1
# Copyright (C) 2005, 2006, 2007 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 cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
38
revision as _mod_revision,
43
from bzrlib.bundle import serializer
44
from bzrlib.revisiontree import RevisionTree
45
from bzrlib.store.versioned import VersionedFileStore
46
from bzrlib.store.text import TextStore
47
from bzrlib.testament import Testament
51
from bzrlib.decorators import needs_read_lock, needs_write_lock
52
from bzrlib.inter import InterObject
53
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
54
from bzrlib.symbol_versioning import (
58
from bzrlib.trace import mutter, note, warning
61
# Old formats display a warning, but only once
62
_deprecation_warning_done = False
65
######################################################################
68
class Repository(object):
69
"""Repository holding history for one or more branches.
71
The repository holds and retrieves historical information including
72
revisions and file history. It's normally accessed only by the Branch,
73
which views a particular line of development through that history.
75
The Repository builds on top of Stores and a Transport, which respectively
76
describe the disk data format and the way of accessing the (possibly
80
_file_ids_altered_regex = lazy_regex.lazy_compile(
81
r'file_id="(?P<file_id>[^"]+)"'
82
r'.*revision="(?P<revision_id>[^"]+)"'
86
def add_inventory(self, revision_id, inv, parents):
87
"""Add the inventory inv to the repository as revision_id.
89
:param parents: The revision ids of the parents that revision_id
90
is known to have and are in the repository already.
92
returns the sha1 of the serialized inventory.
94
revision_id = osutils.safe_revision_id(revision_id)
95
_mod_revision.check_not_reserved_id(revision_id)
96
assert inv.revision_id is None or inv.revision_id == revision_id, \
97
"Mismatch between inventory revision" \
98
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
99
assert inv.root is not None
100
inv_text = self.serialise_inventory(inv)
101
inv_sha1 = osutils.sha_string(inv_text)
102
inv_vf = self.control_weaves.get_weave('inventory',
103
self.get_transaction())
104
self._inventory_add_lines(inv_vf, revision_id, parents,
105
osutils.split_lines(inv_text))
108
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines):
110
for parent in parents:
112
final_parents.append(parent)
114
inv_vf.add_lines(revision_id, final_parents, lines)
117
def add_revision(self, revision_id, rev, inv=None, config=None):
118
"""Add rev to the revision store as revision_id.
120
:param revision_id: the revision id to use.
121
:param rev: The revision object.
122
:param inv: The inventory for the revision. if None, it will be looked
123
up in the inventory storer
124
:param config: If None no digital signature will be created.
125
If supplied its signature_needed method will be used
126
to determine if a signature should be made.
128
revision_id = osutils.safe_revision_id(revision_id)
129
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
131
_mod_revision.check_not_reserved_id(revision_id)
132
if config is not None and config.signature_needed():
134
inv = self.get_inventory(revision_id)
135
plaintext = Testament(rev, inv).as_short_text()
136
self.store_revision_signature(
137
gpg.GPGStrategy(config), plaintext, revision_id)
138
if not revision_id in self.get_inventory_weave():
140
raise errors.WeaveRevisionNotPresent(revision_id,
141
self.get_inventory_weave())
143
# yes, this is not suitable for adding with ghosts.
144
self.add_inventory(revision_id, inv, rev.parent_ids)
145
self._revision_store.add_revision(rev, self.get_transaction())
147
def _add_revision_text(self, revision_id, text):
148
revision = self._revision_store._serializer.read_revision_from_string(
150
self._revision_store._add_revision(revision, StringIO(text),
151
self.get_transaction())
154
def _all_possible_ids(self):
155
"""Return all the possible revisions that we could find."""
156
return self.get_inventory_weave().versions()
158
def all_revision_ids(self):
159
"""Returns a list of all the revision ids in the repository.
161
This is deprecated because code should generally work on the graph
162
reachable from a particular revision, and ignore any other revisions
163
that might be present. There is no direct replacement method.
165
return self._all_revision_ids()
168
def _all_revision_ids(self):
169
"""Returns a list of all the revision ids in the repository.
171
These are in as much topological order as the underlying store can
172
present: for weaves ghosts may lead to a lack of correctness until
173
the reweave updates the parents list.
175
if self._revision_store.text_store.listable():
176
return self._revision_store.all_revision_ids(self.get_transaction())
177
result = self._all_possible_ids()
178
# TODO: jam 20070210 Ensure that _all_possible_ids returns non-unicode
179
# ids. (It should, since _revision_store's API should change to
180
# return utf8 revision_ids)
181
return self._eliminate_revisions_not_present(result)
183
def break_lock(self):
184
"""Break a lock if one is present from another instance.
186
Uses the ui factory to ask for confirmation if the lock may be from
189
self.control_files.break_lock()
192
def _eliminate_revisions_not_present(self, revision_ids):
193
"""Check every revision id in revision_ids to see if we have it.
195
Returns a set of the present revisions.
198
for id in revision_ids:
199
if self.has_revision(id):
204
def create(a_bzrdir):
205
"""Construct the current default format repository in a_bzrdir."""
206
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
208
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
209
"""instantiate a Repository.
211
:param _format: The format of the repository on disk.
212
:param a_bzrdir: The BzrDir of the repository.
214
In the future we will have a single api for all stores for
215
getting file texts, inventories and revisions, then
216
this construct will accept instances of those things.
218
super(Repository, self).__init__()
219
self._format = _format
220
# the following are part of the public API for Repository:
221
self.bzrdir = a_bzrdir
222
self.control_files = control_files
223
self._revision_store = _revision_store
224
# backwards compatibility
225
self.weave_store = text_store
226
# not right yet - should be more semantically clear ?
228
self.control_store = control_store
229
self.control_weaves = control_store
230
# TODO: make sure to construct the right store classes, etc, depending
231
# on whether escaping is required.
232
self._warn_if_deprecated()
235
return '%s(%r)' % (self.__class__.__name__,
236
self.bzrdir.transport.base)
239
return self.control_files.is_locked()
241
def lock_write(self, token=None):
242
"""Lock this repository for writing.
244
:param token: if this is already locked, then lock_write will fail
245
unless the token matches the existing lock.
246
:returns: a token if this instance supports tokens, otherwise None.
247
:raises TokenLockingNotSupported: when a token is given but this
248
instance doesn't support using token locks.
249
:raises MismatchedToken: if the specified token doesn't match the token
250
of the existing lock.
252
A token should be passed in if you know that you have locked the object
253
some other way, and need to synchronise this object's state with that
256
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
258
return self.control_files.lock_write(token=token)
261
self.control_files.lock_read()
263
def get_physical_lock_status(self):
264
return self.control_files.get_physical_lock_status()
266
def leave_lock_in_place(self):
267
"""Tell this repository not to release the physical lock when this
270
If lock_write doesn't return a token, then this method is not supported.
272
self.control_files.leave_in_place()
274
def dont_leave_lock_in_place(self):
275
"""Tell this repository to release the physical lock when this
276
object is unlocked, even if it didn't originally acquire it.
278
If lock_write doesn't return a token, then this method is not supported.
280
self.control_files.dont_leave_in_place()
283
def gather_stats(self, revid=None, committers=None):
284
"""Gather statistics from a revision id.
286
:param revid: The revision id to gather statistics from, if None, then
287
no revision specific statistics are gathered.
288
:param committers: Optional parameter controlling whether to grab
289
a count of committers from the revision specific statistics.
290
:return: A dictionary of statistics. Currently this contains:
291
committers: The number of committers if requested.
292
firstrev: A tuple with timestamp, timezone for the penultimate left
293
most ancestor of revid, if revid is not the NULL_REVISION.
294
latestrev: A tuple with timestamp, timezone for revid, if revid is
295
not the NULL_REVISION.
296
revisions: The total revision count in the repository.
297
size: An estimate disk size of the repository in bytes.
