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.revisiontree import RevisionTree
44
from bzrlib.store.versioned import VersionedFileStore
45
from bzrlib.store.text import TextStore
46
from bzrlib.testament import Testament
50
from bzrlib.decorators import needs_read_lock, needs_write_lock
51
from bzrlib.inter import InterObject
52
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
53
from bzrlib.symbol_versioning import (
57
from bzrlib.trace import mutter, note, warning
60
# Old formats display a warning, but only once
61
_deprecation_warning_done = False
64
######################################################################
67
class Repository(object):
68
"""Repository holding history for one or more branches.
70
The repository holds and retrieves historical information including
71
revisions and file history. It's normally accessed only by the Branch,
72
which views a particular line of development through that history.
74
The Repository builds on top of Stores and a Transport, which respectively
75
describe the disk data format and the way of accessing the (possibly
79
_file_ids_altered_regex = lazy_regex.lazy_compile(
80
r'file_id="(?P<file_id>[^"]+)"'
81
r'.*revision="(?P<revision_id>[^"]+)"'
84
def abort_write_group(self):
85
"""Commit the contents accrued within the current write group.
87
:seealso: start_write_group.
89
if self._write_group is not self.get_transaction():
90
# has an unlock or relock occured ?
91
raise errors.BzrError('mismatched lock context and write group.')
92
self._abort_write_group()
93
self._write_group = None
95
def _abort_write_group(self):
96
"""Template method for per-repository write group cleanup.
98
This is called during abort before the write group is considered to be
99
finished and should cleanup any internal state accrued during the write
100
group. There is no requirement that data handed to the repository be
101
*not* made available - this is not a rollback - but neither should any
102
attempt be made to ensure that data added is fully commited. Abort is
103
invoked when an error has occured so futher disk or network operations
104
may not be possible or may error and if possible should not be
109
def add_inventory(self, revision_id, inv, parents):
110
"""Add the inventory inv to the repository as revision_id.
112
:param parents: The revision ids of the parents that revision_id
113
is known to have and are in the repository already.
115
returns the sha1 of the serialized inventory.
117
revision_id = osutils.safe_revision_id(revision_id)
118
_mod_revision.check_not_reserved_id(revision_id)
119
assert inv.revision_id is None or inv.revision_id == revision_id, \
120
"Mismatch between inventory revision" \
121
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
122
assert inv.root is not None
123
inv_text = self.serialise_inventory(inv)
124
inv_sha1 = osutils.sha_string(inv_text)
125
inv_vf = self.control_weaves.get_weave('inventory',
126
self.get_transaction())
127
self._inventory_add_lines(inv_vf, revision_id, parents,
128
osutils.split_lines(inv_text))
131
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines):
133
for parent in parents:
135
final_parents.append(parent)
137
inv_vf.add_lines(revision_id, final_parents, lines)
140
def add_revision(self, revision_id, rev, inv=None, config=None):
141
"""Add rev to the revision store as revision_id.
143
:param revision_id: the revision id to use.
144
:param rev: The revision object.
145
:param inv: The inventory for the revision. if None, it will be looked
146
up in the inventory storer
147
:param config: If None no digital signature will be created.
148
If supplied its signature_needed method will be used
149
to determine if a signature should be made.
151
revision_id = osutils.safe_revision_id(revision_id)
152
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
154
_mod_revision.check_not_reserved_id(revision_id)
155
if config is not None and config.signature_needed():
157
inv = self.get_inventory(revision_id)
158
plaintext = Testament(rev, inv).as_short_text()
159
self.store_revision_signature(
160
gpg.GPGStrategy(config), plaintext, revision_id)
161
if not revision_id in self.get_inventory_weave():
163
raise errors.WeaveRevisionNotPresent(revision_id,
164
self.get_inventory_weave())
166
# yes, this is not suitable for adding with ghosts.
167
self.add_inventory(revision_id, inv, rev.parent_ids)
168
self._revision_store.add_revision(rev, self.get_transaction())
171
def _all_possible_ids(self):
172
"""Return all the possible revisions that we could find."""
173
return self.get_inventory_weave().versions()
175
def all_revision_ids(self):
176
"""Returns a list of all the revision ids in the repository.
178
This is deprecated because code should generally work on the graph
179
reachable from a particular revision, and ignore any other revisions
180
that might be present. There is no direct replacement method.
182
return self._all_revision_ids()
185
def _all_revision_ids(self):
186
"""Returns a list of all the revision ids in the repository.
188
These are in as much topological order as the underlying store can
189
present: for weaves ghosts may lead to a lack of correctness until
190
the reweave updates the parents list.
192
if self._revision_store.text_store.listable():
193
return self._revision_store.all_revision_ids(self.get_transaction())
194
result = self._all_possible_ids()
195
# TODO: jam 20070210 Ensure that _all_possible_ids returns non-unicode
196
# ids. (It should, since _revision_store's API should change to
197
# return utf8 revision_ids)
198
return self._eliminate_revisions_not_present(result)
200
def break_lock(self):
201
"""Break a lock if one is present from another instance.
203
Uses the ui factory to ask for confirmation if the lock may be from
206
self.control_files.break_lock()
209
def _eliminate_revisions_not_present(self, revision_ids):
210
"""Check every revision id in revision_ids to see if we have it.
212
Returns a set of the present revisions.
215
for id in revision_ids:
216
if self.has_revision(id):
221
def create(a_bzrdir):
222
"""Construct the current default format repository in a_bzrdir."""
223
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
225
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
226
"""instantiate a Repository.
228
:param _format: The format of the repository on disk.
229
:param a_bzrdir: The BzrDir of the repository.
231
In the future we will have a single api for all stores for
232
getting file texts, inventories and revisions, then
233
this construct will accept instances of those things.
235
super(Repository, self).__init__()
236
self._format = _format
237
# the following are part of the public API for Repository:
238
self.bzrdir = a_bzrdir
239
self.control_files = control_files
240
self._revision_store = _revision_store
241
self.text_store = text_store
242
# backwards compatibility
243
self.weave_store = text_store
244
# not right yet - should be more semantically clear ?
246
self.control_store = control_store
247
self.control_weaves = control_store
248
# TODO: make sure to construct the right store classes, etc, depending
249
# on whether escaping is required.
250
self._warn_if_deprecated()
251
self._write_group = None
254
return '%s(%r)' % (self.__class__.__name__,
255
self.bzrdir.transport.base)
257
def is_in_write_group(self):
258
"""Return True if there is an open write group.
260
:seealso: start_write_group.
262
return self._write_group is not None
265
return self.control_files.is_locked()
267
def lock_write(self, token=None):
268
"""Lock this repository for writing.
270
:param token: if this is already locked, then lock_write will fail
271
unless the token matches the existing lock.
272
:returns: a token if this instance supports tokens, otherwise None.
273
:raises TokenLockingNotSupported: when a token is given but this
274
instance doesn't support using token locks.
275
:raises MismatchedToken: if the specified token doesn't match the token
276
of the existing lock.
278
A token should be passed in if you know that you have locked the object
279
some other way, and need to synchronise this object's state with that
282
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
284
result = self.control_files.lock_write(token=token)
289
self.control_files.lock_read()
292
def get_physical_lock_status(self):
293
return self.control_files.get_physical_lock_status()
295
def leave_lock_in_place(self):
296
"""Tell this repository not to release the physical lock when this
299
If lock_write doesn't return a token, then this method is not supported.
301
self.control_files.leave_in_place()
303
def dont_leave_lock_in_place(self):
304
"""Tell this repository to release the physical lock when this
305
object is unlocked, even if it didn't originally acquire it.
307
If lock_write doesn't return a token, then this method is not supported.
309
self.control_files.dont_leave_in_place()
312
def gather_stats(self, revid=None, committers=None):
313
"""Gather statistics from a revision id.
315
:param revid: The revision id to gather statistics from, if None, then
316
no revision specific statistics are gathered.
317
:param committers: Optional parameter controlling whether to grab
318
a count of committers from the revision specific statistics.
319
:return: A dictionary of statistics. Currently this contains:
320
committers: The number of committers if requested.
321
firstrev: A tuple with timestamp, timezone for the penultimate left
322
most ancestor of revid, if revid is not the NULL_REVISION.
323
latestrev: A tuple with timestamp, timezone for revid, if revid is
324
not the NULL_REVISION.
325
revisions: The total revision count in the repository.
326
size: An estimate disk size of the repository in bytes.
