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 binascii import hexlify
18
from copy import deepcopy
19
from cStringIO import StringIO
22
from unittest import TestSuite
24
from bzrlib import bzrdir, check, delta, gpg, errors, xml5, ui, transactions, osutils
25
from bzrlib.decorators import needs_read_lock, needs_write_lock
26
from bzrlib.errors import InvalidRevisionId
27
from bzrlib.graph import Graph
28
from bzrlib.inter import InterObject
29
from bzrlib.inventory import Inventory
30
from bzrlib.knit import KnitVersionedFile, KnitPlainFactory
31
from bzrlib.lockable_files import LockableFiles, TransportLock
32
from bzrlib.lockdir import LockDir
33
from bzrlib.osutils import (safe_unicode, rand_bytes, compact_date,
35
from bzrlib.revision import NULL_REVISION, Revision
36
from bzrlib.revisiontree import RevisionTree
37
from bzrlib.store.versioned import VersionedFileStore, WeaveStore
38
from bzrlib.store.text import TextStore
39
from bzrlib.symbol_versioning import (deprecated_method,
42
from bzrlib.testament import Testament
43
from bzrlib.trace import mutter, note, warning
44
from bzrlib.tsort import topo_sort
45
from bzrlib.weave import WeaveFile
48
# Old formats display a warning, but only once
49
_deprecation_warning_done = False
52
class Repository(object):
53
"""Repository holding history for one or more branches.
55
The repository holds and retrieves historical information including
56
revisions and file history. It's normally accessed only by the Branch,
57
which views a particular line of development through that history.
59
The Repository builds on top of Stores and a Transport, which respectively
60
describe the disk data format and the way of accessing the (possibly
65
def add_inventory(self, revid, inv, parents):
66
"""Add the inventory inv to the repository as revid.
68
:param parents: The revision ids of the parents that revid
69
is known to have and are in the repository already.
71
returns the sha1 of the serialized inventory.
73
assert inv.revision_id is None or inv.revision_id == revid, \
74
"Mismatch between inventory revision" \
75
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
76
inv_text = xml5.serializer_v5.write_inventory_to_string(inv)
77
inv_sha1 = osutils.sha_string(inv_text)
78
inv_vf = self.control_weaves.get_weave('inventory',
79
self.get_transaction())
80
self._inventory_add_lines(inv_vf, revid, parents, osutils.split_lines(inv_text))
83
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
85
for parent in parents:
87
final_parents.append(parent)
89
inv_vf.add_lines(revid, final_parents, lines)
92
def add_revision(self, rev_id, rev, inv=None, config=None):
93
"""Add rev to the revision store as rev_id.
95
:param rev_id: the revision id to use.
96
:param rev: The revision object.
97
:param inv: The inventory for the revision. if None, it will be looked
98
up in the inventory storer
99
:param config: If None no digital signature will be created.
100
If supplied its signature_needed method will be used
101
to determine if a signature should be made.
103
if config is not None and config.signature_needed():
105
inv = self.get_inventory(rev_id)
106
plaintext = Testament(rev, inv).as_short_text()
107
self.store_revision_signature(
108
gpg.GPGStrategy(config), plaintext, rev_id)
109
if not rev_id in self.get_inventory_weave():
111
raise errors.WeaveRevisionNotPresent(rev_id,
112
self.get_inventory_weave())
114
# yes, this is not suitable for adding with ghosts.
115
self.add_inventory(rev_id, inv, rev.parent_ids)
116
self._revision_store.add_revision(rev, self.get_transaction())
119
def _all_possible_ids(self):
120
"""Return all the possible revisions that we could find."""
121
return self.get_inventory_weave().versions()
123
def all_revision_ids(self):
124
"""Returns a list of all the revision ids in the repository.
126
This is deprecated because code should generally work on the graph
127
reachable from a particular revision, and ignore any other revisions
128
that might be present. There is no direct replacement method.
130
return self._all_revision_ids()
133
def _all_revision_ids(self):
134
"""Returns a list of all the revision ids in the repository.
136
These are in as much topological order as the underlying store can
137
present: for weaves ghosts may lead to a lack of correctness until
138
the reweave updates the parents list.
140
if self._revision_store.text_store.listable():
141
return self._revision_store.all_revision_ids(self.get_transaction())
142
result = self._all_possible_ids()
143
return self._eliminate_revisions_not_present(result)
145
def break_lock(self):
146
"""Break a lock if one is present from another instance.
148
Uses the ui factory to ask for confirmation if the lock may be from
151
self.control_files.break_lock()
154
def _eliminate_revisions_not_present(self, revision_ids):
155
"""Check every revision id in revision_ids to see if we have it.
157
Returns a set of the present revisions.
160
for id in revision_ids:
161
if self.has_revision(id):
166
def create(a_bzrdir):
167
"""Construct the current default format repository in a_bzrdir."""
168
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
170
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
171
"""instantiate a Repository.
173
:param _format: The format of the repository on disk.
174
:param a_bzrdir: The BzrDir of the repository.
176
In the future we will have a single api for all stores for
177
getting file texts, inventories and revisions, then
178
this construct will accept instances of those things.
180
super(Repository, self).__init__()
181
self._format = _format
182
# the following are part of the public API for Repository:
183
self.bzrdir = a_bzrdir
184
self.control_files = control_files
185
self._revision_store = _revision_store
186
self.text_store = text_store
187
# backwards compatibility
188
self.weave_store = text_store
189
# not right yet - should be more semantically clear ?
191
self.control_store = control_store
192
self.control_weaves = control_store
193
# TODO: make sure to construct the right store classes, etc, depending
194
# on whether escaping is required.
195
self._warn_if_deprecated()
198
return '%s(%r)' % (self.__class__.__name__,
199
self.bzrdir.transport.base)
202
return self.control_files.is_locked()
204
def lock_write(self):
205
self.control_files.lock_write()
208
self.control_files.lock_read()
210
def get_physical_lock_status(self):
211
return self.control_files.get_physical_lock_status()
214
def missing_revision_ids(self, other, revision_id=None):
215
"""Return the revision ids that other has that this does not.
217
These are returned in topological order.
219
revision_id: only return revision ids included by revision_id.
221
return InterRepository.get(other, self).missing_revision_ids(revision_id)
225
"""Open the repository rooted at base.
227
For instance, if the repository is at URL/.bzr/repository,
228
Repository.open(URL) -> a Repository instance.
230
control = bzrdir.BzrDir.open(base)
231
return control.open_repository()
233
def copy_content_into(self, destination, revision_id=None, basis=None):
234
"""Make a complete copy of the content in self into destination.
236
This is a destructive operation! Do not use it on existing
239
return InterRepository.get(self, destination).copy_content(revision_id, basis)
241
def fetch(self, source, revision_id=None, pb=None):
242
"""Fetch the content required to construct revision_id from source.
244
If revision_id is None all content is copied.
246
return InterRepository.get(source, self).fetch(revision_id=revision_id,
249
def get_commit_builder(self, branch, parents, config, timestamp=None,
250
timezone=None, committer=None, revprops=None,
252
"""Obtain a CommitBuilder for this repository.
254
:param branch: Branch to commit to.
255
:param parents: Revision ids of the parents of the new revision.
256
:param config: Configuration to use.
257
:param timestamp: Optional timestamp recorded for commit.
258
:param timezone: Optional timezone for timestamp.
259
:param committer: Optional committer to set for commit.
260
:param revprops: Optional dictionary of revision properties.
261
:param revision_id: Optional revision id.
263
return CommitBuilder(self, parents, config, timestamp, timezone,
264
committer, revprops, revision_id)
267
self.control_files.unlock()
270
def clone(self, a_bzrdir, revision_id=None, basis=None):
271
"""Clone this repository into a_bzrdir using the current format.
273
Currently no check is made that the format of this repository and
274
the bzrdir format are compatible. FIXME RBC 20060201.
276
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
277
# use target default format.
278
result = a_bzrdir.create_repository()
279
# FIXME RBC 20060209 split out the repository type to avoid this check ?
280
elif isinstance(a_bzrdir._format,
281
(bzrdir.BzrDirFormat4,
282
bzrdir.BzrDirFormat5,
283
bzrdir.BzrDirFormat6)):
284
result = a_bzrdir.open_repository()
286
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
287
self.copy_content_into(result, revision_id, basis)
291
def has_revision(self, revision_id):
292
"""True if this repository has a copy of the revision."""
