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
######################################################################
66
class Repository(object):
67
"""Repository holding history for one or more branches.
69
The repository holds and retrieves historical information including
70
revisions and file history. It's normally accessed only by the Branch,
71
which views a particular line of development through that history.
73
The Repository builds on top of Stores and a Transport, which respectively
74
describe the disk data format and the way of accessing the (possibly
78
_file_ids_altered_regex = lazy_regex.lazy_compile(
79
r'file_id="(?P<file_id>[^"]+)"'
80
r'.*revision="(?P<revision_id>[^"]+)"'
84
def add_inventory(self, revision_id, inv, parents):
85
"""Add the inventory inv to the repository as revision_id.
87
:param parents: The revision ids of the parents that revision_id
88
is known to have and are in the repository already.
90
returns the sha1 of the serialized inventory.
92
revision_id = osutils.safe_revision_id(revision_id)
93
_mod_revision.check_not_reserved_id(revision_id)
94
assert inv.revision_id is None or inv.revision_id == revision_id, \
95
"Mismatch between inventory revision" \
96
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
97
assert inv.root is not None
98
inv_text = self.serialise_inventory(inv)
99
inv_sha1 = osutils.sha_string(inv_text)
100
inv_vf = self.control_weaves.get_weave('inventory',
101
self.get_transaction())
102
self._inventory_add_lines(inv_vf, revision_id, parents,
103
osutils.split_lines(inv_text))
106
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines):
108
for parent in parents:
110
final_parents.append(parent)
112
inv_vf.add_lines(revision_id, final_parents, lines)
115
def add_revision(self, revision_id, rev, inv=None, config=None):
116
"""Add rev to the revision store as revision_id.
118
:param revision_id: the revision id to use.
119
:param rev: The revision object.
120
:param inv: The inventory for the revision. if None, it will be looked
121
up in the inventory storer
122
:param config: If None no digital signature will be created.
123
If supplied its signature_needed method will be used
124
to determine if a signature should be made.
126
revision_id = osutils.safe_revision_id(revision_id)
127
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
129
_mod_revision.check_not_reserved_id(revision_id)
130
if config is not None and config.signature_needed():
132
inv = self.get_inventory(revision_id)
133
plaintext = Testament(rev, inv).as_short_text()
134
self.store_revision_signature(
135
gpg.GPGStrategy(config), plaintext, revision_id)
136
if not revision_id in self.get_inventory_weave():
138
raise errors.WeaveRevisionNotPresent(revision_id,
139
self.get_inventory_weave())
141
# yes, this is not suitable for adding with ghosts.
142
self.add_inventory(revision_id, inv, rev.parent_ids)
143
self._revision_store.add_revision(rev, self.get_transaction())
146
def _all_possible_ids(self):
147
"""Return all the possible revisions that we could find."""
148
return self.get_inventory_weave().versions()
150
def all_revision_ids(self):
151
"""Returns a list of all the revision ids in the repository.
153
This is deprecated because code should generally work on the graph
154
reachable from a particular revision, and ignore any other revisions
155
that might be present. There is no direct replacement method.
157
return self._all_revision_ids()
160
def _all_revision_ids(self):
161
"""Returns a list of all the revision ids in the repository.
163
These are in as much topological order as the underlying store can
164
present: for weaves ghosts may lead to a lack of correctness until
165
the reweave updates the parents list.
167
if self._revision_store.text_store.listable():
168
return self._revision_store.all_revision_ids(self.get_transaction())
169
result = self._all_possible_ids()
170
# TODO: jam 20070210 Ensure that _all_possible_ids returns non-unicode
171
# ids. (It should, since _revision_store's API should change to
172
# return utf8 revision_ids)
173
return self._eliminate_revisions_not_present(result)
175
def break_lock(self):
176
"""Break a lock if one is present from another instance.
178
Uses the ui factory to ask for confirmation if the lock may be from
181
self.control_files.break_lock()
184
def _eliminate_revisions_not_present(self, revision_ids):
185
"""Check every revision id in revision_ids to see if we have it.
187
Returns a set of the present revisions.
190
for id in revision_ids:
191
if self.has_revision(id):
196
def create(a_bzrdir):
197
"""Construct the current default format repository in a_bzrdir."""
198
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
200
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
201
"""instantiate a Repository.
203
:param _format: The format of the repository on disk.
204
:param a_bzrdir: The BzrDir of the repository.
206
In the future we will have a single api for all stores for
207
getting file texts, inventories and revisions, then
208
this construct will accept instances of those things.
210
super(Repository, self).__init__()
211
self._format = _format
212
# the following are part of the public API for Repository:
213
self.bzrdir = a_bzrdir
214
self.control_files = control_files
215
self._revision_store = _revision_store
216
self.text_store = text_store
217
# backwards compatibility
218
self.weave_store = text_store
219
# not right yet - should be more semantically clear ?
221
self.control_store = control_store
222
self.control_weaves = control_store
223
# TODO: make sure to construct the right store classes, etc, depending
224
# on whether escaping is required.
225
self._warn_if_deprecated()
228
return '%s(%r)' % (self.__class__.__name__,
229
self.bzrdir.transport.base)
232
return self.control_files.is_locked()
234
def lock_write(self):
235
self.control_files.lock_write()
238
self.control_files.lock_read()
240
def get_physical_lock_status(self):
241
return self.control_files.get_physical_lock_status()
244
def gather_stats(self, revid=None, committers=None):
245
"""Gather statistics from a revision id.
247
:param revid: The revision id to gather statistics from, if None, then
248
no revision specific statistics are gathered.
249
:param committers: Optional parameter controlling whether to grab
250
a count of committers from the revision specific statistics.
251
:return: A dictionary of statistics. Currently this contains:
252
committers: The number of committers if requested.
253
firstrev: A tuple with timestamp, timezone for the penultimate left
254
most ancestor of revid, if revid is not the NULL_REVISION.
255
latestrev: A tuple with timestamp, timezone for revid, if revid is
256
not the NULL_REVISION.
257
revisions: The total revision count in the repository.
258
size: An estimate disk size of the repository in bytes.
261
if revid and committers:
262
result['committers'] = 0
263
if revid and revid != _mod_revision.NULL_REVISION:
265
all_committers = set()
266
revisions = self.get_ancestry(revid)
267
# pop the leading None
269
first_revision = None
271
# ignore the revisions in the middle - just grab first and last
272
revisions = revisions[0], revisions[-1]
273
for revision in self.get_revisions(revisions):
274
if not first_revision:
275
first_revision = revision
277
all_committers.add(revision.committer)
278
last_revision = revision
280
result['committers'] = len(all_committers)
281
result['firstrev'] = (first_revision.timestamp,
282
first_revision.timezone)
283
result['latestrev'] = (last_revision.timestamp,
284
last_revision.timezone)
286
# now gather global repository information
287
if self.bzrdir.root_transport.listable():
288
c, t = self._revision_store.total_size(self.get_transaction())
289
result['revisions'] = c
294
def missing_revision_ids(self, other, revision_id=None):
295
"""Return the revision ids that other has that this does not.
297
These are returned in topological order.
299
revision_id: only return revision ids included by revision_id.
301
revision_id = osutils.safe_revision_id(revision_id)
302
return InterRepository.get(other, self).missing_revision_ids(revision_id)
306
"""Open the repository rooted at base.
308
For instance, if the repository is at URL/.bzr/repository,
309
Repository.open(URL) -> a Repository instance.
311
control = bzrdir.BzrDir.open(base)
312
return control.open_repository()
314
def copy_content_into(self, destination, revision_id=None, basis=None):
315
"""Make a complete copy of the content in self into destination.
317
This is a destructive operation! Do not use it on existing
320
revision_id = osutils.safe_revision_id(revision_id)
321
return InterRepository.get(self, destination).copy_content(revision_id, basis)
323
def fetch(self, source, revision_id=None, pb=None):
324
"""Fetch the content required to construct revision_id from source.
