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),
828
m = __import__(from_module, globals(), locals(), [name])
830
return getattr(m, name)
831
except AttributeError:
832
raise AttributeError('module %s has no name %s'
834
globals()[name] = _deprecated_repository_forwarder
837
'AllInOneRepository',
838
'WeaveMetaDirRepository',
839
'PreSplitOutRepositoryFormat',
845
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
850
'RepositoryFormatKnit',
851
'RepositoryFormatKnit1',
852
'RepositoryFormatKnit2',
854
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
857
def install_revision(repository, rev, revision_tree):
858
"""Install all revision data into a repository."""
861
for p_id in rev.parent_ids:
862
if repository.has_revision(p_id):
863
present_parents.append(p_id)
864
parent_trees[p_id] = repository.revision_tree(p_id)
866
parent_trees[p_id] = repository.revision_tree(None)
868
inv = revision_tree.inventory
869
entries = inv.iter_entries()
870
# backwards compatability hack: skip the root id.
871
if not repository.supports_rich_root():
872
path, root = entries.next()
873
if root.revision != rev.revision_id:
874
raise errors.IncompatibleRevision(repr(repository))
875
# Add the texts that are not already present
876
for path, ie in entries:
877
w = repository.weave_store.get_weave_or_empty(ie.file_id,
878
repository.get_transaction())
879
if ie.revision not in w:
881
# FIXME: TODO: The following loop *may* be overlapping/duplicate
882
# with InventoryEntry.find_previous_heads(). if it is, then there
883
# is a latent bug here where the parents may have ancestors of each
885
for revision, tree in parent_trees.iteritems():
886
if ie.file_id not in tree:
888
parent_id = tree.inventory[ie.file_id].revision
889
if parent_id in text_parents:
891
text_parents.append(parent_id)
893
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
894
repository.get_transaction())
895
lines = revision_tree.get_file(ie.file_id).readlines()
896
vfile.add_lines(rev.revision_id, text_parents, lines)
898
# install the inventory
899
repository.add_inventory(rev.revision_id, inv, present_parents)
900
except errors.RevisionAlreadyPresent:
902
repository.add_revision(rev.revision_id, rev, inv)
905
class MetaDirRepository(Repository):
906
"""Repositories in the new meta-dir layout."""
908
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
909
super(MetaDirRepository, self).__init__(_format,
915
dir_mode = self.control_files._dir_mode
916
file_mode = self.control_files._file_mode
920
"""Return True if this repository is flagged as a shared repository."""
921
return self.control_files._transport.has('shared-storage')
924
def set_make_working_trees(self, new_value):
925
"""Set the policy flag for making working trees when creating branches.
927
This only applies to branches that use this repository.
929
The default is 'True'.
930
:param new_value: True to restore the default, False to disable making
935
self.control_files._transport.delete('no-working-trees')
936
except errors.NoSuchFile:
939
self.control_files.put_utf8('no-working-trees', '')
941
def make_working_trees(self):
942
"""Returns the policy for making working trees on new branches."""
943
return not self.control_files._transport.has('no-working-trees')
946
class RepositoryFormatRegistry(registry.Registry):
947
"""Registry of RepositoryFormats.
950
def get(self, format_string):
951
r = registry.Registry.get(self, format_string)
957
format_registry = RepositoryFormatRegistry()
958
"""Registry of formats, indexed by their identifying format string.
960
This can contain either format instances themselves, or classes/factories that
961
can be called to obtain one.
965
class RepositoryFormat(object):
966
"""A repository format.
968
Formats provide three things:
969
* An initialization routine to construct repository data on disk.
970
* a format string which is used when the BzrDir supports versioned
972
* an open routine which returns a Repository instance.
974
Formats are placed in an dict by their format string for reference
975
during opening. These should be subclasses of RepositoryFormat
978
Once a format is deprecated, just deprecate the initialize and open
979
methods on the format class. Do not deprecate the object, as the
980
object will be created every system load.
982
Common instance attributes:
983
_matchingbzrdir - the bzrdir format that the repository format was
984
originally written to work with. This can be used if manually
985
constructing a bzrdir and repository, or more commonly for test suite
990
return "<%s>" % self.__class__.__name__
992
def __eq__(self, other):
993
# format objects are generally stateless
994
return isinstance(other, self.__class__)
997
def find_format(klass, a_bzrdir):
998
"""Return the format for the repository object in a_bzrdir.
