1
# Copyright (C) 2005, 2006 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(), """
21
from binascii import hexlify
22
from copy import deepcopy
40
revision as _mod_revision,
49
from bzrlib.osutils import (
54
from bzrlib.revisiontree import RevisionTree
55
from bzrlib.store.versioned import VersionedFileStore
56
from bzrlib.store.text import TextStore
57
from bzrlib.testament import Testament
60
from bzrlib.decorators import needs_read_lock, needs_write_lock
61
from bzrlib.inter import InterObject
62
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
63
from bzrlib.symbol_versioning import (
67
from bzrlib.trace import mutter, note, warning
70
# Old formats display a warning, but only once
71
_deprecation_warning_done = False
74
class Repository(object):
75
"""Repository holding history for one or more branches.
77
The repository holds and retrieves historical information including
78
revisions and file history. It's normally accessed only by the Branch,
79
which views a particular line of development through that history.
81
The Repository builds on top of Stores and a Transport, which respectively
82
describe the disk data format and the way of accessing the (possibly
86
_file_ids_altered_regex = lazy_regex.lazy_compile(
87
r'file_id="(?P<file_id>[^"]+)"'
88
r'.*revision="(?P<revision_id>[^"]+)"'
92
def add_inventory(self, revid, inv, parents):
93
"""Add the inventory inv to the repository as revid.
95
:param parents: The revision ids of the parents that revid
96
is known to have and are in the repository already.
98
returns the sha1 of the serialized inventory.
100
assert inv.revision_id is None or inv.revision_id == revid, \
101
"Mismatch between inventory revision" \
102
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
103
assert inv.root is not None
104
inv_text = self.serialise_inventory(inv)
105
inv_sha1 = osutils.sha_string(inv_text)
106
inv_vf = self.control_weaves.get_weave('inventory',
107
self.get_transaction())
108
self._inventory_add_lines(inv_vf, revid, parents, osutils.split_lines(inv_text))
111
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
113
for parent in parents:
115
final_parents.append(parent)
117
inv_vf.add_lines(revid, final_parents, lines)
120
def add_revision(self, rev_id, rev, inv=None, config=None):
121
"""Add rev to the revision store as rev_id.
123
:param rev_id: the revision id to use.
124
:param rev: The revision object.
125
:param inv: The inventory for the revision. if None, it will be looked
126
up in the inventory storer
127
:param config: If None no digital signature will be created.
128
If supplied its signature_needed method will be used
129
to determine if a signature should be made.
131
if config is not None and config.signature_needed():
133
inv = self.get_inventory(rev_id)
134
plaintext = Testament(rev, inv).as_short_text()
135
self.store_revision_signature(
136
gpg.GPGStrategy(config), plaintext, rev_id)
137
if not rev_id in self.get_inventory_weave():
139
raise errors.WeaveRevisionNotPresent(rev_id,
140
self.get_inventory_weave())
142
# yes, this is not suitable for adding with ghosts.
143
self.add_inventory(rev_id, inv, rev.parent_ids)
144
self._revision_store.add_revision(rev, self.get_transaction())
147
def _all_possible_ids(self):
148
"""Return all the possible revisions that we could find."""
149
return self.get_inventory_weave().versions()
151
def all_revision_ids(self):
152
"""Returns a list of all the revision ids in the repository.
154
This is deprecated because code should generally work on the graph
155
reachable from a particular revision, and ignore any other revisions
156
that might be present. There is no direct replacement method.
158
return self._all_revision_ids()
161
def _all_revision_ids(self):
162
"""Returns a list of all the revision ids in the repository.
164
These are in as much topological order as the underlying store can
165
present: for weaves ghosts may lead to a lack of correctness until
166
the reweave updates the parents list.
168
if self._revision_store.text_store.listable():
169
return self._revision_store.all_revision_ids(self.get_transaction())
170
result = self._all_possible_ids()
171
return self._eliminate_revisions_not_present(result)
173
def break_lock(self):
174
"""Break a lock if one is present from another instance.
176
Uses the ui factory to ask for confirmation if the lock may be from
179
self.control_files.break_lock()
182
def _eliminate_revisions_not_present(self, revision_ids):
183
"""Check every revision id in revision_ids to see if we have it.
185
Returns a set of the present revisions.
188
for id in revision_ids:
189
if self.has_revision(id):
194
def create(a_bzrdir):
195
"""Construct the current default format repository in a_bzrdir."""
196
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
198
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
199
"""instantiate a Repository.
201
:param _format: The format of the repository on disk.
202
:param a_bzrdir: The BzrDir of the repository.
204
In the future we will have a single api for all stores for
205
getting file texts, inventories and revisions, then
206
this construct will accept instances of those things.
208
super(Repository, self).__init__()
209
self._format = _format
210
# the following are part of the public API for Repository:
211
self.bzrdir = a_bzrdir
212
self.control_files = control_files
213
self._revision_store = _revision_store
214
self.text_store = text_store
215
# backwards compatibility
216
self.weave_store = text_store
217
# not right yet - should be more semantically clear ?
219
self.control_store = control_store
220
self.control_weaves = control_store
221
# TODO: make sure to construct the right store classes, etc, depending
222
# on whether escaping is required.
223
self._warn_if_deprecated()
224
self._serializer = xml5.serializer_v5
227
return '%s(%r)' % (self.__class__.__name__,
228
self.bzrdir.transport.base)
231
return self.control_files.is_locked()
233
def lock_write(self):
234
self.control_files.lock_write()
237
self.control_files.lock_read()
239
def get_physical_lock_status(self):
240
return self.control_files.get_physical_lock_status()
243
def missing_revision_ids(self, other, revision_id=None):
244
"""Return the revision ids that other has that this does not.
246
These are returned in topological order.
248
revision_id: only return revision ids included by revision_id.
250
return InterRepository.get(other, self).missing_revision_ids(revision_id)
254
"""Open the repository rooted at base.
256
For instance, if the repository is at URL/.bzr/repository,
257
Repository.open(URL) -> a Repository instance.
259
control = bzrdir.BzrDir.open(base)
260
return control.open_repository()
262
def copy_content_into(self, destination, revision_id=None, basis=None):
263
"""Make a complete copy of the content in self into destination.
265
This is a destructive operation! Do not use it on existing
268
return InterRepository.get(self, destination).copy_content(revision_id, basis)
270
def fetch(self, source, revision_id=None, pb=None):
271
"""Fetch the content required to construct revision_id from source.
273
If revision_id is None all content is copied.
275
return InterRepository.get(source, self).fetch(revision_id=revision_id,
278
def get_commit_builder(self, branch, parents, config, timestamp=None,
279
timezone=None, committer=None, revprops=None,
281
"""Obtain a CommitBuilder for this repository.
283
:param branch: Branch to commit to.
284
:param parents: Revision ids of the parents of the new revision.
285
:param config: Configuration to use.
286
:param timestamp: Optional timestamp recorded for commit.
287
:param timezone: Optional timezone for timestamp.
288
:param committer: Optional committer to set for commit.
289
:param revprops: Optional dictionary of revision properties.
290
:param revision_id: Optional revision id.
292
return _CommitBuilder(self, parents, config, timestamp, timezone,
293
committer, revprops, revision_id)
296
self.control_files.unlock()
299
def clone(self, a_bzrdir, revision_id=None, basis=None):
300
"""Clone this repository into a_bzrdir using the current format.
302
Currently no check is made that the format of this repository and
303
the bzrdir format are compatible. FIXME RBC 20060201.
305
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
306
# use target default format.
307
result = a_bzrdir.create_repository()
308
# FIXME RBC 20060209 split out the repository type to avoid this check ?
309
elif isinstance(a_bzrdir._format,
310
(bzrdir.BzrDirFormat4,
311
bzrdir.BzrDirFormat5,
312
bzrdir.BzrDirFormat6)):
313
result = a_bzrdir.open_repository()
315
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
316
self.copy_content_into(result, revision_id, basis)
320
def has_revision(self, revision_id):
321
"""True if this repository has a copy of the revision."""
322
return self._revision_store.has_revision_id(revision_id,
323
self.get_transaction())
326
def get_revision_reconcile(self, revision_id):
327
"""'reconcile' helper routine that allows access to a revision always.
329
This variant of get_revision does not cross check the weave graph
330
against the revision one as get_revision does: but it should only
331
be used by reconcile, or reconcile-alike commands that are correcting
332
or testing the revision graph.
334
if not revision_id or not isinstance(revision_id, basestring):
335
raise errors.InvalidRevisionId(revision_id=revision_id,
337
return self._revision_store.get_revisions([revision_id],
338
self.get_transaction())[0]
340
def get_revisions(self, revision_ids):
341
return self._revision_store.get_revisions(revision_ids,
342
self.get_transaction())
345
def get_revision_xml(self, revision_id):
346
rev = self.get_revision(revision_id)
348
# the current serializer..
349
self._revision_store._serializer.write_revision(rev, rev_tmp)
351
return rev_tmp.getvalue()
354
def get_revision(self, revision_id):
355
"""Return the Revision object for a named revision"""
356
r = self.get_revision_reconcile(revision_id)
357
# weave corruption can lead to absent revision markers that should be
359
# the following test is reasonably cheap (it needs a single weave read)
360
# and the weave is cached in read transactions. In write transactions
361
# it is not cached but typically we only read a small number of
362
# revisions. For knits when they are introduced we will probably want
363
# to ensure that caching write transactions are in use.
364
inv = self.get_inventory_weave()
365
self._check_revision_parents(r, inv)
369
def get_deltas_for_revisions(self, revisions):
370
"""Produce a generator of revision deltas.
372
Note that the input is a sequence of REVISIONS, not revision_ids.
373
Trees will be held in memory until the generator exits.
374
Each delta is relative to the revision's lefthand predecessor.