326
If revision_id is None all content is copied.
328
revision_id = osutils.safe_revision_id(revision_id)
329
return InterRepository.get(source, self).fetch(revision_id=revision_id,
332
def get_commit_builder(self, branch, parents, config, timestamp=None,
333
timezone=None, committer=None, revprops=None,
335
"""Obtain a CommitBuilder for this repository.
337
:param branch: Branch to commit to.
338
:param parents: Revision ids of the parents of the new revision.
339
:param config: Configuration to use.
340
:param timestamp: Optional timestamp recorded for commit.
341
:param timezone: Optional timezone for timestamp.
342
:param committer: Optional committer to set for commit.
343
:param revprops: Optional dictionary of revision properties.
344
:param revision_id: Optional revision id.
346
revision_id = osutils.safe_revision_id(revision_id)
347
return _CommitBuilder(self, parents, config, timestamp, timezone,
348
committer, revprops, revision_id)
351
self.control_files.unlock()
354
def clone(self, a_bzrdir, revision_id=None, basis=None):
355
"""Clone this repository into a_bzrdir using the current format.
357
Currently no check is made that the format of this repository and
358
the bzrdir format are compatible. FIXME RBC 20060201.
360
:return: The newly created destination repository.
362
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
363
# use target default format.
364
dest_repo = a_bzrdir.create_repository()
366
# Most control formats need the repository to be specifically
367
# created, but on some old all-in-one formats it's not needed
369
dest_repo = self._format.initialize(a_bzrdir, shared=self.is_shared())
370
except errors.UninitializableFormat:
371
dest_repo = a_bzrdir.open_repository()
372
self.copy_content_into(dest_repo, revision_id, basis)
376
def has_revision(self, revision_id):
377
"""True if this repository has a copy of the revision."""
378
revision_id = osutils.safe_revision_id(revision_id)
379
return self._revision_store.has_revision_id(revision_id,
380
self.get_transaction())
383
def get_revision_reconcile(self, revision_id):
384
"""'reconcile' helper routine that allows access to a revision always.
386
This variant of get_revision does not cross check the weave graph
387
against the revision one as get_revision does: but it should only
388
be used by reconcile, or reconcile-alike commands that are correcting
389
or testing the revision graph.
391
if not revision_id or not isinstance(revision_id, basestring):
392
raise errors.InvalidRevisionId(revision_id=revision_id,
394
return self.get_revisions([revision_id])[0]
397
def get_revisions(self, revision_ids):
398
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
399
revs = self._revision_store.get_revisions(revision_ids,
400
self.get_transaction())
402
assert not isinstance(rev.revision_id, unicode)
403
for parent_id in rev.parent_ids:
404
assert not isinstance(parent_id, unicode)
408
def get_revision_xml(self, revision_id):
409
# TODO: jam 20070210 This shouldn't be necessary since get_revision
410
# would have already do it.
411
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
412
revision_id = osutils.safe_revision_id(revision_id)
413
rev = self.get_revision(revision_id)
415
# the current serializer..
416
self._revision_store._serializer.write_revision(rev, rev_tmp)
418
return rev_tmp.getvalue()
421
def get_revision(self, revision_id):
422
"""Return the Revision object for a named revision"""
423
# TODO: jam 20070210 get_revision_reconcile should do this for us
424
revision_id = osutils.safe_revision_id(revision_id)
425
r = self.get_revision_reconcile(revision_id)
426
# weave corruption can lead to absent revision markers that should be
428
# the following test is reasonably cheap (it needs a single weave read)
429
# and the weave is cached in read transactions. In write transactions
430
# it is not cached but typically we only read a small number of
431
# revisions. For knits when they are introduced we will probably want
432
# to ensure that caching write transactions are in use.
433
inv = self.get_inventory_weave()
434
self._check_revision_parents(r, inv)
438
def get_deltas_for_revisions(self, revisions):
439
"""Produce a generator of revision deltas.
441
Note that the input is a sequence of REVISIONS, not revision_ids.
442
Trees will be held in memory until the generator exits.
443
Each delta is relative to the revision's lefthand predecessor.
445
required_trees = set()
446
for revision in revisions:
447
required_trees.add(revision.revision_id)
448
required_trees.update(revision.parent_ids[:1])
449
trees = dict((t.get_revision_id(), t) for
450
t in self.revision_trees(required_trees))
451
for revision in revisions:
452
if not revision.parent_ids:
453
old_tree = self.revision_tree(None)
455
old_tree = trees[revision.parent_ids[0]]
456
yield trees[revision.revision_id].changes_from(old_tree)
459
def get_revision_delta(self, revision_id):
460
"""Return the delta for one revision.
462
The delta is relative to the left-hand predecessor of the
465
r = self.get_revision(revision_id)
466
return list(self.get_deltas_for_revisions([r]))[0]
468
def _check_revision_parents(self, revision, inventory):
469
"""Private to Repository and Fetch.
471
This checks the parentage of revision in an inventory weave for
472
consistency and is only applicable to inventory-weave-for-ancestry
473
using repository formats & fetchers.
475
weave_parents = inventory.get_parents(revision.revision_id)
476
weave_names = inventory.versions()
477
for parent_id in revision.parent_ids:
478
if parent_id in weave_names:
479
# this parent must not be a ghost.
480
if not parent_id in weave_parents:
482
raise errors.CorruptRepository(self)
485
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
486
revision_id = osutils.safe_revision_id(revision_id)
487
signature = gpg_strategy.sign(plaintext)
488
self._revision_store.add_revision_signature_text(revision_id,
490
self.get_transaction())
492
def fileids_altered_by_revision_ids(self, revision_ids):
493
"""Find the file ids and versions affected by revisions.
495
:param revisions: an iterable containing revision ids.
496
:return: a dictionary mapping altered file-ids to an iterable of
497
revision_ids. Each altered file-ids has the exact revision_ids that
498
altered it listed explicitly.
500
assert self._serializer.support_altered_by_hack, \
501
("fileids_altered_by_revision_ids only supported for branches "
502
"which store inventory as unnested xml, not on %r" % self)
503
selected_revision_ids = set(osutils.safe_revision_id(r)
504
for r in revision_ids)
505
w = self.get_inventory_weave()
508
# this code needs to read every new line in every inventory for the
509
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
510
# not present in one of those inventories is unnecessary but not
511
# harmful because we are filtering by the revision id marker in the
512
# inventory lines : we only select file ids altered in one of those
513
# revisions. We don't need to see all lines in the inventory because
514
# only those added in an inventory in rev X can contain a revision=X
516
unescape_revid_cache = {}
517
unescape_fileid_cache = {}
519
# jam 20061218 In a big fetch, this handles hundreds of thousands
520
# of lines, so it has had a lot of inlining and optimizing done.
521
# Sorry that it is a little bit messy.
522
# Move several functions to be local variables, since this is a long
524
search = self._file_ids_altered_regex.search
525
unescape = _unescape_xml
526
setdefault = result.setdefault
527
pb = ui.ui_factory.nested_progress_bar()
529
for line in w.iter_lines_added_or_present_in_versions(
530
selected_revision_ids, pb=pb):
534
# One call to match.group() returning multiple items is quite a
535
# bit faster than 2 calls to match.group() each returning 1
536
file_id, revision_id = match.group('file_id', 'revision_id')
538
# Inlining the cache lookups helps a lot when you make 170,000
539
# lines and 350k ids, versus 8.4 unique ids.
540
# Using a cache helps in 2 ways:
541
# 1) Avoids unnecessary decoding calls
542
# 2) Re-uses cached strings, which helps in future set and
544
# (2) is enough that removing encoding entirely along with
545
# the cache (so we are using plain strings) results in no
546
# performance improvement.
548
revision_id = unescape_revid_cache[revision_id]
550
unescaped = unescape(revision_id)
551
unescape_revid_cache[revision_id] = unescaped
552
revision_id = unescaped
554
if revision_id in selected_revision_ids:
556
file_id = unescape_fileid_cache[file_id]
558
unescaped = unescape(file_id)
559
unescape_fileid_cache[file_id] = unescaped
561
setdefault(file_id, set()).add(revision_id)
567
def get_inventory_weave(self):
568
return self.control_weaves.get_weave('inventory',
569
self.get_transaction())
572
def get_inventory(self, revision_id):
573
"""Get Inventory object by hash."""