300
if revid and committers:
301
result['committers'] = 0
302
if revid and revid != _mod_revision.NULL_REVISION:
304
all_committers = set()
305
revisions = self.get_ancestry(revid)
306
# pop the leading None
308
first_revision = None
310
# ignore the revisions in the middle - just grab first and last
311
revisions = revisions[0], revisions[-1]
312
for revision in self.get_revisions(revisions):
313
if not first_revision:
314
first_revision = revision
316
all_committers.add(revision.committer)
317
last_revision = revision
319
result['committers'] = len(all_committers)
320
result['firstrev'] = (first_revision.timestamp,
321
first_revision.timezone)
322
result['latestrev'] = (last_revision.timestamp,
323
last_revision.timezone)
325
# now gather global repository information
326
if self.bzrdir.root_transport.listable():
327
c, t = self._revision_store.total_size(self.get_transaction())
328
result['revisions'] = c
333
def missing_revision_ids(self, other, revision_id=None):
334
"""Return the revision ids that other has that this does not.
336
These are returned in topological order.
338
revision_id: only return revision ids included by revision_id.
340
revision_id = osutils.safe_revision_id(revision_id)
341
return InterRepository.get(other, self).missing_revision_ids(revision_id)
345
"""Open the repository rooted at base.
347
For instance, if the repository is at URL/.bzr/repository,
348
Repository.open(URL) -> a Repository instance.
350
control = bzrdir.BzrDir.open(base)
351
return control.open_repository()
353
def copy_content_into(self, destination, revision_id=None):
354
"""Make a complete copy of the content in self into destination.
356
This is a destructive operation! Do not use it on existing
359
revision_id = osutils.safe_revision_id(revision_id)
360
return InterRepository.get(self, destination).copy_content(revision_id)
362
def fetch(self, source, revision_id=None, pb=None):
363
"""Fetch the content required to construct revision_id from source.
365
If revision_id is None all content is copied.
367
revision_id = osutils.safe_revision_id(revision_id)
368
inter = InterRepository.get(source, self)
370
return inter.fetch(revision_id=revision_id, pb=pb)
371
except NotImplementedError:
372
raise errors.IncompatibleRepositories(source, self)
374
def create_bundle(self, target, base, fileobj, format=None):
375
return serializer.write_bundle(self, target, base, fileobj, format)
377
def get_commit_builder(self, branch, parents, config, timestamp=None,
378
timezone=None, committer=None, revprops=None,
380
"""Obtain a CommitBuilder for this repository.
382
:param branch: Branch to commit to.
383
:param parents: Revision ids of the parents of the new revision.
384
:param config: Configuration to use.
385
:param timestamp: Optional timestamp recorded for commit.
386
:param timezone: Optional timezone for timestamp.
387
:param committer: Optional committer to set for commit.
388
:param revprops: Optional dictionary of revision properties.
389
:param revision_id: Optional revision id.
391
revision_id = osutils.safe_revision_id(revision_id)
392
return _CommitBuilder(self, parents, config, timestamp, timezone,
393
committer, revprops, revision_id)
396
self.control_files.unlock()
399
def clone(self, a_bzrdir, revision_id=None):
400
"""Clone this repository into a_bzrdir using the current format.
402
Currently no check is made that the format of this repository and
403
the bzrdir format are compatible. FIXME RBC 20060201.
405
:return: The newly created destination repository.
407
# TODO: deprecate after 0.16; cloning this with all its settings is
408
# probably not very useful -- mbp 20070423
409
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
410
self.copy_content_into(dest_repo, revision_id)
414
def sprout(self, to_bzrdir, revision_id=None):
415
"""Create a descendent repository for new development.
417
Unlike clone, this does not copy the settings of the repository.
419
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
420
dest_repo.fetch(self, revision_id=revision_id)
423
def _create_sprouting_repo(self, a_bzrdir, shared):
424
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
425
# use target default format.
426
dest_repo = a_bzrdir.create_repository()
428
# Most control formats need the repository to be specifically
429
# created, but on some old all-in-one formats it's not needed
431
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
432
except errors.UninitializableFormat:
433
dest_repo = a_bzrdir.open_repository()
437
def has_revision(self, revision_id):
438
"""True if this repository has a copy of the revision."""
439
revision_id = osutils.safe_revision_id(revision_id)
440
return self._revision_store.has_revision_id(revision_id,
441
self.get_transaction())
444
def get_revision_reconcile(self, revision_id):
445
"""'reconcile' helper routine that allows access to a revision always.
447
This variant of get_revision does not cross check the weave graph
448
against the revision one as get_revision does: but it should only
449
be used by reconcile, or reconcile-alike commands that are correcting
450
or testing the revision graph.
452
if not revision_id or not isinstance(revision_id, basestring):
453
raise errors.InvalidRevisionId(revision_id=revision_id,
455
return self.get_revisions([revision_id])[0]
458
def get_revisions(self, revision_ids):
459
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
460
revs = self._revision_store.get_revisions(revision_ids,
461
self.get_transaction())
463
assert not isinstance(rev.revision_id, unicode)
464
for parent_id in rev.parent_ids:
465
assert not isinstance(parent_id, unicode)
469
def get_revision_xml(self, revision_id):
470
# TODO: jam 20070210 This shouldn't be necessary since get_revision
471
# would have already do it.
472
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
473
revision_id = osutils.safe_revision_id(revision_id)
474
rev = self.get_revision(revision_id)
476
# the current serializer..
477
self._revision_store._serializer.write_revision(rev, rev_tmp)
479
return rev_tmp.getvalue()
482
def get_revision(self, revision_id):
483
"""Return the Revision object for a named revision"""
484
# TODO: jam 20070210 get_revision_reconcile should do this for us
485
revision_id = osutils.safe_revision_id(revision_id)
486
r = self.get_revision_reconcile(revision_id)
487
# weave corruption can lead to absent revision markers that should be
489
# the following test is reasonably cheap (it needs a single weave read)
490
# and the weave is cached in read transactions. In write transactions
491
# it is not cached but typically we only read a small number of
492
# revisions. For knits when they are introduced we will probably want
493
# to ensure that caching write transactions are in use.
494
inv = self.get_inventory_weave()
495
self._check_revision_parents(r, inv)
499
def get_deltas_for_revisions(self, revisions):
500
"""Produce a generator of revision deltas.
502
Note that the input is a sequence of REVISIONS, not revision_ids.
503
Trees will be held in memory until the generator exits.
504
Each delta is relative to the revision's lefthand predecessor.
506
required_trees = set()
507
for revision in revisions:
508
required_trees.add(revision.revision_id)
509
required_trees.update(revision.parent_ids[:1])
510
trees = dict((t.get_revision_id(), t) for
511
t in self.revision_trees(required_trees))
512
for revision in revisions:
513
if not revision.parent_ids:
514
old_tree = self.revision_tree(None)
516
old_tree = trees[revision.parent_ids[0]]
517
yield trees[revision.revision_id].changes_from(old_tree)
520
def get_revision_delta(self, revision_id):
521
"""Return the delta for one revision.
523
The delta is relative to the left-hand predecessor of the
526
r = self.get_revision(revision_id)
527
return list(self.get_deltas_for_revisions([r]))[0]
529
def _check_revision_parents(self, revision, inventory):
530
"""Private to Repository and Fetch.
532
This checks the parentage of revision in an inventory weave for
533
consistency and is only applicable to inventory-weave-for-ancestry
534
using repository formats & fetchers.
536
weave_parents = inventory.get_parents(revision.revision_id)
537
weave_names = inventory.versions()
538
for parent_id in revision.parent_ids:
539
if parent_id in weave_names:
540
# this parent must not be a ghost.
541
if not parent_id in weave_parents:
543
raise errors.CorruptRepository(self)
546
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
547
revision_id = osutils.safe_revision_id(revision_id)
548
signature = gpg_strategy.sign(plaintext)
549
self._revision_store.add_revision_signature_text(revision_id,
551
self.get_transaction())
553
def fileids_altered_by_revision_ids(self, revision_ids):
554
"""Find the file ids and versions affected by revisions.
556
:param revisions: an iterable containing revision ids.
557
:return: a dictionary mapping altered file-ids to an iterable of
558
revision_ids. Each altered file-ids has the exact revision_ids that
559
altered it listed explicitly.