376
required_trees = set()
377
for revision in revisions:
378
required_trees.add(revision.revision_id)
379
required_trees.update(revision.parent_ids[:1])
380
trees = dict((t.get_revision_id(), t) for
381
t in self.revision_trees(required_trees))
382
for revision in revisions:
383
if not revision.parent_ids:
384
old_tree = self.revision_tree(None)
386
old_tree = trees[revision.parent_ids[0]]
387
yield trees[revision.revision_id].changes_from(old_tree)
390
def get_revision_delta(self, revision_id):
391
"""Return the delta for one revision.
393
The delta is relative to the left-hand predecessor of the
396
r = self.get_revision(revision_id)
397
return list(self.get_deltas_for_revisions([r]))[0]
399
def _check_revision_parents(self, revision, inventory):
400
"""Private to Repository and Fetch.
402
This checks the parentage of revision in an inventory weave for
403
consistency and is only applicable to inventory-weave-for-ancestry
404
using repository formats & fetchers.
406
weave_parents = inventory.get_parents(revision.revision_id)
407
weave_names = inventory.versions()
408
for parent_id in revision.parent_ids:
409
if parent_id in weave_names:
410
# this parent must not be a ghost.
411
if not parent_id in weave_parents:
413
raise errors.CorruptRepository(self)
416
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
417
signature = gpg_strategy.sign(plaintext)
418
self._revision_store.add_revision_signature_text(revision_id,
420
self.get_transaction())
422
def fileids_altered_by_revision_ids(self, revision_ids):
423
"""Find the file ids and versions affected by revisions.
425
:param revisions: an iterable containing revision ids.
426
:return: a dictionary mapping altered file-ids to an iterable of
427
revision_ids. Each altered file-ids has the exact revision_ids that
428
altered it listed explicitly.
430
assert self._serializer.support_altered_by_hack, \
431
("fileids_altered_by_revision_ids only supported for branches "
432
"which store inventory as unnested xml, not on %r" % self)
433
selected_revision_ids = set(revision_ids)
434
w = self.get_inventory_weave()
437
# this code needs to read every new line in every inventory for the
438
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
439
# not present in one of those inventories is unnecessary but not
440
# harmful because we are filtering by the revision id marker in the
441
# inventory lines : we only select file ids altered in one of those
442
# revisions. We don't need to see all lines in the inventory because
443
# only those added in an inventory in rev X can contain a revision=X
445
unescape_revid_cache = {}
446
unescape_fileid_cache = {}
448
# Move several functions to be local variables, since this is a long
450
search = self._file_ids_altered_regex.search
451
unescape = _unescape_xml_cached
452
setdefault = result.setdefault
453
pb = ui.ui_factory.nested_progress_bar()
455
for line in w.iter_lines_added_or_present_in_versions(
456
selected_revision_ids, pb=pb):
460
file_id, revision_id = match.group('file_id', 'revision_id')
461
revision_id = unescape(revision_id, unescape_revid_cache)
462
if revision_id in selected_revision_ids:
463
file_id = unescape(file_id, unescape_fileid_cache)
464
setdefault(file_id, set()).add(revision_id)
470
def get_inventory_weave(self):
471
return self.control_weaves.get_weave('inventory',
472
self.get_transaction())
475
def get_inventory(self, revision_id):
476
"""Get Inventory object by hash."""
477
return self.deserialise_inventory(
478
revision_id, self.get_inventory_xml(revision_id))
480
def deserialise_inventory(self, revision_id, xml):
481
"""Transform the xml into an inventory object.
483
:param revision_id: The expected revision id of the inventory.
484
:param xml: A serialised inventory.
486
result = self._serializer.read_inventory_from_string(xml)
487
result.root.revision = revision_id
490
def serialise_inventory(self, inv):
491
return self._serializer.write_inventory_to_string(inv)
494
def get_inventory_xml(self, revision_id):
495
"""Get inventory XML as a file object."""
497
assert isinstance(revision_id, basestring), type(revision_id)
498
iw = self.get_inventory_weave()
499
return iw.get_text(revision_id)
501
raise errors.HistoryMissing(self, 'inventory', revision_id)
504
def get_inventory_sha1(self, revision_id):
505
"""Return the sha1 hash of the inventory entry
507
return self.get_revision(revision_id).inventory_sha1
510
def get_revision_graph(self, revision_id=None):
511
"""Return a dictionary containing the revision graph.
513
:param revision_id: The revision_id to get a graph from. If None, then
514
the entire revision graph is returned. This is a deprecated mode of
515
operation and will be removed in the future.
516
:return: a dictionary of revision_id->revision_parents_list.
518
# special case NULL_REVISION
519
if revision_id == _mod_revision.NULL_REVISION:
521
a_weave = self.get_inventory_weave()
522
all_revisions = self._eliminate_revisions_not_present(
524
entire_graph = dict([(node, a_weave.get_parents(node)) for
525
node in all_revisions])
526
if revision_id is None:
528
elif revision_id not in entire_graph:
529
raise errors.NoSuchRevision(self, revision_id)
531
# add what can be reached from revision_id
533
pending = set([revision_id])
534
while len(pending) > 0:
536
result[node] = entire_graph[node]
537
for revision_id in result[node]:
538
if revision_id not in result:
539
pending.add(revision_id)
543
def get_revision_graph_with_ghosts(self, revision_ids=None):
544
"""Return a graph of the revisions with ghosts marked as applicable.
546
:param revision_ids: an iterable of revisions to graph or None for all.
547
:return: a Graph object with the graph reachable from revision_ids.
549
result = graph.Graph()
551
pending = set(self.all_revision_ids())
554
pending = set(revision_ids)
555
# special case NULL_REVISION
556
if _mod_revision.NULL_REVISION in pending:
557
pending.remove(_mod_revision.NULL_REVISION)
558
required = set(pending)
561
revision_id = pending.pop()
563
rev = self.get_revision(revision_id)
564
except errors.NoSuchRevision:
565
if revision_id in required:
568
result.add_ghost(revision_id)
570
for parent_id in rev.parent_ids:
571
# is this queued or done ?
572
if (parent_id not in pending and
573
parent_id not in done):
575
pending.add(parent_id)
576
result.add_node(revision_id, rev.parent_ids)
577
done.add(revision_id)
581
def get_revision_inventory(self, revision_id):
582
"""Return inventory of a past revision."""
583
# TODO: Unify this with get_inventory()
584
# bzr 0.0.6 and later imposes the constraint that the inventory_id
585
# must be the same as its revision, so this is trivial.
586
if revision_id is None:
587
# This does not make sense: if there is no revision,
588
# then it is the current tree inventory surely ?!
589
# and thus get_root_id() is something that looks at the last
590
# commit on the branch, and the get_root_id is an inventory check.
591
raise NotImplementedError
592
# return Inventory(self.get_root_id())
594
return self.get_inventory(revision_id)
598
"""Return True if this repository is flagged as a shared repository."""
599
raise NotImplementedError(self.is_shared)
602
def reconcile(self, other=None, thorough=False):
603
"""Reconcile this repository."""
604
from bzrlib.reconcile import RepoReconciler
605
reconciler = RepoReconciler(self, thorough=thorough)
606
reconciler.reconcile()
610
def revision_tree(self, revision_id):
611
"""Return Tree for a revision on this branch.
613
`revision_id` may be None for the empty tree revision.
615
# TODO: refactor this to use an existing revision object
616
# so we don't need to read it in twice.
617
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
618
return RevisionTree(self, Inventory(root_id=None),
619
_mod_revision.NULL_REVISION)
621
inv = self.get_revision_inventory(revision_id)
622
return RevisionTree(self, inv, revision_id)
625
def revision_trees(self, revision_ids):
626
"""Return Tree for a revision on this branch.
628
`revision_id` may not be None or 'null:'"""
629
assert None not in revision_ids
630
assert _mod_revision.NULL_REVISION not in revision_ids
631
texts = self.get_inventory_weave().get_texts(revision_ids)
632
for text, revision_id in zip(texts, revision_ids):
633
inv = self.deserialise_inventory(revision_id, text)
634
yield RevisionTree(self, inv, revision_id)
637
def get_ancestry(self, revision_id):
638
"""Return a list of revision-ids integrated by a revision.
640
The first element of the list is always None, indicating the origin
641
revision. This might change when we have history horizons, or
642
perhaps we should have a new API.
644
This is topologically sorted.
646
if revision_id is None:
648
if not self.has_revision(revision_id):
649
raise errors.NoSuchRevision(self, revision_id)
650
w = self.get_inventory_weave()
651
candidates = w.get_ancestry(revision_id)
652
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
655
def print_file(self, file, revision_id):
656
"""Print `file` to stdout.
658
FIXME RBC 20060125 as John Meinel points out this is a bad api
659
- it writes to stdout, it assumes that that is valid etc. Fix
660
by creating a new more flexible convenience function.
662
tree = self.revision_tree(revision_id)
663
# use inventory as it was in that revision
664
file_id = tree.inventory.path2id(file)
666
# TODO: jam 20060427 Write a test for this code path
667
# it had a bug in it, and was raising the wrong
669
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
670
tree.print_file(file_id)
672
def get_transaction(self):
673
return self.control_files.get_transaction()
675
def revision_parents(self, revid):
676
return self.get_inventory_weave().parent_names(revid)
679
def set_make_working_trees(self, new_value):
680
"""Set the policy flag for making working trees when creating branches.
682
This only applies to branches that use this repository.
684
The default is 'True'.
685
:param new_value: True to restore the default, False to disable making
688
raise NotImplementedError(self.set_make_working_trees)
690
def make_working_trees(self):
691
"""Returns the policy for making working trees on new branches."""
