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(), """
37
revision as _mod_revision,
42
from bzrlib.revisiontree import RevisionTree
43
from bzrlib.store.versioned import VersionedFileStore
44
from bzrlib.store.text import TextStore
45
from bzrlib.testament import Testament
49
from bzrlib.decorators import needs_read_lock, needs_write_lock
50
from bzrlib.inter import InterObject
51
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
52
from bzrlib.symbol_versioning import (
56
from bzrlib.trace import mutter, note, warning
59
# Old formats display a warning, but only once
60
_deprecation_warning_done = False
63
class Repository(object):
64
"""Repository holding history for one or more branches.
66
The repository holds and retrieves historical information including
67
revisions and file history. It's normally accessed only by the Branch,
68
which views a particular line of development through that history.
70
The Repository builds on top of Stores and a Transport, which respectively
71
describe the disk data format and the way of accessing the (possibly
75
_file_ids_altered_regex = lazy_regex.lazy_compile(
76
r'file_id="(?P<file_id>[^"]+)"'
77
r'.*revision="(?P<revision_id>[^"]+)"'
81
def add_inventory(self, revision_id, inv, parents):
82
"""Add the inventory inv to the repository as revision_id.
84
:param parents: The revision ids of the parents that revision_id
85
is known to have and are in the repository already.
87
returns the sha1 of the serialized inventory.
89
revision_id = osutils.safe_revision_id(revision_id)
90
_mod_revision.check_not_reserved_id(revision_id)
91
assert inv.revision_id is None or inv.revision_id == revision_id, \
92
"Mismatch between inventory revision" \
93
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
94
assert inv.root is not None
95
inv_text = self.serialise_inventory(inv)
96
inv_sha1 = osutils.sha_string(inv_text)
97
inv_vf = self.control_weaves.get_weave('inventory',
98
self.get_transaction())
99
self._inventory_add_lines(inv_vf, revision_id, parents,
100
osutils.split_lines(inv_text))
103
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines):
105
for parent in parents:
107
final_parents.append(parent)
109
inv_vf.add_lines(revision_id, final_parents, lines)
112
def add_revision(self, revision_id, rev, inv=None, config=None):
113
"""Add rev to the revision store as revision_id.
115
:param revision_id: the revision id to use.
116
:param rev: The revision object.
117
:param inv: The inventory for the revision. if None, it will be looked
118
up in the inventory storer
119
:param config: If None no digital signature will be created.
120
If supplied its signature_needed method will be used
121
to determine if a signature should be made.
123
revision_id = osutils.safe_revision_id(revision_id)
124
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
126
_mod_revision.check_not_reserved_id(revision_id)
127
if config is not None and config.signature_needed():
129
inv = self.get_inventory(revision_id)
130
plaintext = Testament(rev, inv).as_short_text()
131
self.store_revision_signature(
132
gpg.GPGStrategy(config), plaintext, revision_id)
133
if not revision_id in self.get_inventory_weave():
135
raise errors.WeaveRevisionNotPresent(revision_id,
136
self.get_inventory_weave())
138
# yes, this is not suitable for adding with ghosts.
139
self.add_inventory(revision_id, inv, rev.parent_ids)
140
self._revision_store.add_revision(rev, self.get_transaction())
143
def _all_possible_ids(self):
144
"""Return all the possible revisions that we could find."""
145
return self.get_inventory_weave().versions()
147
def all_revision_ids(self):
148
"""Returns a list of all the revision ids in the repository.
150
This is deprecated because code should generally work on the graph
151
reachable from a particular revision, and ignore any other revisions
152
that might be present. There is no direct replacement method.
154
return self._all_revision_ids()
157
def _all_revision_ids(self):
158
"""Returns a list of all the revision ids in the repository.
160
These are in as much topological order as the underlying store can
161
present: for weaves ghosts may lead to a lack of correctness until
162
the reweave updates the parents list.
164
if self._revision_store.text_store.listable():
165
return self._revision_store.all_revision_ids(self.get_transaction())
166
result = self._all_possible_ids()
167
# TODO: jam 20070210 Ensure that _all_possible_ids returns non-unicode
168
# ids. (It should, since _revision_store's API should change to
169
# return utf8 revision_ids)
170
return self._eliminate_revisions_not_present(result)
172
def break_lock(self):
173
"""Break a lock if one is present from another instance.
175
Uses the ui factory to ask for confirmation if the lock may be from
178
self.control_files.break_lock()
181
def _eliminate_revisions_not_present(self, revision_ids):
182
"""Check every revision id in revision_ids to see if we have it.
184
Returns a set of the present revisions.
187
for id in revision_ids:
188
if self.has_revision(id):
193
def create(a_bzrdir):
194
"""Construct the current default format repository in a_bzrdir."""
195
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
197
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
198
"""instantiate a Repository.
200
:param _format: The format of the repository on disk.
201
:param a_bzrdir: The BzrDir of the repository.
203
In the future we will have a single api for all stores for
204
getting file texts, inventories and revisions, then
205
this construct will accept instances of those things.
207
super(Repository, self).__init__()
208
self._format = _format
209
# the following are part of the public API for Repository:
210
self.bzrdir = a_bzrdir
211
self.control_files = control_files
212
self._revision_store = _revision_store
213
self.text_store = text_store
214
# backwards compatibility
215
self.weave_store = text_store
216
# not right yet - should be more semantically clear ?
218
self.control_store = control_store
219
self.control_weaves = control_store
220
# TODO: make sure to construct the right store classes, etc, depending
221
# on whether escaping is required.
222
self._warn_if_deprecated()
225
return '%s(%r)' % (self.__class__.__name__,
226
self.bzrdir.transport.base)
229
return self.control_files.is_locked()
231
def lock_write(self):
232
self.control_files.lock_write()
235
self.control_files.lock_read()
237
def get_physical_lock_status(self):
238
return self.control_files.get_physical_lock_status()
241
def gather_stats(self, revid=None, committers=None):
242
"""Gather statistics from a revision id.
244
:param revid: The revision id to gather statistics from, if None, then
245
no revision specific statistics are gathered.
246
:param committers: Optional parameter controlling whether to grab
247
a count of committers from the revision specific statistics.
248
:return: A dictionary of statistics. Currently this contains:
249
committers: The number of committers if requested.
250
firstrev: A tuple with timestamp, timezone for the penultimate left
251
most ancestor of revid, if revid is not the NULL_REVISION.
252
latestrev: A tuple with timestamp, timezone for revid, if revid is
253
not the NULL_REVISION.
254
revisions: The total revision count in the repository.
255
size: An estimate disk size of the repository in bytes.
258
if revid and committers:
259
result['committers'] = 0
260
if revid and revid != _mod_revision.NULL_REVISION:
262
all_committers = set()
263
revisions = self.get_ancestry(revid)
264
# pop the leading None
266
first_revision = None
268
# ignore the revisions in the middle - just grab first and last
269
revisions = revisions[0], revisions[-1]
270
for revision in self.get_revisions(revisions):
271
if not first_revision:
272
first_revision = revision
274
all_committers.add(revision.committer)
275
last_revision = revision
277
result['committers'] = len(all_committers)
278
result['firstrev'] = (first_revision.timestamp,
279
first_revision.timezone)
280
result['latestrev'] = (last_revision.timestamp,
281
last_revision.timezone)
283
# now gather global repository information
284
if self.bzrdir.root_transport.listable():
285
c, t = self._revision_store.total_size(self.get_transaction())
286
result['revisions'] = c
291
def missing_revision_ids(self, other, revision_id=None):
292
"""Return the revision ids that other has that this does not.
294
These are returned in topological order.
296
revision_id: only return revision ids included by revision_id.
298
revision_id = osutils.safe_revision_id(revision_id)
299
return InterRepository.get(other, self).missing_revision_ids(revision_id)
303
"""Open the repository rooted at base.
305
For instance, if the repository is at URL/.bzr/repository,
306
Repository.open(URL) -> a Repository instance.