329
if revid and committers:
330
result['committers'] = 0
331
if revid and revid != _mod_revision.NULL_REVISION:
333
all_committers = set()
334
revisions = self.get_ancestry(revid)
335
# pop the leading None
337
first_revision = None
339
# ignore the revisions in the middle - just grab first and last
340
revisions = revisions[0], revisions[-1]
341
for revision in self.get_revisions(revisions):
342
if not first_revision:
343
first_revision = revision
345
all_committers.add(revision.committer)
346
last_revision = revision
348
result['committers'] = len(all_committers)
349
result['firstrev'] = (first_revision.timestamp,
350
first_revision.timezone)
351
result['latestrev'] = (last_revision.timestamp,
352
last_revision.timezone)
354
# now gather global repository information
355
if self.bzrdir.root_transport.listable():
356
c, t = self._revision_store.total_size(self.get_transaction())
357
result['revisions'] = c
362
def missing_revision_ids(self, other, revision_id=None):
363
"""Return the revision ids that other has that this does not.
365
These are returned in topological order.
367
revision_id: only return revision ids included by revision_id.
369
revision_id = osutils.safe_revision_id(revision_id)
370
return InterRepository.get(other, self).missing_revision_ids(revision_id)
374
"""Open the repository rooted at base.
376
For instance, if the repository is at URL/.bzr/repository,
377
Repository.open(URL) -> a Repository instance.
379
control = bzrdir.BzrDir.open(base)
380
return control.open_repository()
382
def copy_content_into(self, destination, revision_id=None):
383
"""Make a complete copy of the content in self into destination.
385
This is a destructive operation! Do not use it on existing
388
revision_id = osutils.safe_revision_id(revision_id)
389
return InterRepository.get(self, destination).copy_content(revision_id)
391
def commit_write_group(self):
392
"""Commit the contents accrued within the current write group.
394
:seealso: start_write_group.
396
if self._write_group is not self.get_transaction():
397
# has an unlock or relock occured ?
398
raise errors.BzrError('mismatched lock context %r and '
400
(self.get_transaction(), self._write_group))
401
self._commit_write_group()
402
self._write_group = None
404
def _commit_write_group(self):
405
"""Template method for per-repository write group cleanup.
407
This is called before the write group is considered to be
408
finished and should ensure that all data handed to the repository
409
for writing during the write group is safely committed (to the
410
extent possible considering file system caching etc).
413
def fetch(self, source, revision_id=None, pb=None):
414
"""Fetch the content required to construct revision_id from source.
416
If revision_id is None all content is copied.
418
revision_id = osutils.safe_revision_id(revision_id)
419
inter = InterRepository.get(source, self)
421
return inter.fetch(revision_id=revision_id, pb=pb)
422
except NotImplementedError:
423
raise errors.IncompatibleRepositories(source, self)
425
def get_commit_builder(self, branch, parents, config, timestamp=None,
426
timezone=None, committer=None, revprops=None,
428
"""Obtain a CommitBuilder for this repository.
430
:param branch: Branch to commit to.
431
:param parents: Revision ids of the parents of the new revision.
432
:param config: Configuration to use.
433
:param timestamp: Optional timestamp recorded for commit.
434
:param timezone: Optional timezone for timestamp.
435
:param committer: Optional committer to set for commit.
436
:param revprops: Optional dictionary of revision properties.
437
:param revision_id: Optional revision id.
439
revision_id = osutils.safe_revision_id(revision_id)
440
result =_CommitBuilder(self, parents, config, timestamp, timezone,
441
committer, revprops, revision_id)
442
self.start_write_group()
446
if (self.control_files._lock_count == 1 and
447
self.control_files._lock_mode == 'w'):
448
if self._write_group is not None:
449
raise errors.BzrError(
450
'Must end write groups before releasing write locks.')
451
self.control_files.unlock()
454
def clone(self, a_bzrdir, revision_id=None):
455
"""Clone this repository into a_bzrdir using the current format.
457
Currently no check is made that the format of this repository and
458
the bzrdir format are compatible. FIXME RBC 20060201.
460
:return: The newly created destination repository.
462
# TODO: deprecate after 0.16; cloning this with all its settings is
463
# probably not very useful -- mbp 20070423
464
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
465
self.copy_content_into(dest_repo, revision_id)
468
def start_write_group(self):
469
"""Start a write group in the repository.
471
Write groups are used by repositories which do not have a 1:1 mapping
472
between file ids and backend store to manage the insertion of data from
473
both fetch and commit operations.
475
A write lock is required around the start_write_group/commit_write_group
476
for the support of lock-requiring repository formats.
478
if not self.is_locked() or self.control_files._lock_mode != 'w':
479
raise errors.NotWriteLocked(self)
480
if self._write_group:
481
raise errors.BzrError('already in a write group')
482
self._start_write_group()
483
# so we can detect unlock/relock - the write group is now entered.
484
self._write_group = self.get_transaction()
486
def _start_write_group(self):
487
"""Template method for per-repository write group startup.
489
This is called before the write group is considered to be
494
def sprout(self, to_bzrdir, revision_id=None):
495
"""Create a descendent repository for new development.
497
Unlike clone, this does not copy the settings of the repository.
499
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
500
dest_repo.fetch(self, revision_id=revision_id)
503
def _create_sprouting_repo(self, a_bzrdir, shared):
504
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
505
# use target default format.
506
dest_repo = a_bzrdir.create_repository()
508
# Most control formats need the repository to be specifically
509
# created, but on some old all-in-one formats it's not needed
511
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
512
except errors.UninitializableFormat:
513
dest_repo = a_bzrdir.open_repository()
517
def has_revision(self, revision_id):
518
"""True if this repository has a copy of the revision."""
519
revision_id = osutils.safe_revision_id(revision_id)
520
return self._revision_store.has_revision_id(revision_id,
521
self.get_transaction())
524
def get_revision_reconcile(self, revision_id):
525
"""'reconcile' helper routine that allows access to a revision always.
527
This variant of get_revision does not cross check the weave graph
528
against the revision one as get_revision does: but it should only
529
be used by reconcile, or reconcile-alike commands that are correcting
530
or testing the revision graph.
532
if not revision_id or not isinstance(revision_id, basestring):
533
raise errors.InvalidRevisionId(revision_id=revision_id,
535
return self.get_revisions([revision_id])[0]
538
def get_revisions(self, revision_ids):
539
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
540
revs = self._revision_store.get_revisions(revision_ids,
541
self.get_transaction())
543
assert not isinstance(rev.revision_id, unicode)
544
for parent_id in rev.parent_ids:
545
assert not isinstance(parent_id, unicode)
549
def get_revision_xml(self, revision_id):
550
# TODO: jam 20070210 This shouldn't be necessary since get_revision
551
# would have already do it.
552
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
553
revision_id = osutils.safe_revision_id(revision_id)
554
rev = self.get_revision(revision_id)
556
# the current serializer..
557
self._revision_store._serializer.write_revision(rev, rev_tmp)
559
return rev_tmp.getvalue()
562
def get_revision(self, revision_id):
563
"""Return the Revision object for a named revision"""
564
# TODO: jam 20070210 get_revision_reconcile should do this for us
565
revision_id = osutils.safe_revision_id(revision_id)
566
r = self.get_revision_reconcile(revision_id)
567
# weave corruption can lead to absent revision markers that should be
569
# the following test is reasonably cheap (it needs a single weave read)
570
# and the weave is cached in read transactions. In write transactions
571
# it is not cached but typically we only read a small number of
572
# revisions. For knits when they are introduced we will probably want
573
# to ensure that caching write transactions are in use.
574
inv = self.get_inventory_weave()
575
self._check_revision_parents(r, inv)
579
def get_deltas_for_revisions(self, revisions):
580
"""Produce a generator of revision deltas.
582
Note that the input is a sequence of REVISIONS, not revision_ids.
583
Trees will be held in memory until the generator exits.
584
Each delta is relative to the revision's lefthand predecessor.
586
required_trees = set()
587
for revision in revisions:
588
required_trees.add(revision.revision_id)
589
required_trees.update(revision.parent_ids[:1])
590
trees = dict((t.get_revision_id(), t) for
591
t in self.revision_trees(required_trees))
592
for revision in revisions:
593
if not revision.parent_ids:
594
old_tree = self.revision_tree(None)
596
old_tree = trees[revision.parent_ids[0]]
597
yield trees[revision.revision_id].changes_from(old_tree)
600
def get_revision_delta(self, revision_id):
601
"""Return the delta for one revision.
