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, revid, inv, parents):
82
"""Add the inventory inv to the repository as revid.
84
:param parents: The revision ids of the parents that revid
85
is known to have and are in the repository already.
87
returns the sha1 of the serialized inventory.
89
_mod_revision.check_not_reserved_id(revid)
90
assert inv.revision_id is None or inv.revision_id == revid, \
91
"Mismatch between inventory revision" \
92
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
93
assert inv.root is not None
94
inv_text = self.serialise_inventory(inv)
95
inv_sha1 = osutils.sha_string(inv_text)
96
inv_vf = self.control_weaves.get_weave('inventory',
97
self.get_transaction())
98
self._inventory_add_lines(inv_vf, revid, parents, osutils.split_lines(inv_text))
101
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
103
for parent in parents:
105
final_parents.append(parent)
107
inv_vf.add_lines(revid, final_parents, lines)
110
def add_revision(self, rev_id, rev, inv=None, config=None):
111
"""Add rev to the revision store as rev_id.
113
:param rev_id: the revision id to use.
114
:param rev: The revision object.
115
:param inv: The inventory for the revision. if None, it will be looked
116
up in the inventory storer
117
:param config: If None no digital signature will be created.
118
If supplied its signature_needed method will be used
119
to determine if a signature should be made.
121
_mod_revision.check_not_reserved_id(rev_id)
122
if config is not None and config.signature_needed():
124
inv = self.get_inventory(rev_id)
125
plaintext = Testament(rev, inv).as_short_text()
126
self.store_revision_signature(
127
gpg.GPGStrategy(config), plaintext, rev_id)
128
if not rev_id in self.get_inventory_weave():
130
raise errors.WeaveRevisionNotPresent(rev_id,
131
self.get_inventory_weave())
133
# yes, this is not suitable for adding with ghosts.
134
self.add_inventory(rev_id, inv, rev.parent_ids)
135
self._revision_store.add_revision(rev, self.get_transaction())
138
def _all_possible_ids(self):
139
"""Return all the possible revisions that we could find."""
140
return self.get_inventory_weave().versions()
142
def all_revision_ids(self):
143
"""Returns a list of all the revision ids in the repository.
145
This is deprecated because code should generally work on the graph
146
reachable from a particular revision, and ignore any other revisions
147
that might be present. There is no direct replacement method.
149
return self._all_revision_ids()
152
def _all_revision_ids(self):
153
"""Returns a list of all the revision ids in the repository.
155
These are in as much topological order as the underlying store can
156
present: for weaves ghosts may lead to a lack of correctness until
157
the reweave updates the parents list.
159
if self._revision_store.text_store.listable():
160
return self._revision_store.all_revision_ids(self.get_transaction())
161
result = self._all_possible_ids()
162
return self._eliminate_revisions_not_present(result)
164
def break_lock(self):
165
"""Break a lock if one is present from another instance.
167
Uses the ui factory to ask for confirmation if the lock may be from
170
self.control_files.break_lock()
173
def _eliminate_revisions_not_present(self, revision_ids):
174
"""Check every revision id in revision_ids to see if we have it.
176
Returns a set of the present revisions.
179
for id in revision_ids:
180
if self.has_revision(id):
185
def create(a_bzrdir):
186
"""Construct the current default format repository in a_bzrdir."""
187
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
189
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
190
"""instantiate a Repository.
192
:param _format: The format of the repository on disk.
193
:param a_bzrdir: The BzrDir of the repository.
195
In the future we will have a single api for all stores for
196
getting file texts, inventories and revisions, then
197
this construct will accept instances of those things.
199
super(Repository, self).__init__()
200
self._format = _format
201
# the following are part of the public API for Repository:
202
self.bzrdir = a_bzrdir
203
self.control_files = control_files
204
self._revision_store = _revision_store
205
self.text_store = text_store
206
# backwards compatibility
207
self.weave_store = text_store
208
# not right yet - should be more semantically clear ?
210
self.control_store = control_store
211
self.control_weaves = control_store
212
# TODO: make sure to construct the right store classes, etc, depending
213
# on whether escaping is required.
214
self._warn_if_deprecated()
217
return '%s(%r)' % (self.__class__.__name__,
218
self.bzrdir.transport.base)
221
return self.control_files.is_locked()
223
def lock_write(self):
224
self.control_files.lock_write()
227
self.control_files.lock_read()
229
def get_physical_lock_status(self):
230
return self.control_files.get_physical_lock_status()
233
def gather_stats(self, revid=None, committers=None):
234
"""Gather statistics from a revision id.
236
:param revid: The revision id to gather statistics from, if None, then
237
no revision specific statistics are gathered.
238
:param committers: Optional parameter controlling whether to grab
239
a count of committers from the revision specific statistics.
240
:return: A dictionary of statistics. Currently this contains:
241
committers: The number of committers if requested.
242
firstrev: A tuple with timestamp, timezone for the penultimate left
243
most ancestor of revid, if revid is not the NULL_REVISION.
244
latestrev: A tuple with timestamp, timezone for revid, if revid is
245
not the NULL_REVISION.
246
revisions: The total revision count in the repository.
247
size: An estimate disk size of the repository in bytes.
250
if revid and committers:
251
result['committers'] = 0
252
if revid and revid != _mod_revision.NULL_REVISION:
254
all_committers = set()
255
revisions = self.get_ancestry(revid)
256
# pop the leading None
258
first_revision = None
260
# ignore the revisions in the middle - just grab first and last
261
revisions = revisions[0], revisions[-1]
262
for revision in self.get_revisions(revisions):
263
if not first_revision:
264
first_revision = revision
266
all_committers.add(revision.committer)
267
last_revision = revision
269
result['committers'] = len(all_committers)
270
result['firstrev'] = (first_revision.timestamp,
271
first_revision.timezone)
272
result['latestrev'] = (last_revision.timestamp,
273
last_revision.timezone)
275
# now gather global repository information
276
if self.bzrdir.root_transport.listable():
277
c, t = self._revision_store.total_size(self.get_transaction())
278
result['revisions'] = c
283
def missing_revision_ids(self, other, revision_id=None):
284
"""Return the revision ids that other has that this does not.
286
These are returned in topological order.
288
revision_id: only return revision ids included by revision_id.
290
return InterRepository.get(other, self).missing_revision_ids(revision_id)
294
"""Open the repository rooted at base.
296
For instance, if the repository is at URL/.bzr/repository,
297
Repository.open(URL) -> a Repository instance.
299
control = bzrdir.BzrDir.open(base)
300
return control.open_repository()
302
def copy_content_into(self, destination, revision_id=None, basis=None):
303
"""Make a complete copy of the content in self into destination.
305
This is a destructive operation! Do not use it on existing
308
return InterRepository.get(self, destination).copy_content(revision_id, basis)
310
def fetch(self, source, revision_id=None, pb=None):
311
"""Fetch the content required to construct revision_id from source.
313
If revision_id is None all content is copied.
315
return InterRepository.get(source, self).fetch(revision_id=revision_id,
318
def get_commit_builder(self, branch, parents, config, timestamp=None,
319
timezone=None, committer=None, revprops=None,
321
"""Obtain a CommitBuilder for this repository.
323
:param branch: Branch to commit to.
324
:param parents: Revision ids of the parents of the new revision.
325
:param config: Configuration to use.