293
return self._revision_store.has_revision_id(revision_id,
294
self.get_transaction())
297
def get_revision_reconcile(self, revision_id):
298
"""'reconcile' helper routine that allows access to a revision always.
300
This variant of get_revision does not cross check the weave graph
301
against the revision one as get_revision does: but it should only
302
be used by reconcile, or reconcile-alike commands that are correcting
303
or testing the revision graph.
305
if not revision_id or not isinstance(revision_id, basestring):
306
raise InvalidRevisionId(revision_id=revision_id, branch=self)
307
return self._revision_store.get_revisions([revision_id],
308
self.get_transaction())[0]
310
def get_revisions(self, revision_ids):
311
return self._revision_store.get_revisions(revision_ids,
312
self.get_transaction())
315
def get_revision_xml(self, revision_id):
316
rev = self.get_revision(revision_id)
318
# the current serializer..
319
self._revision_store._serializer.write_revision(rev, rev_tmp)
321
return rev_tmp.getvalue()
324
def get_revision(self, revision_id):
325
"""Return the Revision object for a named revision"""
326
r = self.get_revision_reconcile(revision_id)
327
# weave corruption can lead to absent revision markers that should be
329
# the following test is reasonably cheap (it needs a single weave read)
330
# and the weave is cached in read transactions. In write transactions
331
# it is not cached but typically we only read a small number of
332
# revisions. For knits when they are introduced we will probably want
333
# to ensure that caching write transactions are in use.
334
inv = self.get_inventory_weave()
335
self._check_revision_parents(r, inv)
339
def get_deltas_for_revisions(self, revisions):
340
"""Produce a generator of revision deltas.
342
Note that the input is a sequence of REVISIONS, not revision_ids.
343
Trees will be held in memory until the generator exits.
344
Each delta is relative to the revision's lefthand predecessor.
346
required_trees = set()
347
for revision in revisions:
348
required_trees.add(revision.revision_id)
349
required_trees.update(revision.parent_ids[:1])
350
trees = dict((t.get_revision_id(), t) for
351
t in self.revision_trees(required_trees))
352
for revision in revisions:
353
if not revision.parent_ids:
354
old_tree = self.revision_tree(None)
356
old_tree = trees[revision.parent_ids[0]]
357
yield trees[revision.revision_id].changes_from(old_tree)
360
def get_revision_delta(self, revision_id):
361
"""Return the delta for one revision.
363
The delta is relative to the left-hand predecessor of the
366
r = self.get_revision(revision_id)
367
return list(self.get_deltas_for_revisions([r]))[0]
369
def _check_revision_parents(self, revision, inventory):
370
"""Private to Repository and Fetch.
372
This checks the parentage of revision in an inventory weave for
373
consistency and is only applicable to inventory-weave-for-ancestry
374
using repository formats & fetchers.
376
weave_parents = inventory.get_parents(revision.revision_id)
377
weave_names = inventory.versions()
378
for parent_id in revision.parent_ids:
379
if parent_id in weave_names:
380
# this parent must not be a ghost.
381
if not parent_id in weave_parents:
383
raise errors.CorruptRepository(self)
386
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
387
signature = gpg_strategy.sign(plaintext)
388
self._revision_store.add_revision_signature_text(revision_id,
390
self.get_transaction())
392
def fileids_altered_by_revision_ids(self, revision_ids):
393
"""Find the file ids and versions affected by revisions.
395
:param revisions: an iterable containing revision ids.
396
:return: a dictionary mapping altered file-ids to an iterable of
397
revision_ids. Each altered file-ids has the exact revision_ids that
398
altered it listed explicitly.
400
assert isinstance(self._format, (RepositoryFormat5,
403
RepositoryFormatKnit1)), \
404
("fileids_altered_by_revision_ids only supported for branches "
405
"which store inventory as unnested xml, not on %r" % self)
406
selected_revision_ids = set(revision_ids)
407
w = self.get_inventory_weave()
410
# this code needs to read every new line in every inventory for the
411
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
412
# not present in one of those inventories is unnecessary but not
413
# harmful because we are filtering by the revision id marker in the
414
# inventory lines : we only select file ids altered in one of those
415
# revisions. We don't need to see all lines in the inventory because
416
# only those added in an inventory in rev X can contain a revision=X
418
for line in w.iter_lines_added_or_present_in_versions(selected_revision_ids):
419
start = line.find('file_id="')+9
420
if start < 9: continue
421
end = line.find('"', start)
423
file_id = _unescape_xml(line[start:end])
425
start = line.find('revision="')+10
426
if start < 10: continue
427
end = line.find('"', start)
429
revision_id = _unescape_xml(line[start:end])
430
if revision_id in selected_revision_ids:
431
result.setdefault(file_id, set()).add(revision_id)
435
def get_inventory_weave(self):
436
return self.control_weaves.get_weave('inventory',
437
self.get_transaction())
440
def get_inventory(self, revision_id):
441
"""Get Inventory object by hash."""
442
return self.deserialise_inventory(
443
revision_id, self.get_inventory_xml(revision_id))
445
def deserialise_inventory(self, revision_id, xml):
446
"""Transform the xml into an inventory object.
448
:param revision_id: The expected revision id of the inventory.
449
:param xml: A serialised inventory.
451
return xml5.serializer_v5.read_inventory_from_string(xml)
454
def get_inventory_xml(self, revision_id):
455
"""Get inventory XML as a file object."""
457
assert isinstance(revision_id, basestring), type(revision_id)
458
iw = self.get_inventory_weave()
459
return iw.get_text(revision_id)
461
raise errors.HistoryMissing(self, 'inventory', revision_id)
464
def get_inventory_sha1(self, revision_id):
465
"""Return the sha1 hash of the inventory entry
467
return self.get_revision(revision_id).inventory_sha1
470
def get_revision_graph(self, revision_id=None):
471
"""Return a dictionary containing the revision graph.
473
:param revision_id: The revision_id to get a graph from. If None, then
474
the entire revision graph is returned. This is a deprecated mode of
475
operation and will be removed in the future.
476
:return: a dictionary of revision_id->revision_parents_list.
478
# special case NULL_REVISION
479
if revision_id == NULL_REVISION:
481
weave = self.get_inventory_weave()
482
all_revisions = self._eliminate_revisions_not_present(weave.versions())
483
entire_graph = dict([(node, weave.get_parents(node)) for
484
node in all_revisions])
485
if revision_id is None:
487
elif revision_id not in entire_graph:
488
raise errors.NoSuchRevision(self, revision_id)
490
# add what can be reached from revision_id
492
pending = set([revision_id])
493
while len(pending) > 0:
495
result[node] = entire_graph[node]
496
for revision_id in result[node]:
497
if revision_id not in result:
498
pending.add(revision_id)
502
def get_revision_graph_with_ghosts(self, revision_ids=None):
503
"""Return a graph of the revisions with ghosts marked as applicable.
505
:param revision_ids: an iterable of revisions to graph or None for all.
506
:return: a Graph object with the graph reachable from revision_ids.
510
pending = set(self.all_revision_ids())
513
pending = set(revision_ids)
514
# special case NULL_REVISION
515
if NULL_REVISION in pending:
516
pending.remove(NULL_REVISION)
517
required = set(pending)
520
revision_id = pending.pop()
522
rev = self.get_revision(revision_id)
523
except errors.NoSuchRevision:
524
if revision_id in required:
527
result.add_ghost(revision_id)
529
for parent_id in rev.parent_ids:
530
# is this queued or done ?
531
if (parent_id not in pending and
532
parent_id not in done):
534
pending.add(parent_id)
535
result.add_node(revision_id, rev.parent_ids)
536
done.add(revision_id)
540
def get_revision_inventory(self, revision_id):
541
"""Return inventory of a past revision."""