308
control = bzrdir.BzrDir.open(base)
309
return control.open_repository()
311
def copy_content_into(self, destination, revision_id=None, basis=None):
312
"""Make a complete copy of the content in self into destination.
314
This is a destructive operation! Do not use it on existing
317
revision_id = osutils.safe_revision_id(revision_id)
318
return InterRepository.get(self, destination).copy_content(revision_id, basis)
320
def fetch(self, source, revision_id=None, pb=None):
321
"""Fetch the content required to construct revision_id from source.
323
If revision_id is None all content is copied.
325
revision_id = osutils.safe_revision_id(revision_id)
326
return InterRepository.get(source, self).fetch(revision_id=revision_id,
329
def get_commit_builder(self, branch, parents, config, timestamp=None,
330
timezone=None, committer=None, revprops=None,
332
"""Obtain a CommitBuilder for this repository.
334
:param branch: Branch to commit to.
335
:param parents: Revision ids of the parents of the new revision.
336
:param config: Configuration to use.
337
:param timestamp: Optional timestamp recorded for commit.
338
:param timezone: Optional timezone for timestamp.
339
:param committer: Optional committer to set for commit.
340
:param revprops: Optional dictionary of revision properties.
341
:param revision_id: Optional revision id.
343
revision_id = osutils.safe_revision_id(revision_id)
344
return _CommitBuilder(self, parents, config, timestamp, timezone,
345
committer, revprops, revision_id)
348
self.control_files.unlock()
351
def clone(self, a_bzrdir, revision_id=None, basis=None):
352
"""Clone this repository into a_bzrdir using the current format.
354
Currently no check is made that the format of this repository and
355
the bzrdir format are compatible. FIXME RBC 20060201.
357
:return: The newly created destination repository.
359
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
360
# use target default format.
361
dest_repo = a_bzrdir.create_repository()
363
# Most control formats need the repository to be specifically
364
# created, but on some old all-in-one formats it's not needed
366
dest_repo = self._format.initialize(a_bzrdir, shared=self.is_shared())
367
except errors.UninitializableFormat:
368
dest_repo = a_bzrdir.open_repository()
369
self.copy_content_into(dest_repo, revision_id, basis)
373
def has_revision(self, revision_id):
374
"""True if this repository has a copy of the revision."""
375
revision_id = osutils.safe_revision_id(revision_id)
376
return self._revision_store.has_revision_id(revision_id,
377
self.get_transaction())
380
def get_revision_reconcile(self, revision_id):
381
"""'reconcile' helper routine that allows access to a revision always.
383
This variant of get_revision does not cross check the weave graph
384
against the revision one as get_revision does: but it should only
385
be used by reconcile, or reconcile-alike commands that are correcting
386
or testing the revision graph.
388
if not revision_id or not isinstance(revision_id, basestring):
389
raise errors.InvalidRevisionId(revision_id=revision_id,
391
return self.get_revisions([revision_id])[0]
394
def get_revisions(self, revision_ids):
395
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
396
revs = self._revision_store.get_revisions(revision_ids,
397
self.get_transaction())
399
assert not isinstance(rev.revision_id, unicode)
400
for parent_id in rev.parent_ids:
401
assert not isinstance(parent_id, unicode)
405
def get_revision_xml(self, revision_id):
406
# TODO: jam 20070210 This shouldn't be necessary since get_revision
407
# would have already do it.
408
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
409
revision_id = osutils.safe_revision_id(revision_id)
410
rev = self.get_revision(revision_id)
412
# the current serializer..
413
self._revision_store._serializer.write_revision(rev, rev_tmp)
415
return rev_tmp.getvalue()
418
def get_revision(self, revision_id):
419
"""Return the Revision object for a named revision"""
420
# TODO: jam 20070210 get_revision_reconcile should do this for us
421
revision_id = osutils.safe_revision_id(revision_id)
422
r = self.get_revision_reconcile(revision_id)
423
# weave corruption can lead to absent revision markers that should be
425
# the following test is reasonably cheap (it needs a single weave read)
426
# and the weave is cached in read transactions. In write transactions
427
# it is not cached but typically we only read a small number of
428
# revisions. For knits when they are introduced we will probably want
429
# to ensure that caching write transactions are in use.
430
inv = self.get_inventory_weave()
431
self._check_revision_parents(r, inv)
435
def get_deltas_for_revisions(self, revisions):
436
"""Produce a generator of revision deltas.
438
Note that the input is a sequence of REVISIONS, not revision_ids.
439
Trees will be held in memory until the generator exits.
440
Each delta is relative to the revision's lefthand predecessor.
442
required_trees = set()
443
for revision in revisions:
444
required_trees.add(revision.revision_id)
445
required_trees.update(revision.parent_ids[:1])
446
trees = dict((t.get_revision_id(), t) for
447
t in self.revision_trees(required_trees))
448
for revision in revisions:
449
if not revision.parent_ids:
450
old_tree = self.revision_tree(None)
452
old_tree = trees[revision.parent_ids[0]]
453
yield trees[revision.revision_id].changes_from(old_tree)
456
def get_revision_delta(self, revision_id):
457
"""Return the delta for one revision.
459
The delta is relative to the left-hand predecessor of the
462
r = self.get_revision(revision_id)
463
return list(self.get_deltas_for_revisions([r]))[0]
465
def _check_revision_parents(self, revision, inventory):
466
"""Private to Repository and Fetch.
468
This checks the parentage of revision in an inventory weave for
469
consistency and is only applicable to inventory-weave-for-ancestry
470
using repository formats & fetchers.
472
weave_parents = inventory.get_parents(revision.revision_id)
473
weave_names = inventory.versions()
474
for parent_id in revision.parent_ids:
475
if parent_id in weave_names:
476
# this parent must not be a ghost.
477
if not parent_id in weave_parents:
479
raise errors.CorruptRepository(self)
482
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
483
revision_id = osutils.safe_revision_id(revision_id)
484
signature = gpg_strategy.sign(plaintext)
485
self._revision_store.add_revision_signature_text(revision_id,
487
self.get_transaction())
489
def fileids_altered_by_revision_ids(self, revision_ids):
490
"""Find the file ids and versions affected by revisions.
492
:param revisions: an iterable containing revision ids.
493
:return: a dictionary mapping altered file-ids to an iterable of
494
revision_ids. Each altered file-ids has the exact revision_ids that
495
altered it listed explicitly.
497
assert self._serializer.support_altered_by_hack, \
498
("fileids_altered_by_revision_ids only supported for branches "
499
"which store inventory as unnested xml, not on %r" % self)
500
selected_revision_ids = set(osutils.safe_revision_id(r)
501
for r in revision_ids)
502
w = self.get_inventory_weave()
505
# this code needs to read every new line in every inventory for the
506
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
507
# not present in one of those inventories is unnecessary but not
508
# harmful because we are filtering by the revision id marker in the
509
# inventory lines : we only select file ids altered in one of those
510
# revisions. We don't need to see all lines in the inventory because
511
# only those added in an inventory in rev X can contain a revision=X
513
unescape_revid_cache = {}
514
unescape_fileid_cache = {}
516
# jam 20061218 In a big fetch, this handles hundreds of thousands
517
# of lines, so it has had a lot of inlining and optimizing done.
518
# Sorry that it is a little bit messy.
519
# Move several functions to be local variables, since this is a long
521
search = self._file_ids_altered_regex.search
522
unescape = _unescape_xml
523
setdefault = result.setdefault
524
pb = ui.ui_factory.nested_progress_bar()
526
for line in w.iter_lines_added_or_present_in_versions(
527
selected_revision_ids, pb=pb):
531
# One call to match.group() returning multiple items is quite a
532
# bit faster than 2 calls to match.group() each returning 1
533
file_id, revision_id = match.group('file_id', 'revision_id')
535
# Inlining the cache lookups helps a lot when you make 170,000
536
# lines and 350k ids, versus 8.4 unique ids.
537
# Using a cache helps in 2 ways:
538
# 1) Avoids unnecessary decoding calls
539
# 2) Re-uses cached strings, which helps in future set and
541
# (2) is enough that removing encoding entirely along with
542
# the cache (so we are using plain strings) results in no
543
# performance improvement.
545
revision_id = unescape_revid_cache[revision_id]
547
# TODO: jam 20070217 For now, _unescape_xml return Unicode
548
# revision ids, once we make file_ids utf8, then it
549
# will return utf8 strings instead, and we can get
551
unescaped = osutils.safe_revision_id(unescape(revision_id))
552
unescape_revid_cache[revision_id] = unescaped
553
revision_id = unescaped
555
if revision_id in selected_revision_ids:
557
file_id = unescape_fileid_cache[file_id]
559
unescaped = unescape(file_id)
560
unescape_fileid_cache[file_id] = unescaped
562
setdefault(file_id, set()).add(revision_id)
568
def get_inventory_weave(self):
569
return self.control_weaves.get_weave('inventory',
570
self.get_transaction())
573
def get_inventory(self, revision_id):
574
"""Get Inventory object by hash."""