603
The delta is relative to the left-hand predecessor of the
606
r = self.get_revision(revision_id)
607
return list(self.get_deltas_for_revisions([r]))[0]
609
def _check_revision_parents(self, revision, inventory):
610
"""Private to Repository and Fetch.
612
This checks the parentage of revision in an inventory weave for
613
consistency and is only applicable to inventory-weave-for-ancestry
614
using repository formats & fetchers.
616
weave_parents = inventory.get_parents(revision.revision_id)
617
weave_names = inventory.versions()
618
for parent_id in revision.parent_ids:
619
if parent_id in weave_names:
620
# this parent must not be a ghost.
621
if not parent_id in weave_parents:
623
raise errors.CorruptRepository(self)
626
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
627
revision_id = osutils.safe_revision_id(revision_id)
628
signature = gpg_strategy.sign(plaintext)
629
self._revision_store.add_revision_signature_text(revision_id,
631
self.get_transaction())
633
def fileids_altered_by_revision_ids(self, revision_ids):
634
"""Find the file ids and versions affected by revisions.
636
:param revisions: an iterable containing revision ids.
637
:return: a dictionary mapping altered file-ids to an iterable of
638
revision_ids. Each altered file-ids has the exact revision_ids that
639
altered it listed explicitly.
641
assert self._serializer.support_altered_by_hack, \
642
("fileids_altered_by_revision_ids only supported for branches "
643
"which store inventory as unnested xml, not on %r" % self)
644
selected_revision_ids = set(osutils.safe_revision_id(r)
645
for r in revision_ids)
646
w = self.get_inventory_weave()
649
# this code needs to read every new line in every inventory for the
650
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
651
# not present in one of those inventories is unnecessary but not
652
# harmful because we are filtering by the revision id marker in the
653
# inventory lines : we only select file ids altered in one of those
654
# revisions. We don't need to see all lines in the inventory because
655
# only those added in an inventory in rev X can contain a revision=X
657
unescape_revid_cache = {}
658
unescape_fileid_cache = {}
660
# jam 20061218 In a big fetch, this handles hundreds of thousands
661
# of lines, so it has had a lot of inlining and optimizing done.
662
# Sorry that it is a little bit messy.
663
# Move several functions to be local variables, since this is a long
665
search = self._file_ids_altered_regex.search
666
unescape = _unescape_xml
667
setdefault = result.setdefault
668
pb = ui.ui_factory.nested_progress_bar()
670
for line in w.iter_lines_added_or_present_in_versions(
671
selected_revision_ids, pb=pb):
675
# One call to match.group() returning multiple items is quite a
676
# bit faster than 2 calls to match.group() each returning 1
677
file_id, revision_id = match.group('file_id', 'revision_id')
679
# Inlining the cache lookups helps a lot when you make 170,000
680
# lines and 350k ids, versus 8.4 unique ids.
681
# Using a cache helps in 2 ways:
682
# 1) Avoids unnecessary decoding calls
683
# 2) Re-uses cached strings, which helps in future set and
685
# (2) is enough that removing encoding entirely along with
686
# the cache (so we are using plain strings) results in no
687
# performance improvement.
689
revision_id = unescape_revid_cache[revision_id]
691
unescaped = unescape(revision_id)
692
unescape_revid_cache[revision_id] = unescaped
693
revision_id = unescaped
695
if revision_id in selected_revision_ids:
697
file_id = unescape_fileid_cache[file_id]
699
unescaped = unescape(file_id)
700
unescape_fileid_cache[file_id] = unescaped
702
setdefault(file_id, set()).add(revision_id)
708
def get_inventory_weave(self):
709
return self.control_weaves.get_weave('inventory',
710
self.get_transaction())
713
def get_inventory(self, revision_id):
714
"""Get Inventory object by hash."""
715
# TODO: jam 20070210 Technically we don't need to sanitize, since all
716
# called functions must sanitize.
717
revision_id = osutils.safe_revision_id(revision_id)
718
return self.deserialise_inventory(
719
revision_id, self.get_inventory_xml(revision_id))
721
def deserialise_inventory(self, revision_id, xml):
722
"""Transform the xml into an inventory object.
724
:param revision_id: The expected revision id of the inventory.
725
:param xml: A serialised inventory.
727
revision_id = osutils.safe_revision_id(revision_id)
728
result = self._serializer.read_inventory_from_string(xml)
729
result.root.revision = revision_id
732
def serialise_inventory(self, inv):
733
return self._serializer.write_inventory_to_string(inv)
736
def get_inventory_xml(self, revision_id):
737
"""Get inventory XML as a file object."""
738
revision_id = osutils.safe_revision_id(revision_id)
740
assert isinstance(revision_id, str), type(revision_id)
741
iw = self.get_inventory_weave()
742
return iw.get_text(revision_id)
744
raise errors.HistoryMissing(self, 'inventory', revision_id)
747
def get_inventory_sha1(self, revision_id):
748
"""Return the sha1 hash of the inventory entry
750
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
751
revision_id = osutils.safe_revision_id(revision_id)
752
return self.get_revision(revision_id).inventory_sha1
755
def get_revision_graph(self, revision_id=None):
756
"""Return a dictionary containing the revision graph.
758
:param revision_id: The revision_id to get a graph from. If None, then
759
the entire revision graph is returned. This is a deprecated mode of
760
operation and will be removed in the future.
761
:return: a dictionary of revision_id->revision_parents_list.
763
# special case NULL_REVISION
764
if revision_id == _mod_revision.NULL_REVISION:
766
revision_id = osutils.safe_revision_id(revision_id)
767
a_weave = self.get_inventory_weave()
768
all_revisions = self._eliminate_revisions_not_present(
770
entire_graph = dict([(node, tuple(a_weave.get_parents(node))) for
771
node in all_revisions])
772
if revision_id is None:
774
elif revision_id not in entire_graph:
775
raise errors.NoSuchRevision(self, revision_id)
777
# add what can be reached from revision_id
779
pending = set([revision_id])
780
while len(pending) > 0:
782
result[node] = entire_graph[node]
783
for revision_id in result[node]:
784
if revision_id not in result:
785
pending.add(revision_id)
789
def get_revision_graph_with_ghosts(self, revision_ids=None):
790
"""Return a graph of the revisions with ghosts marked as applicable.
792
:param revision_ids: an iterable of revisions to graph or None for all.
793
:return: a Graph object with the graph reachable from revision_ids.
795
result = deprecated_graph.Graph()
797
pending = set(self.all_revision_ids())
800
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
801
# special case NULL_REVISION
802
if _mod_revision.NULL_REVISION in pending:
803
pending.remove(_mod_revision.NULL_REVISION)
804
required = set(pending)
807
revision_id = pending.pop()
809
rev = self.get_revision(revision_id)
810
except errors.NoSuchRevision:
811
if revision_id in required:
814
result.add_ghost(revision_id)
816
for parent_id in rev.parent_ids:
817
# is this queued or done ?
818
if (parent_id not in pending and
819
parent_id not in done):
821
pending.add(parent_id)
822
result.add_node(revision_id, rev.parent_ids)
823
done.add(revision_id)
826
def _get_history_vf(self):
827
"""Get a versionedfile whose history graph reflects all revisions.
829
For weave repositories, this is the inventory weave.
831
return self.get_inventory_weave()
833
def iter_reverse_revision_history(self, revision_id):
834
"""Iterate backwards through revision ids in the lefthand history
836
:param revision_id: The revision id to start with. All its lefthand
837
ancestors will be traversed.
839
revision_id = osutils.safe_revision_id(revision_id)
840
if revision_id in (None, _mod_revision.NULL_REVISION):
842
next_id = revision_id
843
versionedfile = self._get_history_vf()
846
parents = versionedfile.get_parents(next_id)
847
if len(parents) == 0:
853
def get_revision_inventory(self, revision_id):
854
"""Return inventory of a past revision."""
855
# TODO: Unify this with get_inventory()
856
# bzr 0.0.6 and later imposes the constraint that the inventory_id
857
# must be the same as its revision, so this is trivial.
858
if revision_id is None:
859
# This does not make sense: if there is no revision,
860
# then it is the current tree inventory surely ?!
861
# and thus get_root_id() is something that looks at the last
862
# commit on the branch, and the get_root_id is an inventory check.
863
raise NotImplementedError
864
# return Inventory(self.get_root_id())
866
return self.get_inventory(revision_id)
870
"""Return True if this repository is flagged as a shared repository."""
