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)
119
def break_lock(self):
120
"""Break a lock if one is present from another instance.
122
Uses the ui factory to ask for confirmation if the lock may be from
125
self.control_files.break_lock()
128
def _eliminate_revisions_not_present(self, revision_ids):
129
"""Check every revision id in revision_ids to see if we have it.
131
Returns a set of the present revisions.
134
for id in revision_ids:
135
if self.has_revision(id):
140
def create(a_bzrdir):
141
"""Construct the current default format repository in a_bzrdir."""
142
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
144
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
145
"""instantiate a Repository.
147
:param _format: The format of the repository on disk.
148
:param a_bzrdir: The BzrDir of the repository.
150
In the future we will have a single api for all stores for
151
getting file texts, inventories and revisions, then
152
this construct will accept instances of those things.
154
super(Repository, self).__init__()
155
self._format = _format
156
# the following are part of the public API for Repository:
157
self.bzrdir = a_bzrdir
158
self.control_files = control_files
159
self._revision_store = _revision_store
160
self.text_store = text_store
161
# backwards compatability
162
self.weave_store = text_store
163
# not right yet - should be more semantically clear ?
165
self.control_store = control_store
166
self.control_weaves = control_store
167
# TODO: make sure to construct the right store classes, etc, depending
168
# on whether escaping is required.
171
return self.control_files.is_locked()
173
def lock_write(self):
174
self.control_files.lock_write()
177
self.control_files.lock_read()
179
def get_physical_lock_status(self):
180
return self.control_files.get_physical_lock_status()
183
def missing_revision_ids(self, other, revision_id=None):
184
"""Return the revision ids that other has that this does not.
186
These are returned in topological order.
188
revision_id: only return revision ids included by revision_id.
190
return InterRepository.get(other, self).missing_revision_ids(revision_id)
194
"""Open the repository rooted at base.
196
For instance, if the repository is at URL/.bzr/repository,
197
Repository.open(URL) -> a Repository instance.
199
control = bzrlib.bzrdir.BzrDir.open(base)
200
return control.open_repository()
202
def copy_content_into(self, destination, revision_id=None, basis=None):
203
"""Make a complete copy of the content in self into destination.
205
This is a destructive operation! Do not use it on existing
208
return InterRepository.get(self, destination).copy_content(revision_id, basis)
210
def fetch(self, source, revision_id=None, pb=None):
211
"""Fetch the content required to construct revision_id from source.
213
If revision_id is None all content is copied.
215
return InterRepository.get(source, self).fetch(revision_id=revision_id,
219
self.control_files.unlock()
222
def clone(self, a_bzrdir, revision_id=None, basis=None):
223
"""Clone this repository into a_bzrdir using the current format.
225
Currently no check is made that the format of this repository and
226
the bzrdir format are compatible. FIXME RBC 20060201.
228
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
229
# use target default format.
230
result = a_bzrdir.create_repository()
231
# FIXME RBC 20060209 split out the repository type to avoid this check ?
232
elif isinstance(a_bzrdir._format,
233
(bzrlib.bzrdir.BzrDirFormat4,
234
bzrlib.bzrdir.BzrDirFormat5,
235
bzrlib.bzrdir.BzrDirFormat6)):
236
result = a_bzrdir.open_repository()
238
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
239
self.copy_content_into(result, revision_id, basis)
243
def has_revision(self, revision_id):
244
"""True if this repository has a copy of the revision."""
245
return self._revision_store.has_revision_id(revision_id,
246
self.get_transaction())
249
def get_revision_reconcile(self, revision_id):
250
"""'reconcile' helper routine that allows access to a revision always.
252
This variant of get_revision does not cross check the weave graph
253
against the revision one as get_revision does: but it should only
254
be used by reconcile, or reconcile-alike commands that are correcting
255
or testing the revision graph.
257
if not revision_id or not isinstance(revision_id, basestring):
258
raise InvalidRevisionId(revision_id=revision_id, branch=self)
259
return self._revision_store.get_revision(revision_id,
260
self.get_transaction())
263
def get_revision_xml(self, revision_id):
264
rev = self.get_revision(revision_id)
266
# the current serializer..
267
self._revision_store._serializer.write_revision(rev, rev_tmp)
269
return rev_tmp.getvalue()
272
def get_revision(self, revision_id):
273
"""Return the Revision object for a named revision"""
274
r = self.get_revision_reconcile(revision_id)
275
# weave corruption can lead to absent revision markers that should be
277
# the following test is reasonably cheap (it needs a single weave read)
278
# and the weave is cached in read transactions. In write transactions
279
# it is not cached but typically we only read a small number of
280
# revisions. For knits when they are introduced we will probably want
281
# to ensure that caching write transactions are in use.
282
inv = self.get_inventory_weave()
283
self._check_revision_parents(r, inv)
286
def _check_revision_parents(self, revision, inventory):
287
"""Private to Repository and Fetch.
289
This checks the parentage of revision in an inventory weave for
290
consistency and is only applicable to inventory-weave-for-ancestry
291
using repository formats & fetchers.
293
weave_parents = inventory.get_parents(revision.revision_id)
294
weave_names = inventory.versions()
295
for parent_id in revision.parent_ids:
296
if parent_id in weave_names:
297
# this parent must not be a ghost.
298
if not parent_id in weave_parents:
300
raise errors.CorruptRepository(self)
303
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
304
signature = gpg_strategy.sign(plaintext)
305
self._revision_store.add_revision_signature_text(revision_id,
307
self.get_transaction())
309
def fileids_altered_by_revision_ids(self, revision_ids):
310
"""Find the file ids and versions affected by revisions.
312
:param revisions: an iterable containing revision ids.
313
:return: a dictionary mapping altered file-ids to an iterable of
314
revision_ids. Each altered file-ids has the exact revision_ids that
315
altered it listed explicitly.