692
raise NotImplementedError(self.make_working_trees)
695
def sign_revision(self, revision_id, gpg_strategy):
696
plaintext = Testament.from_revision(self, revision_id).as_short_text()
697
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
700
def has_signature_for_revision_id(self, revision_id):
701
"""Query for a revision signature for revision_id in the repository."""
702
return self._revision_store.has_signature(revision_id,
703
self.get_transaction())
706
def get_signature_text(self, revision_id):
707
"""Return the text for a signature."""
708
return self._revision_store.get_signature_text(revision_id,
709
self.get_transaction())
712
def check(self, revision_ids):
713
"""Check consistency of all history of given revision_ids.
715
Different repository implementations should override _check().
717
:param revision_ids: A non-empty list of revision_ids whose ancestry
718
will be checked. Typically the last revision_id of a branch.
721
raise ValueError("revision_ids must be non-empty in %s.check"
723
return self._check(revision_ids)
725
def _check(self, revision_ids):
726
result = check.Check(self)
730
def _warn_if_deprecated(self):
731
global _deprecation_warning_done
732
if _deprecation_warning_done:
734
_deprecation_warning_done = True
735
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
736
% (self._format, self.bzrdir.transport.base))
738
def supports_rich_root(self):
739
return self._format.rich_root_data
741
def _check_ascii_revisionid(self, revision_id, method):
742
"""Private helper for ascii-only repositories."""
743
# weave repositories refuse to store revisionids that are non-ascii.
744
if revision_id is not None:
745
# weaves require ascii revision ids.
746
if isinstance(revision_id, unicode):
748
revision_id.encode('ascii')
749
except UnicodeEncodeError:
750
raise errors.NonAsciiRevisionId(method, self)
753
class AllInOneRepository(Repository):
754
"""Legacy support - the repository behaviour for all-in-one branches."""
756
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
757
# we reuse one control files instance.
758
dir_mode = a_bzrdir._control_files._dir_mode
759
file_mode = a_bzrdir._control_files._file_mode
761
def get_store(name, compressed=True, prefixed=False):
762
# FIXME: This approach of assuming stores are all entirely compressed
763
# or entirely uncompressed is tidy, but breaks upgrade from
764
# some existing branches where there's a mixture; we probably
765
# still want the option to look for both.
766
relpath = a_bzrdir._control_files._escape(name)
767
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
768
prefixed=prefixed, compressed=compressed,
771
#if self._transport.should_cache():
772
# cache_path = os.path.join(self.cache_root, name)
773
# os.mkdir(cache_path)
774
# store = bzrlib.store.CachedStore(store, cache_path)
777
# not broken out yet because the controlweaves|inventory_store
778
# and text_store | weave_store bits are still different.
779
if isinstance(_format, RepositoryFormat4):
780
# cannot remove these - there is still no consistent api
781
# which allows access to this old info.
782
self.inventory_store = get_store('inventory-store')
783
text_store = get_store('text-store')
784
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
786
def get_commit_builder(self, branch, parents, config, timestamp=None,
787
timezone=None, committer=None, revprops=None,
789
self._check_ascii_revisionid(revision_id, self.get_commit_builder)
790
return Repository.get_commit_builder(self, branch, parents, config,
791
timestamp, timezone, committer, revprops, revision_id)
795
"""AllInOne repositories cannot be shared."""
799
def set_make_working_trees(self, new_value):
800
"""Set the policy flag for making working trees when creating branches.
802
This only applies to branches that use this repository.
804
The default is 'True'.
805
:param new_value: True to restore the default, False to disable making
808
raise NotImplementedError(self.set_make_working_trees)
810
def make_working_trees(self):
811
"""Returns the policy for making working trees on new branches."""
815
def install_revision(repository, rev, revision_tree):
816
"""Install all revision data into a repository."""
819
for p_id in rev.parent_ids:
820
if repository.has_revision(p_id):
821
present_parents.append(p_id)
822
parent_trees[p_id] = repository.revision_tree(p_id)
824
parent_trees[p_id] = repository.revision_tree(None)
826
inv = revision_tree.inventory
827
entries = inv.iter_entries()
828
# backwards compatability hack: skip the root id.
829
if not repository.supports_rich_root():
830
path, root = entries.next()
831
if root.revision != rev.revision_id:
832
raise errors.IncompatibleRevision(repr(repository))
833
# Add the texts that are not already present
834
for path, ie in entries:
835
w = repository.weave_store.get_weave_or_empty(ie.file_id,
836
repository.get_transaction())
837
if ie.revision not in w:
839
# FIXME: TODO: The following loop *may* be overlapping/duplicate
840
# with InventoryEntry.find_previous_heads(). if it is, then there
841
# is a latent bug here where the parents may have ancestors of each
843
for revision, tree in parent_trees.iteritems():
844
if ie.file_id not in tree:
846
parent_id = tree.inventory[ie.file_id].revision
847
if parent_id in text_parents:
849
text_parents.append(parent_id)
851
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
852
repository.get_transaction())
853
lines = revision_tree.get_file(ie.file_id).readlines()
854
vfile.add_lines(rev.revision_id, text_parents, lines)
856
# install the inventory
857
repository.add_inventory(rev.revision_id, inv, present_parents)
858
except errors.RevisionAlreadyPresent:
860
repository.add_revision(rev.revision_id, rev, inv)
863
class MetaDirRepository(Repository):
864
"""Repositories in the new meta-dir layout."""
866
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
867
super(MetaDirRepository, self).__init__(_format,
873
dir_mode = self.control_files._dir_mode
874
file_mode = self.control_files._file_mode
878
"""Return True if this repository is flagged as a shared repository."""
879
return self.control_files._transport.has('shared-storage')
882
def set_make_working_trees(self, new_value):
883
"""Set the policy flag for making working trees when creating branches.
885
This only applies to branches that use this repository.
887
The default is 'True'.
888
:param new_value: True to restore the default, False to disable making
893
self.control_files._transport.delete('no-working-trees')
894
except errors.NoSuchFile:
897
self.control_files.put_utf8('no-working-trees', '')
899
def make_working_trees(self):
900
"""Returns the policy for making working trees on new branches."""
901
return not self.control_files._transport.has('no-working-trees')
904
class WeaveMetaDirRepository(MetaDirRepository):
905
"""A subclass of MetaDirRepository to set weave specific policy."""
907
def get_commit_builder(self, branch, parents, config, timestamp=None,
908
timezone=None, committer=None, revprops=None,
910
self._check_ascii_revisionid(revision_id, self.get_commit_builder)
911
return MetaDirRepository.get_commit_builder(self, branch, parents,
912
config, timestamp, timezone, committer, revprops, revision_id)
915
class KnitRepository(MetaDirRepository):
916
"""Knit format repository."""
918
def _warn_if_deprecated(self):
919
# This class isn't deprecated
922
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
923
inv_vf.add_lines_with_ghosts(revid, parents, lines)
926
def _all_revision_ids(self):
927
"""See Repository.all_revision_ids()."""
928
# Knits get the revision graph from the index of the revision knit, so
929
# it's always possible even if they're on an unlistable transport.
930
return self._revision_store.all_revision_ids(self.get_transaction())
932
def fileid_involved_between_revs(self, from_revid, to_revid):
933
"""Find file_id(s) which are involved in the changes between revisions.
935
This determines the set of revisions which are involved, and then
936
finds all file ids affected by those revisions.
938
vf = self._get_revision_vf()
939
from_set = set(vf.get_ancestry(from_revid))
940
to_set = set(vf.get_ancestry(to_revid))
941
changed = to_set.difference(from_set)
942
return self._fileid_involved_by_set(changed)
944
def fileid_involved(self, last_revid=None):
945
"""Find all file_ids modified in the ancestry of last_revid.
947
:param last_revid: If None, last_revision() will be used.
950
changed = set(self.all_revision_ids())
952
changed = set(self.get_ancestry(last_revid))
955
return self._fileid_involved_by_set(changed)
958
def get_ancestry(self, revision_id):
959
"""Return a list of revision-ids integrated by a revision.
961
This is topologically sorted.
963
if revision_id is None:
965
vf = self._get_revision_vf()
967
return [None] + vf.get_ancestry(revision_id)
968
except errors.RevisionNotPresent:
969
raise errors.NoSuchRevision(self, revision_id)
972
def get_revision(self, revision_id):
973
"""Return the Revision object for a named revision"""
974
return self.get_revision_reconcile(revision_id)
977
def get_revision_graph(self, revision_id=None):
978
"""Return a dictionary containing the revision graph.
980
:param revision_id: The revision_id to get a graph from. If None, then
981
the entire revision graph is returned. This is a deprecated mode of
982
operation and will be removed in the future.
983
:return: a dictionary of revision_id->revision_parents_list.
985
# special case NULL_REVISION
986
if revision_id == _mod_revision.NULL_REVISION:
988
a_weave = self._get_revision_vf()
989
entire_graph = a_weave.get_graph()
990
if revision_id is None:
991
return a_weave.get_graph()
992
elif revision_id not in a_weave:
993
raise errors.NoSuchRevision(self, revision_id)
995
# add what can be reached from revision_id
997
pending = set([revision_id])
998
while len(pending) > 0:
1000
result[node] = a_weave.get_parents(node)
1001
for revision_id in result[node]:
1002
if revision_id not in result:
1003
pending.add(revision_id)
1007
def get_revision_graph_with_ghosts(self, revision_ids=None):
1008
"""Return a graph of the revisions with ghosts marked as applicable.
1010
:param revision_ids: an iterable of revisions to graph or None for all.
1011
:return: a Graph object with the graph reachable from revision_ids.