575
# TODO: jam 20070210 Technically we don't need to sanitize, since all
576
# called functions must sanitize.
577
revision_id = osutils.safe_revision_id(revision_id)
578
return self.deserialise_inventory(
579
revision_id, self.get_inventory_xml(revision_id))
581
def deserialise_inventory(self, revision_id, xml):
582
"""Transform the xml into an inventory object.
584
:param revision_id: The expected revision id of the inventory.
585
:param xml: A serialised inventory.
587
revision_id = osutils.safe_revision_id(revision_id)
588
result = self._serializer.read_inventory_from_string(xml)
589
result.root.revision = revision_id
592
def serialise_inventory(self, inv):
593
return self._serializer.write_inventory_to_string(inv)
596
def get_inventory_xml(self, revision_id):
597
"""Get inventory XML as a file object."""
598
revision_id = osutils.safe_revision_id(revision_id)
600
assert isinstance(revision_id, str), type(revision_id)
601
iw = self.get_inventory_weave()
602
return iw.get_text(revision_id)
604
raise errors.HistoryMissing(self, 'inventory', revision_id)
607
def get_inventory_sha1(self, revision_id):
608
"""Return the sha1 hash of the inventory entry
610
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
611
revision_id = osutils.safe_revision_id(revision_id)
612
return self.get_revision(revision_id).inventory_sha1
615
def get_revision_graph(self, revision_id=None):
616
"""Return a dictionary containing the revision graph.
618
:param revision_id: The revision_id to get a graph from. If None, then
619
the entire revision graph is returned. This is a deprecated mode of
620
operation and will be removed in the future.
621
:return: a dictionary of revision_id->revision_parents_list.
623
# special case NULL_REVISION
624
if revision_id == _mod_revision.NULL_REVISION:
626
revision_id = osutils.safe_revision_id(revision_id)
627
a_weave = self.get_inventory_weave()
628
all_revisions = self._eliminate_revisions_not_present(
630
entire_graph = dict([(node, a_weave.get_parents(node)) for
631
node in all_revisions])
632
if revision_id is None:
634
elif revision_id not in entire_graph:
635
raise errors.NoSuchRevision(self, revision_id)
637
# add what can be reached from revision_id
639
pending = set([revision_id])
640
while len(pending) > 0:
642
result[node] = entire_graph[node]
643
for revision_id in result[node]:
644
if revision_id not in result:
645
pending.add(revision_id)
649
def get_revision_graph_with_ghosts(self, revision_ids=None):
650
"""Return a graph of the revisions with ghosts marked as applicable.
652
:param revision_ids: an iterable of revisions to graph or None for all.
653
:return: a Graph object with the graph reachable from revision_ids.
655
result = graph.Graph()
657
pending = set(self.all_revision_ids())
660
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
661
# special case NULL_REVISION
662
if _mod_revision.NULL_REVISION in pending:
663
pending.remove(_mod_revision.NULL_REVISION)
664
required = set(pending)
667
revision_id = pending.pop()
669
rev = self.get_revision(revision_id)
670
except errors.NoSuchRevision:
671
if revision_id in required:
674
result.add_ghost(revision_id)
676
for parent_id in rev.parent_ids:
677
# is this queued or done ?
678
if (parent_id not in pending and
679
parent_id not in done):
681
pending.add(parent_id)
682
result.add_node(revision_id, rev.parent_ids)
683
done.add(revision_id)
686
def _get_history_vf(self):
687
"""Get a versionedfile whose history graph reflects all revisions.
689
For weave repositories, this is the inventory weave.
691
return self.get_inventory_weave()
693
def iter_reverse_revision_history(self, revision_id):
694
"""Iterate backwards through revision ids in the lefthand history
696
:param revision_id: The revision id to start with. All its lefthand
697
ancestors will be traversed.
699
revision_id = osutils.safe_revision_id(revision_id)
700
if revision_id in (None, _mod_revision.NULL_REVISION):
702
next_id = revision_id
703
versionedfile = self._get_history_vf()
706
parents = versionedfile.get_parents(next_id)
707
if len(parents) == 0:
713
def get_revision_inventory(self, revision_id):
714
"""Return inventory of a past revision."""
715
# TODO: Unify this with get_inventory()
716
# bzr 0.0.6 and later imposes the constraint that the inventory_id
717
# must be the same as its revision, so this is trivial.
718
if revision_id is None:
719
# This does not make sense: if there is no revision,
720
# then it is the current tree inventory surely ?!
721
# and thus get_root_id() is something that looks at the last
722
# commit on the branch, and the get_root_id is an inventory check.
723
raise NotImplementedError
724
# return Inventory(self.get_root_id())
726
return self.get_inventory(revision_id)
730
"""Return True if this repository is flagged as a shared repository."""
731
raise NotImplementedError(self.is_shared)
734
def reconcile(self, other=None, thorough=False):
735
"""Reconcile this repository."""
736
from bzrlib.reconcile import RepoReconciler
737
reconciler = RepoReconciler(self, thorough=thorough)
738
reconciler.reconcile()
742
def revision_tree(self, revision_id):
743
"""Return Tree for a revision on this branch.
745
`revision_id` may be None for the empty tree revision.
747
# TODO: refactor this to use an existing revision object
748
# so we don't need to read it in twice.
749
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
750
return RevisionTree(self, Inventory(root_id=None),
751
_mod_revision.NULL_REVISION)
753
revision_id = osutils.safe_revision_id(revision_id)
754
inv = self.get_revision_inventory(revision_id)
755
return RevisionTree(self, inv, revision_id)
758
def revision_trees(self, revision_ids):
759
"""Return Tree for a revision on this branch.
761
`revision_id` may not be None or 'null:'"""
762
assert None not in revision_ids
763
assert _mod_revision.NULL_REVISION not in revision_ids
764
texts = self.get_inventory_weave().get_texts(revision_ids)
765
for text, revision_id in zip(texts, revision_ids):
766
inv = self.deserialise_inventory(revision_id, text)
767
yield RevisionTree(self, inv, revision_id)
770
def get_ancestry(self, revision_id):
771
"""Return a list of revision-ids integrated by a revision.
773
The first element of the list is always None, indicating the origin
774
revision. This might change when we have history horizons, or
775
perhaps we should have a new API.
777
This is topologically sorted.
779
if revision_id is None:
781
revision_id = osutils.safe_revision_id(revision_id)
782
if not self.has_revision(revision_id):
783
raise errors.NoSuchRevision(self, revision_id)
784
w = self.get_inventory_weave()
785
candidates = w.get_ancestry(revision_id)
786
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
789
def print_file(self, file, revision_id):
790
"""Print `file` to stdout.
792
FIXME RBC 20060125 as John Meinel points out this is a bad api
793
- it writes to stdout, it assumes that that is valid etc. Fix
794
by creating a new more flexible convenience function.
