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 binascii import hexlify
18
from copy import deepcopy
19
from cStringIO import StringIO
22
from unittest import TestSuite
36
from bzrlib.decorators import needs_read_lock, needs_write_lock
37
from bzrlib.errors import InvalidRevisionId
38
from bzrlib.graph import Graph
39
from bzrlib.inter import InterObject
40
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
41
from bzrlib.knit import KnitVersionedFile, KnitPlainFactory
42
from bzrlib.lockable_files import LockableFiles, TransportLock
43
from bzrlib.lockdir import LockDir
44
from bzrlib.osutils import (safe_unicode, rand_bytes, compact_date,
46
from bzrlib.revision import NULL_REVISION, Revision
47
from bzrlib.revisiontree import RevisionTree
48
from bzrlib.store.versioned import VersionedFileStore, WeaveStore
49
from bzrlib.store.text import TextStore
50
from bzrlib import symbol_versioning
51
from bzrlib.symbol_versioning import (deprecated_method,
54
from bzrlib.testament import Testament
55
from bzrlib.trace import mutter, note, warning
56
from bzrlib.tsort import topo_sort
57
from bzrlib.weave import WeaveFile
60
# Old formats display a warning, but only once
61
_deprecation_warning_done = False
64
class Repository(object):
65
"""Repository holding history for one or more branches.
67
The repository holds and retrieves historical information including
68
revisions and file history. It's normally accessed only by the Branch,
69
which views a particular line of development through that history.
71
The Repository builds on top of Stores and a Transport, which respectively
72
describe the disk data format and the way of accessing the (possibly
77
def add_inventory(self, revid, inv, parents):
78
"""Add the inventory inv to the repository as revid.
80
:param parents: The revision ids of the parents that revid
81
is known to have and are in the repository already.
83
returns the sha1 of the serialized inventory.
85
assert inv.revision_id is None or inv.revision_id == revid, \
86
"Mismatch between inventory revision" \
87
" id and insertion revid (%r, %r)" % (inv.revision_id, revid)
88
assert inv.root is not None
89
inv_text = self.serialise_inventory(inv)
90
inv_sha1 = osutils.sha_string(inv_text)
91
inv_vf = self.control_weaves.get_weave('inventory',
92
self.get_transaction())
93
self._inventory_add_lines(inv_vf, revid, parents, osutils.split_lines(inv_text))
96
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
98
for parent in parents:
100
final_parents.append(parent)
102
inv_vf.add_lines(revid, final_parents, lines)
105
def add_revision(self, rev_id, rev, inv=None, config=None):
106
"""Add rev to the revision store as rev_id.
108
:param rev_id: the revision id to use.
109
:param rev: The revision object.
110
:param inv: The inventory for the revision. if None, it will be looked
111
up in the inventory storer
112
:param config: If None no digital signature will be created.
113
If supplied its signature_needed method will be used
114
to determine if a signature should be made.
116
if config is not None and config.signature_needed():
118
inv = self.get_inventory(rev_id)
119
plaintext = Testament(rev, inv).as_short_text()
120
self.store_revision_signature(
121
gpg.GPGStrategy(config), plaintext, rev_id)
122
if not rev_id in self.get_inventory_weave():
124
raise errors.WeaveRevisionNotPresent(rev_id,
125
self.get_inventory_weave())
127
# yes, this is not suitable for adding with ghosts.
128
self.add_inventory(rev_id, inv, rev.parent_ids)
129
self._revision_store.add_revision(rev, self.get_transaction())
132
def _all_possible_ids(self):
133
"""Return all the possible revisions that we could find."""
134
return self.get_inventory_weave().versions()
136
def all_revision_ids(self):
137
"""Returns a list of all the revision ids in the repository.
139
This is deprecated because code should generally work on the graph
140
reachable from a particular revision, and ignore any other revisions
141
that might be present. There is no direct replacement method.
143
return self._all_revision_ids()
146
def _all_revision_ids(self):
147
"""Returns a list of all the revision ids in the repository.
149
These are in as much topological order as the underlying store can
150
present: for weaves ghosts may lead to a lack of correctness until
151
the reweave updates the parents list.
153
if self._revision_store.text_store.listable():
154
return self._revision_store.all_revision_ids(self.get_transaction())
155
result = self._all_possible_ids()
156
return self._eliminate_revisions_not_present(result)
158
def break_lock(self):
159
"""Break a lock if one is present from another instance.
161
Uses the ui factory to ask for confirmation if the lock may be from
164
self.control_files.break_lock()
167
def _eliminate_revisions_not_present(self, revision_ids):
168
"""Check every revision id in revision_ids to see if we have it.
170
Returns a set of the present revisions.
173
for id in revision_ids:
174
if self.has_revision(id):
179
def create(a_bzrdir):
180
"""Construct the current default format repository in a_bzrdir."""
181
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
183
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
184
"""instantiate a Repository.
186
:param _format: The format of the repository on disk.
187
:param a_bzrdir: The BzrDir of the repository.
189
In the future we will have a single api for all stores for
190
getting file texts, inventories and revisions, then
191
this construct will accept instances of those things.
193
super(Repository, self).__init__()
194
self._format = _format
195
# the following are part of the public API for Repository:
196
self.bzrdir = a_bzrdir
197
self.control_files = control_files
198
self._revision_store = _revision_store
199
self.text_store = text_store
200
# backwards compatibility
201
self.weave_store = text_store
202
# not right yet - should be more semantically clear ?
204
self.control_store = control_store
205
self.control_weaves = control_store
206
# TODO: make sure to construct the right store classes, etc, depending
207
# on whether escaping is required.
208
self._warn_if_deprecated()
209
self._serializer = xml5.serializer_v5
212
return '%s(%r)' % (self.__class__.__name__,
213
self.bzrdir.transport.base)
216
return self.control_files.is_locked()
218
def lock_write(self):
219
self.control_files.lock_write()
222
self.control_files.lock_read()
224
def get_physical_lock_status(self):
225
return self.control_files.get_physical_lock_status()
228
def missing_revision_ids(self, other, revision_id=None):
229
"""Return the revision ids that other has that this does not.
231
These are returned in topological order.
233
revision_id: only return revision ids included by revision_id.
235
return InterRepository.get(other, self).missing_revision_ids(revision_id)
239
"""Open the repository rooted at base.
241
For instance, if the repository is at URL/.bzr/repository,
242
Repository.open(URL) -> a Repository instance.
244
control = bzrdir.BzrDir.open(base)
245
return control.open_repository()
247
def copy_content_into(self, destination, revision_id=None, basis=None):
248
"""Make a complete copy of the content in self into destination.
250
This is a destructive operation! Do not use it on existing
253
return InterRepository.get(self, destination).copy_content(revision_id, basis)
255
def fetch(self, source, revision_id=None, pb=None):
256
"""Fetch the content required to construct revision_id from source.
258
If revision_id is None all content is copied.
260
return InterRepository.get(source, self).fetch(revision_id=revision_id,
263
def get_commit_builder(self, branch, parents, config, timestamp=None,
264
timezone=None, committer=None, revprops=None,
266
"""Obtain a CommitBuilder for this repository.
268
:param branch: Branch to commit to.
269
:param parents: Revision ids of the parents of the new revision.
270
:param config: Configuration to use.
271
:param timestamp: Optional timestamp recorded for commit.
272
:param timezone: Optional timezone for timestamp.
273
:param committer: Optional committer to set for commit.
274
:param revprops: Optional dictionary of revision properties.
275
:param revision_id: Optional revision id.
277
return _CommitBuilder(self, parents, config, timestamp, timezone,
278
committer, revprops, revision_id)
281
self.control_files.unlock()
284
def clone(self, a_bzrdir, revision_id=None, basis=None):
285
"""Clone this repository into a_bzrdir using the current format.
287
Currently no check is made that the format of this repository and
288
the bzrdir format are compatible. FIXME RBC 20060201.
290
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
291
# use target default format.
292
result = a_bzrdir.create_repository()
293
# FIXME RBC 20060209 split out the repository type to avoid this check ?
294
elif isinstance(a_bzrdir._format,
295
(bzrdir.BzrDirFormat4,
296
bzrdir.BzrDirFormat5,
297
bzrdir.BzrDirFormat6)):
298
result = a_bzrdir.open_repository()
300
result = self._format.initialize(a_bzrdir, shared=self.is_shared())
301
self.copy_content_into(result, revision_id, basis)
305
def has_revision(self, revision_id):
306
"""True if this repository has a copy of the revision."""
