bzrdir._format,
240
(bzrlib.bzrdir.BzrDirFormat4,
241
bzrlib.bzrdir.BzrDirFormat5,
242
bzrlib.bzrdir.BzrDirFormat6)):
243
result = a_bzrdir.open_repository()
245
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
246
self.copy_content_into(result, revision_id, basis)
250
def has_revision(self, revision_id):
251
"""True if this repository has a copy of the revision."""
252
return self._revision_store.has_revision_id(revision_id,
253
self.get_transaction())
256
def get_revision_reconcile(self, revision_id):
257
"""'reconcile' helper routine that allows access to a revision always.
259
This variant of get_revision does not cross check the weave graph
260
against the revision one as get_revision does: but it should only
261
be used by reconcile, or reconcile-alike commands that are correcting
262
or testing the revision graph.
264
if not revision_id or not isinstance(revision_id, basestring):
265
raise InvalidRevisionId(revision_id=revision_id, branch=self)
266
return self._revision_store.get_revision(revision_id,
267
self.get_transaction())
270
def get_revision_xml(self, revision_id):
271
rev = self.get_revision(revision_id)
273
# the current serializer..
274
self._revision_store._serializer.write_revision(rev, rev_tmp)
276
return rev_tmp.getvalue()
279
def get_revision(self, revision_id):
280
"""Return the Revision object for a named revision"""
281
r = self.get_revision_reconcile(revision_id)
282
# weave corruption can lead to absent revision markers that should be
284
# the following test is reasonably cheap (it needs a single weave read)
285
# and the weave is cached in read transactions. In write transactions
286
# it is not cached but typically we only read a small number of
287
# revisions. For knits when they are introduced we will probably want
288
# to ensure that caching write transactions are in use.
289
inv = self.get_inventory_weave()
290
self._check_revision_parents(r, inv)
293
def _check_revision_parents(self, revision, inventory):
294
"""Private to Repository and Fetch.
296
This checks the parentage of revision in an inventory weave for
297
consistency and is only applicable to inventory-weave-for-ancestry
298
using repository formats & fetchers.
300
weave_parents = inventory.get_parents(revision.revision_id)
301
weave_names = inventory.versions()
302
for parent_id in revision.parent_ids:
303
if parent_id in weave_names:
304
# this parent must not be a ghost.
305
if not parent_id in weave_parents:
307
raise errors.CorruptRepository(self)
310
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
311
signature = gpg_strategy.sign(plaintext)
312
self._revision_store.add_revision_signature_text(revision_id,
314
self.get_transaction())
316
def fileids_altered_by_revision_ids(self, revision_ids):
317
"""Find the file ids and versions affected by revisions.
319
:param revisions: an iterable containing revision ids.
320
:return: a dictionary mapping altered file-ids to an iterable of
321
revision_ids. Each altered file-ids has the exact revision_ids that
322
altered it listed explicitly.
324
assert isinstance(self._format, (RepositoryFormat5,
327
RepositoryFormatKnit1)), \
328
"fileid_involved only supported for branches which store inventory as unnested xml"
329
selected_revision_ids = set(revision_ids)
330
w = self.get_inventory_weave()
333
# this code needs to read every new line in every inventory for the
334
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
335
# not pesent in one of those inventories is unnecessary but not
336
# harmful because we are filtering by the revision id marker in the
337
# inventory lines : we only select file ids altered in one of those
338
# revisions. We dont need to see all lines in the inventory because
339
# only those added in an inventory in rev X can contain a revision=X
341
for line in w.iter_lines_added_or_present_in_versions(selected_revision_ids):
342
start = line.find('file_id="')+9
343
if start < 9: continue
344
end = line.find('"', start)
346
file_id = _unescape_xml(line[start:end])
348
start = line.find('revision="')+10
349
if start < 10: continue
350
end = line.find('"', start)
352
revision_id = _unescape_xml(line[start:end])
353
if revision_id in selected_revision_ids:
354
result.setdefault(file_id, set()).add(revision_id)
358
def get_inventory_weave(self):
359
return self.control_weaves.get_weave('inventory',
360
self.get_transaction())
363
def get_inventory(self, revision_id):
364
"""Get Inventory object by hash."""
365
return self.deserialise_inventory(
366
revision_id, self.get_inventory_xml(revision_id))
368
def deserialise_inventory(self, revision_id, xml):
369
"""Transform the xml into an inventory object.
371
:param revision_id: The expected revision id of the inventory.
372
:param xml: A serialised inventory.
374
return bzrlib.xml5.serializer_v5.read_inventory_from_string(xml)
377
def get_inventory_xml(self, revision_id):
378
"""Get inventory XML as a file object."""
380
assert isinstance(revision_id, basestring), type(revision_id)
381
iw = self.get_inventory_weave()
382
return iw.get_text(revision_id)
384
raise bzrlib.errors.HistoryMissing(self, 'inventory', revision_id)
387
def get_inventory_sha1(self, revision_id):
388
"""Return the sha1 hash of the inventory entry
390
return self.get_revision(revision_id).inventory_sha1
393
def get_revision_graph(self, revision_id=None):
394
"""Return a dictionary containing the revision graph.
396
:return: a dictionary of revision_id->revision_parents_list.
398
weave = self.get_inventory_weave()
399
all_revisions = self._eliminate_revisions_not_present(weave.versions())
400
entire_graph = dict([(node, weave.get_parents(node)) for
401
node in all_revisions])
402
if revision_id is None:
404
elif revision_id not in entire_graph:
405
raise errors.NoSuchRevision(self, revision_id)
407
# add what can be reached from revision_id
409
pending = set([revision_id])
410
while len(pending) > 0:
412
result[node] = entire_graph[node]
413
for revision_id in result[node]:
414
if revision_id not in result:
415
pending.add(revision_id)
419
def get_revision_graph_with_ghosts(self, revision_ids=None):
420
"""Return a graph of the revisions with ghosts marked as applicable.
422
:param revision_ids: an iterable of revisions to graph or None for all.
423
:return: a Graph object with the graph reachable from revision_ids.
