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)
238
def has_same_location(self, other):
239
"""Returns a boolean indicating if this repository is at the same
240
location as another repository.
242
This might return False even when two repository objects are accessing
243
the same physical repository via different URLs.
245
if self.__class__ is not other.__class__:
247
return (self.control_files._transport.base ==
248
other.control_files._transport.base)
251
return self.control_files.is_locked()
253
def lock_write(self, token=None):
254
"""Lock this repository for writing.
256
:param token: if this is already locked, then lock_write will fail
257
unless the token matches the existing lock.
258
:returns: a token if this instance supports tokens, otherwise None.
259
:raises TokenLockingNotSupported: when a token is given but this
260
instance doesn't support using token locks.
261
:raises MismatchedToken: if the specified token doesn't match the token
262
of the existing lock.
264
A token should be passed in if you know that you have locked the object
265
some other way, and need to synchronise this object's state with that
268
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
270
return self.control_files.lock_write(token=token)
273
self.control_files.lock_read()
275
def get_physical_lock_status(self):
276
return self.control_files.get_physical_lock_status()
278
def leave_lock_in_place(self):
279
"""Tell this repository not to release the physical lock when this
282
If lock_write doesn't return a token, then this method is not supported.
284
self.control_files.leave_in_place()
286
def dont_leave_lock_in_place(self):
287
"""Tell this repository to release the physical lock when this
288
object is unlocked, even if it didn't originally acquire it.
290
If lock_write doesn't return a token, then this method is not supported.
292
self.control_files.dont_leave_in_place()
295
def gather_stats(self, revid=None, committers=None):
296
"""Gather statistics from a revision id.
298
:param revid: The revision id to gather statistics from, if None, then
299
no revision specific statistics are gathered.
300
:param committers: Optional parameter controlling whether to grab
301
a count of committers from the revision specific statistics.
302
:return: A dictionary of statistics. Currently this contains:
303
committers: The number of committers if requested.
304
firstrev: A tuple with timestamp, timezone for the penultimate left
305
most ancestor of revid, if revid is not the NULL_REVISION.
306
latestrev: A tuple with timestamp, timezone for revid, if revid is
307
not the NULL_REVISION.
308
revisions: The total revision count in the repository.
309
size: An estimate disk size of the repository in bytes.
312
if revid and committers:
313
result['committers'] = 0
314
if revid and revid != _mod_revision.NULL_REVISION:
316
all_committers = set()
317
revisions = self.get_ancestry(revid)
318
# pop the leading None
320
first_revision = None
322
# ignore the revisions in the middle - just grab first and last
323
revisions = revisions[0], revisions[-1]
324
for revision in self.get_revisions(revisions):
325
if not first_revision:
326
first_revision = revision
328
all_committers.add(revision.committer)
329
last_revision = revision
331
result['committers'] = len(all_committers)
332
result['firstrev'] = (first_revision.timestamp,
333
first_revision.timezone)
334
result['latestrev'] = (last_revision.timestamp,
335
last_revision.timezone)
337
# now gather global repository information
338
if self.bzrdir.root_transport.listable():
339
c, t = self._revision_store.total_size(self.get_transaction())
340
result['revisions'] = c
345
def missing_revision_ids(self, other, revision_id=None):
346
"""Return the revision ids that other has that this does not.
348
These are returned in topological order.
350
revision_id: only return revision ids included by revision_id.
352
revision_id = osutils.safe_revision_id(revision_id)
353
return InterRepository.get(other, self).missing_revision_ids(revision_id)
357
"""Open the repository rooted at base.
359
For instance, if the repository is at URL/.bzr/repository,
360
Repository.open(URL) -> a Repository instance.
362
control = bzrdir.BzrDir.open(base)
363
return control.open_repository()
365
def copy_content_into(self, destination, revision_id=None):
366
"""Make a complete copy of the content in self into destination.
368
This is a destructive operation! Do not use it on existing
371
revision_id = osutils.safe_revision_id(revision_id)
372
return InterRepository.get(self, destination).copy_content(revision_id)
374
def fetch(self, source, revision_id=None, pb=None):
375
"""Fetch the content required to construct revision_id from source.
377
If revision_id is None all content is copied.
379
revision_id = osutils.safe_revision_id(revision_id)
380
inter = InterRepository.get(source, self)
382
return inter.fetch(revision_id=revision_id, pb=pb)
383
except NotImplementedError:
384
raise errors.IncompatibleRepositories(source, self)
386
def create_bundle(self, target, base, fileobj, format=None):
387
return serializer.write_bundle(self, target, base, fileobj, format)
389
def get_commit_builder(self, branch, parents, config, timestamp=None,
390
timezone=None, committer=None, revprops=None,
392
"""Obtain a CommitBuilder for this repository.
394
:param branch: Branch to commit to.
395
:param parents: Revision ids of the parents of the new revision.
396
:param config: Configuration to use.
397
:param timestamp: Optional timestamp recorded for commit.
398
:param timezone: Optional timezone for timestamp.
399
:param committer: Optional committer to set for commit.
400
:param revprops: Optional dictionary of revision properties.
401
:param revision_id: Optional revision id.
403
revision_id = osutils.safe_revision_id(revision_id)
404
return _CommitBuilder(self, parents, config, timestamp, timezone,
405
committer, revprops, revision_id)
408
self.control_files.unlock()
411
def clone(self, a_bzrdir, revision_id=None):
412
"""Clone this repository into a_bzrdir using the current format.
414
Currently no check is made that the format of this repository and
415
the bzrdir format are compatible. FIXME RBC 20060201.
417
:return: The newly created destination repository.
419
# TODO: deprecate after 0.16; cloning this with all its settings is
420
# probably not very useful -- mbp 20070423
421
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
422
self.copy_content_into(dest_repo, revision_id)
426
def sprout(self, to_bzrdir, revision_id=None):
427
"""Create a descendent repository for new development.
429
Unlike clone, this does not copy the settings of the repository.
431
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
432
dest_repo.fetch(self, revision_id=revision_id)
435
def _create_sprouting_repo(self, a_bzrdir, shared):
436
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
437
# use target default format.
438
dest_repo = a_bzrdir.create_repository()
440
# Most control formats need the repository to be specifically
441
# created, but on some old all-in-one formats it's not needed
443
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
444
except errors.UninitializableFormat:
445
dest_repo = a_bzrdir.open_repository()
449
def has_revision(self, revision_id):
450
"""True if this repository has a copy of the revision."""
451
revision_id = osutils.safe_revision_id(revision_id)
452
return self._revision_store.has_revision_id(revision_id,
453
self.get_transaction())
456
def get_revision_reconcile(self, revision_id):
457
"""'reconcile' helper routine that allows access to a revision always.
459
This variant of get_revision does not cross check the weave graph
460
against the revision one as get_revision does: but it should only
461
be used by reconcile, or reconcile-alike commands that are correcting
462
or testing the revision graph.
