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
unescaped = unescape(revision_id)
548
unescape_revid_cache[revision_id] = unescaped
549
revision_id = unescaped
551
if revision_id in selected_revision_ids:
553
file_id = unescape_fileid_cache[file_id]
555
unescaped = unescape(file_id)
556
unescape_fileid_cache[file_id] = unescaped
558
setdefault(file_id, set()).add(revision_id)
564
def get_inventory_weave(self):
565
return self.control_weaves.get_weave('inventory',
566
self.get_transaction())
569
def get_inventory(self, revision_id):
570
"""Get Inventory object by hash."""
571
# TODO: jam 20070210 Technically we don't need to sanitize, since all
572
# called functions must sanitize.
573
revision_id = osutils.safe_revision_id(revision_id)
574
return self.deserialise_inventory(
575
revision_id, self.get_inventory_xml(revision_id))
577
def deserialise_inventory(self, revision_id, xml):
578
"""Transform the xml into an inventory object.
580
:param revision_id: The expected revision id of the inventory.
581
:param xml: A serialised inventory.
583
revision_id = osutils.safe_revision_id(revision_id)
584
result = self._serializer.read_inventory_from_string(xml)
585
result.root.revision = revision_id
588
def serialise_inventory(self, inv):
589
return self._serializer.write_inventory_to_string(inv)
592
def get_inventory_xml(self, revision_id):
593
"""Get inventory XML as a file object."""
594
revision_id = osutils.safe_revision_id(revision_id)
596
assert isinstance(revision_id, str), type(revision_id)
597
iw = self.get_inventory_weave()
598
return iw.get_text(revision_id)
600
raise errors.HistoryMissing(self, 'inventory', revision_id)
603
def get_inventory_sha1(self, revision_id):
604
"""Return the sha1 hash of the inventory entry
606
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
607
revision_id = osutils.safe_revision_id(revision_id)
608
return self.get_revision(revision_id).inventory_sha1
611
def get_revision_graph(self, revision_id=None):
612
"""Return a dictionary containing the revision graph.
614
:param revision_id: The revision_id to get a graph from. If None, then
615
the entire revision graph is returned. This is a deprecated mode of
616
operation and will be removed in the future.
617
:return: a dictionary of revision_id->revision_parents_list.
619
# special case NULL_REVISION
620
if revision_id == _mod_revision.NULL_REVISION:
622
revision_id = osutils.safe_revision_id(revision_id)
623
a_weave = self.get_inventory_weave()
624
all_revisions = self._eliminate_revisions_not_present(
626
entire_graph = dict([(node, a_weave.get_parents(node)) for
627
node in all_revisions])
628
if revision_id is None:
630
elif revision_id not in entire_graph:
631
raise errors.NoSuchRevision(self, revision_id)
633
# add what can be reached from revision_id
635
pending = set([revision_id])
636
while len(pending) > 0:
638
result[node] = entire_graph[node]
639
for revision_id in result[node]:
640
if revision_id not in result:
641
pending.add(revision_id)
645
def get_revision_graph_with_ghosts(self, revision_ids=None):
646
"""Return a graph of the revisions with ghosts marked as applicable.
648
:param revision_ids: an iterable of revisions to graph or None for all.
649
:return: a Graph object with the graph reachable from revision_ids.
651
result = graph.Graph()
653
pending = set(self.all_revision_ids())
656
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
657
# special case NULL_REVISION
658
if _mod_revision.NULL_REVISION in pending:
659
pending.remove(_mod_revision.NULL_REVISION)
660
required = set(pending)
663
revision_id = pending.pop()
665
rev = self.get_revision(revision_id)
666
except errors.NoSuchRevision:
667
if revision_id in required:
670
result.add_ghost(revision_id)
672
for parent_id in rev.parent_ids:
673
# is this queued or done ?
674
if (parent_id not in pending and
675
parent_id not in done):
677
pending.add(parent_id)
678
result.add_node(revision_id, rev.parent_ids)
679
done.add(revision_id)
682
def _get_history_vf(self):
683
"""Get a versionedfile whose history graph reflects all revisions.
685
For weave repositories, this is the inventory weave.
687
return self.get_inventory_weave()
689
def iter_reverse_revision_history(self, revision_id):
690
"""Iterate backwards through revision ids in the lefthand history
692
:param revision_id: The revision id to start with. All its lefthand
693
ancestors will be traversed.
695
revision_id = osutils.safe_revision_id(revision_id)
696
if revision_id in (None, _mod_revision.NULL_REVISION):
698
next_id = revision_id
699
versionedfile = self._get_history_vf()
702
parents = versionedfile.get_parents(next_id)
703
if len(parents) == 0:
709
def get_revision_inventory(self, revision_id):
710
"""Return inventory of a past revision."""
711
# TODO: Unify this with get_inventory()
712
# bzr 0.0.6 and later imposes the constraint that the inventory_id
713
# must be the same as its revision, so this is trivial.
714
if revision_id is None:
715
# This does not make sense: if there is no revision,
716
# then it is the current tree inventory surely ?!
717
# and thus get_root_id() is something that looks at the last
718
# commit on the branch, and the get_root_id is an inventory check.
719
raise NotImplementedError
720
# return Inventory(self.get_root_id())
722
return self.get_inventory(revision_id)
726
"""Return True if this repository is flagged as a shared repository."""
727
raise NotImplementedError(self.is_shared)
730
def reconcile(self, other=None, thorough=False):
731
"""Reconcile this repository."""
732
from bzrlib.reconcile import RepoReconciler
733
reconciler = RepoReconciler(self, thorough=thorough)
734
reconciler.reconcile()
738
def revision_tree(self, revision_id):
739
"""Return Tree for a revision on this branch.
741
`revision_id` may be None for the empty tree revision.