1013
result = graph.Graph()
1014
vf = self._get_revision_vf()
1015
versions = set(vf.versions())
1016
if not revision_ids:
1017
pending = set(self.all_revision_ids())
1020
pending = set(revision_ids)
1021
# special case NULL_REVISION
1022
if _mod_revision.NULL_REVISION in pending:
1023
pending.remove(_mod_revision.NULL_REVISION)
1024
required = set(pending)
1027
revision_id = pending.pop()
1028
if not revision_id in versions:
1029
if revision_id in required:
1030
raise errors.NoSuchRevision(self, revision_id)
1032
result.add_ghost(revision_id)
1033
# mark it as done so we don't try for it again.
1034
done.add(revision_id)
1036
parent_ids = vf.get_parents_with_ghosts(revision_id)
1037
for parent_id in parent_ids:
1038
# is this queued or done ?
1039
if (parent_id not in pending and
1040
parent_id not in done):
1042
pending.add(parent_id)
1043
result.add_node(revision_id, parent_ids)
1044
done.add(revision_id)
1047
def _get_revision_vf(self):
1048
""":return: a versioned file containing the revisions."""
1049
vf = self._revision_store.get_revision_file(self.get_transaction())
1053
def reconcile(self, other=None, thorough=False):
1054
"""Reconcile this repository."""
1055
from bzrlib.reconcile import KnitReconciler
1056
reconciler = KnitReconciler(self, thorough=thorough)
1057
reconciler.reconcile()
1060
def revision_parents(self, revision_id):
1061
return self._get_revision_vf().get_parents(revision_id)
1064
class KnitRepository2(KnitRepository):
1066
def __init__(self, _format, a_bzrdir, control_files, _revision_store,
1067
control_store, text_store):
1068
KnitRepository.__init__(self, _format, a_bzrdir, control_files,
1069
_revision_store, control_store, text_store)
1070
self._serializer = xml6.serializer_v6
1072
def deserialise_inventory(self, revision_id, xml):
1073
"""Transform the xml into an inventory object.
1075
:param revision_id: The expected revision id of the inventory.
1076
:param xml: A serialised inventory.
1078
result = self._serializer.read_inventory_from_string(xml)
1079
assert result.root.revision is not None
1082
def serialise_inventory(self, inv):
1083
"""Transform the inventory object into XML text.
1085
:param revision_id: The expected revision id of the inventory.
1086
:param xml: A serialised inventory.
1088
assert inv.revision_id is not None
1089
assert inv.root.revision is not None
1090
return KnitRepository.serialise_inventory(self, inv)
1092
def get_commit_builder(self, branch, parents, config, timestamp=None,
1093
timezone=None, committer=None, revprops=None,
1095
"""Obtain a CommitBuilder for this repository.
1097
:param branch: Branch to commit to.
1098
:param parents: Revision ids of the parents of the new revision.
1099
:param config: Configuration to use.
1100
:param timestamp: Optional timestamp recorded for commit.
1101
:param timezone: Optional timezone for timestamp.
1102
:param committer: Optional committer to set for commit.
1103
:param revprops: Optional dictionary of revision properties.
1104
:param revision_id: Optional revision id.
1106
return RootCommitBuilder(self, parents, config, timestamp, timezone,
1107
committer, revprops, revision_id)
1110
class RepositoryFormat(object):
1111
"""A repository format.
1113
Formats provide three things:
1114
* An initialization routine to construct repository data on disk.
1115
* a format string which is used when the BzrDir supports versioned
1117
* an open routine which returns a Repository instance.
1119
Formats are placed in an dict by their format string for reference
1120
during opening. These should be subclasses of RepositoryFormat
1123
Once a format is deprecated, just deprecate the initialize and open
1124
methods on the format class. Do not deprecate the object, as the
1125
object will be created every system load.
1127
Common instance attributes:
1128
_matchingbzrdir - the bzrdir format that the repository format was
1129
originally written to work with. This can be used if manually
1130
constructing a bzrdir and repository, or more commonly for test suite
1134
_default_format = None
1135
"""The default format used for new repositories."""
1138
"""The known formats."""
1141
return "<%s>" % self.__class__.__name__
1144
def find_format(klass, a_bzrdir):
1145
"""Return the format for the repository object in a_bzrdir."""
1147
transport = a_bzrdir.get_repository_transport(None)
1148
format_string = transport.get("format").read()
1149
return klass._formats[format_string]
1150
except errors.NoSuchFile:
1151
raise errors.NoRepositoryPresent(a_bzrdir)
1153
raise errors.UnknownFormatError(format=format_string)
1155
def _get_control_store(self, repo_transport, control_files):
1156
"""Return the control store for this repository."""
1157
raise NotImplementedError(self._get_control_store)
1160
def get_default_format(klass):
1161
"""Return the current default format."""
1162
return klass._default_format
1164
def get_format_string(self):
1165
"""Return the ASCII format string that identifies this format.
1167
Note that in pre format ?? repositories the format string is
1168
not permitted nor written to disk.
1170
raise NotImplementedError(self.get_format_string)
1172
def get_format_description(self):
1173
"""Return the short description for this format."""
1174
raise NotImplementedError(self.get_format_description)
1176
def _get_revision_store(self, repo_transport, control_files):
1177
"""Return the revision store object for this a_bzrdir."""
1178
raise NotImplementedError(self._get_revision_store)
1180
def _get_text_rev_store(self,
1187
"""Common logic for getting a revision store for a repository.
1189
see self._get_revision_store for the subclass-overridable method to
1190
get the store for a repository.
1192
from bzrlib.store.revision.text import TextRevisionStore
1193
dir_mode = control_files._dir_mode
1194
file_mode = control_files._file_mode
1195
text_store =TextStore(transport.clone(name),
1197
compressed=compressed,
1199
file_mode=file_mode)
1200
_revision_store = TextRevisionStore(text_store, serializer)
1201
return _revision_store
1203
def _get_versioned_file_store(self,
1208
versionedfile_class=weave.WeaveFile,
1209
versionedfile_kwargs={},
1211
weave_transport = control_files._transport.clone(name)
1212
dir_mode = control_files._dir_mode
1213
file_mode = control_files._file_mode
1214
return VersionedFileStore(weave_transport, prefixed=prefixed,
1216
file_mode=file_mode,
1217
versionedfile_class=versionedfile_class,
1218
versionedfile_kwargs=versionedfile_kwargs,
1221
def initialize(self, a_bzrdir, shared=False):
1222
"""Initialize a repository of this format in a_bzrdir.
1224
:param a_bzrdir: The bzrdir to put the new repository in it.
1225
:param shared: The repository should be initialized as a sharable one.
1226
:returns: The new repository object.
1228
This may raise UninitializableFormat if shared repository are not
1229
compatible the a_bzrdir.
1231
raise NotImplementedError(self.initialize)
1233
def is_supported(self):
1234
"""Is this format supported?
1236
Supported formats must be initializable and openable.
1237
Unsupported formats may not support initialization or committing or
1238
some other features depending on the reason for not being supported.
1242
def check_conversion_target(self, target_format):
1243
raise NotImplementedError(self.check_conversion_target)
1245
def open(self, a_bzrdir, _found=False):
1246
"""Return an instance of this format for the bzrdir a_bzrdir.
1248
_found is a private parameter, do not use it.
1250
raise NotImplementedError(self.open)
1253
def register_format(klass, format):
1254
klass._formats[format.get_format_string()] = format
1257
def set_default_format(klass, format):
1258
klass._default_format = format
1261
def unregister_format(klass, format):
1262
assert klass._formats[format.get_format_string()] is format
1263
del klass._formats[format.get_format_string()]
1266
class PreSplitOutRepositoryFormat(RepositoryFormat):
1267
"""Base class for the pre split out repository formats."""
1269
rich_root_data = False
1271
def initialize(self, a_bzrdir, shared=False, _internal=False):
1272
"""Create a weave repository.
1274
TODO: when creating split out bzr branch formats, move this to a common
1275
base for Format5, Format6. or something like that.
1278
raise errors.IncompatibleFormat(self, a_bzrdir._format)
1281
# always initialized when the bzrdir is.
1282
return self.open(a_bzrdir, _found=True)
1284
# Create an empty weave
1286
weavefile.write_weave_v5(weave.Weave(), sio)
1287
empty_weave = sio.getvalue()
1289
mutter('creating repository in %s.', a_bzrdir.transport.base)
1290
dirs = ['revision-store', 'weaves']
1291
files = [('inventory.weave', StringIO(empty_weave)),
1294
# FIXME: RBC 20060125 don't peek under the covers
1295
# NB: no need to escape relative paths that are url safe.
1296
control_files = lockable_files.LockableFiles(a_bzrdir.transport,
1297
'branch-lock', lockable_files.TransportLock)
1298
control_files.create_lock()
1299
control_files.lock_write()
1300
control_files._transport.mkdir_multi(dirs,
1301
mode=control_files._dir_mode)
1303
for file, content in files:
1304
control_files.put(file, content)
1306
control_files.unlock()
1307
return self.open(a_bzrdir, _found=True)
1309
def _get_control_store(self, repo_transport, control_files):
1310
"""Return the control store for this repository."""
1311
return self._get_versioned_file_store('',
1316
def _get_text_store(self, transport, control_files):
1317
"""Get a store for file texts for this format."""
1318
raise NotImplementedError(self._get_text_store)
1320
def open(self, a_bzrdir, _found=False):
1321
"""See RepositoryFormat.open()."""
1323
# we are being called directly and must probe.
1324
raise NotImplementedError
1326
repo_transport = a_bzrdir.get_repository_transport(None)
1327
control_files = a_bzrdir._control_files
1328
text_store = self._get_text_store(repo_transport, control_files)
1329
control_store = self._get_control_store(repo_transport, control_files)
1330
_revision_store = self._get_revision_store(repo_transport, control_files)
1331
return AllInOneRepository(_format=self,
1333
_revision_store=_revision_store,
1334
control_store=control_store,
1335
text_store=text_store)
1337
def check_conversion_target(self, target_format):
1341
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1342
"""Bzr repository format 4.