542
# TODO: Unify this with get_inventory()
543
# bzr 0.0.6 and later imposes the constraint that the inventory_id
544
# must be the same as its revision, so this is trivial.
545
if revision_id is None:
546
# This does not make sense: if there is no revision,
547
# then it is the current tree inventory surely ?!
548
# and thus get_root_id() is something that looks at the last
549
# commit on the branch, and the get_root_id is an inventory check.
550
raise NotImplementedError
551
# return Inventory(self.get_root_id())
553
return self.get_inventory(revision_id)
557
"""Return True if this repository is flagged as a shared repository."""
558
raise NotImplementedError(self.is_shared)
561
def reconcile(self, other=None, thorough=False):
562
"""Reconcile this repository."""
563
from bzrlib.reconcile import RepoReconciler
564
reconciler = RepoReconciler(self, thorough=thorough)
565
reconciler.reconcile()
569
def revision_tree(self, revision_id):
570
"""Return Tree for a revision on this branch.
572
`revision_id` may be None for the empty tree revision.
574
# TODO: refactor this to use an existing revision object
575
# so we don't need to read it in twice.
576
if revision_id is None or revision_id == NULL_REVISION:
577
return RevisionTree(self, Inventory(), NULL_REVISION)
579
inv = self.get_revision_inventory(revision_id)
580
return RevisionTree(self, inv, revision_id)
583
def revision_trees(self, revision_ids):
584
"""Return Tree for a revision on this branch.
586
`revision_id` may not be None or 'null:'"""
587
assert None not in revision_ids
588
assert NULL_REVISION not in revision_ids
589
texts = self.get_inventory_weave().get_texts(revision_ids)
590
for text, revision_id in zip(texts, revision_ids):
591
inv = self.deserialise_inventory(revision_id, text)
592
yield RevisionTree(self, inv, revision_id)
595
def get_ancestry(self, revision_id):
596
"""Return a list of revision-ids integrated by a revision.
598
The first element of the list is always None, indicating the origin
599
revision. This might change when we have history horizons, or
600
perhaps we should have a new API.
602
This is topologically sorted.
604
if revision_id is None:
606
if not self.has_revision(revision_id):
607
raise errors.NoSuchRevision(self, revision_id)
608
w = self.get_inventory_weave()
609
candidates = w.get_ancestry(revision_id)
610
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
613
def print_file(self, file, revision_id):
614
"""Print `file` to stdout.
616
FIXME RBC 20060125 as John Meinel points out this is a bad api
617
- it writes to stdout, it assumes that that is valid etc. Fix
618
by creating a new more flexible convenience function.
620
tree = self.revision_tree(revision_id)
621
# use inventory as it was in that revision
622
file_id = tree.inventory.path2id(file)
624
# TODO: jam 20060427 Write a test for this code path
625
# it had a bug in it, and was raising the wrong
627
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
628
tree.print_file(file_id)
630
def get_transaction(self):
631
return self.control_files.get_transaction()
633
def revision_parents(self, revid):
634
return self.get_inventory_weave().parent_names(revid)
637
def set_make_working_trees(self, new_value):
638
"""Set the policy flag for making working trees when creating branches.
640
This only applies to branches that use this repository.
642
The default is 'True'.
643
:param new_value: True to restore the default, False to disable making
646
raise NotImplementedError(self.set_make_working_trees)
648
def make_working_trees(self):
649
"""Returns the policy for making working trees on new branches."""
650
raise NotImplementedError(self.make_working_trees)
653
def sign_revision(self, revision_id, gpg_strategy):
654
plaintext = Testament.from_revision(self, revision_id).as_short_text()
655
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
658
def has_signature_for_revision_id(self, revision_id):
659
"""Query for a revision signature for revision_id in the repository."""
660
return self._revision_store.has_signature(revision_id,
661
self.get_transaction())
664
def get_signature_text(self, revision_id):
665
"""Return the text for a signature."""
666
return self._revision_store.get_signature_text(revision_id,
667
self.get_transaction())
670
def check(self, revision_ids):
671
"""Check consistency of all history of given revision_ids.
673
Different repository implementations should override _check().
675
:param revision_ids: A non-empty list of revision_ids whose ancestry
676
will be checked. Typically the last revision_id of a branch.
679
raise ValueError("revision_ids must be non-empty in %s.check"
681
return self._check(revision_ids)
683
def _check(self, revision_ids):
684
result = check.Check(self)
688
def _warn_if_deprecated(self):
689
global _deprecation_warning_done
690
if _deprecation_warning_done:
692
_deprecation_warning_done = True
693
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
694
% (self._format, self.bzrdir.transport.base))
697
class AllInOneRepository(Repository):
698
"""Legacy support - the repository behaviour for all-in-one branches."""
700
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
701
# we reuse one control files instance.
702
dir_mode = a_bzrdir._control_files._dir_mode
703
file_mode = a_bzrdir._control_files._file_mode
705
def get_store(name, compressed=True, prefixed=False):
706
# FIXME: This approach of assuming stores are all entirely compressed
707
# or entirely uncompressed is tidy, but breaks upgrade from
708
# some existing branches where there's a mixture; we probably
709
# still want the option to look for both.
710
relpath = a_bzrdir._control_files._escape(name)
711
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
712
prefixed=prefixed, compressed=compressed,
715
#if self._transport.should_cache():
716
# cache_path = os.path.join(self.cache_root, name)
717
# os.mkdir(cache_path)
718
# store = bzrlib.store.CachedStore(store, cache_path)
721
# not broken out yet because the controlweaves|inventory_store
722
# and text_store | weave_store bits are still different.
723
if isinstance(_format, RepositoryFormat4):
724
# cannot remove these - there is still no consistent api
725
# which allows access to this old info.
726
self.inventory_store = get_store('inventory-store')
727
text_store = get_store('text-store')
728
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
732
"""AllInOne repositories cannot be shared."""
736
def set_make_working_trees(self, new_value):
737
"""Set the policy flag for making working trees when creating branches.
739
This only applies to branches that use this repository.
741
The default is 'True'.
742
:param new_value: True to restore the default, False to disable making
745
raise NotImplementedError(self.set_make_working_trees)
747
def make_working_trees(self):
748
"""Returns the policy for making working trees on new branches."""
752
def install_revision(repository, rev, revision_tree):
753
"""Install all revision data into a repository."""
756
for p_id in rev.parent_ids:
757
if repository.has_revision(p_id):
758
present_parents.append(p_id)
759
parent_trees[p_id] = repository.revision_tree(p_id)
761
parent_trees[p_id] = repository.revision_tree(None)
763
inv = revision_tree.inventory
765
# backwards compatability hack: skip the root id.
766
entries = inv.iter_entries()
768
# Add the texts that are not already present
769
for path, ie in entries:
770
w = repository.weave_store.get_weave_or_empty(ie.file_id,
771
repository.get_transaction())
772
if ie.revision not in w:
774
# FIXME: TODO: The following loop *may* be overlapping/duplicate
775
# with InventoryEntry.find_previous_heads(). if it is, then there
776
# is a latent bug here where the parents may have ancestors of each
778
for revision, tree in parent_trees.iteritems():
779
if ie.file_id not in tree:
781
parent_id = tree.inventory[ie.file_id].revision
782
if parent_id in text_parents:
784
text_parents.append(parent_id)
786
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
787
repository.get_transaction())
788
lines = revision_tree.get_file(ie.file_id).readlines()
789
vfile.add_lines(rev.revision_id, text_parents, lines)
791
# install the inventory
792
repository.add_inventory(rev.revision_id, inv, present_parents)
793
except errors.RevisionAlreadyPresent:
795
repository.add_revision(rev.revision_id, rev, inv)
798
class MetaDirRepository(Repository):
799
"""Repositories in the new meta-dir layout."""
801
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
802
super(MetaDirRepository, self).__init__(_format,
808
dir_mode = self.control_files._dir_mode
809
file_mode = self.control_files._file_mode
813
"""Return True if this repository is flagged as a shared repository."""
814
return self.control_files._transport.has('shared-storage')
817
def set_make_working_trees(self, new_value):
818
"""Set the policy flag for making working trees when creating branches.
820
This only applies to branches that use this repository.
822
The default is 'True'.
823
:param new_value: True to restore the default, False to disable making
828
self.control_files._transport.delete('no-working-trees')
829
except errors.NoSuchFile:
832
self.control_files.put_utf8('no-working-trees', '')
834
def make_working_trees(self):
835
"""Returns the policy for making working trees on new branches."""