743
# TODO: refactor this to use an existing revision object
744
# so we don't need to read it in twice.
745
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
746
return RevisionTree(self, Inventory(root_id=None),
747
_mod_revision.NULL_REVISION)
749
revision_id = osutils.safe_revision_id(revision_id)
750
inv = self.get_revision_inventory(revision_id)
751
return RevisionTree(self, inv, revision_id)
754
def revision_trees(self, revision_ids):
755
"""Return Tree for a revision on this branch.
757
`revision_id` may not be None or 'null:'"""
758
assert None not in revision_ids
759
assert _mod_revision.NULL_REVISION not in revision_ids
760
texts = self.get_inventory_weave().get_texts(revision_ids)
761
for text, revision_id in zip(texts, revision_ids):
762
inv = self.deserialise_inventory(revision_id, text)
763
yield RevisionTree(self, inv, revision_id)
766
def get_ancestry(self, revision_id):
767
"""Return a list of revision-ids integrated by a revision.
769
The first element of the list is always None, indicating the origin
770
revision. This might change when we have history horizons, or
771
perhaps we should have a new API.
773
This is topologically sorted.
775
if revision_id is None:
777
revision_id = osutils.safe_revision_id(revision_id)
778
if not self.has_revision(revision_id):
779
raise errors.NoSuchRevision(self, revision_id)
780
w = self.get_inventory_weave()
781
candidates = w.get_ancestry(revision_id)
782
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
785
def print_file(self, file, revision_id):
786
"""Print `file` to stdout.
788
FIXME RBC 20060125 as John Meinel points out this is a bad api
789
- it writes to stdout, it assumes that that is valid etc. Fix
790
by creating a new more flexible convenience function.
792
revision_id = osutils.safe_revision_id(revision_id)
793
tree = self.revision_tree(revision_id)
794
# use inventory as it was in that revision
795
file_id = tree.inventory.path2id(file)
797
# TODO: jam 20060427 Write a test for this code path
798
# it had a bug in it, and was raising the wrong
800
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
801
tree.print_file(file_id)
803
def get_transaction(self):
804
return self.control_files.get_transaction()
806
def revision_parents(self, revision_id):
807
revision_id = osutils.safe_revision_id(revision_id)
808
return self.get_inventory_weave().parent_names(revision_id)
811
def set_make_working_trees(self, new_value):
812
"""Set the policy flag for making working trees when creating branches.
814
This only applies to branches that use this repository.
816
The default is 'True'.
817
:param new_value: True to restore the default, False to disable making
820
raise NotImplementedError(self.set_make_working_trees)
822
def make_working_trees(self):
823
"""Returns the policy for making working trees on new branches."""
824
raise NotImplementedError(self.make_working_trees)
827
def sign_revision(self, revision_id, gpg_strategy):
828
revision_id = osutils.safe_revision_id(revision_id)
829
plaintext = Testament.from_revision(self, revision_id).as_short_text()
830
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
833
def has_signature_for_revision_id(self, revision_id):
834
"""Query for a revision signature for revision_id in the repository."""
835
revision_id = osutils.safe_revision_id(revision_id)
836
return self._revision_store.has_signature(revision_id,
837
self.get_transaction())
840
def get_signature_text(self, revision_id):
841
"""Return the text for a signature."""
842
revision_id = osutils.safe_revision_id(revision_id)
843
return self._revision_store.get_signature_text(revision_id,
844
self.get_transaction())
847
def check(self, revision_ids):
848
"""Check consistency of all history of given revision_ids.
850
Different repository implementations should override _check().
852
:param revision_ids: A non-empty list of revision_ids whose ancestry
853
will be checked. Typically the last revision_id of a branch.
856
raise ValueError("revision_ids must be non-empty in %s.check"
858
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
859
return self._check(revision_ids)
861
def _check(self, revision_ids):
862
result = check.Check(self)
866
def _warn_if_deprecated(self):
867
global _deprecation_warning_done
868
if _deprecation_warning_done:
870
_deprecation_warning_done = True
871
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
872
% (self._format, self.bzrdir.transport.base))
874
def supports_rich_root(self):
875
return self._format.rich_root_data
877
def _check_ascii_revisionid(self, revision_id, method):
878
"""Private helper for ascii-only repositories."""
879
# weave repositories refuse to store revisionids that are non-ascii.
880
if revision_id is not None:
881
# weaves require ascii revision ids.
882
if isinstance(revision_id, unicode):
884
revision_id.encode('ascii')
885
except UnicodeEncodeError:
886
raise errors.NonAsciiRevisionId(method, self)
889
revision_id.decode('ascii')
890
except UnicodeDecodeError:
891
raise errors.NonAsciiRevisionId(method, self)
895
# remove these delegates a while after bzr 0.15
896
def __make_delegated(name, from_module):
897
def _deprecated_repository_forwarder():
898
symbol_versioning.warn('%s moved to %s in bzr 0.15'
899
% (name, from_module),
902
m = __import__(from_module, globals(), locals(), [name])
904
return getattr(m, name)
905
except AttributeError:
906
raise AttributeError('module %s has no name %s'
908
globals()[name] = _deprecated_repository_forwarder
911
'AllInOneRepository',
912
'WeaveMetaDirRepository',
913
'PreSplitOutRepositoryFormat',
919
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
924
'RepositoryFormatKnit',
925
'RepositoryFormatKnit1',
926
'RepositoryFormatKnit2',
928
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
931
def install_revision(repository, rev, revision_tree):
932
"""Install all revision data into a repository."""