796
revision_id = osutils.safe_revision_id(revision_id)
797
tree = self.revision_tree(revision_id)
798
# use inventory as it was in that revision
799
file_id = tree.inventory.path2id(file)
801
# TODO: jam 20060427 Write a test for this code path
802
# it had a bug in it, and was raising the wrong
804
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
805
tree.print_file(file_id)
807
def get_transaction(self):
808
return self.control_files.get_transaction()
810
def revision_parents(self, revision_id):
811
revision_id = osutils.safe_revision_id(revision_id)
812
return self.get_inventory_weave().parent_names(revision_id)
815
def set_make_working_trees(self, new_value):
816
"""Set the policy flag for making working trees when creating branches.
818
This only applies to branches that use this repository.
820
The default is 'True'.
821
:param new_value: True to restore the default, False to disable making
824
raise NotImplementedError(self.set_make_working_trees)
826
def make_working_trees(self):
827
"""Returns the policy for making working trees on new branches."""
828
raise NotImplementedError(self.make_working_trees)
831
def sign_revision(self, revision_id, gpg_strategy):
832
revision_id = osutils.safe_revision_id(revision_id)
833
plaintext = Testament.from_revision(self, revision_id).as_short_text()
834
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
837
def has_signature_for_revision_id(self, revision_id):
838
"""Query for a revision signature for revision_id in the repository."""
839
revision_id = osutils.safe_revision_id(revision_id)
840
return self._revision_store.has_signature(revision_id,
841
self.get_transaction())
844
def get_signature_text(self, revision_id):
845
"""Return the text for a signature."""
846
revision_id = osutils.safe_revision_id(revision_id)
847
return self._revision_store.get_signature_text(revision_id,
848
self.get_transaction())
851
def check(self, revision_ids):
852
"""Check consistency of all history of given revision_ids.
854
Different repository implementations should override _check().
856
:param revision_ids: A non-empty list of revision_ids whose ancestry
857
will be checked. Typically the last revision_id of a branch.
860
raise ValueError("revision_ids must be non-empty in %s.check"
862
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
863
return self._check(revision_ids)
865
def _check(self, revision_ids):
866
result = check.Check(self)
870
def _warn_if_deprecated(self):
871
global _deprecation_warning_done
872
if _deprecation_warning_done:
874
_deprecation_warning_done = True
875
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
876
% (self._format, self.bzrdir.transport.base))
878
def supports_rich_root(self):
879
return self._format.rich_root_data
881
def _check_ascii_revisionid(self, revision_id, method):
882
"""Private helper for ascii-only repositories."""
883
# weave repositories refuse to store revisionids that are non-ascii.
884
if revision_id is not None:
885
# weaves require ascii revision ids.
886
if isinstance(revision_id, unicode):
888
revision_id.encode('ascii')
889
except UnicodeEncodeError:
890
raise errors.NonAsciiRevisionId(method, self)
893
revision_id.decode('ascii')
894
except UnicodeDecodeError:
895
raise errors.NonAsciiRevisionId(method, self)
899
# remove these delegates a while after bzr 0.15
900
def __make_delegated(name, from_module):
901
def _deprecated_repository_forwarder():
902
symbol_versioning.warn('%s moved to %s in bzr 0.15'
903
% (name, from_module),
906
m = __import__(from_module, globals(), locals(), [name])
908
return getattr(m, name)
909
except AttributeError:
910
raise AttributeError('module %s has no name %s'
912
globals()[name] = _deprecated_repository_forwarder
915
'AllInOneRepository',
916
'WeaveMetaDirRepository',
917
'PreSplitOutRepositoryFormat',
923
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
928
'RepositoryFormatKnit',
929
'RepositoryFormatKnit1',
930
'RepositoryFormatKnit2',
932
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
935
def install_revision(repository, rev, revision_tree):
936
"""Install all revision data into a repository."""
939
for p_id in rev.parent_ids:
940
if repository.has_revision(p_id):
941
present_parents.append(p_id)
942
parent_trees[p_id] = repository.revision_tree(p_id)
944
parent_trees[p_id] = repository.revision_tree(None)
946
inv = revision_tree.inventory
947
entries = inv.iter_entries()
948
# backwards compatability hack: skip the root id.
949
if not repository.supports_rich_root():
950
path, root = entries.next()
951
if root.revision != rev.revision_id:
952
raise errors.IncompatibleRevision(repr(repository))
953
# Add the texts that are not already present
954
for path, ie in entries:
955
w = repository.weave_store.get_weave_or_empty(ie.file_id,
956
repository.get_transaction())
957
if ie.revision not in w:
959
# FIXME: TODO: The following loop *may* be overlapping/duplicate
960
# with InventoryEntry.find_previous_heads(). if it is, then there
961
# is a latent bug here where the parents may have ancestors of each
963
for revision, tree in parent_trees.iteritems():
964
if ie.file_id not in tree:
966
parent_id = tree.inventory[ie.file_id].revision
967
if parent_id in text_parents:
969
text_parents.append(parent_id)
971
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
972
repository.get_transaction())
973
lines = revision_tree.get_file(ie.file_id).readlines()
974
vfile.add_lines(rev.revision_id, text_parents, lines)
976
# install the inventory
977
repository.add_inventory(rev.revision_id, inv, present_parents)
978
except errors.RevisionAlreadyPresent:
980
repository.add_revision(rev.revision_id, rev, inv)
983
class MetaDirRepository(Repository):
984
"""Repositories in the new meta-dir layout."""
986
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
987
super(MetaDirRepository, self).__init__(_format,
993
dir_mode = self.control_files._dir_mode
994
file_mode = self.control_files._file_mode
998
"""Return True if this repository is flagged as a shared repository."""
999
return self.control_files._transport.has('shared-storage')
1002
def set_make_working_trees(self, new_value):
1003
"""Set the policy flag for making working trees when creating branches.
1005
This only applies to branches that use this repository.
1007
The default is 'True'.
1008
:param new_value: True to restore the default, False to disable making
1013
self.control_files._transport.delete('no-working-trees')
1014
except errors.NoSuchFile:
1017
self.control_files.put_utf8('no-working-trees', '')
1019
def make_working_trees(self):
1020
"""Returns the policy for making working trees on new branches."""
1021
return not self.control_files._transport.has('no-working-trees')
1024
class RepositoryFormatRegistry(registry.Registry):
1025
"""Registry of RepositoryFormats.
1028
def get(self, format_string):
1029
r = registry.Registry.get(self, format_string)
1035
format_registry = RepositoryFormatRegistry()
1036
"""Registry of formats, indexed by their identifying format string.
1038
This can contain either format instances themselves, or classes/factories that
1039
can be called to obtain one.
1043
class RepositoryFormat(object):
1044
"""A repository format.
1046
Formats provide three things:
1047
* An initialization routine to construct repository data on disk.
1048
* a format string which is used when the BzrDir supports versioned
1050
* an open routine which returns a Repository instance.
1052
Formats are placed in an dict by their format string for reference
1053
during opening. These should be subclasses of RepositoryFormat
1056
Once a format is deprecated, just deprecate the initialize and open
1057
methods on the format class. Do not deprecate the object, as the
1058
object will be created every system load.
1060
Common instance attributes:
1061
_matchingbzrdir - the bzrdir format that the repository format was
1062
originally written to work with. This can be used if manually
1063
constructing a bzrdir and repository, or more commonly for test suite
1068
return "<%s>" % self.__class__.__name__
1070
def __eq__(self, other):
1071
# format objects are generally stateless
1072
return isinstance(other, self.__class__)
1075
def find_format(klass, a_bzrdir):
1076
"""Return the format for the repository object in a_bzrdir.
1078
This is used by bzr native formats that have a "format" file in
1079
the repository. Other methods may be used by different types of
1083
transport = a_bzrdir.get_repository_transport(None)
1084
format_string = transport.get("format").read()
1085
return format_registry.get(format_string)
1086
except errors.NoSuchFile:
1087
raise errors.NoRepositoryPresent(a_bzrdir)
1089
raise errors.UnknownFormatError(format=format_string)
1092
def register_format(klass, format):
1093
format_registry.register(format.get_format_string(), format)
1096
def unregister_format(klass, format):
1097
format_registry.remove(format.get_format_string())
1100
def get_default_format(klass):
1101
"""Return the current default format."""