317
assert isinstance(self._format, (RepositoryFormat5,
320
RepositoryFormatKnit1)), \
321
"fileid_involved only supported for branches which store inventory as unnested xml"
322
selected_revision_ids = set(revision_ids)
323
w = self.get_inventory_weave()
326
# this code needs to read every new line in every inventory for the
327
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
328
# not pesent in one of those inventories is unnecessary but not
329
# harmful because we are filtering by the revision id marker in the
330
# inventory lines : we only select file ids altered in one of those
331
# revisions. We dont need to see all lines in the inventory because
332
# only those added in an inventory in rev X can contain a revision=X
334
for line in w.iter_lines_added_or_present_in_versions(selected_revision_ids):
335
start = line.find('file_id="')+9
336
if start < 9: continue
337
end = line.find('"', start)
339
file_id = _unescape_xml(line[start:end])
341
start = line.find('revision="')+10
342
if start < 10: continue
343
end = line.find('"', start)
345
revision_id = _unescape_xml(line[start:end])
346
if revision_id in selected_revision_ids:
347
result.setdefault(file_id, set()).add(revision_id)
351
def get_inventory_weave(self):
352
return self.control_weaves.get_weave('inventory',
353
self.get_transaction())
356
def get_inventory(self, revision_id):
357
"""Get Inventory object by hash."""
358
xml = self.get_inventory_xml(revision_id)
359
return bzrlib.xml5.serializer_v5.read_inventory_from_string(xml)
362
def get_inventory_xml(self, revision_id):
363
"""Get inventory XML as a file object."""
365
assert isinstance(revision_id, basestring), type(revision_id)
366
iw = self.get_inventory_weave()
367
return iw.get_text(revision_id)
369
raise bzrlib.errors.HistoryMissing(self, 'inventory', revision_id)
372
def get_inventory_sha1(self, revision_id):
373
"""Return the sha1 hash of the inventory entry
375
return self.get_revision(revision_id).inventory_sha1
378
def get_revision_graph(self, revision_id=None):
379
"""Return a dictionary containing the revision graph.
381
:return: a dictionary of revision_id->revision_parents_list.
383
weave = self.get_inventory_weave()
384
all_revisions = self._eliminate_revisions_not_present(weave.versions())
385
entire_graph = dict([(node, weave.get_parents(node)) for
386
node in all_revisions])
387
if revision_id is None:
389
elif revision_id not in entire_graph:
390
raise errors.NoSuchRevision(self, revision_id)
392
# add what can be reached from revision_id
394
pending = set([revision_id])
395
while len(pending) > 0:
397
result[node] = entire_graph[node]
398
for revision_id in result[node]:
399
if revision_id not in result:
400
pending.add(revision_id)
404
def get_revision_graph_with_ghosts(self, revision_ids=None):
405
"""Return a graph of the revisions with ghosts marked as applicable.
407
:param revision_ids: an iterable of revisions to graph or None for all.
408
:return: a Graph object with the graph reachable from revision_ids.
412
pending = set(self.all_revision_ids())
415
pending = set(revision_ids)
416
required = set(revision_ids)
419
revision_id = pending.pop()
421
rev = self.get_revision(revision_id)
422
except errors.NoSuchRevision:
423
if revision_id in required:
426
result.add_ghost(revision_id)
428
for parent_id in rev.parent_ids:
429
# is this queued or done ?
430
if (parent_id not in pending and
431
parent_id not in done):
433
pending.add(parent_id)
434
result.add_node(revision_id, rev.parent_ids)
435
done.add(revision_id)
439
def get_revision_inventory(self, revision_id):
440
"""Return inventory of a past revision."""
441
# TODO: Unify this with get_inventory()
442
# bzr 0.0.6 and later imposes the constraint that the inventory_id
443
# must be the same as its revision, so this is trivial.
444
if revision_id is None:
445
# This does not make sense: if there is no revision,
446
# then it is the current tree inventory surely ?!
447
# and thus get_root_id() is something that looks at the last
448
# commit on the branch, and the get_root_id is an inventory check.
449
raise NotImplementedError
450
# return Inventory(self.get_root_id())
452
return self.get_inventory(revision_id)
456
"""Return True if this repository is flagged as a shared repository."""
457
raise NotImplementedError(self.is_shared)
460
def reconcile(self, other=None, thorough=False):
461
"""Reconcile this repository."""
462
from bzrlib.reconcile import RepoReconciler
463
reconciler = RepoReconciler(self, thorough=thorough)
464
reconciler.reconcile()
468
def revision_tree(self, revision_id):
469
"""Return Tree for a revision on this branch.
471
`revision_id` may be None for the null revision, in which case
472
an `EmptyTree` is returned."""
473
# TODO: refactor this to use an existing revision object
474
# so we don't need to read it in twice.
475
if revision_id is None or revision_id == NULL_REVISION:
478
inv = self.get_revision_inventory(revision_id)
479
return RevisionTree(self, inv, revision_id)
482
def get_ancestry(self, revision_id):
483
"""Return a list of revision-ids integrated by a revision.
485
This is topologically sorted.
487
if revision_id is None:
489
if not self.has_revision(revision_id):
490
raise errors.NoSuchRevision(self, revision_id)
491
w = self.get_inventory_weave()
492
candidates = w.get_ancestry(revision_id)
493
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
496
def print_file(self, file, revision_id):
497
"""Print `file` to stdout.
499
FIXME RBC 20060125 as John Meinel points out this is a bad api
500
- it writes to stdout, it assumes that that is valid etc. Fix
501
by creating a new more flexible convenience function.
503
tree = self.revision_tree(revision_id)
504
# use inventory as it was in that revision
505
file_id = tree.inventory.path2id(file)
507
# TODO: jam 20060427 Write a test for this code path
508
# it had a bug in it, and was raising the wrong
510
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
511
tree.print_file(file_id)
513
def get_transaction(self):
514
return self.control_files.get_transaction()
516
def revision_parents(self, revid):
517
return self.get_inventory_weave().parent_names(revid)
520
def set_make_working_trees(self, new_value):
521
"""Set the policy flag for making working trees when creating branches.
523
This only applies to branches that use this repository.
525
The default is 'True'.
526
:param new_value: True to restore the default, False to disable making
529
raise NotImplementedError(self.set_make_working_trees)
531
def make_working_trees(self):
532
"""Returns the policy for making working trees on new branches."""
533
raise NotImplementedError(self.make_working_trees)
536
def sign_revision(self, revision_id, gpg_strategy):
537
plaintext = Testament.from_revision(self, revision_id).as_short_text()
538
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
541
def has_signature_for_revision_id(self, revision_id):
542
"""Query for a revision signature for revision_id in the repository."""
543
return self._revision_store.has_signature(revision_id,
544
self.get_transaction())
547
def get_signature_text(self, revision_id):
548
"""Return the text for a signature."""