1000
This is used by bzr native formats that have a "format" file in
1001
the repository. Other methods may be used by different types of
1005
transport = a_bzrdir.get_repository_transport(None)
1006
format_string = transport.get("format").read()
1007
return format_registry.get(format_string)
1008
except errors.NoSuchFile:
1009
raise errors.NoRepositoryPresent(a_bzrdir)
1011
raise errors.UnknownFormatError(format=format_string)
1014
def register_format(klass, format):
1015
format_registry.register(format.get_format_string(), format)
1018
def unregister_format(klass, format):
1019
format_registry.remove(format.get_format_string())
1022
def get_default_format(klass):
1023
"""Return the current default format."""
1024
from bzrlib import bzrdir
1025
return bzrdir.format_registry.make_bzrdir('default').repository_format
1027
def _get_control_store(self, repo_transport, control_files):
1028
"""Return the control store for this repository."""
1029
raise NotImplementedError(self._get_control_store)
1031
def get_format_string(self):
1032
"""Return the ASCII format string that identifies this format.
1034
Note that in pre format ?? repositories the format string is
1035
not permitted nor written to disk.
1037
raise NotImplementedError(self.get_format_string)
1039
def get_format_description(self):
1040
"""Return the short description for this format."""
1041
raise NotImplementedError(self.get_format_description)
1043
def _get_revision_store(self, repo_transport, control_files):
1044
"""Return the revision store object for this a_bzrdir."""
1045
raise NotImplementedError(self._get_revision_store)
1047
def _get_text_rev_store(self,
1054
"""Common logic for getting a revision store for a repository.
1056
see self._get_revision_store for the subclass-overridable method to
1057
get the store for a repository.
1059
from bzrlib.store.revision.text import TextRevisionStore
1060
dir_mode = control_files._dir_mode
1061
file_mode = control_files._file_mode
1062
text_store =TextStore(transport.clone(name),
1064
compressed=compressed,
1066
file_mode=file_mode)
1067
_revision_store = TextRevisionStore(text_store, serializer)
1068
return _revision_store
1070
# TODO: this shouldn't be in the base class, it's specific to things that
1071
# use weaves or knits -- mbp 20070207
1072
def _get_versioned_file_store(self,
1077
versionedfile_class=None,
1078
versionedfile_kwargs={},
1080
if versionedfile_class is None:
1081
versionedfile_class = self._versionedfile_class
1082
weave_transport = control_files._transport.clone(name)
1083
dir_mode = control_files._dir_mode
1084
file_mode = control_files._file_mode
1085
return VersionedFileStore(weave_transport, prefixed=prefixed,
1087
file_mode=file_mode,
1088
versionedfile_class=versionedfile_class,
1089
versionedfile_kwargs=versionedfile_kwargs,
1092
def initialize(self, a_bzrdir, shared=False):
1093
"""Initialize a repository of this format in a_bzrdir.
1095
:param a_bzrdir: The bzrdir to put the new repository in it.
1096
:param shared: The repository should be initialized as a sharable one.
1098
This may raise UninitializableFormat if shared repository are not
1099
compatible the a_bzrdir.
1102
def is_supported(self):
1103
"""Is this format supported?
1105
Supported formats must be initializable and openable.
1106
Unsupported formats may not support initialization or committing or
1107
some other features depending on the reason for not being supported.
1111
def check_conversion_target(self, target_format):
1112
raise NotImplementedError(self.check_conversion_target)
1114
def open(self, a_bzrdir, _found=False):
1115
"""Return an instance of this format for the bzrdir a_bzrdir.
1117
_found is a private parameter, do not use it.
1119
raise NotImplementedError(self.open)
1122
class MetaDirRepositoryFormat(RepositoryFormat):
1123
"""Common base class for the new repositories using the metadir layout."""
1125
rich_root_data = False
1126
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1129
super(MetaDirRepositoryFormat, self).__init__()
1131
def _create_control_files(self, a_bzrdir):
1132
"""Create the required files and the initial control_files object."""
1133
# FIXME: RBC 20060125 don't peek under the covers
1134
# NB: no need to escape relative paths that are url safe.
1135
repository_transport = a_bzrdir.get_repository_transport(self)
1136
control_files = lockable_files.LockableFiles(repository_transport,
1137
'lock', lockdir.LockDir)
1138
control_files.create_lock()
1139
return control_files
1141
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1142
"""Upload the initial blank content."""