836
return not self.control_files._transport.has('no-working-trees')
839
class KnitRepository(MetaDirRepository):
840
"""Knit format repository."""
842
def _warn_if_deprecated(self):
843
# This class isn't deprecated
846
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
847
inv_vf.add_lines_with_ghosts(revid, parents, lines)
850
def _all_revision_ids(self):
851
"""See Repository.all_revision_ids()."""
852
# Knits get the revision graph from the index of the revision knit, so
853
# it's always possible even if they're on an unlistable transport.
854
return self._revision_store.all_revision_ids(self.get_transaction())
856
def fileid_involved_between_revs(self, from_revid, to_revid):
857
"""Find file_id(s) which are involved in the changes between revisions.
859
This determines the set of revisions which are involved, and then
860
finds all file ids affected by those revisions.
862
vf = self._get_revision_vf()
863
from_set = set(vf.get_ancestry(from_revid))
864
to_set = set(vf.get_ancestry(to_revid))
865
changed = to_set.difference(from_set)
866
return self._fileid_involved_by_set(changed)
868
def fileid_involved(self, last_revid=None):
869
"""Find all file_ids modified in the ancestry of last_revid.
871
:param last_revid: If None, last_revision() will be used.
874
changed = set(self.all_revision_ids())
876
changed = set(self.get_ancestry(last_revid))
879
return self._fileid_involved_by_set(changed)
882
def get_ancestry(self, revision_id):
883
"""Return a list of revision-ids integrated by a revision.
885
This is topologically sorted.
887
if revision_id is None:
889
vf = self._get_revision_vf()
891
return [None] + vf.get_ancestry(revision_id)
892
except errors.RevisionNotPresent:
893
raise errors.NoSuchRevision(self, revision_id)
896
def get_revision(self, revision_id):
897
"""Return the Revision object for a named revision"""
898
return self.get_revision_reconcile(revision_id)
901
def get_revision_graph(self, revision_id=None):
902
"""Return a dictionary containing the revision graph.
904
:param revision_id: The revision_id to get a graph from. If None, then
905
the entire revision graph is returned. This is a deprecated mode of
906
operation and will be removed in the future.
907
:return: a dictionary of revision_id->revision_parents_list.
909
# special case NULL_REVISION
910
if revision_id == NULL_REVISION:
912
weave = self._get_revision_vf()
913
entire_graph = weave.get_graph()
914
if revision_id is None:
915
return weave.get_graph()
916
elif revision_id not in weave:
917
raise errors.NoSuchRevision(self, revision_id)
919
# add what can be reached from revision_id
921
pending = set([revision_id])
922
while len(pending) > 0:
924
result[node] = weave.get_parents(node)
925
for revision_id in result[node]:
926
if revision_id not in result:
927
pending.add(revision_id)
931
def get_revision_graph_with_ghosts(self, revision_ids=None):
932
"""Return a graph of the revisions with ghosts marked as applicable.
934
:param revision_ids: an iterable of revisions to graph or None for all.
935
:return: a Graph object with the graph reachable from revision_ids.
938
vf = self._get_revision_vf()
939
versions = set(vf.versions())
941
pending = set(self.all_revision_ids())
944
pending = set(revision_ids)
945
# special case NULL_REVISION
946
if NULL_REVISION in pending:
947
pending.remove(NULL_REVISION)
948
required = set(pending)
951
revision_id = pending.pop()
952
if not revision_id in versions:
953
if revision_id in required:
954
raise errors.NoSuchRevision(self, revision_id)
956
result.add_ghost(revision_id)
957
# mark it as done so we don't try for it again.
958
done.add(revision_id)
960
parent_ids = vf.get_parents_with_ghosts(revision_id)
961
for parent_id in parent_ids:
962
# is this queued or done ?
963
if (parent_id not in pending and
964
parent_id not in done):
966
pending.add(parent_id)
967
result.add_node(revision_id, parent_ids)
968
done.add(revision_id)
971
def _get_revision_vf(self):
972
""":return: a versioned file containing the revisions."""
973
vf = self._revision_store.get_revision_file(self.get_transaction())
977
def reconcile(self, other=None, thorough=False):
978
"""Reconcile this repository."""
979
from bzrlib.reconcile import KnitReconciler
980
reconciler = KnitReconciler(self, thorough=thorough)
981
reconciler.reconcile()
984
def revision_parents(self, revision_id):
985
return self._get_revision_vf().get_parents(revision_id)
988
class RepositoryFormat(object):
989
"""A repository format.
991
Formats provide three things:
992
* An initialization routine to construct repository data on disk.
993
* a format string which is used when the BzrDir supports versioned
995
* an open routine which returns a Repository instance.
997
Formats are placed in an dict by their format string for reference
998
during opening. These should be subclasses of RepositoryFormat
1001
Once a format is deprecated, just deprecate the initialize and open
1002
methods on the format class. Do not deprecate the object, as the
1003
object will be created every system load.
1005
Common instance attributes:
1006
_matchingbzrdir - the bzrdir format that the repository format was
1007
originally written to work with. This can be used if manually
1008
constructing a bzrdir and repository, or more commonly for test suite
1012
_default_format = None
1013
"""The default format used for new repositories."""
1016
"""The known formats."""
1019
return "<%s>" % self.__class__.__name__
1022
def find_format(klass, a_bzrdir):
1023
"""Return the format for the repository object in a_bzrdir."""
1025
transport = a_bzrdir.get_repository_transport(None)
1026
format_string = transport.get("format").read()
1027
return klass._formats[format_string]
1028
except errors.NoSuchFile:
1029
raise errors.NoRepositoryPresent(a_bzrdir)
1031
raise errors.UnknownFormatError(format=format_string)
1033
def _get_control_store(self, repo_transport, control_files):
1034
"""Return the control store for this repository."""
1035
raise NotImplementedError(self._get_control_store)
1038
def get_default_format(klass):
1039
"""Return the current default format."""
1040
return klass._default_format
1042
def get_format_string(self):
1043
"""Return the ASCII format string that identifies this format.
1045
Note that in pre format ?? repositories the format string is
1046
not permitted nor written to disk.
1048
raise NotImplementedError(self.get_format_string)
1050
def get_format_description(self):
1051
"""Return the short description for this format."""
1052
raise NotImplementedError(self.get_format_description)
1054
def _get_revision_store(self, repo_transport, control_files):
1055
"""Return the revision store object for this a_bzrdir."""
1056
raise NotImplementedError(self._get_revision_store)
1058
def _get_text_rev_store(self,
1065
"""Common logic for getting a revision store for a repository.
1067
see self._get_revision_store for the subclass-overridable method to
1068
get the store for a repository.
1070
from bzrlib.store.revision.text import TextRevisionStore
1071
dir_mode = control_files._dir_mode
1072
file_mode = control_files._file_mode
1073
text_store =TextStore(transport.clone(name),
1075
compressed=compressed,
1077
file_mode=file_mode)
1078
_revision_store = TextRevisionStore(text_store, serializer)
1079
return _revision_store
1081
def _get_versioned_file_store(self,
1086
versionedfile_class=WeaveFile,
1088
weave_transport = control_files._transport.clone(name)
1089
dir_mode = control_files._dir_mode
1090
file_mode = control_files._file_mode
1091
return VersionedFileStore(weave_transport, prefixed=prefixed,
1093
file_mode=file_mode,
1094
versionedfile_class=versionedfile_class,
1097
def initialize(self, a_bzrdir, shared=False):
1098
"""Initialize a repository of this format in a_bzrdir.
1100
:param a_bzrdir: The bzrdir to put the new repository in it.
1101
:param shared: The repository should be initialized as a sharable one.
1103
This may raise UninitializableFormat if shared repository are not
1104
compatible the a_bzrdir.
1107
def is_supported(self):
1108
"""Is this format supported?
1110
Supported formats must be initializable and openable.
1111
Unsupported formats may not support initialization or committing or
1112
some other features depending on the reason for not being supported.
1116
def open(self, a_bzrdir, _found=False):
1117
"""Return an instance of this format for the bzrdir a_bzrdir.
1119
_found is a private parameter, do not use it.