326
:param timestamp: Optional timestamp recorded for commit.
327
:param timezone: Optional timezone for timestamp.
328
:param committer: Optional committer to set for commit.
329
:param revprops: Optional dictionary of revision properties.
330
:param revision_id: Optional revision id.
332
return _CommitBuilder(self, parents, config, timestamp, timezone,
333
committer, revprops, revision_id)
336
self.control_files.unlock()
339
def clone(self, a_bzrdir, revision_id=None, basis=None):
340
"""Clone this repository into a_bzrdir using the current format.
342
Currently no check is made that the format of this repository and
343
the bzrdir format are compatible. FIXME RBC 20060201.
345
:return: The newly created destination repository.
347
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
348
# use target default format.
349
dest_repo = a_bzrdir.create_repository()
351
# Most control formats need the repository to be specifically
352
# created, but on some old all-in-one formats it's not needed
354
dest_repo = self._format.initialize(a_bzrdir, shared=self.is_shared())
355
except errors.UninitializableFormat:
356
dest_repo = a_bzrdir.open_repository()
357
self.copy_content_into(dest_repo, revision_id, basis)
361
def has_revision(self, revision_id):
362
"""True if this repository has a copy of the revision."""
363
return self._revision_store.has_revision_id(revision_id,
364
self.get_transaction())
367
def get_revision_reconcile(self, revision_id):
368
"""'reconcile' helper routine that allows access to a revision always.
370
This variant of get_revision does not cross check the weave graph
371
against the revision one as get_revision does: but it should only
372
be used by reconcile, or reconcile-alike commands that are correcting
373
or testing the revision graph.
375
if not revision_id or not isinstance(revision_id, basestring):
376
raise errors.InvalidRevisionId(revision_id=revision_id,
378
return self._revision_store.get_revisions([revision_id],
379
self.get_transaction())[0]
381
def get_revisions(self, revision_ids):
382
return self._revision_store.get_revisions(revision_ids,
383
self.get_transaction())
386
def get_revision_xml(self, revision_id):
387
rev = self.get_revision(revision_id)
389
# the current serializer..
390
self._revision_store._serializer.write_revision(rev, rev_tmp)
392
return rev_tmp.getvalue()
395
def get_revision(self, revision_id):
396
"""Return the Revision object for a named revision"""
397
r = self.get_revision_reconcile(revision_id)
398
# weave corruption can lead to absent revision markers that should be
400
# the following test is reasonably cheap (it needs a single weave read)
401
# and the weave is cached in read transactions. In write transactions
402
# it is not cached but typically we only read a small number of
403
# revisions. For knits when they are introduced we will probably want
404
# to ensure that caching write transactions are in use.
405
inv = self.get_inventory_weave()
406
self._check_revision_parents(r, inv)
410
def get_deltas_for_revisions(self, revisions):
411
"""Produce a generator of revision deltas.
413
Note that the input is a sequence of REVISIONS, not revision_ids.
414
Trees will be held in memory until the generator exits.
415
Each delta is relative to the revision's lefthand predecessor.
417
required_trees = set()
418
for revision in revisions:
419
required_trees.add(revision.revision_id)
420
required_trees.update(revision.parent_ids[:1])
421
trees = dict((t.get_revision_id(), t) for
422
t in self.revision_trees(required_trees))
423
for revision in revisions:
424
if not revision.parent_ids:
425
old_tree = self.revision_tree(None)
427
old_tree = trees[revision.parent_ids[0]]
428
yield trees[revision.revision_id].changes_from(old_tree)
431
def get_revision_delta(self, revision_id):
432
"""Return the delta for one revision.
434
The delta is relative to the left-hand predecessor of the
437
r = self.get_revision(revision_id)
438
return list(self.get_deltas_for_revisions([r]))[0]
440
def _check_revision_parents(self, revision, inventory):
441
"""Private to Repository and Fetch.
443
This checks the parentage of revision in an inventory weave for
444
consistency and is only applicable to inventory-weave-for-ancestry
445
using repository formats & fetchers.
447
weave_parents = inventory.get_parents(revision.revision_id)
448
weave_names = inventory.versions()
449
for parent_id in revision.parent_ids:
450
if parent_id in weave_names:
451
# this parent must not be a ghost.
452
if not parent_id in weave_parents:
454
raise errors.CorruptRepository(self)
457
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
458
signature = gpg_strategy.sign(plaintext)
459
self._revision_store.add_revision_signature_text(revision_id,
461
self.get_transaction())
463
def fileids_altered_by_revision_ids(self, revision_ids):
464
"""Find the file ids and versions affected by revisions.
466
:param revisions: an iterable containing revision ids.
467
:return: a dictionary mapping altered file-ids to an iterable of
468
revision_ids. Each altered file-ids has the exact revision_ids that
469
altered it listed explicitly.
471
assert self._serializer.support_altered_by_hack, \
472
("fileids_altered_by_revision_ids only supported for branches "
473
"which store inventory as unnested xml, not on %r" % self)
474
selected_revision_ids = set(revision_ids)
475
w = self.get_inventory_weave()
478
# this code needs to read every new line in every inventory for the
479
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
480
# not present in one of those inventories is unnecessary but not
481
# harmful because we are filtering by the revision id marker in the
482
# inventory lines : we only select file ids altered in one of those
483
# revisions. We don't need to see all lines in the inventory because
484
# only those added in an inventory in rev X can contain a revision=X
486
unescape_revid_cache = {}
487
unescape_fileid_cache = {}
489
# jam 20061218 In a big fetch, this handles hundreds of thousands
490
# of lines, so it has had a lot of inlining and optimizing done.
491
# Sorry that it is a little bit messy.
492
# Move several functions to be local variables, since this is a long
494
search = self._file_ids_altered_regex.search
495
unescape = _unescape_xml
496
setdefault = result.setdefault
497
pb = ui.ui_factory.nested_progress_bar()
499
for line in w.iter_lines_added_or_present_in_versions(
500
selected_revision_ids, pb=pb):
504
# One call to match.group() returning multiple items is quite a
505
# bit faster than 2 calls to match.group() each returning 1
506
file_id, revision_id = match.group('file_id', 'revision_id')
508
# Inlining the cache lookups helps a lot when you make 170,000
509
# lines and 350k ids, versus 8.4 unique ids.
510
# Using a cache helps in 2 ways:
511
# 1) Avoids unnecessary decoding calls
512
# 2) Re-uses cached strings, which helps in future set and
514
# (2) is enough that removing encoding entirely along with
515
# the cache (so we are using plain strings) results in no
516
# performance improvement.
518
revision_id = unescape_revid_cache[revision_id]
520
unescaped = unescape(revision_id)
521
unescape_revid_cache[revision_id] = unescaped
522
revision_id = unescaped
524
if revision_id in selected_revision_ids:
526
file_id = unescape_fileid_cache[file_id]
528
unescaped = unescape(file_id)
529
unescape_fileid_cache[file_id] = unescaped
531
setdefault(file_id, set()).add(revision_id)
537
def get_inventory_weave(self):
538
return self.control_weaves.get_weave('inventory',
539
self.get_transaction())
542
def get_inventory(self, revision_id):
543
"""Get Inventory object by hash."""