1344
This repository format has:
1346
- TextStores for texts, inventories,revisions.
1348
This format is deprecated: it indexes texts using a text id which is
1349
removed in format 5; initialization and write support for this format
1354
super(RepositoryFormat4, self).__init__()
1355
self._matchingbzrdir = bzrdir.BzrDirFormat4()
1357
def get_format_description(self):
1358
"""See RepositoryFormat.get_format_description()."""
1359
return "Repository format 4"
1361
def initialize(self, url, shared=False, _internal=False):
1362
"""Format 4 branches cannot be created."""
1363
raise errors.UninitializableFormat(self)
1365
def is_supported(self):
1366
"""Format 4 is not supported.
1368
It is not supported because the model changed from 4 to 5 and the
1369
conversion logic is expensive - so doing it on the fly was not
1374
def _get_control_store(self, repo_transport, control_files):
1375
"""Format 4 repositories have no formal control store at this point.
1377
This will cause any control-file-needing apis to fail - this is desired.
1381
def _get_revision_store(self, repo_transport, control_files):
1382
"""See RepositoryFormat._get_revision_store()."""
1383
from bzrlib.xml4 import serializer_v4
1384
return self._get_text_rev_store(repo_transport,
1387
serializer=serializer_v4)
1389
def _get_text_store(self, transport, control_files):
1390
"""See RepositoryFormat._get_text_store()."""
1393
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1394
"""Bzr control format 5.
1396
This repository format has:
1397
- weaves for file texts and inventory
1399
- TextStores for revisions and signatures.
1403
super(RepositoryFormat5, self).__init__()
1404
self._matchingbzrdir = bzrdir.BzrDirFormat5()
1406
def get_format_description(self):
1407
"""See RepositoryFormat.get_format_description()."""
1408
return "Weave repository format 5"
1410
def _get_revision_store(self, repo_transport, control_files):
1411
"""See RepositoryFormat._get_revision_store()."""
1412
"""Return the revision store object for this a_bzrdir."""
1413
return self._get_text_rev_store(repo_transport,
1418
def _get_text_store(self, transport, control_files):
1419
"""See RepositoryFormat._get_text_store()."""
1420
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1423
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1424
"""Bzr control format 6.
1426
This repository format has:
1427
- weaves for file texts and inventory
1428
- hash subdirectory based stores.
1429
- TextStores for revisions and signatures.
1433
super(RepositoryFormat6, self).__init__()
1434
self._matchingbzrdir = bzrdir.BzrDirFormat6()
1436
def get_format_description(self):
1437
"""See RepositoryFormat.get_format_description()."""
1438
return "Weave repository format 6"
1440
def _get_revision_store(self, repo_transport, control_files):
1441
"""See RepositoryFormat._get_revision_store()."""
1442
return self._get_text_rev_store(repo_transport,
1448
def _get_text_store(self, transport, control_files):
1449
"""See RepositoryFormat._get_text_store()."""
1450
return self._get_versioned_file_store('weaves', transport, control_files)
1453
class MetaDirRepositoryFormat(RepositoryFormat):
1454
"""Common base class for the new repositories using the metadir layout."""
1456
rich_root_data = False
1459
super(MetaDirRepositoryFormat, self).__init__()
1460
self._matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1462
def _create_control_files(self, a_bzrdir):
1463
"""Create the required files and the initial control_files object."""
1464
# FIXME: RBC 20060125 don't peek under the covers
1465
# NB: no need to escape relative paths that are url safe.
1466
repository_transport = a_bzrdir.get_repository_transport(self)
1467
control_files = lockable_files.LockableFiles(repository_transport,
1468
'lock', lockdir.LockDir)
1469
control_files.create_lock()
1470
return control_files
1472
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1473
"""Upload the initial blank content."""
1474
control_files = self._create_control_files(a_bzrdir)
1475
control_files.lock_write()
1477
control_files._transport.mkdir_multi(dirs,
1478
mode=control_files._dir_mode)
1479
for file, content in files:
1480
control_files.put(file, content)
1481
for file, content in utf8_files:
1482
control_files.put_utf8(file, content)
1484
control_files.put_utf8('shared-storage', '')
1486
control_files.unlock()
1489
class RepositoryFormat7(MetaDirRepositoryFormat):
1490
"""Bzr repository 7.
1492
This repository format has:
1493
- weaves for file texts and inventory
1494
- hash subdirectory based stores.
1495
- TextStores for revisions and signatures.
1496
- a format marker of its own
1497
- an optional 'shared-storage' flag
1498
- an optional 'no-working-trees' flag
1501
def _get_control_store(self, repo_transport, control_files):
1502
"""Return the control store for this repository."""
1503
return self._get_versioned_file_store('',
1508
def get_format_string(self):
1509
"""See RepositoryFormat.get_format_string()."""
1510
return "Bazaar-NG Repository format 7"
1512
def get_format_description(self):
1513
"""See RepositoryFormat.get_format_description()."""
1514
return "Weave repository format 7"
1516
def check_conversion_target(self, target_format):
1519
def _get_revision_store(self, repo_transport, control_files):
1520
"""See RepositoryFormat._get_revision_store()."""
1521
return self._get_text_rev_store(repo_transport,
1528
def _get_text_store(self, transport, control_files):
1529
"""See RepositoryFormat._get_text_store()."""
1530
return self._get_versioned_file_store('weaves',
1534
def initialize(self, a_bzrdir, shared=False):
1535
"""Create a weave repository.
1537
:param shared: If true the repository will be initialized as a shared
1540
# Create an empty weave
1542
weavefile.write_weave_v5(weave.Weave(), sio)
1543
empty_weave = sio.getvalue()
1545
mutter('creating repository in %s.', a_bzrdir.transport.base)
1546
dirs = ['revision-store', 'weaves']
1547
files = [('inventory.weave', StringIO(empty_weave)),
1549
utf8_files = [('format', self.get_format_string())]
1551
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1552
return self.open(a_bzrdir=a_bzrdir, _found=True)
1554
def open(self, a_bzrdir, _found=False, _override_transport=None):
1555
"""See RepositoryFormat.open().
1557
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1558
repository at a slightly different url
1559
than normal. I.e. during 'upgrade'.
1562
format = RepositoryFormat.find_format(a_bzrdir)
1563
assert format.__class__ == self.__class__
1564
if _override_transport is not None:
1565
repo_transport = _override_transport
1567
repo_transport = a_bzrdir.get_repository_transport(None)
1568
control_files = lockable_files.LockableFiles(repo_transport,
1569
'lock', lockdir.LockDir)
1570
text_store = self._get_text_store(repo_transport, control_files)
1571
control_store = self._get_control_store(repo_transport, control_files)
1572
_revision_store = self._get_revision_store(repo_transport, control_files)
1573
return WeaveMetaDirRepository(_format=self,
1575
control_files=control_files,
1576
_revision_store=_revision_store,
1577
control_store=control_store,
1578
text_store=text_store)
1581
class RepositoryFormatKnit(MetaDirRepositoryFormat):
1582
"""Bzr repository knit format (generalized).
1584
This repository format has:
1585
- knits for file texts and inventory
1586
- hash subdirectory based stores.
1587
- knits for revisions and signatures
1588
- TextStores for revisions and signatures.
1589
- a format marker of its own
1590
- an optional 'shared-storage' flag
1591
- an optional 'no-working-trees' flag
1595
def _get_control_store(self, repo_transport, control_files):
1596
"""Return the control store for this repository."""
1597
return VersionedFileStore(
1600
file_mode=control_files._file_mode,
1601
versionedfile_class=knit.KnitVersionedFile,
1602
versionedfile_kwargs={'factory':knit.KnitPlainFactory()},
1605
def _get_revision_store(self, repo_transport, control_files):
1606
"""See RepositoryFormat._get_revision_store()."""
1607
from bzrlib.store.revision.knit import KnitRevisionStore
1608
versioned_file_store = VersionedFileStore(
1610
file_mode=control_files._file_mode,
1613
versionedfile_class=knit.KnitVersionedFile,
1614
versionedfile_kwargs={'delta':False,
1615
'factory':knit.KnitPlainFactory(),
1619
return KnitRevisionStore(versioned_file_store)
1621
def _get_text_store(self, transport, control_files):
1622
"""See RepositoryFormat._get_text_store()."""
1623
return self._get_versioned_file_store('knits',
1626
versionedfile_class=knit.KnitVersionedFile,
1627
versionedfile_kwargs={
1628
'create_parent_dir':True,
1629
'delay_create':True,
1630
'dir_mode':control_files._dir_mode,
1634
def initialize(self, a_bzrdir, shared=False):
1635
"""Create a knit format 1 repository.
1637
:param a_bzrdir: bzrdir to contain the new repository; must already
1639
:param shared: If true the repository will be initialized as a shared
1642
mutter('creating repository in %s.', a_bzrdir.transport.base)
1643
dirs = ['revision-store', 'knits']
1645
utf8_files = [('format', self.get_format_string())]
1647
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1648
repo_transport = a_bzrdir.get_repository_transport(None)
1649
control_files = lockable_files.LockableFiles(repo_transport,
1650
'lock', lockdir.LockDir)
1651
control_store = self._get_control_store(repo_transport, control_files)
1652
transaction = transactions.WriteTransaction()
1653
# trigger a write of the inventory store.
1654
control_store.get_weave_or_empty('inventory', transaction)
1655
_revision_store = self._get_revision_store(repo_transport, control_files)
1656
# the revision id here is irrelevant: it will not be stored, and cannot
1658
_revision_store.has_revision_id('A', transaction)
1659
_revision_store.get_signature_file(transaction)
1660
return self.open(a_bzrdir=a_bzrdir, _found=True)
1662
def open(self, a_bzrdir, _found=False, _override_transport=None):
1663
"""See RepositoryFormat.open().