561
assert self._serializer.support_altered_by_hack, \
562
("fileids_altered_by_revision_ids only supported for branches "
563
"which store inventory as unnested xml, not on %r" % self)
564
selected_revision_ids = set(osutils.safe_revision_id(r)
565
for r in revision_ids)
566
w = self.get_inventory_weave()
569
# this code needs to read every new line in every inventory for the
570
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
571
# not present in one of those inventories is unnecessary but not
572
# harmful because we are filtering by the revision id marker in the
573
# inventory lines : we only select file ids altered in one of those
574
# revisions. We don't need to see all lines in the inventory because
575
# only those added in an inventory in rev X can contain a revision=X
577
unescape_revid_cache = {}
578
unescape_fileid_cache = {}
580
# jam 20061218 In a big fetch, this handles hundreds of thousands
581
# of lines, so it has had a lot of inlining and optimizing done.
582
# Sorry that it is a little bit messy.
583
# Move several functions to be local variables, since this is a long
585
search = self._file_ids_altered_regex.search
586
unescape = _unescape_xml
587
setdefault = result.setdefault
588
pb = ui.ui_factory.nested_progress_bar()
590
for line in w.iter_lines_added_or_present_in_versions(
591
selected_revision_ids, pb=pb):
595
# One call to match.group() returning multiple items is quite a
596
# bit faster than 2 calls to match.group() each returning 1
597
file_id, revision_id = match.group('file_id', 'revision_id')
599
# Inlining the cache lookups helps a lot when you make 170,000
600
# lines and 350k ids, versus 8.4 unique ids.
601
# Using a cache helps in 2 ways:
602
# 1) Avoids unnecessary decoding calls
603
# 2) Re-uses cached strings, which helps in future set and
605
# (2) is enough that removing encoding entirely along with
606
# the cache (so we are using plain strings) results in no
607
# performance improvement.
609
revision_id = unescape_revid_cache[revision_id]
611
unescaped = unescape(revision_id)
612
unescape_revid_cache[revision_id] = unescaped
613
revision_id = unescaped
615
if revision_id in selected_revision_ids:
617
file_id = unescape_fileid_cache[file_id]
619
unescaped = unescape(file_id)
620
unescape_fileid_cache[file_id] = unescaped
622
setdefault(file_id, set()).add(revision_id)
628
def get_inventory_weave(self):
629
return self.control_weaves.get_weave('inventory',
630
self.get_transaction())
633
def get_inventory(self, revision_id):
634
"""Get Inventory object by hash."""
635
# TODO: jam 20070210 Technically we don't need to sanitize, since all
636
# called functions must sanitize.
637
revision_id = osutils.safe_revision_id(revision_id)
638
return self.deserialise_inventory(
639
revision_id, self.get_inventory_xml(revision_id))
641
def deserialise_inventory(self, revision_id, xml):
642
"""Transform the xml into an inventory object.
644
:param revision_id: The expected revision id of the inventory.
645
:param xml: A serialised inventory.
647
revision_id = osutils.safe_revision_id(revision_id)
648
result = self._serializer.read_inventory_from_string(xml)
649
result.root.revision = revision_id
652
def serialise_inventory(self, inv):
653
return self._serializer.write_inventory_to_string(inv)
655
def get_serializer_format(self):
656
return self._serializer.format_num
659
def get_inventory_xml(self, revision_id):
660
"""Get inventory XML as a file object."""
661
revision_id = osutils.safe_revision_id(revision_id)
663
assert isinstance(revision_id, str), type(revision_id)
664
iw = self.get_inventory_weave()
665
return iw.get_text(revision_id)
667
raise errors.HistoryMissing(self, 'inventory', revision_id)
670
def get_inventory_sha1(self, revision_id):
671
"""Return the sha1 hash of the inventory entry
673
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
674
revision_id = osutils.safe_revision_id(revision_id)
675
return self.get_revision(revision_id).inventory_sha1
678
def get_revision_graph(self, revision_id=None):
679
"""Return a dictionary containing the revision graph.
681
:param revision_id: The revision_id to get a graph from. If None, then
682
the entire revision graph is returned. This is a deprecated mode of
683
operation and will be removed in the future.
684
:return: a dictionary of revision_id->revision_parents_list.
686
# special case NULL_REVISION
687
if revision_id == _mod_revision.NULL_REVISION:
689
revision_id = osutils.safe_revision_id(revision_id)
690
a_weave = self.get_inventory_weave()
691
all_revisions = self._eliminate_revisions_not_present(
693
entire_graph = dict([(node, tuple(a_weave.get_parents(node))) for
694
node in all_revisions])
695
if revision_id is None:
697
elif revision_id not in entire_graph:
698
raise errors.NoSuchRevision(self, revision_id)
700
# add what can be reached from revision_id
702
pending = set([revision_id])
703
while len(pending) > 0:
705
result[node] = entire_graph[node]
706
for revision_id in result[node]:
707
if revision_id not in result:
708
pending.add(revision_id)
712
def get_revision_graph_with_ghosts(self, revision_ids=None):
713
"""Return a graph of the revisions with ghosts marked as applicable.
715
:param revision_ids: an iterable of revisions to graph or None for all.
716
:return: a Graph object with the graph reachable from revision_ids.
718
result = deprecated_graph.Graph()
720
pending = set(self.all_revision_ids())
723
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
724
# special case NULL_REVISION
725
if _mod_revision.NULL_REVISION in pending:
726
pending.remove(_mod_revision.NULL_REVISION)
727
required = set(pending)
730
revision_id = pending.pop()
732
rev = self.get_revision(revision_id)
733
except errors.NoSuchRevision:
734
if revision_id in required:
737
result.add_ghost(revision_id)
739
for parent_id in rev.parent_ids:
740
# is this queued or done ?
741
if (parent_id not in pending and
742
parent_id not in done):
744
pending.add(parent_id)
745
result.add_node(revision_id, rev.parent_ids)
746
done.add(revision_id)
749
def _get_history_vf(self):
750
"""Get a versionedfile whose history graph reflects all revisions.
752
For weave repositories, this is the inventory weave.
754
return self.get_inventory_weave()
756
def iter_reverse_revision_history(self, revision_id):
757
"""Iterate backwards through revision ids in the lefthand history
759
:param revision_id: The revision id to start with. All its lefthand
760
ancestors will be traversed.
762
revision_id = osutils.safe_revision_id(revision_id)
763
if revision_id in (None, _mod_revision.NULL_REVISION):
765
next_id = revision_id
766
versionedfile = self._get_history_vf()
769
parents = versionedfile.get_parents(next_id)
770
if len(parents) == 0:
776
def get_revision_inventory(self, revision_id):
777
"""Return inventory of a past revision."""
778
# TODO: Unify this with get_inventory()
779
# bzr 0.0.6 and later imposes the constraint that the inventory_id
780
# must be the same as its revision, so this is trivial.
781
if revision_id is None:
782
# This does not make sense: if there is no revision,
783
# then it is the current tree inventory surely ?!
784
# and thus get_root_id() is something that looks at the last
785
# commit on the branch, and the get_root_id is an inventory check.
786
raise NotImplementedError
787
# return Inventory(self.get_root_id())
789
return self.get_inventory(revision_id)
793
"""Return True if this repository is flagged as a shared repository."""
794
raise NotImplementedError(self.is_shared)
797
def reconcile(self, other=None, thorough=False):
798
"""Reconcile this repository."""
799
from bzrlib.reconcile import RepoReconciler
800
reconciler = RepoReconciler(self, thorough=thorough)
801
reconciler.reconcile()
805
def revision_tree(self, revision_id):
806
"""Return Tree for a revision on this branch.
808
`revision_id` may be None for the empty tree revision.
810
# TODO: refactor this to use an existing revision object
811
# so we don't need to read it in twice.
812
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
813
return RevisionTree(self, Inventory(root_id=None),
814
_mod_revision.NULL_REVISION)
816
revision_id = osutils.safe_revision_id(revision_id)
817
inv = self.get_revision_inventory(revision_id)
818
return RevisionTree(self, inv, revision_id)
821
def revision_trees(self, revision_ids):
822
"""Return Tree for a revision on this branch.