871
raise NotImplementedError(self.is_shared)
874
def reconcile(self, other=None, thorough=False):
875
"""Reconcile this repository."""
876
from bzrlib.reconcile import RepoReconciler
877
reconciler = RepoReconciler(self, thorough=thorough)
878
reconciler.reconcile()
881
def _refresh_data(self):
882
"""Helper called from lock_* to ensure coherency with disk.
884
The default implementation does nothing; it is however possible
885
for repositories to maintain loaded indices across multiple locks
886
by checking inside their implementation of this method to see
887
whether their indices are still valid. This depends of course on
888
the disk format being validatable in this manner.
892
def revision_tree(self, revision_id):
893
"""Return Tree for a revision on this branch.
895
`revision_id` may be None for the empty tree revision.
897
# TODO: refactor this to use an existing revision object
898
# so we don't need to read it in twice.
899
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
900
return RevisionTree(self, Inventory(root_id=None),
901
_mod_revision.NULL_REVISION)
903
revision_id = osutils.safe_revision_id(revision_id)
904
inv = self.get_revision_inventory(revision_id)
905
return RevisionTree(self, inv, revision_id)
908
def revision_trees(self, revision_ids):
909
"""Return Tree for a revision on this branch.
911
`revision_id` may not be None or 'null:'"""
912
assert None not in revision_ids
913
assert _mod_revision.NULL_REVISION not in revision_ids
914
texts = self.get_inventory_weave().get_texts(revision_ids)
915
for text, revision_id in zip(texts, revision_ids):
916
inv = self.deserialise_inventory(revision_id, text)
917
yield RevisionTree(self, inv, revision_id)
920
def get_ancestry(self, revision_id, topo_sorted=True):
921
"""Return a list of revision-ids integrated by a revision.
923
The first element of the list is always None, indicating the origin
924
revision. This might change when we have history horizons, or
925
perhaps we should have a new API.
927
This is topologically sorted.
929
if _mod_revision.is_null(revision_id):
931
revision_id = osutils.safe_revision_id(revision_id)
932
if not self.has_revision(revision_id):
933
raise errors.NoSuchRevision(self, revision_id)
934
w = self.get_inventory_weave()
935
candidates = w.get_ancestry(revision_id, topo_sorted)
936
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
939
"""Compress the data within the repository.
941
This operation only makes sense for some repository types. For other
942
types it should be a no-op that just returns.
944
This stub method does not require a lock, but subclasses should use
945
@needs_write_lock as this is a long running call its reasonable to
946
implicitly lock for the user.
950
def print_file(self, file, revision_id):
951
"""Print `file` to stdout.
953
FIXME RBC 20060125 as John Meinel points out this is a bad api
954
- it writes to stdout, it assumes that that is valid etc. Fix
955
by creating a new more flexible convenience function.
957
revision_id = osutils.safe_revision_id(revision_id)
958
tree = self.revision_tree(revision_id)
959
# use inventory as it was in that revision
960
file_id = tree.inventory.path2id(file)
962
# TODO: jam 20060427 Write a test for this code path
963
# it had a bug in it, and was raising the wrong
965
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
966
tree.print_file(file_id)
968
def get_transaction(self):
969
return self.control_files.get_transaction()
971
def revision_parents(self, revision_id):
972
revision_id = osutils.safe_revision_id(revision_id)
973
return self.get_inventory_weave().parent_names(revision_id)
975
def get_parents(self, revision_ids):
976
"""See StackedParentsProvider.get_parents"""
978
for revision_id in revision_ids:
979
if revision_id == _mod_revision.NULL_REVISION:
983
parents = self.get_revision(revision_id).parent_ids
984
except errors.NoSuchRevision:
987
if len(parents) == 0:
988
parents = [_mod_revision.NULL_REVISION]
989
parents_list.append(parents)
992
def _make_parents_provider(self):
995
def get_graph(self, other_repository=None):
996
"""Return the graph walker for this repository format"""
997
parents_provider = self._make_parents_provider()
998
if (other_repository is not None and
999
other_repository.bzrdir.transport.base !=
1000
self.bzrdir.transport.base):
1001
parents_provider = graph._StackedParentsProvider(
1002
[parents_provider, other_repository._make_parents_provider()])
1003
return graph.Graph(parents_provider)
1006
def set_make_working_trees(self, new_value):
1007
"""Set the policy flag for making working trees when creating branches.
1009
This only applies to branches that use this repository.
1011
The default is 'True'.
1012
:param new_value: True to restore the default, False to disable making
1015
raise NotImplementedError(self.set_make_working_trees)
1017
def make_working_trees(self):
1018
"""Returns the policy for making working trees on new branches."""
1019
raise NotImplementedError(self.make_working_trees)
1022
def sign_revision(self, revision_id, gpg_strategy):
1023
revision_id = osutils.safe_revision_id(revision_id)
1024
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1025
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1028
def has_signature_for_revision_id(self, revision_id):
1029
"""Query for a revision signature for revision_id in the repository."""
1030
revision_id = osutils.safe_revision_id(revision_id)
1031
return self._revision_store.has_signature(revision_id,
1032
self.get_transaction())
1035
def get_signature_text(self, revision_id):
1036
"""Return the text for a signature."""
1037
revision_id = osutils.safe_revision_id(revision_id)
1038
return self._revision_store.get_signature_text(revision_id,
1039
self.get_transaction())
1042
def check(self, revision_ids):
1043
"""Check consistency of all history of given revision_ids.
1045
Different repository implementations should override _check().
1047
:param revision_ids: A non-empty list of revision_ids whose ancestry
1048
will be checked. Typically the last revision_id of a branch.
1050
if not revision_ids:
1051
raise ValueError("revision_ids must be non-empty in %s.check"
1053
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
1054
return self._check(revision_ids)
1056
def _check(self, revision_ids):
1057
result = check.Check(self)
1061
def _warn_if_deprecated(self):
1062
global _deprecation_warning_done
1063
if _deprecation_warning_done:
1065
_deprecation_warning_done = True
1066
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1067
% (self._format, self.bzrdir.transport.base))
1069
def supports_rich_root(self):
1070
return self._format.rich_root_data
1072
def _check_ascii_revisionid(self, revision_id, method):
1073
"""Private helper for ascii-only repositories."""
1074
# weave repositories refuse to store revisionids that are non-ascii.
1075
if revision_id is not None:
1076
# weaves require ascii revision ids.
1077
if isinstance(revision_id, unicode):
1079
revision_id.encode('ascii')
1080
except UnicodeEncodeError:
1081
raise errors.NonAsciiRevisionId(method, self)
1084
revision_id.decode('ascii')
1085
except UnicodeDecodeError:
1086
raise errors.NonAsciiRevisionId(method, self)
1090
# remove these delegates a while after bzr 0.15
1091
def __make_delegated(name, from_module):
1092
def _deprecated_repository_forwarder():
1093
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1094
% (name, from_module),
1097
m = __import__(from_module, globals(), locals(), [name])
1099
return getattr(m, name)
1100
except AttributeError:
1101
raise AttributeError('module %s has no name %s'
1103
globals()[name] = _deprecated_repository_forwarder
1106
'AllInOneRepository',
1107
'WeaveMetaDirRepository',
1108
'PreSplitOutRepositoryFormat',
1109
'RepositoryFormat4',
1110
'RepositoryFormat5',
1111
'RepositoryFormat6',
1112
'RepositoryFormat7',
1114
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1118
'RepositoryFormatKnit',
1119
'RepositoryFormatKnit1',
1121
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1124
def install_revision(repository, rev, revision_tree):
1125
"""Install all revision data into a repository."""