427
pending = set(self.all_revision_ids())
430
pending = set(revision_ids)
431
required = set(revision_ids)
434
revision_id = pending.pop()
436
rev = self.get_revision(revision_id)
437
except errors.NoSuchRevision:
438
if revision_id in required:
441
result.add_ghost(revision_id)
443
for parent_id in rev.parent_ids:
444
# is this queued or done ?
445
if (parent_id not in pending and
446
parent_id not in done):
448
pending.add(parent_id)
449
result.add_node(revision_id, rev.parent_ids)
450
done.add(revision_id)
454
def get_revision_inventory(self, revision_id):
455
"""Return inventory of a past revision."""
456
# TODO: Unify this with get_inventory()
457
# bzr 0.0.6 and later imposes the constraint that the inventory_id
458
# must be the same as its revision, so this is trivial.
459
if revision_id is None:
460
# This does not make sense: if there is no revision,
461
# then it is the current tree inventory surely ?!
462
# and thus get_root_id() is something that looks at the last
463
# commit on the branch, and the get_root_id is an inventory check.
464
raise NotImplementedError
465
# return Inventory(self.get_root_id())
467
return self.get_inventory(revision_id)
471
"""Return True if this repository is flagged as a shared repository."""
472
raise NotImplementedError(self.is_shared)
475
def reconcile(self, other=None, thorough=False):
476
"""Reconcile this repository."""
477
from bzrlib.reconcile import RepoReconciler
478
reconciler = RepoReconciler(self, thorough=thorough)
479
reconciler.reconcile()
483
def revision_tree(self, revision_id):
484
"""Return Tree for a revision on this branch.
486
`revision_id` may be None for the null revision, in which case
487
an `EmptyTree` is returned."""
488
# TODO: refactor this to use an existing revision object
489
# so we don't need to read it in twice.
490
if revision_id is None or revision_id == NULL_REVISION:
493
inv = self.get_revision_inventory(revision_id)
494
return RevisionTree(self, inv, revision_id)
497
def get_ancestry(self, revision_id):
498
"""Return a list of revision-ids integrated by a revision.
500
This is topologically sorted.
502
if revision_id is None:
504
if not self.has_revision(revision_id):
505
raise errors.NoSuchRevision(self, revision_id)
506
w = self.get_inventory_weave()
507
candidates = w.get_ancestry(revision_id)
508
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
511
def print_file(self, file, revision_id):
512
"""Print `file` to stdout.
514
FIXME RBC 20060125 as John Meinel points out this is a bad api
515
- it writes to stdout, it assumes that that is valid etc. Fix
516
by creating a new more flexible convenience function.
518
tree = self.revision_tree(revision_id)
519
# use inventory as it was in that revision
520
file_id = tree.inventory.path2id(file)
522
# TODO: jam 20060427 Write a test for this code path
523
# it had a bug in it, and was raising the wrong
525
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
526
tree.print_file(file_id)
528
def get_transaction(self):
529
return self.control_files.get_transaction()
531
def revision_parents(self, revid):
532
return self.get_inventory_weave().parent_names(revid)
535
def set_make_working_trees(self, new_value):
536
"""Set the policy flag for making working trees when creating branches.
538
This only applies to branches that use this repository.
540
The default is 'True'.
541
:param new_value: True to restore the default, False to disable making
544
raise NotImplementedError(self.set_make_working_trees)
546
def make_working_trees(self):
547
"""Returns the policy for making working trees on new branches."""
548
raise NotImplementedError(self.make_working_trees)
551
def sign_revision(self, revision_id, gpg_strategy):
552
plaintext = Testament.from_revision(self, revision_id).as_short_text()
553
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
556
def has_signature_for_revision_id(self, revision_id):
557
"""Query for a revision signature for revision_id in the repository."""
558
return self._revision_store.has_signature(revision_id,
559
self.get_transaction())
562
def get_signature_text(self, revision_id):
563
"""Return the text for a signature."""
564
return self._revision_store.get_signature_text(revision_id,
565
self.get_transaction())
568
class AllInOneRepository(Repository):
569
"""Legacy support - the repository behaviour for all-in-one branches."""
571
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
572
# we reuse one control files instance.
573
dir_mode = a_bzrdir._control_files._dir_mode
574
file_mode = a_bzrdir._control_files._file_mode
576
def get_store(name, compressed=True, prefixed=False):
577
# FIXME: This approach of assuming stores are all entirely compressed
578
# or entirely uncompressed is tidy, but breaks upgrade from
579
# some existing branches where there's a mixture; we probably
580
# still want the option to look for both.
581
relpath = a_bzrdir._control_files._escape(name)
582
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
583
prefixed=prefixed, compressed=compressed,
586
#if self._transport.should_cache():
587
# cache_path = os.path.join(self.cache_root, name)
588
# os.mkdir(cache_path)
589
# store = bzrlib.store.CachedStore(store, cache_path)
592
# not broken out yet because the controlweaves|inventory_store
593
# and text_store | weave_store bits are still different.
594
if isinstance(_format, RepositoryFormat4):
595
# cannot remove these - there is still no consistent api
596
# which allows access to this old info.
597
self.inventory_store = get_store('inventory-store')
598
text_store = get_store('text-store')
599
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
603
"""AllInOne repositories cannot be shared."""
607
def set_make_working_trees(self, new_value):
608
"""Set the policy flag for making working trees when creating branches.
610
This only applies to branches that use this repository.
612
The default is 'True'.
613
:param new_value: True to restore the default, False to disable making
616
raise NotImplementedError(self.set_make_working_trees)
618
def make_working_trees(self):
619
"""Returns the policy for making working trees on new branches."""
623
def install_revision(repository, rev, revision_tree):
624
"""Install all revision data into a repository."""