1121
raise NotImplementedError(self.open)
1124
def register_format(klass, format):
1125
klass._formats[format.get_format_string()] = format
1128
def set_default_format(klass, format):
1129
klass._default_format = format
1132
def unregister_format(klass, format):
1133
assert klass._formats[format.get_format_string()] is format
1134
del klass._formats[format.get_format_string()]
1137
class PreSplitOutRepositoryFormat(RepositoryFormat):
1138
"""Base class for the pre split out repository formats."""
1140
def initialize(self, a_bzrdir, shared=False, _internal=False):
1141
"""Create a weave repository.
1143
TODO: when creating split out bzr branch formats, move this to a common
1144
base for Format5, Format6. or something like that.
1146
from bzrlib.weavefile import write_weave_v5
1147
from bzrlib.weave import Weave
1150
raise errors.IncompatibleFormat(self, a_bzrdir._format)
1153
# always initialized when the bzrdir is.
1154
return self.open(a_bzrdir, _found=True)
1156
# Create an empty weave
1158
write_weave_v5(Weave(), sio)
1159
empty_weave = sio.getvalue()
1161
mutter('creating repository in %s.', a_bzrdir.transport.base)
1162
dirs = ['revision-store', 'weaves']
1163
files = [('inventory.weave', StringIO(empty_weave)),
1166
# FIXME: RBC 20060125 don't peek under the covers
1167
# NB: no need to escape relative paths that are url safe.
1168
control_files = LockableFiles(a_bzrdir.transport, 'branch-lock',
1170
control_files.create_lock()
1171
control_files.lock_write()
1172
control_files._transport.mkdir_multi(dirs,
1173
mode=control_files._dir_mode)
1175
for file, content in files:
1176
control_files.put(file, content)
1178
control_files.unlock()
1179
return self.open(a_bzrdir, _found=True)
1181
def _get_control_store(self, repo_transport, control_files):
1182
"""Return the control store for this repository."""
1183
return self._get_versioned_file_store('',
1188
def _get_text_store(self, transport, control_files):
1189
"""Get a store for file texts for this format."""
1190
raise NotImplementedError(self._get_text_store)
1192
def open(self, a_bzrdir, _found=False):
1193
"""See RepositoryFormat.open()."""
1195
# we are being called directly and must probe.
1196
raise NotImplementedError
1198
repo_transport = a_bzrdir.get_repository_transport(None)
1199
control_files = a_bzrdir._control_files
1200
text_store = self._get_text_store(repo_transport, control_files)
1201
control_store = self._get_control_store(repo_transport, control_files)
1202
_revision_store = self._get_revision_store(repo_transport, control_files)
1203
return AllInOneRepository(_format=self,
1205
_revision_store=_revision_store,
1206
control_store=control_store,
1207
text_store=text_store)
1210
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1211
"""Bzr repository format 4.
1213
This repository format has:
1215
- TextStores for texts, inventories,revisions.
1217
This format is deprecated: it indexes texts using a text id which is
1218
removed in format 5; initialization and write support for this format
1223
super(RepositoryFormat4, self).__init__()
1224
self._matchingbzrdir = bzrdir.BzrDirFormat4()
1226
def get_format_description(self):
1227
"""See RepositoryFormat.get_format_description()."""
1228
return "Repository format 4"
1230
def initialize(self, url, shared=False, _internal=False):
1231
"""Format 4 branches cannot be created."""
1232
raise errors.UninitializableFormat(self)
1234
def is_supported(self):
1235
"""Format 4 is not supported.
1237
It is not supported because the model changed from 4 to 5 and the
1238
conversion logic is expensive - so doing it on the fly was not
1243
def _get_control_store(self, repo_transport, control_files):
1244
"""Format 4 repositories have no formal control store at this point.
1246
This will cause any control-file-needing apis to fail - this is desired.
1250
def _get_revision_store(self, repo_transport, control_files):
1251
"""See RepositoryFormat._get_revision_store()."""
1252
from bzrlib.xml4 import serializer_v4
1253
return self._get_text_rev_store(repo_transport,
1256
serializer=serializer_v4)
1258
def _get_text_store(self, transport, control_files):
1259
"""See RepositoryFormat._get_text_store()."""
1262
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1263
"""Bzr control format 5.
1265
This repository format has:
1266
- weaves for file texts and inventory
1268
- TextStores for revisions and signatures.
1272
super(RepositoryFormat5, self).__init__()
1273
self._matchingbzrdir = bzrdir.BzrDirFormat5()
1275
def get_format_description(self):
1276
"""See RepositoryFormat.get_format_description()."""
1277
return "Weave repository format 5"
1279
def _get_revision_store(self, repo_transport, control_files):
1280
"""See RepositoryFormat._get_revision_store()."""
1281
"""Return the revision store object for this a_bzrdir."""
1282
return self._get_text_rev_store(repo_transport,
1287
def _get_text_store(self, transport, control_files):
1288
"""See RepositoryFormat._get_text_store()."""
1289
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1292
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1293
"""Bzr control format 6.
1295
This repository format has:
1296
- weaves for file texts and inventory
1297
- hash subdirectory based stores.
1298
- TextStores for revisions and signatures.
1302
super(RepositoryFormat6, self).__init__()
1303
self._matchingbzrdir = bzrdir.BzrDirFormat6()
1305
def get_format_description(self):
1306
"""See RepositoryFormat.get_format_description()."""
1307
return "Weave repository format 6"
1309
def _get_revision_store(self, repo_transport, control_files):
1310
"""See RepositoryFormat._get_revision_store()."""
1311
return self._get_text_rev_store(repo_transport,
1317
def _get_text_store(self, transport, control_files):
1318
"""See RepositoryFormat._get_text_store()."""
1319
return self._get_versioned_file_store('weaves', transport, control_files)
1322
class MetaDirRepositoryFormat(RepositoryFormat):
1323
"""Common base class for the new repositories using the metadir layout."""
1326
super(MetaDirRepositoryFormat, self).__init__()
1327
self._matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1329
def _create_control_files(self, a_bzrdir):
1330
"""Create the required files and the initial control_files object."""
1331
# FIXME: RBC 20060125 don't peek under the covers
1332
# NB: no need to escape relative paths that are url safe.
1333
repository_transport = a_bzrdir.get_repository_transport(self)
1334
control_files = LockableFiles(repository_transport, 'lock', LockDir)
1335
control_files.create_lock()
1336
return control_files
1338
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1339
"""Upload the initial blank content."""
1340
control_files = self._create_control_files(a_bzrdir)
1341
control_files.lock_write()
1343
control_files._transport.mkdir_multi(dirs,
1344
mode=control_files._dir_mode)
1345
for file, content in files:
1346
control_files.put(file, content)
1347
for file, content in utf8_files:
1348
control_files.put_utf8(file, content)
1350
control_files.put_utf8('shared-storage', '')
1352
control_files.unlock()
1355
class RepositoryFormat7(MetaDirRepositoryFormat):
1356
"""Bzr repository 7.
1358
This repository format has:
1359
- weaves for file texts and inventory
1360
- hash subdirectory based stores.
1361
- TextStores for revisions and signatures.
1362
- a format marker of its own
1363
- an optional 'shared-storage' flag
1364
- an optional 'no-working-trees' flag
1367
def _get_control_store(self, repo_transport, control_files):
1368
"""Return the control store for this repository."""
1369
return self._get_versioned_file_store('',
1374
def get_format_string(self):
1375
"""See RepositoryFormat.get_format_string()."""
1376
return "Bazaar-NG Repository format 7"
1378
def get_format_description(self):
1379
"""See RepositoryFormat.get_format_description()."""
1380
return "Weave repository format 7"
1382
def _get_revision_store(self, repo_transport, control_files):
1383
"""See RepositoryFormat._get_revision_store()."""
1384
return self._get_text_rev_store(repo_transport,
1391
def _get_text_store(self, transport, control_files):
1392
"""See RepositoryFormat._get_text_store()."""
1393
return self._get_versioned_file_store('weaves',
1397
def initialize(self, a_bzrdir, shared=False):
1398
"""Create a weave repository.