464
if not revision_id or not isinstance(revision_id, basestring):
465
raise errors.InvalidRevisionId(revision_id=revision_id,
467
return self.get_revisions([revision_id])[0]
470
def get_revisions(self, revision_ids):
471
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
472
revs = self._revision_store.get_revisions(revision_ids,
473
self.get_transaction())
475
assert not isinstance(rev.revision_id, unicode)
476
for parent_id in rev.parent_ids:
477
assert not isinstance(parent_id, unicode)
481
def get_revision_xml(self, revision_id):
482
# TODO: jam 20070210 This shouldn't be necessary since get_revision
483
# would have already do it.
484
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
485
revision_id = osutils.safe_revision_id(revision_id)
486
rev = self.get_revision(revision_id)
488
# the current serializer..
489
self._revision_store._serializer.write_revision(rev, rev_tmp)
491
return rev_tmp.getvalue()
494
def get_revision(self, revision_id):
495
"""Return the Revision object for a named revision"""
496
# TODO: jam 20070210 get_revision_reconcile should do this for us
497
revision_id = osutils.safe_revision_id(revision_id)
498
r = self.get_revision_reconcile(revision_id)
499
# weave corruption can lead to absent revision markers that should be
501
# the following test is reasonably cheap (it needs a single weave read)
502
# and the weave is cached in read transactions. In write transactions
503
# it is not cached but typically we only read a small number of
504
# revisions. For knits when they are introduced we will probably want
505
# to ensure that caching write transactions are in use.
506
inv = self.get_inventory_weave()
507
self._check_revision_parents(r, inv)
511
def get_deltas_for_revisions(self, revisions):
512
"""Produce a generator of revision deltas.
514
Note that the input is a sequence of REVISIONS, not revision_ids.
515
Trees will be held in memory until the generator exits.
516
Each delta is relative to the revision's lefthand predecessor.
518
required_trees = set()
519
for revision in revisions:
520
required_trees.add(revision.revision_id)
521
required_trees.update(revision.parent_ids[:1])
522
trees = dict((t.get_revision_id(), t) for
523
t in self.revision_trees(required_trees))
524
for revision in revisions:
525
if not revision.parent_ids:
526
old_tree = self.revision_tree(None)
528
old_tree = trees[revision.parent_ids[0]]
529
yield trees[revision.revision_id].changes_from(old_tree)
532
def get_revision_delta(self, revision_id):
533
"""Return the delta for one revision.
535
The delta is relative to the left-hand predecessor of the
538
r = self.get_revision(revision_id)
539
return list(self.get_deltas_for_revisions([r]))[0]
541
def _check_revision_parents(self, revision, inventory):
542
"""Private to Repository and Fetch.
544
This checks the parentage of revision in an inventory weave for
545
consistency and is only applicable to inventory-weave-for-ancestry
546
using repository formats & fetchers.
548
weave_parents = inventory.get_parents(revision.revision_id)
549
weave_names = inventory.versions()
550
for parent_id in revision.parent_ids:
551
if parent_id in weave_names:
552
# this parent must not be a ghost.
553
if not parent_id in weave_parents:
555
raise errors.CorruptRepository(self)
558
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
559
revision_id = osutils.safe_revision_id(revision_id)
560
signature = gpg_strategy.sign(plaintext)
561
self._revision_store.add_revision_signature_text(revision_id,
563
self.get_transaction())
565
def fileids_altered_by_revision_ids(self, revision_ids):
566
"""Find the file ids and versions affected by revisions.
568
:param revisions: an iterable containing revision ids.
569
:return: a dictionary mapping altered file-ids to an iterable of
570
revision_ids. Each altered file-ids has the exact revision_ids that
571
altered it listed explicitly.
573
assert self._serializer.support_altered_by_hack, \
574
("fileids_altered_by_revision_ids only supported for branches "
575
"which store inventory as unnested xml, not on %r" % self)
576
selected_revision_ids = set(osutils.safe_revision_id(r)
577
for r in revision_ids)
578
w = self.get_inventory_weave()
581
# this code needs to read every new line in every inventory for the
582
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
583
# not present in one of those inventories is unnecessary but not
584
# harmful because we are filtering by the revision id marker in the
585
# inventory lines : we only select file ids altered in one of those
586
# revisions. We don't need to see all lines in the inventory because
587
# only those added in an inventory in rev X can contain a revision=X
589
unescape_revid_cache = {}
590
unescape_fileid_cache = {}
592
# jam 20061218 In a big fetch, this handles hundreds of thousands
593
# of lines, so it has had a lot of inlining and optimizing done.
594
# Sorry that it is a little bit messy.
595
# Move several functions to be local variables, since this is a long
597
search = self._file_ids_altered_regex.search
598
unescape = _unescape_xml
599
setdefault = result.setdefault
600
pb = ui.ui_factory.nested_progress_bar()
602
for line in w.iter_lines_added_or_present_in_versions(
603
selected_revision_ids, pb=pb):
607
# One call to match.group() returning multiple items is quite a
608
# bit faster than 2 calls to match.group() each returning 1
609
file_id, revision_id = match.group('file_id', 'revision_id')
611
# Inlining the cache lookups helps a lot when you make 170,000
612
# lines and 350k ids, versus 8.4 unique ids.
613
# Using a cache helps in 2 ways:
614
# 1) Avoids unnecessary decoding calls
615
# 2) Re-uses cached strings, which helps in future set and
617
# (2) is enough that removing encoding entirely along with
618
# the cache (so we are using plain strings) results in no
619
# performance improvement.
621
revision_id = unescape_revid_cache[revision_id]
623
unescaped = unescape(revision_id)
624
unescape_revid_cache[revision_id] = unescaped
625
revision_id = unescaped
627
if revision_id in selected_revision_ids:
629
file_id = unescape_fileid_cache[file_id]
631
unescaped = unescape(file_id)
632
unescape_fileid_cache[file_id] = unescaped
634
setdefault(file_id, set()).add(revision_id)
640
def get_inventory_weave(self):
641
return self.control_weaves.get_weave('inventory',
642
self.get_transaction())
645
def get_inventory(self, revision_id):
646
"""Get Inventory object by hash."""
647
# TODO: jam 20070210 Technically we don't need to sanitize, since all
648
# called functions must sanitize.
649
revision_id = osutils.safe_revision_id(revision_id)
650
return self.deserialise_inventory(
651
revision_id, self.get_inventory_xml(revision_id))
653
def deserialise_inventory(self, revision_id, xml):
654
"""Transform the xml into an inventory object.
656
:param revision_id: The expected revision id of the inventory.
657
:param xml: A serialised inventory.
659
revision_id = osutils.safe_revision_id(revision_id)
660
result = self._serializer.read_inventory_from_string(xml)
661
result.root.revision = revision_id
664
def serialise_inventory(self, inv):
665
return self._serializer.write_inventory_to_string(inv)
667
def get_serializer_format(self):
668
return self._serializer.format_num
671
def get_inventory_xml(self, revision_id):
672
"""Get inventory XML as a file object."""
673
revision_id = osutils.safe_revision_id(revision_id)
675
assert isinstance(revision_id, str), type(revision_id)
676
iw = self.get_inventory_weave()
677
return iw.get_text(revision_id)
679
raise errors.HistoryMissing(self, 'inventory', revision_id)
682
def get_inventory_sha1(self, revision_id):
683
"""Return the sha1 hash of the inventory entry
685
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
686
revision_id = osutils.safe_revision_id(revision_id)
687
return self.get_revision(revision_id).inventory_sha1
690
def get_revision_graph(self, revision_id=None):
691
"""Return a dictionary containing the revision graph.
693
:param revision_id: The revision_id to get a graph from. If None, then
694
the entire revision graph is returned. This is a deprecated mode of
695
operation and will be removed in the future.
696
:return: a dictionary of revision_id->revision_parents_list.