307
return self._revision_store.has_revision_id(revision_id,
308
self.get_transaction())
311
def get_revision_reconcile(self, revision_id):
312
"""'reconcile' helper routine that allows access to a revision always.
314
This variant of get_revision does not cross check the weave graph
315
against the revision one as get_revision does: but it should only
316
be used by reconcile, or reconcile-alike commands that are correcting
317
or testing the revision graph.
319
if not revision_id or not isinstance(revision_id, basestring):
320
raise InvalidRevisionId(revision_id=revision_id, branch=self)
321
return self._revision_store.get_revisions([revision_id],
322
self.get_transaction())[0]
324
def get_revisions(self, revision_ids):
325
return self._revision_store.get_revisions(revision_ids,
326
self.get_transaction())
329
def get_revision_xml(self, revision_id):
330
rev = self.get_revision(revision_id)
332
# the current serializer..
333
self._revision_store._serializer.write_revision(rev, rev_tmp)
335
return rev_tmp.getvalue()
338
def get_revision(self, revision_id):
339
"""Return the Revision object for a named revision"""
340
r = self.get_revision_reconcile(revision_id)
341
# weave corruption can lead to absent revision markers that should be
343
# the following test is reasonably cheap (it needs a single weave read)
344
# and the weave is cached in read transactions. In write transactions
345
# it is not cached but typically we only read a small number of
346
# revisions. For knits when they are introduced we will probably want
347
# to ensure that caching write transactions are in use.
348
inv = self.get_inventory_weave()
349
self._check_revision_parents(r, inv)
353
def get_deltas_for_revisions(self, revisions):
354
"""Produce a generator of revision deltas.
356
Note that the input is a sequence of REVISIONS, not revision_ids.
357
Trees will be held in memory until the generator exits.
358
Each delta is relative to the revision's lefthand predecessor.
360
required_trees = set()
361
for revision in revisions:
362
required_trees.add(revision.revision_id)
363
required_trees.update(revision.parent_ids[:1])
364
trees = dict((t.get_revision_id(), t) for
365
t in self.revision_trees(required_trees))
366
for revision in revisions:
367
if not revision.parent_ids:
368
old_tree = self.revision_tree(None)
370
old_tree = trees[revision.parent_ids[0]]
371
yield trees[revision.revision_id].changes_from(old_tree)
374
def get_revision_delta(self, revision_id):
375
"""Return the delta for one revision.
377
The delta is relative to the left-hand predecessor of the
380
r = self.get_revision(revision_id)
381
return list(self.get_deltas_for_revisions([r]))[0]
383
def _check_revision_parents(self, revision, inventory):
384
"""Private to Repository and Fetch.
386
This checks the parentage of revision in an inventory weave for
387
consistency and is only applicable to inventory-weave-for-ancestry
388
using repository formats & fetchers.
390
weave_parents = inventory.get_parents(revision.revision_id)
391
weave_names = inventory.versions()
392
for parent_id in revision.parent_ids:
393
if parent_id in weave_names:
394
# this parent must not be a ghost.
395
if not parent_id in weave_parents:
397
raise errors.CorruptRepository(self)
400
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
401
signature = gpg_strategy.sign(plaintext)
402
self._revision_store.add_revision_signature_text(revision_id,
404
self.get_transaction())
406
def fileids_altered_by_revision_ids(self, revision_ids):
407
"""Find the file ids and versions affected by revisions.
409
:param revisions: an iterable containing revision ids.
410
:return: a dictionary mapping altered file-ids to an iterable of
411
revision_ids. Each altered file-ids has the exact revision_ids that
412
altered it listed explicitly.
414
assert self._serializer.support_altered_by_hack, \
415
("fileids_altered_by_revision_ids only supported for branches "
416
"which store inventory as unnested xml, not on %r" % self)
417
selected_revision_ids = set(revision_ids)
418
w = self.get_inventory_weave()
421
# this code needs to read every new line in every inventory for the
422
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
423
# not present in one of those inventories is unnecessary but not
424
# harmful because we are filtering by the revision id marker in the
425
# inventory lines : we only select file ids altered in one of those
426
# revisions. We don't need to see all lines in the inventory because
427
# only those added in an inventory in rev X can contain a revision=X
429
pb = ui.ui_factory.nested_progress_bar()
431
for line in w.iter_lines_added_or_present_in_versions(
432
selected_revision_ids, pb=pb):
433
start = line.find('file_id="')+9
434
if start < 9: continue
435
end = line.find('"', start)
437
file_id = _unescape_xml(line[start:end])
439
start = line.find('revision="')+10
440
if start < 10: continue
441
end = line.find('"', start)
443
revision_id = _unescape_xml(line[start:end])
444
if revision_id in selected_revision_ids:
445
result.setdefault(file_id, set()).add(revision_id)
451
def get_inventory_weave(self):
452
return self.control_weaves.get_weave('inventory',
453
self.get_transaction())
456
def get_inventory(self, revision_id):
457
"""Get Inventory object by hash."""
458
return self.deserialise_inventory(
459
revision_id, self.get_inventory_xml(revision_id))
461
def deserialise_inventory(self, revision_id, xml):
462
"""Transform the xml into an inventory object.
464
:param revision_id: The expected revision id of the inventory.
465
:param xml: A serialised inventory.
467
result = self._serializer.read_inventory_from_string(xml)
468
result.root.revision = revision_id
471
def serialise_inventory(self, inv):
472
return self._serializer.write_inventory_to_string(inv)
475
def get_inventory_xml(self, revision_id):
476
"""Get inventory XML as a file object."""
478
assert isinstance(revision_id, basestring), type(revision_id)
479
iw = self.get_inventory_weave()
480
return iw.get_text(revision_id)
482
raise errors.HistoryMissing(self, 'inventory', revision_id)
485
def get_inventory_sha1(self, revision_id):
486
"""Return the sha1 hash of the inventory entry
488
return self.get_revision(revision_id).inventory_sha1
491
def get_revision_graph(self, revision_id=None):
492
"""Return a dictionary containing the revision graph.
494
:param revision_id: The revision_id to get a graph from. If None, then
495
the entire revision graph is returned. This is a deprecated mode of
496
operation and will be removed in the future.
497
:return: a dictionary of revision_id->revision_parents_list.
499
# special case NULL_REVISION
500
if revision_id == NULL_REVISION:
502
weave = self.get_inventory_weave()
503
all_revisions = self._eliminate_revisions_not_present(weave.versions())
504
entire_graph = dict([(node, weave.get_parents(node)) for
505
node in all_revisions])
506
if revision_id is None:
508
elif revision_id not in entire_graph:
509
raise errors.NoSuchRevision(self, revision_id)
511
# add what can be reached from revision_id
513
pending = set([revision_id])
514
while len(pending) > 0:
516
result[node] = entire_graph[node]
517
for revision_id in result[node]:
518
if revision_id not in result:
519
pending.add(revision_id)
523
def get_revision_graph_with_ghosts(self, revision_ids=None):
524
"""Return a graph of the revisions with ghosts marked as applicable.
526
:param revision_ids: an iterable of revisions to graph or None for all.
527
:return: a Graph object with the graph reachable from revision_ids.
531
pending = set(self.all_revision_ids())
534
pending = set(revision_ids)
535
# special case NULL_REVISION
536
if NULL_REVISION in pending:
537
pending.remove(NULL_REVISION)
538
required = set(pending)
541
revision_id = pending.pop()
543
rev = self.get_revision(revision_id)
544
except errors.NoSuchRevision:
545
if revision_id in required:
548
result.add_ghost(revision_id)
550
for parent_id in rev.parent_ids:
551
# is this queued or done ?
552
if (parent_id not in pending and
553
parent_id not in done):
555
pending.add(parent_id)
556
result.add_node(revision_id, rev.parent_ids)
557
done.add(revision_id)
561
def get_revision_inventory(self, revision_id):
562
"""Return inventory of a past revision."""
563
# TODO: Unify this with get_inventory()
564
# bzr 0.0.6 and later imposes the constraint that the inventory_id
565
# must be the same as its revision, so this is trivial.
566
if revision_id is None:
567
# This does not make sense: if there is no revision,
568
# then it is the current tree inventory surely ?!
569
# and thus get_root_id() is something that looks at the last
570
# commit on the branch, and the get_root_id is an inventory check.
571
raise NotImplementedError
572
# return Inventory(self.get_root_id())
574
return self.get_inventory(revision_id)
578
"""Return True if this repository is flagged as a shared repository."""