1143
control_files = self._create_control_files(a_bzrdir)
1144
control_files.lock_write()
1146
control_files._transport.mkdir_multi(dirs,
1147
mode=control_files._dir_mode)
1148
for file, content in files:
1149
control_files.put(file, content)
1150
for file, content in utf8_files:
1151
control_files.put_utf8(file, content)
1153
control_files.put_utf8('shared-storage', '')
1155
control_files.unlock()
1158
# formats which have no format string are not discoverable
1159
# and not independently creatable, so are not registered. They're
1160
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1161
# needed, it's constructed directly by the BzrDir. Non-native formats where
1162
# the repository is not separately opened are similar.
1164
format_registry.register_lazy(
1165
'Bazaar-NG Repository format 7',
1166
'bzrlib.repofmt.weaverepo',
1169
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1170
# default control directory format
1172
format_registry.register_lazy(
1173
'Bazaar-NG Knit Repository Format 1',
1174
'bzrlib.repofmt.knitrepo',
1175
'RepositoryFormatKnit1',
1177
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1179
format_registry.register_lazy(
1180
'Bazaar Knit Repository Format 2\n',
1181
'bzrlib.repofmt.knitrepo',
1182
'RepositoryFormatKnit2',
1186
class InterRepository(InterObject):
1187
"""This class represents operations taking place between two repositories.
1189
Its instances have methods like copy_content and fetch, and contain
1190
references to the source and target repositories these operations can be
1193
Often we will provide convenience methods on 'repository' which carry out
1194
operations with another repository - they will always forward to
1195
InterRepository.get(other).method_name(parameters).
1199
"""The available optimised InterRepository types."""
1201
def copy_content(self, revision_id=None, basis=None):
1202
raise NotImplementedError(self.copy_content)
1204
def fetch(self, revision_id=None, pb=None):
1205
"""Fetch the content required to construct revision_id.
1207
The content is copied from self.source to self.target.
1209
:param revision_id: if None all content is copied, if NULL_REVISION no
1211
:param pb: optional progress bar to use for progress reports. If not
1212
provided a default one will be created.
1214
Returns the copied revision count and the failed revisions in a tuple:
1217
raise NotImplementedError(self.fetch)
1220
def missing_revision_ids(self, revision_id=None):
1221
"""Return the revision ids that source has that target does not.
1223
These are returned in topological order.
1225
:param revision_id: only return revision ids included by this
1228
# generic, possibly worst case, slow code path.
1229
target_ids = set(self.target.all_revision_ids())
1230
if revision_id is not None:
1231
source_ids = self.source.get_ancestry(revision_id)
1232
assert source_ids[0] is None
1235
source_ids = self.source.all_revision_ids()
1236
result_set = set(source_ids).difference(target_ids)
1237
# this may look like a no-op: its not. It preserves the ordering
1238
# other_ids had while only returning the members from other_ids
1239
# that we've decided we need.
1240
return [rev_id for rev_id in source_ids if rev_id in result_set]
1243
class InterSameDataRepository(InterRepository):
1244
"""Code for converting between repositories that represent the same data.
1246
Data format and model must match for this to work.
1250
def _get_repo_format_to_test(self):
1251
"""Repository format for testing with."""
1252
return RepositoryFormat.get_default_format()
1255
def is_compatible(source, target):
1256
if not isinstance(source, Repository):
1258
if not isinstance(target, Repository):
1260
if source._format.rich_root_data == target._format.rich_root_data:
1266
def copy_content(self, revision_id=None, basis=None):
1267
"""Make a complete copy of the content in self into destination.
1269
This is a destructive operation! Do not use it on existing
1272
:param revision_id: Only copy the content needed to construct
1273
revision_id and its parents.
1274
:param basis: Copy the needed data preferentially from basis.
1277
self.target.set_make_working_trees(self.source.make_working_trees())
1278
except NotImplementedError:
1280
# grab the basis available data
1281
if basis is not None:
1282
self.target.fetch(basis, revision_id=revision_id)
1283
# but don't bother fetching if we have the needed data now.
1284
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1285
self.target.has_revision(revision_id)):
1287
self.target.fetch(self.source, revision_id=revision_id)
1290
def fetch(self, revision_id=None, pb=None):
1291
"""See InterRepository.fetch()."""
1292
from bzrlib.fetch import GenericRepoFetcher
1293
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1294
self.source, self.source._format, self.target,
1295
self.target._format)
1296
f = GenericRepoFetcher(to_repository=self.target,
1297
from_repository=self.source,
1298
last_revision=revision_id,
1300
return f.count_copied, f.failed_revisions
1303
class InterWeaveRepo(InterSameDataRepository):
1304
"""Optimised code paths between Weave based repositories."""