549
return self._revision_store.get_signature_text(revision_id,
550
self.get_transaction())
553
class AllInOneRepository(Repository):
554
"""Legacy support - the repository behaviour for all-in-one branches."""
556
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
557
# we reuse one control files instance.
558
dir_mode = a_bzrdir._control_files._dir_mode
559
file_mode = a_bzrdir._control_files._file_mode
561
def get_store(name, compressed=True, prefixed=False):
562
# FIXME: This approach of assuming stores are all entirely compressed
563
# or entirely uncompressed is tidy, but breaks upgrade from
564
# some existing branches where there's a mixture; we probably
565
# still want the option to look for both.
566
relpath = a_bzrdir._control_files._escape(name)
567
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
568
prefixed=prefixed, compressed=compressed,
571
#if self._transport.should_cache():
572
# cache_path = os.path.join(self.cache_root, name)
573
# os.mkdir(cache_path)
574
# store = bzrlib.store.CachedStore(store, cache_path)
577
# not broken out yet because the controlweaves|inventory_store
578
# and text_store | weave_store bits are still different.
579
if isinstance(_format, RepositoryFormat4):
580
# cannot remove these - there is still no consistent api
581
# which allows access to this old info.
582
self.inventory_store = get_store('inventory-store')
583
text_store = get_store('text-store')
584
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
588
"""AllInOne repositories cannot be shared."""
592
def set_make_working_trees(self, new_value):
593
"""Set the policy flag for making working trees when creating branches.
595
This only applies to branches that use this repository.
597
The default is 'True'.
598
:param new_value: True to restore the default, False to disable making
601
raise NotImplementedError(self.set_make_working_trees)
603
def make_working_trees(self):
604
"""Returns the policy for making working trees on new branches."""
608
class MetaDirRepository(Repository):
609
"""Repositories in the new meta-dir layout."""
611
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
612
super(MetaDirRepository, self).__init__(_format,
619
dir_mode = self.control_files._dir_mode
620
file_mode = self.control_files._file_mode
624
"""Return True if this repository is flagged as a shared repository."""
625
return self.control_files._transport.has('shared-storage')
628
def set_make_working_trees(self, new_value):
629
"""Set the policy flag for making working trees when creating branches.
631
This only applies to branches that use this repository.
633
The default is 'True'.
634
:param new_value: True to restore the default, False to disable making
639
self.control_files._transport.delete('no-working-trees')
640
except errors.NoSuchFile:
643
self.control_files.put_utf8('no-working-trees', '')
645
def make_working_trees(self):
646
"""Returns the policy for making working trees on new branches."""
647
return not self.control_files._transport.has('no-working-trees')
650
class KnitRepository(MetaDirRepository):
651
"""Knit format repository."""
654
def all_revision_ids(self):
655
"""See Repository.all_revision_ids()."""
656
return self._revision_store.all_revision_ids(self.get_transaction())
658
def fileid_involved_between_revs(self, from_revid, to_revid):
659
"""Find file_id(s) which are involved in the changes between revisions.
661
This determines the set of revisions which are involved, and then
662
finds all file ids affected by those revisions.
664
vf = self._get_revision_vf()
665
from_set = set(vf.get_ancestry(from_revid))
666
to_set = set(vf.get_ancestry(to_revid))
667
changed = to_set.difference(from_set)
668
return self._fileid_involved_by_set(changed)
670
def fileid_involved(self, last_revid=None):
671
"""Find all file_ids modified in the ancestry of last_revid.
673
:param last_revid: If None, last_revision() will be used.
676
changed = set(self.all_revision_ids())
678
changed = set(self.get_ancestry(last_revid))
681
return self._fileid_involved_by_set(changed)
684
def get_ancestry(self, revision_id):
685
"""Return a list of revision-ids integrated by a revision.
687
This is topologically sorted.
689
if revision_id is None:
691
vf = self._get_revision_vf()
693
return [None] + vf.get_ancestry(revision_id)
694
except errors.RevisionNotPresent:
695
raise errors.NoSuchRevision(self, revision_id)
698
def get_revision(self, revision_id):
699
"""Return the Revision object for a named revision"""
700
return self.get_revision_reconcile(revision_id)
703
def get_revision_graph(self, revision_id=None):
704
"""Return a dictionary containing the revision graph.
706
:return: a dictionary of revision_id->revision_parents_list.
708
weave = self._get_revision_vf()
709
entire_graph = weave.get_graph()
710
if revision_id is None:
711
return weave.get_graph()
712
elif revision_id not in weave:
713
raise errors.NoSuchRevision(self, revision_id)
715
# add what can be reached from revision_id
717
pending = set([revision_id])
718
while len(pending) > 0:
720
result[node] = weave.get_parents(node)
721
for revision_id in result[node]:
722
if revision_id not in result:
723
pending.add(revision_id)
727
def get_revision_graph_with_ghosts(self, revision_ids=None):
728
"""Return a graph of the revisions with ghosts marked as applicable.
730
:param revision_ids: an iterable of revisions to graph or None for all.
731
:return: a Graph object with the graph reachable from revision_ids.
734
vf = self._get_revision_vf()
735
versions = set(vf.versions())
737
pending = set(self.all_revision_ids())
740
pending = set(revision_ids)
741
required = set(revision_ids)
744
revision_id = pending.pop()
745
if not revision_id in versions:
746
if revision_id in required:
747
raise errors.NoSuchRevision(self, revision_id)
749
result.add_ghost(revision_id)
750
# mark it as done so we dont try for it again.
751
done.add(revision_id)
753
parent_ids = vf.get_parents_with_ghosts(revision_id)
754
for parent_id in parent_ids:
755
# is this queued or done ?
756
if (parent_id not in pending and
757
parent_id not in done):
759
pending.add(parent_id)
760
result.add_node(revision_id, parent_ids)
761
done.add(revision_id)
764
def _get_revision_vf(self):
765
""":return: a versioned file containing the revisions."""
766
vf = self._revision_store.get_revision_file(self.get_transaction())
770
def reconcile(self, other=None, thorough=False):
771
"""Reconcile this repository."""
772
from bzrlib.reconcile import KnitReconciler
773
reconciler = KnitReconciler(self, thorough=thorough)
774
reconciler.reconcile()
777
def revision_parents(self, revid):
778
return self._get_revision_vf().get_parents(rev_id)
780
class RepositoryFormat(object):
781
"""A repository format.