579
raise NotImplementedError(self.is_shared)
582
def reconcile(self, other=None, thorough=False):
583
"""Reconcile this repository."""
584
from bzrlib.reconcile import RepoReconciler
585
reconciler = RepoReconciler(self, thorough=thorough)
586
reconciler.reconcile()
590
def revision_tree(self, revision_id):
591
"""Return Tree for a revision on this branch.
593
`revision_id` may be None for the empty tree revision.
595
# TODO: refactor this to use an existing revision object
596
# so we don't need to read it in twice.
597
if revision_id is None or revision_id == NULL_REVISION:
598
return RevisionTree(self, Inventory(), NULL_REVISION)
600
inv = self.get_revision_inventory(revision_id)
601
return RevisionTree(self, inv, revision_id)
604
def revision_trees(self, revision_ids):
605
"""Return Tree for a revision on this branch.
607
`revision_id` may not be None or 'null:'"""
608
assert None not in revision_ids
609
assert NULL_REVISION not in revision_ids
610
texts = self.get_inventory_weave().get_texts(revision_ids)
611
for text, revision_id in zip(texts, revision_ids):
612
inv = self.deserialise_inventory(revision_id, text)
613
yield RevisionTree(self, inv, revision_id)
616
def get_ancestry(self, revision_id):
617
"""Return a list of revision-ids integrated by a revision.
619
The first element of the list is always None, indicating the origin
620
revision. This might change when we have history horizons, or
621
perhaps we should have a new API.
623
This is topologically sorted.
625
if revision_id is None:
627
if not self.has_revision(revision_id):
628
raise errors.NoSuchRevision(self, revision_id)
629
w = self.get_inventory_weave()
630
candidates = w.get_ancestry(revision_id)
631
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
634
def print_file(self, file, revision_id):
635
"""Print `file` to stdout.
637
FIXME RBC 20060125 as John Meinel points out this is a bad api
638
- it writes to stdout, it assumes that that is valid etc. Fix
639
by creating a new more flexible convenience function.
641
tree = self.revision_tree(revision_id)
642
# use inventory as it was in that revision
643
file_id = tree.inventory.path2id(file)
645
# TODO: jam 20060427 Write a test for this code path
646
# it had a bug in it, and was raising the wrong
648
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
649
tree.print_file(file_id)
651
def get_transaction(self):
652
return self.control_files.get_transaction()
654
def revision_parents(self, revid):
655
return self.get_inventory_weave().parent_names(revid)
658
def set_make_working_trees(self, new_value):
659
"""Set the policy flag for making working trees when creating branches.
661
This only applies to branches that use this repository.
663
The default is 'True'.
664
:param new_value: True to restore the default, False to disable making
667
raise NotImplementedError(self.set_make_working_trees)
669
def make_working_trees(self):
670
"""Returns the policy for making working trees on new branches."""
671
raise NotImplementedError(self.make_working_trees)
674
def sign_revision(self, revision_id, gpg_strategy):
675
plaintext = Testament.from_revision(self, revision_id).as_short_text()
676
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
679
def has_signature_for_revision_id(self, revision_id):
680
"""Query for a revision signature for revision_id in the repository."""
681
return self._revision_store.has_signature(revision_id,
682
self.get_transaction())
685
def get_signature_text(self, revision_id):
686
"""Return the text for a signature."""
687
return self._revision_store.get_signature_text(revision_id,
688
self.get_transaction())
691
def check(self, revision_ids):
692
"""Check consistency of all history of given revision_ids.
694
Different repository implementations should override _check().
696
:param revision_ids: A non-empty list of revision_ids whose ancestry
697
will be checked. Typically the last revision_id of a branch.
700
raise ValueError("revision_ids must be non-empty in %s.check"
702
return self._check(revision_ids)
704
def _check(self, revision_ids):
705
result = check.Check(self)
709
def _warn_if_deprecated(self):
710
global _deprecation_warning_done
711
if _deprecation_warning_done:
713
_deprecation_warning_done = True
714
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
715
% (self._format, self.bzrdir.transport.base))
718
class AllInOneRepository(Repository):
719
"""Legacy support - the repository behaviour for all-in-one branches."""
721
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
722
# we reuse one control files instance.
723
dir_mode = a_bzrdir._control_files._dir_mode
724
file_mode = a_bzrdir._control_files._file_mode
726
def get_store(name, compressed=True, prefixed=False):
727
# FIXME: This approach of assuming stores are all entirely compressed
728
# or entirely uncompressed is tidy, but breaks upgrade from
729
# some existing branches where there's a mixture; we probably
730
# still want the option to look for both.
731
relpath = a_bzrdir._control_files._escape(name)
732
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
733
prefixed=prefixed, compressed=compressed,
736
#if self._transport.should_cache():
737
# cache_path = os.path.join(self.cache_root, name)
738
# os.mkdir(cache_path)
739
# store = bzrlib.store.CachedStore(store, cache_path)
742
# not broken out yet because the controlweaves|inventory_store
743
# and text_store | weave_store bits are still different.
744
if isinstance(_format, RepositoryFormat4):
745
# cannot remove these - there is still no consistent api
746
# which allows access to this old info.
747
self.inventory_store = get_store('inventory-store')
748
text_store = get_store('text-store')
749
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
753
"""AllInOne repositories cannot be shared."""
757
def set_make_working_trees(self, new_value):
758
"""Set the policy flag for making working trees when creating branches.
760
This only applies to branches that use this repository.
762
The default is 'True'.
763
:param new_value: True to restore the default, False to disable making
766
raise NotImplementedError(self.set_make_working_trees)
768
def make_working_trees(self):
769
"""Returns the policy for making working trees on new branches."""
773
def install_revision(repository, rev, revision_tree):
774
"""Install all revision data into a repository."""
777
for p_id in rev.parent_ids:
778
if repository.has_revision(p_id):
779
present_parents.append(p_id)
780
parent_trees[p_id] = repository.revision_tree(p_id)
782
parent_trees[p_id] = repository.revision_tree(None)
784
inv = revision_tree.inventory
786
# backwards compatability hack: skip the root id.
787
entries = inv.iter_entries()
789
# Add the texts that are not already present
790
for path, ie in entries:
791
w = repository.weave_store.get_weave_or_empty(ie.file_id,
792
repository.get_transaction())
793
if ie.revision not in w:
795
# FIXME: TODO: The following loop *may* be overlapping/duplicate
796
# with InventoryEntry.find_previous_heads(). if it is, then there
797
# is a latent bug here where the parents may have ancestors of each
799
for revision, tree in parent_trees.iteritems():
800
if ie.file_id not in tree:
802
parent_id = tree.inventory[ie.file_id].revision
803
if parent_id in text_parents:
805
text_parents.append(parent_id)
807
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
808
repository.get_transaction())
809
lines = revision_tree.get_file(ie.file_id).readlines()
810
vfile.add_lines(rev.revision_id, text_parents, lines)
812
# install the inventory
813
repository.add_inventory(rev.revision_id, inv, present_parents)
814
except errors.RevisionAlreadyPresent:
816
repository.add_revision(rev.revision_id, rev, inv)
819
class MetaDirRepository(Repository):
820
"""Repositories in the new meta-dir layout."""
822
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
823
super(MetaDirRepository, self).__init__(_format,
829
dir_mode = self.control_files._dir_mode
830
file_mode = self.control_files._file_mode
834
"""Return True if this repository is flagged as a shared repository."""
835
return self.control_files._transport.has('shared-storage')
838
def set_make_working_trees(self, new_value):
839
"""Set the policy flag for making working trees when creating branches.
841
This only applies to branches that use this repository.
843
The default is 'True'.
844
:param new_value: True to restore the default, False to disable making
849
self.control_files._transport.delete('no-working-trees')
850
except errors.NoSuchFile:
853
self.control_files.put_utf8('no-working-trees', '')
855
def make_working_trees(self):
856
"""Returns the policy for making working trees on new branches."""
857
return not self.control_files._transport.has('no-working-trees')
860
class KnitRepository(MetaDirRepository):
861
"""Knit format repository."""
863
def _warn_if_deprecated(self):
864
# This class isn't deprecated
867
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
868
inv_vf.add_lines_with_ghosts(revid, parents, lines)
871
def _all_revision_ids(self):
872
"""See Repository.all_revision_ids()."""
873
# Knits get the revision graph from the index of the revision knit, so
874
# it's always possible even if they're on an unlistable transport.
875
return self._revision_store.all_revision_ids(self.get_transaction())
877
def fileid_involved_between_revs(self, from_revid, to_revid):
878
"""Find file_id(s) which are involved in the changes between revisions.