1307
def _get_repo_format_to_test(self):
1308
from bzrlib.repofmt import weaverepo
1309
return weaverepo.RepositoryFormat7()
1312
def is_compatible(source, target):
1313
"""Be compatible with known Weave formats.
1315
We don't test for the stores being of specific types because that
1316
could lead to confusing results, and there is no need to be
1319
from bzrlib.repofmt.weaverepo import (
1325
return (isinstance(source._format, (RepositoryFormat5,
1327
RepositoryFormat7)) and
1328
isinstance(target._format, (RepositoryFormat5,
1330
RepositoryFormat7)))
1331
except AttributeError:
1335
def copy_content(self, revision_id=None, basis=None):
1336
"""See InterRepository.copy_content()."""
1337
# weave specific optimised path:
1338
if basis is not None:
1339
# copy the basis in, then fetch remaining data.
1340
basis.copy_content_into(self.target, revision_id)
1341
# the basis copy_content_into could miss-set this.
1343
self.target.set_make_working_trees(self.source.make_working_trees())
1344
except NotImplementedError:
1346
self.target.fetch(self.source, revision_id=revision_id)
1349
self.target.set_make_working_trees(self.source.make_working_trees())
1350
except NotImplementedError:
1352
# FIXME do not peek!
1353
if self.source.control_files._transport.listable():
1354
pb = ui.ui_factory.nested_progress_bar()
1356
self.target.weave_store.copy_all_ids(
1357
self.source.weave_store,
1359
from_transaction=self.source.get_transaction(),
1360
to_transaction=self.target.get_transaction())
1361
pb.update('copying inventory', 0, 1)
1362
self.target.control_weaves.copy_multi(
1363
self.source.control_weaves, ['inventory'],
1364
from_transaction=self.source.get_transaction(),
1365
to_transaction=self.target.get_transaction())
1366
self.target._revision_store.text_store.copy_all_ids(
1367
self.source._revision_store.text_store,
1372
self.target.fetch(self.source, revision_id=revision_id)
1375
def fetch(self, revision_id=None, pb=None):
1376
"""See InterRepository.fetch()."""
1377
from bzrlib.fetch import GenericRepoFetcher
1378
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1379
self.source, self.source._format, self.target, self.target._format)
1380
f = GenericRepoFetcher(to_repository=self.target,
1381
from_repository=self.source,
1382
last_revision=revision_id,
1384
return f.count_copied, f.failed_revisions
1387
def missing_revision_ids(self, revision_id=None):
1388
"""See InterRepository.missing_revision_ids()."""
1389
# we want all revisions to satisfy revision_id in source.
1390
# but we don't want to stat every file here and there.
1391
# we want then, all revisions other needs to satisfy revision_id
1392
# checked, but not those that we have locally.
1393
# so the first thing is to get a subset of the revisions to
1394
# satisfy revision_id in source, and then eliminate those that
1395
# we do already have.
1396
# this is slow on high latency connection to self, but as as this
1397
# disk format scales terribly for push anyway due to rewriting
1398
# inventory.weave, this is considered acceptable.
1400
if revision_id is not None:
1401
source_ids = self.source.get_ancestry(revision_id)
1402
assert source_ids[0] is None
1405
source_ids = self.source._all_possible_ids()
1406
source_ids_set = set(source_ids)
1407
# source_ids is the worst possible case we may need to pull.
1408
# now we want to filter source_ids against what we actually
1409
# have in target, but don't try to check for existence where we know
1410
# we do not have a revision as that would be pointless.
1411
target_ids = set(self.target._all_possible_ids())
1412
possibly_present_revisions = target_ids.intersection(source_ids_set)
1413
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1414
required_revisions = source_ids_set.difference(actually_present_revisions)
1415
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1416
if revision_id is not None:
1417
# we used get_ancestry to determine source_ids then we are assured all
1418
# revisions referenced are present as they are installed in topological order.
1419
# and the tip revision was validated by get_ancestry.
1420
return required_topo_revisions
1422
# if we just grabbed the possibly available ids, then
1423
# we only have an estimate of whats available and need to validate
1424
# that against the revision records.
1425
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1428
class InterKnitRepo(InterSameDataRepository):
1429
"""Optimised code paths between Knit based repositories."""