1102
from bzrlib import bzrdir
1103
return bzrdir.format_registry.make_bzrdir('default').repository_format
1105
def _get_control_store(self, repo_transport, control_files):
1106
"""Return the control store for this repository."""
1107
raise NotImplementedError(self._get_control_store)
1109
def get_format_string(self):
1110
"""Return the ASCII format string that identifies this format.
1112
Note that in pre format ?? repositories the format string is
1113
not permitted nor written to disk.
1115
raise NotImplementedError(self.get_format_string)
1117
def get_format_description(self):
1118
"""Return the short description for this format."""
1119
raise NotImplementedError(self.get_format_description)
1121
def _get_revision_store(self, repo_transport, control_files):
1122
"""Return the revision store object for this a_bzrdir."""
1123
raise NotImplementedError(self._get_revision_store)
1125
def _get_text_rev_store(self,
1132
"""Common logic for getting a revision store for a repository.
1134
see self._get_revision_store for the subclass-overridable method to
1135
get the store for a repository.
1137
from bzrlib.store.revision.text import TextRevisionStore
1138
dir_mode = control_files._dir_mode
1139
file_mode = control_files._file_mode
1140
text_store =TextStore(transport.clone(name),
1142
compressed=compressed,
1144
file_mode=file_mode)
1145
_revision_store = TextRevisionStore(text_store, serializer)
1146
return _revision_store
1148
# TODO: this shouldn't be in the base class, it's specific to things that
1149
# use weaves or knits -- mbp 20070207
1150
def _get_versioned_file_store(self,
1155
versionedfile_class=None,
1156
versionedfile_kwargs={},
1158
if versionedfile_class is None:
1159
versionedfile_class = self._versionedfile_class
1160
weave_transport = control_files._transport.clone(name)
1161
dir_mode = control_files._dir_mode
1162
file_mode = control_files._file_mode
1163
return VersionedFileStore(weave_transport, prefixed=prefixed,
1165
file_mode=file_mode,
1166
versionedfile_class=versionedfile_class,
1167
versionedfile_kwargs=versionedfile_kwargs,
1170
def initialize(self, a_bzrdir, shared=False):
1171
"""Initialize a repository of this format in a_bzrdir.
1173
:param a_bzrdir: The bzrdir to put the new repository in it.
1174
:param shared: The repository should be initialized as a sharable one.
1176
This may raise UninitializableFormat if shared repository are not
1177
compatible the a_bzrdir.
1180
def is_supported(self):
1181
"""Is this format supported?
1183
Supported formats must be initializable and openable.
1184
Unsupported formats may not support initialization or committing or
1185
some other features depending on the reason for not being supported.
1189
def check_conversion_target(self, target_format):
1190
raise NotImplementedError(self.check_conversion_target)
1192
def open(self, a_bzrdir, _found=False):
1193
"""Return an instance of this format for the bzrdir a_bzrdir.
1195
_found is a private parameter, do not use it.
1197
raise NotImplementedError(self.open)
1200
class MetaDirRepositoryFormat(RepositoryFormat):
1201
"""Common base class for the new repositories using the metadir layout."""
1203
rich_root_data = False
1204
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1207
super(MetaDirRepositoryFormat, self).__init__()
1209
def _create_control_files(self, a_bzrdir):
1210
"""Create the required files and the initial control_files object."""
1211
# FIXME: RBC 20060125 don't peek under the covers
1212
# NB: no need to escape relative paths that are url safe.
1213
repository_transport = a_bzrdir.get_repository_transport(self)
1214
control_files = lockable_files.LockableFiles(repository_transport,
1215
'lock', lockdir.LockDir)
1216
control_files.create_lock()
1217
return control_files
1219
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1220
"""Upload the initial blank content."""
1221
control_files = self._create_control_files(a_bzrdir)
1222
control_files.lock_write()
1224
control_files._transport.mkdir_multi(dirs,
1225
mode=control_files._dir_mode)
1226
for file, content in files:
1227
control_files.put(file, content)
1228
for file, content in utf8_files:
1229
control_files.put_utf8(file, content)
1231
control_files.put_utf8('shared-storage', '')
1233
control_files.unlock()
1236
# formats which have no format string are not discoverable
1237
# and not independently creatable, so are not registered. They're
1238
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1239
# needed, it's constructed directly by the BzrDir. Non-native formats where
1240
# the repository is not separately opened are similar.
1242
format_registry.register_lazy(
1243
'Bazaar-NG Repository format 7',
1244
'bzrlib.repofmt.weaverepo',
1247
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1248
# default control directory format
1250
format_registry.register_lazy(
1251
'Bazaar-NG Knit Repository Format 1',
1252
'bzrlib.repofmt.knitrepo',
1253
'RepositoryFormatKnit1',
1255
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1257
format_registry.register_lazy(
1258
'Bazaar Knit Repository Format 2\n',
1259
'bzrlib.repofmt.knitrepo',
1260
'RepositoryFormatKnit2',
1264
class InterRepository(InterObject):
1265
"""This class represents operations taking place between two repositories.
1267
Its instances have methods like copy_content and fetch, and contain
1268
references to the source and target repositories these operations can be
1271
Often we will provide convenience methods on 'repository' which carry out
1272
operations with another repository - they will always forward to
1273
InterRepository.get(other).method_name(parameters).
1277
"""The available optimised InterRepository types."""
1279
def copy_content(self, revision_id=None, basis=None):
1280
raise NotImplementedError(self.copy_content)
1282
def fetch(self, revision_id=None, pb=None):
1283
"""Fetch the content required to construct revision_id.
1285
The content is copied from self.source to self.target.
1287
:param revision_id: if None all content is copied, if NULL_REVISION no
1289
:param pb: optional progress bar to use for progress reports. If not
1290
provided a default one will be created.
1292
Returns the copied revision count and the failed revisions in a tuple:
1295
raise NotImplementedError(self.fetch)
1298
def missing_revision_ids(self, revision_id=None):
1299
"""Return the revision ids that source has that target does not.
1301
These are returned in topological order.
1303
:param revision_id: only return revision ids included by this
1306
# generic, possibly worst case, slow code path.
1307
target_ids = set(self.target.all_revision_ids())
1308
if revision_id is not None:
1309
# TODO: jam 20070210 InterRepository is internal enough that it
1310
# should assume revision_ids are already utf-8
1311
revision_id = osutils.safe_revision_id(revision_id)
1312
source_ids = self.source.get_ancestry(revision_id)
1313
assert source_ids[0] is None
1316
source_ids = self.source.all_revision_ids()
1317
result_set = set(source_ids).difference(target_ids)
1318
# this may look like a no-op: its not. It preserves the ordering
1319
# other_ids had while only returning the members from other_ids
1320
# that we've decided we need.
1321
return [rev_id for rev_id in source_ids if rev_id in result_set]
1324
class InterSameDataRepository(InterRepository):
1325
"""Code for converting between repositories that represent the same data.
1327
Data format and model must match for this to work.
1331
def _get_repo_format_to_test(self):
1332
"""Repository format for testing with."""
1333
return RepositoryFormat.get_default_format()
1336
def is_compatible(source, target):
1337
if not isinstance(source, Repository):
1339
if not isinstance(target, Repository):
1341
if source._format.rich_root_data == target._format.rich_root_data:
1347
def copy_content(self, revision_id=None, basis=None):
1348
"""Make a complete copy of the content in self into destination.
1350
This is a destructive operation! Do not use it on existing
1353
:param revision_id: Only copy the content needed to construct
1354
revision_id and its parents.
1355
:param basis: Copy the needed data preferentially from basis.