824
`revision_id` may not be None or 'null:'"""
825
assert None not in revision_ids
826
assert _mod_revision.NULL_REVISION not in revision_ids
827
texts = self.get_inventory_weave().get_texts(revision_ids)
828
for text, revision_id in zip(texts, revision_ids):
829
inv = self.deserialise_inventory(revision_id, text)
830
yield RevisionTree(self, inv, revision_id)
833
def get_ancestry(self, revision_id, topo_sorted=True):
834
"""Return a list of revision-ids integrated by a revision.
836
The first element of the list is always None, indicating the origin
837
revision. This might change when we have history horizons, or
838
perhaps we should have a new API.
840
This is topologically sorted.
842
if _mod_revision.is_null(revision_id):
844
revision_id = osutils.safe_revision_id(revision_id)
845
if not self.has_revision(revision_id):
846
raise errors.NoSuchRevision(self, revision_id)
847
w = self.get_inventory_weave()
848
candidates = w.get_ancestry(revision_id, topo_sorted)
849
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
852
"""Compress the data within the repository.
854
This operation only makes sense for some repository types. For other
855
types it should be a no-op that just returns.
857
This stub method does not require a lock, but subclasses should use
858
@needs_write_lock as this is a long running call its reasonable to
859
implicitly lock for the user.
863
def print_file(self, file, revision_id):
864
"""Print `file` to stdout.
866
FIXME RBC 20060125 as John Meinel points out this is a bad api
867
- it writes to stdout, it assumes that that is valid etc. Fix
868
by creating a new more flexible convenience function.
870
revision_id = osutils.safe_revision_id(revision_id)
871
tree = self.revision_tree(revision_id)
872
# use inventory as it was in that revision
873
file_id = tree.inventory.path2id(file)
875
# TODO: jam 20060427 Write a test for this code path
876
# it had a bug in it, and was raising the wrong
878
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
879
tree.print_file(file_id)
881
def get_transaction(self):
882
return self.control_files.get_transaction()
884
def revision_parents(self, revision_id):
885
revision_id = osutils.safe_revision_id(revision_id)
886
return self.get_inventory_weave().parent_names(revision_id)
888
def get_parents(self, revision_ids):
889
"""See StackedParentsProvider.get_parents"""
891
for revision_id in revision_ids:
892
if revision_id == _mod_revision.NULL_REVISION:
896
parents = self.get_revision(revision_id).parent_ids
897
except errors.NoSuchRevision:
900
if len(parents) == 0:
901
parents = [_mod_revision.NULL_REVISION]
902
parents_list.append(parents)
905
def _make_parents_provider(self):
908
def get_graph(self, other_repository=None):
909
"""Return the graph walker for this repository format"""
910
parents_provider = self._make_parents_provider()
911
if (other_repository is not None and
912
other_repository.bzrdir.transport.base !=
913
self.bzrdir.transport.base):
914
parents_provider = graph._StackedParentsProvider(
915
[parents_provider, other_repository._make_parents_provider()])
916
return graph.Graph(parents_provider)
919
def set_make_working_trees(self, new_value):
920
"""Set the policy flag for making working trees when creating branches.
922
This only applies to branches that use this repository.
924
The default is 'True'.
925
:param new_value: True to restore the default, False to disable making
928
raise NotImplementedError(self.set_make_working_trees)
930
def make_working_trees(self):
931
"""Returns the policy for making working trees on new branches."""
932
raise NotImplementedError(self.make_working_trees)
935
def sign_revision(self, revision_id, gpg_strategy):
936
revision_id = osutils.safe_revision_id(revision_id)
937
plaintext = Testament.from_revision(self, revision_id).as_short_text()
938
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
941
def has_signature_for_revision_id(self, revision_id):
942
"""Query for a revision signature for revision_id in the repository."""
943
revision_id = osutils.safe_revision_id(revision_id)
944
return self._revision_store.has_signature(revision_id,
945
self.get_transaction())
948
def get_signature_text(self, revision_id):
949
"""Return the text for a signature."""
950
revision_id = osutils.safe_revision_id(revision_id)
951
return self._revision_store.get_signature_text(revision_id,
952
self.get_transaction())
955
def check(self, revision_ids):
956
"""Check consistency of all history of given revision_ids.
958
Different repository implementations should override _check().
960
:param revision_ids: A non-empty list of revision_ids whose ancestry
961
will be checked. Typically the last revision_id of a branch.
964
raise ValueError("revision_ids must be non-empty in %s.check"
966
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
967
return self._check(revision_ids)
969
def _check(self, revision_ids):
970
result = check.Check(self)
974
def _warn_if_deprecated(self):
975
global _deprecation_warning_done
976
if _deprecation_warning_done:
978
_deprecation_warning_done = True
979
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
980
% (self._format, self.bzrdir.transport.base))
982
def supports_rich_root(self):
983
return self._format.rich_root_data
985
def _check_ascii_revisionid(self, revision_id, method):
986
"""Private helper for ascii-only repositories."""
987
# weave repositories refuse to store revisionids that are non-ascii.
988
if revision_id is not None:
989
# weaves require ascii revision ids.
990
if isinstance(revision_id, unicode):
992
revision_id.encode('ascii')
993
except UnicodeEncodeError:
994
raise errors.NonAsciiRevisionId(method, self)
997
revision_id.decode('ascii')
998
except UnicodeDecodeError:
999
raise errors.NonAsciiRevisionId(method, self)
1003
# remove these delegates a while after bzr 0.15
1004
def __make_delegated(name, from_module):
1005
def _deprecated_repository_forwarder():
1006
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1007
% (name, from_module),
1010
m = __import__(from_module, globals(), locals(), [name])
1012
return getattr(m, name)
1013
except AttributeError:
1014
raise AttributeError('module %s has no name %s'
1016
globals()[name] = _deprecated_repository_forwarder
1019
'AllInOneRepository',
1020
'WeaveMetaDirRepository',
1021
'PreSplitOutRepositoryFormat',
1022
'RepositoryFormat4',
1023
'RepositoryFormat5',
1024
'RepositoryFormat6',
1025
'RepositoryFormat7',
1027
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1031
'RepositoryFormatKnit',
1032
'RepositoryFormatKnit1',
1034
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1037
def install_revision(repository, rev, revision_tree):
1038
"""Install all revision data into a repository."""
1039
present_parents = []
1041
for p_id in rev.parent_ids:
1042
if repository.has_revision(p_id):
1043
present_parents.append(p_id)
1044
parent_trees[p_id] = repository.revision_tree(p_id)
1046
parent_trees[p_id] = repository.revision_tree(None)
1048
inv = revision_tree.inventory
1049
entries = inv.iter_entries()
1050
# backwards compatability hack: skip the root id.
1051
if not repository.supports_rich_root():
1052
path, root = entries.next()
1053
if root.revision != rev.revision_id:
1054
raise errors.IncompatibleRevision(repr(repository))
1055
# Add the texts that are not already present
1056
for path, ie in entries:
1057
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1058
repository.get_transaction())
1059
if ie.revision not in w:
1061
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1062
# with InventoryEntry.find_previous_heads(). if it is, then there
1063
# is a latent bug here where the parents may have ancestors of each
1065
for revision, tree in parent_trees.iteritems():
1066
if ie.file_id not in tree:
1068
parent_id = tree.inventory[ie.file_id].revision
1069
if parent_id in text_parents:
1071
text_parents.append(parent_id)
1073
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1074
repository.get_transaction())
1075
lines = revision_tree.get_file(ie.file_id).readlines()
1076
vfile.add_lines(rev.revision_id, text_parents, lines)
1078
# install the inventory
1079
repository.add_inventory(rev.revision_id, inv, present_parents)
1080
except errors.RevisionAlreadyPresent:
1082
repository.add_revision(rev.revision_id, rev, inv)
1085
class MetaDirRepository(Repository):
1086
"""Repositories in the new meta-dir layout."""