1432
def _get_repo_format_to_test(self):
1433
from bzrlib.repofmt import knitrepo
1434
return knitrepo.RepositoryFormatKnit1()
1437
def is_compatible(source, target):
1438
"""Be compatible with known Knit formats.
1440
We don't test for the stores being of specific types because that
1441
could lead to confusing results, and there is no need to be
1444
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1446
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1447
isinstance(target._format, (RepositoryFormatKnit1)))
1448
except AttributeError:
1452
def fetch(self, revision_id=None, pb=None):
1453
"""See InterRepository.fetch()."""
1454
from bzrlib.fetch import KnitRepoFetcher
1455
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1456
self.source, self.source._format, self.target, self.target._format)
1457
f = KnitRepoFetcher(to_repository=self.target,
1458
from_repository=self.source,
1459
last_revision=revision_id,
1461
return f.count_copied, f.failed_revisions
1464
def missing_revision_ids(self, revision_id=None):
1465
"""See InterRepository.missing_revision_ids()."""
1466
if revision_id is not None:
1467
source_ids = self.source.get_ancestry(revision_id)
1468
assert source_ids[0] is None
1471
source_ids = self.source._all_possible_ids()
1472
source_ids_set = set(source_ids)
1473
# source_ids is the worst possible case we may need to pull.
1474
# now we want to filter source_ids against what we actually
1475
# have in target, but don't try to check for existence where we know
1476
# we do not have a revision as that would be pointless.
1477
target_ids = set(self.target._all_possible_ids())
1478
possibly_present_revisions = target_ids.intersection(source_ids_set)
1479
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1480
required_revisions = source_ids_set.difference(actually_present_revisions)
1481
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1482
if revision_id is not None:
1483
# we used get_ancestry to determine source_ids then we are assured all
1484
# revisions referenced are present as they are installed in topological order.
1485
# and the tip revision was validated by get_ancestry.
1486
return required_topo_revisions
1488
# if we just grabbed the possibly available ids, then
1489
# we only have an estimate of whats available and need to validate
1490
# that against the revision records.
1491
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1494
class InterModel1and2(InterRepository):
1497
def _get_repo_format_to_test(self):
1501
def is_compatible(source, target):
1502
if not isinstance(source, Repository):
1504
if not isinstance(target, Repository):
1506
if not source._format.rich_root_data and target._format.rich_root_data:
1512
def fetch(self, revision_id=None, pb=None):
1513
"""See InterRepository.fetch()."""
1514
from bzrlib.fetch import Model1toKnit2Fetcher
1515
f = Model1toKnit2Fetcher(to_repository=self.target,
1516
from_repository=self.source,
1517
last_revision=revision_id,
1519
return f.count_copied, f.failed_revisions
1522
def copy_content(self, revision_id=None, basis=None):
1523
"""Make a complete copy of the content in self into destination.
1525
This is a destructive operation! Do not use it on existing
1528
:param revision_id: Only copy the content needed to construct
1529
revision_id and its parents.
1530
:param basis: Copy the needed data preferentially from basis.
1533
self.target.set_make_working_trees(self.source.make_working_trees())
1534
except NotImplementedError:
1536
# grab the basis available data
1537
if basis is not None:
1538
self.target.fetch(basis, revision_id=revision_id)
1539
# but don't bother fetching if we have the needed data now.
1540
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1541
self.target.has_revision(revision_id)):
1543
self.target.fetch(self.source, revision_id=revision_id)
1546
class InterKnit1and2(InterKnitRepo):
1549
def _get_repo_format_to_test(self):
1553
def is_compatible(source, target):
1554
"""Be compatible with Knit1 source and Knit2 target"""
1555
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit2
1557
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1558
RepositoryFormatKnit2
1559
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1560
isinstance(target._format, (RepositoryFormatKnit2)))
1561
except AttributeError:
1565
def fetch(self, revision_id=None, pb=None):
1566
"""See InterRepository.fetch()."""
1567
from bzrlib.fetch import Knit1to2Fetcher
1568
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1569
self.source, self.source._format, self.target,
1570
self.target._format)
1571
f = Knit1to2Fetcher(to_repository=self.target,
1572
from_repository=self.source,
1573
last_revision=revision_id,
1575
return f.count_copied, f.failed_revisions
1578
InterRepository.register_optimiser(InterSameDataRepository)
1579
InterRepository.register_optimiser(InterWeaveRepo)
1580
InterRepository.register_optimiser(InterKnitRepo)
1581
InterRepository.register_optimiser(InterModel1and2)
1582
InterRepository.register_optimiser(InterKnit1and2)
1585
class RepositoryTestProviderAdapter(object):
1586
"""A tool to generate a suite testing multiple repository formats at once.