627
for p_id in rev.parent_ids:
628
if repository.has_revision(p_id):
629
present_parents.append(p_id)
630
parent_trees[p_id] = repository.revision_tree(p_id)
632
parent_trees[p_id] = EmptyTree()
634
inv = revision_tree.inventory
636
# Add the texts that are not already present
637
for path, ie in inv.iter_entries():
638
w = repository.weave_store.get_weave_or_empty(ie.file_id,
639
repository.get_transaction())
640
if ie.revision not in w:
642
# FIXME: TODO: The following loop *may* be overlapping/duplicate
643
# with inventoryEntry.find_previous_heads(). if it is, then there
644
# is a latent bug here where the parents may have ancestors of each
646
for revision, tree in parent_trees.iteritems():
647
if ie.file_id not in tree:
649
parent_id = tree.inventory[ie.file_id].revision
650
if parent_id in text_parents:
652
text_parents.append(parent_id)
654
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
655
repository.get_transaction())
656
lines = revision_tree.get_file(ie.file_id).readlines()
657
vfile.add_lines(rev.revision_id, text_parents, lines)
659
# install the inventory
660
repository.add_inventory(rev.revision_id, inv, present_parents)
661
except errors.RevisionAlreadyPresent:
663
repository.add_revision(rev.revision_id, rev, inv)
666
class MetaDirRepository(Repository):
667
"""Repositories in the new meta-dir layout."""
669
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
670
super(MetaDirRepository, self).__init__(_format,
677
dir_mode = self.control_files._dir_mode
678
file_mode = self.control_files._file_mode
682
"""Return True if this repository is flagged as a shared repository."""
683
return self.control_files._transport.has('shared-storage')
686
def set_make_working_trees(self, new_value):
687
"""Set the policy flag for making working trees when creating branches.
689
This only applies to branches that use this repository.
691
The default is 'True'.
692
:param new_value: True to restore the default, False to disable making
697
self.control_files._transport.delete('no-working-trees')
698
except errors.NoSuchFile:
701
self.control_files.put_utf8('no-working-trees', '')
703
def make_working_trees(self):
704
"""Returns the policy for making working trees on new branches."""
705
return not self.control_files._transport.has('no-working-trees')
708
class KnitRepository(MetaDirRepository):
709
"""Knit format repository."""
712
def all_revision_ids(self):
713
"""See Repository.all_revision_ids()."""
714
return self._revision_store.all_revision_ids(self.get_transaction())
716
def fileid_involved_between_revs(self, from_revid, to_revid):
717
"""Find file_id(s) which are involved in the changes between revisions.
719
This determines the set of revisions which are involved, and then
720
finds all file ids affected by those revisions.
722
vf = self._get_revision_vf()
723
from_set = set(vf.get_ancestry(from_revid))
724
to_set = set(vf.get_ancestry(to_revid))
725
changed = to_set.difference(from_set)
726
return self._fileid_involved_by_set(changed)
728
def fileid_involved(self, last_revid=None):
729
"""Find all file_ids modified in the ancestry of last_revid.
731
:param last_revid: If None, last_revision() will be used.
734
changed = set(self.all_revision_ids())
736
changed = set(self.get_ancestry(last_revid))
739
return self._fileid_involved_by_set(changed)
742
def get_ancestry(self, revision_id):
743
"""Return a list of revision-ids integrated by a revision.
745
This is topologically sorted.
747
if revision_id is None:
749
vf = self._get_revision_vf()
751
return [None] + vf.get_ancestry(revision_id)
752
except errors.RevisionNotPresent:
753
raise errors.NoSuchRevision(self, revision_id)
756
def get_revision(self, revision_id):
757
"""Return the Revision object for a named revision"""
758
return self.get_revision_reconcile(revision_id)
761
def get_revision_graph(self, revision_id=None):
762
"""Return a dictionary containing the revision graph.
764
:return: a dictionary of revision_id->revision_parents_list.
766
weave = self._get_revision_vf()
767
entire_graph = weave.get_graph()
768
if revision_id is None:
769
return weave.get_graph()
770
elif revision_id not in weave:
771
raise errors.NoSuchRevision(self, revision_id)
773
# add what can be reached from revision_id
775
pending = set([revision_id])
776
while len(pending) > 0:
778
result[node] = weave.get_parents(node)
779
for revision_id in result[node]:
780
if revision_id not in result:
781
pending.add(revision_id)
785
def get_revision_graph_with_ghosts(self, revision_ids=None):
786
"""Return a graph of the revisions with ghosts marked as applicable.
788
:param revision_ids: an iterable of revisions to graph or None for all.
789
:return: a Graph object with the graph reachable from revision_ids.
792
vf = self._get_revision_vf()
793
versions = set(vf.versions())
795
pending = set(self.all_revision_ids())
798
pending = set(revision_ids)
799
required = set(revision_ids)
802
revision_id = pending.pop()
803
if not revision_id in versions:
804
if revision_id in required:
805
raise errors.NoSuchRevision(self, revision_id)
807
result.add_ghost(revision_id)
808
# mark it as done so we dont try for it again.
809
done.add(revision_id)
811
parent_ids = vf.get_parents_with_ghosts(revision_id)
812
for parent_id in parent_ids:
813
# is this queued or done ?
814
if (parent_id not in pending and
815
parent_id not in done):
817
pending.add(parent_id)
818
result.add_node(revision_id, parent_ids)
819
done.add(revision_id)
822
def _get_revision_vf(self):
823
""":return: a versioned file containing the revisions."""
824
vf = self._revision_store.get_revision_file(self.get_transaction())
828
def reconcile(self, other=None, thorough=False):
829
"""Reconcile this repository."""
830
from bzrlib.reconcile import KnitReconciler
831
reconciler = KnitReconciler(self, thorough=thorough)
832
reconciler.reconcile()
835
def revision_parents(self, revid):
836
return self._get_revision_vf().get_parents(rev_id)
838
class RepositoryFormat(object):
839
"""A repository format.
841
Formats provide three things:
842
* An initialization routine to construct repository data on disk.
843
* a format string which is used when the BzrDir supports versioned
845
* an open routine which returns a Repository instance.
847
Formats are placed in an dict by their format string for reference
848
during opening. These should be subclasses of RepositoryFormat
851
Once a format is deprecated, just deprecate the initialize and open
852
methods on the format class. Do not deprecate the object, as the
853
object will be created every system load.
855
Common instance attributes:
856
_matchingbzrdir - the bzrdir format that the repository format was
857
originally written to work with. This can be used if manually
858
constructing a bzrdir and repository, or more commonly for test suite
862
_default_format = None
863
"""The default format used for new repositories."""
866
"""The known formats."""