1126
present_parents = []
1128
for p_id in rev.parent_ids:
1129
if repository.has_revision(p_id):
1130
present_parents.append(p_id)
1131
parent_trees[p_id] = repository.revision_tree(p_id)
1133
parent_trees[p_id] = repository.revision_tree(None)
1135
inv = revision_tree.inventory
1136
entries = inv.iter_entries()
1137
# backwards compatability hack: skip the root id.
1138
if not repository.supports_rich_root():
1139
path, root = entries.next()
1140
if root.revision != rev.revision_id:
1141
raise errors.IncompatibleRevision(repr(repository))
1142
# Add the texts that are not already present
1143
for path, ie in entries:
1144
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1145
repository.get_transaction())
1146
if ie.revision not in w:
1148
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1149
# with InventoryEntry.find_previous_heads(). if it is, then there
1150
# is a latent bug here where the parents may have ancestors of each
1152
for revision, tree in parent_trees.iteritems():
1153
if ie.file_id not in tree:
1155
parent_id = tree.inventory[ie.file_id].revision
1156
if parent_id in text_parents:
1158
text_parents.append(parent_id)
1160
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1161
repository.get_transaction())
1162
lines = revision_tree.get_file(ie.file_id).readlines()
1163
vfile.add_lines(rev.revision_id, text_parents, lines)
1165
# install the inventory
1166
repository.add_inventory(rev.revision_id, inv, present_parents)
1167
except errors.RevisionAlreadyPresent:
1169
repository.add_revision(rev.revision_id, rev, inv)
1172
class MetaDirRepository(Repository):
1173
"""Repositories in the new meta-dir layout."""
1175
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1176
super(MetaDirRepository, self).__init__(_format,
1182
dir_mode = self.control_files._dir_mode
1183
file_mode = self.control_files._file_mode
1186
def is_shared(self):
1187
"""Return True if this repository is flagged as a shared repository."""
1188
return self.control_files._transport.has('shared-storage')
1191
def set_make_working_trees(self, new_value):
1192
"""Set the policy flag for making working trees when creating branches.
1194
This only applies to branches that use this repository.
1196
The default is 'True'.
1197
:param new_value: True to restore the default, False to disable making
1202
self.control_files._transport.delete('no-working-trees')
1203
except errors.NoSuchFile:
1206
self.control_files.put_utf8('no-working-trees', '')
1208
def make_working_trees(self):
1209
"""Returns the policy for making working trees on new branches."""
1210
return not self.control_files._transport.has('no-working-trees')
1213
class RepositoryFormatRegistry(registry.Registry):
1214
"""Registry of RepositoryFormats.
1217
def get(self, format_string):
1218
r = registry.Registry.get(self, format_string)
1224
format_registry = RepositoryFormatRegistry()
1225
"""Registry of formats, indexed by their identifying format string.
1227
This can contain either format instances themselves, or classes/factories that
1228
can be called to obtain one.
1232
#####################################################################
1233
# Repository Formats
1235
class RepositoryFormat(object):
1236
"""A repository format.
1238
Formats provide three things:
1239
* An initialization routine to construct repository data on disk.
1240
* a format string which is used when the BzrDir supports versioned
1242
* an open routine which returns a Repository instance.
1244
Formats are placed in an dict by their format string for reference
1245
during opening. These should be subclasses of RepositoryFormat
1248
Once a format is deprecated, just deprecate the initialize and open
1249
methods on the format class. Do not deprecate the object, as the
1250
object will be created every system load.
1252
Common instance attributes:
1253
_matchingbzrdir - the bzrdir format that the repository format was
1254
originally written to work with. This can be used if manually
1255
constructing a bzrdir and repository, or more commonly for test suite
1260
return "<%s>" % self.__class__.__name__
1262
def __eq__(self, other):
1263
# format objects are generally stateless
1264
return isinstance(other, self.__class__)
1266
def __ne__(self, other):
1267
return not self == other
1270
def find_format(klass, a_bzrdir):
1271
"""Return the format for the repository object in a_bzrdir.
1273
This is used by bzr native formats that have a "format" file in
1274
the repository. Other methods may be used by different types of
1278
transport = a_bzrdir.get_repository_transport(None)
1279
format_string = transport.get("format").read()
1280
return format_registry.get(format_string)
1281
except errors.NoSuchFile:
1282
raise errors.NoRepositoryPresent(a_bzrdir)
1284
raise errors.UnknownFormatError(format=format_string)
1287
def register_format(klass, format):
1288
format_registry.register(format.get_format_string(), format)
1291
def unregister_format(klass, format):
1292
format_registry.remove(format.get_format_string())
1295
def get_default_format(klass):
1296
"""Return the current default format."""
1297
from bzrlib import bzrdir
1298
return bzrdir.format_registry.make_bzrdir('default').repository_format
1300
def _get_control_store(self, repo_transport, control_files):
1301
"""Return the control store for this repository."""
1302
raise NotImplementedError(self._get_control_store)
1304
def get_format_string(self):
1305
"""Return the ASCII format string that identifies this format.
1307
Note that in pre format ?? repositories the format string is
1308
not permitted nor written to disk.
1310
raise NotImplementedError(self.get_format_string)
1312
def get_format_description(self):
1313
"""Return the short description for this format."""
1314
raise NotImplementedError(self.get_format_description)
1316
def _get_revision_store(self, repo_transport, control_files):
1317
"""Return the revision store object for this a_bzrdir."""
1318
raise NotImplementedError(self._get_revision_store)
1320
def _get_text_rev_store(self,
1327
"""Common logic for getting a revision store for a repository.
1329
see self._get_revision_store for the subclass-overridable method to
1330
get the store for a repository.
1332
from bzrlib.store.revision.text import TextRevisionStore
1333
dir_mode = control_files._dir_mode
1334
file_mode = control_files._file_mode
1335
text_store = TextStore(transport.clone(name),
1337
compressed=compressed,
1339
file_mode=file_mode)
1340
_revision_store = TextRevisionStore(text_store, serializer)
1341
return _revision_store
1343
# TODO: this shouldn't be in the base class, it's specific to things that
1344
# use weaves or knits -- mbp 20070207
1345
def _get_versioned_file_store(self,
1350
versionedfile_class=None,
1351
versionedfile_kwargs={},
1353
if versionedfile_class is None:
1354
versionedfile_class = self._versionedfile_class
1355
weave_transport = control_files._transport.clone(name)
1356
dir_mode = control_files._dir_mode
1357
file_mode = control_files._file_mode
1358
return VersionedFileStore(weave_transport, prefixed=prefixed,
1360
file_mode=file_mode,
1361
versionedfile_class=versionedfile_class,
1362
versionedfile_kwargs=versionedfile_kwargs,
1365
def initialize(self, a_bzrdir, shared=False):
1366
"""Initialize a repository of this format in a_bzrdir.
1368
:param a_bzrdir: The bzrdir to put the new repository in it.
1369
:param shared: The repository should be initialized as a sharable one.
1370
:returns: The new repository object.
1372
This may raise UninitializableFormat if shared repository are not
1373
compatible the a_bzrdir.
1375
raise NotImplementedError(self.initialize)
1377
def is_supported(self):
1378
"""Is this format supported?
1380
Supported formats must be initializable and openable.
1381
Unsupported formats may not support initialization or committing or
1382
some other features depending on the reason for not being supported.
1386
def check_conversion_target(self, target_format):
1387
raise NotImplementedError(self.check_conversion_target)
1389
def open(self, a_bzrdir, _found=False):
1390
"""Return an instance of this format for the bzrdir a_bzrdir.
1392
_found is a private parameter, do not use it.
1394
raise NotImplementedError(self.open)
1397
class MetaDirRepositoryFormat(RepositoryFormat):
1398
"""Common base class for the new repositories using the metadir layout."""
1400
rich_root_data = False
1401
supports_tree_reference = False
1402
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1405
super(MetaDirRepositoryFormat, self).__init__()
1407
def _create_control_files(self, a_bzrdir):
1408
"""Create the required files and the initial control_files object."""
1409
# FIXME: RBC 20060125 don't peek under the covers
1410
# NB: no need to escape relative paths that are url safe.
1411
repository_transport = a_bzrdir.get_repository_transport(self)
1412
control_files = lockable_files.LockableFiles(repository_transport,
1413
'lock', lockdir.LockDir)
1414
control_files.create_lock()
1415
return control_files
1417
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1418
"""Upload the initial blank content."""