544
return self.deserialise_inventory(
545
revision_id, self.get_inventory_xml(revision_id))
547
def deserialise_inventory(self, revision_id, xml):
548
"""Transform the xml into an inventory object.
550
:param revision_id: The expected revision id of the inventory.
551
:param xml: A serialised inventory.
553
result = self._serializer.read_inventory_from_string(xml)
554
result.root.revision = revision_id
557
def serialise_inventory(self, inv):
558
return self._serializer.write_inventory_to_string(inv)
561
def get_inventory_xml(self, revision_id):
562
"""Get inventory XML as a file object."""
564
assert isinstance(revision_id, basestring), type(revision_id)
565
iw = self.get_inventory_weave()
566
return iw.get_text(revision_id)
568
raise errors.HistoryMissing(self, 'inventory', revision_id)
571
def get_inventory_sha1(self, revision_id):
572
"""Return the sha1 hash of the inventory entry
574
return self.get_revision(revision_id).inventory_sha1
577
def get_revision_graph(self, revision_id=None):
578
"""Return a dictionary containing the revision graph.
580
:param revision_id: The revision_id to get a graph from. If None, then
581
the entire revision graph is returned. This is a deprecated mode of
582
operation and will be removed in the future.
583
:return: a dictionary of revision_id->revision_parents_list.
585
# special case NULL_REVISION
586
if revision_id == _mod_revision.NULL_REVISION:
588
a_weave = self.get_inventory_weave()
589
all_revisions = self._eliminate_revisions_not_present(
591
entire_graph = dict([(node, a_weave.get_parents(node)) for
592
node in all_revisions])
593
if revision_id is None:
595
elif revision_id not in entire_graph:
596
raise errors.NoSuchRevision(self, revision_id)
598
# add what can be reached from revision_id
600
pending = set([revision_id])
601
while len(pending) > 0:
603
result[node] = entire_graph[node]
604
for revision_id in result[node]:
605
if revision_id not in result:
606
pending.add(revision_id)
610
def get_revision_graph_with_ghosts(self, revision_ids=None):
611
"""Return a graph of the revisions with ghosts marked as applicable.
613
:param revision_ids: an iterable of revisions to graph or None for all.
614
:return: a Graph object with the graph reachable from revision_ids.
616
result = graph.Graph()
618
pending = set(self.all_revision_ids())
621
pending = set(revision_ids)
622
# special case NULL_REVISION
623
if _mod_revision.NULL_REVISION in pending:
624
pending.remove(_mod_revision.NULL_REVISION)
625
required = set(pending)
628
revision_id = pending.pop()
630
rev = self.get_revision(revision_id)
631
except errors.NoSuchRevision:
632
if revision_id in required:
635
result.add_ghost(revision_id)
637
for parent_id in rev.parent_ids:
638
# is this queued or done ?
639
if (parent_id not in pending and
640
parent_id not in done):
642
pending.add(parent_id)
643
result.add_node(revision_id, rev.parent_ids)
644
done.add(revision_id)
648
def get_revision_inventory(self, revision_id):
649
"""Return inventory of a past revision."""
650
# TODO: Unify this with get_inventory()
651
# bzr 0.0.6 and later imposes the constraint that the inventory_id
652
# must be the same as its revision, so this is trivial.
653
if revision_id is None:
654
# This does not make sense: if there is no revision,
655
# then it is the current tree inventory surely ?!
656
# and thus get_root_id() is something that looks at the last
657
# commit on the branch, and the get_root_id is an inventory check.
658
raise NotImplementedError
659
# return Inventory(self.get_root_id())
661
return self.get_inventory(revision_id)
665
"""Return True if this repository is flagged as a shared repository."""
666
raise NotImplementedError(self.is_shared)
669
def reconcile(self, other=None, thorough=False):
670
"""Reconcile this repository."""
671
from bzrlib.reconcile import RepoReconciler
672
reconciler = RepoReconciler(self, thorough=thorough)
673
reconciler.reconcile()
677
def revision_tree(self, revision_id):
678
"""Return Tree for a revision on this branch.
680
`revision_id` may be None for the empty tree revision.
682
# TODO: refactor this to use an existing revision object
683
# so we don't need to read it in twice.
684
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
685
return RevisionTree(self, Inventory(root_id=None),
686
_mod_revision.NULL_REVISION)
688
inv = self.get_revision_inventory(revision_id)
689
return RevisionTree(self, inv, revision_id)
692
def revision_trees(self, revision_ids):
693
"""Return Tree for a revision on this branch.
695
`revision_id` may not be None or 'null:'"""
696
assert None not in revision_ids
697
assert _mod_revision.NULL_REVISION not in revision_ids
698
texts = self.get_inventory_weave().get_texts(revision_ids)
699
for text, revision_id in zip(texts, revision_ids):
700
inv = self.deserialise_inventory(revision_id, text)
701
yield RevisionTree(self, inv, revision_id)
704
def get_ancestry(self, revision_id):
705
"""Return a list of revision-ids integrated by a revision.
707
The first element of the list is always None, indicating the origin
708
revision. This might change when we have history horizons, or
709
perhaps we should have a new API.
711
This is topologically sorted.
713
if revision_id is None:
715
if not self.has_revision(revision_id):
716
raise errors.NoSuchRevision(self, revision_id)
717
w = self.get_inventory_weave()
718
candidates = w.get_ancestry(revision_id)
719
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
722
def print_file(self, file, revision_id):
723
"""Print `file` to stdout.
725
FIXME RBC 20060125 as John Meinel points out this is a bad api
726
- it writes to stdout, it assumes that that is valid etc. Fix
727
by creating a new more flexible convenience function.
729
tree = self.revision_tree(revision_id)
730
# use inventory as it was in that revision
731
file_id = tree.inventory.path2id(file)
733
# TODO: jam 20060427 Write a test for this code path
734
# it had a bug in it, and was raising the wrong
736
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
737
tree.print_file(file_id)
739
def get_transaction(self):
740
return self.control_files.get_transaction()
742
def revision_parents(self, revid):
743
return self.get_inventory_weave().parent_names(revid)
746
def set_make_working_trees(self, new_value):
747
"""Set the policy flag for making working trees when creating branches.
749
This only applies to branches that use this repository.
751
The default is 'True'.
752
:param new_value: True to restore the default, False to disable making
755
raise NotImplementedError(self.set_make_working_trees)
757
def make_working_trees(self):
758
"""Returns the policy for making working trees on new branches."""
759
raise NotImplementedError(self.make_working_trees)
762
def sign_revision(self, revision_id, gpg_strategy):
763
plaintext = Testament.from_revision(self, revision_id).as_short_text()
764
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
767
def has_signature_for_revision_id(self, revision_id):
768
"""Query for a revision signature for revision_id in the repository."""
769
return self._revision_store.has_signature(revision_id,
770
self.get_transaction())
773
def get_signature_text(self, revision_id):
774
"""Return the text for a signature."""
775
return self._revision_store.get_signature_text(revision_id,
776
self.get_transaction())
779
def check(self, revision_ids):
780
"""Check consistency of all history of given revision_ids.
782
Different repository implementations should override _check().
784
:param revision_ids: A non-empty list of revision_ids whose ancestry
785
will be checked. Typically the last revision_id of a branch.