869
def find_format(klass, a_bzrdir):
870
"""Return the format for the repository object in a_bzrdir."""
872
transport = a_bzrdir.get_repository_transport(None)
873
format_string = transport.get("format").read()
874
return klass._formats[format_string]
875
except errors.NoSuchFile:
876
raise errors.NoRepositoryPresent(a_bzrdir)
878
raise errors.UnknownFormatError(format_string)
880
def _get_control_store(self, repo_transport, control_files):
881
"""Return the control store for this repository."""
882
raise NotImplementedError(self._get_control_store)
885
def get_default_format(klass):
886
"""Return the current default format."""
887
return klass._default_format
889
def get_format_string(self):
890
"""Return the ASCII format string that identifies this format.
892
Note that in pre format ?? repositories the format string is
893
not permitted nor written to disk.
895
raise NotImplementedError(self.get_format_string)
897
def get_format_description(self):
898
"""Return the short desciption for this format."""
899
raise NotImplementedError(self.get_format_description)
901
def _get_revision_store(self, repo_transport, control_files):
902
"""Return the revision store object for this a_bzrdir."""
903
raise NotImplementedError(self._get_revision_store)
905
def _get_text_rev_store(self,
912
"""Common logic for getting a revision store for a repository.
914
see self._get_revision_store for the subclass-overridable method to
915
get the store for a repository.
917
from bzrlib.store.revision.text import TextRevisionStore
918
dir_mode = control_files._dir_mode
919
file_mode = control_files._file_mode
920
text_store =TextStore(transport.clone(name),
922
compressed=compressed,
925
_revision_store = TextRevisionStore(text_store, serializer)
926
return _revision_store
928
def _get_versioned_file_store(self,
933
versionedfile_class=WeaveFile,
935
weave_transport = control_files._transport.clone(name)
936
dir_mode = control_files._dir_mode
937
file_mode = control_files._file_mode
938
return VersionedFileStore(weave_transport, prefixed=prefixed,
941
versionedfile_class=versionedfile_class,
944
def initialize(self, a_bzrdir, shared=False):
945
"""Initialize a repository of this format in a_bzrdir.
947
:param a_bzrdir: The bzrdir to put the new repository in it.
948
:param shared: The repository should be initialized as a sharable one.
950
This may raise UninitializableFormat if shared repository are not
951
compatible the a_bzrdir.
954
def is_supported(self):
955
"""Is this format supported?
957
Supported formats must be initializable and openable.
958
Unsupported formats may not support initialization or committing or
959
some other features depending on the reason for not being supported.
963
def open(self, a_bzrdir, _found=False):
964
"""Return an instance of this format for the bzrdir a_bzrdir.
966
_found is a private parameter, do not use it.
968
raise NotImplementedError(self.open)
971
def register_format(klass, format):
972
klass._formats[format.get_format_string()] = format
975
def set_default_format(klass, format):
976
klass._default_format = format
979
def unregister_format(klass, format):
980
assert klass._formats[format.get_format_string()] is format
981
del klass._formats[format.get_format_string()]
984
class PreSplitOutRepositoryFormat(RepositoryFormat):
985
"""Base class for the pre split out repository formats."""
987
def initialize(self, a_bzrdir, shared=False, _internal=False):
988
"""Create a weave repository.
990
TODO: when creating split out bzr branch formats, move this to a common
991
base for Format5, Format6. or something like that.
993
from bzrlib.weavefile import write_weave_v5
994
from bzrlib.weave import Weave
997
raise errors.IncompatibleFormat(self, a_bzrdir._format)
1000
# always initialized when the bzrdir is.
1001
return self.open(a_bzrdir, _found=True)
1003
# Create an empty weave
1005
bzrlib.weavefile.write_weave_v5(Weave(), sio)
1006
empty_weave = sio.getvalue()
1008
mutter('creating repository in %s.', a_bzrdir.transport.base)
1009
dirs = ['revision-store', 'weaves']
1010
files = [('inventory.weave', StringIO(empty_weave)),
1013
# FIXME: RBC 20060125 dont peek under the covers
1014
# NB: no need to escape relative paths that are url safe.
1015
control_files = LockableFiles(a_bzrdir.transport, 'branch-lock',
1017
control_files.create_lock()
1018
control_files.lock_write()
1019
control_files._transport.mkdir_multi(dirs,
1020
mode=control_files._dir_mode)
1022
for file, content in files:
1023
control_files.put(file, content)
1025
control_files.unlock()
1026
return self.open(a_bzrdir, _found=True)
1028
def _get_control_store(self, repo_transport, control_files):
1029
"""Return the control store for this repository."""
1030
return self._get_versioned_file_store('',
1035
def _get_text_store(self, transport, control_files):
1036
"""Get a store for file texts for this format."""
1037
raise NotImplementedError(self._get_text_store)
1039
def open(self, a_bzrdir, _found=False):
1040
"""See RepositoryFormat.open()."""
1042
# we are being called directly and must probe.
1043
raise NotImplementedError
1045
repo_transport = a_bzrdir.get_repository_transport(None)
1046
control_files = a_bzrdir._control_files
1047
text_store = self._get_text_store(repo_transport, control_files)
1048
control_store = self._get_control_store(repo_transport, control_files)
1049
_revision_store = self._get_revision_store(repo_transport, control_files)
1050
return AllInOneRepository(_format=self,
1052
_revision_store=_revision_store,
1053
control_store=control_store,
1054
text_store=text_store)
1057
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1058
"""Bzr repository format 4.
1060
This repository format has:
1062
- TextStores for texts, inventories,revisions.
1064
This format is deprecated: it indexes texts using a text id which is
1065
removed in format 5; initializationa and write support for this format
1070
super(RepositoryFormat4, self).__init__()
1071
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat4()
1073
def get_format_description(self):
1074
"""See RepositoryFormat.get_format_description()."""
1075
return "Repository format 4"
1077
def initialize(self, url, shared=False, _internal=False):
1078
"""Format 4 branches cannot be created."""
1079
raise errors.UninitializableFormat(self)
1081
def is_supported(self):
1082
"""Format 4 is not supported.