574
# TODO: jam 20070210 Technically we don't need to sanitize, since all
575
# called functions must sanitize.
576
revision_id = osutils.safe_revision_id(revision_id)
577
return self.deserialise_inventory(
578
revision_id, self.get_inventory_xml(revision_id))
580
def deserialise_inventory(self, revision_id, xml):
581
"""Transform the xml into an inventory object.
583
:param revision_id: The expected revision id of the inventory.
584
:param xml: A serialised inventory.
586
revision_id = osutils.safe_revision_id(revision_id)
587
result = self._serializer.read_inventory_from_string(xml)
588
result.root.revision = revision_id
591
def serialise_inventory(self, inv):
592
return self._serializer.write_inventory_to_string(inv)
595
def get_inventory_xml(self, revision_id):
596
"""Get inventory XML as a file object."""
597
revision_id = osutils.safe_revision_id(revision_id)
599
assert isinstance(revision_id, str), type(revision_id)
600
iw = self.get_inventory_weave()
601
return iw.get_text(revision_id)
603
raise errors.HistoryMissing(self, 'inventory', revision_id)
606
def get_inventory_sha1(self, revision_id):
607
"""Return the sha1 hash of the inventory entry
609
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
610
revision_id = osutils.safe_revision_id(revision_id)
611
return self.get_revision(revision_id).inventory_sha1
614
def get_revision_graph(self, revision_id=None):
615
"""Return a dictionary containing the revision graph.
617
:param revision_id: The revision_id to get a graph from. If None, then
618
the entire revision graph is returned. This is a deprecated mode of
619
operation and will be removed in the future.
620
:return: a dictionary of revision_id->revision_parents_list.
622
# special case NULL_REVISION
623
if revision_id == _mod_revision.NULL_REVISION:
625
revision_id = osutils.safe_revision_id(revision_id)
626
a_weave = self.get_inventory_weave()
627
all_revisions = self._eliminate_revisions_not_present(
629
entire_graph = dict([(node, a_weave.get_parents(node)) for
630
node in all_revisions])
631
if revision_id is None:
633
elif revision_id not in entire_graph:
634
raise errors.NoSuchRevision(self, revision_id)
636
# add what can be reached from revision_id
638
pending = set([revision_id])
639
while len(pending) > 0:
641
result[node] = entire_graph[node]
642
for revision_id in result[node]:
643
if revision_id not in result:
644
pending.add(revision_id)
648
def get_revision_graph_with_ghosts(self, revision_ids=None):
649
"""Return a graph of the revisions with ghosts marked as applicable.
651
:param revision_ids: an iterable of revisions to graph or None for all.
652
:return: a Graph object with the graph reachable from revision_ids.
654
result = graph.Graph()
656
pending = set(self.all_revision_ids())
659
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
660
# special case NULL_REVISION
661
if _mod_revision.NULL_REVISION in pending:
662
pending.remove(_mod_revision.NULL_REVISION)
663
required = set(pending)
666
revision_id = pending.pop()
668
rev = self.get_revision(revision_id)
669
except errors.NoSuchRevision:
670
if revision_id in required:
673
result.add_ghost(revision_id)
675
for parent_id in rev.parent_ids:
676
# is this queued or done ?
677
if (parent_id not in pending and
678
parent_id not in done):
680
pending.add(parent_id)
681
result.add_node(revision_id, rev.parent_ids)
682
done.add(revision_id)
685
def _get_history_vf(self):
686
"""Get a versionedfile whose history graph reflects all revisions.
688
For weave repositories, this is the inventory weave.
690
return self.get_inventory_weave()
692
def iter_reverse_revision_history(self, revision_id):
693
"""Iterate backwards through revision ids in the lefthand history
695
:param revision_id: The revision id to start with. All its lefthand
696
ancestors will be traversed.
698
revision_id = osutils.safe_revision_id(revision_id)
699
if revision_id in (None, _mod_revision.NULL_REVISION):
701
next_id = revision_id
702
versionedfile = self._get_history_vf()
705
parents = versionedfile.get_parents(next_id)
706
if len(parents) == 0:
712
def get_revision_inventory(self, revision_id):
713
"""Return inventory of a past revision."""
714
# TODO: Unify this with get_inventory()
715
# bzr 0.0.6 and later imposes the constraint that the inventory_id
716
# must be the same as its revision, so this is trivial.
717
if revision_id is None:
718
# This does not make sense: if there is no revision,
719
# then it is the current tree inventory surely ?!
720
# and thus get_root_id() is something that looks at the last
721
# commit on the branch, and the get_root_id is an inventory check.
722
raise NotImplementedError
723
# return Inventory(self.get_root_id())
725
return self.get_inventory(revision_id)
729
"""Return True if this repository is flagged as a shared repository."""
730
raise NotImplementedError(self.is_shared)
733
def reconcile(self, other=None, thorough=False):
734
"""Reconcile this repository."""
735
from bzrlib.reconcile import RepoReconciler
736
reconciler = RepoReconciler(self, thorough=thorough)
737
reconciler.reconcile()
741
def revision_tree(self, revision_id):
742
"""Return Tree for a revision on this branch.
744
`revision_id` may be None for the empty tree revision.
746
# TODO: refactor this to use an existing revision object
747
# so we don't need to read it in twice.
748
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
749
return RevisionTree(self, Inventory(root_id=None),
750
_mod_revision.NULL_REVISION)
752
revision_id = osutils.safe_revision_id(revision_id)
753
inv = self.get_revision_inventory(revision_id)
754
return RevisionTree(self, inv, revision_id)
757
def revision_trees(self, revision_ids):
758
"""Return Tree for a revision on this branch.
760
`revision_id` may not be None or 'null:'"""
761
assert None not in revision_ids
762
assert _mod_revision.NULL_REVISION not in revision_ids
763
texts = self.get_inventory_weave().get_texts(revision_ids)
764
for text, revision_id in zip(texts, revision_ids):
765
inv = self.deserialise_inventory(revision_id, text)
766
yield RevisionTree(self, inv, revision_id)
769
def get_ancestry(self, revision_id):
770
"""Return a list of revision-ids integrated by a revision.
772
The first element of the list is always None, indicating the origin
773
revision. This might change when we have history horizons, or
774
perhaps we should have a new API.
776
This is topologically sorted.
778
if revision_id is None:
780
revision_id = osutils.safe_revision_id(revision_id)
781
if not self.has_revision(revision_id):
782
raise errors.NoSuchRevision(self, revision_id)
783
w = self.get_inventory_weave()
784
candidates = w.get_ancestry(revision_id)
785
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
788
def print_file(self, file, revision_id):
789
"""Print `file` to stdout.
791
FIXME RBC 20060125 as John Meinel points out this is a bad api
792
- it writes to stdout, it assumes that that is valid etc. Fix
793
by creating a new more flexible convenience function.
795
revision_id = osutils.safe_revision_id(revision_id)
796
tree = self.revision_tree(revision_id)
797
# use inventory as it was in that revision
798
file_id = tree.inventory.path2id(file)
800
# TODO: jam 20060427 Write a test for this code path
801
# it had a bug in it, and was raising the wrong
803
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
804
tree.print_file(file_id)
806
def get_transaction(self):
807
return self.control_files.get_transaction()
809
def revision_parents(self, revision_id):
810
revision_id = osutils.safe_revision_id(revision_id)
811
return self.get_inventory_weave().parent_names(revision_id)
814
def set_make_working_trees(self, new_value):
815
"""Set the policy flag for making working trees when creating branches.