1088
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1089
super(MetaDirRepository, self).__init__(_format,
1095
dir_mode = self.control_files._dir_mode
1096
file_mode = self.control_files._file_mode
1099
def is_shared(self):
1100
"""Return True if this repository is flagged as a shared repository."""
1101
return self.control_files._transport.has('shared-storage')
1104
def set_make_working_trees(self, new_value):
1105
"""Set the policy flag for making working trees when creating branches.
1107
This only applies to branches that use this repository.
1109
The default is 'True'.
1110
:param new_value: True to restore the default, False to disable making
1115
self.control_files._transport.delete('no-working-trees')
1116
except errors.NoSuchFile:
1119
self.control_files.put_utf8('no-working-trees', '')
1121
def make_working_trees(self):
1122
"""Returns the policy for making working trees on new branches."""
1123
return not self.control_files._transport.has('no-working-trees')
1126
class RepositoryFormatRegistry(registry.Registry):
1127
"""Registry of RepositoryFormats.
1130
def get(self, format_string):
1131
r = registry.Registry.get(self, format_string)
1137
format_registry = RepositoryFormatRegistry()
1138
"""Registry of formats, indexed by their identifying format string.
1140
This can contain either format instances themselves, or classes/factories that
1141
can be called to obtain one.
1145
#####################################################################
1146
# Repository Formats
1148
class RepositoryFormat(object):
1149
"""A repository format.
1151
Formats provide three things:
1152
* An initialization routine to construct repository data on disk.
1153
* a format string which is used when the BzrDir supports versioned
1155
* an open routine which returns a Repository instance.
1157
Formats are placed in an dict by their format string for reference
1158
during opening. These should be subclasses of RepositoryFormat
1161
Once a format is deprecated, just deprecate the initialize and open
1162
methods on the format class. Do not deprecate the object, as the
1163
object will be created every system load.
1165
Common instance attributes:
1166
_matchingbzrdir - the bzrdir format that the repository format was
1167
originally written to work with. This can be used if manually
1168
constructing a bzrdir and repository, or more commonly for test suite
1173
return "<%s>" % self.__class__.__name__
1175
def __eq__(self, other):
1176
# format objects are generally stateless
1177
return isinstance(other, self.__class__)
1179
def __ne__(self, other):
1180
return not self == other
1183
def find_format(klass, a_bzrdir):
1184
"""Return the format for the repository object in a_bzrdir.
1186
This is used by bzr native formats that have a "format" file in
1187
the repository. Other methods may be used by different types of
1191
transport = a_bzrdir.get_repository_transport(None)
1192
format_string = transport.get("format").read()
1193
return format_registry.get(format_string)
1194
except errors.NoSuchFile:
1195
raise errors.NoRepositoryPresent(a_bzrdir)
1197
raise errors.UnknownFormatError(format=format_string)
1200
def register_format(klass, format):
1201
format_registry.register(format.get_format_string(), format)
1204
def unregister_format(klass, format):
1205
format_registry.remove(format.get_format_string())
1208
def get_default_format(klass):
1209
"""Return the current default format."""
1210
from bzrlib import bzrdir
1211
return bzrdir.format_registry.make_bzrdir('default').repository_format
1213
def _get_control_store(self, repo_transport, control_files):
1214
"""Return the control store for this repository."""
1215
raise NotImplementedError(self._get_control_store)
1217
def get_format_string(self):
1218
"""Return the ASCII format string that identifies this format.
1220
Note that in pre format ?? repositories the format string is
1221
not permitted nor written to disk.
1223
raise NotImplementedError(self.get_format_string)
1225
def get_format_description(self):
1226
"""Return the short description for this format."""
1227
raise NotImplementedError(self.get_format_description)
1229
def _get_revision_store(self, repo_transport, control_files):
1230
"""Return the revision store object for this a_bzrdir."""
1231
raise NotImplementedError(self._get_revision_store)
1233
def _get_text_rev_store(self,
1240
"""Common logic for getting a revision store for a repository.
1242
see self._get_revision_store for the subclass-overridable method to
1243
get the store for a repository.
1245
from bzrlib.store.revision.text import TextRevisionStore
1246
dir_mode = control_files._dir_mode
1247
file_mode = control_files._file_mode
1248
text_store = TextStore(transport.clone(name),
1250
compressed=compressed,
1252
file_mode=file_mode)
1253
_revision_store = TextRevisionStore(text_store, serializer)
1254
return _revision_store
1256
# TODO: this shouldn't be in the base class, it's specific to things that
1257
# use weaves or knits -- mbp 20070207
1258
def _get_versioned_file_store(self,
1263
versionedfile_class=None,
1264
versionedfile_kwargs={},
1266
if versionedfile_class is None:
1267
versionedfile_class = self._versionedfile_class
1268
weave_transport = control_files._transport.clone(name)
1269
dir_mode = control_files._dir_mode
1270
file_mode = control_files._file_mode
1271
return VersionedFileStore(weave_transport, prefixed=prefixed,
1273
file_mode=file_mode,
1274
versionedfile_class=versionedfile_class,
1275
versionedfile_kwargs=versionedfile_kwargs,
1278
def initialize(self, a_bzrdir, shared=False):
1279
"""Initialize a repository of this format in a_bzrdir.
1281
:param a_bzrdir: The bzrdir to put the new repository in it.
1282
:param shared: The repository should be initialized as a sharable one.
1283
:returns: The new repository object.
1285
This may raise UninitializableFormat if shared repository are not
1286
compatible the a_bzrdir.
1288
raise NotImplementedError(self.initialize)
1290
def is_supported(self):
1291
"""Is this format supported?
1293
Supported formats must be initializable and openable.
1294
Unsupported formats may not support initialization or committing or
1295
some other features depending on the reason for not being supported.
1299
def check_conversion_target(self, target_format):
1300
raise NotImplementedError(self.check_conversion_target)
1302
def open(self, a_bzrdir, _found=False):
1303
"""Return an instance of this format for the bzrdir a_bzrdir.
1305
_found is a private parameter, do not use it.
1307
raise NotImplementedError(self.open)
1310
class MetaDirRepositoryFormat(RepositoryFormat):
1311
"""Common base class for the new repositories using the metadir layout."""
1313
rich_root_data = False
1314
supports_tree_reference = False
1315
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1318
super(MetaDirRepositoryFormat, self).__init__()
1320
def _create_control_files(self, a_bzrdir):
1321
"""Create the required files and the initial control_files object."""
1322
# FIXME: RBC 20060125 don't peek under the covers
1323
# NB: no need to escape relative paths that are url safe.
1324
repository_transport = a_bzrdir.get_repository_transport(self)
1325
control_files = lockable_files.LockableFiles(repository_transport,
1326
'lock', lockdir.LockDir)
1327
control_files.create_lock()
1328
return control_files
1330
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1331
"""Upload the initial blank content."""
1332
control_files = self._create_control_files(a_bzrdir)
1333
control_files.lock_write()
1335
control_files._transport.mkdir_multi(dirs,
1336
mode=control_files._dir_mode)
1337
for file, content in files:
1338
control_files.put(file, content)
1339
for file, content in utf8_files:
1340
control_files.put_utf8(file, content)
1342
control_files.put_utf8('shared-storage', '')
1344
control_files.unlock()
1347
# formats which have no format string are not discoverable
1348
# and not independently creatable, so are not registered. They're
1349
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1350
# needed, it's constructed directly by the BzrDir. Non-native formats where
1351
# the repository is not separately opened are similar.
1353
format_registry.register_lazy(
1354
'Bazaar-NG Repository format 7',
1355
'bzrlib.repofmt.weaverepo',
1358
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1359
# default control directory format
1361
format_registry.register_lazy(
1362
'Bazaar-NG Knit Repository Format 1',
1363
'bzrlib.repofmt.knitrepo',
1364
'RepositoryFormatKnit1',
1366
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1368
format_registry.register_lazy(
1369
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1370
'bzrlib.repofmt.knitrepo',
1371
'RepositoryFormatKnit3',
1375
class InterRepository(InterObject):
1376
"""This class represents operations taking place between two repositories.