698
# special case NULL_REVISION
699
if revision_id == _mod_revision.NULL_REVISION:
701
revision_id = osutils.safe_revision_id(revision_id)
702
a_weave = self.get_inventory_weave()
703
all_revisions = self._eliminate_revisions_not_present(
705
entire_graph = dict([(node, tuple(a_weave.get_parents(node))) for
706
node in all_revisions])
707
if revision_id is None:
709
elif revision_id not in entire_graph:
710
raise errors.NoSuchRevision(self, revision_id)
712
# add what can be reached from revision_id
714
pending = set([revision_id])
715
while len(pending) > 0:
717
result[node] = entire_graph[node]
718
for revision_id in result[node]:
719
if revision_id not in result:
720
pending.add(revision_id)
724
def get_revision_graph_with_ghosts(self, revision_ids=None):
725
"""Return a graph of the revisions with ghosts marked as applicable.
727
:param revision_ids: an iterable of revisions to graph or None for all.
728
:return: a Graph object with the graph reachable from revision_ids.
730
result = deprecated_graph.Graph()
732
pending = set(self.all_revision_ids())
735
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
736
# special case NULL_REVISION
737
if _mod_revision.NULL_REVISION in pending:
738
pending.remove(_mod_revision.NULL_REVISION)
739
required = set(pending)
742
revision_id = pending.pop()
744
rev = self.get_revision(revision_id)
745
except errors.NoSuchRevision:
746
if revision_id in required:
749
result.add_ghost(revision_id)
751
for parent_id in rev.parent_ids:
752
# is this queued or done ?
753
if (parent_id not in pending and
754
parent_id not in done):
756
pending.add(parent_id)
757
result.add_node(revision_id, rev.parent_ids)
758
done.add(revision_id)
761
def _get_history_vf(self):
762
"""Get a versionedfile whose history graph reflects all revisions.
764
For weave repositories, this is the inventory weave.
766
return self.get_inventory_weave()
768
def iter_reverse_revision_history(self, revision_id):
769
"""Iterate backwards through revision ids in the lefthand history
771
:param revision_id: The revision id to start with. All its lefthand
772
ancestors will be traversed.
774
revision_id = osutils.safe_revision_id(revision_id)
775
if revision_id in (None, _mod_revision.NULL_REVISION):
777
next_id = revision_id
778
versionedfile = self._get_history_vf()
781
parents = versionedfile.get_parents(next_id)
782
if len(parents) == 0:
788
def get_revision_inventory(self, revision_id):
789
"""Return inventory of a past revision."""
790
# TODO: Unify this with get_inventory()
791
# bzr 0.0.6 and later imposes the constraint that the inventory_id
792
# must be the same as its revision, so this is trivial.
793
if revision_id is None:
794
# This does not make sense: if there is no revision,
795
# then it is the current tree inventory surely ?!
796
# and thus get_root_id() is something that looks at the last
797
# commit on the branch, and the get_root_id is an inventory check.
798
raise NotImplementedError
799
# return Inventory(self.get_root_id())
801
return self.get_inventory(revision_id)
805
"""Return True if this repository is flagged as a shared repository."""
806
raise NotImplementedError(self.is_shared)
809
def reconcile(self, other=None, thorough=False):
810
"""Reconcile this repository."""
811
from bzrlib.reconcile import RepoReconciler
812
reconciler = RepoReconciler(self, thorough=thorough)
813
reconciler.reconcile()
817
def revision_tree(self, revision_id):
818
"""Return Tree for a revision on this branch.
820
`revision_id` may be None for the empty tree revision.
822
# TODO: refactor this to use an existing revision object
823
# so we don't need to read it in twice.
824
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
825
return RevisionTree(self, Inventory(root_id=None),
826
_mod_revision.NULL_REVISION)
828
revision_id = osutils.safe_revision_id(revision_id)
829
inv = self.get_revision_inventory(revision_id)
830
return RevisionTree(self, inv, revision_id)
833
def revision_trees(self, revision_ids):
834
"""Return Tree for a revision on this branch.
836
`revision_id` may not be None or 'null:'"""
837
assert None not in revision_ids
838
assert _mod_revision.NULL_REVISION not in revision_ids
839
texts = self.get_inventory_weave().get_texts(revision_ids)
840
for text, revision_id in zip(texts, revision_ids):
841
inv = self.deserialise_inventory(revision_id, text)
842
yield RevisionTree(self, inv, revision_id)
845
def get_ancestry(self, revision_id, topo_sorted=True):
846
"""Return a list of revision-ids integrated by a revision.
848
The first element of the list is always None, indicating the origin
849
revision. This might change when we have history horizons, or
850
perhaps we should have a new API.
852
This is topologically sorted.
854
if _mod_revision.is_null(revision_id):
856
revision_id = osutils.safe_revision_id(revision_id)
857
if not self.has_revision(revision_id):
858
raise errors.NoSuchRevision(self, revision_id)
859
w = self.get_inventory_weave()
860
candidates = w.get_ancestry(revision_id, topo_sorted)
861
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
864
"""Compress the data within the repository.
866
This operation only makes sense for some repository types. For other
867
types it should be a no-op that just returns.
869
This stub method does not require a lock, but subclasses should use
870
@needs_write_lock as this is a long running call its reasonable to
871
implicitly lock for the user.
875
def print_file(self, file, revision_id):
876
"""Print `file` to stdout.
878
FIXME RBC 20060125 as John Meinel points out this is a bad api
879
- it writes to stdout, it assumes that that is valid etc. Fix
880
by creating a new more flexible convenience function.
882
revision_id = osutils.safe_revision_id(revision_id)
883
tree = self.revision_tree(revision_id)
884
# use inventory as it was in that revision
885
file_id = tree.inventory.path2id(file)
887
# TODO: jam 20060427 Write a test for this code path
888
# it had a bug in it, and was raising the wrong
890
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
891
tree.print_file(file_id)
893
def get_transaction(self):
894
return self.control_files.get_transaction()
896
def revision_parents(self, revision_id):
897
revision_id = osutils.safe_revision_id(revision_id)
898
return self.get_inventory_weave().parent_names(revision_id)
900
def get_parents(self, revision_ids):
901
"""See StackedParentsProvider.get_parents"""
903
for revision_id in revision_ids:
904
if revision_id == _mod_revision.NULL_REVISION:
908
parents = self.get_revision(revision_id).parent_ids
909
except errors.NoSuchRevision:
912
if len(parents) == 0:
913
parents = [_mod_revision.NULL_REVISION]
914
parents_list.append(parents)
917
def _make_parents_provider(self):
920
def get_graph(self, other_repository=None):
921
"""Return the graph walker for this repository format"""
922
parents_provider = self._make_parents_provider()
923
if (other_repository is not None and
924
other_repository.bzrdir.transport.base !=
925
self.bzrdir.transport.base):
926
parents_provider = graph._StackedParentsProvider(
927
[parents_provider, other_repository._make_parents_provider()])
928
return graph.Graph(parents_provider)
931
def set_make_working_trees(self, new_value):
932
"""Set the policy flag for making working trees when creating branches.
934
This only applies to branches that use this repository.
936
The default is 'True'.
937
:param new_value: True to restore the default, False to disable making
940
raise NotImplementedError(self.set_make_working_trees)
942
def make_working_trees(self):
943
"""Returns the policy for making working trees on new branches."""