1084
It is not supported because the model changed from 4 to 5 and the
1085
conversion logic is expensive - so doing it on the fly was not
1090
def _get_control_store(self, repo_transport, control_files):
1091
"""Format 4 repositories have no formal control store at this point.
1093
This will cause any control-file-needing apis to fail - this is desired.
1097
def _get_revision_store(self, repo_transport, control_files):
1098
"""See RepositoryFormat._get_revision_store()."""
1099
from bzrlib.xml4 import serializer_v4
1100
return self._get_text_rev_store(repo_transport,
1103
serializer=serializer_v4)
1105
def _get_text_store(self, transport, control_files):
1106
"""See RepositoryFormat._get_text_store()."""
1109
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1110
"""Bzr control format 5.
1112
This repository format has:
1113
- weaves for file texts and inventory
1115
- TextStores for revisions and signatures.
1119
super(RepositoryFormat5, self).__init__()
1120
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat5()
1122
def get_format_description(self):
1123
"""See RepositoryFormat.get_format_description()."""
1124
return "Weave repository format 5"
1126
def _get_revision_store(self, repo_transport, control_files):
1127
"""See RepositoryFormat._get_revision_store()."""
1128
"""Return the revision store object for this a_bzrdir."""
1129
return self._get_text_rev_store(repo_transport,
1134
def _get_text_store(self, transport, control_files):
1135
"""See RepositoryFormat._get_text_store()."""
1136
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1139
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1140
"""Bzr control format 6.
1142
This repository format has:
1143
- weaves for file texts and inventory
1144
- hash subdirectory based stores.
1145
- TextStores for revisions and signatures.
1149
super(RepositoryFormat6, self).__init__()
1150
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat6()
1152
def get_format_description(self):
1153
"""See RepositoryFormat.get_format_description()."""
1154
return "Weave repository format 6"
1156
def _get_revision_store(self, repo_transport, control_files):
1157
"""See RepositoryFormat._get_revision_store()."""
1158
return self._get_text_rev_store(repo_transport,
1164
def _get_text_store(self, transport, control_files):
1165
"""See RepositoryFormat._get_text_store()."""
1166
return self._get_versioned_file_store('weaves', transport, control_files)
1169
class MetaDirRepositoryFormat(RepositoryFormat):
1170
"""Common base class for the new repositories using the metadir layour."""
1173
super(MetaDirRepositoryFormat, self).__init__()
1174
self._matchingbzrdir = bzrlib.bzrdir.BzrDirMetaFormat1()
1176
def _create_control_files(self, a_bzrdir):
1177
"""Create the required files and the initial control_files object."""
1178
# FIXME: RBC 20060125 dont peek under the covers
1179
# NB: no need to escape relative paths that are url safe.
1180
repository_transport = a_bzrdir.get_repository_transport(self)
1181
control_files = LockableFiles(repository_transport, 'lock', LockDir)
1182
control_files.create_lock()
1183
return control_files
1185
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1186
"""Upload the initial blank content."""
1187
control_files = self._create_control_files(a_bzrdir)
1188
control_files.lock_write()
1190
control_files._transport.mkdir_multi(dirs,
1191
mode=control_files._dir_mode)
1192
for file, content in files:
1193
control_files.put(file, content)
1194
for file, content in utf8_files:
1195
control_files.put_utf8(file, content)
1197
control_files.put_utf8('shared-storage', '')
1199
control_files.unlock()
1202
class RepositoryFormat7(MetaDirRepositoryFormat):
1203
"""Bzr repository 7.
1205
This repository format has:
1206
- weaves for file texts and inventory
1207
- hash subdirectory based stores.
1208
- TextStores for revisions and signatures.
1209
- a format marker of its own
1210
- an optional 'shared-storage' flag
1211
- an optional 'no-working-trees' flag
1214
def _get_control_store(self, repo_transport, control_files):
1215
"""Return the control store for this repository."""
1216
return self._get_versioned_file_store('',
1221
def get_format_string(self):
1222
"""See RepositoryFormat.get_format_string()."""
1223
return "Bazaar-NG Repository format 7"
1225
def get_format_description(self):
1226
"""See RepositoryFormat.get_format_description()."""
1227
return "Weave repository format 7"
1229
def _get_revision_store(self, repo_transport, control_files):
1230
"""See RepositoryFormat._get_revision_store()."""
1231
return self._get_text_rev_store(repo_transport,
1238
def _get_text_store(self, transport, control_files):
1239
"""See RepositoryFormat._get_text_store()."""
1240
return self._get_versioned_file_store('weaves',
1244
def initialize(self, a_bzrdir, shared=False):
1245
"""Create a weave repository.
1247
:param shared: If true the repository will be initialized as a shared
1250
from bzrlib.weavefile import write_weave_v5
1251
from bzrlib.weave import Weave
1253
# Create an empty weave
1255
bzrlib.weavefile.write_weave_v5(Weave(), sio)
1256
empty_weave = sio.getvalue()
1258
mutter('creating repository in %s.', a_bzrdir.transport.base)
1259
dirs = ['revision-store', 'weaves']
1260
files = [('inventory.weave', StringIO(empty_weave)),
1262
utf8_files = [('format', self.get_format_string())]
1264
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1265
return self.open(a_bzrdir=a_bzrdir, _found=True)
1267
def open(self, a_bzrdir, _found=False, _override_transport=None):
1268
"""See RepositoryFormat.open().
1270
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1271
repository at a slightly different url
1272
than normal. I.e. during 'upgrade'.
1275
format = RepositoryFormat.find_format(a_bzrdir)
1276
assert format.__class__ == self.__class__
1277
if _override_transport is not None:
1278
repo_transport = _override_transport
1280
repo_transport = a_bzrdir.get_repository_transport(None)
1281
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1282
text_store = self._get_text_store(repo_transport, control_files)
1283
control_store = self._get_control_store(repo_transport, control_files)
1284
_revision_store = self._get_revision_store(repo_transport, control_files)
1285
return MetaDirRepository(_format=self,
1287
control_files=control_files,
1288
_revision_store=_revision_store,
1289
control_store=control_store,
1290
text_store=text_store)
1293
class RepositoryFormatKnit1(MetaDirRepositoryFormat):
1294
"""Bzr repository knit format 1.