1358
self.target.set_make_working_trees(self.source.make_working_trees())
1359
except NotImplementedError:
1361
# TODO: jam 20070210 This is fairly internal, so we should probably
1362
# just assert that revision_id is not unicode.
1363
revision_id = osutils.safe_revision_id(revision_id)
1364
# grab the basis available data
1365
if basis is not None:
1366
self.target.fetch(basis, revision_id=revision_id)
1367
# but don't bother fetching if we have the needed data now.
1368
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1369
self.target.has_revision(revision_id)):
1371
self.target.fetch(self.source, revision_id=revision_id)
1374
def fetch(self, revision_id=None, pb=None):
1375
"""See InterRepository.fetch()."""
1376
from bzrlib.fetch import GenericRepoFetcher
1377
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1378
self.source, self.source._format, self.target,
1379
self.target._format)
1380
# TODO: jam 20070210 This should be an assert, not a translate
1381
revision_id = osutils.safe_revision_id(revision_id)
1382
f = GenericRepoFetcher(to_repository=self.target,
1383
from_repository=self.source,
1384
last_revision=revision_id,
1386
return f.count_copied, f.failed_revisions
1389
class InterWeaveRepo(InterSameDataRepository):
1390
"""Optimised code paths between Weave based repositories."""
1393
def _get_repo_format_to_test(self):
1394
from bzrlib.repofmt import weaverepo
1395
return weaverepo.RepositoryFormat7()
1398
def is_compatible(source, target):
1399
"""Be compatible with known Weave formats.
1401
We don't test for the stores being of specific types because that
1402
could lead to confusing results, and there is no need to be
1405
from bzrlib.repofmt.weaverepo import (
1411
return (isinstance(source._format, (RepositoryFormat5,
1413
RepositoryFormat7)) and
1414
isinstance(target._format, (RepositoryFormat5,
1416
RepositoryFormat7)))
1417
except AttributeError:
1421
def copy_content(self, revision_id=None, basis=None):
1422
"""See InterRepository.copy_content()."""
1423
# weave specific optimised path:
1424
# TODO: jam 20070210 Internal, should be an assert, not translate
1425
revision_id = osutils.safe_revision_id(revision_id)
1426
if basis is not None:
1427
# copy the basis in, then fetch remaining data.
1428
basis.copy_content_into(self.target, revision_id)
1429
# the basis copy_content_into could miss-set this.
1431
self.target.set_make_working_trees(self.source.make_working_trees())
1432
except NotImplementedError:
1434
self.target.fetch(self.source, revision_id=revision_id)
1437
self.target.set_make_working_trees(self.source.make_working_trees())
1438
except NotImplementedError:
1440
# FIXME do not peek!
1441
if self.source.control_files._transport.listable():
1442
pb = ui.ui_factory.nested_progress_bar()
1444
self.target.weave_store.copy_all_ids(
1445
self.source.weave_store,
1447
from_transaction=self.source.get_transaction(),
1448
to_transaction=self.target.get_transaction())
1449
pb.update('copying inventory', 0, 1)
1450
self.target.control_weaves.copy_multi(
1451
self.source.control_weaves, ['inventory'],
1452
from_transaction=self.source.get_transaction(),
1453
to_transaction=self.target.get_transaction())
1454
self.target._revision_store.text_store.copy_all_ids(
1455
self.source._revision_store.text_store,
1460
self.target.fetch(self.source, revision_id=revision_id)
1463
def fetch(self, revision_id=None, pb=None):
1464
"""See InterRepository.fetch()."""
1465
from bzrlib.fetch import GenericRepoFetcher
1466
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1467
self.source, self.source._format, self.target, self.target._format)
1468
# TODO: jam 20070210 This should be an assert, not a translate
1469
revision_id = osutils.safe_revision_id(revision_id)
1470
f = GenericRepoFetcher(to_repository=self.target,
1471
from_repository=self.source,
1472
last_revision=revision_id,
1474
return f.count_copied, f.failed_revisions
1477
def missing_revision_ids(self, revision_id=None):
1478
"""See InterRepository.missing_revision_ids()."""
1479
# we want all revisions to satisfy revision_id in source.
1480
# but we don't want to stat every file here and there.
1481
# we want then, all revisions other needs to satisfy revision_id
1482
# checked, but not those that we have locally.
1483
# so the first thing is to get a subset of the revisions to
1484
# satisfy revision_id in source, and then eliminate those that
1485
# we do already have.
1486
# this is slow on high latency connection to self, but as as this
1487
# disk format scales terribly for push anyway due to rewriting
1488
# inventory.weave, this is considered acceptable.
1490
if revision_id is not None:
1491
source_ids = self.source.get_ancestry(revision_id)
1492
assert source_ids[0] is None
1495
source_ids = self.source._all_possible_ids()
1496
source_ids_set = set(source_ids)
1497
# source_ids is the worst possible case we may need to pull.
1498
# now we want to filter source_ids against what we actually
1499
# have in target, but don't try to check for existence where we know
1500
# we do not have a revision as that would be pointless.
1501
target_ids = set(self.target._all_possible_ids())
1502
possibly_present_revisions = target_ids.intersection(source_ids_set)
1503
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1504
required_revisions = source_ids_set.difference(actually_present_revisions)
1505
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1506
if revision_id is not None:
1507
# we used get_ancestry to determine source_ids then we are assured all
1508
# revisions referenced are present as they are installed in topological order.
1509
# and the tip revision was validated by get_ancestry.
1510
return required_topo_revisions
1512
# if we just grabbed the possibly available ids, then
1513
# we only have an estimate of whats available and need to validate
1514
# that against the revision records.
1515
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1518
class InterKnitRepo(InterSameDataRepository):
1519
"""Optimised code paths between Knit based repositories."""
1522
def _get_repo_format_to_test(self):
1523
from bzrlib.repofmt import knitrepo
1524
return knitrepo.RepositoryFormatKnit1()
1527
def is_compatible(source, target):
1528
"""Be compatible with known Knit formats.
1530
We don't test for the stores being of specific types because that
1531
could lead to confusing results, and there is no need to be
1534
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1536
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1537
isinstance(target._format, (RepositoryFormatKnit1)))
1538
except AttributeError:
1542
def fetch(self, revision_id=None, pb=None):
1543
"""See InterRepository.fetch()."""
1544
from bzrlib.fetch import KnitRepoFetcher
1545
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1546
self.source, self.source._format, self.target, self.target._format)
1547
# TODO: jam 20070210 This should be an assert, not a translate
1548
revision_id = osutils.safe_revision_id(revision_id)
1549
f = KnitRepoFetcher(to_repository=self.target,
1550
from_repository=self.source,
1551
last_revision=revision_id,
1553
return f.count_copied, f.failed_revisions
1556
def missing_revision_ids(self, revision_id=None):
1557
"""See InterRepository.missing_revision_ids()."""
1558
if revision_id is not None:
1559
source_ids = self.source.get_ancestry(revision_id)
1560
assert source_ids[0] is None
1563
source_ids = self.source._all_possible_ids()
1564
source_ids_set = set(source_ids)
1565
# source_ids is the worst possible case we may need to pull.
1566
# now we want to filter source_ids against what we actually
1567
# have in target, but don't try to check for existence where we know
1568
# we do not have a revision as that would be pointless.
1569
target_ids = set(self.target._all_possible_ids())
1570
possibly_present_revisions = target_ids.intersection(source_ids_set)
1571
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1572
required_revisions = source_ids_set.difference(actually_present_revisions)
1573
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1574
if revision_id is not None:
1575
# we used get_ancestry to determine source_ids then we are assured all
1576
# revisions referenced are present as they are installed in topological order.
1577
# and the tip revision was validated by get_ancestry.
1578
return required_topo_revisions
1580
# if we just grabbed the possibly available ids, then
1581
# we only have an estimate of whats available and need to validate
1582
# that against the revision records.
1583
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1586
class InterModel1and2(InterRepository):
1589
def _get_repo_format_to_test(self):
1593
def is_compatible(source, target):
1594
if not isinstance(source, Repository):
1596
if not isinstance(target, Repository):
1598
if not source._format.rich_root_data and target._format.rich_root_data:
1604
def fetch(self, revision_id=None, pb=None):
1605
"""See InterRepository.fetch()."""