788
raise ValueError("revision_ids must be non-empty in %s.check"
790
return self._check(revision_ids)
792
def _check(self, revision_ids):
793
result = check.Check(self)
797
def _warn_if_deprecated(self):
798
global _deprecation_warning_done
799
if _deprecation_warning_done:
801
_deprecation_warning_done = True
802
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
803
% (self._format, self.bzrdir.transport.base))
805
def supports_rich_root(self):
806
return self._format.rich_root_data
808
def _check_ascii_revisionid(self, revision_id, method):
809
"""Private helper for ascii-only repositories."""
810
# weave repositories refuse to store revisionids that are non-ascii.
811
if revision_id is not None:
812
# weaves require ascii revision ids.
813
if isinstance(revision_id, unicode):
815
revision_id.encode('ascii')
816
except UnicodeEncodeError:
817
raise errors.NonAsciiRevisionId(method, self)
821
# remove these delegates a while after bzr 0.15
822
def __make_delegated(name, from_module):
823
def _deprecated_repository_forwarder():
824
symbol_versioning.warn('%s moved to %s in bzr 0.15'
825
% (name, from_module),
827
m = __import__(from_module, globals(), locals(), [name])
829
return getattr(m, name)
830
except AttributeError:
831
raise AttributeError('module %s has no name %s'
833
globals()[name] = _deprecated_repository_forwarder
836
'AllInOneRepository',
837
'WeaveMetaDirRepository',
838
'PreSplitOutRepositoryFormat',
844
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
849
'RepositoryFormatKnit',
850
'RepositoryFormatKnit1',
851
'RepositoryFormatKnit2',
853
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
856
def install_revision(repository, rev, revision_tree):
857
"""Install all revision data into a repository."""
860
for p_id in rev.parent_ids:
861
if repository.has_revision(p_id):
862
present_parents.append(p_id)
863
parent_trees[p_id] = repository.revision_tree(p_id)
865
parent_trees[p_id] = repository.revision_tree(None)
867
inv = revision_tree.inventory
868
entries = inv.iter_entries()
869
# backwards compatability hack: skip the root id.
870
if not repository.supports_rich_root():
871
path, root = entries.next()
872
if root.revision != rev.revision_id:
873
raise errors.IncompatibleRevision(repr(repository))
874
# Add the texts that are not already present
875
for path, ie in entries:
876
w = repository.weave_store.get_weave_or_empty(ie.file_id,
877
repository.get_transaction())
878
if ie.revision not in w:
880
# FIXME: TODO: The following loop *may* be overlapping/duplicate
881
# with InventoryEntry.find_previous_heads(). if it is, then there
882
# is a latent bug here where the parents may have ancestors of each
884
for revision, tree in parent_trees.iteritems():
885
if ie.file_id not in tree:
887
parent_id = tree.inventory[ie.file_id].revision
888
if parent_id in text_parents:
890
text_parents.append(parent_id)
892
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
893
repository.get_transaction())
894
lines = revision_tree.get_file(ie.file_id).readlines()
895
vfile.add_lines(rev.revision_id, text_parents, lines)
897
# install the inventory
898
repository.add_inventory(rev.revision_id, inv, present_parents)
899
except errors.RevisionAlreadyPresent:
901
repository.add_revision(rev.revision_id, rev, inv)
904
class MetaDirRepository(Repository):
905
"""Repositories in the new meta-dir layout."""
907
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
908
super(MetaDirRepository, self).__init__(_format,
914
dir_mode = self.control_files._dir_mode
915
file_mode = self.control_files._file_mode
919
"""Return True if this repository is flagged as a shared repository."""
920
return self.control_files._transport.has('shared-storage')
923
def set_make_working_trees(self, new_value):
924
"""Set the policy flag for making working trees when creating branches.
926
This only applies to branches that use this repository.
928
The default is 'True'.
929
:param new_value: True to restore the default, False to disable making
934
self.control_files._transport.delete('no-working-trees')
935
except errors.NoSuchFile:
938
self.control_files.put_utf8('no-working-trees', '')
940
def make_working_trees(self):
941
"""Returns the policy for making working trees on new branches."""
942
return not self.control_files._transport.has('no-working-trees')
945
class RepositoryFormatRegistry(registry.Registry):
946
"""Registry of RepositoryFormats.
949
def get(self, format_string):
950
r = registry.Registry.get(self, format_string)
956
format_registry = RepositoryFormatRegistry()
957
"""Registry of formats, indexed by their identifying format string.
959
This can contain either format instances themselves, or classes/factories that
960
can be called to obtain one.
964
class RepositoryFormat(object):
965
"""A repository format.
967
Formats provide three things:
968
* An initialization routine to construct repository data on disk.
969
* a format string which is used when the BzrDir supports versioned
971
* an open routine which returns a Repository instance.
973
Formats are placed in an dict by their format string for reference
974
during opening. These should be subclasses of RepositoryFormat
977
Once a format is deprecated, just deprecate the initialize and open
978
methods on the format class. Do not deprecate the object, as the
979
object will be created every system load.
981
Common instance attributes:
982
_matchingbzrdir - the bzrdir format that the repository format was
983
originally written to work with. This can be used if manually
984
constructing a bzrdir and repository, or more commonly for test suite
989
return "<%s>" % self.__class__.__name__
991
def __eq__(self, other):
992
# format objects are generally stateless
993
return isinstance(other, self.__class__)
996
def find_format(klass, a_bzrdir):
997
"""Return the format for the repository object in a_bzrdir.
999
This is used by bzr native formats that have a "format" file in
1000
the repository. Other methods may be used by different types of
1004
transport = a_bzrdir.get_repository_transport(None)
1005
format_string = transport.get("format").read()
1006
return format_registry.get(format_string)
1007
except errors.NoSuchFile:
1008
raise errors.NoRepositoryPresent(a_bzrdir)
1010
raise errors.UnknownFormatError(format=format_string)
1013
def register_format(klass, format):
1014
format_registry.register(format.get_format_string(), format)
1017
def unregister_format(klass, format):
1018
format_registry.remove(format.get_format_string())
1021
def get_default_format(klass):
1022
"""Return the current default format."""
1023
from bzrlib import bzrdir
1024
return bzrdir.format_registry.make_bzrdir('default').repository_format
1026
def _get_control_store(self, repo_transport, control_files):
1027
"""Return the control store for this repository."""
1028
raise NotImplementedError(self._get_control_store)
1030
def get_format_string(self):
1031
"""Return the ASCII format string that identifies this format.
1033
Note that in pre format ?? repositories the format string is
1034
not permitted nor written to disk.
1036
raise NotImplementedError(self.get_format_string)
1038
def get_format_description(self):
1039
"""Return the short description for this format."""
1040
raise NotImplementedError(self.get_format_description)
1042
def _get_revision_store(self, repo_transport, control_files):
1043
"""Return the revision store object for this a_bzrdir."""
1044
raise NotImplementedError(self._get_revision_store)
1046
def _get_text_rev_store(self,
1053
"""Common logic for getting a revision store for a repository.
1055
see self._get_revision_store for the subclass-overridable method to
1056
get the store for a repository.