817
This only applies to branches that use this repository.
819
The default is 'True'.
820
:param new_value: True to restore the default, False to disable making
823
raise NotImplementedError(self.set_make_working_trees)
825
def make_working_trees(self):
826
"""Returns the policy for making working trees on new branches."""
827
raise NotImplementedError(self.make_working_trees)
830
def sign_revision(self, revision_id, gpg_strategy):
831
revision_id = osutils.safe_revision_id(revision_id)
832
plaintext = Testament.from_revision(self, revision_id).as_short_text()
833
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
836
def has_signature_for_revision_id(self, revision_id):
837
"""Query for a revision signature for revision_id in the repository."""
838
revision_id = osutils.safe_revision_id(revision_id)
839
return self._revision_store.has_signature(revision_id,
840
self.get_transaction())
843
def get_signature_text(self, revision_id):
844
"""Return the text for a signature."""
845
revision_id = osutils.safe_revision_id(revision_id)
846
return self._revision_store.get_signature_text(revision_id,
847
self.get_transaction())
850
def check(self, revision_ids):
851
"""Check consistency of all history of given revision_ids.
853
Different repository implementations should override _check().
855
:param revision_ids: A non-empty list of revision_ids whose ancestry
856
will be checked. Typically the last revision_id of a branch.
859
raise ValueError("revision_ids must be non-empty in %s.check"
861
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
862
return self._check(revision_ids)
864
def _check(self, revision_ids):
865
result = check.Check(self)
869
def _warn_if_deprecated(self):
870
global _deprecation_warning_done
871
if _deprecation_warning_done:
873
_deprecation_warning_done = True
874
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
875
% (self._format, self.bzrdir.transport.base))
877
def supports_rich_root(self):
878
return self._format.rich_root_data
880
def _check_ascii_revisionid(self, revision_id, method):
881
"""Private helper for ascii-only repositories."""
882
# weave repositories refuse to store revisionids that are non-ascii.
883
if revision_id is not None:
884
# weaves require ascii revision ids.
885
if isinstance(revision_id, unicode):
887
revision_id.encode('ascii')
888
except UnicodeEncodeError:
889
raise errors.NonAsciiRevisionId(method, self)
892
revision_id.decode('ascii')
893
except UnicodeDecodeError:
894
raise errors.NonAsciiRevisionId(method, self)
898
# remove these delegates a while after bzr 0.15
899
def __make_delegated(name, from_module):
900
def _deprecated_repository_forwarder():
901
symbol_versioning.warn('%s moved to %s in bzr 0.15'
902
% (name, from_module),
905
m = __import__(from_module, globals(), locals(), [name])
907
return getattr(m, name)
908
except AttributeError:
909
raise AttributeError('module %s has no name %s'
911
globals()[name] = _deprecated_repository_forwarder
914
'AllInOneRepository',
915
'WeaveMetaDirRepository',
916
'PreSplitOutRepositoryFormat',
922
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
927
'RepositoryFormatKnit',
928
'RepositoryFormatKnit1',
929
'RepositoryFormatKnit2',
931
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
934
def install_revision(repository, rev, revision_tree):
935
"""Install all revision data into a repository."""
938
for p_id in rev.parent_ids:
939
if repository.has_revision(p_id):
940
present_parents.append(p_id)
941
parent_trees[p_id] = repository.revision_tree(p_id)
943
parent_trees[p_id] = repository.revision_tree(None)
945
inv = revision_tree.inventory
946
entries = inv.iter_entries()
947
# backwards compatability hack: skip the root id.
948
if not repository.supports_rich_root():
949
path, root = entries.next()
950
if root.revision != rev.revision_id:
951
raise errors.IncompatibleRevision(repr(repository))
952
# Add the texts that are not already present
953
for path, ie in entries:
954
w = repository.weave_store.get_weave_or_empty(ie.file_id,
955
repository.get_transaction())
956
if ie.revision not in w:
958
# FIXME: TODO: The following loop *may* be overlapping/duplicate
959
# with InventoryEntry.find_previous_heads(). if it is, then there
960
# is a latent bug here where the parents may have ancestors of each
962
for revision, tree in parent_trees.iteritems():
963
if ie.file_id not in tree:
965
parent_id = tree.inventory[ie.file_id].revision
966
if parent_id in text_parents:
968
text_parents.append(parent_id)
970
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
971
repository.get_transaction())
972
lines = revision_tree.get_file(ie.file_id).readlines()
973
vfile.add_lines(rev.revision_id, text_parents, lines)
975
# install the inventory
976
repository.add_inventory(rev.revision_id, inv, present_parents)
977
except errors.RevisionAlreadyPresent:
979
repository.add_revision(rev.revision_id, rev, inv)
982
class MetaDirRepository(Repository):
983
"""Repositories in the new meta-dir layout."""
985
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
986
super(MetaDirRepository, self).__init__(_format,
992
dir_mode = self.control_files._dir_mode
993
file_mode = self.control_files._file_mode
997
"""Return True if this repository is flagged as a shared repository."""
998
return self.control_files._transport.has('shared-storage')
1001
def set_make_working_trees(self, new_value):
1002
"""Set the policy flag for making working trees when creating branches.
1004
This only applies to branches that use this repository.
1006
The default is 'True'.
1007
:param new_value: True to restore the default, False to disable making
1012
self.control_files._transport.delete('no-working-trees')
1013
except errors.NoSuchFile:
1016
self.control_files.put_utf8('no-working-trees', '')
1018
def make_working_trees(self):
1019
"""Returns the policy for making working trees on new branches."""
1020
return not self.control_files._transport.has('no-working-trees')
1023
class RepositoryFormatRegistry(registry.Registry):
1024
"""Registry of RepositoryFormats.
1027
def get(self, format_string):
1028
r = registry.Registry.get(self, format_string)
1034
format_registry = RepositoryFormatRegistry()
1035
"""Registry of formats, indexed by their identifying format string.
1037
This can contain either format instances themselves, or classes/factories that
1038
can be called to obtain one.
1042
#####################################################################
1043
# Repository Formats
1045
class RepositoryFormat(object):
1046
"""A repository format.
1048
Formats provide three things:
1049
* An initialization routine to construct repository data on disk.
1050
* a format string which is used when the BzrDir supports versioned
1052
* an open routine which returns a Repository instance.
1054
Formats are placed in an dict by their format string for reference
1055
during opening. These should be subclasses of RepositoryFormat
1058
Once a format is deprecated, just deprecate the initialize and open
1059
methods on the format class. Do not deprecate the object, as the
1060
object will be created every system load.
1062
Common instance attributes:
1063
_matchingbzrdir - the bzrdir format that the repository format was
1064
originally written to work with. This can be used if manually
1065
constructing a bzrdir and repository, or more commonly for test suite
1070
return "<%s>" % self.__class__.__name__
1072
def __eq__(self, other):
1073
# format objects are generally stateless
1074
return isinstance(other, self.__class__)
1077
def find_format(klass, a_bzrdir):
1078
"""Return the format for the repository object in a_bzrdir.
1080
This is used by bzr native formats that have a "format" file in
1081
the repository. Other methods may be used by different types of
1085
transport = a_bzrdir.get_repository_transport(None)
1086
format_string = transport.get("format").read()
1087
return format_registry.get(format_string)
1088
except errors.NoSuchFile:
1089
raise errors.NoRepositoryPresent(a_bzrdir)
1091
raise errors.UnknownFormatError(format=format_string)
1094
def register_format(klass, format):
1095
format_registry.register(format.get_format_string(), format)
1098
def unregister_format(klass, format):
1099
format_registry.remove(format.get_format_string())
1102
def get_default_format(klass):
1103
"""Return the current default format."""
1104
from bzrlib import bzrdir
1105
return bzrdir.format_registry.make_bzrdir('default').repository_format
1107
def _get_control_store(self, repo_transport, control_files):
1108
"""Return the control store for this repository."""