1665
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1666
repository at a slightly different url
1667
than normal. I.e. during 'upgrade'.
1670
format = RepositoryFormat.find_format(a_bzrdir)
1671
assert format.__class__ == self.__class__
1672
if _override_transport is not None:
1673
repo_transport = _override_transport
1675
repo_transport = a_bzrdir.get_repository_transport(None)
1676
control_files = lockable_files.LockableFiles(repo_transport,
1677
'lock', lockdir.LockDir)
1678
text_store = self._get_text_store(repo_transport, control_files)
1679
control_store = self._get_control_store(repo_transport, control_files)
1680
_revision_store = self._get_revision_store(repo_transport, control_files)
1681
return KnitRepository(_format=self,
1683
control_files=control_files,
1684
_revision_store=_revision_store,
1685
control_store=control_store,
1686
text_store=text_store)
1689
class RepositoryFormatKnit1(RepositoryFormatKnit):
1690
"""Bzr repository knit format 1.
1692
This repository format has:
1693
- knits for file texts and inventory
1694
- hash subdirectory based stores.
1695
- knits for revisions and signatures
1696
- TextStores for revisions and signatures.
1697
- a format marker of its own
1698
- an optional 'shared-storage' flag
1699
- an optional 'no-working-trees' flag
1702
This format was introduced in bzr 0.8.
1704
def get_format_string(self):
1705
"""See RepositoryFormat.get_format_string()."""
1706
return "Bazaar-NG Knit Repository Format 1"
1708
def get_format_description(self):
1709
"""See RepositoryFormat.get_format_description()."""
1710
return "Knit repository format 1"
1712
def check_conversion_target(self, target_format):
1716
class RepositoryFormatKnit2(RepositoryFormatKnit):
1717
"""Bzr repository knit format 2.
1719
THIS FORMAT IS EXPERIMENTAL
1720
This repository format has:
1721
- knits for file texts and inventory
1722
- hash subdirectory based stores.
1723
- knits for revisions and signatures
1724
- TextStores for revisions and signatures.
1725
- a format marker of its own
1726
- an optional 'shared-storage' flag
1727
- an optional 'no-working-trees' flag
1729
- Support for recording full info about the tree root
1733
rich_root_data = True
1735
def get_format_string(self):
1736
"""See RepositoryFormat.get_format_string()."""
1737
return "Bazaar Knit Repository Format 2\n"
1739
def get_format_description(self):
1740
"""See RepositoryFormat.get_format_description()."""
1741
return "Knit repository format 2"
1743
def check_conversion_target(self, target_format):
1744
if not target_format.rich_root_data:
1745
raise errors.BadConversionTarget(
1746
'Does not support rich root data.', target_format)
1748
def open(self, a_bzrdir, _found=False, _override_transport=None):
1749
"""See RepositoryFormat.open().
1751
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1752
repository at a slightly different url
1753
than normal. I.e. during 'upgrade'.
1756
format = RepositoryFormat.find_format(a_bzrdir)
1757
assert format.__class__ == self.__class__
1758
if _override_transport is not None:
1759
repo_transport = _override_transport
1761
repo_transport = a_bzrdir.get_repository_transport(None)
1762
control_files = lockable_files.LockableFiles(repo_transport, 'lock',
1764
text_store = self._get_text_store(repo_transport, control_files)
1765
control_store = self._get_control_store(repo_transport, control_files)
1766
_revision_store = self._get_revision_store(repo_transport, control_files)
1767
return KnitRepository2(_format=self,
1769
control_files=control_files,
1770
_revision_store=_revision_store,
1771
control_store=control_store,
1772
text_store=text_store)
1776
# formats which have no format string are not discoverable
1777
# and not independently creatable, so are not registered.
1778
RepositoryFormat.register_format(RepositoryFormat7())
1779
_default_format = RepositoryFormatKnit1()
1780
RepositoryFormat.register_format(_default_format)
1781
RepositoryFormat.register_format(RepositoryFormatKnit2())
1782
RepositoryFormat.set_default_format(_default_format)
1783
_legacy_formats = [RepositoryFormat4(),
1784
RepositoryFormat5(),
1785
RepositoryFormat6()]
1788
class InterRepository(InterObject):
1789
"""This class represents operations taking place between two repositories.
1791
Its instances have methods like copy_content and fetch, and contain
1792
references to the source and target repositories these operations can be
1795
Often we will provide convenience methods on 'repository' which carry out
1796
operations with another repository - they will always forward to
1797
InterRepository.get(other).method_name(parameters).
1801
"""The available optimised InterRepository types."""
1803
def copy_content(self, revision_id=None, basis=None):
1804
raise NotImplementedError(self.copy_content)
1806
def fetch(self, revision_id=None, pb=None):
1807
"""Fetch the content required to construct revision_id.
1809
The content is copied from self.source to self.target.
1811
:param revision_id: if None all content is copied, if NULL_REVISION no
1813
:param pb: optional progress bar to use for progress reports. If not
1814
provided a default one will be created.
1816
Returns the copied revision count and the failed revisions in a tuple:
1819
raise NotImplementedError(self.fetch)
1822
def missing_revision_ids(self, revision_id=None):
1823
"""Return the revision ids that source has that target does not.
1825
These are returned in topological order.
1827
:param revision_id: only return revision ids included by this
1830
# generic, possibly worst case, slow code path.
1831
target_ids = set(self.target.all_revision_ids())
1832
if revision_id is not None:
1833
source_ids = self.source.get_ancestry(revision_id)
1834
assert source_ids[0] is None
1837
source_ids = self.source.all_revision_ids()
1838
result_set = set(source_ids).difference(target_ids)
1839
# this may look like a no-op: its not. It preserves the ordering
1840
# other_ids had while only returning the members from other_ids
1841
# that we've decided we need.
1842
return [rev_id for rev_id in source_ids if rev_id in result_set]
1845
class InterSameDataRepository(InterRepository):
1846
"""Code for converting between repositories that represent the same data.
1848
Data format and model must match for this to work.
1851
_matching_repo_format = RepositoryFormat4()
1852
"""Repository format for testing with."""
1855
def is_compatible(source, target):
1856
if source._format.rich_root_data == target._format.rich_root_data:
1862
def copy_content(self, revision_id=None, basis=None):
1863
"""Make a complete copy of the content in self into destination.
1865
This is a destructive operation! Do not use it on existing
1868
:param revision_id: Only copy the content needed to construct
1869
revision_id and its parents.
1870
:param basis: Copy the needed data preferentially from basis.
1873
self.target.set_make_working_trees(self.source.make_working_trees())
1874
except NotImplementedError:
1876
# grab the basis available data
1877
if basis is not None:
1878
self.target.fetch(basis, revision_id=revision_id)
1879
# but don't bother fetching if we have the needed data now.
1880
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1881
self.target.has_revision(revision_id)):
1883
self.target.fetch(self.source, revision_id=revision_id)
1886
def fetch(self, revision_id=None, pb=None):
1887
"""See InterRepository.fetch()."""
1888
from bzrlib.fetch import GenericRepoFetcher
1889
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1890
self.source, self.source._format, self.target,
1891
self.target._format)
1892
f = GenericRepoFetcher(to_repository=self.target,
1893
from_repository=self.source,
1894
last_revision=revision_id,
1896
return f.count_copied, f.failed_revisions
1899
class InterWeaveRepo(InterSameDataRepository):
1900
"""Optimised code paths between Weave based repositories."""
1902
_matching_repo_format = RepositoryFormat7()
1903
"""Repository format for testing with."""
1906
def is_compatible(source, target):
1907
"""Be compatible with known Weave formats.
1909
We don't test for the stores being of specific types because that
1910
could lead to confusing results, and there is no need to be
1914
return (isinstance(source._format, (RepositoryFormat5,
1916
RepositoryFormat7)) and
1917
isinstance(target._format, (RepositoryFormat5,
1919
RepositoryFormat7)))
1920
except AttributeError:
1924
def copy_content(self, revision_id=None, basis=None):
1925
"""See InterRepository.copy_content()."""
1926
# weave specific optimised path:
1927
if basis is not None:
1928
# copy the basis in, then fetch remaining data.
1929
basis.copy_content_into(self.target, revision_id)
1930
# the basis copy_content_into could miss-set this.
1932
self.target.set_make_working_trees(self.source.make_working_trees())
1933
except NotImplementedError:
1935
self.target.fetch(self.source, revision_id=revision_id)
1938
self.target.set_make_working_trees(self.source.make_working_trees())
1939
except NotImplementedError:
1941
# FIXME do not peek!
1942
if self.source.control_files._transport.listable():
1943
pb = ui.ui_factory.nested_progress_bar()
1945
self.target.weave_store.copy_all_ids(
1946
self.source.weave_store,
1948
from_transaction=self.source.get_transaction(),
1949
to_transaction=self.target.get_transaction())
1950
pb.update('copying inventory', 0, 1)
1951
self.target.control_weaves.copy_multi(
1952
self.source.control_weaves, ['inventory'],
1953
from_transaction=self.source.get_transaction(),
1954
to_transaction=self.target.get_transaction())
1955
self.target._revision_store.text_store.copy_all_ids(
1956
self.source._revision_store.text_store,
1961
self.target.fetch(self.source, revision_id=revision_id)
1964
def fetch(self, revision_id=None, pb=None):
1965
"""See InterRepository.fetch()."""