1588
This is done by copying the test once for each transport and injecting
1589
the transport_server, transport_readonly_server, and bzrdir_format and
1590
repository_format classes into each copy. Each copy is also given a new id()
1591
to make it easy to identify.
1594
def __init__(self, transport_server, transport_readonly_server, formats):
1595
self._transport_server = transport_server
1596
self._transport_readonly_server = transport_readonly_server
1597
self._formats = formats
1599
def adapt(self, test):
1600
result = unittest.TestSuite()
1601
for repository_format, bzrdir_format in self._formats:
1602
from copy import deepcopy
1603
new_test = deepcopy(test)
1604
new_test.transport_server = self._transport_server
1605
new_test.transport_readonly_server = self._transport_readonly_server
1606
new_test.bzrdir_format = bzrdir_format
1607
new_test.repository_format = repository_format
1608
def make_new_test_id():
1609
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
1610
return lambda: new_id
1611
new_test.id = make_new_test_id()
1612
result.addTest(new_test)
1616
class InterRepositoryTestProviderAdapter(object):
1617
"""A tool to generate a suite testing multiple inter repository formats.
1619
This is done by copying the test once for each interrepo provider and injecting
1620
the transport_server, transport_readonly_server, repository_format and
1621
repository_to_format classes into each copy.
1622
Each copy is also given a new id() to make it easy to identify.
1625
def __init__(self, transport_server, transport_readonly_server, formats):
1626
self._transport_server = transport_server
1627
self._transport_readonly_server = transport_readonly_server
1628
self._formats = formats
1630
def adapt(self, test):
1631
result = unittest.TestSuite()
1632
for interrepo_class, repository_format, repository_format_to in self._formats:
1633
from copy import deepcopy
1634
new_test = deepcopy(test)
1635
new_test.transport_server = self._transport_server
1636
new_test.transport_readonly_server = self._transport_readonly_server
1637
new_test.interrepo_class = interrepo_class
1638
new_test.repository_format = repository_format
1639
new_test.repository_format_to = repository_format_to
1640
def make_new_test_id():
1641
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
1642
return lambda: new_id
1643
new_test.id = make_new_test_id()
1644
result.addTest(new_test)
1648
def default_test_list():
1649
"""Generate the default list of interrepo permutations to test."""
1650
from bzrlib.repofmt import knitrepo, weaverepo
1652
# test the default InterRepository between format 6 and the current
1654
# XXX: robertc 20060220 reinstate this when there are two supported
1655
# formats which do not have an optimal code path between them.
1656
#result.append((InterRepository,
1657
# RepositoryFormat6(),
1658
# RepositoryFormatKnit1()))
1659
for optimiser_class in InterRepository._optimisers:
1660
format_to_test = optimiser_class._get_repo_format_to_test()
1661
if format_to_test is not None:
1662
result.append((optimiser_class,
1663
format_to_test, format_to_test))
1664
# if there are specific combinations we want to use, we can add them
1666
result.append((InterModel1and2,
1667
weaverepo.RepositoryFormat5(),
1668
knitrepo.RepositoryFormatKnit2()))
1669
result.append((InterKnit1and2,
1670
knitrepo.RepositoryFormatKnit1(),
1671
knitrepo.RepositoryFormatKnit2()))
1675
class CopyConverter(object):
1676
"""A repository conversion tool which just performs a copy of the content.
1678
This is slow but quite reliable.
1681
def __init__(self, target_format):
1682
"""Create a CopyConverter.
1684
:param target_format: The format the resulting repository should be.
1686
self.target_format = target_format
1688
def convert(self, repo, pb):
1689
"""Perform the conversion of to_convert, giving feedback via pb.
1691
:param to_convert: The disk object to convert.
1692
:param pb: a progress bar to use for progress information.
1697
# this is only useful with metadir layouts - separated repo content.
1698
# trigger an assertion if not such
1699
repo._format.get_format_string()
1700
self.repo_dir = repo.bzrdir
1701
self.step('Moving repository to repository.backup')
1702
self.repo_dir.transport.move('repository', 'repository.backup')
1703
backup_transport = self.repo_dir.transport.clone('repository.backup')
1704
repo._format.check_conversion_target(self.target_format)
1705
self.source_repo = repo._format.open(self.repo_dir,
1707
_override_transport=backup_transport)
1708
self.step('Creating new repository')
1709
converted = self.target_format.initialize(self.repo_dir,
1710
self.source_repo.is_shared())
1711
converted.lock_write()
1713
self.step('Copying content into repository.')