1109
raise NotImplementedError(self._get_control_store)
1111
def get_format_string(self):
1112
"""Return the ASCII format string that identifies this format.
1114
Note that in pre format ?? repositories the format string is
1115
not permitted nor written to disk.
1117
raise NotImplementedError(self.get_format_string)
1119
def get_format_description(self):
1120
"""Return the short description for this format."""
1121
raise NotImplementedError(self.get_format_description)
1123
def _get_revision_store(self, repo_transport, control_files):
1124
"""Return the revision store object for this a_bzrdir."""
1125
raise NotImplementedError(self._get_revision_store)
1127
def _get_text_rev_store(self,
1134
"""Common logic for getting a revision store for a repository.
1136
see self._get_revision_store for the subclass-overridable method to
1137
get the store for a repository.
1139
from bzrlib.store.revision.text import TextRevisionStore
1140
dir_mode = control_files._dir_mode
1141
file_mode = control_files._file_mode
1142
text_store = TextStore(transport.clone(name),
1144
compressed=compressed,
1146
file_mode=file_mode)
1147
_revision_store = TextRevisionStore(text_store, serializer)
1148
return _revision_store
1150
# TODO: this shouldn't be in the base class, it's specific to things that
1151
# use weaves or knits -- mbp 20070207
1152
def _get_versioned_file_store(self,
1157
versionedfile_class=None,
1158
versionedfile_kwargs={},
1160
if versionedfile_class is None:
1161
versionedfile_class = self._versionedfile_class
1162
weave_transport = control_files._transport.clone(name)
1163
dir_mode = control_files._dir_mode
1164
file_mode = control_files._file_mode
1165
return VersionedFileStore(weave_transport, prefixed=prefixed,
1167
file_mode=file_mode,
1168
versionedfile_class=versionedfile_class,
1169
versionedfile_kwargs=versionedfile_kwargs,
1172
def initialize(self, a_bzrdir, shared=False):
1173
"""Initialize a repository of this format in a_bzrdir.
1175
:param a_bzrdir: The bzrdir to put the new repository in it.
1176
:param shared: The repository should be initialized as a sharable one.
1178
This may raise UninitializableFormat if shared repository are not
1179
compatible the a_bzrdir.
1182
def is_supported(self):
1183
"""Is this format supported?
1185
Supported formats must be initializable and openable.
1186
Unsupported formats may not support initialization or committing or
1187
some other features depending on the reason for not being supported.
1191
def check_conversion_target(self, target_format):
1192
raise NotImplementedError(self.check_conversion_target)
1194
def open(self, a_bzrdir, _found=False):
1195
"""Return an instance of this format for the bzrdir a_bzrdir.
1197
_found is a private parameter, do not use it.
1199
raise NotImplementedError(self.open)
1202
class MetaDirRepositoryFormat(RepositoryFormat):
1203
"""Common base class for the new repositories using the metadir layout."""
1205
rich_root_data = False
1206
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1209
super(MetaDirRepositoryFormat, self).__init__()
1211
def _create_control_files(self, a_bzrdir):
1212
"""Create the required files and the initial control_files object."""
1213
# FIXME: RBC 20060125 don't peek under the covers
1214
# NB: no need to escape relative paths that are url safe.
1215
repository_transport = a_bzrdir.get_repository_transport(self)
1216
control_files = lockable_files.LockableFiles(repository_transport,
1217
'lock', lockdir.LockDir)
1218
control_files.create_lock()
1219
return control_files
1221
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1222
"""Upload the initial blank content."""
1223
control_files = self._create_control_files(a_bzrdir)
1224
control_files.lock_write()
1226
control_files._transport.mkdir_multi(dirs,
1227
mode=control_files._dir_mode)
1228
for file, content in files:
1229
control_files.put(file, content)
1230
for file, content in utf8_files:
1231
control_files.put_utf8(file, content)
1233
control_files.put_utf8('shared-storage', '')
1235
control_files.unlock()
1238
# formats which have no format string are not discoverable
1239
# and not independently creatable, so are not registered. They're
1240
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1241
# needed, it's constructed directly by the BzrDir. Non-native formats where
1242
# the repository is not separately opened are similar.
1244
format_registry.register_lazy(
1245
'Bazaar-NG Repository format 7',
1246
'bzrlib.repofmt.weaverepo',
1249
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1250
# default control directory format
1252
format_registry.register_lazy(
1253
'Bazaar-NG Knit Repository Format 1',
1254
'bzrlib.repofmt.knitrepo',
1255
'RepositoryFormatKnit1',
1257
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1259
format_registry.register_lazy(
1260
'Bazaar Knit Repository Format 2\n',
1261
'bzrlib.repofmt.knitrepo',
1262
'RepositoryFormatKnit2',
1266
class InterRepository(InterObject):
1267
"""This class represents operations taking place between two repositories.
1269
Its instances have methods like copy_content and fetch, and contain
1270
references to the source and target repositories these operations can be
1273
Often we will provide convenience methods on 'repository' which carry out
1274
operations with another repository - they will always forward to
1275
InterRepository.get(other).method_name(parameters).
1279
"""The available optimised InterRepository types."""
1281
def copy_content(self, revision_id=None, basis=None):
1282
raise NotImplementedError(self.copy_content)
1284
def fetch(self, revision_id=None, pb=None):
1285
"""Fetch the content required to construct revision_id.
1287
The content is copied from self.source to self.target.
1289
:param revision_id: if None all content is copied, if NULL_REVISION no
1291
:param pb: optional progress bar to use for progress reports. If not
1292
provided a default one will be created.
1294
Returns the copied revision count and the failed revisions in a tuple:
1297
raise NotImplementedError(self.fetch)
1300
def missing_revision_ids(self, revision_id=None):
1301
"""Return the revision ids that source has that target does not.
1303
These are returned in topological order.
1305
:param revision_id: only return revision ids included by this
1308
# generic, possibly worst case, slow code path.
1309
target_ids = set(self.target.all_revision_ids())
1310
if revision_id is not None:
1311
# TODO: jam 20070210 InterRepository is internal enough that it
1312
# should assume revision_ids are already utf-8
1313
revision_id = osutils.safe_revision_id(revision_id)
1314
source_ids = self.source.get_ancestry(revision_id)
1315
assert source_ids[0] is None
1318
source_ids = self.source.all_revision_ids()
1319
result_set = set(source_ids).difference(target_ids)
1320
# this may look like a no-op: its not. It preserves the ordering
1321
# other_ids had while only returning the members from other_ids
1322
# that we've decided we need.
1323
return [rev_id for rev_id in source_ids if rev_id in result_set]
1326
class InterSameDataRepository(InterRepository):
1327
"""Code for converting between repositories that represent the same data.
1329
Data format and model must match for this to work.
1333
def _get_repo_format_to_test(self):
1334
"""Repository format for testing with."""
1335
return RepositoryFormat.get_default_format()
1338
def is_compatible(source, target):
1339
if not isinstance(source, Repository):
1341
if not isinstance(target, Repository):
1343
if source._format.rich_root_data == target._format.rich_root_data:
1349
def copy_content(self, revision_id=None, basis=None):
1350
"""Make a complete copy of the content in self into destination.
1352
This is a destructive operation! Do not use it on existing
1355
:param revision_id: Only copy the content needed to construct
1356
revision_id and its parents.
1357
:param basis: Copy the needed data preferentially from basis.
1360
self.target.set_make_working_trees(self.source.make_working_trees())
1361
except NotImplementedError:
1363
# TODO: jam 20070210 This is fairly internal, so we should probably
1364
# just assert that revision_id is not unicode.
1365
revision_id = osutils.safe_revision_id(revision_id)
1366
# grab the basis available data
1367
if basis is not None:
1368
self.target.fetch(basis, revision_id=revision_id)
1369
# but don't bother fetching if we have the needed data now.
1370
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1371
self.target.has_revision(revision_id)):
1373
self.target.fetch(self.source, revision_id=revision_id)
1376
def fetch(self, revision_id=None, pb=None):
1377
"""See InterRepository.fetch()."""