1296
This repository format has:
1297
- knits for file texts and inventory
1298
- hash subdirectory based stores.
1299
- knits for revisions and signatures
1300
- TextStores for revisions and signatures.
1301
- a format marker of its own
1302
- an optional 'shared-storage' flag
1303
- an optional 'no-working-trees' flag
1306
This format was introduced in bzr 0.8.
1309
def _get_control_store(self, repo_transport, control_files):
1310
"""Return the control store for this repository."""
1311
return VersionedFileStore(
1314
file_mode=control_files._file_mode,
1315
versionedfile_class=KnitVersionedFile,
1316
versionedfile_kwargs={'factory':KnitPlainFactory()},
1319
def get_format_string(self):
1320
"""See RepositoryFormat.get_format_string()."""
1321
return "Bazaar-NG Knit Repository Format 1"
1323
def get_format_description(self):
1324
"""See RepositoryFormat.get_format_description()."""
1325
return "Knit repository format 1"
1327
def _get_revision_store(self, repo_transport, control_files):
1328
"""See RepositoryFormat._get_revision_store()."""
1329
from bzrlib.store.revision.knit import KnitRevisionStore
1330
versioned_file_store = VersionedFileStore(
1332
file_mode=control_files._file_mode,
1335
versionedfile_class=KnitVersionedFile,
1336
versionedfile_kwargs={'delta':False, 'factory':KnitPlainFactory()},
1339
return KnitRevisionStore(versioned_file_store)
1341
def _get_text_store(self, transport, control_files):
1342
"""See RepositoryFormat._get_text_store()."""
1343
return self._get_versioned_file_store('knits',
1346
versionedfile_class=KnitVersionedFile,
1349
def initialize(self, a_bzrdir, shared=False):
1350
"""Create a knit format 1 repository.
1352
:param a_bzrdir: bzrdir to contain the new repository; must already
1354
:param shared: If true the repository will be initialized as a shared
1357
mutter('creating repository in %s.', a_bzrdir.transport.base)
1358
dirs = ['revision-store', 'knits']
1360
utf8_files = [('format', self.get_format_string())]
1362
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1363
repo_transport = a_bzrdir.get_repository_transport(None)
1364
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1365
control_store = self._get_control_store(repo_transport, control_files)
1366
transaction = bzrlib.transactions.WriteTransaction()
1367
# trigger a write of the inventory store.
1368
control_store.get_weave_or_empty('inventory', transaction)
1369
_revision_store = self._get_revision_store(repo_transport, control_files)
1370
_revision_store.has_revision_id('A', transaction)
1371
_revision_store.get_signature_file(transaction)
1372
return self.open(a_bzrdir=a_bzrdir, _found=True)
1374
def open(self, a_bzrdir, _found=False, _override_transport=None):
1375
"""See RepositoryFormat.open().
1377
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1378
repository at a slightly different url
1379
than normal. I.e. during 'upgrade'.
1382
format = RepositoryFormat.find_format(a_bzrdir)
1383
assert format.__class__ == self.__class__
1384
if _override_transport is not None:
1385
repo_transport = _override_transport
1387
repo_transport = a_bzrdir.get_repository_transport(None)
1388
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1389
text_store = self._get_text_store(repo_transport, control_files)
1390
control_store = self._get_control_store(repo_transport, control_files)
1391
_revision_store = self._get_revision_store(repo_transport, control_files)
1392
return KnitRepository(_format=self,
1394
control_files=control_files,
1395
_revision_store=_revision_store,
1396
control_store=control_store,
1397
text_store=text_store)
1400
# formats which have no format string are not discoverable
1401
# and not independently creatable, so are not registered.
1402
RepositoryFormat.register_format(RepositoryFormat7())
1403
_default_format = RepositoryFormatKnit1()
1404
RepositoryFormat.register_format(_default_format)
1405
RepositoryFormat.set_default_format(_default_format)
1406
_legacy_formats = [RepositoryFormat4(),
1407
RepositoryFormat5(),
1408
RepositoryFormat6()]
1411
class InterRepository(InterObject):
1412
"""This class represents operations taking place between two repositories.
1414
Its instances have methods like copy_content and fetch, and contain
1415
references to the source and target repositories these operations can be
1418
Often we will provide convenience methods on 'repository' which carry out
1419
operations with another repository - they will always forward to
1420
InterRepository.get(other).method_name(parameters).
1424
"""The available optimised InterRepository types."""
1427
def copy_content(self, revision_id=None, basis=None):
1428
"""Make a complete copy of the content in self into destination.
1430
This is a destructive operation! Do not use it on existing
1433
:param revision_id: Only copy the content needed to construct
1434
revision_id and its parents.
1435
:param basis: Copy the needed data preferentially from basis.
1438
self.target.set_make_working_trees(self.source.make_working_trees())
1439
except NotImplementedError:
1441
# grab the basis available data
1442
if basis is not None:
1443
self.target.fetch(basis, revision_id=revision_id)
1444
# but dont bother fetching if we have the needed data now.
1445
if (revision_id not in (None, NULL_REVISION) and
1446
self.target.has_revision(revision_id)):
1448
self.target.fetch(self.source, revision_id=revision_id)
1450
def _double_lock(self, lock_source, lock_target):
1451
"""Take out too locks, rolling back the first if the second throws."""
1456
# we want to ensure that we don't leave source locked by mistake.
1457
# and any error on target should not confuse source.
1458
self.source.unlock()
1462
def fetch(self, revision_id=None, pb=None):
1463
"""Fetch the content required to construct revision_id.
1465
The content is copied from source to target.
1467
:param revision_id: if None all content is copied, if NULL_REVISION no
1469
:param pb: optional progress bar to use for progress reports. If not
1470
provided a default one will be created.
1472
Returns the copied revision count and the failed revisions in a tuple:
1475
from bzrlib.fetch import GenericRepoFetcher
1476
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1477
self.source, self.source._format, self.target, self.target._format)
1478
f = GenericRepoFetcher(to_repository=self.target,
1479
from_repository=self.source,
1480
last_revision=revision_id,
1482
return f.count_copied, f.failed_revisions
1484
def lock_read(self):
1485
"""Take out a logical read lock.