1400
:param shared: If true the repository will be initialized as a shared
1403
from bzrlib.weavefile import write_weave_v5
1404
from bzrlib.weave import Weave
1406
# Create an empty weave
1408
write_weave_v5(Weave(), sio)
1409
empty_weave = sio.getvalue()
1411
mutter('creating repository in %s.', a_bzrdir.transport.base)
1412
dirs = ['revision-store', 'weaves']
1413
files = [('inventory.weave', StringIO(empty_weave)),
1415
utf8_files = [('format', self.get_format_string())]
1417
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1418
return self.open(a_bzrdir=a_bzrdir, _found=True)
1420
def open(self, a_bzrdir, _found=False, _override_transport=None):
1421
"""See RepositoryFormat.open().
1423
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1424
repository at a slightly different url
1425
than normal. I.e. during 'upgrade'.
1428
format = RepositoryFormat.find_format(a_bzrdir)
1429
assert format.__class__ == self.__class__
1430
if _override_transport is not None:
1431
repo_transport = _override_transport
1433
repo_transport = a_bzrdir.get_repository_transport(None)
1434
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1435
text_store = self._get_text_store(repo_transport, control_files)
1436
control_store = self._get_control_store(repo_transport, control_files)
1437
_revision_store = self._get_revision_store(repo_transport, control_files)
1438
return MetaDirRepository(_format=self,
1440
control_files=control_files,
1441
_revision_store=_revision_store,
1442
control_store=control_store,
1443
text_store=text_store)
1446
class RepositoryFormatKnit1(MetaDirRepositoryFormat):
1447
"""Bzr repository knit format 1.
1449
This repository format has:
1450
- knits for file texts and inventory
1451
- hash subdirectory based stores.
1452
- knits for revisions and signatures
1453
- TextStores for revisions and signatures.
1454
- a format marker of its own
1455
- an optional 'shared-storage' flag
1456
- an optional 'no-working-trees' flag
1459
This format was introduced in bzr 0.8.
1462
def _get_control_store(self, repo_transport, control_files):
1463
"""Return the control store for this repository."""
1464
return VersionedFileStore(
1467
file_mode=control_files._file_mode,
1468
versionedfile_class=KnitVersionedFile,
1469
versionedfile_kwargs={'factory':KnitPlainFactory()},
1472
def get_format_string(self):
1473
"""See RepositoryFormat.get_format_string()."""
1474
return "Bazaar-NG Knit Repository Format 1"
1476
def get_format_description(self):
1477
"""See RepositoryFormat.get_format_description()."""
1478
return "Knit repository format 1"
1480
def _get_revision_store(self, repo_transport, control_files):
1481
"""See RepositoryFormat._get_revision_store()."""
1482
from bzrlib.store.revision.knit import KnitRevisionStore
1483
versioned_file_store = VersionedFileStore(
1485
file_mode=control_files._file_mode,
1488
versionedfile_class=KnitVersionedFile,
1489
versionedfile_kwargs={'delta':False, 'factory':KnitPlainFactory()},
1492
return KnitRevisionStore(versioned_file_store)
1494
def _get_text_store(self, transport, control_files):
1495
"""See RepositoryFormat._get_text_store()."""
1496
return self._get_versioned_file_store('knits',
1499
versionedfile_class=KnitVersionedFile,
1502
def initialize(self, a_bzrdir, shared=False):
1503
"""Create a knit format 1 repository.
1505
:param a_bzrdir: bzrdir to contain the new repository; must already
1507
:param shared: If true the repository will be initialized as a shared
1510
mutter('creating repository in %s.', a_bzrdir.transport.base)
1511
dirs = ['revision-store', 'knits']
1513
utf8_files = [('format', self.get_format_string())]
1515
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1516
repo_transport = a_bzrdir.get_repository_transport(None)
1517
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1518
control_store = self._get_control_store(repo_transport, control_files)
1519
transaction = transactions.WriteTransaction()
1520
# trigger a write of the inventory store.
1521
control_store.get_weave_or_empty('inventory', transaction)
1522
_revision_store = self._get_revision_store(repo_transport, control_files)
1523
_revision_store.has_revision_id('A', transaction)
1524
_revision_store.get_signature_file(transaction)
1525
return self.open(a_bzrdir=a_bzrdir, _found=True)
1527
def open(self, a_bzrdir, _found=False, _override_transport=None):
1528
"""See RepositoryFormat.open().
1530
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1531
repository at a slightly different url
1532
than normal. I.e. during 'upgrade'.
1535
format = RepositoryFormat.find_format(a_bzrdir)
1536
assert format.__class__ == self.__class__
1537
if _override_transport is not None:
1538
repo_transport = _override_transport
1540
repo_transport = a_bzrdir.get_repository_transport(None)
1541
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1542
text_store = self._get_text_store(repo_transport, control_files)
1543
control_store = self._get_control_store(repo_transport, control_files)
1544
_revision_store = self._get_revision_store(repo_transport, control_files)
1545
return KnitRepository(_format=self,
1547
control_files=control_files,
1548
_revision_store=_revision_store,
1549
control_store=control_store,
1550
text_store=text_store)
1553
# formats which have no format string are not discoverable
1554
# and not independently creatable, so are not registered.
1555
RepositoryFormat.register_format(RepositoryFormat7())
1556
_default_format = RepositoryFormatKnit1()
1557
RepositoryFormat.register_format(_default_format)
1558
RepositoryFormat.set_default_format(_default_format)
1559
_legacy_formats = [RepositoryFormat4(),
1560
RepositoryFormat5(),
1561
RepositoryFormat6()]
1564
class InterRepository(InterObject):
1565
"""This class represents operations taking place between two repositories.
1567
Its instances have methods like copy_content and fetch, and contain
1568
references to the source and target repositories these operations can be
1571
Often we will provide convenience methods on 'repository' which carry out
1572
operations with another repository - they will always forward to
1573
InterRepository.get(other).method_name(parameters).
1577
"""The available optimised InterRepository types."""
1580
def copy_content(self, revision_id=None, basis=None):
1581
"""Make a complete copy of the content in self into destination.
1583
This is a destructive operation! Do not use it on existing
1586
:param revision_id: Only copy the content needed to construct
1587
revision_id and its parents.
1588
:param basis: Copy the needed data preferentially from basis.
1591
self.target.set_make_working_trees(self.source.make_working_trees())
1592
except NotImplementedError:
1594
# grab the basis available data
1595
if basis is not None:
1596
self.target.fetch(basis, revision_id=revision_id)
1597
# but don't bother fetching if we have the needed data now.
1598
if (revision_id not in (None, NULL_REVISION) and
1599
self.target.has_revision(revision_id)):
1601
self.target.fetch(self.source, revision_id=revision_id)
1604
def fetch(self, revision_id=None, pb=None):
1605
"""Fetch the content required to construct revision_id.
1607
The content is copied from source to target.
1609
:param revision_id: if None all content is copied, if NULL_REVISION no
1611
:param pb: optional progress bar to use for progress reports. If not
1612
provided a default one will be created.
1614
Returns the copied revision count and the failed revisions in a tuple:
1617
from bzrlib.fetch import GenericRepoFetcher
1618
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1619
self.source, self.source._format, self.target, self.target._format)
1620
f = GenericRepoFetcher(to_repository=self.target,
1621
from_repository=self.source,
1622
last_revision=revision_id,
1624
return f.count_copied, f.failed_revisions
1627
def missing_revision_ids(self, revision_id=None):
1628
"""Return the revision ids that source has that target does not.
1630
These are returned in topological order.
1632
:param revision_id: only return revision ids included by this
1635
# generic, possibly worst case, slow code path.
1636
target_ids = set(self.target.all_revision_ids())
1637
if revision_id is not None:
1638
source_ids = self.source.get_ancestry(revision_id)
1639
assert source_ids[0] == None
1642
source_ids = self.source.all_revision_ids()
1643
result_set = set(source_ids).difference(target_ids)
1644
# this may look like a no-op: its not. It preserves the ordering
1645
# other_ids had while only returning the members from other_ids
1646
# that we've decided we need.
1647
return [rev_id for rev_id in source_ids if rev_id in result_set]
1650
class InterWeaveRepo(InterRepository):
1651
"""Optimised code paths between Weave based repositories."""
1653
_matching_repo_format = RepositoryFormat7()
1654
"""Repository format for testing with."""