935
for p_id in rev.parent_ids:
936
if repository.has_revision(p_id):
937
present_parents.append(p_id)
938
parent_trees[p_id] = repository.revision_tree(p_id)
940
parent_trees[p_id] = repository.revision_tree(None)
942
inv = revision_tree.inventory
943
entries = inv.iter_entries()
944
# backwards compatability hack: skip the root id.
945
if not repository.supports_rich_root():
946
path, root = entries.next()
947
if root.revision != rev.revision_id:
948
raise errors.IncompatibleRevision(repr(repository))
949
# Add the texts that are not already present
950
for path, ie in entries:
951
w = repository.weave_store.get_weave_or_empty(ie.file_id,
952
repository.get_transaction())
953
if ie.revision not in w:
955
# FIXME: TODO: The following loop *may* be overlapping/duplicate
956
# with InventoryEntry.find_previous_heads(). if it is, then there
957
# is a latent bug here where the parents may have ancestors of each
959
for revision, tree in parent_trees.iteritems():
960
if ie.file_id not in tree:
962
parent_id = tree.inventory[ie.file_id].revision
963
if parent_id in text_parents:
965
text_parents.append(parent_id)
967
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
968
repository.get_transaction())
969
lines = revision_tree.get_file(ie.file_id).readlines()
970
vfile.add_lines(rev.revision_id, text_parents, lines)
972
# install the inventory
973
repository.add_inventory(rev.revision_id, inv, present_parents)
974
except errors.RevisionAlreadyPresent:
976
repository.add_revision(rev.revision_id, rev, inv)
979
class MetaDirRepository(Repository):
980
"""Repositories in the new meta-dir layout."""
982
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
983
super(MetaDirRepository, self).__init__(_format,
989
dir_mode = self.control_files._dir_mode
990
file_mode = self.control_files._file_mode
994
"""Return True if this repository is flagged as a shared repository."""
995
return self.control_files._transport.has('shared-storage')
998
def set_make_working_trees(self, new_value):
999
"""Set the policy flag for making working trees when creating branches.
1001
This only applies to branches that use this repository.
1003
The default is 'True'.
1004
:param new_value: True to restore the default, False to disable making
1009
self.control_files._transport.delete('no-working-trees')
1010
except errors.NoSuchFile:
1013
self.control_files.put_utf8('no-working-trees', '')
1015
def make_working_trees(self):
1016
"""Returns the policy for making working trees on new branches."""
1017
return not self.control_files._transport.has('no-working-trees')
1020
class RepositoryFormatRegistry(registry.Registry):
1021
"""Registry of RepositoryFormats.
1024
def get(self, format_string):
1025
r = registry.Registry.get(self, format_string)
1031
format_registry = RepositoryFormatRegistry()
1032
"""Registry of formats, indexed by their identifying format string.
1034
This can contain either format instances themselves, or classes/factories that
1035
can be called to obtain one.
1039
class RepositoryFormat(object):
1040
"""A repository format.
1042
Formats provide three things:
1043
* An initialization routine to construct repository data on disk.
1044
* a format string which is used when the BzrDir supports versioned
1046
* an open routine which returns a Repository instance.
1048
Formats are placed in an dict by their format string for reference
1049
during opening. These should be subclasses of RepositoryFormat
1052
Once a format is deprecated, just deprecate the initialize and open
1053
methods on the format class. Do not deprecate the object, as the
1054
object will be created every system load.
1056
Common instance attributes:
1057
_matchingbzrdir - the bzrdir format that the repository format was
1058
originally written to work with. This can be used if manually
1059
constructing a bzrdir and repository, or more commonly for test suite
1064
return "<%s>" % self.__class__.__name__
1066
def __eq__(self, other):
1067
# format objects are generally stateless
1068
return isinstance(other, self.__class__)
1071
def find_format(klass, a_bzrdir):
1072
"""Return the format for the repository object in a_bzrdir.
1074
This is used by bzr native formats that have a "format" file in
1075
the repository. Other methods may be used by different types of
1079
transport = a_bzrdir.get_repository_transport(None)
1080
format_string = transport.get("format").read()
1081
return format_registry.get(format_string)
1082
except errors.NoSuchFile:
1083
raise errors.NoRepositoryPresent(a_bzrdir)
1085
raise errors.UnknownFormatError(format=format_string)
1088
def register_format(klass, format):
1089
format_registry.register(format.get_format_string(), format)
1092
def unregister_format(klass, format):
1093
format_registry.remove(format.get_format_string())
1096
def get_default_format(klass):
1097
"""Return the current default format."""
1098
from bzrlib import bzrdir
1099
return bzrdir.format_registry.make_bzrdir('default').repository_format
1101
def _get_control_store(self, repo_transport, control_files):
1102
"""Return the control store for this repository."""
1103
raise NotImplementedError(self._get_control_store)
1105
def get_format_string(self):
1106
"""Return the ASCII format string that identifies this format.
1108
Note that in pre format ?? repositories the format string is
1109
not permitted nor written to disk.
1111
raise NotImplementedError(self.get_format_string)
1113
def get_format_description(self):
1114
"""Return the short description for this format."""
1115
raise NotImplementedError(self.get_format_description)
1117
def _get_revision_store(self, repo_transport, control_files):
1118
"""Return the revision store object for this a_bzrdir."""
1119
raise NotImplementedError(self._get_revision_store)
1121
def _get_text_rev_store(self,
1128
"""Common logic for getting a revision store for a repository.
1130
see self._get_revision_store for the subclass-overridable method to
1131
get the store for a repository.
1133
from bzrlib.store.revision.text import TextRevisionStore
1134
dir_mode = control_files._dir_mode
1135
file_mode = control_files._file_mode
1136
text_store =TextStore(transport.clone(name),
1138
compressed=compressed,
1140
file_mode=file_mode)
1141
_revision_store = TextRevisionStore(text_store, serializer)
1142
return _revision_store
1144
# TODO: this shouldn't be in the base class, it's specific to things that
1145
# use weaves or knits -- mbp 20070207
1146
def _get_versioned_file_store(self,
1151
versionedfile_class=None,
1152
versionedfile_kwargs={},
1154
if versionedfile_class is None:
1155
versionedfile_class = self._versionedfile_class
1156
weave_transport = control_files._transport.clone(name)
1157
dir_mode = control_files._dir_mode
1158
file_mode = control_files._file_mode
1159
return VersionedFileStore(weave_transport, prefixed=prefixed,
1161
file_mode=file_mode,
1162
versionedfile_class=versionedfile_class,
1163
versionedfile_kwargs=versionedfile_kwargs,
1166
def initialize(self, a_bzrdir, shared=False):
1167
"""Initialize a repository of this format in a_bzrdir.