783
Formats provide three things:
784
* An initialization routine to construct repository data on disk.
785
* a format string which is used when the BzrDir supports versioned
787
* an open routine which returns a Repository instance.
789
Formats are placed in an dict by their format string for reference
790
during opening. These should be subclasses of RepositoryFormat
793
Once a format is deprecated, just deprecate the initialize and open
794
methods on the format class. Do not deprecate the object, as the
795
object will be created every system load.
797
Common instance attributes:
798
_matchingbzrdir - the bzrdir format that the repository format was
799
originally written to work with. This can be used if manually
800
constructing a bzrdir and repository, or more commonly for test suite
804
_default_format = None
805
"""The default format used for new repositories."""
808
"""The known formats."""
811
def find_format(klass, a_bzrdir):
812
"""Return the format for the repository object in a_bzrdir."""
814
transport = a_bzrdir.get_repository_transport(None)
815
format_string = transport.get("format").read()
816
return klass._formats[format_string]
817
except errors.NoSuchFile:
818
raise errors.NoRepositoryPresent(a_bzrdir)
820
raise errors.UnknownFormatError(format_string)
822
def _get_control_store(self, repo_transport, control_files):
823
"""Return the control store for this repository."""
824
raise NotImplementedError(self._get_control_store)
827
def get_default_format(klass):
828
"""Return the current default format."""
829
return klass._default_format
831
def get_format_string(self):
832
"""Return the ASCII format string that identifies this format.
834
Note that in pre format ?? repositories the format string is
835
not permitted nor written to disk.
837
raise NotImplementedError(self.get_format_string)
839
def get_format_description(self):
840
"""Return the short desciption for this format."""
841
raise NotImplementedError(self.get_format_description)
843
def _get_revision_store(self, repo_transport, control_files):
844
"""Return the revision store object for this a_bzrdir."""
845
raise NotImplementedError(self._get_revision_store)
847
def _get_text_rev_store(self,
854
"""Common logic for getting a revision store for a repository.
856
see self._get_revision_store for the subclass-overridable method to
857
get the store for a repository.
859
from bzrlib.store.revision.text import TextRevisionStore
860
dir_mode = control_files._dir_mode
861
file_mode = control_files._file_mode
862
text_store =TextStore(transport.clone(name),
864
compressed=compressed,
867
_revision_store = TextRevisionStore(text_store, serializer)
868
return _revision_store
870
def _get_versioned_file_store(self,
875
versionedfile_class=WeaveFile,
877
weave_transport = control_files._transport.clone(name)
878
dir_mode = control_files._dir_mode
879
file_mode = control_files._file_mode
880
return VersionedFileStore(weave_transport, prefixed=prefixed,
883
versionedfile_class=versionedfile_class,
886
def initialize(self, a_bzrdir, shared=False):
887
"""Initialize a repository of this format in a_bzrdir.
889
:param a_bzrdir: The bzrdir to put the new repository in it.
890
:param shared: The repository should be initialized as a sharable one.
892
This may raise UninitializableFormat if shared repository are not
893
compatible the a_bzrdir.
896
def is_supported(self):
897
"""Is this format supported?
899
Supported formats must be initializable and openable.
900
Unsupported formats may not support initialization or committing or
901
some other features depending on the reason for not being supported.
905
def open(self, a_bzrdir, _found=False):
906
"""Return an instance of this format for the bzrdir a_bzrdir.
908
_found is a private parameter, do not use it.
910
raise NotImplementedError(self.open)
913
def register_format(klass, format):
914
klass._formats[format.get_format_string()] = format
917
def set_default_format(klass, format):
918
klass._default_format = format
921
def unregister_format(klass, format):
922
assert klass._formats[format.get_format_string()] is format
923
del klass._formats[format.get_format_string()]
926
class PreSplitOutRepositoryFormat(RepositoryFormat):
927
"""Base class for the pre split out repository formats."""
929
def initialize(self, a_bzrdir, shared=False, _internal=False):
930
"""Create a weave repository.
932
TODO: when creating split out bzr branch formats, move this to a common
933
base for Format5, Format6. or something like that.
935
from bzrlib.weavefile import write_weave_v5
936
from bzrlib.weave import Weave
939
raise errors.IncompatibleFormat(self, a_bzrdir._format)
942
# always initialized when the bzrdir is.
943
return self.open(a_bzrdir, _found=True)
945
# Create an empty weave
947
bzrlib.weavefile.write_weave_v5(Weave(), sio)
948
empty_weave = sio.getvalue()
950
mutter('creating repository in %s.', a_bzrdir.transport.base)
951
dirs = ['revision-store', 'weaves']
952
files = [('inventory.weave', StringIO(empty_weave)),
955
# FIXME: RBC 20060125 dont peek under the covers
956
# NB: no need to escape relative paths that are url safe.
957
control_files = LockableFiles(a_bzrdir.transport, 'branch-lock',
959
control_files.create_lock()
960
control_files.lock_write()
961
control_files._transport.mkdir_multi(dirs,
962
mode=control_files._dir_mode)
964
for file, content in files:
965
control_files.put(file, content)
967
control_files.unlock()
968
return self.open(a_bzrdir, _found=True)
970
def _get_control_store(self, repo_transport, control_files):
971
"""Return the control store for this repository."""
972
return self._get_versioned_file_store('',
977
def _get_text_store(self, transport, control_files):
978
"""Get a store for file texts for this format."""
979
raise NotImplementedError(self._get_text_store)
981
def open(self, a_bzrdir, _found=False):
982
"""See RepositoryFormat.open()."""
984
# we are being called directly and must probe.
985
raise NotImplementedError
987
repo_transport = a_bzrdir.get_repository_transport(None)
988
control_files = a_bzrdir._control_files
989
text_store = self._get_text_store(repo_transport, control_files)
990
control_store = self._get_control_store(repo_transport, control_files)
991
_revision_store = self._get_revision_store(repo_transport, control_files)
992
return AllInOneRepository(_format=self,
994
_revision_store=_revision_store,
995
control_store=control_store,
996
text_store=text_store)
999
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1000
"""Bzr repository format 4.
1002
This repository format has:
1004
- TextStores for texts, inventories,revisions.