1966
from bzrlib.fetch import GenericRepoFetcher
1967
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1968
self.source, self.source._format, self.target, self.target._format)
1969
f = GenericRepoFetcher(to_repository=self.target,
1970
from_repository=self.source,
1971
last_revision=revision_id,
1973
return f.count_copied, f.failed_revisions
1976
def missing_revision_ids(self, revision_id=None):
1977
"""See InterRepository.missing_revision_ids()."""
1978
# we want all revisions to satisfy revision_id in source.
1979
# but we don't want to stat every file here and there.
1980
# we want then, all revisions other needs to satisfy revision_id
1981
# checked, but not those that we have locally.
1982
# so the first thing is to get a subset of the revisions to
1983
# satisfy revision_id in source, and then eliminate those that
1984
# we do already have.
1985
# this is slow on high latency connection to self, but as as this
1986
# disk format scales terribly for push anyway due to rewriting
1987
# inventory.weave, this is considered acceptable.
1989
if revision_id is not None:
1990
source_ids = self.source.get_ancestry(revision_id)
1991
assert source_ids[0] is None
1994
source_ids = self.source._all_possible_ids()
1995
source_ids_set = set(source_ids)
1996
# source_ids is the worst possible case we may need to pull.
1997
# now we want to filter source_ids against what we actually
1998
# have in target, but don't try to check for existence where we know
1999
# we do not have a revision as that would be pointless.
2000
target_ids = set(self.target._all_possible_ids())
2001
possibly_present_revisions = target_ids.intersection(source_ids_set)
2002
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2003
required_revisions = source_ids_set.difference(actually_present_revisions)
2004
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2005
if revision_id is not None:
2006
# we used get_ancestry to determine source_ids then we are assured all
2007
# revisions referenced are present as they are installed in topological order.
2008
# and the tip revision was validated by get_ancestry.
2009
return required_topo_revisions
2011
# if we just grabbed the possibly available ids, then
2012
# we only have an estimate of whats available and need to validate
2013
# that against the revision records.
2014
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2017
class InterKnitRepo(InterSameDataRepository):
2018
"""Optimised code paths between Knit based repositories."""
2020
_matching_repo_format = RepositoryFormatKnit1()
2021
"""Repository format for testing with."""
2024
def is_compatible(source, target):
2025
"""Be compatible with known Knit formats.
2027
We don't test for the stores being of specific types because that
2028
could lead to confusing results, and there is no need to be
2032
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2033
isinstance(target._format, (RepositoryFormatKnit1)))
2034
except AttributeError:
2038
def fetch(self, revision_id=None, pb=None):
2039
"""See InterRepository.fetch()."""
2040
from bzrlib.fetch import KnitRepoFetcher
2041
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2042
self.source, self.source._format, self.target, self.target._format)
2043
f = KnitRepoFetcher(to_repository=self.target,
2044
from_repository=self.source,
2045
last_revision=revision_id,
2047
return f.count_copied, f.failed_revisions
2050
def missing_revision_ids(self, revision_id=None):
2051
"""See InterRepository.missing_revision_ids()."""
2052
if revision_id is not None:
2053
source_ids = self.source.get_ancestry(revision_id)
2054
assert source_ids[0] is None
2057
source_ids = self.source._all_possible_ids()
2058
source_ids_set = set(source_ids)
2059
# source_ids is the worst possible case we may need to pull.
2060
# now we want to filter source_ids against what we actually
2061
# have in target, but don't try to check for existence where we know
2062
# we do not have a revision as that would be pointless.
2063
target_ids = set(self.target._all_possible_ids())
2064
possibly_present_revisions = target_ids.intersection(source_ids_set)
2065
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2066
required_revisions = source_ids_set.difference(actually_present_revisions)
2067
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2068
if revision_id is not None:
2069
# we used get_ancestry to determine source_ids then we are assured all
2070
# revisions referenced are present as they are installed in topological order.
2071
# and the tip revision was validated by get_ancestry.
2072
return required_topo_revisions
2074
# if we just grabbed the possibly available ids, then
2075
# we only have an estimate of whats available and need to validate
2076
# that against the revision records.
2077
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2080
class InterModel1and2(InterRepository):
2082
_matching_repo_format = None
2085
def is_compatible(source, target):
2086
if not isinstance(source, Repository):
2088
if not isinstance(target, Repository):
2090
if not source._format.rich_root_data and target._format.rich_root_data:
2096
def fetch(self, revision_id=None, pb=None):
2097
"""See InterRepository.fetch()."""
2098
from bzrlib.fetch import Model1toKnit2Fetcher
2099
f = Model1toKnit2Fetcher(to_repository=self.target,
2100
from_repository=self.source,
2101
last_revision=revision_id,
2103
return f.count_copied, f.failed_revisions
2106
def copy_content(self, revision_id=None, basis=None):
2107
"""Make a complete copy of the content in self into destination.
2109
This is a destructive operation! Do not use it on existing
2112
:param revision_id: Only copy the content needed to construct
2113
revision_id and its parents.
2114
:param basis: Copy the needed data preferentially from basis.
2117
self.target.set_make_working_trees(self.source.make_working_trees())
2118
except NotImplementedError:
2120
# grab the basis available data
2121
if basis is not None:
2122
self.target.fetch(basis, revision_id=revision_id)
2123
# but don't bother fetching if we have the needed data now.
2124
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2125
self.target.has_revision(revision_id)):
2127
self.target.fetch(self.source, revision_id=revision_id)
2130
class InterKnit1and2(InterKnitRepo):
2132
_matching_repo_format = None
2135
def is_compatible(source, target):
2136
"""Be compatible with Knit1 source and Knit2 target"""
2138
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2139
isinstance(target._format, (RepositoryFormatKnit2)))
2140
except AttributeError:
2144
def fetch(self, revision_id=None, pb=None):
2145
"""See InterRepository.fetch()."""
2146
from bzrlib.fetch import Knit1to2Fetcher
2147
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2148
self.source, self.source._format, self.target,
2149
self.target._format)
2150
f = Knit1to2Fetcher(to_repository=self.target,
2151
from_repository=self.source,
2152
last_revision=revision_id,
2154
return f.count_copied, f.failed_revisions
2157
InterRepository.register_optimiser(InterSameDataRepository)
2158
InterRepository.register_optimiser(InterWeaveRepo)
2159
InterRepository.register_optimiser(InterKnitRepo)
2160
InterRepository.register_optimiser(InterModel1and2)
2161
InterRepository.register_optimiser(InterKnit1and2)
2164
class RepositoryTestProviderAdapter(object):
2165
"""A tool to generate a suite testing multiple repository formats at once.
2167
This is done by copying the test once for each transport and injecting
2168
the transport_server, transport_readonly_server, and bzrdir_format and
2169
repository_format classes into each copy. Each copy is also given a new id()
2170
to make it easy to identify.
2173
def __init__(self, transport_server, transport_readonly_server, formats):
2174
self._transport_server = transport_server
2175
self._transport_readonly_server = transport_readonly_server
2176
self._formats = formats
2178
def adapt(self, test):
2179
result = unittest.TestSuite()
2180
for repository_format, bzrdir_format in self._formats:
2181
new_test = deepcopy(test)
2182
new_test.transport_server = self._transport_server
2183
new_test.transport_readonly_server = self._transport_readonly_server
2184
new_test.bzrdir_format = bzrdir_format
2185
new_test.repository_format = repository_format
2186
def make_new_test_id():
2187
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
2188
return lambda: new_id
2189
new_test.id = make_new_test_id()
2190
result.addTest(new_test)
2194
class InterRepositoryTestProviderAdapter(object):
2195
"""A tool to generate a suite testing multiple inter repository formats.
2197
This is done by copying the test once for each interrepo provider and injecting
2198
the transport_server, transport_readonly_server, repository_format and
2199
repository_to_format classes into each copy.
2200
Each copy is also given a new id() to make it easy to identify.
2203
def __init__(self, transport_server, transport_readonly_server, formats):
2204
self._transport_server = transport_server
2205
self._transport_readonly_server = transport_readonly_server
2206
self._formats = formats
2208
def adapt(self, test):
2209
result = unittest.TestSuite()
2210
for interrepo_class, repository_format, repository_format_to in self._formats:
2211
new_test = deepcopy(test)
2212
new_test.transport_server = self._transport_server
2213
new_test.transport_readonly_server = self._transport_readonly_server
2214
new_test.interrepo_class = interrepo_class
2215
new_test.repository_format = repository_format
2216
new_test.repository_format_to = repository_format_to
2217
def make_new_test_id():
2218
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
2219
return lambda: new_id
2220
new_test.id = make_new_test_id()
2221
result.addTest(new_test)
2225
def default_test_list():
2226
"""Generate the default list of interrepo permutations to test."""
2228
# test the default InterRepository between format 6 and the current
2230
# XXX: robertc 20060220 reinstate this when there are two supported
2231
# formats which do not have an optimal code path between them.
2232
#result.append((InterRepository,
2233
# RepositoryFormat6(),
2234
# RepositoryFormatKnit1()))
2235
for optimiser in InterRepository._optimisers:
2236
if optimiser._matching_repo_format is not None:
2237
result.append((optimiser,
2238
optimiser._matching_repo_format,
2239
optimiser._matching_repo_format
2241
# if there are specific combinations we want to use, we can add them
2243
result.append((InterModel1and2, RepositoryFormat5(),
2244
RepositoryFormatKnit2()))
2245
result.append((InterKnit1and2, RepositoryFormatKnit1(),
2246
RepositoryFormatKnit2()))
2250
class CopyConverter(object):
2251
"""A repository conversion tool which just performs a copy of the content.
2253
This is slow but quite reliable.
2256
def __init__(self, target_format):
2257
"""Create a CopyConverter.
2259
:param target_format: The format the resulting repository should be.
2261
self.target_format = target_format
2263
def convert(self, repo, pb):
2264
"""Perform the conversion of to_convert, giving feedback via pb.