1657
def is_compatible(source, target):
1658
"""Be compatible with known Weave formats.
1660
We don't test for the stores being of specific types because that
1661
could lead to confusing results, and there is no need to be
1665
return (isinstance(source._format, (RepositoryFormat5,
1667
RepositoryFormat7)) and
1668
isinstance(target._format, (RepositoryFormat5,
1670
RepositoryFormat7)))
1671
except AttributeError:
1675
def copy_content(self, revision_id=None, basis=None):
1676
"""See InterRepository.copy_content()."""
1677
# weave specific optimised path:
1678
if basis is not None:
1679
# copy the basis in, then fetch remaining data.
1680
basis.copy_content_into(self.target, revision_id)
1681
# the basis copy_content_into could miss-set this.
1683
self.target.set_make_working_trees(self.source.make_working_trees())
1684
except NotImplementedError:
1686
self.target.fetch(self.source, revision_id=revision_id)
1689
self.target.set_make_working_trees(self.source.make_working_trees())
1690
except NotImplementedError:
1692
# FIXME do not peek!
1693
if self.source.control_files._transport.listable():
1694
pb = ui.ui_factory.nested_progress_bar()
1696
self.target.weave_store.copy_all_ids(
1697
self.source.weave_store,
1699
from_transaction=self.source.get_transaction(),
1700
to_transaction=self.target.get_transaction())
1701
pb.update('copying inventory', 0, 1)
1702
self.target.control_weaves.copy_multi(
1703
self.source.control_weaves, ['inventory'],
1704
from_transaction=self.source.get_transaction(),
1705
to_transaction=self.target.get_transaction())
1706
self.target._revision_store.text_store.copy_all_ids(
1707
self.source._revision_store.text_store,
1712
self.target.fetch(self.source, revision_id=revision_id)
1715
def fetch(self, revision_id=None, pb=None):
1716
"""See InterRepository.fetch()."""
1717
from bzrlib.fetch import GenericRepoFetcher
1718
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1719
self.source, self.source._format, self.target, self.target._format)
1720
f = GenericRepoFetcher(to_repository=self.target,
1721
from_repository=self.source,
1722
last_revision=revision_id,
1724
return f.count_copied, f.failed_revisions
1727
def missing_revision_ids(self, revision_id=None):
1728
"""See InterRepository.missing_revision_ids()."""
1729
# we want all revisions to satisfy revision_id in source.
1730
# but we don't want to stat every file here and there.
1731
# we want then, all revisions other needs to satisfy revision_id
1732
# checked, but not those that we have locally.
1733
# so the first thing is to get a subset of the revisions to
1734
# satisfy revision_id in source, and then eliminate those that
1735
# we do already have.
1736
# this is slow on high latency connection to self, but as as this
1737
# disk format scales terribly for push anyway due to rewriting
1738
# inventory.weave, this is considered acceptable.
1740
if revision_id is not None:
1741
source_ids = self.source.get_ancestry(revision_id)
1742
assert source_ids[0] == None
1745
source_ids = self.source._all_possible_ids()
1746
source_ids_set = set(source_ids)
1747
# source_ids is the worst possible case we may need to pull.
1748
# now we want to filter source_ids against what we actually
1749
# have in target, but don't try to check for existence where we know
1750
# we do not have a revision as that would be pointless.
1751
target_ids = set(self.target._all_possible_ids())
1752
possibly_present_revisions = target_ids.intersection(source_ids_set)
1753
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1754
required_revisions = source_ids_set.difference(actually_present_revisions)
1755
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1756
if revision_id is not None:
1757
# we used get_ancestry to determine source_ids then we are assured all
1758
# revisions referenced are present as they are installed in topological order.
1759
# and the tip revision was validated by get_ancestry.
1760
return required_topo_revisions
1762
# if we just grabbed the possibly available ids, then
1763
# we only have an estimate of whats available and need to validate
1764
# that against the revision records.
1765
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1768
class InterKnitRepo(InterRepository):
1769
"""Optimised code paths between Knit based repositories."""
1771
_matching_repo_format = RepositoryFormatKnit1()
1772
"""Repository format for testing with."""
1775
def is_compatible(source, target):
1776
"""Be compatible with known Knit formats.
1778
We don't test for the stores being of specific types because that
1779
could lead to confusing results, and there is no need to be
1783
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1784
isinstance(target._format, (RepositoryFormatKnit1)))
1785
except AttributeError:
1789
def fetch(self, revision_id=None, pb=None):
1790
"""See InterRepository.fetch()."""
1791
from bzrlib.fetch import KnitRepoFetcher
1792
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1793
self.source, self.source._format, self.target, self.target._format)
1794
f = KnitRepoFetcher(to_repository=self.target,
1795
from_repository=self.source,
1796
last_revision=revision_id,
1798
return f.count_copied, f.failed_revisions
1801
def missing_revision_ids(self, revision_id=None):
1802
"""See InterRepository.missing_revision_ids()."""
1803
if revision_id is not None:
1804
source_ids = self.source.get_ancestry(revision_id)
1805
assert source_ids[0] == None
1808
source_ids = self.source._all_possible_ids()
1809
source_ids_set = set(source_ids)
1810
# source_ids is the worst possible case we may need to pull.
1811
# now we want to filter source_ids against what we actually
1812
# have in target, but don't try to check for existence where we know
1813
# we do not have a revision as that would be pointless.
1814
target_ids = set(self.target._all_possible_ids())
1815
possibly_present_revisions = target_ids.intersection(source_ids_set)
1816
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1817
required_revisions = source_ids_set.difference(actually_present_revisions)
1818
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1819
if revision_id is not None:
1820
# we used get_ancestry to determine source_ids then we are assured all
1821
# revisions referenced are present as they are installed in topological order.
1822
# and the tip revision was validated by get_ancestry.
1823
return required_topo_revisions
1825
# if we just grabbed the possibly available ids, then
1826
# we only have an estimate of whats available and need to validate
1827
# that against the revision records.
1828
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1830
InterRepository.register_optimiser(InterWeaveRepo)
1831
InterRepository.register_optimiser(InterKnitRepo)
1834
class RepositoryTestProviderAdapter(object):
1835
"""A tool to generate a suite testing multiple repository formats at once.
1837
This is done by copying the test once for each transport and injecting
1838
the transport_server, transport_readonly_server, and bzrdir_format and
1839
repository_format classes into each copy. Each copy is also given a new id()
1840
to make it easy to identify.
1843
def __init__(self, transport_server, transport_readonly_server, formats):
1844
self._transport_server = transport_server
1845
self._transport_readonly_server = transport_readonly_server
1846
self._formats = formats
1848
def adapt(self, test):
1849
result = TestSuite()
1850
for repository_format, bzrdir_format in self._formats:
1851
new_test = deepcopy(test)
1852
new_test.transport_server = self._transport_server
1853
new_test.transport_readonly_server = self._transport_readonly_server
1854
new_test.bzrdir_format = bzrdir_format
1855
new_test.repository_format = repository_format
1856
def make_new_test_id():
1857
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1858
return lambda: new_id
1859
new_test.id = make_new_test_id()
1860
result.addTest(new_test)
1864
class InterRepositoryTestProviderAdapter(object):
1865
"""A tool to generate a suite testing multiple inter repository formats.
1867
This is done by copying the test once for each interrepo provider and injecting
1868
the transport_server, transport_readonly_server, repository_format and
1869
repository_to_format classes into each copy.
1870
Each copy is also given a new id() to make it easy to identify.
1873
def __init__(self, transport_server, transport_readonly_server, formats):
1874
self._transport_server = transport_server
1875
self._transport_readonly_server = transport_readonly_server
1876
self._formats = formats
1878
def adapt(self, test):
1879
result = TestSuite()
1880
for interrepo_class, repository_format, repository_format_to in self._formats:
1881
new_test = deepcopy(test)
1882
new_test.transport_server = self._transport_server
1883
new_test.transport_readonly_server = self._transport_readonly_server
1884
new_test.interrepo_class = interrepo_class
1885
new_test.repository_format = repository_format
1886
new_test.repository_format_to = repository_format_to
1887
def make_new_test_id():
1888
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1889
return lambda: new_id
1890
new_test.id = make_new_test_id()
1891
result.addTest(new_test)
1895
def default_test_list():
1896
"""Generate the default list of interrepo permutations to test."""