944
raise NotImplementedError(self.make_working_trees)
947
def sign_revision(self, revision_id, gpg_strategy):
948
revision_id = osutils.safe_revision_id(revision_id)
949
plaintext = Testament.from_revision(self, revision_id).as_short_text()
950
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
953
def has_signature_for_revision_id(self, revision_id):
954
"""Query for a revision signature for revision_id in the repository."""
955
revision_id = osutils.safe_revision_id(revision_id)
956
return self._revision_store.has_signature(revision_id,
957
self.get_transaction())
960
def get_signature_text(self, revision_id):
961
"""Return the text for a signature."""
962
revision_id = osutils.safe_revision_id(revision_id)
963
return self._revision_store.get_signature_text(revision_id,
964
self.get_transaction())
967
def check(self, revision_ids):
968
"""Check consistency of all history of given revision_ids.
970
Different repository implementations should override _check().
972
:param revision_ids: A non-empty list of revision_ids whose ancestry
973
will be checked. Typically the last revision_id of a branch.
976
raise ValueError("revision_ids must be non-empty in %s.check"
978
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
979
return self._check(revision_ids)
981
def _check(self, revision_ids):
982
result = check.Check(self)
986
def _warn_if_deprecated(self):
987
global _deprecation_warning_done
988
if _deprecation_warning_done:
990
_deprecation_warning_done = True
991
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
992
% (self._format, self.bzrdir.transport.base))
994
def supports_rich_root(self):
995
return self._format.rich_root_data
997
def _check_ascii_revisionid(self, revision_id, method):
998
"""Private helper for ascii-only repositories."""
999
# weave repositories refuse to store revisionids that are non-ascii.
1000
if revision_id is not None:
1001
# weaves require ascii revision ids.
1002
if isinstance(revision_id, unicode):
1004
revision_id.encode('ascii')
1005
except UnicodeEncodeError:
1006
raise errors.NonAsciiRevisionId(method, self)
1009
revision_id.decode('ascii')
1010
except UnicodeDecodeError:
1011
raise errors.NonAsciiRevisionId(method, self)
1015
# remove these delegates a while after bzr 0.15
1016
def __make_delegated(name, from_module):
1017
def _deprecated_repository_forwarder():
1018
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1019
% (name, from_module),
1022
m = __import__(from_module, globals(), locals(), [name])
1024
return getattr(m, name)
1025
except AttributeError:
1026
raise AttributeError('module %s has no name %s'
1028
globals()[name] = _deprecated_repository_forwarder
1031
'AllInOneRepository',
1032
'WeaveMetaDirRepository',
1033
'PreSplitOutRepositoryFormat',
1034
'RepositoryFormat4',
1035
'RepositoryFormat5',
1036
'RepositoryFormat6',
1037
'RepositoryFormat7',
1039
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1043
'RepositoryFormatKnit',
1044
'RepositoryFormatKnit1',
1046
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1049
def install_revision(repository, rev, revision_tree):
1050
"""Install all revision data into a repository."""
1051
present_parents = []
1053
for p_id in rev.parent_ids:
1054
if repository.has_revision(p_id):
1055
present_parents.append(p_id)
1056
parent_trees[p_id] = repository.revision_tree(p_id)
1058
parent_trees[p_id] = repository.revision_tree(None)
1060
inv = revision_tree.inventory
1061
entries = inv.iter_entries()
1062
# backwards compatability hack: skip the root id.
1063
if not repository.supports_rich_root():
1064
path, root = entries.next()
1065
if root.revision != rev.revision_id:
1066
raise errors.IncompatibleRevision(repr(repository))
1067
# Add the texts that are not already present
1068
for path, ie in entries:
1069
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1070
repository.get_transaction())
1071
if ie.revision not in w:
1073
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1074
# with InventoryEntry.find_previous_heads(). if it is, then there
1075
# is a latent bug here where the parents may have ancestors of each
1077
for revision, tree in parent_trees.iteritems():
1078
if ie.file_id not in tree:
1080
parent_id = tree.inventory[ie.file_id].revision
1081
if parent_id in text_parents:
1083
text_parents.append(parent_id)
1085
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1086
repository.get_transaction())
1087
lines = revision_tree.get_file(ie.file_id).readlines()
1088
vfile.add_lines(rev.revision_id, text_parents, lines)
1090
# install the inventory
1091
repository.add_inventory(rev.revision_id, inv, present_parents)
1092
except errors.RevisionAlreadyPresent:
1094
repository.add_revision(rev.revision_id, rev, inv)
1097
class MetaDirRepository(Repository):
1098
"""Repositories in the new meta-dir layout."""
1100
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1101
super(MetaDirRepository, self).__init__(_format,
1107
dir_mode = self.control_files._dir_mode
1108
file_mode = self.control_files._file_mode
1111
def is_shared(self):
1112
"""Return True if this repository is flagged as a shared repository."""
1113
return self.control_files._transport.has('shared-storage')
1116
def set_make_working_trees(self, new_value):
1117
"""Set the policy flag for making working trees when creating branches.
1119
This only applies to branches that use this repository.
1121
The default is 'True'.
1122
:param new_value: True to restore the default, False to disable making
1127
self.control_files._transport.delete('no-working-trees')
1128
except errors.NoSuchFile:
1131
self.control_files.put_utf8('no-working-trees', '')
1133
def make_working_trees(self):
1134
"""Returns the policy for making working trees on new branches."""
1135
return not self.control_files._transport.has('no-working-trees')
1138
class RepositoryFormatRegistry(registry.Registry):
1139
"""Registry of RepositoryFormats.
1142
def get(self, format_string):
1143
r = registry.Registry.get(self, format_string)
1149
format_registry = RepositoryFormatRegistry()
1150
"""Registry of formats, indexed by their identifying format string.
1152
This can contain either format instances themselves, or classes/factories that
1153
can be called to obtain one.
1157
#####################################################################
1158
# Repository Formats
1160
class RepositoryFormat(object):
1161
"""A repository format.
1163
Formats provide three things:
1164
* An initialization routine to construct repository data on disk.
1165
* a format string which is used when the BzrDir supports versioned
1167
* an open routine which returns a Repository instance.
1169
Formats are placed in an dict by their format string for reference
1170
during opening. These should be subclasses of RepositoryFormat
1173
Once a format is deprecated, just deprecate the initialize and open
1174
methods on the format class. Do not deprecate the object, as the
1175
object will be created every system load.
1177
Common instance attributes:
1178
_matchingbzrdir - the bzrdir format that the repository format was
1179
originally written to work with. This can be used if manually
1180
constructing a bzrdir and repository, or more commonly for test suite
1185
return "<%s>" % self.__class__.__name__
1187
def __eq__(self, other):
1188
# format objects are generally stateless
1189
return isinstance(other, self.__class__)
1191
def __ne__(self, other):
1192
return not self == other
1195
def find_format(klass, a_bzrdir):
1196
"""Return the format for the repository object in a_bzrdir.
1198
This is used by bzr native formats that have a "format" file in
1199
the repository. Other methods may be used by different types of
1203
transport = a_bzrdir.get_repository_transport(None)
1204
format_string = transport.get("format").read()
1205
return format_registry.get(format_string)
1206
except errors.NoSuchFile:
1207
raise errors.NoRepositoryPresent(a_bzrdir)
1209
raise errors.UnknownFormatError(format=format_string)
1212
def register_format(klass, format):
1213
format_registry.register(format.get_format_string(), format)
1216
def unregister_format(klass, format):
1217
format_registry.remove(format.get_format_string())
1220
def get_default_format(klass):
1221
"""Return the current default format."""