2266
:param to_convert: The disk object to convert.
2267
:param pb: a progress bar to use for progress information.
2272
# this is only useful with metadir layouts - separated repo content.
2273
# trigger an assertion if not such
2274
repo._format.get_format_string()
2275
self.repo_dir = repo.bzrdir
2276
self.step('Moving repository to repository.backup')
2277
self.repo_dir.transport.move('repository', 'repository.backup')
2278
backup_transport = self.repo_dir.transport.clone('repository.backup')
2279
repo._format.check_conversion_target(self.target_format)
2280
self.source_repo = repo._format.open(self.repo_dir,
2282
_override_transport=backup_transport)
2283
self.step('Creating new repository')
2284
converted = self.target_format.initialize(self.repo_dir,
2285
self.source_repo.is_shared())
2286
converted.lock_write()
2288
self.step('Copying content into repository.')
2289
self.source_repo.copy_content_into(converted)
2292
self.step('Deleting old repository content.')
2293
self.repo_dir.transport.delete_tree('repository.backup')
2294
self.pb.note('repository converted')
2296
def step(self, message):
2297
"""Update the pb by a step."""
2299
self.pb.update(message, self.count, self.total)
2302
class CommitBuilder(object):
2303
"""Provides an interface to build up a commit.
2305
This allows describing a tree to be committed without needing to
2306
know the internals of the format of the repository.
2309
record_root_entry = False
2310
def __init__(self, repository, parents, config, timestamp=None,
2311
timezone=None, committer=None, revprops=None,
2313
"""Initiate a CommitBuilder.
2315
:param repository: Repository to commit to.
2316
:param parents: Revision ids of the parents of the new revision.
2317
:param config: Configuration to use.
2318
:param timestamp: Optional timestamp recorded for commit.
2319
:param timezone: Optional timezone for timestamp.
2320
:param committer: Optional committer to set for commit.
2321
:param revprops: Optional dictionary of revision properties.
2322
:param revision_id: Optional revision id.
2324
self._config = config
2326
if committer is None:
2327
self._committer = self._config.username()
2329
assert isinstance(committer, basestring), type(committer)
2330
self._committer = committer
2332
self.new_inventory = Inventory(None)
2333
self._new_revision_id = revision_id
2334
self.parents = parents
2335
self.repository = repository
2338
if revprops is not None:
2339
self._revprops.update(revprops)
2341
if timestamp is None:
2342
timestamp = time.time()
2343
# Restrict resolution to 1ms
2344
self._timestamp = round(timestamp, 3)
2346
if timezone is None:
2347
self._timezone = local_time_offset()
2349
self._timezone = int(timezone)
2351
self._generate_revision_if_needed()
2353
def commit(self, message):
2354
"""Make the actual commit.
2356
:return: The revision id of the recorded revision.
2358
rev = _mod_revision.Revision(
2359
timestamp=self._timestamp,
2360
timezone=self._timezone,
2361
committer=self._committer,
2363
inventory_sha1=self.inv_sha1,
2364
revision_id=self._new_revision_id,
2365
properties=self._revprops)
2366
rev.parent_ids = self.parents
2367
self.repository.add_revision(self._new_revision_id, rev,
2368
self.new_inventory, self._config)
2369
return self._new_revision_id
2371
def revision_tree(self):
2372
"""Return the tree that was just committed.
2374
After calling commit() this can be called to get a RevisionTree
2375
representing the newly committed tree. This is preferred to
2376
calling Repository.revision_tree() because that may require
2377
deserializing the inventory, while we already have a copy in
2380
return RevisionTree(self.repository, self.new_inventory,
2381
self._new_revision_id)
2383
def finish_inventory(self):
2384
"""Tell the builder that the inventory is finished."""
2385
if self.new_inventory.root is None:
2386
symbol_versioning.warn('Root entry should be supplied to'
2387
' record_entry_contents, as of bzr 0.10.',
2388
DeprecationWarning, stacklevel=2)
2389
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
2390
self.new_inventory.revision_id = self._new_revision_id
2391
self.inv_sha1 = self.repository.add_inventory(
2392
self._new_revision_id,
2397
def _gen_revision_id(self):
2398
"""Return new revision-id."""
2399
return generate_ids.gen_revision_id(self._config.username(),
2402
def _generate_revision_if_needed(self):
2403
"""Create a revision id if None was supplied.
2405
If the repository can not support user-specified revision ids
2406
they should override this function and raise CannotSetRevisionId
2407
if _new_revision_id is not None.
2409
:raises: CannotSetRevisionId
2411
if self._new_revision_id is None:
2412
self._new_revision_id = self._gen_revision_id()
2414
def record_entry_contents(self, ie, parent_invs, path, tree):
2415
"""Record the content of ie from tree into the commit if needed.
2417
Side effect: sets ie.revision when unchanged
2419
:param ie: An inventory entry present in the commit.
2420
:param parent_invs: The inventories of the parent revisions of the
2422
:param path: The path the entry is at in the tree.
2423
:param tree: The tree which contains this entry and should be used to
2426
if self.new_inventory.root is None and ie.parent_id is not None:
2427
symbol_versioning.warn('Root entry should be supplied to'
2428
' record_entry_contents, as of bzr 0.10.',
2429
DeprecationWarning, stacklevel=2)
2430
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
2432
self.new_inventory.add(ie)
2434
# ie.revision is always None if the InventoryEntry is considered
2435
# for committing. ie.snapshot will record the correct revision
2436
# which may be the sole parent if it is untouched.
2437
if ie.revision is not None:
2440
# In this revision format, root entries have no knit or weave
2441
if ie is self.new_inventory.root:
2442
# When serializing out to disk and back in
2443
# root.revision is always _new_revision_id
2444
ie.revision = self._new_revision_id
2446
previous_entries = ie.find_previous_heads(
2448
self.repository.weave_store,
2449
self.repository.get_transaction())
2450
# we are creating a new revision for ie in the history store
2452
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2454
def modified_directory(self, file_id, file_parents):
2455
"""Record the presence of a symbolic link.
2457
:param file_id: The file_id of the link to record.
2458
:param file_parents: The per-file parent revision ids.
2460
self._add_text_to_weave(file_id, [], file_parents.keys())
2462
def modified_file_text(self, file_id, file_parents,
2463
get_content_byte_lines, text_sha1=None,
2465
"""Record the text of file file_id
2467
:param file_id: The file_id of the file to record the text of.
2468
:param file_parents: The per-file parent revision ids.
2469
:param get_content_byte_lines: A callable which will return the byte
2471
:param text_sha1: Optional SHA1 of the file contents.
2472
:param text_size: Optional size of the file contents.
2474
# mutter('storing text of file {%s} in revision {%s} into %r',
2475
# file_id, self._new_revision_id, self.repository.weave_store)
2476
# special case to avoid diffing on renames or
2478
if (len(file_parents) == 1
2479
and text_sha1 == file_parents.values()[0].text_sha1
2480
and text_size == file_parents.values()[0].text_size):
2481
previous_ie = file_parents.values()[0]
2482
versionedfile = self.repository.weave_store.get_weave(file_id,
2483
self.repository.get_transaction())
2484
versionedfile.clone_text(self._new_revision_id,
2485
previous_ie.revision, file_parents.keys())
2486
return text_sha1, text_size
2488
new_lines = get_content_byte_lines()
2489
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2490
# should return the SHA1 and size
2491
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2492
return osutils.sha_strings(new_lines), \
2493
sum(map(len, new_lines))
2495
def modified_link(self, file_id, file_parents, link_target):
2496
"""Record the presence of a symbolic link.
2498
:param file_id: The file_id of the link to record.
2499
:param file_parents: The per-file parent revision ids.
2500
:param link_target: Target location of this link.
2502
self._add_text_to_weave(file_id, [], file_parents.keys())
2504
def _add_text_to_weave(self, file_id, new_lines, parents):
2505
versionedfile = self.repository.weave_store.get_weave_or_empty(
2506
file_id, self.repository.get_transaction())
2507
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2508
versionedfile.clear_cache()
2511
class _CommitBuilder(CommitBuilder):
2512
"""Temporary class so old CommitBuilders are detected properly
2514
Note: CommitBuilder works whether or not root entry is recorded.
2517
record_root_entry = True
2520
class RootCommitBuilder(CommitBuilder):
2521
"""This commitbuilder actually records the root id"""
2523
record_root_entry = True
2525
def record_entry_contents(self, ie, parent_invs, path, tree):
2526
"""Record the content of ie from tree into the commit if needed.
2528
Side effect: sets ie.revision when unchanged
2530
:param ie: An inventory entry present in the commit.
2531
:param parent_invs: The inventories of the parent revisions of the
2533
:param path: The path the entry is at in the tree.
2534
:param tree: The tree which contains this entry and should be used to
2537
assert self.new_inventory.root is not None or ie.parent_id is None
2538
self.new_inventory.add(ie)
2540
# ie.revision is always None if the InventoryEntry is considered
2541
# for committing. ie.snapshot will record the correct revision
2542
# which may be the sole parent if it is untouched.
2543
if ie.revision is not None:
2546
previous_entries = ie.find_previous_heads(
2548
self.repository.weave_store,
2549
self.repository.get_transaction())
2550
# we are creating a new revision for ie in the history store
2552
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2564
def _unescaper(match, _map=_unescape_map):
2565
return _map[match.group(1)]
2571
def _unescape_xml(data):
2572
"""Unescape predefined XML entities in a string of data."""
2574
if _unescape_re is None:
2575
_unescape_re = re.compile('\&([^;]*);')
2576
return _unescape_re.sub(_unescaper, data)
2579
def _unescape_xml_cached(data, cache):
2583
unescaped = _unescape_xml(data)
2584
cache[data] = unescaped
added
removed
bzrlib/repository.py