1419
control_files = self._create_control_files(a_bzrdir)
1420
control_files.lock_write()
1422
control_files._transport.mkdir_multi(dirs,
1423
mode=control_files._dir_mode)
1424
for file, content in files:
1425
control_files.put(file, content)
1426
for file, content in utf8_files:
1427
control_files.put_utf8(file, content)
1429
control_files.put_utf8('shared-storage', '')
1431
control_files.unlock()
1434
# formats which have no format string are not discoverable
1435
# and not independently creatable, so are not registered. They're
1436
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1437
# needed, it's constructed directly by the BzrDir. Non-native formats where
1438
# the repository is not separately opened are similar.
1440
format_registry.register_lazy(
1441
'Bazaar-NG Repository format 7',
1442
'bzrlib.repofmt.weaverepo',
1446
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1447
# default control directory format
1448
format_registry.register_lazy(
1449
'Bazaar-NG Knit Repository Format 1',
1450
'bzrlib.repofmt.knitrepo',
1451
'RepositoryFormatKnit1',
1453
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1455
format_registry.register_lazy(
1456
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1457
'bzrlib.repofmt.knitrepo',
1458
'RepositoryFormatKnit3',
1461
# Experimental formats. These make no guarantee about data stability.
1462
# There is one format for pre-subtrees, and one for post-subtrees to
1463
# allow ease of testing.
1464
format_registry.register_lazy(
1465
'Bazaar Experimental no-subtrees\n',
1466
'bzrlib.repofmt.knitrepo',
1467
'RepositoryFormatGraphKnit1',
1469
format_registry.register_lazy(
1470
'Bazaar Experimental subtrees\n',
1471
'bzrlib.repofmt.knitrepo',
1472
'RepositoryFormatGraphKnit3',
1476
class InterRepository(InterObject):
1477
"""This class represents operations taking place between two repositories.
1479
Its instances have methods like copy_content and fetch, and contain
1480
references to the source and target repositories these operations can be
1483
Often we will provide convenience methods on 'repository' which carry out
1484
operations with another repository - they will always forward to
1485
InterRepository.get(other).method_name(parameters).
1489
"""The available optimised InterRepository types."""
1491
def copy_content(self, revision_id=None):
1492
raise NotImplementedError(self.copy_content)
1494
def fetch(self, revision_id=None, pb=None):
1495
"""Fetch the content required to construct revision_id.
1497
The content is copied from self.source to self.target.
1499
:param revision_id: if None all content is copied, if NULL_REVISION no
1501
:param pb: optional progress bar to use for progress reports. If not
1502
provided a default one will be created.
1504
Returns the copied revision count and the failed revisions in a tuple:
1507
raise NotImplementedError(self.fetch)
1510
def missing_revision_ids(self, revision_id=None):
1511
"""Return the revision ids that source has that target does not.
1513
These are returned in topological order.
1515
:param revision_id: only return revision ids included by this
1518
# generic, possibly worst case, slow code path.
1519
target_ids = set(self.target.all_revision_ids())
1520
if revision_id is not None:
1521
# TODO: jam 20070210 InterRepository is internal enough that it
1522
# should assume revision_ids are already utf-8
1523
revision_id = osutils.safe_revision_id(revision_id)
1524
source_ids = self.source.get_ancestry(revision_id)
1525
assert source_ids[0] is None
1528
source_ids = self.source.all_revision_ids()
1529
result_set = set(source_ids).difference(target_ids)
1530
# this may look like a no-op: its not. It preserves the ordering
1531
# other_ids had while only returning the members from other_ids
1532
# that we've decided we need.
1533
return [rev_id for rev_id in source_ids if rev_id in result_set]
1536
def _same_model(source, target):
1537
"""True if source and target have the same data representation."""
1538
if source.supports_rich_root() != target.supports_rich_root():
1540
if source._serializer != target._serializer:
1545
class InterSameDataRepository(InterRepository):
1546
"""Code for converting between repositories that represent the same data.
1548
Data format and model must match for this to work.
1552
def _get_repo_format_to_test(self):
1553
"""Repository format for testing with."""
1554
return RepositoryFormat.get_default_format()
1557
def is_compatible(source, target):
1558
return InterRepository._same_model(source, target)
1561
def copy_content(self, revision_id=None):
1562
"""Make a complete copy of the content in self into destination.
1564
This copies both the repository's revision data, and configuration information
1565
such as the make_working_trees setting.
1567
This is a destructive operation! Do not use it on existing
1570
:param revision_id: Only copy the content needed to construct
1571
revision_id and its parents.
1574
self.target.set_make_working_trees(self.source.make_working_trees())
1575
except NotImplementedError:
1577
# TODO: jam 20070210 This is fairly internal, so we should probably
1578
# just assert that revision_id is not unicode.
1579
revision_id = osutils.safe_revision_id(revision_id)
1580
# but don't bother fetching if we have the needed data now.
1581
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1582
self.target.has_revision(revision_id)):
1584
self.target.fetch(self.source, revision_id=revision_id)
1587
def fetch(self, revision_id=None, pb=None):
1588
"""See InterRepository.fetch()."""
1589
from bzrlib.fetch import GenericRepoFetcher
1590
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1591
self.source, self.source._format, self.target,
1592
self.target._format)
1593
# TODO: jam 20070210 This should be an assert, not a translate
1594
revision_id = osutils.safe_revision_id(revision_id)
1595
f = GenericRepoFetcher(to_repository=self.target,
1596
from_repository=self.source,
1597
last_revision=revision_id,
1599
return f.count_copied, f.failed_revisions
1602
class InterWeaveRepo(InterSameDataRepository):
1603
"""Optimised code paths between Weave based repositories."""
1606
def _get_repo_format_to_test(self):
1607
from bzrlib.repofmt import weaverepo
1608
return weaverepo.RepositoryFormat7()
1611
def is_compatible(source, target):
1612
"""Be compatible with known Weave formats.
1614
We don't test for the stores being of specific types because that
1615
could lead to confusing results, and there is no need to be
1618
from bzrlib.repofmt.weaverepo import (
1624
return (isinstance(source._format, (RepositoryFormat5,
1626
RepositoryFormat7)) and
1627
isinstance(target._format, (RepositoryFormat5,
1629
RepositoryFormat7)))
1630
except AttributeError:
1634
def copy_content(self, revision_id=None):
1635
"""See InterRepository.copy_content()."""
1636
# weave specific optimised path:
1637
# TODO: jam 20070210 Internal, should be an assert, not translate
1638
revision_id = osutils.safe_revision_id(revision_id)
1640
self.target.set_make_working_trees(self.source.make_working_trees())
1641
except NotImplementedError:
1643
# FIXME do not peek!
1644
if self.source.control_files._transport.listable():
1645
pb = ui.ui_factory.nested_progress_bar()
1647
self.target.weave_store.copy_all_ids(
1648
self.source.weave_store,
1650
from_transaction=self.source.get_transaction(),
1651
to_transaction=self.target.get_transaction())
1652
pb.update('copying inventory', 0, 1)
1653
self.target.control_weaves.copy_multi(
1654
self.source.control_weaves, ['inventory'],
1655
from_transaction=self.source.get_transaction(),
1656
to_transaction=self.target.get_transaction())
1657
self.target._revision_store.text_store.copy_all_ids(
1658
self.source._revision_store.text_store,
1663
self.target.fetch(self.source, revision_id=revision_id)
1666
def fetch(self, revision_id=None, pb=None):
1667
"""See InterRepository.fetch()."""
1668
from bzrlib.fetch import GenericRepoFetcher
1669
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1670
self.source, self.source._format, self.target, self.target._format)
1671
# TODO: jam 20070210 This should be an assert, not a translate
1672
revision_id = osutils.safe_revision_id(revision_id)
1673
f = GenericRepoFetcher(to_repository=self.target,
1674
from_repository=self.source,
1675
last_revision=revision_id,
1677
return f.count_copied, f.failed_revisions
1680
def missing_revision_ids(self, revision_id=None):
1681
"""See InterRepository.missing_revision_ids()."""
1682
# we want all revisions to satisfy revision_id in source.
1683
# but we don't want to stat every file here and there.
1684
# we want then, all revisions other needs to satisfy revision_id
1685
# checked, but not those that we have locally.
1686
# so the first thing is to get a subset of the revisions to
1687
# satisfy revision_id in source, and then eliminate those that
1688
# we do already have.
1689
# this is slow on high latency connection to self, but as as this
1690
# disk format scales terribly for push anyway due to rewriting
1691
# inventory.weave, this is considered acceptable.
1693
if revision_id is not None:
1694
source_ids = self.source.get_ancestry(revision_id)
1695
assert source_ids[0] is None
1698
source_ids = self.source._all_possible_ids()
1699
source_ids_set = set(source_ids)
1700
# source_ids is the worst possible case we may need to pull.
1701
# now we want to filter source_ids against what we actually
1702
# have in target, but don't try to check for existence where we know
1703
# we do not have a revision as that would be pointless.
1704
target_ids = set(self.target._all_possible_ids())
1705
possibly_present_revisions = target_ids.intersection(source_ids_set)
1706
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1707
required_revisions = source_ids_set.difference(actually_present_revisions)
1708
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1709
if revision_id is not None:
1710
# we used get_ancestry to determine source_ids then we are assured all
1711
# revisions referenced are present as they are installed in topological order.
1712
# and the tip revision was validated by get_ancestry.
1713
return required_topo_revisions
1715
# if we just grabbed the possibly available ids, then
1716
# we only have an estimate of whats available and need to validate
1717
# that against the revision records.
1718
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1721
class InterKnitRepo(InterSameDataRepository):
1722
"""Optimised code paths between Knit based repositories."""