1222
from bzrlib import bzrdir
1223
return bzrdir.format_registry.make_bzrdir('default').repository_format
1225
def _get_control_store(self, repo_transport, control_files):
1226
"""Return the control store for this repository."""
1227
raise NotImplementedError(self._get_control_store)
1229
def get_format_string(self):
1230
"""Return the ASCII format string that identifies this format.
1232
Note that in pre format ?? repositories the format string is
1233
not permitted nor written to disk.
1235
raise NotImplementedError(self.get_format_string)
1237
def get_format_description(self):
1238
"""Return the short description for this format."""
1239
raise NotImplementedError(self.get_format_description)
1241
def _get_revision_store(self, repo_transport, control_files):
1242
"""Return the revision store object for this a_bzrdir."""
1243
raise NotImplementedError(self._get_revision_store)
1245
def _get_text_rev_store(self,
1252
"""Common logic for getting a revision store for a repository.
1254
see self._get_revision_store for the subclass-overridable method to
1255
get the store for a repository.
1257
from bzrlib.store.revision.text import TextRevisionStore
1258
dir_mode = control_files._dir_mode
1259
file_mode = control_files._file_mode
1260
text_store = TextStore(transport.clone(name),
1262
compressed=compressed,
1264
file_mode=file_mode)
1265
_revision_store = TextRevisionStore(text_store, serializer)
1266
return _revision_store
1268
# TODO: this shouldn't be in the base class, it's specific to things that
1269
# use weaves or knits -- mbp 20070207
1270
def _get_versioned_file_store(self,
1275
versionedfile_class=None,
1276
versionedfile_kwargs={},
1278
if versionedfile_class is None:
1279
versionedfile_class = self._versionedfile_class
1280
weave_transport = control_files._transport.clone(name)
1281
dir_mode = control_files._dir_mode
1282
file_mode = control_files._file_mode
1283
return VersionedFileStore(weave_transport, prefixed=prefixed,
1285
file_mode=file_mode,
1286
versionedfile_class=versionedfile_class,
1287
versionedfile_kwargs=versionedfile_kwargs,
1290
def initialize(self, a_bzrdir, shared=False):
1291
"""Initialize a repository of this format in a_bzrdir.
1293
:param a_bzrdir: The bzrdir to put the new repository in it.
1294
:param shared: The repository should be initialized as a sharable one.
1295
:returns: The new repository object.
1297
This may raise UninitializableFormat if shared repository are not
1298
compatible the a_bzrdir.
1300
raise NotImplementedError(self.initialize)
1302
def is_supported(self):
1303
"""Is this format supported?
1305
Supported formats must be initializable and openable.
1306
Unsupported formats may not support initialization or committing or
1307
some other features depending on the reason for not being supported.
1311
def check_conversion_target(self, target_format):
1312
raise NotImplementedError(self.check_conversion_target)
1314
def open(self, a_bzrdir, _found=False):
1315
"""Return an instance of this format for the bzrdir a_bzrdir.
1317
_found is a private parameter, do not use it.
1319
raise NotImplementedError(self.open)
1322
class MetaDirRepositoryFormat(RepositoryFormat):
1323
"""Common base class for the new repositories using the metadir layout."""
1325
rich_root_data = False
1326
supports_tree_reference = False
1327
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1330
super(MetaDirRepositoryFormat, self).__init__()
1332
def _create_control_files(self, a_bzrdir):
1333
"""Create the required files and the initial control_files object."""
1334
# FIXME: RBC 20060125 don't peek under the covers
1335
# NB: no need to escape relative paths that are url safe.
1336
repository_transport = a_bzrdir.get_repository_transport(self)
1337
control_files = lockable_files.LockableFiles(repository_transport,
1338
'lock', lockdir.LockDir)
1339
control_files.create_lock()
1340
return control_files
1342
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1343
"""Upload the initial blank content."""
1344
control_files = self._create_control_files(a_bzrdir)
1345
control_files.lock_write()
1347
control_files._transport.mkdir_multi(dirs,
1348
mode=control_files._dir_mode)
1349
for file, content in files:
1350
control_files.put(file, content)
1351
for file, content in utf8_files:
1352
control_files.put_utf8(file, content)
1354
control_files.put_utf8('shared-storage', '')
1356
control_files.unlock()
1359
# formats which have no format string are not discoverable
1360
# and not independently creatable, so are not registered. They're
1361
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1362
# needed, it's constructed directly by the BzrDir. Non-native formats where
1363
# the repository is not separately opened are similar.
1365
format_registry.register_lazy(
1366
'Bazaar-NG Repository format 7',
1367
'bzrlib.repofmt.weaverepo',
1370
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1371
# default control directory format
1373
format_registry.register_lazy(
1374
'Bazaar-NG Knit Repository Format 1',
1375
'bzrlib.repofmt.knitrepo',
1376
'RepositoryFormatKnit1',
1378
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1380
format_registry.register_lazy(
1381
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1382
'bzrlib.repofmt.knitrepo',
1383
'RepositoryFormatKnit3',
1387
class InterRepository(InterObject):
1388
"""This class represents operations taking place between two repositories.
1390
Its instances have methods like copy_content and fetch, and contain
1391
references to the source and target repositories these operations can be
1394
Often we will provide convenience methods on 'repository' which carry out
1395
operations with another repository - they will always forward to
1396
InterRepository.get(other).method_name(parameters).
1400
"""The available optimised InterRepository types."""
1402
def copy_content(self, revision_id=None):
1403
raise NotImplementedError(self.copy_content)
1405
def fetch(self, revision_id=None, pb=None):
1406
"""Fetch the content required to construct revision_id.
1408
The content is copied from self.source to self.target.
1410
:param revision_id: if None all content is copied, if NULL_REVISION no
1412
:param pb: optional progress bar to use for progress reports. If not
1413
provided a default one will be created.
1415
Returns the copied revision count and the failed revisions in a tuple:
1418
raise NotImplementedError(self.fetch)
1421
def missing_revision_ids(self, revision_id=None):
1422
"""Return the revision ids that source has that target does not.
1424
These are returned in topological order.
1426
:param revision_id: only return revision ids included by this
1429
# generic, possibly worst case, slow code path.
1430
target_ids = set(self.target.all_revision_ids())
1431
if revision_id is not None:
1432
# TODO: jam 20070210 InterRepository is internal enough that it
1433
# should assume revision_ids are already utf-8
1434
revision_id = osutils.safe_revision_id(revision_id)
1435
source_ids = self.source.get_ancestry(revision_id)
1436
assert source_ids[0] is None
1439
source_ids = self.source.all_revision_ids()
1440
result_set = set(source_ids).difference(target_ids)
1441
# this may look like a no-op: its not. It preserves the ordering
1442
# other_ids had while only returning the members from other_ids
1443
# that we've decided we need.
1444
return [rev_id for rev_id in source_ids if rev_id in result_set]
1447
class InterSameDataRepository(InterRepository):
1448
"""Code for converting between repositories that represent the same data.
1450
Data format and model must match for this to work.
1454
def _get_repo_format_to_test(self):
1455
"""Repository format for testing with."""