1169
:param a_bzrdir: The bzrdir to put the new repository in it.
1170
:param shared: The repository should be initialized as a sharable one.
1172
This may raise UninitializableFormat if shared repository are not
1173
compatible the a_bzrdir.
1176
def is_supported(self):
1177
"""Is this format supported?
1179
Supported formats must be initializable and openable.
1180
Unsupported formats may not support initialization or committing or
1181
some other features depending on the reason for not being supported.
1185
def check_conversion_target(self, target_format):
1186
raise NotImplementedError(self.check_conversion_target)
1188
def open(self, a_bzrdir, _found=False):
1189
"""Return an instance of this format for the bzrdir a_bzrdir.
1191
_found is a private parameter, do not use it.
1193
raise NotImplementedError(self.open)
1196
class MetaDirRepositoryFormat(RepositoryFormat):
1197
"""Common base class for the new repositories using the metadir layout."""
1199
rich_root_data = False
1200
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1203
super(MetaDirRepositoryFormat, self).__init__()
1205
def _create_control_files(self, a_bzrdir):
1206
"""Create the required files and the initial control_files object."""
1207
# FIXME: RBC 20060125 don't peek under the covers
1208
# NB: no need to escape relative paths that are url safe.
1209
repository_transport = a_bzrdir.get_repository_transport(self)
1210
control_files = lockable_files.LockableFiles(repository_transport,
1211
'lock', lockdir.LockDir)
1212
control_files.create_lock()
1213
return control_files
1215
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1216
"""Upload the initial blank content."""
1217
control_files = self._create_control_files(a_bzrdir)
1218
control_files.lock_write()
1220
control_files._transport.mkdir_multi(dirs,
1221
mode=control_files._dir_mode)
1222
for file, content in files:
1223
control_files.put(file, content)
1224
for file, content in utf8_files:
1225
control_files.put_utf8(file, content)
1227
control_files.put_utf8('shared-storage', '')
1229
control_files.unlock()
1232
# formats which have no format string are not discoverable
1233
# and not independently creatable, so are not registered. They're
1234
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1235
# needed, it's constructed directly by the BzrDir. Non-native formats where
1236
# the repository is not separately opened are similar.
1238
format_registry.register_lazy(
1239
'Bazaar-NG Repository format 7',
1240
'bzrlib.repofmt.weaverepo',
1243
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1244
# default control directory format
1246
format_registry.register_lazy(
1247
'Bazaar-NG Knit Repository Format 1',
1248
'bzrlib.repofmt.knitrepo',
1249
'RepositoryFormatKnit1',
1251
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1253
format_registry.register_lazy(
1254
'Bazaar Knit Repository Format 2\n',
1255
'bzrlib.repofmt.knitrepo',
1256
'RepositoryFormatKnit2',
1260
class InterRepository(InterObject):
1261
"""This class represents operations taking place between two repositories.
1263
Its instances have methods like copy_content and fetch, and contain
1264
references to the source and target repositories these operations can be
1267
Often we will provide convenience methods on 'repository' which carry out
1268
operations with another repository - they will always forward to
1269
InterRepository.get(other).method_name(parameters).
1273
"""The available optimised InterRepository types."""
1275
def copy_content(self, revision_id=None, basis=None):
1276
raise NotImplementedError(self.copy_content)
1278
def fetch(self, revision_id=None, pb=None):
1279
"""Fetch the content required to construct revision_id.
1281
The content is copied from self.source to self.target.
1283
:param revision_id: if None all content is copied, if NULL_REVISION no
1285
:param pb: optional progress bar to use for progress reports. If not
1286
provided a default one will be created.
1288
Returns the copied revision count and the failed revisions in a tuple:
1291
raise NotImplementedError(self.fetch)
1294
def missing_revision_ids(self, revision_id=None):
1295
"""Return the revision ids that source has that target does not.
1297
These are returned in topological order.
1299
:param revision_id: only return revision ids included by this
1302
# generic, possibly worst case, slow code path.
1303
target_ids = set(self.target.all_revision_ids())
1304
if revision_id is not None:
1305
# TODO: jam 20070210 InterRepository is internal enough that it
1306
# should assume revision_ids are already utf-8
1307
revision_id = osutils.safe_revision_id(revision_id)
1308
source_ids = self.source.get_ancestry(revision_id)
1309
assert source_ids[0] is None
1312
source_ids = self.source.all_revision_ids()
1313
result_set = set(source_ids).difference(target_ids)
1314
# this may look like a no-op: its not. It preserves the ordering
1315
# other_ids had while only returning the members from other_ids
1316
# that we've decided we need.
1317
return [rev_id for rev_id in source_ids if rev_id in result_set]
1320
class InterSameDataRepository(InterRepository):
1321
"""Code for converting between repositories that represent the same data.
1323
Data format and model must match for this to work.
1327
def _get_repo_format_to_test(self):
1328
"""Repository format for testing with."""
1329
return RepositoryFormat.get_default_format()
1332
def is_compatible(source, target):
1333
if not isinstance(source, Repository):
1335
if not isinstance(target, Repository):
1337
if source._format.rich_root_data == target._format.rich_root_data:
1343
def copy_content(self, revision_id=None, basis=None):
1344
"""Make a complete copy of the content in self into destination.
1346
This is a destructive operation! Do not use it on existing
1349
:param revision_id: Only copy the content needed to construct
1350
revision_id and its parents.
1351
:param basis: Copy the needed data preferentially from basis.
1354
self.target.set_make_working_trees(self.source.make_working_trees())
1355
except NotImplementedError:
1357
# TODO: jam 20070210 This is fairly internal, so we should probably
1358
# just assert that revision_id is not unicode.
1359
revision_id = osutils.safe_revision_id(revision_id)
1360
# grab the basis available data
1361
if basis is not None:
1362
self.target.fetch(basis, revision_id=revision_id)
1363
# but don't bother fetching if we have the needed data now.
1364
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1365
self.target.has_revision(revision_id)):
1367
self.target.fetch(self.source, revision_id=revision_id)
1370
def fetch(self, revision_id=None, pb=None):
1371
"""See InterRepository.fetch()."""