1378
Its instances have methods like copy_content and fetch, and contain
1379
references to the source and target repositories these operations can be
1382
Often we will provide convenience methods on 'repository' which carry out
1383
operations with another repository - they will always forward to
1384
InterRepository.get(other).method_name(parameters).
1388
"""The available optimised InterRepository types."""
1390
def copy_content(self, revision_id=None):
1391
raise NotImplementedError(self.copy_content)
1393
def fetch(self, revision_id=None, pb=None):
1394
"""Fetch the content required to construct revision_id.
1396
The content is copied from self.source to self.target.
1398
:param revision_id: if None all content is copied, if NULL_REVISION no
1400
:param pb: optional progress bar to use for progress reports. If not
1401
provided a default one will be created.
1403
Returns the copied revision count and the failed revisions in a tuple:
1406
raise NotImplementedError(self.fetch)
1409
def missing_revision_ids(self, revision_id=None):
1410
"""Return the revision ids that source has that target does not.
1412
These are returned in topological order.
1414
:param revision_id: only return revision ids included by this
1417
# generic, possibly worst case, slow code path.
1418
target_ids = set(self.target.all_revision_ids())
1419
if revision_id is not None:
1420
# TODO: jam 20070210 InterRepository is internal enough that it
1421
# should assume revision_ids are already utf-8
1422
revision_id = osutils.safe_revision_id(revision_id)
1423
source_ids = self.source.get_ancestry(revision_id)
1424
assert source_ids[0] is None
1427
source_ids = self.source.all_revision_ids()
1428
result_set = set(source_ids).difference(target_ids)
1429
# this may look like a no-op: its not. It preserves the ordering
1430
# other_ids had while only returning the members from other_ids
1431
# that we've decided we need.
1432
return [rev_id for rev_id in source_ids if rev_id in result_set]
1435
class InterSameDataRepository(InterRepository):
1436
"""Code for converting between repositories that represent the same data.
1438
Data format and model must match for this to work.
1442
def _get_repo_format_to_test(self):
1443
"""Repository format for testing with."""
1444
return RepositoryFormat.get_default_format()
1447
def is_compatible(source, target):
1448
if source.supports_rich_root() != target.supports_rich_root():
1450
if source._serializer != target._serializer:
1455
def copy_content(self, revision_id=None):
1456
"""Make a complete copy of the content in self into destination.
1458
This copies both the repository's revision data, and configuration information
1459
such as the make_working_trees setting.
1461
This is a destructive operation! Do not use it on existing
1464
:param revision_id: Only copy the content needed to construct
1465
revision_id and its parents.
1468
self.target.set_make_working_trees(self.source.make_working_trees())
1469
except NotImplementedError:
1471
# TODO: jam 20070210 This is fairly internal, so we should probably
1472
# just assert that revision_id is not unicode.
1473
revision_id = osutils.safe_revision_id(revision_id)
1474
# but don't bother fetching if we have the needed data now.
1475
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1476
self.target.has_revision(revision_id)):
1478
self.target.fetch(self.source, revision_id=revision_id)
1481
def fetch(self, revision_id=None, pb=None):
1482
"""See InterRepository.fetch()."""
1483
from bzrlib.fetch import GenericRepoFetcher
1484
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1485
self.source, self.source._format, self.target,
1486
self.target._format)
1487
# TODO: jam 20070210 This should be an assert, not a translate
1488
revision_id = osutils.safe_revision_id(revision_id)
1489
f = GenericRepoFetcher(to_repository=self.target,
1490
from_repository=self.source,
1491
last_revision=revision_id,
1493
return f.count_copied, f.failed_revisions
1496
class InterWeaveRepo(InterSameDataRepository):
1497
"""Optimised code paths between Weave based repositories."""
1500
def _get_repo_format_to_test(self):
1501
from bzrlib.repofmt import weaverepo
1502
return weaverepo.RepositoryFormat7()
1505
def is_compatible(source, target):
1506
"""Be compatible with known Weave formats.
1508
We don't test for the stores being of specific types because that
1509
could lead to confusing results, and there is no need to be
1512
from bzrlib.repofmt.weaverepo import (
1518
return (isinstance(source._format, (RepositoryFormat5,
1520
RepositoryFormat7)) and
1521
isinstance(target._format, (RepositoryFormat5,
1523
RepositoryFormat7)))
1524
except AttributeError:
1528
def copy_content(self, revision_id=None):
1529
"""See InterRepository.copy_content()."""
1530
# weave specific optimised path:
1531
# TODO: jam 20070210 Internal, should be an assert, not translate
1532
revision_id = osutils.safe_revision_id(revision_id)
1534
self.target.set_make_working_trees(self.source.make_working_trees())
1535
except NotImplementedError:
1537
# FIXME do not peek!
1538
if self.source.control_files._transport.listable():
1539
pb = ui.ui_factory.nested_progress_bar()
1541
self.target.weave_store.copy_all_ids(
1542
self.source.weave_store,
1544
from_transaction=self.source.get_transaction(),
1545
to_transaction=self.target.get_transaction())
1546
pb.update('copying inventory', 0, 1)
1547
self.target.control_weaves.copy_multi(
1548
self.source.control_weaves, ['inventory'],
1549
from_transaction=self.source.get_transaction(),
1550
to_transaction=self.target.get_transaction())
1551
self.target._revision_store.text_store.copy_all_ids(
1552
self.source._revision_store.text_store,
1557
self.target.fetch(self.source, revision_id=revision_id)
1560
def fetch(self, revision_id=None, pb=None):
1561
"""See InterRepository.fetch()."""
1562
from bzrlib.fetch import GenericRepoFetcher
1563
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1564
self.source, self.source._format, self.target, self.target._format)
1565
# TODO: jam 20070210 This should be an assert, not a translate
1566
revision_id = osutils.safe_revision_id(revision_id)
1567
f = GenericRepoFetcher(to_repository=self.target,
1568
from_repository=self.source,
1569
last_revision=revision_id,
1571
return f.count_copied, f.failed_revisions
1574
def missing_revision_ids(self, revision_id=None):
1575
"""See InterRepository.missing_revision_ids()."""
1576
# we want all revisions to satisfy revision_id in source.
1577
# but we don't want to stat every file here and there.
1578
# we want then, all revisions other needs to satisfy revision_id
1579
# checked, but not those that we have locally.
1580
# so the first thing is to get a subset of the revisions to
1581
# satisfy revision_id in source, and then eliminate those that
1582
# we do already have.
1583
# this is slow on high latency connection to self, but as as this
1584
# disk format scales terribly for push anyway due to rewriting
1585
# inventory.weave, this is considered acceptable.
1587
if revision_id is not None:
1588
source_ids = self.source.get_ancestry(revision_id)
1589
assert source_ids[0] is None
1592
source_ids = self.source._all_possible_ids()
1593
source_ids_set = set(source_ids)
1594
# source_ids is the worst possible case we may need to pull.
1595
# now we want to filter source_ids against what we actually
1596
# have in target, but don't try to check for existence where we know
1597
# we do not have a revision as that would be pointless.
1598
target_ids = set(self.target._all_possible_ids())
1599
possibly_present_revisions = target_ids.intersection(source_ids_set)
1600
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1601
required_revisions = source_ids_set.difference(actually_present_revisions)
1602
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1603
if revision_id is not None:
1604
# we used get_ancestry to determine source_ids then we are assured all
1605
# revisions referenced are present as they are installed in topological order.
1606
# and the tip revision was validated by get_ancestry.
1607
return required_topo_revisions
1609
# if we just grabbed the possibly available ids, then
1610
# we only have an estimate of whats available and need to validate
1611
# that against the revision records.
1612
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1615
class InterKnitRepo(InterSameDataRepository):
1616
"""Optimised code paths between Knit based repositories."""
1619
def _get_repo_format_to_test(self):
1620
from bzrlib.repofmt import knitrepo
1621
return knitrepo.RepositoryFormatKnit1()
1624
def is_compatible(source, target):
1625
"""Be compatible with known Knit formats.