880
This determines the set of revisions which are involved, and then
881
finds all file ids affected by those revisions.
883
vf = self._get_revision_vf()
884
from_set = set(vf.get_ancestry(from_revid))
885
to_set = set(vf.get_ancestry(to_revid))
886
changed = to_set.difference(from_set)
887
return self._fileid_involved_by_set(changed)
889
def fileid_involved(self, last_revid=None):
890
"""Find all file_ids modified in the ancestry of last_revid.
892
:param last_revid: If None, last_revision() will be used.
895
changed = set(self.all_revision_ids())
897
changed = set(self.get_ancestry(last_revid))
900
return self._fileid_involved_by_set(changed)
903
def get_ancestry(self, revision_id):
904
"""Return a list of revision-ids integrated by a revision.
906
This is topologically sorted.
908
if revision_id is None:
910
vf = self._get_revision_vf()
912
return [None] + vf.get_ancestry(revision_id)
913
except errors.RevisionNotPresent:
914
raise errors.NoSuchRevision(self, revision_id)
917
def get_revision(self, revision_id):
918
"""Return the Revision object for a named revision"""
919
return self.get_revision_reconcile(revision_id)
922
def get_revision_graph(self, revision_id=None):
923
"""Return a dictionary containing the revision graph.
925
:param revision_id: The revision_id to get a graph from. If None, then
926
the entire revision graph is returned. This is a deprecated mode of
927
operation and will be removed in the future.
928
:return: a dictionary of revision_id->revision_parents_list.
930
# special case NULL_REVISION
931
if revision_id == NULL_REVISION:
933
weave = self._get_revision_vf()
934
entire_graph = weave.get_graph()
935
if revision_id is None:
936
return weave.get_graph()
937
elif revision_id not in weave:
938
raise errors.NoSuchRevision(self, revision_id)
940
# add what can be reached from revision_id
942
pending = set([revision_id])
943
while len(pending) > 0:
945
result[node] = weave.get_parents(node)
946
for revision_id in result[node]:
947
if revision_id not in result:
948
pending.add(revision_id)
952
def get_revision_graph_with_ghosts(self, revision_ids=None):
953
"""Return a graph of the revisions with ghosts marked as applicable.
955
:param revision_ids: an iterable of revisions to graph or None for all.
956
:return: a Graph object with the graph reachable from revision_ids.
959
vf = self._get_revision_vf()
960
versions = set(vf.versions())
962
pending = set(self.all_revision_ids())
965
pending = set(revision_ids)
966
# special case NULL_REVISION
967
if NULL_REVISION in pending:
968
pending.remove(NULL_REVISION)
969
required = set(pending)
972
revision_id = pending.pop()
973
if not revision_id in versions:
974
if revision_id in required:
975
raise errors.NoSuchRevision(self, revision_id)
977
result.add_ghost(revision_id)
978
# mark it as done so we don't try for it again.
979
done.add(revision_id)
981
parent_ids = vf.get_parents_with_ghosts(revision_id)
982
for parent_id in parent_ids:
983
# is this queued or done ?
984
if (parent_id not in pending and
985
parent_id not in done):
987
pending.add(parent_id)
988
result.add_node(revision_id, parent_ids)
989
done.add(revision_id)
992
def _get_revision_vf(self):
993
""":return: a versioned file containing the revisions."""
994
vf = self._revision_store.get_revision_file(self.get_transaction())
998
def reconcile(self, other=None, thorough=False):
999
"""Reconcile this repository."""
1000
from bzrlib.reconcile import KnitReconciler
1001
reconciler = KnitReconciler(self, thorough=thorough)
1002
reconciler.reconcile()
1005
def revision_parents(self, revision_id):
1006
return self._get_revision_vf().get_parents(revision_id)
1009
class KnitRepository2(KnitRepository):
1011
def __init__(self, _format, a_bzrdir, control_files, _revision_store,
1012
control_store, text_store):
1013
KnitRepository.__init__(self, _format, a_bzrdir, control_files,
1014
_revision_store, control_store, text_store)
1015
self._serializer = xml6.serializer_v6
1017
def deserialise_inventory(self, revision_id, xml):
1018
"""Transform the xml into an inventory object.
1020
:param revision_id: The expected revision id of the inventory.
1021
:param xml: A serialised inventory.
1023
result = self._serializer.read_inventory_from_string(xml)
1024
assert result.root.revision is not None
1027
def serialise_inventory(self, inv):
1028
"""Transform the inventory object into XML text.
1030
:param revision_id: The expected revision id of the inventory.
1031
:param xml: A serialised inventory.
1033
assert inv.revision_id is not None
1034
assert inv.root.revision is not None
1035
return KnitRepository.serialise_inventory(self, inv)
1037
def get_commit_builder(self, branch, parents, config, timestamp=None,
1038
timezone=None, committer=None, revprops=None,
1040
"""Obtain a CommitBuilder for this repository.
1042
:param branch: Branch to commit to.
1043
:param parents: Revision ids of the parents of the new revision.
1044
:param config: Configuration to use.
1045
:param timestamp: Optional timestamp recorded for commit.
1046
:param timezone: Optional timezone for timestamp.
1047
:param committer: Optional committer to set for commit.
1048
:param revprops: Optional dictionary of revision properties.
1049
:param revision_id: Optional revision id.
1051
return RootCommitBuilder(self, parents, config, timestamp, timezone,
1052
committer, revprops, revision_id)
1055
class RepositoryFormat(object):
1056
"""A repository format.
1058
Formats provide three things:
1059
* An initialization routine to construct repository data on disk.
1060
* a format string which is used when the BzrDir supports versioned
1062
* an open routine which returns a Repository instance.
1064
Formats are placed in an dict by their format string for reference
1065
during opening. These should be subclasses of RepositoryFormat
1068
Once a format is deprecated, just deprecate the initialize and open
1069
methods on the format class. Do not deprecate the object, as the
1070
object will be created every system load.
1072
Common instance attributes:
1073
_matchingbzrdir - the bzrdir format that the repository format was
1074
originally written to work with. This can be used if manually
1075
constructing a bzrdir and repository, or more commonly for test suite
1079
_default_format = None
1080
"""The default format used for new repositories."""
1083
"""The known formats."""
1086
return "<%s>" % self.__class__.__name__
1089
def find_format(klass, a_bzrdir):
1090
"""Return the format for the repository object in a_bzrdir."""
1092
transport = a_bzrdir.get_repository_transport(None)
1093
format_string = transport.get("format").read()
1094
return klass._formats[format_string]
1095
except errors.NoSuchFile:
1096
raise errors.NoRepositoryPresent(a_bzrdir)
1098
raise errors.UnknownFormatError(format=format_string)
1100
def _get_control_store(self, repo_transport, control_files):
1101
"""Return the control store for this repository."""
1102
raise NotImplementedError(self._get_control_store)
1105
def get_default_format(klass):
1106
"""Return the current default format."""
1107
return klass._default_format
1109
def get_format_string(self):
1110
"""Return the ASCII format string that identifies this format.
1112
Note that in pre format ?? repositories the format string is
1113
not permitted nor written to disk.
1115
raise NotImplementedError(self.get_format_string)
1117
def get_format_description(self):
1118
"""Return the short description for this format."""
1119
raise NotImplementedError(self.get_format_description)
1121
def _get_revision_store(self, repo_transport, control_files):
1122
"""Return the revision store object for this a_bzrdir."""
1123
raise NotImplementedError(self._get_revision_store)
1125
def _get_text_rev_store(self,
1132
"""Common logic for getting a revision store for a repository.
1134
see self._get_revision_store for the subclass-overridable method to
1135
get the store for a repository.
1137
from bzrlib.store.revision.text import TextRevisionStore
1138
dir_mode = control_files._dir_mode
1139
file_mode = control_files._file_mode
1140
text_store =TextStore(transport.clone(name),
1142
compressed=compressed,
1144
file_mode=file_mode)
1145
_revision_store = TextRevisionStore(text_store, serializer)
1146
return _revision_store
1148
def _get_versioned_file_store(self,
1153
versionedfile_class=WeaveFile,
1154
versionedfile_kwargs={},
1156
weave_transport = control_files._transport.clone(name)
1157
dir_mode = control_files._dir_mode
1158
file_mode = control_files._file_mode
1159
return VersionedFileStore(weave_transport, prefixed=prefixed,
1161
file_mode=file_mode,
1162
versionedfile_class=versionedfile_class,
1163
versionedfile_kwargs=versionedfile_kwargs,
1166
def initialize(self, a_bzrdir, shared=False):
1167
"""Initialize a repository of this format in a_bzrdir.