1714
self.source_repo.copy_content_into(converted)
1717
self.step('Deleting old repository content.')
1718
self.repo_dir.transport.delete_tree('repository.backup')
1719
self.pb.note('repository converted')
1721
def step(self, message):
1722
"""Update the pb by a step."""
1724
self.pb.update(message, self.count, self.total)
1727
class CommitBuilder(object):
1728
"""Provides an interface to build up a commit.
1730
This allows describing a tree to be committed without needing to
1731
know the internals of the format of the repository.
1734
record_root_entry = False
1735
def __init__(self, repository, parents, config, timestamp=None,
1736
timezone=None, committer=None, revprops=None,
1738
"""Initiate a CommitBuilder.
1740
:param repository: Repository to commit to.
1741
:param parents: Revision ids of the parents of the new revision.
1742
:param config: Configuration to use.
1743
:param timestamp: Optional timestamp recorded for commit.
1744
:param timezone: Optional timezone for timestamp.
1745
:param committer: Optional committer to set for commit.
1746
:param revprops: Optional dictionary of revision properties.
1747
:param revision_id: Optional revision id.
1749
self._config = config
1751
if committer is None:
1752
self._committer = self._config.username()
1754
assert isinstance(committer, basestring), type(committer)
1755
self._committer = committer
1757
self.new_inventory = Inventory(None)
1758
self._new_revision_id = revision_id
1759
self.parents = parents
1760
self.repository = repository
1763
if revprops is not None:
1764
self._revprops.update(revprops)
1766
if timestamp is None:
1767
timestamp = time.time()
1768
# Restrict resolution to 1ms
1769
self._timestamp = round(timestamp, 3)
1771
if timezone is None:
1772
self._timezone = osutils.local_time_offset()
1774
self._timezone = int(timezone)
1776
self._generate_revision_if_needed()
1778
def commit(self, message):
1779
"""Make the actual commit.
1781
:return: The revision id of the recorded revision.
1783
rev = _mod_revision.Revision(
1784
timestamp=self._timestamp,
1785
timezone=self._timezone,
1786
committer=self._committer,
1788
inventory_sha1=self.inv_sha1,
1789
revision_id=self._new_revision_id,
1790
properties=self._revprops)
1791
rev.parent_ids = self.parents
1792
self.repository.add_revision(self._new_revision_id, rev,
1793
self.new_inventory, self._config)
1794
return self._new_revision_id
1796
def revision_tree(self):
1797
"""Return the tree that was just committed.
1799
After calling commit() this can be called to get a RevisionTree
1800
representing the newly committed tree. This is preferred to
1801
calling Repository.revision_tree() because that may require
1802
deserializing the inventory, while we already have a copy in
1805
return RevisionTree(self.repository, self.new_inventory,
1806
self._new_revision_id)
1808
def finish_inventory(self):
1809
"""Tell the builder that the inventory is finished."""
1810
if self.new_inventory.root is None:
1811
symbol_versioning.warn('Root entry should be supplied to'
1812
' record_entry_contents, as of bzr 0.10.',
1813
DeprecationWarning, stacklevel=2)
1814
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
1815
self.new_inventory.revision_id = self._new_revision_id
1816
self.inv_sha1 = self.repository.add_inventory(
1817
self._new_revision_id,
1822
def _gen_revision_id(self):
1823
"""Return new revision-id."""
1824
return generate_ids.gen_revision_id(self._config.username(),
1827
def _generate_revision_if_needed(self):
1828
"""Create a revision id if None was supplied.
1830
If the repository can not support user-specified revision ids
1831
they should override this function and raise CannotSetRevisionId
1832
if _new_revision_id is not None.
1834
:raises: CannotSetRevisionId
1836
if self._new_revision_id is None:
1837
self._new_revision_id = self._gen_revision_id()
1839
def record_entry_contents(self, ie, parent_invs, path, tree):
1840
"""Record the content of ie from tree into the commit if needed.
1842
Side effect: sets ie.revision when unchanged
1844
:param ie: An inventory entry present in the commit.
1845
:param parent_invs: The inventories of the parent revisions of the
1847
:param path: The path the entry is at in the tree.