1627
We don't test for the stores being of specific types because that
1628
could lead to confusing results, and there is no need to be
1631
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1633
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1634
isinstance(target._format, (RepositoryFormatKnit1)))
1635
except AttributeError:
1639
def fetch(self, revision_id=None, pb=None):
1640
"""See InterRepository.fetch()."""
1641
from bzrlib.fetch import KnitRepoFetcher
1642
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1643
self.source, self.source._format, self.target, self.target._format)
1644
# TODO: jam 20070210 This should be an assert, not a translate
1645
revision_id = osutils.safe_revision_id(revision_id)
1646
f = KnitRepoFetcher(to_repository=self.target,
1647
from_repository=self.source,
1648
last_revision=revision_id,
1650
return f.count_copied, f.failed_revisions
1653
def missing_revision_ids(self, revision_id=None):
1654
"""See InterRepository.missing_revision_ids()."""
1655
if revision_id is not None:
1656
source_ids = self.source.get_ancestry(revision_id)
1657
assert source_ids[0] is None
1660
source_ids = self.source._all_possible_ids()
1661
source_ids_set = set(source_ids)
1662
# source_ids is the worst possible case we may need to pull.
1663
# now we want to filter source_ids against what we actually
1664
# have in target, but don't try to check for existence where we know
1665
# we do not have a revision as that would be pointless.
1666
target_ids = set(self.target._all_possible_ids())
1667
possibly_present_revisions = target_ids.intersection(source_ids_set)
1668
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1669
required_revisions = source_ids_set.difference(actually_present_revisions)
1670
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1671
if revision_id is not None:
1672
# we used get_ancestry to determine source_ids then we are assured all
1673
# revisions referenced are present as they are installed in topological order.
1674
# and the tip revision was validated by get_ancestry.
1675
return required_topo_revisions
1677
# if we just grabbed the possibly available ids, then
1678
# we only have an estimate of whats available and need to validate
1679
# that against the revision records.
1680
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1683
class InterModel1and2(InterRepository):
1686
def _get_repo_format_to_test(self):
1690
def is_compatible(source, target):
1691
if not source.supports_rich_root() and target.supports_rich_root():
1697
def fetch(self, revision_id=None, pb=None):
1698
"""See InterRepository.fetch()."""
1699
from bzrlib.fetch import Model1toKnit2Fetcher
1700
# TODO: jam 20070210 This should be an assert, not a translate
1701
revision_id = osutils.safe_revision_id(revision_id)
1702
f = Model1toKnit2Fetcher(to_repository=self.target,
1703
from_repository=self.source,
1704
last_revision=revision_id,
1706
return f.count_copied, f.failed_revisions
1709
def copy_content(self, revision_id=None):
1710
"""Make a complete copy of the content in self into destination.
1712
This is a destructive operation! Do not use it on existing
1715
:param revision_id: Only copy the content needed to construct
1716
revision_id and its parents.
1719
self.target.set_make_working_trees(self.source.make_working_trees())
1720
except NotImplementedError:
1722
# TODO: jam 20070210 Internal, assert, don't translate
1723
revision_id = osutils.safe_revision_id(revision_id)
1724
# but don't bother fetching if we have the needed data now.
1725
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1726
self.target.has_revision(revision_id)):
1728
self.target.fetch(self.source, revision_id=revision_id)
1731
class InterKnit1and2(InterKnitRepo):
1734
def _get_repo_format_to_test(self):
1738
def is_compatible(source, target):
1739
"""Be compatible with Knit1 source and Knit3 target"""
1740
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
1742
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1743
RepositoryFormatKnit3
1744
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1745
isinstance(target._format, (RepositoryFormatKnit3)))
1746
except AttributeError:
1750
def fetch(self, revision_id=None, pb=None):
1751
"""See InterRepository.fetch()."""
1752
from bzrlib.fetch import Knit1to2Fetcher
1753
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1754
self.source, self.source._format, self.target,
1755
self.target._format)
1756
# TODO: jam 20070210 This should be an assert, not a translate
1757
revision_id = osutils.safe_revision_id(revision_id)
1758
f = Knit1to2Fetcher(to_repository=self.target,
1759
from_repository=self.source,
1760
last_revision=revision_id,
1762
return f.count_copied, f.failed_revisions
1765
class InterRemoteRepository(InterRepository):
1766
"""Code for converting between RemoteRepository objects.
1768
This just gets an non-remote repository from the RemoteRepository, and calls
1769
InterRepository.get again.
1772
def __init__(self, source, target):
1773
if isinstance(source, remote.RemoteRepository):
1774
source._ensure_real()
1775
real_source = source._real_repository
1777
real_source = source
1778
if isinstance(target, remote.RemoteRepository):
1779
target._ensure_real()
1780
real_target = target._real_repository
1782
real_target = target
1783
self.real_inter = InterRepository.get(real_source, real_target)
1786
def is_compatible(source, target):
1787
if isinstance(source, remote.RemoteRepository):
1789
if isinstance(target, remote.RemoteRepository):
1793
def copy_content(self, revision_id=None):
1794
self.real_inter.copy_content(revision_id=revision_id)
1796
def fetch(self, revision_id=None, pb=None):
1797
self.real_inter.fetch(revision_id=revision_id, pb=pb)
1800
def _get_repo_format_to_test(self):
1804
InterRepository.register_optimiser(InterSameDataRepository)
1805
InterRepository.register_optimiser(InterWeaveRepo)
1806
InterRepository.register_optimiser(InterKnitRepo)
1807
InterRepository.register_optimiser(InterModel1and2)
1808
InterRepository.register_optimiser(InterKnit1and2)
1809
InterRepository.register_optimiser(InterRemoteRepository)
1812
class CopyConverter(object):
1813
"""A repository conversion tool which just performs a copy of the content.
1815
This is slow but quite reliable.
1818
def __init__(self, target_format):
1819
"""Create a CopyConverter.
1821
:param target_format: The format the resulting repository should be.
1823
self.target_format = target_format
1825
def convert(self, repo, pb):
1826
"""Perform the conversion of to_convert, giving feedback via pb.
1828
:param to_convert: The disk object to convert.
1829
:param pb: a progress bar to use for progress information.
1834
# this is only useful with metadir layouts - separated repo content.
1835
# trigger an assertion if not such
1836
repo._format.get_format_string()
1837
self.repo_dir = repo.bzrdir
1838
self.step('Moving repository to repository.backup')
1839
self.repo_dir.transport.move('repository', 'repository.backup')
1840
backup_transport = self.repo_dir.transport.clone('repository.backup')
1841
repo._format.check_conversion_target(self.target_format)
1842
self.source_repo = repo._format.open(self.repo_dir,
1844
_override_transport=backup_transport)
1845
self.step('Creating new repository')
1846
converted = self.target_format.initialize(self.repo_dir,
1847
self.source_repo.is_shared())
1848
converted.lock_write()
1850
self.step('Copying content into repository.')
1851
self.source_repo.copy_content_into(converted)
1854
self.step('Deleting old repository content.')
1855
self.repo_dir.transport.delete_tree('repository.backup')
1856
self.pb.note('repository converted')
1858
def step(self, message):
1859
"""Update the pb by a step."""
1861
self.pb.update(message, self.count, self.total)
1864
class CommitBuilder(object):
1865
"""Provides an interface to build up a commit.
1867
This allows describing a tree to be committed without needing to
1868
know the internals of the format of the repository.
1871
record_root_entry = False
1872
def __init__(self, repository, parents, config, timestamp=None,
1873
timezone=None, committer=None, revprops=None,
1875
"""Initiate a CommitBuilder.
1877
:param repository: Repository to commit to.
1878
:param parents: Revision ids of the parents of the new revision.
1879
:param config: Configuration to use.
1880
:param timestamp: Optional timestamp recorded for commit.
1881
:param timezone: Optional timezone for timestamp.
1882
:param committer: Optional committer to set for commit.
1883
:param revprops: Optional dictionary of revision properties.
1884
:param revision_id: Optional revision id.