1006
This format is deprecated: it indexes texts using a text id which is
1007
removed in format 5; initializationa and write support for this format
1012
super(RepositoryFormat4, self).__init__()
1013
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat4()
1015
def get_format_description(self):
1016
"""See RepositoryFormat.get_format_description()."""
1017
return "Repository format 4"
1019
def initialize(self, url, shared=False, _internal=False):
1020
"""Format 4 branches cannot be created."""
1021
raise errors.UninitializableFormat(self)
1023
def is_supported(self):
1024
"""Format 4 is not supported.
1026
It is not supported because the model changed from 4 to 5 and the
1027
conversion logic is expensive - so doing it on the fly was not
1032
def _get_control_store(self, repo_transport, control_files):
1033
"""Format 4 repositories have no formal control store at this point.
1035
This will cause any control-file-needing apis to fail - this is desired.
1039
def _get_revision_store(self, repo_transport, control_files):
1040
"""See RepositoryFormat._get_revision_store()."""
1041
from bzrlib.xml4 import serializer_v4
1042
return self._get_text_rev_store(repo_transport,
1045
serializer=serializer_v4)
1047
def _get_text_store(self, transport, control_files):
1048
"""See RepositoryFormat._get_text_store()."""
1051
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1052
"""Bzr control format 5.
1054
This repository format has:
1055
- weaves for file texts and inventory
1057
- TextStores for revisions and signatures.
1061
super(RepositoryFormat5, self).__init__()
1062
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat5()
1064
def get_format_description(self):
1065
"""See RepositoryFormat.get_format_description()."""
1066
return "Weave repository format 5"
1068
def _get_revision_store(self, repo_transport, control_files):
1069
"""See RepositoryFormat._get_revision_store()."""
1070
"""Return the revision store object for this a_bzrdir."""
1071
return self._get_text_rev_store(repo_transport,
1076
def _get_text_store(self, transport, control_files):
1077
"""See RepositoryFormat._get_text_store()."""
1078
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1081
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1082
"""Bzr control format 6.
1084
This repository format has:
1085
- weaves for file texts and inventory
1086
- hash subdirectory based stores.
1087
- TextStores for revisions and signatures.
1091
super(RepositoryFormat6, self).__init__()
1092
self._matchingbzrdir = bzrlib.bzrdir.BzrDirFormat6()
1094
def get_format_description(self):
1095
"""See RepositoryFormat.get_format_description()."""
1096
return "Weave repository format 6"
1098
def _get_revision_store(self, repo_transport, control_files):
1099
"""See RepositoryFormat._get_revision_store()."""
1100
return self._get_text_rev_store(repo_transport,
1106
def _get_text_store(self, transport, control_files):
1107
"""See RepositoryFormat._get_text_store()."""
1108
return self._get_versioned_file_store('weaves', transport, control_files)
1111
class MetaDirRepositoryFormat(RepositoryFormat):
1112
"""Common base class for the new repositories using the metadir layour."""
1115
super(MetaDirRepositoryFormat, self).__init__()
1116
self._matchingbzrdir = bzrlib.bzrdir.BzrDirMetaFormat1()
1118
def _create_control_files(self, a_bzrdir):
1119
"""Create the required files and the initial control_files object."""
1120
# FIXME: RBC 20060125 dont peek under the covers
1121
# NB: no need to escape relative paths that are url safe.
1122
repository_transport = a_bzrdir.get_repository_transport(self)
1123
control_files = LockableFiles(repository_transport, 'lock', LockDir)
1124
control_files.create_lock()
1125
return control_files
1127
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1128
"""Upload the initial blank content."""
1129
control_files = self._create_control_files(a_bzrdir)
1130
control_files.lock_write()
1132
control_files._transport.mkdir_multi(dirs,
1133
mode=control_files._dir_mode)
1134
for file, content in files:
1135
control_files.put(file, content)
1136
for file, content in utf8_files:
1137
control_files.put_utf8(file, content)
1139
control_files.put_utf8('shared-storage', '')
1141
control_files.unlock()
1144
class RepositoryFormat7(MetaDirRepositoryFormat):
1145
"""Bzr repository 7.
1147
This repository format has:
1148
- weaves for file texts and inventory
1149
- hash subdirectory based stores.
1150
- TextStores for revisions and signatures.
1151
- a format marker of its own
1152
- an optional 'shared-storage' flag
1153
- an optional 'no-working-trees' flag
1156
def _get_control_store(self, repo_transport, control_files):
1157
"""Return the control store for this repository."""
1158
return self._get_versioned_file_store('',
1163
def get_format_string(self):
1164
"""See RepositoryFormat.get_format_string()."""
1165
return "Bazaar-NG Repository format 7"
1167
def get_format_description(self):
1168
"""See RepositoryFormat.get_format_description()."""
1169
return "Weave repository format 7"
1171
def _get_revision_store(self, repo_transport, control_files):
1172
"""See RepositoryFormat._get_revision_store()."""
1173
return self._get_text_rev_store(repo_transport,
1180
def _get_text_store(self, transport, control_files):
1181
"""See RepositoryFormat._get_text_store()."""
1182
return self._get_versioned_file_store('weaves',
1186
def initialize(self, a_bzrdir, shared=False):
1187
"""Create a weave repository.
1189
:param shared: If true the repository will be initialized as a shared
1192
from bzrlib.weavefile import write_weave_v5
1193
from bzrlib.weave import Weave
1195
# Create an empty weave
1197
bzrlib.weavefile.write_weave_v5(Weave(), sio)
1198
empty_weave = sio.getvalue()
1200
mutter('creating repository in %s.', a_bzrdir.transport.base)
1201
dirs = ['revision-store', 'weaves']
1202
files = [('inventory.weave', StringIO(empty_weave)),
1204
utf8_files = [('format', self.get_format_string())]
1206
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1207
return self.open(a_bzrdir=a_bzrdir, _found=True)
1209
def open(self, a_bzrdir, _found=False, _override_transport=None):
1210
"""See RepositoryFormat.open().
1212
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1213
repository at a slightly different url
1214
than normal. I.e. during 'upgrade'.