1725
def _get_repo_format_to_test(self):
1726
from bzrlib.repofmt import knitrepo
1727
return knitrepo.RepositoryFormatKnit1()
1730
def is_compatible(source, target):
1731
"""Be compatible with known Knit formats.
1733
We don't test for the stores being of specific types because that
1734
could lead to confusing results, and there is no need to be
1737
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
1739
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
1740
isinstance(target._format, RepositoryFormatKnit))
1741
except AttributeError:
1743
return are_knits and InterRepository._same_model(source, target)
1746
def fetch(self, revision_id=None, pb=None):
1747
"""See InterRepository.fetch()."""
1748
from bzrlib.fetch import KnitRepoFetcher
1749
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1750
self.source, self.source._format, self.target, self.target._format)
1751
# TODO: jam 20070210 This should be an assert, not a translate
1752
revision_id = osutils.safe_revision_id(revision_id)
1753
f = KnitRepoFetcher(to_repository=self.target,
1754
from_repository=self.source,
1755
last_revision=revision_id,
1757
return f.count_copied, f.failed_revisions
1760
def missing_revision_ids(self, revision_id=None):
1761
"""See InterRepository.missing_revision_ids()."""
1762
if revision_id is not None:
1763
source_ids = self.source.get_ancestry(revision_id)
1764
assert source_ids[0] is None
1767
source_ids = self.source._all_possible_ids()
1768
source_ids_set = set(source_ids)
1769
# source_ids is the worst possible case we may need to pull.
1770
# now we want to filter source_ids against what we actually
1771
# have in target, but don't try to check for existence where we know
1772
# we do not have a revision as that would be pointless.
1773
target_ids = set(self.target._all_possible_ids())
1774
possibly_present_revisions = target_ids.intersection(source_ids_set)
1775
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1776
required_revisions = source_ids_set.difference(actually_present_revisions)
1777
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1778
if revision_id is not None:
1779
# we used get_ancestry to determine source_ids then we are assured all
1780
# revisions referenced are present as they are installed in topological order.
1781
# and the tip revision was validated by get_ancestry.
1782
return required_topo_revisions
1784
# if we just grabbed the possibly available ids, then
1785
# we only have an estimate of whats available and need to validate
1786
# that against the revision records.
1787
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1790
class InterModel1and2(InterRepository):
1793
def _get_repo_format_to_test(self):
1797
def is_compatible(source, target):
1798
if not source.supports_rich_root() and target.supports_rich_root():
1804
def fetch(self, revision_id=None, pb=None):
1805
"""See InterRepository.fetch()."""
1806
from bzrlib.fetch import Model1toKnit2Fetcher
1807
# TODO: jam 20070210 This should be an assert, not a translate
1808
revision_id = osutils.safe_revision_id(revision_id)
1809
f = Model1toKnit2Fetcher(to_repository=self.target,
1810
from_repository=self.source,
1811
last_revision=revision_id,
1813
return f.count_copied, f.failed_revisions
1816
def copy_content(self, revision_id=None):
1817
"""Make a complete copy of the content in self into destination.
1819
This is a destructive operation! Do not use it on existing
1822
:param revision_id: Only copy the content needed to construct
1823
revision_id and its parents.
1826
self.target.set_make_working_trees(self.source.make_working_trees())
1827
except NotImplementedError:
1829
# TODO: jam 20070210 Internal, assert, don't translate
1830
revision_id = osutils.safe_revision_id(revision_id)
1831
# but don't bother fetching if we have the needed data now.
1832
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1833
self.target.has_revision(revision_id)):
1835
self.target.fetch(self.source, revision_id=revision_id)
1838
class InterKnit1and2(InterKnitRepo):
1841
def _get_repo_format_to_test(self):
1845
def is_compatible(source, target):
1846
"""Be compatible with Knit1 source and Knit3 target"""
1847
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
1849
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1850
RepositoryFormatKnit3
1851
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1852
isinstance(target._format, (RepositoryFormatKnit3)))
1853
except AttributeError:
1857
def fetch(self, revision_id=None, pb=None):
1858
"""See InterRepository.fetch()."""
1859
from bzrlib.fetch import Knit1to2Fetcher
1860
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1861
self.source, self.source._format, self.target,
1862
self.target._format)
1863
# TODO: jam 20070210 This should be an assert, not a translate
1864
revision_id = osutils.safe_revision_id(revision_id)
1865
f = Knit1to2Fetcher(to_repository=self.target,
1866
from_repository=self.source,
1867
last_revision=revision_id,
1869
return f.count_copied, f.failed_revisions
1872
class InterRemoteRepository(InterRepository):
1873
"""Code for converting between RemoteRepository objects.
1875
This just gets an non-remote repository from the RemoteRepository, and calls
1876
InterRepository.get again.
1879
def __init__(self, source, target):
1880
if isinstance(source, remote.RemoteRepository):
1881
source._ensure_real()
1882
real_source = source._real_repository
1884
real_source = source
1885
if isinstance(target, remote.RemoteRepository):
1886
target._ensure_real()
1887
real_target = target._real_repository
1889
real_target = target
1890
self.real_inter = InterRepository.get(real_source, real_target)
1893
def is_compatible(source, target):
1894
if isinstance(source, remote.RemoteRepository):
1896
if isinstance(target, remote.RemoteRepository):
1900
def copy_content(self, revision_id=None):
1901
self.real_inter.copy_content(revision_id=revision_id)
1903
def fetch(self, revision_id=None, pb=None):
1904
self.real_inter.fetch(revision_id=revision_id, pb=pb)
1907
def _get_repo_format_to_test(self):
1911
InterRepository.register_optimiser(InterSameDataRepository)
1912
InterRepository.register_optimiser(InterWeaveRepo)
1913
InterRepository.register_optimiser(InterKnitRepo)
1914
InterRepository.register_optimiser(InterModel1and2)
1915
InterRepository.register_optimiser(InterKnit1and2)
1916
InterRepository.register_optimiser(InterRemoteRepository)
1919
class CopyConverter(object):
1920
"""A repository conversion tool which just performs a copy of the content.
1922
This is slow but quite reliable.
1925
def __init__(self, target_format):
1926
"""Create a CopyConverter.
1928
:param target_format: The format the resulting repository should be.
1930
self.target_format = target_format
1932
def convert(self, repo, pb):
1933
"""Perform the conversion of to_convert, giving feedback via pb.
1935
:param to_convert: The disk object to convert.
1936
:param pb: a progress bar to use for progress information.
1941
# this is only useful with metadir layouts - separated repo content.
1942
# trigger an assertion if not such
1943
repo._format.get_format_string()
1944
self.repo_dir = repo.bzrdir
1945
self.step('Moving repository to repository.backup')
1946
self.repo_dir.transport.move('repository', 'repository.backup')
1947
backup_transport = self.repo_dir.transport.clone('repository.backup')
1948
repo._format.check_conversion_target(self.target_format)
1949
self.source_repo = repo._format.open(self.repo_dir,
1951
_override_transport=backup_transport)
1952
self.step('Creating new repository')
1953
converted = self.target_format.initialize(self.repo_dir,
1954
self.source_repo.is_shared())
1955
converted.lock_write()
1957
self.step('Copying content into repository.')
1958
self.source_repo.copy_content_into(converted)
1961
self.step('Deleting old repository content.')
1962
self.repo_dir.transport.delete_tree('repository.backup')
1963
self.pb.note('repository converted')
1965
def step(self, message):
1966
"""Update the pb by a step."""
1968
self.pb.update(message, self.count, self.total)
1971
class CommitBuilder(object):
1972
"""Provides an interface to build up a commit.
1974
This allows describing a tree to be committed without needing to
1975
know the internals of the format of the repository.