1487
This will lock the source branch and the target branch. The source gets
1488
a read lock and the target a read lock.
1490
self._double_lock(self.source.lock_read, self.target.lock_read)
1492
def lock_write(self):
1493
"""Take out a logical write lock.
1495
This will lock the source branch and the target branch. The source gets
1496
a read lock and the target a write lock.
1498
self._double_lock(self.source.lock_read, self.target.lock_write)
1501
def missing_revision_ids(self, revision_id=None):
1502
"""Return the revision ids that source has that target does not.
1504
These are returned in topological order.
1506
:param revision_id: only return revision ids included by this
1509
# generic, possibly worst case, slow code path.
1510
target_ids = set(self.target.all_revision_ids())
1511
if revision_id is not None:
1512
source_ids = self.source.get_ancestry(revision_id)
1513
assert source_ids[0] == None
1516
source_ids = self.source.all_revision_ids()
1517
result_set = set(source_ids).difference(target_ids)
1518
# this may look like a no-op: its not. It preserves the ordering
1519
# other_ids had while only returning the members from other_ids
1520
# that we've decided we need.
1521
return [rev_id for rev_id in source_ids if rev_id in result_set]
1524
"""Release the locks on source and target."""
1526
self.target.unlock()
1528
self.source.unlock()
1531
class InterWeaveRepo(InterRepository):
1532
"""Optimised code paths between Weave based repositories."""
1534
_matching_repo_format = RepositoryFormat7()
1535
"""Repository format for testing with."""
1538
def is_compatible(source, target):
1539
"""Be compatible with known Weave formats.
1541
We dont test for the stores being of specific types becase that
1542
could lead to confusing results, and there is no need to be
1546
return (isinstance(source._format, (RepositoryFormat5,
1548
RepositoryFormat7)) and
1549
isinstance(target._format, (RepositoryFormat5,
1551
RepositoryFormat7)))
1552
except AttributeError:
1556
def copy_content(self, revision_id=None, basis=None):
1557
"""See InterRepository.copy_content()."""
1558
# weave specific optimised path:
1559
if basis is not None:
1560
# copy the basis in, then fetch remaining data.
1561
basis.copy_content_into(self.target, revision_id)
1562
# the basis copy_content_into could misset this.
1564
self.target.set_make_working_trees(self.source.make_working_trees())
1565
except NotImplementedError:
1567
self.target.fetch(self.source, revision_id=revision_id)
1570
self.target.set_make_working_trees(self.source.make_working_trees())
1571
except NotImplementedError:
1573
# FIXME do not peek!
1574
if self.source.control_files._transport.listable():
1575
pb = bzrlib.ui.ui_factory.nested_progress_bar()
1577
self.target.weave_store.copy_all_ids(
1578
self.source.weave_store,
1580
from_transaction=self.source.get_transaction(),
1581
to_transaction=self.target.get_transaction())
1582
pb.update('copying inventory', 0, 1)
1583
self.target.control_weaves.copy_multi(
1584
self.source.control_weaves, ['inventory'],
1585
from_transaction=self.source.get_transaction(),
1586
to_transaction=self.target.get_transaction())
1587
self.target._revision_store.text_store.copy_all_ids(
1588
self.source._revision_store.text_store,
1593
self.target.fetch(self.source, revision_id=revision_id)
1596
def fetch(self, revision_id=None, pb=None):
1597
"""See InterRepository.fetch()."""
1598
from bzrlib.fetch import GenericRepoFetcher
1599
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1600
self.source, self.source._format, self.target, self.target._format)
1601
f = GenericRepoFetcher(to_repository=self.target,
1602
from_repository=self.source,
1603
last_revision=revision_id,
1605
return f.count_copied, f.failed_revisions
1608
def missing_revision_ids(self, revision_id=None):
1609
"""See InterRepository.missing_revision_ids()."""
1610
# we want all revisions to satisfy revision_id in source.
1611
# but we dont want to stat every file here and there.
1612
# we want then, all revisions other needs to satisfy revision_id
1613
# checked, but not those that we have locally.
1614
# so the first thing is to get a subset of the revisions to
1615
# satisfy revision_id in source, and then eliminate those that
1616
# we do already have.
1617
# this is slow on high latency connection to self, but as as this
1618
# disk format scales terribly for push anyway due to rewriting
1619
# inventory.weave, this is considered acceptable.
1621
if revision_id is not None:
1622
source_ids = self.source.get_ancestry(revision_id)
1623
assert source_ids[0] == None
1626
source_ids = self.source._all_possible_ids()
1627
source_ids_set = set(source_ids)
1628
# source_ids is the worst possible case we may need to pull.
1629
# now we want to filter source_ids against what we actually
1630
# have in target, but dont try to check for existence where we know
1631
# we do not have a revision as that would be pointless.
1632
target_ids = set(self.target._all_possible_ids())
1633
possibly_present_revisions = target_ids.intersection(source_ids_set)
1634
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1635
required_revisions = source_ids_set.difference(actually_present_revisions)
1636
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1637
if revision_id is not None:
1638
# we used get_ancestry to determine source_ids then we are assured all
1639
# revisions referenced are present as they are installed in topological order.
1640
# and the tip revision was validated by get_ancestry.
1641
return required_topo_revisions
1643
# if we just grabbed the possibly available ids, then
1644
# we only have an estimate of whats available and need to validate
1645
# that against the revision records.
1646
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1649
class InterKnitRepo(InterRepository):
1650
"""Optimised code paths between Knit based repositories."""
1652
_matching_repo_format = RepositoryFormatKnit1()
1653
"""Repository format for testing with."""
1656
def is_compatible(source, target):
1657
"""Be compatible with known Knit formats.
1659
We dont test for the stores being of specific types becase that
1660
could lead to confusing results, and there is no need to be
1664
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1665
isinstance(target._format, (RepositoryFormatKnit1)))
1666
except AttributeError:
1670
def fetch(self, revision_id=None, pb=None):
1671
"""See InterRepository.fetch()."""