1058
from bzrlib.store.revision.text import TextRevisionStore
1059
dir_mode = control_files._dir_mode
1060
file_mode = control_files._file_mode
1061
text_store =TextStore(transport.clone(name),
1063
compressed=compressed,
1065
file_mode=file_mode)
1066
_revision_store = TextRevisionStore(text_store, serializer)
1067
return _revision_store
1069
# TODO: this shouldn't be in the base class, it's specific to things that
1070
# use weaves or knits -- mbp 20070207
1071
def _get_versioned_file_store(self,
1076
versionedfile_class=None,
1077
versionedfile_kwargs={},
1079
if versionedfile_class is None:
1080
versionedfile_class = self._versionedfile_class
1081
weave_transport = control_files._transport.clone(name)
1082
dir_mode = control_files._dir_mode
1083
file_mode = control_files._file_mode
1084
return VersionedFileStore(weave_transport, prefixed=prefixed,
1086
file_mode=file_mode,
1087
versionedfile_class=versionedfile_class,
1088
versionedfile_kwargs=versionedfile_kwargs,
1091
def initialize(self, a_bzrdir, shared=False):
1092
"""Initialize a repository of this format in a_bzrdir.
1094
:param a_bzrdir: The bzrdir to put the new repository in it.
1095
:param shared: The repository should be initialized as a sharable one.
1097
This may raise UninitializableFormat if shared repository are not
1098
compatible the a_bzrdir.
1101
def is_supported(self):
1102
"""Is this format supported?
1104
Supported formats must be initializable and openable.
1105
Unsupported formats may not support initialization or committing or
1106
some other features depending on the reason for not being supported.
1110
def check_conversion_target(self, target_format):
1111
raise NotImplementedError(self.check_conversion_target)
1113
def open(self, a_bzrdir, _found=False):
1114
"""Return an instance of this format for the bzrdir a_bzrdir.
1116
_found is a private parameter, do not use it.
1118
raise NotImplementedError(self.open)
1121
class MetaDirRepositoryFormat(RepositoryFormat):
1122
"""Common base class for the new repositories using the metadir layout."""
1124
rich_root_data = False
1125
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1128
super(MetaDirRepositoryFormat, self).__init__()
1130
def _create_control_files(self, a_bzrdir):
1131
"""Create the required files and the initial control_files object."""
1132
# FIXME: RBC 20060125 don't peek under the covers
1133
# NB: no need to escape relative paths that are url safe.
1134
repository_transport = a_bzrdir.get_repository_transport(self)
1135
control_files = lockable_files.LockableFiles(repository_transport,
1136
'lock', lockdir.LockDir)
1137
control_files.create_lock()
1138
return control_files
1140
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1141
"""Upload the initial blank content."""
1142
control_files = self._create_control_files(a_bzrdir)
1143
control_files.lock_write()
1145
control_files._transport.mkdir_multi(dirs,
1146
mode=control_files._dir_mode)
1147
for file, content in files:
1148
control_files.put(file, content)
1149
for file, content in utf8_files:
1150
control_files.put_utf8(file, content)
1152
control_files.put_utf8('shared-storage', '')
1154
control_files.unlock()
1157
# formats which have no format string are not discoverable
1158
# and not independently creatable, so are not registered. They're
1159
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1160
# needed, it's constructed directly by the BzrDir. Non-native formats where
1161
# the repository is not separately opened are similar.
1163
format_registry.register_lazy(
1164
'Bazaar-NG Repository format 7',
1165
'bzrlib.repofmt.weaverepo',
1168
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1169
# default control directory format
1171
format_registry.register_lazy(
1172
'Bazaar-NG Knit Repository Format 1',
1173
'bzrlib.repofmt.knitrepo',
1174
'RepositoryFormatKnit1',
1176
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1178
format_registry.register_lazy(
1179
'Bazaar Knit Repository Format 2\n',
1180
'bzrlib.repofmt.knitrepo',
1181
'RepositoryFormatKnit2',
1185
class InterRepository(InterObject):
1186
"""This class represents operations taking place between two repositories.
1188
Its instances have methods like copy_content and fetch, and contain
1189
references to the source and target repositories these operations can be
1192
Often we will provide convenience methods on 'repository' which carry out
1193
operations with another repository - they will always forward to
1194
InterRepository.get(other).method_name(parameters).
1198
"""The available optimised InterRepository types."""
1200
def copy_content(self, revision_id=None, basis=None):
1201
raise NotImplementedError(self.copy_content)
1203
def fetch(self, revision_id=None, pb=None):
1204
"""Fetch the content required to construct revision_id.
1206
The content is copied from self.source to self.target.
1208
:param revision_id: if None all content is copied, if NULL_REVISION no
1210
:param pb: optional progress bar to use for progress reports. If not
1211
provided a default one will be created.
1213
Returns the copied revision count and the failed revisions in a tuple:
1216
raise NotImplementedError(self.fetch)
1219
def missing_revision_ids(self, revision_id=None):
1220
"""Return the revision ids that source has that target does not.
1222
These are returned in topological order.
1224
:param revision_id: only return revision ids included by this
1227
# generic, possibly worst case, slow code path.
1228
target_ids = set(self.target.all_revision_ids())
1229
if revision_id is not None:
1230
source_ids = self.source.get_ancestry(revision_id)
1231
assert source_ids[0] is None
1234
source_ids = self.source.all_revision_ids()
1235
result_set = set(source_ids).difference(target_ids)
1236
# this may look like a no-op: its not. It preserves the ordering
1237
# other_ids had while only returning the members from other_ids
1238
# that we've decided we need.
1239
return [rev_id for rev_id in source_ids if rev_id in result_set]
1242
class InterSameDataRepository(InterRepository):
1243
"""Code for converting between repositories that represent the same data.
1245
Data format and model must match for this to work.
1249
def _get_repo_format_to_test(self):
1250
"""Repository format for testing with."""
1251
return RepositoryFormat.get_default_format()
1254
def is_compatible(source, target):
1255
if not isinstance(source, Repository):
1257
if not isinstance(target, Repository):
1259
if source._format.rich_root_data == target._format.rich_root_data:
1265
def copy_content(self, revision_id=None, basis=None):
1266
"""Make a complete copy of the content in self into destination.
1268
This is a destructive operation! Do not use it on existing
1271
:param revision_id: Only copy the content needed to construct
1272
revision_id and its parents.
1273
:param basis: Copy the needed data preferentially from basis.
1276
self.target.set_make_working_trees(self.source.make_working_trees())
1277
except NotImplementedError:
1279
# grab the basis available data
1280
if basis is not None:
1281
self.target.fetch(basis, revision_id=revision_id)
1282
# but don't bother fetching if we have the needed data now.
1283
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1284
self.target.has_revision(revision_id)):
1286
self.target.fetch(self.source, revision_id=revision_id)
1289
def fetch(self, revision_id=None, pb=None):
1290
"""See InterRepository.fetch()."""
1291
from bzrlib.fetch import GenericRepoFetcher
1292
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1293
self.source, self.source._format, self.target,
1294
self.target._format)
1295
f = GenericRepoFetcher(to_repository=self.target,
1296
from_repository=self.source,
1297
last_revision=revision_id,
1299
return f.count_copied, f.failed_revisions
1302
class InterWeaveRepo(InterSameDataRepository):
1303
"""Optimised code paths between Weave based repositories."""
1306
def _get_repo_format_to_test(self):
1307
from bzrlib.repofmt import weaverepo
1308
return weaverepo.RepositoryFormat7()
1311
def is_compatible(source, target):
1312
"""Be compatible with known Weave formats.