1169
:param a_bzrdir: The bzrdir to put the new repository in it.
1170
:param shared: The repository should be initialized as a sharable one.
1172
This may raise UninitializableFormat if shared repository are not
1173
compatible the a_bzrdir.
1176
def is_supported(self):
1177
"""Is this format supported?
1179
Supported formats must be initializable and openable.
1180
Unsupported formats may not support initialization or committing or
1181
some other features depending on the reason for not being supported.
1185
def check_conversion_target(self, target_format):
1186
raise NotImplementedError(self.check_conversion_target)
1188
def open(self, a_bzrdir, _found=False):
1189
"""Return an instance of this format for the bzrdir a_bzrdir.
1191
_found is a private parameter, do not use it.
1193
raise NotImplementedError(self.open)
1196
def register_format(klass, format):
1197
klass._formats[format.get_format_string()] = format
1200
def set_default_format(klass, format):
1201
klass._default_format = format
1204
def unregister_format(klass, format):
1205
assert klass._formats[format.get_format_string()] is format
1206
del klass._formats[format.get_format_string()]
1209
class PreSplitOutRepositoryFormat(RepositoryFormat):
1210
"""Base class for the pre split out repository formats."""
1212
rich_root_data = False
1214
def initialize(self, a_bzrdir, shared=False, _internal=False):
1215
"""Create a weave repository.
1217
TODO: when creating split out bzr branch formats, move this to a common
1218
base for Format5, Format6. or something like that.
1220
from bzrlib.weavefile import write_weave_v5
1221
from bzrlib.weave import Weave
1224
raise errors.IncompatibleFormat(self, a_bzrdir._format)
1227
# always initialized when the bzrdir is.
1228
return self.open(a_bzrdir, _found=True)
1230
# Create an empty weave
1232
write_weave_v5(Weave(), sio)
1233
empty_weave = sio.getvalue()
1235
mutter('creating repository in %s.', a_bzrdir.transport.base)
1236
dirs = ['revision-store', 'weaves']
1237
files = [('inventory.weave', StringIO(empty_weave)),
1240
# FIXME: RBC 20060125 don't peek under the covers
1241
# NB: no need to escape relative paths that are url safe.
1242
control_files = LockableFiles(a_bzrdir.transport, 'branch-lock',
1244
control_files.create_lock()
1245
control_files.lock_write()
1246
control_files._transport.mkdir_multi(dirs,
1247
mode=control_files._dir_mode)
1249
for file, content in files:
1250
control_files.put(file, content)
1252
control_files.unlock()
1253
return self.open(a_bzrdir, _found=True)
1255
def _get_control_store(self, repo_transport, control_files):
1256
"""Return the control store for this repository."""
1257
return self._get_versioned_file_store('',
1262
def _get_text_store(self, transport, control_files):
1263
"""Get a store for file texts for this format."""
1264
raise NotImplementedError(self._get_text_store)
1266
def open(self, a_bzrdir, _found=False):
1267
"""See RepositoryFormat.open()."""
1269
# we are being called directly and must probe.
1270
raise NotImplementedError
1272
repo_transport = a_bzrdir.get_repository_transport(None)
1273
control_files = a_bzrdir._control_files
1274
text_store = self._get_text_store(repo_transport, control_files)
1275
control_store = self._get_control_store(repo_transport, control_files)
1276
_revision_store = self._get_revision_store(repo_transport, control_files)
1277
return AllInOneRepository(_format=self,
1279
_revision_store=_revision_store,
1280
control_store=control_store,
1281
text_store=text_store)
1283
def check_conversion_target(self, target_format):
1287
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1288
"""Bzr repository format 4.
1290
This repository format has:
1292
- TextStores for texts, inventories,revisions.
1294
This format is deprecated: it indexes texts using a text id which is
1295
removed in format 5; initialization and write support for this format
1300
super(RepositoryFormat4, self).__init__()
1301
self._matchingbzrdir = bzrdir.BzrDirFormat4()
1303
def get_format_description(self):
1304
"""See RepositoryFormat.get_format_description()."""
1305
return "Repository format 4"
1307
def initialize(self, url, shared=False, _internal=False):
1308
"""Format 4 branches cannot be created."""
1309
raise errors.UninitializableFormat(self)
1311
def is_supported(self):
1312
"""Format 4 is not supported.
1314
It is not supported because the model changed from 4 to 5 and the
1315
conversion logic is expensive - so doing it on the fly was not
1320
def _get_control_store(self, repo_transport, control_files):
1321
"""Format 4 repositories have no formal control store at this point.
1323
This will cause any control-file-needing apis to fail - this is desired.
1327
def _get_revision_store(self, repo_transport, control_files):
1328
"""See RepositoryFormat._get_revision_store()."""
1329
from bzrlib.xml4 import serializer_v4
1330
return self._get_text_rev_store(repo_transport,
1333
serializer=serializer_v4)
1335
def _get_text_store(self, transport, control_files):
1336
"""See RepositoryFormat._get_text_store()."""
1339
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1340
"""Bzr control format 5.
1342
This repository format has:
1343
- weaves for file texts and inventory
1345
- TextStores for revisions and signatures.
1349
super(RepositoryFormat5, self).__init__()
1350
self._matchingbzrdir = bzrdir.BzrDirFormat5()
1352
def get_format_description(self):
1353
"""See RepositoryFormat.get_format_description()."""
1354
return "Weave repository format 5"
1356
def _get_revision_store(self, repo_transport, control_files):
1357
"""See RepositoryFormat._get_revision_store()."""
1358
"""Return the revision store object for this a_bzrdir."""
1359
return self._get_text_rev_store(repo_transport,
1364
def _get_text_store(self, transport, control_files):
1365
"""See RepositoryFormat._get_text_store()."""
1366
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1369
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1370
"""Bzr control format 6.
1372
This repository format has:
1373
- weaves for file texts and inventory
1374
- hash subdirectory based stores.
1375
- TextStores for revisions and signatures.
1379
super(RepositoryFormat6, self).__init__()
1380
self._matchingbzrdir = bzrdir.BzrDirFormat6()
1382
def get_format_description(self):
1383
"""See RepositoryFormat.get_format_description()."""
1384
return "Weave repository format 6"
1386
def _get_revision_store(self, repo_transport, control_files):
1387
"""See RepositoryFormat._get_revision_store()."""
1388
return self._get_text_rev_store(repo_transport,
1394
def _get_text_store(self, transport, control_files):
1395
"""See RepositoryFormat._get_text_store()."""
1396
return self._get_versioned_file_store('weaves', transport, control_files)
1399
class MetaDirRepositoryFormat(RepositoryFormat):
1400
"""Common base class for the new repositories using the metadir layout."""
1402
rich_root_data = False
1405
super(MetaDirRepositoryFormat, self).__init__()
1406
self._matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1408
def _create_control_files(self, a_bzrdir):
1409
"""Create the required files and the initial control_files object."""
1410
# FIXME: RBC 20060125 don't peek under the covers
1411
# NB: no need to escape relative paths that are url safe.
1412
repository_transport = a_bzrdir.get_repository_transport(self)
1413
control_files = LockableFiles(repository_transport, 'lock', LockDir)
1414
control_files.create_lock()
1415
return control_files
1417
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1418
"""Upload the initial blank content."""
1419
control_files = self._create_control_files(a_bzrdir)
1420
control_files.lock_write()
1422
control_files._transport.mkdir_multi(dirs,
1423
mode=control_files._dir_mode)
1424
for file, content in files:
1425
control_files.put(file, content)
1426
for file, content in utf8_files:
1427
control_files.put_utf8(file, content)
1429
control_files.put_utf8('shared-storage', '')
1431
control_files.unlock()
1434
class RepositoryFormat7(MetaDirRepositoryFormat):
1435
"""Bzr repository 7.
1437
This repository format has:
1438
- weaves for file texts and inventory
1439
- hash subdirectory based stores.
1440
- TextStores for revisions and signatures.
1441
- a format marker of its own
1442
- an optional 'shared-storage' flag
1443
- an optional 'no-working-trees' flag
1446
def _get_control_store(self, repo_transport, control_files):
1447
"""Return the control store for this repository."""
1448
return self._get_versioned_file_store('',
1453
def get_format_string(self):
1454
"""See RepositoryFormat.get_format_string()."""