1456
return RepositoryFormat.get_default_format()
1459
def is_compatible(source, target):
1460
if source.supports_rich_root() != target.supports_rich_root():
1462
if source._serializer != target._serializer:
1467
def copy_content(self, revision_id=None):
1468
"""Make a complete copy of the content in self into destination.
1470
This copies both the repository's revision data, and configuration information
1471
such as the make_working_trees setting.
1473
This is a destructive operation! Do not use it on existing
1476
:param revision_id: Only copy the content needed to construct
1477
revision_id and its parents.
1480
self.target.set_make_working_trees(self.source.make_working_trees())
1481
except NotImplementedError:
1483
# TODO: jam 20070210 This is fairly internal, so we should probably
1484
# just assert that revision_id is not unicode.
1485
revision_id = osutils.safe_revision_id(revision_id)
1486
# but don't bother fetching if we have the needed data now.
1487
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1488
self.target.has_revision(revision_id)):
1490
self.target.fetch(self.source, revision_id=revision_id)
1493
def fetch(self, revision_id=None, pb=None):
1494
"""See InterRepository.fetch()."""
1495
from bzrlib.fetch import GenericRepoFetcher
1496
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1497
self.source, self.source._format, self.target,
1498
self.target._format)
1499
# TODO: jam 20070210 This should be an assert, not a translate
1500
revision_id = osutils.safe_revision_id(revision_id)
1501
f = GenericRepoFetcher(to_repository=self.target,
1502
from_repository=self.source,
1503
last_revision=revision_id,
1505
return f.count_copied, f.failed_revisions
1508
class InterWeaveRepo(InterSameDataRepository):
1509
"""Optimised code paths between Weave based repositories."""
1512
def _get_repo_format_to_test(self):
1513
from bzrlib.repofmt import weaverepo
1514
return weaverepo.RepositoryFormat7()
1517
def is_compatible(source, target):
1518
"""Be compatible with known Weave formats.
1520
We don't test for the stores being of specific types because that
1521
could lead to confusing results, and there is no need to be
1524
from bzrlib.repofmt.weaverepo import (
1530
return (isinstance(source._format, (RepositoryFormat5,
1532
RepositoryFormat7)) and
1533
isinstance(target._format, (RepositoryFormat5,
1535
RepositoryFormat7)))
1536
except AttributeError:
1540
def copy_content(self, revision_id=None):
1541
"""See InterRepository.copy_content()."""
1542
# weave specific optimised path:
1543
# TODO: jam 20070210 Internal, should be an assert, not translate
1544
revision_id = osutils.safe_revision_id(revision_id)
1546
self.target.set_make_working_trees(self.source.make_working_trees())
1547
except NotImplementedError:
1549
# FIXME do not peek!
1550
if self.source.control_files._transport.listable():
1551
pb = ui.ui_factory.nested_progress_bar()
1553
self.target.weave_store.copy_all_ids(
1554
self.source.weave_store,
1556
from_transaction=self.source.get_transaction(),
1557
to_transaction=self.target.get_transaction())
1558
pb.update('copying inventory', 0, 1)
1559
self.target.control_weaves.copy_multi(
1560
self.source.control_weaves, ['inventory'],
1561
from_transaction=self.source.get_transaction(),
1562
to_transaction=self.target.get_transaction())
1563
self.target._revision_store.text_store.copy_all_ids(
1564
self.source._revision_store.text_store,
1569
self.target.fetch(self.source, revision_id=revision_id)
1572
def fetch(self, revision_id=None, pb=None):
1573
"""See InterRepository.fetch()."""
1574
from bzrlib.fetch import GenericRepoFetcher
1575
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1576
self.source, self.source._format, self.target, self.target._format)
1577
# TODO: jam 20070210 This should be an assert, not a translate
1578
revision_id = osutils.safe_revision_id(revision_id)
1579
f = GenericRepoFetcher(to_repository=self.target,
1580
from_repository=self.source,
1581
last_revision=revision_id,
1583
return f.count_copied, f.failed_revisions
1586
def missing_revision_ids(self, revision_id=None):
1587
"""See InterRepository.missing_revision_ids()."""
1588
# we want all revisions to satisfy revision_id in source.
1589
# but we don't want to stat every file here and there.
1590
# we want then, all revisions other needs to satisfy revision_id
1591
# checked, but not those that we have locally.
1592
# so the first thing is to get a subset of the revisions to
1593
# satisfy revision_id in source, and then eliminate those that
1594
# we do already have.
1595
# this is slow on high latency connection to self, but as as this
1596
# disk format scales terribly for push anyway due to rewriting
1597
# inventory.weave, this is considered acceptable.
1599
if revision_id is not None:
1600
source_ids = self.source.get_ancestry(revision_id)
1601
assert source_ids[0] is None
1604
source_ids = self.source._all_possible_ids()
1605
source_ids_set = set(source_ids)
1606
# source_ids is the worst possible case we may need to pull.
1607
# now we want to filter source_ids against what we actually
1608
# have in target, but don't try to check for existence where we know
1609
# we do not have a revision as that would be pointless.
1610
target_ids = set(self.target._all_possible_ids())
1611
possibly_present_revisions = target_ids.intersection(source_ids_set)
1612
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1613
required_revisions = source_ids_set.difference(actually_present_revisions)
1614
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1615
if revision_id is not None:
1616
# we used get_ancestry to determine source_ids then we are assured all
1617
# revisions referenced are present as they are installed in topological order.
1618
# and the tip revision was validated by get_ancestry.
1619
return required_topo_revisions
1621
# if we just grabbed the possibly available ids, then
1622
# we only have an estimate of whats available and need to validate
1623
# that against the revision records.
1624
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1627
class InterKnitRepo(InterSameDataRepository):
1628
"""Optimised code paths between Knit based repositories."""
1631
def _get_repo_format_to_test(self):
1632
from bzrlib.repofmt import knitrepo
1633
return knitrepo.RepositoryFormatKnit1()
1636
def is_compatible(source, target):
1637
"""Be compatible with known Knit formats.
1639
We don't test for the stores being of specific types because that
1640
could lead to confusing results, and there is no need to be
1643
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1645
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1646
isinstance(target._format, (RepositoryFormatKnit1)))
1647
except AttributeError:
1651
def fetch(self, revision_id=None, pb=None):
1652
"""See InterRepository.fetch()."""
1653
from bzrlib.fetch import KnitRepoFetcher
1654
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1655
self.source, self.source._format, self.target, self.target._format)
1656
# TODO: jam 20070210 This should be an assert, not a translate
1657
revision_id = osutils.safe_revision_id(revision_id)
1658
f = KnitRepoFetcher(to_repository=self.target,
1659
from_repository=self.source,
1660
last_revision=revision_id,
1662
return f.count_copied, f.failed_revisions
1665
def missing_revision_ids(self, revision_id=None):
1666
"""See InterRepository.missing_revision_ids()."""