1672
from bzrlib.fetch import KnitRepoFetcher
1673
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1674
self.source, self.source._format, self.target, self.target._format)
1675
f = KnitRepoFetcher(to_repository=self.target,
1676
from_repository=self.source,
1677
last_revision=revision_id,
1679
return f.count_copied, f.failed_revisions
1682
def missing_revision_ids(self, revision_id=None):
1683
"""See InterRepository.missing_revision_ids()."""
1684
if revision_id is not None:
1685
source_ids = self.source.get_ancestry(revision_id)
1686
assert source_ids[0] == None
1689
source_ids = self.source._all_possible_ids()
1690
source_ids_set = set(source_ids)
1691
# source_ids is the worst possible case we may need to pull.
1692
# now we want to filter source_ids against what we actually
1693
# have in target, but dont try to check for existence where we know
1694
# we do not have a revision as that would be pointless.
1695
target_ids = set(self.target._all_possible_ids())
1696
possibly_present_revisions = target_ids.intersection(source_ids_set)
1697
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1698
required_revisions = source_ids_set.difference(actually_present_revisions)
1699
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1700
if revision_id is not None:
1701
# we used get_ancestry to determine source_ids then we are assured all
1702
# revisions referenced are present as they are installed in topological order.
1703
# and the tip revision was validated by get_ancestry.
1704
return required_topo_revisions
1706
# if we just grabbed the possibly available ids, then
1707
# we only have an estimate of whats available and need to validate
1708
# that against the revision records.
1709
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1711
InterRepository.register_optimiser(InterWeaveRepo)
1712
InterRepository.register_optimiser(InterKnitRepo)
1715
class RepositoryTestProviderAdapter(object):
1716
"""A tool to generate a suite testing multiple repository formats at once.
1718
This is done by copying the test once for each transport and injecting
1719
the transport_server, transport_readonly_server, and bzrdir_format and
1720
repository_format classes into each copy. Each copy is also given a new id()
1721
to make it easy to identify.
1724
def __init__(self, transport_server, transport_readonly_server, formats):
1725
self._transport_server = transport_server
1726
self._transport_readonly_server = transport_readonly_server
1727
self._formats = formats
1729
def adapt(self, test):
1730
result = TestSuite()
1731
for repository_format, bzrdir_format in self._formats:
1732
new_test = deepcopy(test)
1733
new_test.transport_server = self._transport_server
1734
new_test.transport_readonly_server = self._transport_readonly_server
1735
new_test.bzrdir_format = bzrdir_format
1736
new_test.repository_format = repository_format
1737
def make_new_test_id():
1738
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1739
return lambda: new_id
1740
new_test.id = make_new_test_id()
1741
result.addTest(new_test)
1745
class InterRepositoryTestProviderAdapter(object):
1746
"""A tool to generate a suite testing multiple inter repository formats.
1748
This is done by copying the test once for each interrepo provider and injecting
1749
the transport_server, transport_readonly_server, repository_format and
1750
repository_to_format classes into each copy.
1751
Each copy is also given a new id() to make it easy to identify.
1754
def __init__(self, transport_server, transport_readonly_server, formats):
1755
self._transport_server = transport_server
1756
self._transport_readonly_server = transport_readonly_server
1757
self._formats = formats
1759
def adapt(self, test):
1760
result = TestSuite()
1761
for interrepo_class, repository_format, repository_format_to in self._formats:
1762
new_test = deepcopy(test)
1763
new_test.transport_server = self._transport_server
1764
new_test.transport_readonly_server = self._transport_readonly_server
1765
new_test.interrepo_class = interrepo_class
1766
new_test.repository_format = repository_format
1767
new_test.repository_format_to = repository_format_to
1768
def make_new_test_id():
1769
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1770
return lambda: new_id
1771
new_test.id = make_new_test_id()
1772
result.addTest(new_test)
1776
def default_test_list():
1777
"""Generate the default list of interrepo permutations to test."""
1779
# test the default InterRepository between format 6 and the current
1781
# XXX: robertc 20060220 reinstate this when there are two supported
1782
# formats which do not have an optimal code path between them.
1783
result.append((InterRepository,
1784
RepositoryFormat6(),
1785
RepositoryFormatKnit1()))
1786
for optimiser in InterRepository._optimisers:
1787
result.append((optimiser,
1788
optimiser._matching_repo_format,
1789
optimiser._matching_repo_format
1791
# if there are specific combinations we want to use, we can add them
1796
class CopyConverter(object):
1797
"""A repository conversion tool which just performs a copy of the content.
1799
This is slow but quite reliable.
1802
def __init__(self, target_format):
1803
"""Create a CopyConverter.
1805
:param target_format: The format the resulting repository should be.
1807
self.target_format = target_format
1809
def convert(self, repo, pb):
1810
"""Perform the conversion of to_convert, giving feedback via pb.
1812
:param to_convert: The disk object to convert.
1813
:param pb: a progress bar to use for progress information.
1818
# this is only useful with metadir layouts - separated repo content.
1819
# trigger an assertion if not such
1820
repo._format.get_format_string()
1821
self.repo_dir = repo.bzrdir
1822
self.step('Moving repository to repository.backup')
1823
self.repo_dir.transport.move('repository', 'repository.backup')
1824
backup_transport = self.repo_dir.transport.clone('repository.backup')
1825
self.source_repo = repo._format.open(self.repo_dir,
1827
_override_transport=backup_transport)
1828
self.step('Creating new repository')
1829
converted = self.target_format.initialize(self.repo_dir,
1830
self.source_repo.is_shared())
1831
converted.lock_write()
1833
self.step('Copying content into repository.')
1834
self.source_repo.copy_content_into(converted)
1837
self.step('Deleting old repository content.')
1838
self.repo_dir.transport.delete_tree('repository.backup')
1839
self.pb.note('repository converted')
1841
def step(self, message):
1842
"""Update the pb by a step."""
1844
self.pb.update(message, self.count, self.total)
1847
# Copied from xml.sax.saxutils
1848
def _unescape_xml(data):
1849
"""Unescape &, <, and > in a string of data.
1851
data = data.replace("<", "<")
1852
data = data.replace(">", ">")
1853
# must do ampersand last
1854
return data.replace("&", "&")
added
removed
bzrlib/repository.py