1455
return "Bazaar-NG Repository format 7"
1457
def get_format_description(self):
1458
"""See RepositoryFormat.get_format_description()."""
1459
return "Weave repository format 7"
1461
def check_conversion_target(self, target_format):
1464
def _get_revision_store(self, repo_transport, control_files):
1465
"""See RepositoryFormat._get_revision_store()."""
1466
return self._get_text_rev_store(repo_transport,
1473
def _get_text_store(self, transport, control_files):
1474
"""See RepositoryFormat._get_text_store()."""
1475
return self._get_versioned_file_store('weaves',
1479
def initialize(self, a_bzrdir, shared=False):
1480
"""Create a weave repository.
1482
:param shared: If true the repository will be initialized as a shared
1485
from bzrlib.weavefile import write_weave_v5
1486
from bzrlib.weave import Weave
1488
# Create an empty weave
1490
write_weave_v5(Weave(), sio)
1491
empty_weave = sio.getvalue()
1493
mutter('creating repository in %s.', a_bzrdir.transport.base)
1494
dirs = ['revision-store', 'weaves']
1495
files = [('inventory.weave', StringIO(empty_weave)),
1497
utf8_files = [('format', self.get_format_string())]
1499
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1500
return self.open(a_bzrdir=a_bzrdir, _found=True)
1502
def open(self, a_bzrdir, _found=False, _override_transport=None):
1503
"""See RepositoryFormat.open().
1505
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1506
repository at a slightly different url
1507
than normal. I.e. during 'upgrade'.
1510
format = RepositoryFormat.find_format(a_bzrdir)
1511
assert format.__class__ == self.__class__
1512
if _override_transport is not None:
1513
repo_transport = _override_transport
1515
repo_transport = a_bzrdir.get_repository_transport(None)
1516
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1517
text_store = self._get_text_store(repo_transport, control_files)
1518
control_store = self._get_control_store(repo_transport, control_files)
1519
_revision_store = self._get_revision_store(repo_transport, control_files)
1520
return MetaDirRepository(_format=self,
1522
control_files=control_files,
1523
_revision_store=_revision_store,
1524
control_store=control_store,
1525
text_store=text_store)
1528
class RepositoryFormatKnit(MetaDirRepositoryFormat):
1529
"""Bzr repository knit format (generalized).
1531
This repository format has:
1532
- knits for file texts and inventory
1533
- hash subdirectory based stores.
1534
- knits for revisions and signatures
1535
- TextStores for revisions and signatures.
1536
- a format marker of its own
1537
- an optional 'shared-storage' flag
1538
- an optional 'no-working-trees' flag
1542
def _get_control_store(self, repo_transport, control_files):
1543
"""Return the control store for this repository."""
1544
return VersionedFileStore(
1547
file_mode=control_files._file_mode,
1548
versionedfile_class=KnitVersionedFile,
1549
versionedfile_kwargs={'factory':KnitPlainFactory()},
1552
def _get_revision_store(self, repo_transport, control_files):
1553
"""See RepositoryFormat._get_revision_store()."""
1554
from bzrlib.store.revision.knit import KnitRevisionStore
1555
versioned_file_store = VersionedFileStore(
1557
file_mode=control_files._file_mode,
1560
versionedfile_class=KnitVersionedFile,
1561
versionedfile_kwargs={'delta':False, 'factory':KnitPlainFactory(),},
1564
return KnitRevisionStore(versioned_file_store)
1566
def _get_text_store(self, transport, control_files):
1567
"""See RepositoryFormat._get_text_store()."""
1568
return self._get_versioned_file_store('knits',
1571
versionedfile_class=KnitVersionedFile,
1572
versionedfile_kwargs={
1573
'create_parent_dir':True,
1574
'delay_create':True,
1575
'dir_mode':control_files._dir_mode,
1579
def initialize(self, a_bzrdir, shared=False):
1580
"""Create a knit format 1 repository.
1582
:param a_bzrdir: bzrdir to contain the new repository; must already
1584
:param shared: If true the repository will be initialized as a shared
1587
mutter('creating repository in %s.', a_bzrdir.transport.base)
1588
dirs = ['revision-store', 'knits']
1590
utf8_files = [('format', self.get_format_string())]
1592
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1593
repo_transport = a_bzrdir.get_repository_transport(None)
1594
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1595
control_store = self._get_control_store(repo_transport, control_files)
1596
transaction = transactions.WriteTransaction()
1597
# trigger a write of the inventory store.
1598
control_store.get_weave_or_empty('inventory', transaction)
1599
_revision_store = self._get_revision_store(repo_transport, control_files)
1600
_revision_store.has_revision_id('A', transaction)
1601
_revision_store.get_signature_file(transaction)
1602
return self.open(a_bzrdir=a_bzrdir, _found=True)
1604
def open(self, a_bzrdir, _found=False, _override_transport=None):
1605
"""See RepositoryFormat.open().
1607
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1608
repository at a slightly different url
1609
than normal. I.e. during 'upgrade'.
1612
format = RepositoryFormat.find_format(a_bzrdir)
1613
assert format.__class__ == self.__class__
1614
if _override_transport is not None:
1615
repo_transport = _override_transport
1617
repo_transport = a_bzrdir.get_repository_transport(None)
1618
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1619
text_store = self._get_text_store(repo_transport, control_files)
1620
control_store = self._get_control_store(repo_transport, control_files)
1621
_revision_store = self._get_revision_store(repo_transport, control_files)
1622
return KnitRepository(_format=self,
1624
control_files=control_files,
1625
_revision_store=_revision_store,
1626
control_store=control_store,
1627
text_store=text_store)
1630
class RepositoryFormatKnit1(RepositoryFormatKnit):
1631
"""Bzr repository knit format 1.
1633
This repository format has:
1634
- knits for file texts and inventory
1635
- hash subdirectory based stores.
1636
- knits for revisions and signatures
1637
- TextStores for revisions and signatures.
1638
- a format marker of its own
1639
- an optional 'shared-storage' flag
1640
- an optional 'no-working-trees' flag
1643
This format was introduced in bzr 0.8.
1645
def get_format_string(self):
1646
"""See RepositoryFormat.get_format_string()."""
1647
return "Bazaar-NG Knit Repository Format 1"
1649
def get_format_description(self):
1650
"""See RepositoryFormat.get_format_description()."""
1651
return "Knit repository format 1"
1653
def check_conversion_target(self, target_format):
1657
class RepositoryFormatKnit2(RepositoryFormatKnit):
1658
"""Bzr repository knit format 2.
1660
THIS FORMAT IS EXPERIMENTAL
1661
This repository format has:
1662
- knits for file texts and inventory
1663
- hash subdirectory based stores.
1664
- knits for revisions and signatures
1665
- TextStores for revisions and signatures.
1666
- a format marker of its own
1667
- an optional 'shared-storage' flag
1668
- an optional 'no-working-trees' flag
1670
- Support for recording full info about the tree root
1674
rich_root_data = True
1676
def get_format_string(self):
1677
"""See RepositoryFormat.get_format_string()."""
1678
return "Bazaar Knit Repository Format 2\n"
1680
def get_format_description(self):
1681
"""See RepositoryFormat.get_format_description()."""
1682
return "Knit repository format 2"
1684
def check_conversion_target(self, target_format):
1685
if not target_format.rich_root_data:
1686
raise errors.BadConversionTarget(
1687
'Does not support rich root data.', target_format)
1689
def open(self, a_bzrdir, _found=False, _override_transport=None):
1690
"""See RepositoryFormat.open().
1692
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1693
repository at a slightly different url
1694
than normal. I.e. during 'upgrade'.
1697
format = RepositoryFormat.find_format(a_bzrdir)
1698
assert format.__class__ == self.__class__
1699
if _override_transport is not None:
1700
repo_transport = _override_transport
1702
repo_transport = a_bzrdir.get_repository_transport(None)
1703
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1704
text_store = self._get_text_store(repo_transport, control_files)
1705
control_store = self._get_control_store(repo_transport, control_files)
1706
_revision_store = self._get_revision_store(repo_transport, control_files)
1707
return KnitRepository2(_format=self,
1709
control_files=control_files,
1710
_revision_store=_revision_store,
1711
control_store=control_store,
1712
text_store=text_store)
1716
# formats which have no format string are not discoverable
1717
# and not independently creatable, so are not registered.
1718
RepositoryFormat.register_format(RepositoryFormat7())
1719
_default_format = RepositoryFormatKnit1()
1720
RepositoryFormat.register_format(_default_format)
1721
RepositoryFormat.register_format(RepositoryFormatKnit2())
1722
RepositoryFormat.set_default_format(_default_format)
1723
_legacy_formats = [RepositoryFormat4(),
1724
RepositoryFormat5(),
1725
RepositoryFormat6()]
1728
class InterRepository(InterObject):
1729
"""This class represents operations taking place between two repositories.