1372
from bzrlib.fetch import GenericRepoFetcher
1373
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1374
self.source, self.source._format, self.target,
1375
self.target._format)
1376
# TODO: jam 20070210 This should be an assert, not a translate
1377
revision_id = osutils.safe_revision_id(revision_id)
1378
f = GenericRepoFetcher(to_repository=self.target,
1379
from_repository=self.source,
1380
last_revision=revision_id,
1382
return f.count_copied, f.failed_revisions
1385
class InterWeaveRepo(InterSameDataRepository):
1386
"""Optimised code paths between Weave based repositories."""
1389
def _get_repo_format_to_test(self):
1390
from bzrlib.repofmt import weaverepo
1391
return weaverepo.RepositoryFormat7()
1394
def is_compatible(source, target):
1395
"""Be compatible with known Weave formats.
1397
We don't test for the stores being of specific types because that
1398
could lead to confusing results, and there is no need to be
1401
from bzrlib.repofmt.weaverepo import (
1407
return (isinstance(source._format, (RepositoryFormat5,
1409
RepositoryFormat7)) and
1410
isinstance(target._format, (RepositoryFormat5,
1412
RepositoryFormat7)))
1413
except AttributeError:
1417
def copy_content(self, revision_id=None, basis=None):
1418
"""See InterRepository.copy_content()."""
1419
# weave specific optimised path:
1420
# TODO: jam 20070210 Internal, should be an assert, not translate
1421
revision_id = osutils.safe_revision_id(revision_id)
1422
if basis is not None:
1423
# copy the basis in, then fetch remaining data.
1424
basis.copy_content_into(self.target, revision_id)
1425
# the basis copy_content_into could miss-set this.
1427
self.target.set_make_working_trees(self.source.make_working_trees())
1428
except NotImplementedError:
1430
self.target.fetch(self.source, revision_id=revision_id)
1433
self.target.set_make_working_trees(self.source.make_working_trees())
1434
except NotImplementedError:
1436
# FIXME do not peek!
1437
if self.source.control_files._transport.listable():
1438
pb = ui.ui_factory.nested_progress_bar()
1440
self.target.weave_store.copy_all_ids(
1441
self.source.weave_store,
1443
from_transaction=self.source.get_transaction(),
1444
to_transaction=self.target.get_transaction())
1445
pb.update('copying inventory', 0, 1)
1446
self.target.control_weaves.copy_multi(
1447
self.source.control_weaves, ['inventory'],
1448
from_transaction=self.source.get_transaction(),
1449
to_transaction=self.target.get_transaction())
1450
self.target._revision_store.text_store.copy_all_ids(
1451
self.source._revision_store.text_store,
1456
self.target.fetch(self.source, revision_id=revision_id)
1459
def fetch(self, revision_id=None, pb=None):
1460
"""See InterRepository.fetch()."""
1461
from bzrlib.fetch import GenericRepoFetcher
1462
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1463
self.source, self.source._format, self.target, self.target._format)
1464
# TODO: jam 20070210 This should be an assert, not a translate
1465
revision_id = osutils.safe_revision_id(revision_id)
1466
f = GenericRepoFetcher(to_repository=self.target,
1467
from_repository=self.source,
1468
last_revision=revision_id,
1470
return f.count_copied, f.failed_revisions
1473
def missing_revision_ids(self, revision_id=None):
1474
"""See InterRepository.missing_revision_ids()."""
1475
# we want all revisions to satisfy revision_id in source.
1476
# but we don't want to stat every file here and there.
1477
# we want then, all revisions other needs to satisfy revision_id
1478
# checked, but not those that we have locally.
1479
# so the first thing is to get a subset of the revisions to
1480
# satisfy revision_id in source, and then eliminate those that
1481
# we do already have.
1482
# this is slow on high latency connection to self, but as as this
1483
# disk format scales terribly for push anyway due to rewriting
1484
# inventory.weave, this is considered acceptable.
1486
if revision_id is not None:
1487
source_ids = self.source.get_ancestry(revision_id)
1488
assert source_ids[0] is None
1491
source_ids = self.source._all_possible_ids()
1492
source_ids_set = set(source_ids)
1493
# source_ids is the worst possible case we may need to pull.
1494
# now we want to filter source_ids against what we actually
1495
# have in target, but don't try to check for existence where we know
1496
# we do not have a revision as that would be pointless.
1497
target_ids = set(self.target._all_possible_ids())
1498
possibly_present_revisions = target_ids.intersection(source_ids_set)
1499
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1500
required_revisions = source_ids_set.difference(actually_present_revisions)
1501
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1502
if revision_id is not None:
1503
# we used get_ancestry to determine source_ids then we are assured all
1504
# revisions referenced are present as they are installed in topological order.
1505
# and the tip revision was validated by get_ancestry.
1506
return required_topo_revisions
1508
# if we just grabbed the possibly available ids, then
1509
# we only have an estimate of whats available and need to validate
1510
# that against the revision records.
1511
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1514
class InterKnitRepo(InterSameDataRepository):
1515
"""Optimised code paths between Knit based repositories."""
1518
def _get_repo_format_to_test(self):
1519
from bzrlib.repofmt import knitrepo
1520
return knitrepo.RepositoryFormatKnit1()
1523
def is_compatible(source, target):
1524
"""Be compatible with known Knit formats.
1526
We don't test for the stores being of specific types because that
1527
could lead to confusing results, and there is no need to be
1530
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1532
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1533
isinstance(target._format, (RepositoryFormatKnit1)))
1534
except AttributeError:
1538
def fetch(self, revision_id=None, pb=None):
1539
"""See InterRepository.fetch()."""