1606
from bzrlib.fetch import Model1toKnit2Fetcher
1607
# TODO: jam 20070210 This should be an assert, not a translate
1608
revision_id = osutils.safe_revision_id(revision_id)
1609
f = Model1toKnit2Fetcher(to_repository=self.target,
1610
from_repository=self.source,
1611
last_revision=revision_id,
1613
return f.count_copied, f.failed_revisions
1616
def copy_content(self, revision_id=None, basis=None):
1617
"""Make a complete copy of the content in self into destination.
1619
This is a destructive operation! Do not use it on existing
1622
:param revision_id: Only copy the content needed to construct
1623
revision_id and its parents.
1624
:param basis: Copy the needed data preferentially from basis.
1627
self.target.set_make_working_trees(self.source.make_working_trees())
1628
except NotImplementedError:
1630
# TODO: jam 20070210 Internal, assert, don't translate
1631
revision_id = osutils.safe_revision_id(revision_id)
1632
# grab the basis available data
1633
if basis is not None:
1634
self.target.fetch(basis, revision_id=revision_id)
1635
# but don't bother fetching if we have the needed data now.
1636
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1637
self.target.has_revision(revision_id)):
1639
self.target.fetch(self.source, revision_id=revision_id)
1642
class InterKnit1and2(InterKnitRepo):
1645
def _get_repo_format_to_test(self):
1649
def is_compatible(source, target):
1650
"""Be compatible with Knit1 source and Knit2 target"""
1651
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit2
1653
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1654
RepositoryFormatKnit2
1655
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1656
isinstance(target._format, (RepositoryFormatKnit2)))
1657
except AttributeError:
1661
def fetch(self, revision_id=None, pb=None):
1662
"""See InterRepository.fetch()."""
1663
from bzrlib.fetch import Knit1to2Fetcher
1664
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1665
self.source, self.source._format, self.target,
1666
self.target._format)
1667
# TODO: jam 20070210 This should be an assert, not a translate
1668
revision_id = osutils.safe_revision_id(revision_id)
1669
f = Knit1to2Fetcher(to_repository=self.target,
1670
from_repository=self.source,
1671
last_revision=revision_id,
1673
return f.count_copied, f.failed_revisions
1676
InterRepository.register_optimiser(InterSameDataRepository)
1677
InterRepository.register_optimiser(InterWeaveRepo)
1678
InterRepository.register_optimiser(InterKnitRepo)
1679
InterRepository.register_optimiser(InterModel1and2)
1680
InterRepository.register_optimiser(InterKnit1and2)
1683
class RepositoryTestProviderAdapter(object):
1684
"""A tool to generate a suite testing multiple repository formats at once.
1686
This is done by copying the test once for each transport and injecting
1687
the transport_server, transport_readonly_server, and bzrdir_format and
1688
repository_format classes into each copy. Each copy is also given a new id()
1689
to make it easy to identify.
1692
def __init__(self, transport_server, transport_readonly_server, formats):
1693
self._transport_server = transport_server
1694
self._transport_readonly_server = transport_readonly_server
1695
self._formats = formats
1697
def adapt(self, test):
1698
result = unittest.TestSuite()
1699
for repository_format, bzrdir_format in self._formats:
1700
from copy import deepcopy
1701
new_test = deepcopy(test)
1702
new_test.transport_server = self._transport_server
1703
new_test.transport_readonly_server = self._transport_readonly_server
1704
new_test.bzrdir_format = bzrdir_format
1705
new_test.repository_format = repository_format
1706
def make_new_test_id():
1707
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1708
return lambda: new_id
1709
new_test.id = make_new_test_id()
1710
result.addTest(new_test)
1714
class InterRepositoryTestProviderAdapter(object):
1715
"""A tool to generate a suite testing multiple inter repository formats.
1717
This is done by copying the test once for each interrepo provider and injecting
1718
the transport_server, transport_readonly_server, repository_format and
1719
repository_to_format classes into each copy.
1720
Each copy is also given a new id() to make it easy to identify.
1723
def __init__(self, transport_server, transport_readonly_server, formats):
1724
self._transport_server = transport_server
1725
self._transport_readonly_server = transport_readonly_server
1726
self._formats = formats
1728
def adapt(self, test):
1729
result = unittest.TestSuite()
1730
for interrepo_class, repository_format, repository_format_to in self._formats:
1731
from copy import deepcopy
1732
new_test = deepcopy(test)
1733
new_test.transport_server = self._transport_server
1734
new_test.transport_readonly_server = self._transport_readonly_server
1735
new_test.interrepo_class = interrepo_class
1736
new_test.repository_format = repository_format
1737
new_test.repository_format_to = repository_format_to
1738
def make_new_test_id():
1739
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1740
return lambda: new_id
1741
new_test.id = make_new_test_id()
1742
result.addTest(new_test)
1746
def default_test_list():
1747
"""Generate the default list of interrepo permutations to test."""
1748
from bzrlib.repofmt import knitrepo, weaverepo
1750
# test the default InterRepository between format 6 and the current
1752
# XXX: robertc 20060220 reinstate this when there are two supported
1753
# formats which do not have an optimal code path between them.
1754
#result.append((InterRepository,
1755
# RepositoryFormat6(),
1756
# RepositoryFormatKnit1()))
1757
for optimiser_class in InterRepository._optimisers:
1758
format_to_test = optimiser_class._get_repo_format_to_test()
1759
if format_to_test is not None:
1760
result.append((optimiser_class,
1761
format_to_test, format_to_test))
1762
# if there are specific combinations we want to use, we can add them
1764
result.append((InterModel1and2,
1765
weaverepo.RepositoryFormat5(),
1766
knitrepo.RepositoryFormatKnit2()))
1767
result.append((InterKnit1and2,
1768
knitrepo.RepositoryFormatKnit1(),
1769
knitrepo.RepositoryFormatKnit2()))
1773
class CopyConverter(object):
1774
"""A repository conversion tool which just performs a copy of the content.
1776
This is slow but quite reliable.
1779
def __init__(self, target_format):
1780
"""Create a CopyConverter.
1782
:param target_format: The format the resulting repository should be.
1784
self.target_format = target_format
1786
def convert(self, repo, pb):
1787
"""Perform the conversion of to_convert, giving feedback via pb.
1789
:param to_convert: The disk object to convert.
1790
:param pb: a progress bar to use for progress information.
1795
# this is only useful with metadir layouts - separated repo content.
1796
# trigger an assertion if not such
1797
repo._format.get_format_string()
1798
self.repo_dir = repo.bzrdir
1799
self.step('Moving repository to repository.backup')
1800
self.repo_dir.transport.move('repository', 'repository.backup')
1801
backup_transport = self.repo_dir.transport.clone('repository.backup')
1802
repo._format.check_conversion_target(self.target_format)
1803
self.source_repo = repo._format.open(self.repo_dir,
1805
_override_transport=backup_transport)
1806
self.step('Creating new repository')
1807
converted = self.target_format.initialize(self.repo_dir,
1808
self.source_repo.is_shared())
1809
converted.lock_write()
1811
self.step('Copying content into repository.')
1812
self.source_repo.copy_content_into(converted)
1815
self.step('Deleting old repository content.')
1816
self.repo_dir.transport.delete_tree('repository.backup')
1817
self.pb.note('repository converted')
1819
def step(self, message):
1820
"""Update the pb by a step."""
1822
self.pb.update(message, self.count, self.total)
1825
class CommitBuilder(object):
1826
"""Provides an interface to build up a commit.
1828
This allows describing a tree to be committed without needing to
1829
know the internals of the format of the repository.
1832
record_root_entry = False
1833
def __init__(self, repository, parents, config, timestamp=None,
1834
timezone=None, committer=None, revprops=None,
1836
"""Initiate a CommitBuilder.
1838
:param repository: Repository to commit to.
1839
:param parents: Revision ids of the parents of the new revision.
1840
:param config: Configuration to use.
1841
:param timestamp: Optional timestamp recorded for commit.