1667
if revision_id is not None:
1668
source_ids = self.source.get_ancestry(revision_id)
1669
assert source_ids[0] is None
1672
source_ids = self.source._all_possible_ids()
1673
source_ids_set = set(source_ids)
1674
# source_ids is the worst possible case we may need to pull.
1675
# now we want to filter source_ids against what we actually
1676
# have in target, but don't try to check for existence where we know
1677
# we do not have a revision as that would be pointless.
1678
target_ids = set(self.target._all_possible_ids())
1679
possibly_present_revisions = target_ids.intersection(source_ids_set)
1680
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1681
required_revisions = source_ids_set.difference(actually_present_revisions)
1682
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1683
if revision_id is not None:
1684
# we used get_ancestry to determine source_ids then we are assured all
1685
# revisions referenced are present as they are installed in topological order.
1686
# and the tip revision was validated by get_ancestry.
1687
return required_topo_revisions
1689
# if we just grabbed the possibly available ids, then
1690
# we only have an estimate of whats available and need to validate
1691
# that against the revision records.
1692
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1695
class InterModel1and2(InterRepository):
1698
def _get_repo_format_to_test(self):
1702
def is_compatible(source, target):
1703
if not source.supports_rich_root() and target.supports_rich_root():
1709
def fetch(self, revision_id=None, pb=None):
1710
"""See InterRepository.fetch()."""
1711
from bzrlib.fetch import Model1toKnit2Fetcher
1712
# TODO: jam 20070210 This should be an assert, not a translate
1713
revision_id = osutils.safe_revision_id(revision_id)
1714
f = Model1toKnit2Fetcher(to_repository=self.target,
1715
from_repository=self.source,
1716
last_revision=revision_id,
1718
return f.count_copied, f.failed_revisions
1721
def copy_content(self, revision_id=None):
1722
"""Make a complete copy of the content in self into destination.
1724
This is a destructive operation! Do not use it on existing
1727
:param revision_id: Only copy the content needed to construct
1728
revision_id and its parents.
1731
self.target.set_make_working_trees(self.source.make_working_trees())
1732
except NotImplementedError:
1734
# TODO: jam 20070210 Internal, assert, don't translate
1735
revision_id = osutils.safe_revision_id(revision_id)
1736
# but don't bother fetching if we have the needed data now.
1737
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1738
self.target.has_revision(revision_id)):
1740
self.target.fetch(self.source, revision_id=revision_id)
1743
class InterKnit1and2(InterKnitRepo):
1746
def _get_repo_format_to_test(self):
1750
def is_compatible(source, target):
1751
"""Be compatible with Knit1 source and Knit3 target"""
1752
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
1754
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1755
RepositoryFormatKnit3
1756
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1757
isinstance(target._format, (RepositoryFormatKnit3)))
1758
except AttributeError:
1762
def fetch(self, revision_id=None, pb=None):
1763
"""See InterRepository.fetch()."""
1764
from bzrlib.fetch import Knit1to2Fetcher
1765
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1766
self.source, self.source._format, self.target,
1767
self.target._format)
1768
# TODO: jam 20070210 This should be an assert, not a translate
1769
revision_id = osutils.safe_revision_id(revision_id)
1770
f = Knit1to2Fetcher(to_repository=self.target,
1771
from_repository=self.source,
1772
last_revision=revision_id,
1774
return f.count_copied, f.failed_revisions
1777
class InterRemoteRepository(InterRepository):
1778
"""Code for converting between RemoteRepository objects.
1780
This just gets an non-remote repository from the RemoteRepository, and calls
1781
InterRepository.get again.
1784
def __init__(self, source, target):
1785
if isinstance(source, remote.RemoteRepository):
1786
source._ensure_real()
1787
real_source = source._real_repository
1789
real_source = source
1790
if isinstance(target, remote.RemoteRepository):
1791
target._ensure_real()
1792
real_target = target._real_repository
1794
real_target = target
1795
self.real_inter = InterRepository.get(real_source, real_target)
1798
def is_compatible(source, target):
1799
if isinstance(source, remote.RemoteRepository):
1801
if isinstance(target, remote.RemoteRepository):
1805
def copy_content(self, revision_id=None):
1806
self.real_inter.copy_content(revision_id=revision_id)
1808
def fetch(self, revision_id=None, pb=None):
1809
self.real_inter.fetch(revision_id=revision_id, pb=pb)
1812
def _get_repo_format_to_test(self):
1816
InterRepository.register_optimiser(InterSameDataRepository)
1817
InterRepository.register_optimiser(InterWeaveRepo)
1818
InterRepository.register_optimiser(InterKnitRepo)
1819
InterRepository.register_optimiser(InterModel1and2)
1820
InterRepository.register_optimiser(InterKnit1and2)
1821
InterRepository.register_optimiser(InterRemoteRepository)
1824
class CopyConverter(object):
1825
"""A repository conversion tool which just performs a copy of the content.
1827
This is slow but quite reliable.
1830
def __init__(self, target_format):
1831
"""Create a CopyConverter.
1833
:param target_format: The format the resulting repository should be.
1835
self.target_format = target_format
1837
def convert(self, repo, pb):
1838
"""Perform the conversion of to_convert, giving feedback via pb.
1840
:param to_convert: The disk object to convert.
1841
:param pb: a progress bar to use for progress information.
1846
# this is only useful with metadir layouts - separated repo content.
1847
# trigger an assertion if not such
1848
repo._format.get_format_string()
1849
self.repo_dir = repo.bzrdir
1850
self.step('Moving repository to repository.backup')
1851
self.repo_dir.transport.move('repository', 'repository.backup')
1852
backup_transport = self.repo_dir.transport.clone('repository.backup')
1853
repo._format.check_conversion_target(self.target_format)
1854
self.source_repo = repo._format.open(self.repo_dir,
1856
_override_transport=backup_transport)
1857
self.step('Creating new repository')
1858
converted = self.target_format.initialize(self.repo_dir,
1859
self.source_repo.is_shared())
1860
converted.lock_write()
1862
self.step('Copying content into repository.')
1863
self.source_repo.copy_content_into(converted)
1866
self.step('Deleting old repository content.')
1867
self.repo_dir.transport.delete_tree('repository.backup')
1868
self.pb.note('repository converted')
1870
def step(self, message):
1871
"""Update the pb by a step."""
1873
self.pb.update(message, self.count, self.total)
1876
class CommitBuilder(object):
1877
"""Provides an interface to build up a commit.
1879
This allows describing a tree to be committed without needing to
1880
know the internals of the format of the repository.
1883
record_root_entry = False
1884
def __init__(self, repository, parents, config, timestamp=None,
1885
timezone=None, committer=None, revprops=None,
1887
"""Initiate a CommitBuilder.
1889
:param repository: Repository to commit to.
1890
:param parents: Revision ids of the parents of the new revision.
1891
:param config: Configuration to use.
1892
:param timestamp: Optional timestamp recorded for commit.
1893
:param timezone: Optional timezone for timestamp.
1894
:param committer: Optional committer to set for commit.
1895
:param revprops: Optional dictionary of revision properties.
1896
:param revision_id: Optional revision id.
1898
self._config = config
1900
if committer is None:
1901
self._committer = self._config.username()
1903
assert isinstance(committer, basestring), type(committer)
1904
self._committer = committer
1906
self.new_inventory = Inventory(None)
1907
self._new_revision_id = osutils.safe_revision_id(revision_id)
1908
self.parents = parents
1909
self.repository = repository
1912
if revprops is not None:
1913
self._revprops.update(revprops)
1915
if timestamp is None:
1916
timestamp = time.time()
1917
# Restrict resolution to 1ms
1918
self._timestamp = round(timestamp, 3)
1920
if timezone is None:
1921
self._timezone = osutils.local_time_offset()
1923
self._timezone = int(timezone)
1925
self._generate_revision_if_needed()
1927
def commit(self, message):
1928
"""Make the actual commit.