1217
format = RepositoryFormat.find_format(a_bzrdir)
1218
assert format.__class__ == self.__class__
1219
if _override_transport is not None:
1220
repo_transport = _override_transport
1222
repo_transport = a_bzrdir.get_repository_transport(None)
1223
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1224
text_store = self._get_text_store(repo_transport, control_files)
1225
control_store = self._get_control_store(repo_transport, control_files)
1226
_revision_store = self._get_revision_store(repo_transport, control_files)
1227
return MetaDirRepository(_format=self,
1229
control_files=control_files,
1230
_revision_store=_revision_store,
1231
control_store=control_store,
1232
text_store=text_store)
1235
class RepositoryFormatKnit1(MetaDirRepositoryFormat):
1236
"""Bzr repository knit format 1.
1238
This repository format has:
1239
- knits for file texts and inventory
1240
- hash subdirectory based stores.
1241
- knits for revisions and signatures
1242
- TextStores for revisions and signatures.
1243
- a format marker of its own
1244
- an optional 'shared-storage' flag
1245
- an optional 'no-working-trees' flag
1248
This format was introduced in bzr 0.8.
1251
def _get_control_store(self, repo_transport, control_files):
1252
"""Return the control store for this repository."""
1253
return VersionedFileStore(
1256
file_mode=control_files._file_mode,
1257
versionedfile_class=KnitVersionedFile,
1258
versionedfile_kwargs={'factory':KnitPlainFactory()},
1261
def get_format_string(self):
1262
"""See RepositoryFormat.get_format_string()."""
1263
return "Bazaar-NG Knit Repository Format 1"
1265
def get_format_description(self):
1266
"""See RepositoryFormat.get_format_description()."""
1267
return "Knit repository format 1"
1269
def _get_revision_store(self, repo_transport, control_files):
1270
"""See RepositoryFormat._get_revision_store()."""
1271
from bzrlib.store.revision.knit import KnitRevisionStore
1272
versioned_file_store = VersionedFileStore(
1274
file_mode=control_files._file_mode,
1277
versionedfile_class=KnitVersionedFile,
1278
versionedfile_kwargs={'delta':False, 'factory':KnitPlainFactory()},
1281
return KnitRevisionStore(versioned_file_store)
1283
def _get_text_store(self, transport, control_files):
1284
"""See RepositoryFormat._get_text_store()."""
1285
return self._get_versioned_file_store('knits',
1288
versionedfile_class=KnitVersionedFile,
1291
def initialize(self, a_bzrdir, shared=False):
1292
"""Create a knit format 1 repository.
1294
:param a_bzrdir: bzrdir to contain the new repository; must already
1296
:param shared: If true the repository will be initialized as a shared
1299
mutter('creating repository in %s.', a_bzrdir.transport.base)
1300
dirs = ['revision-store', 'knits']
1302
utf8_files = [('format', self.get_format_string())]
1304
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1305
repo_transport = a_bzrdir.get_repository_transport(None)
1306
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1307
control_store = self._get_control_store(repo_transport, control_files)
1308
transaction = bzrlib.transactions.WriteTransaction()
1309
# trigger a write of the inventory store.
1310
control_store.get_weave_or_empty('inventory', transaction)
1311
_revision_store = self._get_revision_store(repo_transport, control_files)
1312
_revision_store.has_revision_id('A', transaction)
1313
_revision_store.get_signature_file(transaction)
1314
return self.open(a_bzrdir=a_bzrdir, _found=True)
1316
def open(self, a_bzrdir, _found=False, _override_transport=None):
1317
"""See RepositoryFormat.open().
1319
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1320
repository at a slightly different url
1321
than normal. I.e. during 'upgrade'.
1324
format = RepositoryFormat.find_format(a_bzrdir)
1325
assert format.__class__ == self.__class__
1326
if _override_transport is not None:
1327
repo_transport = _override_transport
1329
repo_transport = a_bzrdir.get_repository_transport(None)
1330
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1331
text_store = self._get_text_store(repo_transport, control_files)
1332
control_store = self._get_control_store(repo_transport, control_files)
1333
_revision_store = self._get_revision_store(repo_transport, control_files)
1334
return KnitRepository(_format=self,
1336
control_files=control_files,
1337
_revision_store=_revision_store,
1338
control_store=control_store,
1339
text_store=text_store)
1342
# formats which have no format string are not discoverable
1343
# and not independently creatable, so are not registered.
1344
RepositoryFormat.register_format(RepositoryFormat7())
1345
_default_format = RepositoryFormatKnit1()
1346
RepositoryFormat.register_format(_default_format)
1347
RepositoryFormat.set_default_format(_default_format)
1348
_legacy_formats = [RepositoryFormat4(),
1349
RepositoryFormat5(),
1350
RepositoryFormat6()]
1353
class InterRepository(InterObject):
1354
"""This class represents operations taking place between two repositories.
1356
Its instances have methods like copy_content and fetch, and contain
1357
references to the source and target repositories these operations can be
1360
Often we will provide convenience methods on 'repository' which carry out
1361
operations with another repository - they will always forward to
1362
InterRepository.get(other).method_name(parameters).
1366
"""The available optimised InterRepository types."""
1369
def copy_content(self, revision_id=None, basis=None):
1370
"""Make a complete copy of the content in self into destination.
1372
This is a destructive operation! Do not use it on existing
1375
:param revision_id: Only copy the content needed to construct
1376
revision_id and its parents.
1377
:param basis: Copy the needed data preferentially from basis.
1380
self.target.set_make_working_trees(self.source.make_working_trees())
1381
except NotImplementedError:
1383
# grab the basis available data
1384
if basis is not None:
1385
self.target.fetch(basis, revision_id=revision_id)
1386
# but dont bother fetching if we have the needed data now.
1387
if (revision_id not in (None, NULL_REVISION) and
1388
self.target.has_revision(revision_id)):
1390
self.target.fetch(self.source, revision_id=revision_id)
1392
def _double_lock(self, lock_source, lock_target):
1393
"""Take out too locks, rolling back the first if the second throws."""
1398
# we want to ensure that we don't leave source locked by mistake.
1399
# and any error on target should not confuse source.
1400
self.source.unlock()
1404
def fetch(self, revision_id=None, pb=None):
1405
"""Fetch the content required to construct revision_id.
1407
The content is copied from source to target.
1409
:param revision_id: if None all content is copied, if NULL_REVISION no
1411
:param pb: optional progress bar to use for progress reports. If not
1412
provided a default one will be created.