1731
Its instances have methods like copy_content and fetch, and contain
1732
references to the source and target repositories these operations can be
1735
Often we will provide convenience methods on 'repository' which carry out
1736
operations with another repository - they will always forward to
1737
InterRepository.get(other).method_name(parameters).
1741
"""The available optimised InterRepository types."""
1743
def copy_content(self, revision_id=None, basis=None):
1744
raise NotImplementedError(self.copy_content)
1746
def fetch(self, revision_id=None, pb=None):
1747
"""Fetch the content required to construct revision_id.
1749
The content is copied from self.source to self.target.
1751
:param revision_id: if None all content is copied, if NULL_REVISION no
1753
:param pb: optional progress bar to use for progress reports. If not
1754
provided a default one will be created.
1756
Returns the copied revision count and the failed revisions in a tuple:
1759
raise NotImplementedError(self.fetch)
1762
def missing_revision_ids(self, revision_id=None):
1763
"""Return the revision ids that source has that target does not.
1765
These are returned in topological order.
1767
:param revision_id: only return revision ids included by this
1770
# generic, possibly worst case, slow code path.
1771
target_ids = set(self.target.all_revision_ids())
1772
if revision_id is not None:
1773
source_ids = self.source.get_ancestry(revision_id)
1774
assert source_ids[0] is None
1777
source_ids = self.source.all_revision_ids()
1778
result_set = set(source_ids).difference(target_ids)
1779
# this may look like a no-op: its not. It preserves the ordering
1780
# other_ids had while only returning the members from other_ids
1781
# that we've decided we need.
1782
return [rev_id for rev_id in source_ids if rev_id in result_set]
1785
class InterSameDataRepository(InterRepository):
1786
"""Code for converting between repositories that represent the same data.
1788
Data format and model must match for this to work.
1791
_matching_repo_format = RepositoryFormat4()
1792
"""Repository format for testing with."""
1795
def is_compatible(source, target):
1796
if not isinstance(source, Repository):
1798
if not isinstance(target, Repository):
1800
if source._format.rich_root_data == target._format.rich_root_data:
1806
def copy_content(self, revision_id=None, basis=None):
1807
"""Make a complete copy of the content in self into destination.
1809
This is a destructive operation! Do not use it on existing
1812
:param revision_id: Only copy the content needed to construct
1813
revision_id and its parents.
1814
:param basis: Copy the needed data preferentially from basis.
1817
self.target.set_make_working_trees(self.source.make_working_trees())
1818
except NotImplementedError:
1820
# grab the basis available data
1821
if basis is not None:
1822
self.target.fetch(basis, revision_id=revision_id)
1823
# but don't bother fetching if we have the needed data now.
1824
if (revision_id not in (None, NULL_REVISION) and
1825
self.target.has_revision(revision_id)):
1827
self.target.fetch(self.source, revision_id=revision_id)
1830
def fetch(self, revision_id=None, pb=None):
1831
"""See InterRepository.fetch()."""
1832
from bzrlib.fetch import GenericRepoFetcher
1833
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1834
self.source, self.source._format, self.target,
1835
self.target._format)
1836
f = GenericRepoFetcher(to_repository=self.target,
1837
from_repository=self.source,
1838
last_revision=revision_id,
1840
return f.count_copied, f.failed_revisions
1843
class InterWeaveRepo(InterSameDataRepository):
1844
"""Optimised code paths between Weave based repositories."""
1846
_matching_repo_format = RepositoryFormat7()
1847
"""Repository format for testing with."""
1850
def is_compatible(source, target):
1851
"""Be compatible with known Weave formats.
1853
We don't test for the stores being of specific types because that
1854
could lead to confusing results, and there is no need to be
1858
return (isinstance(source._format, (RepositoryFormat5,
1860
RepositoryFormat7)) and
1861
isinstance(target._format, (RepositoryFormat5,
1863
RepositoryFormat7)))
1864
except AttributeError:
1868
def copy_content(self, revision_id=None, basis=None):
1869
"""See InterRepository.copy_content()."""
1870
# weave specific optimised path:
1871
if basis is not None:
1872
# copy the basis in, then fetch remaining data.
1873
basis.copy_content_into(self.target, revision_id)
1874
# the basis copy_content_into could miss-set this.
1876
self.target.set_make_working_trees(self.source.make_working_trees())
1877
except NotImplementedError:
1879
self.target.fetch(self.source, revision_id=revision_id)
1882
self.target.set_make_working_trees(self.source.make_working_trees())
1883
except NotImplementedError:
1885
# FIXME do not peek!
1886
if self.source.control_files._transport.listable():
1887
pb = ui.ui_factory.nested_progress_bar()
1889
self.target.weave_store.copy_all_ids(
1890
self.source.weave_store,
1892
from_transaction=self.source.get_transaction(),
1893
to_transaction=self.target.get_transaction())
1894
pb.update('copying inventory', 0, 1)
1895
self.target.control_weaves.copy_multi(
1896
self.source.control_weaves, ['inventory'],
1897
from_transaction=self.source.get_transaction(),
1898
to_transaction=self.target.get_transaction())
1899
self.target._revision_store.text_store.copy_all_ids(
1900
self.source._revision_store.text_store,
1905
self.target.fetch(self.source, revision_id=revision_id)
1908
def fetch(self, revision_id=None, pb=None):
1909
"""See InterRepository.fetch()."""
1910
from bzrlib.fetch import GenericRepoFetcher
1911
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1912
self.source, self.source._format, self.target, self.target._format)
1913
f = GenericRepoFetcher(to_repository=self.target,
1914
from_repository=self.source,
1915
last_revision=revision_id,
1917
return f.count_copied, f.failed_revisions
1920
def missing_revision_ids(self, revision_id=None):
1921
"""See InterRepository.missing_revision_ids()."""
1922
# we want all revisions to satisfy revision_id in source.
1923
# but we don't want to stat every file here and there.
1924
# we want then, all revisions other needs to satisfy revision_id
1925
# checked, but not those that we have locally.
1926
# so the first thing is to get a subset of the revisions to
1927
# satisfy revision_id in source, and then eliminate those that
1928
# we do already have.
1929
# this is slow on high latency connection to self, but as as this
1930
# disk format scales terribly for push anyway due to rewriting
1931
# inventory.weave, this is considered acceptable.
1933
if revision_id is not None:
1934
source_ids = self.source.get_ancestry(revision_id)
1935
assert source_ids[0] is None
1938
source_ids = self.source._all_possible_ids()
1939
source_ids_set = set(source_ids)
1940
# source_ids is the worst possible case we may need to pull.
1941
# now we want to filter source_ids against what we actually
1942
# have in target, but don't try to check for existence where we know
1943
# we do not have a revision as that would be pointless.
1944
target_ids = set(self.target._all_possible_ids())
1945
possibly_present_revisions = target_ids.intersection(source_ids_set)
1946
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1947
required_revisions = source_ids_set.difference(actually_present_revisions)
1948
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1949
if revision_id is not None:
1950
# we used get_ancestry to determine source_ids then we are assured all
1951
# revisions referenced are present as they are installed in topological order.
1952
# and the tip revision was validated by get_ancestry.
1953
return required_topo_revisions
1955
# if we just grabbed the possibly available ids, then
1956
# we only have an estimate of whats available and need to validate
1957
# that against the revision records.
1958
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1961
class InterKnitRepo(InterSameDataRepository):
1962
"""Optimised code paths between Knit based repositories."""
1964
_matching_repo_format = RepositoryFormatKnit1()
1965
"""Repository format for testing with."""
1968
def is_compatible(source, target):
1969
"""Be compatible with known Knit formats.
1971
We don't test for the stores being of specific types because that
1972
could lead to confusing results, and there is no need to be
1976
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1977
isinstance(target._format, (RepositoryFormatKnit1)))
1978
except AttributeError:
1982
def fetch(self, revision_id=None, pb=None):
1983
"""See InterRepository.fetch()."""
1984
from bzrlib.fetch import KnitRepoFetcher
1985
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1986
self.source, self.source._format, self.target, self.target._format)
1987
f = KnitRepoFetcher(to_repository=self.target,
1988
from_repository=self.source,
1989
last_revision=revision_id,
1991
return f.count_copied, f.failed_revisions
1994
def missing_revision_ids(self, revision_id=None):
1995
"""See InterRepository.missing_revision_ids()."""