1978
record_root_entry = False
1979
def __init__(self, repository, parents, config, timestamp=None,
1980
timezone=None, committer=None, revprops=None,
1982
"""Initiate a CommitBuilder.
1984
:param repository: Repository to commit to.
1985
:param parents: Revision ids of the parents of the new revision.
1986
:param config: Configuration to use.
1987
:param timestamp: Optional timestamp recorded for commit.
1988
:param timezone: Optional timezone for timestamp.
1989
:param committer: Optional committer to set for commit.
1990
:param revprops: Optional dictionary of revision properties.
1991
:param revision_id: Optional revision id.
1993
self._config = config
1995
if committer is None:
1996
self._committer = self._config.username()
1998
assert isinstance(committer, basestring), type(committer)
1999
self._committer = committer
2001
self.new_inventory = Inventory(None)
2002
self._new_revision_id = osutils.safe_revision_id(revision_id)
2003
self.parents = parents
2004
self.repository = repository
2007
if revprops is not None:
2008
self._revprops.update(revprops)
2010
if timestamp is None:
2011
timestamp = time.time()
2012
# Restrict resolution to 1ms
2013
self._timestamp = round(timestamp, 3)
2015
if timezone is None:
2016
self._timezone = osutils.local_time_offset()
2018
self._timezone = int(timezone)
2020
self._generate_revision_if_needed()
2022
def commit(self, message):
2023
"""Make the actual commit.
2025
:return: The revision id of the recorded revision.
2027
rev = _mod_revision.Revision(
2028
timestamp=self._timestamp,
2029
timezone=self._timezone,
2030
committer=self._committer,
2032
inventory_sha1=self.inv_sha1,
2033
revision_id=self._new_revision_id,
2034
properties=self._revprops)
2035
rev.parent_ids = self.parents
2036
self.repository.add_revision(self._new_revision_id, rev,
2037
self.new_inventory, self._config)
2038
self.repository.commit_write_group()
2039
return self._new_revision_id
2041
def revision_tree(self):
2042
"""Return the tree that was just committed.
2044
After calling commit() this can be called to get a RevisionTree
2045
representing the newly committed tree. This is preferred to
2046
calling Repository.revision_tree() because that may require
2047
deserializing the inventory, while we already have a copy in
2050
return RevisionTree(self.repository, self.new_inventory,
2051
self._new_revision_id)
2053
def finish_inventory(self):
2054
"""Tell the builder that the inventory is finished."""
2055
if self.new_inventory.root is None:
2056
symbol_versioning.warn('Root entry should be supplied to'
2057
' record_entry_contents, as of bzr 0.10.',
2058
DeprecationWarning, stacklevel=2)
2059
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
2060
self.new_inventory.revision_id = self._new_revision_id
2061
self.inv_sha1 = self.repository.add_inventory(
2062
self._new_revision_id,
2067
def _gen_revision_id(self):
2068
"""Return new revision-id."""
2069
return generate_ids.gen_revision_id(self._config.username(),
2072
def _generate_revision_if_needed(self):
2073
"""Create a revision id if None was supplied.
2075
If the repository can not support user-specified revision ids
2076
they should override this function and raise CannotSetRevisionId
2077
if _new_revision_id is not None.
2079
:raises: CannotSetRevisionId
2081
if self._new_revision_id is None:
2082
self._new_revision_id = self._gen_revision_id()
2084
def record_entry_contents(self, ie, parent_invs, path, tree):
2085
"""Record the content of ie from tree into the commit if needed.
2087
Side effect: sets ie.revision when unchanged
2089
:param ie: An inventory entry present in the commit.
2090
:param parent_invs: The inventories of the parent revisions of the
2092
:param path: The path the entry is at in the tree.
2093
:param tree: The tree which contains this entry and should be used to
2096
if self.new_inventory.root is None and ie.parent_id is not None:
2097
symbol_versioning.warn('Root entry should be supplied to'
2098
' record_entry_contents, as of bzr 0.10.',
2099
DeprecationWarning, stacklevel=2)
2100
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
2102
self.new_inventory.add(ie)
2104
# ie.revision is always None if the InventoryEntry is considered
2105
# for committing. ie.snapshot will record the correct revision
2106
# which may be the sole parent if it is untouched.
2107
if ie.revision is not None:
2110
# In this revision format, root entries have no knit or weave
2111
if ie is self.new_inventory.root:
2112
# When serializing out to disk and back in
2113
# root.revision is always _new_revision_id
2114
ie.revision = self._new_revision_id
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)
2124
def modified_directory(self, file_id, file_parents):
2125
"""Record the presence of a symbolic link.
2127
:param file_id: The file_id of the link to record.
2128
:param file_parents: The per-file parent revision ids.
2130
self._add_text_to_weave(file_id, [], file_parents.keys())
2132
def modified_reference(self, file_id, file_parents):
2133
"""Record the modification of a reference.
2135
:param file_id: The file_id of the link to record.
2136
:param file_parents: The per-file parent revision ids.
2138
self._add_text_to_weave(file_id, [], file_parents.keys())
2140
def modified_file_text(self, file_id, file_parents,
2141
get_content_byte_lines, text_sha1=None,
2143
"""Record the text of file file_id
2145
:param file_id: The file_id of the file to record the text of.
2146
:param file_parents: The per-file parent revision ids.
2147
:param get_content_byte_lines: A callable which will return the byte
2149
:param text_sha1: Optional SHA1 of the file contents.
2150
:param text_size: Optional size of the file contents.
2152
# mutter('storing text of file {%s} in revision {%s} into %r',
2153
# file_id, self._new_revision_id, self.repository.weave_store)
2154
# special case to avoid diffing on renames or
2156
if (len(file_parents) == 1
2157
and text_sha1 == file_parents.values()[0].text_sha1
2158
and text_size == file_parents.values()[0].text_size):
2159
previous_ie = file_parents.values()[0]
2160
versionedfile = self.repository.weave_store.get_weave(file_id,
2161
self.repository.get_transaction())
2162
versionedfile.clone_text(self._new_revision_id,
2163
previous_ie.revision, file_parents.keys())
2164
return text_sha1, text_size
2166
new_lines = get_content_byte_lines()
2167
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2168
# should return the SHA1 and size
2169
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2170
return osutils.sha_strings(new_lines), \
2171
sum(map(len, new_lines))
2173
def modified_link(self, file_id, file_parents, link_target):
2174
"""Record the presence of a symbolic link.
2176
:param file_id: The file_id of the link to record.
2177
:param file_parents: The per-file parent revision ids.
2178
:param link_target: Target location of this link.
2180
self._add_text_to_weave(file_id, [], file_parents.keys())
2182
def _add_text_to_weave(self, file_id, new_lines, parents):
2183
versionedfile = self.repository.weave_store.get_weave_or_empty(
2184
file_id, self.repository.get_transaction())
2185
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2186
versionedfile.clear_cache()
2189
class _CommitBuilder(CommitBuilder):
2190
"""Temporary class so old CommitBuilders are detected properly
2192
Note: CommitBuilder works whether or not root entry is recorded.
2195
record_root_entry = True
2198
class RootCommitBuilder(CommitBuilder):
2199
"""This commitbuilder actually records the root id"""
2201
record_root_entry = True
2203
def record_entry_contents(self, ie, parent_invs, path, tree):
2204
"""Record the content of ie from tree into the commit if needed.
2206
Side effect: sets ie.revision when unchanged
2208
:param ie: An inventory entry present in the commit.
2209
:param parent_invs: The inventories of the parent revisions of the
2211
:param path: The path the entry is at in the tree.
2212
:param tree: The tree which contains this entry and should be used to
2215
assert self.new_inventory.root is not None or ie.parent_id is None
2216
self.new_inventory.add(ie)
2218
# ie.revision is always None if the InventoryEntry is considered
2219
# for committing. ie.snapshot will record the correct revision
2220
# which may be the sole parent if it is untouched.
2221
if ie.revision is not None:
2224
previous_entries = ie.find_previous_heads(
2226
self.repository.weave_store,
2227
self.repository.get_transaction())
2228
# we are creating a new revision for ie in the history store
2230
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2242
def _unescaper(match, _map=_unescape_map):
2243
code = match.group(1)
2247
if not code.startswith('#'):
2249
return unichr(int(code[1:])).encode('utf8')
2255
def _unescape_xml(data):
2256
"""Unescape predefined XML entities in a string of data."""
2258
if _unescape_re is None:
2259
_unescape_re = re.compile('\&([^;]*);')
2260
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