1414
Returns the copied revision count and the failed revisions in a tuple:
1417
from bzrlib.fetch import GenericRepoFetcher
1418
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1419
self.source, self.source._format, self.target, self.target._format)
1420
f = GenericRepoFetcher(to_repository=self.target,
1421
from_repository=self.source,
1422
last_revision=revision_id,
1424
return f.count_copied, f.failed_revisions
1426
def lock_read(self):
1427
"""Take out a logical read lock.
1429
This will lock the source branch and the target branch. The source gets
1430
a read lock and the target a read lock.
1432
self._double_lock(self.source.lock_read, self.target.lock_read)
1434
def lock_write(self):
1435
"""Take out a logical write lock.
1437
This will lock the source branch and the target branch. The source gets
1438
a read lock and the target a write lock.
1440
self._double_lock(self.source.lock_read, self.target.lock_write)
1443
def missing_revision_ids(self, revision_id=None):
1444
"""Return the revision ids that source has that target does not.
1446
These are returned in topological order.
1448
:param revision_id: only return revision ids included by this
1451
# generic, possibly worst case, slow code path.
1452
target_ids = set(self.target.all_revision_ids())
1453
if revision_id is not None:
1454
source_ids = self.source.get_ancestry(revision_id)
1455
assert source_ids.pop(0) == None
1457
source_ids = self.source.all_revision_ids()
1458
result_set = set(source_ids).difference(target_ids)
1459
# this may look like a no-op: its not. It preserves the ordering
1460
# other_ids had while only returning the members from other_ids
1461
# that we've decided we need.
1462
return [rev_id for rev_id in source_ids if rev_id in result_set]
1465
"""Release the locks on source and target."""
1467
self.target.unlock()
1469
self.source.unlock()
1472
class InterWeaveRepo(InterRepository):
1473
"""Optimised code paths between Weave based repositories."""
1475
_matching_repo_format = RepositoryFormat7()
1476
"""Repository format for testing with."""
1479
def is_compatible(source, target):
1480
"""Be compatible with known Weave formats.
1482
We dont test for the stores being of specific types becase that
1483
could lead to confusing results, and there is no need to be
1487
return (isinstance(source._format, (RepositoryFormat5,
1489
RepositoryFormat7)) and
1490
isinstance(target._format, (RepositoryFormat5,
1492
RepositoryFormat7)))
1493
except AttributeError:
1497
def copy_content(self, revision_id=None, basis=None):
1498
"""See InterRepository.copy_content()."""
1499
# weave specific optimised path:
1500
if basis is not None:
1501
# copy the basis in, then fetch remaining data.
1502
basis.copy_content_into(self.target, revision_id)
1503
# the basis copy_content_into could misset this.
1505
self.target.set_make_working_trees(self.source.make_working_trees())
1506
except NotImplementedError:
1508
self.target.fetch(self.source, revision_id=revision_id)
1511
self.target.set_make_working_trees(self.source.make_working_trees())
1512
except NotImplementedError:
1514
# FIXME do not peek!
1515
if self.source.control_files._transport.listable():
1516
pb = bzrlib.ui.ui_factory.nested_progress_bar()
1518
self.target.weave_store.copy_all_ids(
1519
self.source.weave_store,
1521
from_transaction=self.source.get_transaction(),
1522
to_transaction=self.target.get_transaction())
1523
pb.update('copying inventory', 0, 1)
1524
self.target.control_weaves.copy_multi(
1525
self.source.control_weaves, ['inventory'],
1526
from_transaction=self.source.get_transaction(),
1527
to_transaction=self.target.get_transaction())
1528
self.target._revision_store.text_store.copy_all_ids(
1529
self.source._revision_store.text_store,
1534
self.target.fetch(self.source, revision_id=revision_id)
1537
def fetch(self, revision_id=None, pb=None):
1538
"""See InterRepository.fetch()."""
1539
from bzrlib.fetch import GenericRepoFetcher
1540
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1541
self.source, self.source._format, self.target, self.target._format)
1542
f = GenericRepoFetcher(to_repository=self.target,
1543
from_repository=self.source,
1544
last_revision=revision_id,
1546
return f.count_copied, f.failed_revisions
1549
def missing_revision_ids(self, revision_id=None):
1550
"""See InterRepository.missing_revision_ids()."""
1551
# we want all revisions to satisfy revision_id in source.
1552
# but we dont want to stat every file here and there.
1553
# we want then, all revisions other needs to satisfy revision_id
1554
# checked, but not those that we have locally.
1555
# so the first thing is to get a subset of the revisions to
1556
# satisfy revision_id in source, and then eliminate those that
1557
# we do already have.
1558
# this is slow on high latency connection to self, but as as this
1559
# disk format scales terribly for push anyway due to rewriting
1560
# inventory.weave, this is considered acceptable.
1562
if revision_id is not None:
1563
source_ids = self.source.get_ancestry(revision_id)
1564
assert source_ids.pop(0) == None
1566
source_ids = self.source._all_possible_ids()
1567
source_ids_set = set(source_ids)
1568
# source_ids is the worst possible case we may need to pull.
1569
# now we want to filter source_ids against what we actually
1570
# have in target, but dont try to check for existence where we know
1571
# we do not have a revision as that would be pointless.
1572
target_ids = set(self.target._all_possible_ids())
1573
possibly_present_revisions = target_ids.intersection(source_ids_set)
1574
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1575
required_revisions = source_ids_set.difference(actually_present_revisions)
1576
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1577
if revision_id is not None:
1578
# we used get_ancestry to determine source_ids then we are assured all
1579
# revisions referenced are present as they are installed in topological order.
1580
# and the tip revision was validated by get_ancestry.
1581
return required_topo_revisions
1583
# if we just grabbed the possibly available ids, then
1584
# we only have an estimate of whats available and need to validate
1585
# that against the revision records.
1586
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1589
class InterKnitRepo(InterRepository):
1590
"""Optimised code paths between Knit based repositories."""
1592
_matching_repo_format = RepositoryFormatKnit1()
1593
"""Repository format for testing with."""
1596
def is_compatible(source, target):
1597
"""Be compatible with known Knit formats.
1599
We dont test for the stores being of specific types becase that
1600
could lead to confusing results, and there is no need to be
1604
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1605
isinstance(target._format, (RepositoryFormatKnit1)))
1606
except AttributeError:
1610
def fetch(self, revision_id=None, pb=None):
1611
"""See InterRepository.fetch()."""