1314
We don't test for the stores being of specific types because that
1315
could lead to confusing results, and there is no need to be
1318
from bzrlib.repofmt.weaverepo import (
1324
return (isinstance(source._format, (RepositoryFormat5,
1326
RepositoryFormat7)) and
1327
isinstance(target._format, (RepositoryFormat5,
1329
RepositoryFormat7)))
1330
except AttributeError:
1334
def copy_content(self, revision_id=None, basis=None):
1335
"""See InterRepository.copy_content()."""
1336
# weave specific optimised path:
1337
if basis is not None:
1338
# copy the basis in, then fetch remaining data.
1339
basis.copy_content_into(self.target, revision_id)
1340
# the basis copy_content_into could miss-set this.
1342
self.target.set_make_working_trees(self.source.make_working_trees())
1343
except NotImplementedError:
1345
self.target.fetch(self.source, revision_id=revision_id)
1348
self.target.set_make_working_trees(self.source.make_working_trees())
1349
except NotImplementedError:
1351
# FIXME do not peek!
1352
if self.source.control_files._transport.listable():
1353
pb = ui.ui_factory.nested_progress_bar()
1355
self.target.weave_store.copy_all_ids(
1356
self.source.weave_store,
1358
from_transaction=self.source.get_transaction(),
1359
to_transaction=self.target.get_transaction())
1360
pb.update('copying inventory', 0, 1)
1361
self.target.control_weaves.copy_multi(
1362
self.source.control_weaves, ['inventory'],
1363
from_transaction=self.source.get_transaction(),
1364
to_transaction=self.target.get_transaction())
1365
self.target._revision_store.text_store.copy_all_ids(
1366
self.source._revision_store.text_store,
1371
self.target.fetch(self.source, revision_id=revision_id)
1374
def fetch(self, revision_id=None, pb=None):
1375
"""See InterRepository.fetch()."""
1376
from bzrlib.fetch import GenericRepoFetcher
1377
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1378
self.source, self.source._format, self.target, self.target._format)
1379
f = GenericRepoFetcher(to_repository=self.target,
1380
from_repository=self.source,
1381
last_revision=revision_id,
1383
return f.count_copied, f.failed_revisions
1386
def missing_revision_ids(self, revision_id=None):
1387
"""See InterRepository.missing_revision_ids()."""
1388
# we want all revisions to satisfy revision_id in source.
1389
# but we don't want to stat every file here and there.
1390
# we want then, all revisions other needs to satisfy revision_id
1391
# checked, but not those that we have locally.
1392
# so the first thing is to get a subset of the revisions to
1393
# satisfy revision_id in source, and then eliminate those that
1394
# we do already have.
1395
# this is slow on high latency connection to self, but as as this
1396
# disk format scales terribly for push anyway due to rewriting
1397
# inventory.weave, this is considered acceptable.
1399
if revision_id is not None:
1400
source_ids = self.source.get_ancestry(revision_id)
1401
assert source_ids[0] is None
1404
source_ids = self.source._all_possible_ids()
1405
source_ids_set = set(source_ids)
1406
# source_ids is the worst possible case we may need to pull.
1407
# now we want to filter source_ids against what we actually
1408
# have in target, but don't try to check for existence where we know
1409
# we do not have a revision as that would be pointless.
1410
target_ids = set(self.target._all_possible_ids())
1411
possibly_present_revisions = target_ids.intersection(source_ids_set)
1412
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1413
required_revisions = source_ids_set.difference(actually_present_revisions)
1414
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1415
if revision_id is not None:
1416
# we used get_ancestry to determine source_ids then we are assured all
1417
# revisions referenced are present as they are installed in topological order.
1418
# and the tip revision was validated by get_ancestry.
1419
return required_topo_revisions
1421
# if we just grabbed the possibly available ids, then
1422
# we only have an estimate of whats available and need to validate
1423
# that against the revision records.
1424
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1427
class InterKnitRepo(InterSameDataRepository):
1428
"""Optimised code paths between Knit based repositories."""
1431
def _get_repo_format_to_test(self):
1432
from bzrlib.repofmt import knitrepo
1433
return knitrepo.RepositoryFormatKnit1()
1436
def is_compatible(source, target):
1437
"""Be compatible with known Knit formats.
1439
We don't test for the stores being of specific types because that
1440
could lead to confusing results, and there is no need to be
1443
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1445
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1446
isinstance(target._format, (RepositoryFormatKnit1)))
1447
except AttributeError:
1451
def fetch(self, revision_id=None, pb=None):
1452
"""See InterRepository.fetch()."""
1453
from bzrlib.fetch import KnitRepoFetcher
1454
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1455
self.source, self.source._format, self.target, self.target._format)
1456
f = KnitRepoFetcher(to_repository=self.target,
1457
from_repository=self.source,
1458
last_revision=revision_id,
1460
return f.count_copied, f.failed_revisions
1463
def missing_revision_ids(self, revision_id=None):
1464
"""See InterRepository.missing_revision_ids()."""
1465
if revision_id is not None:
1466
source_ids = self.source.get_ancestry(revision_id)
1467
assert source_ids[0] is None
1470
source_ids = self.source._all_possible_ids()
1471
source_ids_set = set(source_ids)
1472
# source_ids is the worst possible case we may need to pull.
1473
# now we want to filter source_ids against what we actually
1474
# have in target, but don't try to check for existence where we know
1475
# we do not have a revision as that would be pointless.
1476
target_ids = set(self.target._all_possible_ids())
1477
possibly_present_revisions = target_ids.intersection(source_ids_set)
1478
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1479
required_revisions = source_ids_set.difference(actually_present_revisions)
1480
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1481
if revision_id is not None:
1482
# we used get_ancestry to determine source_ids then we are assured all
1483
# revisions referenced are present as they are installed in topological order.
1484
# and the tip revision was validated by get_ancestry.
1485
return required_topo_revisions
1487
# if we just grabbed the possibly available ids, then
1488
# we only have an estimate of whats available and need to validate
1489
# that against the revision records.
1490
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1493
class InterModel1and2(InterRepository):
1496
def _get_repo_format_to_test(self):
1500
def is_compatible(source, target):
1501
if not isinstance(source, Repository):
1503
if not isinstance(target, Repository):
1505
if not source._format.rich_root_data and target._format.rich_root_data:
1511
def fetch(self, revision_id=None, pb=None):
1512
"""See InterRepository.fetch()."""
1513
from bzrlib.fetch import Model1toKnit2Fetcher
1514
f = Model1toKnit2Fetcher(to_repository=self.target,
1515
from_repository=self.source,
1516
last_revision=revision_id,
1518
return f.count_copied, f.failed_revisions
1521
def copy_content(self, revision_id=None, basis=None):
1522
"""Make a complete copy of the content in self into destination.
1524
This is a destructive operation! Do not use it on existing
1527
:param revision_id: Only copy the content needed to construct
1528
revision_id and its parents.
1529
:param basis: Copy the needed data preferentially from basis.