1540
from bzrlib.fetch import KnitRepoFetcher
1541
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1542
self.source, self.source._format, self.target, self.target._format)
1543
# TODO: jam 20070210 This should be an assert, not a translate
1544
revision_id = osutils.safe_revision_id(revision_id)
1545
f = KnitRepoFetcher(to_repository=self.target,
1546
from_repository=self.source,
1547
last_revision=revision_id,
1549
return f.count_copied, f.failed_revisions
1552
def missing_revision_ids(self, revision_id=None):
1553
"""See InterRepository.missing_revision_ids()."""
1554
if revision_id is not None:
1555
source_ids = self.source.get_ancestry(revision_id)
1556
assert source_ids[0] is None
1559
source_ids = self.source._all_possible_ids()
1560
source_ids_set = set(source_ids)
1561
# source_ids is the worst possible case we may need to pull.
1562
# now we want to filter source_ids against what we actually
1563
# have in target, but don't try to check for existence where we know
1564
# we do not have a revision as that would be pointless.
1565
target_ids = set(self.target._all_possible_ids())
1566
possibly_present_revisions = target_ids.intersection(source_ids_set)
1567
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1568
required_revisions = source_ids_set.difference(actually_present_revisions)
1569
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1570
if revision_id is not None:
1571
# we used get_ancestry to determine source_ids then we are assured all
1572
# revisions referenced are present as they are installed in topological order.
1573
# and the tip revision was validated by get_ancestry.
1574
return required_topo_revisions
1576
# if we just grabbed the possibly available ids, then
1577
# we only have an estimate of whats available and need to validate
1578
# that against the revision records.
1579
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1582
class InterModel1and2(InterRepository):
1585
def _get_repo_format_to_test(self):
1589
def is_compatible(source, target):
1590
if not isinstance(source, Repository):
1592
if not isinstance(target, Repository):
1594
if not source._format.rich_root_data and target._format.rich_root_data:
1600
def fetch(self, revision_id=None, pb=None):
1601
"""See InterRepository.fetch()."""
1602
from bzrlib.fetch import Model1toKnit2Fetcher
1603
# TODO: jam 20070210 This should be an assert, not a translate
1604
revision_id = osutils.safe_revision_id(revision_id)
1605
f = Model1toKnit2Fetcher(to_repository=self.target,
1606
from_repository=self.source,
1607
last_revision=revision_id,
1609
return f.count_copied, f.failed_revisions
1612
def copy_content(self, revision_id=None, basis=None):
1613
"""Make a complete copy of the content in self into destination.
1615
This is a destructive operation! Do not use it on existing
1618
:param revision_id: Only copy the content needed to construct
1619
revision_id and its parents.
1620
:param basis: Copy the needed data preferentially from basis.
1623
self.target.set_make_working_trees(self.source.make_working_trees())
1624
except NotImplementedError:
1626
# TODO: jam 20070210 Internal, assert, don't translate
1627
revision_id = osutils.safe_revision_id(revision_id)
1628
# grab the basis available data
1629
if basis is not None:
1630
self.target.fetch(basis, revision_id=revision_id)
1631
# but don't bother fetching if we have the needed data now.
1632
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1633
self.target.has_revision(revision_id)):
1635
self.target.fetch(self.source, revision_id=revision_id)
1638
class InterKnit1and2(InterKnitRepo):
1641
def _get_repo_format_to_test(self):
1645
def is_compatible(source, target):
1646
"""Be compatible with Knit1 source and Knit2 target"""
1647
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit2
1649
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1650
RepositoryFormatKnit2
1651
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1652
isinstance(target._format, (RepositoryFormatKnit2)))
1653
except AttributeError:
1657
def fetch(self, revision_id=None, pb=None):
1658
"""See InterRepository.fetch()."""
1659
from bzrlib.fetch import Knit1to2Fetcher
1660
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1661
self.source, self.source._format, self.target,
1662
self.target._format)
1663
# TODO: jam 20070210 This should be an assert, not a translate
1664
revision_id = osutils.safe_revision_id(revision_id)
1665
f = Knit1to2Fetcher(to_repository=self.target,
1666
from_repository=self.source,
1667
last_revision=revision_id,
1669
return f.count_copied, f.failed_revisions
1672
InterRepository.register_optimiser(InterSameDataRepository)
1673
InterRepository.register_optimiser(InterWeaveRepo)
1674
InterRepository.register_optimiser(InterKnitRepo)
1675
InterRepository.register_optimiser(InterModel1and2)
1676
InterRepository.register_optimiser(InterKnit1and2)
1679
class RepositoryTestProviderAdapter(object):
1680
"""A tool to generate a suite testing multiple repository formats at once.
1682
This is done by copying the test once for each transport and injecting
1683
the transport_server, transport_readonly_server, and bzrdir_format and
1684
repository_format classes into each copy. Each copy is also given a new id()
1685
to make it easy to identify.
1688
def __init__(self, transport_server, transport_readonly_server, formats):
1689
self._transport_server = transport_server
1690
self._transport_readonly_server = transport_readonly_server
1691
self._formats = formats
1693
def adapt(self, test):
1694
result = unittest.TestSuite()
1695
for repository_format, bzrdir_format in self._formats:
1696
from copy import deepcopy
1697
new_test = deepcopy(test)
1698
new_test.transport_server = self._transport_server
1699
new_test.transport_readonly_server = self._transport_readonly_server
1700
new_test.bzrdir_format = bzrdir_format
1701
new_test.repository_format = repository_format
1702
def make_new_test_id():
1703
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1704
return lambda: new_id
1705
new_test.id = make_new_test_id()
1706
result.addTest(new_test)
1710
class InterRepositoryTestProviderAdapter(object):
1711
"""A tool to generate a suite testing multiple inter repository formats.
1713
This is done by copying the test once for each interrepo provider and injecting
1714
the transport_server, transport_readonly_server, repository_format and
1715
repository_to_format classes into each copy.
1716
Each copy is also given a new id() to make it easy to identify.