1612
from bzrlib.fetch import KnitRepoFetcher
1613
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1614
self.source, self.source._format, self.target, self.target._format)
1615
f = KnitRepoFetcher(to_repository=self.target,
1616
from_repository=self.source,
1617
last_revision=revision_id,
1619
return f.count_copied, f.failed_revisions
1622
def missing_revision_ids(self, revision_id=None):
1623
"""See InterRepository.missing_revision_ids()."""
1624
if revision_id is not None:
1625
source_ids = self.source.get_ancestry(revision_id)
1626
assert source_ids.pop(0) == None
1628
source_ids = self.source._all_possible_ids()
1629
source_ids_set = set(source_ids)
1630
# source_ids is the worst possible case we may need to pull.
1631
# now we want to filter source_ids against what we actually
1632
# have in target, but dont try to check for existence where we know
1633
# we do not have a revision as that would be pointless.
1634
target_ids = set(self.target._all_possible_ids())
1635
possibly_present_revisions = target_ids.intersection(source_ids_set)
1636
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1637
required_revisions = source_ids_set.difference(actually_present_revisions)
1638
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1639
if revision_id is not None:
1640
# we used get_ancestry to determine source_ids then we are assured all
1641
# revisions referenced are present as they are installed in topological order.
1642
# and the tip revision was validated by get_ancestry.
1643
return required_topo_revisions
1645
# if we just grabbed the possibly available ids, then
1646
# we only have an estimate of whats available and need to validate
1647
# that against the revision records.
1648
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1650
InterRepository.register_optimiser(InterWeaveRepo)
1651
InterRepository.register_optimiser(InterKnitRepo)
1654
class RepositoryTestProviderAdapter(object):
1655
"""A tool to generate a suite testing multiple repository formats at once.
1657
This is done by copying the test once for each transport and injecting
1658
the transport_server, transport_readonly_server, and bzrdir_format and
1659
repository_format classes into each copy. Each copy is also given a new id()
1660
to make it easy to identify.
1663
def __init__(self, transport_server, transport_readonly_server, formats):
1664
self._transport_server = transport_server
1665
self._transport_readonly_server = transport_readonly_server
1666
self._formats = formats
1668
def adapt(self, test):
1669
result = TestSuite()
1670
for repository_format, bzrdir_format in self._formats:
1671
new_test = deepcopy(test)
1672
new_test.transport_server = self._transport_server
1673
new_test.transport_readonly_server = self._transport_readonly_server
1674
new_test.bzrdir_format = bzrdir_format
1675
new_test.repository_format = repository_format
1676
def make_new_test_id():
1677
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1678
return lambda: new_id
1679
new_test.id = make_new_test_id()
1680
result.addTest(new_test)
1684
class InterRepositoryTestProviderAdapter(object):
1685
"""A tool to generate a suite testing multiple inter repository formats.
1687
This is done by copying the test once for each interrepo provider and injecting
1688
the transport_server, transport_readonly_server, repository_format and
1689
repository_to_format classes into each copy.
1690
Each copy is also given a new id() to make it easy to identify.
1693
def __init__(self, transport_server, transport_readonly_server, formats):
1694
self._transport_server = transport_server
1695
self._transport_readonly_server = transport_readonly_server
1696
self._formats = formats
1698
def adapt(self, test):
1699
result = TestSuite()
1700
for interrepo_class, repository_format, repository_format_to in self._formats:
1701
new_test = deepcopy(test)
1702
new_test.transport_server = self._transport_server
1703
new_test.transport_readonly_server = self._transport_readonly_server
1704
new_test.interrepo_class = interrepo_class
1705
new_test.repository_format = repository_format
1706
new_test.repository_format_to = repository_format_to
1707
def make_new_test_id():
1708
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1709
return lambda: new_id
1710
new_test.id = make_new_test_id()
1711
result.addTest(new_test)
1715
def default_test_list():
1716
"""Generate the default list of interrepo permutations to test."""
1718
# test the default InterRepository between format 6 and the current
1720
# XXX: robertc 20060220 reinstate this when there are two supported
1721
# formats which do not have an optimal code path between them.
1722
result.append((InterRepository,
1723
RepositoryFormat6(),
1724
RepositoryFormatKnit1()))
1725
for optimiser in InterRepository._optimisers:
1726
result.append((optimiser,
1727
optimiser._matching_repo_format,
1728
optimiser._matching_repo_format
1730
# if there are specific combinations we want to use, we can add them
1735
class CopyConverter(object):
1736
"""A repository conversion tool which just performs a copy of the content.
1738
This is slow but quite reliable.
1741
def __init__(self, target_format):
1742
"""Create a CopyConverter.
1744
:param target_format: The format the resulting repository should be.
1746
self.target_format = target_format
1748
def convert(self, repo, pb):
1749
"""Perform the conversion of to_convert, giving feedback via pb.
1751
:param to_convert: The disk object to convert.
1752
:param pb: a progress bar to use for progress information.
1757
# this is only useful with metadir layouts - separated repo content.
1758
# trigger an assertion if not such
1759
repo._format.get_format_string()
1760
self.repo_dir = repo.bzrdir
1761
self.step('Moving repository to repository.backup')
1762
self.repo_dir.transport.move('repository', 'repository.backup')
1763
backup_transport = self.repo_dir.transport.clone('repository.backup')
1764
self.source_repo = repo._format.open(self.repo_dir,
1766
_override_transport=backup_transport)
1767
self.step('Creating new repository')
1768
converted = self.target_format.initialize(self.repo_dir,
1769
self.source_repo.is_shared())
1770
converted.lock_write()
1772
self.step('Copying content into repository.')
1773
self.source_repo.copy_content_into(converted)
1776
self.step('Deleting old repository content.')
1777
self.repo_dir.transport.delete_tree('repository.backup')
1778
self.pb.note('repository converted')
1780
def step(self, message):
1781
"""Update the pb by a step."""
1783
self.pb.update(message, self.count, self.total)
1786
# Copied from xml.sax.saxutils
1787
def _unescape_xml(data):
1788
"""Unescape &, <, and > in a string of data.
1790
data = data.replace("<", "<")
1791
data = data.replace(">", ">")
1792
# must do ampersand last
1793
return data.replace("&", "&")