1930
:return: The revision id of the recorded revision.
1932
rev = _mod_revision.Revision(
1933
timestamp=self._timestamp,
1934
timezone=self._timezone,
1935
committer=self._committer,
1937
inventory_sha1=self.inv_sha1,
1938
revision_id=self._new_revision_id,
1939
properties=self._revprops)
1940
rev.parent_ids = self.parents
1941
self.repository.add_revision(self._new_revision_id, rev,
1942
self.new_inventory, self._config)
1943
return self._new_revision_id
1945
def revision_tree(self):
1946
"""Return the tree that was just committed.
1948
After calling commit() this can be called to get a RevisionTree
1949
representing the newly committed tree. This is preferred to
1950
calling Repository.revision_tree() because that may require
1951
deserializing the inventory, while we already have a copy in
1954
return RevisionTree(self.repository, self.new_inventory,
1955
self._new_revision_id)
1957
def finish_inventory(self):
1958
"""Tell the builder that the inventory is finished."""
1959
if self.new_inventory.root is None:
1960
symbol_versioning.warn('Root entry should be supplied to'
1961
' record_entry_contents, as of bzr 0.10.',
1962
DeprecationWarning, stacklevel=2)
1963
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
1964
self.new_inventory.revision_id = self._new_revision_id
1965
self.inv_sha1 = self.repository.add_inventory(
1966
self._new_revision_id,
1971
def _gen_revision_id(self):
1972
"""Return new revision-id."""
1973
return generate_ids.gen_revision_id(self._config.username(),
1976
def _generate_revision_if_needed(self):
1977
"""Create a revision id if None was supplied.
1979
If the repository can not support user-specified revision ids
1980
they should override this function and raise CannotSetRevisionId
1981
if _new_revision_id is not None.
1983
:raises: CannotSetRevisionId
1985
if self._new_revision_id is None:
1986
self._new_revision_id = self._gen_revision_id()
1988
def record_entry_contents(self, ie, parent_invs, path, tree):
1989
"""Record the content of ie from tree into the commit if needed.
1991
Side effect: sets ie.revision when unchanged
1993
:param ie: An inventory entry present in the commit.
1994
:param parent_invs: The inventories of the parent revisions of the
1996
:param path: The path the entry is at in the tree.
1997
:param tree: The tree which contains this entry and should be used to
2000
if self.new_inventory.root is None and ie.parent_id is not None:
2001
symbol_versioning.warn('Root entry should be supplied to'
2002
' record_entry_contents, as of bzr 0.10.',
2003
DeprecationWarning, stacklevel=2)
2004
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
2006
self.new_inventory.add(ie)
2008
# ie.revision is always None if the InventoryEntry is considered
2009
# for committing. ie.snapshot will record the correct revision
2010
# which may be the sole parent if it is untouched.
2011
if ie.revision is not None:
2014
# In this revision format, root entries have no knit or weave
2015
if ie is self.new_inventory.root:
2016
# When serializing out to disk and back in
2017
# root.revision is always _new_revision_id
2018
ie.revision = self._new_revision_id
2020
previous_entries = ie.find_previous_heads(
2022
self.repository.weave_store,
2023
self.repository.get_transaction())
2024
# we are creating a new revision for ie in the history store
2026
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2028
def modified_directory(self, file_id, file_parents):
2029
"""Record the presence of a symbolic link.
2031
:param file_id: The file_id of the link to record.
2032
:param file_parents: The per-file parent revision ids.
2034
self._add_text_to_weave(file_id, [], file_parents.keys())
2036
def modified_reference(self, file_id, file_parents):
2037
"""Record the modification of a reference.
2039
:param file_id: The file_id of the link to record.
2040
:param file_parents: The per-file parent revision ids.
2042
self._add_text_to_weave(file_id, [], file_parents.keys())
2044
def modified_file_text(self, file_id, file_parents,
2045
get_content_byte_lines, text_sha1=None,
2047
"""Record the text of file file_id
2049
:param file_id: The file_id of the file to record the text of.
2050
:param file_parents: The per-file parent revision ids.
2051
:param get_content_byte_lines: A callable which will return the byte
2053
:param text_sha1: Optional SHA1 of the file contents.
2054
:param text_size: Optional size of the file contents.
2056
# mutter('storing text of file {%s} in revision {%s} into %r',
2057
# file_id, self._new_revision_id, self.repository.weave_store)
2058
# special case to avoid diffing on renames or
2060
if (len(file_parents) == 1
2061
and text_sha1 == file_parents.values()[0].text_sha1
2062
and text_size == file_parents.values()[0].text_size):
2063
previous_ie = file_parents.values()[0]
2064
versionedfile = self.repository.weave_store.get_weave(file_id,
2065
self.repository.get_transaction())
2066
versionedfile.clone_text(self._new_revision_id,
2067
previous_ie.revision, file_parents.keys())
2068
return text_sha1, text_size
2070
new_lines = get_content_byte_lines()
2071
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2072
# should return the SHA1 and size
2073
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2074
return osutils.sha_strings(new_lines), \
2075
sum(map(len, new_lines))
2077
def modified_link(self, file_id, file_parents, link_target):
2078
"""Record the presence of a symbolic link.
2080
:param file_id: The file_id of the link to record.
2081
:param file_parents: The per-file parent revision ids.
2082
:param link_target: Target location of this link.
2084
self._add_text_to_weave(file_id, [], file_parents.keys())
2086
def _add_text_to_weave(self, file_id, new_lines, parents):
2087
versionedfile = self.repository.weave_store.get_weave_or_empty(
2088
file_id, self.repository.get_transaction())
2089
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2090
versionedfile.clear_cache()
2093
class _CommitBuilder(CommitBuilder):
2094
"""Temporary class so old CommitBuilders are detected properly
2096
Note: CommitBuilder works whether or not root entry is recorded.
2099
record_root_entry = True
2102
class RootCommitBuilder(CommitBuilder):
2103
"""This commitbuilder actually records the root id"""
2105
record_root_entry = True
2107
def record_entry_contents(self, ie, parent_invs, path, tree):
2108
"""Record the content of ie from tree into the commit if needed.
2110
Side effect: sets ie.revision when unchanged
2112
:param ie: An inventory entry present in the commit.
2113
:param parent_invs: The inventories of the parent revisions of the
2115
:param path: The path the entry is at in the tree.
2116
:param tree: The tree which contains this entry and should be used to
2119
assert self.new_inventory.root is not None or ie.parent_id is None
2120
self.new_inventory.add(ie)
2122
# ie.revision is always None if the InventoryEntry is considered
2123
# for committing. ie.snapshot will record the correct revision
2124
# which may be the sole parent if it is untouched.
2125
if ie.revision is not None:
2128
previous_entries = ie.find_previous_heads(
2130
self.repository.weave_store,
2131
self.repository.get_transaction())
2132
# we are creating a new revision for ie in the history store
2134
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2146
def _unescaper(match, _map=_unescape_map):
2147
code = match.group(1)
2151
if not code.startswith('#'):
2153
return unichr(int(code[1:])).encode('utf8')
2159
def _unescape_xml(data):
2160
"""Unescape predefined XML entities in a string of data."""
2162
if _unescape_re is None:
2163
_unescape_re = re.compile('\&([^;]*);')
2164
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