1378
from bzrlib.fetch import GenericRepoFetcher
1379
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1380
self.source, self.source._format, self.target,
1381
self.target._format)
1382
# TODO: jam 20070210 This should be an assert, not a translate
1383
revision_id = osutils.safe_revision_id(revision_id)
1384
f = GenericRepoFetcher(to_repository=self.target,
1385
from_repository=self.source,
1386
last_revision=revision_id,
1388
return f.count_copied, f.failed_revisions
1391
class InterWeaveRepo(InterSameDataRepository):
1392
"""Optimised code paths between Weave based repositories."""
1395
def _get_repo_format_to_test(self):
1396
from bzrlib.repofmt import weaverepo
1397
return weaverepo.RepositoryFormat7()
1400
def is_compatible(source, target):
1401
"""Be compatible with known Weave formats.
1403
We don't test for the stores being of specific types because that
1404
could lead to confusing results, and there is no need to be
1407
from bzrlib.repofmt.weaverepo import (
1413
return (isinstance(source._format, (RepositoryFormat5,
1415
RepositoryFormat7)) and
1416
isinstance(target._format, (RepositoryFormat5,
1418
RepositoryFormat7)))
1419
except AttributeError:
1423
def copy_content(self, revision_id=None, basis=None):
1424
"""See InterRepository.copy_content()."""
1425
# weave specific optimised path:
1426
# TODO: jam 20070210 Internal, should be an assert, not translate
1427
revision_id = osutils.safe_revision_id(revision_id)
1428
if basis is not None:
1429
# copy the basis in, then fetch remaining data.
1430
basis.copy_content_into(self.target, revision_id)
1431
# the basis copy_content_into could miss-set this.
1433
self.target.set_make_working_trees(self.source.make_working_trees())
1434
except NotImplementedError:
1436
self.target.fetch(self.source, revision_id=revision_id)
1439
self.target.set_make_working_trees(self.source.make_working_trees())
1440
except NotImplementedError:
1442
# FIXME do not peek!
1443
if self.source.control_files._transport.listable():
1444
pb = ui.ui_factory.nested_progress_bar()
1446
self.target.weave_store.copy_all_ids(
1447
self.source.weave_store,
1449
from_transaction=self.source.get_transaction(),
1450
to_transaction=self.target.get_transaction())
1451
pb.update('copying inventory', 0, 1)
1452
self.target.control_weaves.copy_multi(
1453
self.source.control_weaves, ['inventory'],
1454
from_transaction=self.source.get_transaction(),
1455
to_transaction=self.target.get_transaction())
1456
self.target._revision_store.text_store.copy_all_ids(
1457
self.source._revision_store.text_store,
1462
self.target.fetch(self.source, revision_id=revision_id)
1465
def fetch(self, revision_id=None, pb=None):
1466
"""See InterRepository.fetch()."""
1467
from bzrlib.fetch import GenericRepoFetcher
1468
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1469
self.source, self.source._format, self.target, self.target._format)
1470
# TODO: jam 20070210 This should be an assert, not a translate
1471
revision_id = osutils.safe_revision_id(revision_id)
1472
f = GenericRepoFetcher(to_repository=self.target,
1473
from_repository=self.source,
1474
last_revision=revision_id,
1476
return f.count_copied, f.failed_revisions
1479
def missing_revision_ids(self, revision_id=None):
1480
"""See InterRepository.missing_revision_ids()."""
1481
# we want all revisions to satisfy revision_id in source.
1482
# but we don't want to stat every file here and there.
1483
# we want then, all revisions other needs to satisfy revision_id
1484
# checked, but not those that we have locally.
1485
# so the first thing is to get a subset of the revisions to
1486
# satisfy revision_id in source, and then eliminate those that
1487
# we do already have.
1488
# this is slow on high latency connection to self, but as as this
1489
# disk format scales terribly for push anyway due to rewriting
1490
# inventory.weave, this is considered acceptable.
1492
if revision_id is not None:
1493
source_ids = self.source.get_ancestry(revision_id)
1494
assert source_ids[0] is None
1497
source_ids = self.source._all_possible_ids()
1498
source_ids_set = set(source_ids)
1499
# source_ids is the worst possible case we may need to pull.
1500
# now we want to filter source_ids against what we actually
1501
# have in target, but don't try to check for existence where we know
1502
# we do not have a revision as that would be pointless.
1503
target_ids = set(self.target._all_possible_ids())
1504
possibly_present_revisions = target_ids.intersection(source_ids_set)
1505
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1506
required_revisions = source_ids_set.difference(actually_present_revisions)
1507
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1508
if revision_id is not None:
1509
# we used get_ancestry to determine source_ids then we are assured all
1510
# revisions referenced are present as they are installed in topological order.
1511
# and the tip revision was validated by get_ancestry.
1512
return required_topo_revisions
1514
# if we just grabbed the possibly available ids, then
1515
# we only have an estimate of whats available and need to validate
1516
# that against the revision records.
1517
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1520
class InterKnitRepo(InterSameDataRepository):
1521
"""Optimised code paths between Knit based repositories."""
1524
def _get_repo_format_to_test(self):
1525
from bzrlib.repofmt import knitrepo
1526
return knitrepo.RepositoryFormatKnit1()
1529
def is_compatible(source, target):
1530
"""Be compatible with known Knit formats.
1532
We don't test for the stores being of specific types because that
1533
could lead to confusing results, and there is no need to be
1536
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1538
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1539
isinstance(target._format, (RepositoryFormatKnit1)))
1540
except AttributeError:
1544
def fetch(self, revision_id=None, pb=None):
1545
"""See InterRepository.fetch()."""
1546
from bzrlib.fetch import KnitRepoFetcher
1547
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1548
self.source, self.source._format, self.target, self.target._format)
1549
# TODO: jam 20070210 This should be an assert, not a translate
1550
revision_id = osutils.safe_revision_id(revision_id)
1551
f = KnitRepoFetcher(to_repository=self.target,
1552
from_repository=self.source,
1553
last_revision=revision_id,
1555
return f.count_copied, f.failed_revisions
1558
def missing_revision_ids(self, revision_id=None):
1559
"""See InterRepository.missing_revision_ids()."""
1560
if revision_id is not None:
1561
source_ids = self.source.get_ancestry(revision_id)
1562
assert source_ids[0] is None
1565
source_ids = self.source._all_possible_ids()
1566
source_ids_set = set(source_ids)
1567
# source_ids is the worst possible case we may need to pull.
1568
# now we want to filter source_ids against what we actually
1569
# have in target, but don't try to check for existence where we know
1570
# we do not have a revision as that would be pointless.
1571
target_ids = set(self.target._all_possible_ids())
1572
possibly_present_revisions = target_ids.intersection(source_ids_set)
1573
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1574
required_revisions = source_ids_set.difference(actually_present_revisions)
1575
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1576
if revision_id is not None:
1577
# we used get_ancestry to determine source_ids then we are assured all
1578
# revisions referenced are present as they are installed in topological order.
1579
# and the tip revision was validated by get_ancestry.
1580
return required_topo_revisions
1582
# if we just grabbed the possibly available ids, then
1583
# we only have an estimate of whats available and need to validate
1584
# that against the revision records.
1585
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1588
class InterModel1and2(InterRepository):
1591
def _get_repo_format_to_test(self):
1595
def is_compatible(source, target):
1596
if not isinstance(source, Repository):
1598
if not isinstance(target, Repository):
1600
if not source._format.rich_root_data and target._format.rich_root_data:
1606
def fetch(self, revision_id=None, pb=None):
1607
"""See InterRepository.fetch()."""