1898
# test the default InterRepository between format 6 and the current
1900
# XXX: robertc 20060220 reinstate this when there are two supported
1901
# formats which do not have an optimal code path between them.
1902
result.append((InterRepository,
1903
RepositoryFormat6(),
1904
RepositoryFormatKnit1()))
1905
for optimiser in InterRepository._optimisers:
1906
result.append((optimiser,
1907
optimiser._matching_repo_format,
1908
optimiser._matching_repo_format
1910
# if there are specific combinations we want to use, we can add them
1915
class CopyConverter(object):
1916
"""A repository conversion tool which just performs a copy of the content.
1918
This is slow but quite reliable.
1921
def __init__(self, target_format):
1922
"""Create a CopyConverter.
1924
:param target_format: The format the resulting repository should be.
1926
self.target_format = target_format
1928
def convert(self, repo, pb):
1929
"""Perform the conversion of to_convert, giving feedback via pb.
1931
:param to_convert: The disk object to convert.
1932
:param pb: a progress bar to use for progress information.
1937
# this is only useful with metadir layouts - separated repo content.
1938
# trigger an assertion if not such
1939
repo._format.get_format_string()
1940
self.repo_dir = repo.bzrdir
1941
self.step('Moving repository to repository.backup')
1942
self.repo_dir.transport.move('repository', 'repository.backup')
1943
backup_transport = self.repo_dir.transport.clone('repository.backup')
1944
self.source_repo = repo._format.open(self.repo_dir,
1946
_override_transport=backup_transport)
1947
self.step('Creating new repository')
1948
converted = self.target_format.initialize(self.repo_dir,
1949
self.source_repo.is_shared())
1950
converted.lock_write()
1952
self.step('Copying content into repository.')
1953
self.source_repo.copy_content_into(converted)
1956
self.step('Deleting old repository content.')
1957
self.repo_dir.transport.delete_tree('repository.backup')
1958
self.pb.note('repository converted')
1960
def step(self, message):
1961
"""Update the pb by a step."""
1963
self.pb.update(message, self.count, self.total)
1966
class CommitBuilder(object):
1967
"""Provides an interface to build up a commit.
1969
This allows describing a tree to be committed without needing to
1970
know the internals of the format of the repository.
1972
def __init__(self, repository, parents, config, timestamp=None,
1973
timezone=None, committer=None, revprops=None,
1975
"""Initiate a CommitBuilder.
1977
:param repository: Repository to commit to.
1978
:param parents: Revision ids of the parents of the new revision.
1979
:param config: Configuration to use.
1980
:param timestamp: Optional timestamp recorded for commit.
1981
:param timezone: Optional timezone for timestamp.
1982
:param committer: Optional committer to set for commit.
1983
:param revprops: Optional dictionary of revision properties.
1984
:param revision_id: Optional revision id.
1986
self._config = config
1988
if committer is None:
1989
self._committer = self._config.username()
1991
assert isinstance(committer, basestring), type(committer)
1992
self._committer = committer
1994
self.new_inventory = Inventory()
1995
self._new_revision_id = revision_id
1996
self.parents = parents
1997
self.repository = repository
2000
if revprops is not None:
2001
self._revprops.update(revprops)
2003
if timestamp is None:
2004
timestamp = time.time()
2005
# Restrict resolution to 1ms
2006
self._timestamp = round(timestamp, 3)
2008
if timezone is None:
2009
self._timezone = local_time_offset()
2011
self._timezone = int(timezone)
2013
self._generate_revision_if_needed()
2015
def commit(self, message):
2016
"""Make the actual commit.
2018
:return: The revision id of the recorded revision.
2020
rev = Revision(timestamp=self._timestamp,
2021
timezone=self._timezone,
2022
committer=self._committer,
2024
inventory_sha1=self.inv_sha1,
2025
revision_id=self._new_revision_id,
2026
properties=self._revprops)
2027
rev.parent_ids = self.parents
2028
self.repository.add_revision(self._new_revision_id, rev,
2029
self.new_inventory, self._config)
2030
return self._new_revision_id
2032
def finish_inventory(self):
2033
"""Tell the builder that the inventory is finished."""
2034
self.new_inventory.revision_id = self._new_revision_id
2035
self.inv_sha1 = self.repository.add_inventory(
2036
self._new_revision_id,
2041
def _gen_revision_id(self):
2042
"""Return new revision-id."""
2043
s = '%s-%s-' % (self._config.user_email(),
2044
compact_date(self._timestamp))
2045
s += hexlify(rand_bytes(8))
2048
def _generate_revision_if_needed(self):
2049
"""Create a revision id if None was supplied.
2051
If the repository can not support user-specified revision ids
2052
they should override this function and raise UnsupportedOperation
2053
if _new_revision_id is not None.
2055
:raises: UnsupportedOperation
2057
if self._new_revision_id is None:
2058
self._new_revision_id = self._gen_revision_id()
2060
def record_entry_contents(self, ie, parent_invs, path, tree):
2061
"""Record the content of ie from tree into the commit if needed.
2063
:param ie: An inventory entry present in the commit.
2064
:param parent_invs: The inventories of the parent revisions of the
2066
:param path: The path the entry is at in the tree.
2067
:param tree: The tree which contains this entry and should be used to
2070
self.new_inventory.add(ie)
2072
# ie.revision is always None if the InventoryEntry is considered
2073
# for committing. ie.snapshot will record the correct revision
2074
# which may be the sole parent if it is untouched.
2075
if ie.revision is not None:
2077
previous_entries = ie.find_previous_heads(
2079
self.repository.weave_store,
2080
self.repository.get_transaction())
2081
# we are creating a new revision for ie in the history store
2083
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2085
def modified_directory(self, file_id, file_parents):
2086
"""Record the presence of a symbolic link.
2088
:param file_id: The file_id of the link to record.
2089
:param file_parents: The per-file parent revision ids.
2091
self._add_text_to_weave(file_id, [], file_parents.keys())
2093
def modified_file_text(self, file_id, file_parents,
2094
get_content_byte_lines, text_sha1=None,
2096
"""Record the text of file file_id
2098
:param file_id: The file_id of the file to record the text of.
2099
:param file_parents: The per-file parent revision ids.
2100
:param get_content_byte_lines: A callable which will return the byte
2102
:param text_sha1: Optional SHA1 of the file contents.
2103
:param text_size: Optional size of the file contents.
2105
# mutter('storing text of file {%s} in revision {%s} into %r',
2106
# file_id, self._new_revision_id, self.repository.weave_store)
2107
# special case to avoid diffing on renames or
2109
if (len(file_parents) == 1
2110
and text_sha1 == file_parents.values()[0].text_sha1
2111
and text_size == file_parents.values()[0].text_size):
2112
previous_ie = file_parents.values()[0]
2113
versionedfile = self.repository.weave_store.get_weave(file_id,
2114
self.repository.get_transaction())
2115
versionedfile.clone_text(self._new_revision_id,
2116
previous_ie.revision, file_parents.keys())
2117
return text_sha1, text_size
2119
new_lines = get_content_byte_lines()
2120
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2121
# should return the SHA1 and size
2122
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2123
return osutils.sha_strings(new_lines), \
2124
sum(map(len, new_lines))
2126
def modified_link(self, file_id, file_parents, link_target):
2127
"""Record the presence of a symbolic link.
2129
:param file_id: The file_id of the link to record.
2130
:param file_parents: The per-file parent revision ids.
2131
:param link_target: Target location of this link.
2133
self._add_text_to_weave(file_id, [], file_parents.keys())
2135
def _add_text_to_weave(self, file_id, new_lines, parents):
2136
versionedfile = self.repository.weave_store.get_weave_or_empty(
2137
file_id, self.repository.get_transaction())
2138
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2139
versionedfile.clear_cache()
2151
def _unescaper(match, _map=_unescape_map):
2152
return _map[match.group(1)]
2158
def _unescape_xml(data):
2159
"""Unescape predefined XML entities in a string of data."""
2161
if _unescape_re is None:
2162
_unescape_re = re.compile('\&([^;]*);')
2163
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