1886
self._config = config
1888
if committer is None:
1889
self._committer = self._config.username()
1891
assert isinstance(committer, basestring), type(committer)
1892
self._committer = committer
1894
self.new_inventory = Inventory(None)
1895
self._new_revision_id = osutils.safe_revision_id(revision_id)
1896
self.parents = parents
1897
self.repository = repository
1900
if revprops is not None:
1901
self._revprops.update(revprops)
1903
if timestamp is None:
1904
timestamp = time.time()
1905
# Restrict resolution to 1ms
1906
self._timestamp = round(timestamp, 3)
1908
if timezone is None:
1909
self._timezone = osutils.local_time_offset()
1911
self._timezone = int(timezone)
1913
self._generate_revision_if_needed()
1915
def commit(self, message):
1916
"""Make the actual commit.
1918
:return: The revision id of the recorded revision.
1920
rev = _mod_revision.Revision(
1921
timestamp=self._timestamp,
1922
timezone=self._timezone,
1923
committer=self._committer,
1925
inventory_sha1=self.inv_sha1,
1926
revision_id=self._new_revision_id,
1927
properties=self._revprops)
1928
rev.parent_ids = self.parents
1929
self.repository.add_revision(self._new_revision_id, rev,
1930
self.new_inventory, self._config)
1931
return self._new_revision_id
1933
def revision_tree(self):
1934
"""Return the tree that was just committed.
1936
After calling commit() this can be called to get a RevisionTree
1937
representing the newly committed tree. This is preferred to
1938
calling Repository.revision_tree() because that may require
1939
deserializing the inventory, while we already have a copy in
1942
return RevisionTree(self.repository, self.new_inventory,
1943
self._new_revision_id)
1945
def finish_inventory(self):
1946
"""Tell the builder that the inventory is finished."""
1947
if self.new_inventory.root is None:
1948
symbol_versioning.warn('Root entry should be supplied to'
1949
' record_entry_contents, as of bzr 0.10.',
1950
DeprecationWarning, stacklevel=2)
1951
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
1952
self.new_inventory.revision_id = self._new_revision_id
1953
self.inv_sha1 = self.repository.add_inventory(
1954
self._new_revision_id,
1959
def _gen_revision_id(self):
1960
"""Return new revision-id."""
1961
return generate_ids.gen_revision_id(self._config.username(),
1964
def _generate_revision_if_needed(self):
1965
"""Create a revision id if None was supplied.
1967
If the repository can not support user-specified revision ids
1968
they should override this function and raise CannotSetRevisionId
1969
if _new_revision_id is not None.
1971
:raises: CannotSetRevisionId
1973
if self._new_revision_id is None:
1974
self._new_revision_id = self._gen_revision_id()
1976
def record_entry_contents(self, ie, parent_invs, path, tree):
1977
"""Record the content of ie from tree into the commit if needed.
1979
Side effect: sets ie.revision when unchanged
1981
:param ie: An inventory entry present in the commit.
1982
:param parent_invs: The inventories of the parent revisions of the
1984
:param path: The path the entry is at in the tree.
1985
:param tree: The tree which contains this entry and should be used to
1988
if self.new_inventory.root is None and ie.parent_id is not None:
1989
symbol_versioning.warn('Root entry should be supplied to'
1990
' record_entry_contents, as of bzr 0.10.',
1991
DeprecationWarning, stacklevel=2)
1992
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
1994
self.new_inventory.add(ie)
1996
# ie.revision is always None if the InventoryEntry is considered
1997
# for committing. ie.snapshot will record the correct revision
1998
# which may be the sole parent if it is untouched.
1999
if ie.revision is not None:
2002
# In this revision format, root entries have no knit or weave
2003
if ie is self.new_inventory.root:
2004
# When serializing out to disk and back in
2005
# root.revision is always _new_revision_id
2006
ie.revision = self._new_revision_id
2008
previous_entries = ie.find_previous_heads(
2010
self.repository.weave_store,
2011
self.repository.get_transaction())
2012
# we are creating a new revision for ie in the history store
2014
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2016
def modified_directory(self, file_id, file_parents):
2017
"""Record the presence of a symbolic link.
2019
:param file_id: The file_id of the link to record.
2020
:param file_parents: The per-file parent revision ids.
2022
self._add_text_to_weave(file_id, [], file_parents.keys())
2024
def modified_reference(self, file_id, file_parents):
2025
"""Record the modification of a reference.
2027
:param file_id: The file_id of the link to record.
2028
:param file_parents: The per-file parent revision ids.
2030
self._add_text_to_weave(file_id, [], file_parents.keys())
2032
def modified_file_text(self, file_id, file_parents,
2033
get_content_byte_lines, text_sha1=None,
2035
"""Record the text of file file_id
2037
:param file_id: The file_id of the file to record the text of.
2038
:param file_parents: The per-file parent revision ids.
2039
:param get_content_byte_lines: A callable which will return the byte
2041
:param text_sha1: Optional SHA1 of the file contents.
2042
:param text_size: Optional size of the file contents.
2044
# mutter('storing text of file {%s} in revision {%s} into %r',
2045
# file_id, self._new_revision_id, self.repository.weave_store)
2046
# special case to avoid diffing on renames or
2048
if (len(file_parents) == 1
2049
and text_sha1 == file_parents.values()[0].text_sha1
2050
and text_size == file_parents.values()[0].text_size):
2051
previous_ie = file_parents.values()[0]
2052
versionedfile = self.repository.weave_store.get_weave(file_id,
2053
self.repository.get_transaction())
2054
versionedfile.clone_text(self._new_revision_id,
2055
previous_ie.revision, file_parents.keys())
2056
return text_sha1, text_size
2058
new_lines = get_content_byte_lines()
2059
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2060
# should return the SHA1 and size
2061
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2062
return osutils.sha_strings(new_lines), \
2063
sum(map(len, new_lines))
2065
def modified_link(self, file_id, file_parents, link_target):
2066
"""Record the presence of a symbolic link.
2068
:param file_id: The file_id of the link to record.
2069
:param file_parents: The per-file parent revision ids.
2070
:param link_target: Target location of this link.
2072
self._add_text_to_weave(file_id, [], file_parents.keys())
2074
def _add_text_to_weave(self, file_id, new_lines, parents):
2075
versionedfile = self.repository.weave_store.get_weave_or_empty(
2076
file_id, self.repository.get_transaction())
2077
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2078
versionedfile.clear_cache()
2081
class _CommitBuilder(CommitBuilder):
2082
"""Temporary class so old CommitBuilders are detected properly
2084
Note: CommitBuilder works whether or not root entry is recorded.
2087
record_root_entry = True
2090
class RootCommitBuilder(CommitBuilder):
2091
"""This commitbuilder actually records the root id"""
2093
record_root_entry = True
2095
def record_entry_contents(self, ie, parent_invs, path, tree):
2096
"""Record the content of ie from tree into the commit if needed.
2098
Side effect: sets ie.revision when unchanged
2100
:param ie: An inventory entry present in the commit.
2101
:param parent_invs: The inventories of the parent revisions of the
2103
:param path: The path the entry is at in the tree.
2104
:param tree: The tree which contains this entry and should be used to
2107
assert self.new_inventory.root is not None or ie.parent_id is None
2108
self.new_inventory.add(ie)
2110
# ie.revision is always None if the InventoryEntry is considered
2111
# for committing. ie.snapshot will record the correct revision
2112
# which may be the sole parent if it is untouched.
2113
if ie.revision is not None:
2116
previous_entries = ie.find_previous_heads(
2118
self.repository.weave_store,
2119
self.repository.get_transaction())
2120
# we are creating a new revision for ie in the history store
2122
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2134
def _unescaper(match, _map=_unescape_map):
2135
code = match.group(1)
2139
if not code.startswith('#'):
2141
return unichr(int(code[1:])).encode('utf8')
2147
def _unescape_xml(data):
2148
"""Unescape predefined XML entities in a string of data."""
2150
if _unescape_re is None:
2151
_unescape_re = re.compile('\&([^;]*);')
2152
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