1842
:param timezone: Optional timezone for timestamp.
1843
:param committer: Optional committer to set for commit.
1844
:param revprops: Optional dictionary of revision properties.
1845
:param revision_id: Optional revision id.
1847
self._config = config
1849
if committer is None:
1850
self._committer = self._config.username()
1852
assert isinstance(committer, basestring), type(committer)
1853
self._committer = committer
1855
self.new_inventory = Inventory(None)
1856
self._new_revision_id = osutils.safe_revision_id(revision_id)
1857
self.parents = parents
1858
self.repository = repository
1861
if revprops is not None:
1862
self._revprops.update(revprops)
1864
if timestamp is None:
1865
timestamp = time.time()
1866
# Restrict resolution to 1ms
1867
self._timestamp = round(timestamp, 3)
1869
if timezone is None:
1870
self._timezone = osutils.local_time_offset()
1872
self._timezone = int(timezone)
1874
self._generate_revision_if_needed()
1876
def commit(self, message):
1877
"""Make the actual commit.
1879
:return: The revision id of the recorded revision.
1881
rev = _mod_revision.Revision(
1882
timestamp=self._timestamp,
1883
timezone=self._timezone,
1884
committer=self._committer,
1886
inventory_sha1=self.inv_sha1,
1887
revision_id=self._new_revision_id,
1888
properties=self._revprops)
1889
rev.parent_ids = self.parents
1890
self.repository.add_revision(self._new_revision_id, rev,
1891
self.new_inventory, self._config)
1892
return self._new_revision_id
1894
def revision_tree(self):
1895
"""Return the tree that was just committed.
1897
After calling commit() this can be called to get a RevisionTree
1898
representing the newly committed tree. This is preferred to
1899
calling Repository.revision_tree() because that may require
1900
deserializing the inventory, while we already have a copy in
1903
return RevisionTree(self.repository, self.new_inventory,
1904
self._new_revision_id)
1906
def finish_inventory(self):
1907
"""Tell the builder that the inventory is finished."""
1908
if self.new_inventory.root is None:
1909
symbol_versioning.warn('Root entry should be supplied to'
1910
' record_entry_contents, as of bzr 0.10.',
1911
DeprecationWarning, stacklevel=2)
1912
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
1913
self.new_inventory.revision_id = self._new_revision_id
1914
self.inv_sha1 = self.repository.add_inventory(
1915
self._new_revision_id,
1920
def _gen_revision_id(self):
1921
"""Return new revision-id."""
1922
return generate_ids.gen_revision_id(self._config.username(),
1925
def _generate_revision_if_needed(self):
1926
"""Create a revision id if None was supplied.
1928
If the repository can not support user-specified revision ids
1929
they should override this function and raise CannotSetRevisionId
1930
if _new_revision_id is not None.
1932
:raises: CannotSetRevisionId
1934
if self._new_revision_id is None:
1935
self._new_revision_id = self._gen_revision_id()
1937
def record_entry_contents(self, ie, parent_invs, path, tree):
1938
"""Record the content of ie from tree into the commit if needed.
1940
Side effect: sets ie.revision when unchanged
1942
:param ie: An inventory entry present in the commit.
1943
:param parent_invs: The inventories of the parent revisions of the
1945
:param path: The path the entry is at in the tree.
1946
:param tree: The tree which contains this entry and should be used to
1949
if self.new_inventory.root is None and ie.parent_id is not None:
1950
symbol_versioning.warn('Root entry should be supplied to'
1951
' record_entry_contents, as of bzr 0.10.',
1952
DeprecationWarning, stacklevel=2)
1953
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
1955
self.new_inventory.add(ie)
1957
# ie.revision is always None if the InventoryEntry is considered
1958
# for committing. ie.snapshot will record the correct revision
1959
# which may be the sole parent if it is untouched.
1960
if ie.revision is not None:
1963
# In this revision format, root entries have no knit or weave
1964
if ie is self.new_inventory.root:
1965
# When serializing out to disk and back in
1966
# root.revision is always _new_revision_id
1967
ie.revision = self._new_revision_id
1969
previous_entries = ie.find_previous_heads(
1971
self.repository.weave_store,
1972
self.repository.get_transaction())
1973
# we are creating a new revision for ie in the history store
1975
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
1977
def modified_directory(self, file_id, file_parents):
1978
"""Record the presence of a symbolic link.
1980
:param file_id: The file_id of the link to record.
1981
:param file_parents: The per-file parent revision ids.
1983
self._add_text_to_weave(file_id, [], file_parents.keys())
1985
def modified_file_text(self, file_id, file_parents,
1986
get_content_byte_lines, text_sha1=None,
1988
"""Record the text of file file_id
1990
:param file_id: The file_id of the file to record the text of.
1991
:param file_parents: The per-file parent revision ids.
1992
:param get_content_byte_lines: A callable which will return the byte
1994
:param text_sha1: Optional SHA1 of the file contents.
1995
:param text_size: Optional size of the file contents.
1997
# mutter('storing text of file {%s} in revision {%s} into %r',
1998
# file_id, self._new_revision_id, self.repository.weave_store)
1999
# special case to avoid diffing on renames or
2001
if (len(file_parents) == 1
2002
and text_sha1 == file_parents.values()[0].text_sha1
2003
and text_size == file_parents.values()[0].text_size):
2004
previous_ie = file_parents.values()[0]
2005
versionedfile = self.repository.weave_store.get_weave(file_id,
2006
self.repository.get_transaction())
2007
versionedfile.clone_text(self._new_revision_id,
2008
previous_ie.revision, file_parents.keys())
2009
return text_sha1, text_size
2011
new_lines = get_content_byte_lines()
2012
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2013
# should return the SHA1 and size
2014
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2015
return osutils.sha_strings(new_lines), \
2016
sum(map(len, new_lines))
2018
def modified_link(self, file_id, file_parents, link_target):
2019
"""Record the presence of a symbolic link.
2021
:param file_id: The file_id of the link to record.
2022
:param file_parents: The per-file parent revision ids.
2023
:param link_target: Target location of this link.
2025
self._add_text_to_weave(file_id, [], file_parents.keys())
2027
def _add_text_to_weave(self, file_id, new_lines, parents):
2028
versionedfile = self.repository.weave_store.get_weave_or_empty(
2029
file_id, self.repository.get_transaction())
2030
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2031
versionedfile.clear_cache()
2034
class _CommitBuilder(CommitBuilder):
2035
"""Temporary class so old CommitBuilders are detected properly
2037
Note: CommitBuilder works whether or not root entry is recorded.
2040
record_root_entry = True
2043
class RootCommitBuilder(CommitBuilder):
2044
"""This commitbuilder actually records the root id"""
2046
record_root_entry = True
2048
def record_entry_contents(self, ie, parent_invs, path, tree):
2049
"""Record the content of ie from tree into the commit if needed.
2051
Side effect: sets ie.revision when unchanged
2053
:param ie: An inventory entry present in the commit.
2054
:param parent_invs: The inventories of the parent revisions of the
2056
:param path: The path the entry is at in the tree.
2057
:param tree: The tree which contains this entry and should be used to
2060
assert self.new_inventory.root is not None or ie.parent_id is None
2061
self.new_inventory.add(ie)
2063
# ie.revision is always None if the InventoryEntry is considered
2064
# for committing. ie.snapshot will record the correct revision
2065
# which may be the sole parent if it is untouched.
2066
if ie.revision is not None:
2069
previous_entries = ie.find_previous_heads(
2071
self.repository.weave_store,
2072
self.repository.get_transaction())
2073
# we are creating a new revision for ie in the history store
2075
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2087
def _unescaper(match, _map=_unescape_map):
2088
code = match.group(1)
2092
if not code.startswith('#'):
2094
return unichr(int(code[1:]))
2100
def _unescape_xml(data):
2101
"""Unescape predefined XML entities in a string of data."""
2103
if _unescape_re is None:
2104
_unescape_re = re.compile('\&([^;]*);')
2105
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