1532
self.target.set_make_working_trees(self.source.make_working_trees())
1533
except NotImplementedError:
1535
# grab the basis available data
1536
if basis is not None:
1537
self.target.fetch(basis, revision_id=revision_id)
1538
# but don't bother fetching if we have the needed data now.
1539
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1540
self.target.has_revision(revision_id)):
1542
self.target.fetch(self.source, revision_id=revision_id)
1545
class InterKnit1and2(InterKnitRepo):
1548
def _get_repo_format_to_test(self):
1552
def is_compatible(source, target):
1553
"""Be compatible with Knit1 source and Knit2 target"""
1554
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit2
1556
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1557
RepositoryFormatKnit2
1558
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1559
isinstance(target._format, (RepositoryFormatKnit2)))
1560
except AttributeError:
1564
def fetch(self, revision_id=None, pb=None):
1565
"""See InterRepository.fetch()."""
1566
from bzrlib.fetch import Knit1to2Fetcher
1567
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1568
self.source, self.source._format, self.target,
1569
self.target._format)
1570
f = Knit1to2Fetcher(to_repository=self.target,
1571
from_repository=self.source,
1572
last_revision=revision_id,
1574
return f.count_copied, f.failed_revisions
1577
InterRepository.register_optimiser(InterSameDataRepository)
1578
InterRepository.register_optimiser(InterWeaveRepo)
1579
InterRepository.register_optimiser(InterKnitRepo)
1580
InterRepository.register_optimiser(InterModel1and2)
1581
InterRepository.register_optimiser(InterKnit1and2)
1584
class RepositoryTestProviderAdapter(object):
1585
"""A tool to generate a suite testing multiple repository formats at once.
1587
This is done by copying the test once for each transport and injecting
1588
the transport_server, transport_readonly_server, and bzrdir_format and
1589
repository_format classes into each copy. Each copy is also given a new id()
1590
to make it easy to identify.
1593
def __init__(self, transport_server, transport_readonly_server, formats):
1594
self._transport_server = transport_server
1595
self._transport_readonly_server = transport_readonly_server
1596
self._formats = formats
1598
def adapt(self, test):
1599
result = unittest.TestSuite()
1600
for repository_format, bzrdir_format in self._formats:
1601
from copy import deepcopy
1602
new_test = deepcopy(test)
1603
new_test.transport_server = self._transport_server
1604
new_test.transport_readonly_server = self._transport_readonly_server
1605
new_test.bzrdir_format = bzrdir_format
1606
new_test.repository_format = repository_format
1607
def make_new_test_id():
1608
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1609
return lambda: new_id
1610
new_test.id = make_new_test_id()
1611
result.addTest(new_test)
1615
class InterRepositoryTestProviderAdapter(object):
1616
"""A tool to generate a suite testing multiple inter repository formats.
1618
This is done by copying the test once for each interrepo provider and injecting
1619
the transport_server, transport_readonly_server, repository_format and
1620
repository_to_format classes into each copy.
1621
Each copy is also given a new id() to make it easy to identify.
1624
def __init__(self, transport_server, transport_readonly_server, formats):
1625
self._transport_server = transport_server
1626
self._transport_readonly_server = transport_readonly_server
1627
self._formats = formats
1629
def adapt(self, test):
1630
result = unittest.TestSuite()
1631
for interrepo_class, repository_format, repository_format_to in self._formats:
1632
from copy import deepcopy
1633
new_test = deepcopy(test)
1634
new_test.transport_server = self._transport_server
1635
new_test.transport_readonly_server = self._transport_readonly_server
1636
new_test.interrepo_class = interrepo_class
1637
new_test.repository_format = repository_format
1638
new_test.repository_format_to = repository_format_to
1639
def make_new_test_id():
1640
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1641
return lambda: new_id
1642
new_test.id = make_new_test_id()
1643
result.addTest(new_test)
1647
def default_test_list():
1648
"""Generate the default list of interrepo permutations to test."""
1649
from bzrlib.repofmt import knitrepo, weaverepo
1651
# test the default InterRepository between format 6 and the current
1653
# XXX: robertc 20060220 reinstate this when there are two supported
1654
# formats which do not have an optimal code path between them.
1655
#result.append((InterRepository,
1656
# RepositoryFormat6(),
1657
# RepositoryFormatKnit1()))
1658
for optimiser_class in InterRepository._optimisers:
1659
format_to_test = optimiser_class._get_repo_format_to_test()
1660
if format_to_test is not None:
1661
result.append((optimiser_class,
1662
format_to_test, format_to_test))
1663
# if there are specific combinations we want to use, we can add them
1665
result.append((InterModel1and2,
1666
weaverepo.RepositoryFormat5(),
1667
knitrepo.RepositoryFormatKnit2()))
1668
result.append((InterKnit1and2,
1669
knitrepo.RepositoryFormatKnit1(),
1670
knitrepo.RepositoryFormatKnit2()))
1674
class CopyConverter(object):
1675
"""A repository conversion tool which just performs a copy of the content.
1677
This is slow but quite reliable.
1680
def __init__(self, target_format):
1681
"""Create a CopyConverter.
1683
:param target_format: The format the resulting repository should be.
1685
self.target_format = target_format
1687
def convert(self, repo, pb):
1688
"""Perform the conversion of to_convert, giving feedback via pb.
1690
:param to_convert: The disk object to convert.
1691
:param pb: a progress bar to use for progress information.
1696
# this is only useful with metadir layouts - separated repo content.
1697
# trigger an assertion if not such
1698
repo._format.get_format_string()
1699
self.repo_dir = repo.bzrdir
1700
self.step('Moving repository to repository.backup')
1701
self.repo_dir.transport.move('repository', 'repository.backup')
1702
backup_transport = self.repo_dir.transport.clone('repository.backup')
1703
repo._format.check_conversion_target(self.target_format)
1704
self.source_repo = repo._format.open(self.repo_dir,
1706
_override_transport=backup_transport)
1707
self.step('Creating new repository')
1708
converted = self.target_format.initialize(self.repo_dir,
1709
self.source_repo.is_shared())
1710
converted.lock_write()
1712
self.step('Copying content into repository.')
1713
self.source_repo.copy_content_into(converted)
1716
self.step('Deleting old repository content.')
1717
self.repo_dir.transport.delete_tree('repository.backup')
1718
self.pb.note('repository converted')
1720
def step(self, message):
1721
"""Update the pb by a step."""
1723
self.pb.update(message, self.count, self.total)
1726
class CommitBuilder(object):
1727
"""Provides an interface to build up a commit.
1729
This allows describing a tree to be committed without needing to
1730
know the internals of the format of the repository.
1733
record_root_entry = False
1734
def __init__(self, repository, parents, config, timestamp=None,
1735
timezone=None, committer=None, revprops=None,
1737
"""Initiate a CommitBuilder.
1739
:param repository: Repository to commit to.
1740
:param parents: Revision ids of the parents of the new revision.
1741
:param config: Configuration to use.
1742
:param timestamp: Optional timestamp recorded for commit.
1743
:param timezone: Optional timezone for timestamp.
1744
:param committer: Optional committer to set for commit.
1745
:param revprops: Optional dictionary of revision properties.
1746
:param revision_id: Optional revision id.
1748
self._config = config
1750
if committer is None:
1751
self._committer = self._config.username()
1753
assert isinstance(committer, basestring), type(committer)
1754
self._committer = committer
1756
self.new_inventory = Inventory(None)
1757
self._new_revision_id = revision_id
1758
self.parents = parents
1759
self.repository = repository
1762
if revprops is not None:
1763
self._revprops.update(revprops)
1765
if timestamp is None:
1766
timestamp = time.time()
1767
# Restrict resolution to 1ms
1768
self._timestamp = round(timestamp, 3)
1770
if timezone is None:
1771
self._timezone = osutils.local_time_offset()
1773
self._timezone = int(timezone)
1775
self._generate_revision_if_needed()
1777
def commit(self, message):
1778
"""Make the actual commit.