1848
:param tree: The tree which contains this entry and should be used to
1851
if self.new_inventory.root is None and ie.parent_id is not None:
1852
symbol_versioning.warn('Root entry should be supplied to'
1853
' record_entry_contents, as of bzr 0.10.',
1854
DeprecationWarning, stacklevel=2)
1855
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
1857
self.new_inventory.add(ie)
1859
# ie.revision is always None if the InventoryEntry is considered
1860
# for committing. ie.snapshot will record the correct revision
1861
# which may be the sole parent if it is untouched.
1862
if ie.revision is not None:
1865
# In this revision format, root entries have no knit or weave
1866
if ie is self.new_inventory.root:
1867
# When serializing out to disk and back in
1868
# root.revision is always _new_revision_id
1869
ie.revision = self._new_revision_id
1871
previous_entries = ie.find_previous_heads(
1873
self.repository.weave_store,
1874
self.repository.get_transaction())
1875
# we are creating a new revision for ie in the history store
1877
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
1879
def modified_directory(self, file_id, file_parents):
1880
"""Record the presence of a symbolic link.
1882
:param file_id: The file_id of the link to record.
1883
:param file_parents: The per-file parent revision ids.
1885
self._add_text_to_weave(file_id, [], file_parents.keys())
1887
def modified_file_text(self, file_id, file_parents,
1888
get_content_byte_lines, text_sha1=None,
1890
"""Record the text of file file_id
1892
:param file_id: The file_id of the file to record the text of.
1893
:param file_parents: The per-file parent revision ids.
1894
:param get_content_byte_lines: A callable which will return the byte
1896
:param text_sha1: Optional SHA1 of the file contents.
1897
:param text_size: Optional size of the file contents.
1899
# mutter('storing text of file {%s} in revision {%s} into %r',
1900
# file_id, self._new_revision_id, self.repository.weave_store)
1901
# special case to avoid diffing on renames or
1903
if (len(file_parents) == 1
1904
and text_sha1 == file_parents.values()[0].text_sha1
1905
and text_size == file_parents.values()[0].text_size):
1906
previous_ie = file_parents.values()[0]
1907
versionedfile = self.repository.weave_store.get_weave(file_id,
1908
self.repository.get_transaction())
1909
versionedfile.clone_text(self._new_revision_id,
1910
previous_ie.revision, file_parents.keys())
1911
return text_sha1, text_size
1913
new_lines = get_content_byte_lines()
1914
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
1915
# should return the SHA1 and size
1916
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
1917
return osutils.sha_strings(new_lines), \
1918
sum(map(len, new_lines))
1920
def modified_link(self, file_id, file_parents, link_target):
1921
"""Record the presence of a symbolic link.
1923
:param file_id: The file_id of the link to record.
1924
:param file_parents: The per-file parent revision ids.
1925
:param link_target: Target location of this link.
1927
self._add_text_to_weave(file_id, [], file_parents.keys())
1929
def _add_text_to_weave(self, file_id, new_lines, parents):
1930
versionedfile = self.repository.weave_store.get_weave_or_empty(
1931
file_id, self.repository.get_transaction())
1932
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
1933
versionedfile.clear_cache()
1936
class _CommitBuilder(CommitBuilder):
1937
"""Temporary class so old CommitBuilders are detected properly
1939
Note: CommitBuilder works whether or not root entry is recorded.
1942
record_root_entry = True
1945
class RootCommitBuilder(CommitBuilder):
1946
"""This commitbuilder actually records the root id"""
1948
record_root_entry = True
1950
def record_entry_contents(self, ie, parent_invs, path, tree):
1951
"""Record the content of ie from tree into the commit if needed.
1953
Side effect: sets ie.revision when unchanged
1955
:param ie: An inventory entry present in the commit.
1956
:param parent_invs: The inventories of the parent revisions of the
1958
:param path: The path the entry is at in the tree.
1959
:param tree: The tree which contains this entry and should be used to
1962
assert self.new_inventory.root is not None or ie.parent_id is None
1963
self.new_inventory.add(ie)
1965
# ie.revision is always None if the InventoryEntry is considered
1966
# for committing. ie.snapshot will record the correct revision
1967
# which may be the sole parent if it is untouched.
1968
if ie.revision is not None:
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)
1989
def _unescaper(match, _map=_unescape_map):
1990
return _map[match.group(1)]
1996
def _unescape_xml(data):
1997
"""Unescape predefined XML entities in a string of data."""
1999
if _unescape_re is None:
2000
_unescape_re = re.compile('\&([^;]*);')
2001
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