1719
def __init__(self, transport_server, transport_readonly_server, formats):
1720
self._transport_server = transport_server
1721
self._transport_readonly_server = transport_readonly_server
1722
self._formats = formats
1724
def adapt(self, test):
1725
result = unittest.TestSuite()
1726
for interrepo_class, repository_format, repository_format_to in self._formats:
1727
from copy import deepcopy
1728
new_test = deepcopy(test)
1729
new_test.transport_server = self._transport_server
1730
new_test.transport_readonly_server = self._transport_readonly_server
1731
new_test.interrepo_class = interrepo_class
1732
new_test.repository_format = repository_format
1733
new_test.repository_format_to = repository_format_to
1734
def make_new_test_id():
1735
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1736
return lambda: new_id
1737
new_test.id = make_new_test_id()
1738
result.addTest(new_test)
1742
def default_test_list():
1743
"""Generate the default list of interrepo permutations to test."""
1744
from bzrlib.repofmt import knitrepo, weaverepo
1746
# test the default InterRepository between format 6 and the current
1748
# XXX: robertc 20060220 reinstate this when there are two supported
1749
# formats which do not have an optimal code path between them.
1750
#result.append((InterRepository,
1751
# RepositoryFormat6(),
1752
# RepositoryFormatKnit1()))
1753
for optimiser_class in InterRepository._optimisers:
1754
format_to_test = optimiser_class._get_repo_format_to_test()
1755
if format_to_test is not None:
1756
result.append((optimiser_class,
1757
format_to_test, format_to_test))
1758
# if there are specific combinations we want to use, we can add them
1760
result.append((InterModel1and2,
1761
weaverepo.RepositoryFormat5(),
1762
knitrepo.RepositoryFormatKnit2()))
1763
result.append((InterKnit1and2,
1764
knitrepo.RepositoryFormatKnit1(),
1765
knitrepo.RepositoryFormatKnit2()))
1769
class CopyConverter(object):
1770
"""A repository conversion tool which just performs a copy of the content.
1772
This is slow but quite reliable.
1775
def __init__(self, target_format):
1776
"""Create a CopyConverter.
1778
:param target_format: The format the resulting repository should be.
1780
self.target_format = target_format
1782
def convert(self, repo, pb):
1783
"""Perform the conversion of to_convert, giving feedback via pb.
1785
:param to_convert: The disk object to convert.
1786
:param pb: a progress bar to use for progress information.
1791
# this is only useful with metadir layouts - separated repo content.
1792
# trigger an assertion if not such
1793
repo._format.get_format_string()
1794
self.repo_dir = repo.bzrdir
1795
self.step('Moving repository to repository.backup')
1796
self.repo_dir.transport.move('repository', 'repository.backup')
1797
backup_transport = self.repo_dir.transport.clone('repository.backup')
1798
repo._format.check_conversion_target(self.target_format)
1799
self.source_repo = repo._format.open(self.repo_dir,
1801
_override_transport=backup_transport)
1802
self.step('Creating new repository')
1803
converted = self.target_format.initialize(self.repo_dir,
1804
self.source_repo.is_shared())
1805
converted.lock_write()
1807
self.step('Copying content into repository.')
1808
self.source_repo.copy_content_into(converted)
1811
self.step('Deleting old repository content.')
1812
self.repo_dir.transport.delete_tree('repository.backup')
1813
self.pb.note('repository converted')
1815
def step(self, message):
1816
"""Update the pb by a step."""
1818
self.pb.update(message, self.count, self.total)
1821
class CommitBuilder(object):
1822
"""Provides an interface to build up a commit.
1824
This allows describing a tree to be committed without needing to
1825
know the internals of the format of the repository.
1828
record_root_entry = False
1829
def __init__(self, repository, parents, config, timestamp=None,
1830
timezone=None, committer=None, revprops=None,
1832
"""Initiate a CommitBuilder.
1834
:param repository: Repository to commit to.
1835
:param parents: Revision ids of the parents of the new revision.
1836
:param config: Configuration to use.
1837
:param timestamp: Optional timestamp recorded for commit.
1838
:param timezone: Optional timezone for timestamp.
1839
:param committer: Optional committer to set for commit.
1840
:param revprops: Optional dictionary of revision properties.
1841
:param revision_id: Optional revision id.
1843
self._config = config
1845
if committer is None:
1846
self._committer = self._config.username()
1848
assert isinstance(committer, basestring), type(committer)
1849
self._committer = committer
1851
self.new_inventory = Inventory(None)
1852
self._new_revision_id = osutils.safe_revision_id(revision_id)
1853
self.parents = parents
1854
self.repository = repository
1857
if revprops is not None:
1858
self._revprops.update(revprops)
1860
if timestamp is None:
1861
timestamp = time.time()
1862
# Restrict resolution to 1ms
1863
self._timestamp = round(timestamp, 3)
1865
if timezone is None:
1866
self._timezone = osutils.local_time_offset()
1868
self._timezone = int(timezone)
1870
self._generate_revision_if_needed()
1872
def commit(self, message):
1873
"""Make the actual commit.
1875
:return: The revision id of the recorded revision.
1877
rev = _mod_revision.Revision(
1878
timestamp=self._timestamp,
1879
timezone=self._timezone,
1880
committer=self._committer,
1882
inventory_sha1=self.inv_sha1,
1883
revision_id=self._new_revision_id,
1884
properties=self._revprops)
1885
rev.parent_ids = self.parents
1886
self.repository.add_revision(self._new_revision_id, rev,
1887
self.new_inventory, self._config)
1888
return self._new_revision_id
1890
def revision_tree(self):
1891
"""Return the tree that was just committed.
1893
After calling commit() this can be called to get a RevisionTree
1894
representing the newly committed tree. This is preferred to
1895
calling Repository.revision_tree() because that may require
1896
deserializing the inventory, while we already have a copy in
1899
return RevisionTree(self.repository, self.new_inventory,
1900
self._new_revision_id)
1902
def finish_inventory(self):
1903
"""Tell the builder that the inventory is finished."""