1608
from bzrlib.fetch import Model1toKnit2Fetcher
1609
# TODO: jam 20070210 This should be an assert, not a translate
1610
revision_id = osutils.safe_revision_id(revision_id)
1611
f = Model1toKnit2Fetcher(to_repository=self.target,
1612
from_repository=self.source,
1613
last_revision=revision_id,
1615
return f.count_copied, f.failed_revisions
1618
def copy_content(self, revision_id=None, basis=None):
1619
"""Make a complete copy of the content in self into destination.
1621
This is a destructive operation! Do not use it on existing
1624
:param revision_id: Only copy the content needed to construct
1625
revision_id and its parents.
1626
:param basis: Copy the needed data preferentially from basis.
1629
self.target.set_make_working_trees(self.source.make_working_trees())
1630
except NotImplementedError:
1632
# TODO: jam 20070210 Internal, assert, don't translate
1633
revision_id = osutils.safe_revision_id(revision_id)
1634
# grab the basis available data
1635
if basis is not None:
1636
self.target.fetch(basis, revision_id=revision_id)
1637
# but don't bother fetching if we have the needed data now.
1638
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1639
self.target.has_revision(revision_id)):
1641
self.target.fetch(self.source, revision_id=revision_id)
1644
class InterKnit1and2(InterKnitRepo):
1647
def _get_repo_format_to_test(self):
1651
def is_compatible(source, target):
1652
"""Be compatible with Knit1 source and Knit2 target"""
1653
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit2
1655
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1656
RepositoryFormatKnit2
1657
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1658
isinstance(target._format, (RepositoryFormatKnit2)))
1659
except AttributeError:
1663
def fetch(self, revision_id=None, pb=None):
1664
"""See InterRepository.fetch()."""
1665
from bzrlib.fetch import Knit1to2Fetcher
1666
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1667
self.source, self.source._format, self.target,
1668
self.target._format)
1669
# TODO: jam 20070210 This should be an assert, not a translate
1670
revision_id = osutils.safe_revision_id(revision_id)
1671
f = Knit1to2Fetcher(to_repository=self.target,
1672
from_repository=self.source,
1673
last_revision=revision_id,
1675
return f.count_copied, f.failed_revisions
1678
InterRepository.register_optimiser(InterSameDataRepository)
1679
InterRepository.register_optimiser(InterWeaveRepo)
1680
InterRepository.register_optimiser(InterKnitRepo)
1681
InterRepository.register_optimiser(InterModel1and2)
1682
InterRepository.register_optimiser(InterKnit1and2)
1685
class RepositoryTestProviderAdapter(object):
1686
"""A tool to generate a suite testing multiple repository formats at once.
1688
This is done by copying the test once for each transport and injecting
1689
the transport_server, transport_readonly_server, and bzrdir_format and
1690
repository_format classes into each copy. Each copy is also given a new id()
1691
to make it easy to identify.
1694
def __init__(self, transport_server, transport_readonly_server, formats):
1695
self._transport_server = transport_server
1696
self._transport_readonly_server = transport_readonly_server
1697
self._formats = formats
1699
def adapt(self, test):
1700
result = unittest.TestSuite()
1701
for repository_format, bzrdir_format in self._formats:
1702
from copy import deepcopy
1703
new_test = deepcopy(test)
1704
new_test.transport_server = self._transport_server
1705
new_test.transport_readonly_server = self._transport_readonly_server
1706
new_test.bzrdir_format = bzrdir_format
1707
new_test.repository_format = repository_format
1708
def make_new_test_id():
1709
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1710
return lambda: new_id
1711
new_test.id = make_new_test_id()
1712
result.addTest(new_test)
1716
class InterRepositoryTestProviderAdapter(object):
1717
"""A tool to generate a suite testing multiple inter repository formats.
1719
This is done by copying the test once for each interrepo provider and injecting
1720
the transport_server, transport_readonly_server, repository_format and
1721
repository_to_format classes into each copy.
1722
Each copy is also given a new id() to make it easy to identify.
1725
def __init__(self, transport_server, transport_readonly_server, formats):
1726
self._transport_server = transport_server
1727
self._transport_readonly_server = transport_readonly_server
1728
self._formats = formats
1730
def adapt(self, test):
1731
result = unittest.TestSuite()
1732
for interrepo_class, repository_format, repository_format_to in self._formats:
1733
from copy import deepcopy
1734
new_test = deepcopy(test)
1735
new_test.transport_server = self._transport_server
1736
new_test.transport_readonly_server = self._transport_readonly_server
1737
new_test.interrepo_class = interrepo_class
1738
new_test.repository_format = repository_format
1739
new_test.repository_format_to = repository_format_to
1740
def make_new_test_id():
1741
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1742
return lambda: new_id
1743
new_test.id = make_new_test_id()
1744
result.addTest(new_test)
1748
def default_test_list():
1749
"""Generate the default list of interrepo permutations to test."""
1750
from bzrlib.repofmt import knitrepo, weaverepo
1752
# test the default InterRepository between format 6 and the current
1754
# XXX: robertc 20060220 reinstate this when there are two supported
1755
# formats which do not have an optimal code path between them.
1756
#result.append((InterRepository,
1757
# RepositoryFormat6(),
1758
# RepositoryFormatKnit1()))
1759
for optimiser_class in InterRepository._optimisers:
1760
format_to_test = optimiser_class._get_repo_format_to_test()
1761
if format_to_test is not None:
1762
result.append((optimiser_class,
1763
format_to_test, format_to_test))
1764
# if there are specific combinations we want to use, we can add them
1766
result.append((InterModel1and2,
1767
weaverepo.RepositoryFormat5(),
1768
knitrepo.RepositoryFormatKnit2()))
1769
result.append((InterKnit1and2,
1770
knitrepo.RepositoryFormatKnit1(),
1771
knitrepo.RepositoryFormatKnit2()))
1775
class CopyConverter(object):
1776
"""A repository conversion tool which just performs a copy of the content.
1778
This is slow but quite reliable.
1781
def __init__(self, target_format):
1782
"""Create a CopyConverter.
1784
:param target_format: The format the resulting repository should be.
1786
self.target_format = target_format
1788
def convert(self, repo, pb):
1789
"""Perform the conversion of to_convert, giving feedback via pb.
1791
:param to_convert: The disk object to convert.
1792
:param pb: a progress bar to use for progress information.
1797
# this is only useful with metadir layouts - separated repo content.
1798
# trigger an assertion if not such
1799
repo._format.get_format_string()
1800
self.repo_dir = repo.bzrdir
1801
self.step('Moving repository to repository.backup')
1802
self.repo_dir.transport.move('repository', 'repository.backup')
1803
backup_transport = self.repo_dir.transport.clone('repository.backup')
1804
repo._format.check_conversion_target(self.target_format)
1805
self.source_repo = repo._format.open(self.repo_dir,
1807
_override_transport=backup_transport)
1808
self.step('Creating new repository')
1809
converted = self.target_format.initialize(self.repo_dir,
1810
self.source_repo.is_shared())
1811
converted.lock_write()
1813
self.step('Copying content into repository.')
1814
self.source_repo.copy_content_into(converted)
1817
self.step('Deleting old repository content.')
1818
self.repo_dir.transport.delete_tree('repository.backup')
1819
self.pb.note('repository converted')
1821
def step(self, message):
1822
"""Update the pb by a step."""
1824
self.pb.update(message, self.count, self.total)
1827
class CommitBuilder(object):
1828
"""Provides an interface to build up a commit.
1830
This allows describing a tree to be committed without needing to
1831
know the internals of the format of the repository.
1834
record_root_entry = False
1835
def __init__(self, repository, parents, config, timestamp=None,
1836
timezone=None, committer=None, revprops=None,
1838
"""Initiate a CommitBuilder.
1840
:param repository: Repository to commit to.
1841
:param parents: Revision ids of the parents of the new revision.
1842
:param config: Configuration to use.
1843
:param timestamp: Optional timestamp recorded for commit.
1844
:param timezone: Optional timezone for timestamp.
1845
:param committer: Optional committer to set for commit.