1780
:return: The revision id of the recorded revision.
1782
rev = _mod_revision.Revision(
1783
timestamp=self._timestamp,
1784
timezone=self._timezone,
1785
committer=self._committer,
1787
inventory_sha1=self.inv_sha1,
1788
revision_id=self._new_revision_id,
1789
properties=self._revprops)
1790
rev.parent_ids = self.parents
1791
self.repository.add_revision(self._new_revision_id, rev,
1792
self.new_inventory, self._config)
1793
return self._new_revision_id
1795
def revision_tree(self):
1796
"""Return the tree that was just committed.
1798
After calling commit() this can be called to get a RevisionTree
1799
representing the newly committed tree. This is preferred to
1800
calling Repository.revision_tree() because that may require
1801
deserializing the inventory, while we already have a copy in
1804
return RevisionTree(self.repository, self.new_inventory,
1805
self._new_revision_id)
1807
def finish_inventory(self):
1808
"""Tell the builder that the inventory is finished."""
1809
if self.new_inventory.root is None:
1810
symbol_versioning.warn('Root entry should be supplied to'
1811
' record_entry_contents, as of bzr 0.10.',
1812
DeprecationWarning, stacklevel=2)
1813
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
1814
self.new_inventory.revision_id = self._new_revision_id
1815
self.inv_sha1 = self.repository.add_inventory(
1816
self._new_revision_id,
1821
def _gen_revision_id(self):
1822
"""Return new revision-id."""
1823
return generate_ids.gen_revision_id(self._config.username(),
1826
def _generate_revision_if_needed(self):
1827
"""Create a revision id if None was supplied.
1829
If the repository can not support user-specified revision ids
1830
they should override this function and raise CannotSetRevisionId
1831
if _new_revision_id is not None.
1833
:raises: CannotSetRevisionId
1835
if self._new_revision_id is None:
1836
self._new_revision_id = self._gen_revision_id()
1838
def record_entry_contents(self, ie, parent_invs, path, tree):
1839
"""Record the content of ie from tree into the commit if needed.
1841
Side effect: sets ie.revision when unchanged
1843
:param ie: An inventory entry present in the commit.
1844
:param parent_invs: The inventories of the parent revisions of the
1846
:param path: The path the entry is at in the tree.
1847
:param tree: The tree which contains this entry and should be used to
1850
if self.new_inventory.root is None and ie.parent_id is not None:
1851
symbol_versioning.warn('Root entry should be supplied to'
1852
' record_entry_contents, as of bzr 0.10.',
1853
DeprecationWarning, stacklevel=2)
1854
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
1856
self.new_inventory.add(ie)
1858
# ie.revision is always None if the InventoryEntry is considered
1859
# for committing. ie.snapshot will record the correct revision
1860
# which may be the sole parent if it is untouched.
1861
if ie.revision is not None:
1864
# In this revision format, root entries have no knit or weave
1865
if ie is self.new_inventory.root:
1866
# When serializing out to disk and back in
1867
# root.revision is always _new_revision_id
1868
ie.revision = self._new_revision_id
1870
previous_entries = ie.find_previous_heads(
1872
self.repository.weave_store,
1873
self.repository.get_transaction())
1874
# we are creating a new revision for ie in the history store
1876
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
1878
def modified_directory(self, file_id, file_parents):
1879
"""Record the presence of a symbolic link.
1881
:param file_id: The file_id of the link to record.
1882
:param file_parents: The per-file parent revision ids.
1884
self._add_text_to_weave(file_id, [], file_parents.keys())
1886
def modified_file_text(self, file_id, file_parents,
1887
get_content_byte_lines, text_sha1=None,
1889
"""Record the text of file file_id
1891
:param file_id: The file_id of the file to record the text of.
1892
:param file_parents: The per-file parent revision ids.
1893
:param get_content_byte_lines: A callable which will return the byte
1895
:param text_sha1: Optional SHA1 of the file contents.
1896
:param text_size: Optional size of the file contents.
1898
# mutter('storing text of file {%s} in revision {%s} into %r',
1899
# file_id, self._new_revision_id, self.repository.weave_store)
1900
# special case to avoid diffing on renames or
1902
if (len(file_parents) == 1
1903
and text_sha1 == file_parents.values()[0].text_sha1
1904
and text_size == file_parents.values()[0].text_size):
1905
previous_ie = file_parents.values()[0]
1906
versionedfile = self.repository.weave_store.get_weave(file_id,
1907
self.repository.get_transaction())
1908
versionedfile.clone_text(self._new_revision_id,
1909
previous_ie.revision, file_parents.keys())
1910
return text_sha1, text_size
1912
new_lines = get_content_byte_lines()
1913
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
1914
# should return the SHA1 and size
1915
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
1916
return osutils.sha_strings(new_lines), \
1917
sum(map(len, new_lines))
1919
def modified_link(self, file_id, file_parents, link_target):
1920
"""Record the presence of a symbolic link.
1922
:param file_id: The file_id of the link to record.
1923
:param file_parents: The per-file parent revision ids.
1924
:param link_target: Target location of this link.
1926
self._add_text_to_weave(file_id, [], file_parents.keys())
1928
def _add_text_to_weave(self, file_id, new_lines, parents):
1929
versionedfile = self.repository.weave_store.get_weave_or_empty(
1930
file_id, self.repository.get_transaction())
1931
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
1932
versionedfile.clear_cache()
1935
class _CommitBuilder(CommitBuilder):
1936
"""Temporary class so old CommitBuilders are detected properly
1938
Note: CommitBuilder works whether or not root entry is recorded.
1941
record_root_entry = True
1944
class RootCommitBuilder(CommitBuilder):
1945
"""This commitbuilder actually records the root id"""
1947
record_root_entry = True
1949
def record_entry_contents(self, ie, parent_invs, path, tree):
1950
"""Record the content of ie from tree into the commit if needed.
1952
Side effect: sets ie.revision when unchanged
1954
:param ie: An inventory entry present in the commit.
1955
:param parent_invs: The inventories of the parent revisions of the
1957
:param path: The path the entry is at in the tree.
1958
:param tree: The tree which contains this entry and should be used to
1961
assert self.new_inventory.root is not None or ie.parent_id is None
1962
self.new_inventory.add(ie)
1964
# ie.revision is always None if the InventoryEntry is considered
1965
# for committing. ie.snapshot will record the correct revision
1966
# which may be the sole parent if it is untouched.
1967
if ie.revision is not None:
1970
previous_entries = ie.find_previous_heads(
1972
self.repository.weave_store,
1973
self.repository.get_transaction())
1974
# we are creating a new revision for ie in the history store
1976
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
1988
def _unescaper(match, _map=_unescape_map):
1989
return _map[match.group(1)]
1995
def _unescape_xml(data):
1996
"""Unescape predefined XML entities in a string of data."""
1998
if _unescape_re is None:
1999
_unescape_re = re.compile('\&([^;]*);')
2000
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