1904
if self.new_inventory.root is None:
1905
symbol_versioning.warn('Root entry should be supplied to'
1906
' record_entry_contents, as of bzr 0.10.',
1907
DeprecationWarning, stacklevel=2)
1908
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
1909
self.new_inventory.revision_id = self._new_revision_id
1910
self.inv_sha1 = self.repository.add_inventory(
1911
self._new_revision_id,
1916
def _gen_revision_id(self):
1917
"""Return new revision-id."""
1918
return generate_ids.gen_revision_id(self._config.username(),
1921
def _generate_revision_if_needed(self):
1922
"""Create a revision id if None was supplied.
1924
If the repository can not support user-specified revision ids
1925
they should override this function and raise CannotSetRevisionId
1926
if _new_revision_id is not None.
1928
:raises: CannotSetRevisionId
1930
if self._new_revision_id is None:
1931
self._new_revision_id = self._gen_revision_id()
1933
def record_entry_contents(self, ie, parent_invs, path, tree):
1934
"""Record the content of ie from tree into the commit if needed.
1936
Side effect: sets ie.revision when unchanged
1938
:param ie: An inventory entry present in the commit.
1939
:param parent_invs: The inventories of the parent revisions of the
1941
:param path: The path the entry is at in the tree.
1942
:param tree: The tree which contains this entry and should be used to
1945
if self.new_inventory.root is None and ie.parent_id is not None:
1946
symbol_versioning.warn('Root entry should be supplied to'
1947
' record_entry_contents, as of bzr 0.10.',
1948
DeprecationWarning, stacklevel=2)
1949
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
1951
self.new_inventory.add(ie)
1953
# ie.revision is always None if the InventoryEntry is considered
1954
# for committing. ie.snapshot will record the correct revision
1955
# which may be the sole parent if it is untouched.
1956
if ie.revision is not None:
1959
# In this revision format, root entries have no knit or weave
1960
if ie is self.new_inventory.root:
1961
# When serializing out to disk and back in
1962
# root.revision is always _new_revision_id
1963
ie.revision = self._new_revision_id
1965
previous_entries = ie.find_previous_heads(
1967
self.repository.weave_store,
1968
self.repository.get_transaction())
1969
# we are creating a new revision for ie in the history store
1971
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
1973
def modified_directory(self, file_id, file_parents):
1974
"""Record the presence of a symbolic link.
1976
:param file_id: The file_id of the link to record.
1977
:param file_parents: The per-file parent revision ids.
1979
self._add_text_to_weave(file_id, [], file_parents.keys())
1981
def modified_file_text(self, file_id, file_parents,
1982
get_content_byte_lines, text_sha1=None,
1984
"""Record the text of file file_id
1986
:param file_id: The file_id of the file to record the text of.
1987
:param file_parents: The per-file parent revision ids.
1988
:param get_content_byte_lines: A callable which will return the byte
1990
:param text_sha1: Optional SHA1 of the file contents.
1991
:param text_size: Optional size of the file contents.
1993
# mutter('storing text of file {%s} in revision {%s} into %r',
1994
# file_id, self._new_revision_id, self.repository.weave_store)
1995
# special case to avoid diffing on renames or
1997
if (len(file_parents) == 1
1998
and text_sha1 == file_parents.values()[0].text_sha1
1999
and text_size == file_parents.values()[0].text_size):
2000
previous_ie = file_parents.values()[0]
2001
versionedfile = self.repository.weave_store.get_weave(file_id,
2002
self.repository.get_transaction())
2003
versionedfile.clone_text(self._new_revision_id,
2004
previous_ie.revision, file_parents.keys())
2005
return text_sha1, text_size
2007
new_lines = get_content_byte_lines()
2008
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2009
# should return the SHA1 and size
2010
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2011
return osutils.sha_strings(new_lines), \
2012
sum(map(len, new_lines))
2014
def modified_link(self, file_id, file_parents, link_target):
2015
"""Record the presence of a symbolic link.
2017
:param file_id: The file_id of the link to record.
2018
:param file_parents: The per-file parent revision ids.
2019
:param link_target: Target location of this link.
2021
self._add_text_to_weave(file_id, [], file_parents.keys())
2023
def _add_text_to_weave(self, file_id, new_lines, parents):
2024
versionedfile = self.repository.weave_store.get_weave_or_empty(
2025
file_id, self.repository.get_transaction())
2026
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2027
versionedfile.clear_cache()
2030
class _CommitBuilder(CommitBuilder):
2031
"""Temporary class so old CommitBuilders are detected properly
2033
Note: CommitBuilder works whether or not root entry is recorded.
2036
record_root_entry = True
2039
class RootCommitBuilder(CommitBuilder):
2040
"""This commitbuilder actually records the root id"""
2042
record_root_entry = True
2044
def record_entry_contents(self, ie, parent_invs, path, tree):
2045
"""Record the content of ie from tree into the commit if needed.
2047
Side effect: sets ie.revision when unchanged
2049
:param ie: An inventory entry present in the commit.
2050
:param parent_invs: The inventories of the parent revisions of the
2052
:param path: The path the entry is at in the tree.
2053
:param tree: The tree which contains this entry and should be used to
2056
assert self.new_inventory.root is not None or ie.parent_id is None
2057
self.new_inventory.add(ie)
2059
# ie.revision is always None if the InventoryEntry is considered
2060
# for committing. ie.snapshot will record the correct revision
2061
# which may be the sole parent if it is untouched.
2062
if ie.revision is not None:
2065
previous_entries = ie.find_previous_heads(
2067
self.repository.weave_store,
2068
self.repository.get_transaction())
2069
# we are creating a new revision for ie in the history store
2071
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2083
def _unescaper(match, _map=_unescape_map):
2084
code = match.group(1)
2088
if not code.startswith('#'):
2090
return unichr(int(code[1:])).encode('utf8')
2096
def _unescape_xml(data):
2097
"""Unescape predefined XML entities in a string of data."""
2099
if _unescape_re is None:
2100
_unescape_re = re.compile('\&([^;]*);')
2101
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