1846
:param revprops: Optional dictionary of revision properties.
1847
:param revision_id: Optional revision id.
1849
self._config = config
1851
if committer is None:
1852
self._committer = self._config.username()
1854
assert isinstance(committer, basestring), type(committer)
1855
self._committer = committer
1857
self.new_inventory = Inventory(None)
1858
self._new_revision_id = osutils.safe_revision_id(revision_id)
1859
self.parents = parents
1860
self.repository = repository
1863
if revprops is not None:
1864
self._revprops.update(revprops)
1866
if timestamp is None:
1867
timestamp = time.time()
1868
# Restrict resolution to 1ms
1869
self._timestamp = round(timestamp, 3)
1871
if timezone is None:
1872
self._timezone = osutils.local_time_offset()
1874
self._timezone = int(timezone)
1876
self._generate_revision_if_needed()
1878
def commit(self, message):
1879
"""Make the actual commit.
1881
:return: The revision id of the recorded revision.
1883
rev = _mod_revision.Revision(
1884
timestamp=self._timestamp,
1885
timezone=self._timezone,
1886
committer=self._committer,
1888
inventory_sha1=self.inv_sha1,
1889
revision_id=self._new_revision_id,
1890
properties=self._revprops)
1891
rev.parent_ids = self.parents
1892
self.repository.add_revision(self._new_revision_id, rev,
1893
self.new_inventory, self._config)
1894
return self._new_revision_id
1896
def revision_tree(self):
1897
"""Return the tree that was just committed.
1899
After calling commit() this can be called to get a RevisionTree
1900
representing the newly committed tree. This is preferred to
1901
calling Repository.revision_tree() because that may require
1902
deserializing the inventory, while we already have a copy in
1905
return RevisionTree(self.repository, self.new_inventory,
1906
self._new_revision_id)
1908
def finish_inventory(self):
1909
"""Tell the builder that the inventory is finished."""
1910
if self.new_inventory.root is None:
1911
symbol_versioning.warn('Root entry should be supplied to'
1912
' record_entry_contents, as of bzr 0.10.',
1913
DeprecationWarning, stacklevel=2)
1914
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
1915
self.new_inventory.revision_id = self._new_revision_id
1916
self.inv_sha1 = self.repository.add_inventory(
1917
self._new_revision_id,
1922
def _gen_revision_id(self):
1923
"""Return new revision-id."""
1924
return generate_ids.gen_revision_id(self._config.username(),
1927
def _generate_revision_if_needed(self):
1928
"""Create a revision id if None was supplied.
1930
If the repository can not support user-specified revision ids
1931
they should override this function and raise CannotSetRevisionId
1932
if _new_revision_id is not None.
1934
:raises: CannotSetRevisionId
1936
if self._new_revision_id is None:
1937
self._new_revision_id = self._gen_revision_id()
1939
def record_entry_contents(self, ie, parent_invs, path, tree):
1940
"""Record the content of ie from tree into the commit if needed.
1942
Side effect: sets ie.revision when unchanged
1944
:param ie: An inventory entry present in the commit.
1945
:param parent_invs: The inventories of the parent revisions of the
1947
:param path: The path the entry is at in the tree.
1948
:param tree: The tree which contains this entry and should be used to
1951
if self.new_inventory.root is None and ie.parent_id is not None:
1952
symbol_versioning.warn('Root entry should be supplied to'
1953
' record_entry_contents, as of bzr 0.10.',
1954
DeprecationWarning, stacklevel=2)
1955
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
1957
self.new_inventory.add(ie)
1959
# ie.revision is always None if the InventoryEntry is considered
1960
# for committing. ie.snapshot will record the correct revision
1961
# which may be the sole parent if it is untouched.
1962
if ie.revision is not None:
1965
# In this revision format, root entries have no knit or weave
1966
if ie is self.new_inventory.root:
1967
# When serializing out to disk and back in
1968
# root.revision is always _new_revision_id
1969
ie.revision = self._new_revision_id
1971
previous_entries = ie.find_previous_heads(
1973
self.repository.weave_store,
1974
self.repository.get_transaction())
1975
# we are creating a new revision for ie in the history store
1977
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
1979
def modified_directory(self, file_id, file_parents):
1980
"""Record the presence of a symbolic link.
1982
:param file_id: The file_id of the link to record.
1983
:param file_parents: The per-file parent revision ids.
1985
self._add_text_to_weave(file_id, [], file_parents.keys())
1987
def modified_file_text(self, file_id, file_parents,
1988
get_content_byte_lines, text_sha1=None,
1990
"""Record the text of file file_id
1992
:param file_id: The file_id of the file to record the text of.
1993
:param file_parents: The per-file parent revision ids.
1994
:param get_content_byte_lines: A callable which will return the byte
1996
:param text_sha1: Optional SHA1 of the file contents.
1997
:param text_size: Optional size of the file contents.
1999
# mutter('storing text of file {%s} in revision {%s} into %r',
2000
# file_id, self._new_revision_id, self.repository.weave_store)
2001
# special case to avoid diffing on renames or
2003
if (len(file_parents) == 1
2004
and text_sha1 == file_parents.values()[0].text_sha1
2005
and text_size == file_parents.values()[0].text_size):
2006
previous_ie = file_parents.values()[0]
2007
versionedfile = self.repository.weave_store.get_weave(file_id,
2008
self.repository.get_transaction())
2009
versionedfile.clone_text(self._new_revision_id,
2010
previous_ie.revision, file_parents.keys())
2011
return text_sha1, text_size
2013
new_lines = get_content_byte_lines()
2014
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2015
# should return the SHA1 and size
2016
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2017
return osutils.sha_strings(new_lines), \
2018
sum(map(len, new_lines))
2020
def modified_link(self, file_id, file_parents, link_target):
2021
"""Record the presence of a symbolic link.
2023
:param file_id: The file_id of the link to record.
2024
:param file_parents: The per-file parent revision ids.
2025
:param link_target: Target location of this link.
2027
self._add_text_to_weave(file_id, [], file_parents.keys())
2029
def _add_text_to_weave(self, file_id, new_lines, parents):
2030
versionedfile = self.repository.weave_store.get_weave_or_empty(
2031
file_id, self.repository.get_transaction())
2032
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2033
versionedfile.clear_cache()
2036
class _CommitBuilder(CommitBuilder):
2037
"""Temporary class so old CommitBuilders are detected properly
2039
Note: CommitBuilder works whether or not root entry is recorded.
2042
record_root_entry = True
2045
class RootCommitBuilder(CommitBuilder):
2046
"""This commitbuilder actually records the root id"""
2048
record_root_entry = True
2050
def record_entry_contents(self, ie, parent_invs, path, tree):
2051
"""Record the content of ie from tree into the commit if needed.
2053
Side effect: sets ie.revision when unchanged
2055
:param ie: An inventory entry present in the commit.
2056
:param parent_invs: The inventories of the parent revisions of the
2058
:param path: The path the entry is at in the tree.
2059
:param tree: The tree which contains this entry and should be used to
2062
assert self.new_inventory.root is not None or ie.parent_id is None
2063
self.new_inventory.add(ie)
2065
# ie.revision is always None if the InventoryEntry is considered
2066
# for committing. ie.snapshot will record the correct revision
2067
# which may be the sole parent if it is untouched.
2068
if ie.revision is not None:
2071
previous_entries = ie.find_previous_heads(
2073
self.repository.weave_store,
2074
self.repository.get_transaction())
2075
# we are creating a new revision for ie in the history store
2077
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2089
def _unescaper(match, _map=_unescape_map):
2090
code = match.group(1)
2094
if not code.startswith('#'):
2096
return unichr(int(code[1:])).encode('utf8')
2102
def _unescape_xml(data):
2103
"""Unescape predefined XML entities in a string of data."""
2105
if _unescape_re is None:
2106
_unescape_re = re.compile('\&([^;]*);')
2107
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