1996
if revision_id is not None:
1997
source_ids = self.source.get_ancestry(revision_id)
1998
assert source_ids[0] is None
2001
source_ids = self.source._all_possible_ids()
2002
source_ids_set = set(source_ids)
2003
# source_ids is the worst possible case we may need to pull.
2004
# now we want to filter source_ids against what we actually
2005
# have in target, but don't try to check for existence where we know
2006
# we do not have a revision as that would be pointless.
2007
target_ids = set(self.target._all_possible_ids())
2008
possibly_present_revisions = target_ids.intersection(source_ids_set)
2009
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2010
required_revisions = source_ids_set.difference(actually_present_revisions)
2011
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2012
if revision_id is not None:
2013
# we used get_ancestry to determine source_ids then we are assured all
2014
# revisions referenced are present as they are installed in topological order.
2015
# and the tip revision was validated by get_ancestry.
2016
return required_topo_revisions
2018
# if we just grabbed the possibly available ids, then
2019
# we only have an estimate of whats available and need to validate
2020
# that against the revision records.
2021
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2024
class InterModel1and2(InterRepository):
2026
_matching_repo_format = None
2029
def is_compatible(source, target):
2030
if not isinstance(source, Repository):
2032
if not isinstance(target, Repository):
2034
if not source._format.rich_root_data and target._format.rich_root_data:
2040
def fetch(self, revision_id=None, pb=None):
2041
"""See InterRepository.fetch()."""
2042
from bzrlib.fetch import Model1toKnit2Fetcher
2043
f = Model1toKnit2Fetcher(to_repository=self.target,
2044
from_repository=self.source,
2045
last_revision=revision_id,
2047
return f.count_copied, f.failed_revisions
2050
def copy_content(self, revision_id=None, basis=None):
2051
"""Make a complete copy of the content in self into destination.
2053
This is a destructive operation! Do not use it on existing
2056
:param revision_id: Only copy the content needed to construct
2057
revision_id and its parents.
2058
:param basis: Copy the needed data preferentially from basis.
2061
self.target.set_make_working_trees(self.source.make_working_trees())
2062
except NotImplementedError:
2064
# grab the basis available data
2065
if basis is not None:
2066
self.target.fetch(basis, revision_id=revision_id)
2067
# but don't bother fetching if we have the needed data now.
2068
if (revision_id not in (None, NULL_REVISION) and
2069
self.target.has_revision(revision_id)):
2071
self.target.fetch(self.source, revision_id=revision_id)
2074
class InterKnit1and2(InterKnitRepo):
2076
_matching_repo_format = None
2079
def is_compatible(source, target):
2080
"""Be compatible with Knit1 source and Knit2 target"""
2082
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2083
isinstance(target._format, (RepositoryFormatKnit2)))
2084
except AttributeError:
2088
def fetch(self, revision_id=None, pb=None):
2089
"""See InterRepository.fetch()."""
2090
from bzrlib.fetch import Knit1to2Fetcher
2091
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2092
self.source, self.source._format, self.target,
2093
self.target._format)
2094
f = Knit1to2Fetcher(to_repository=self.target,
2095
from_repository=self.source,
2096
last_revision=revision_id,
2098
return f.count_copied, f.failed_revisions
2101
InterRepository.register_optimiser(InterSameDataRepository)
2102
InterRepository.register_optimiser(InterWeaveRepo)
2103
InterRepository.register_optimiser(InterKnitRepo)
2104
InterRepository.register_optimiser(InterModel1and2)
2105
InterRepository.register_optimiser(InterKnit1and2)
2108
class RepositoryTestProviderAdapter(object):
2109
"""A tool to generate a suite testing multiple repository formats at once.
2111
This is done by copying the test once for each transport and injecting
2112
the transport_server, transport_readonly_server, and bzrdir_format and
2113
repository_format classes into each copy. Each copy is also given a new id()
2114
to make it easy to identify.
2117
def __init__(self, transport_server, transport_readonly_server, formats):
2118
self._transport_server = transport_server
2119
self._transport_readonly_server = transport_readonly_server
2120
self._formats = formats
2122
def adapt(self, test):
2123
result = TestSuite()
2124
for repository_format, bzrdir_format in self._formats:
2125
new_test = deepcopy(test)
2126
new_test.transport_server = self._transport_server
2127
new_test.transport_readonly_server = self._transport_readonly_server
2128
new_test.bzrdir_format = bzrdir_format
2129
new_test.repository_format = repository_format
2130
def make_new_test_id():
2131
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
2132
return lambda: new_id
2133
new_test.id = make_new_test_id()
2134
result.addTest(new_test)
2138
class InterRepositoryTestProviderAdapter(object):
2139
"""A tool to generate a suite testing multiple inter repository formats.
2141
This is done by copying the test once for each interrepo provider and injecting
2142
the transport_server, transport_readonly_server, repository_format and
2143
repository_to_format classes into each copy.
2144
Each copy is also given a new id() to make it easy to identify.
2147
def __init__(self, transport_server, transport_readonly_server, formats):
2148
self._transport_server = transport_server
2149
self._transport_readonly_server = transport_readonly_server
2150
self._formats = formats
2152
def adapt(self, test):
2153
result = TestSuite()
2154
for interrepo_class, repository_format, repository_format_to in self._formats:
2155
new_test = deepcopy(test)
2156
new_test.transport_server = self._transport_server
2157
new_test.transport_readonly_server = self._transport_readonly_server
2158
new_test.interrepo_class = interrepo_class
2159
new_test.repository_format = repository_format
2160
new_test.repository_format_to = repository_format_to
2161
def make_new_test_id():
2162
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
2163
return lambda: new_id
2164
new_test.id = make_new_test_id()
2165
result.addTest(new_test)
2169
def default_test_list():
2170
"""Generate the default list of interrepo permutations to test."""
2172
# test the default InterRepository between format 6 and the current
2174
# XXX: robertc 20060220 reinstate this when there are two supported
2175
# formats which do not have an optimal code path between them.
2176
#result.append((InterRepository,
2177
# RepositoryFormat6(),
2178
# RepositoryFormatKnit1()))
2179
for optimiser in InterRepository._optimisers:
2180
if optimiser._matching_repo_format is not None:
2181
result.append((optimiser,
2182
optimiser._matching_repo_format,
2183
optimiser._matching_repo_format
2185
# if there are specific combinations we want to use, we can add them
2187
result.append((InterModel1and2, RepositoryFormat5(),
2188
RepositoryFormatKnit2()))
2189
result.append((InterKnit1and2, RepositoryFormatKnit1(),
2190
RepositoryFormatKnit2()))
2194
class CopyConverter(object):
2195
"""A repository conversion tool which just performs a copy of the content.
2197
This is slow but quite reliable.
2200
def __init__(self, target_format):
2201
"""Create a CopyConverter.
2203
:param target_format: The format the resulting repository should be.
2205
self.target_format = target_format
2207
def convert(self, repo, pb):
2208
"""Perform the conversion of to_convert, giving feedback via pb.
2210
:param to_convert: The disk object to convert.
2211
:param pb: a progress bar to use for progress information.
2216
# this is only useful with metadir layouts - separated repo content.
2217
# trigger an assertion if not such
2218
repo._format.get_format_string()
2219
self.repo_dir = repo.bzrdir
2220
self.step('Moving repository to repository.backup')
2221
self.repo_dir.transport.move('repository', 'repository.backup')
2222
backup_transport = self.repo_dir.transport.clone('repository.backup')
2223
repo._format.check_conversion_target(self.target_format)
2224
self.source_repo = repo._format.open(self.repo_dir,
2226
_override_transport=backup_transport)
2227
self.step('Creating new repository')
2228
converted = self.target_format.initialize(self.repo_dir,
2229
self.source_repo.is_shared())
2230
converted.lock_write()
2232
self.step('Copying content into repository.')
2233
self.source_repo.copy_content_into(converted)
2236
self.step('Deleting old repository content.')
2237
self.repo_dir.transport.delete_tree('repository.backup')
2238
self.pb.note('repository converted')
2240
def step(self, message):
2241
"""Update the pb by a step."""
2243
self.pb.update(message, self.count, self.total)
2246
class CommitBuilder(object):
2247
"""Provides an interface to build up a commit.
2249
This allows describing a tree to be committed without needing to
2250
know the internals of the format of the repository.
2253
record_root_entry = False
2254
def __init__(self, repository, parents, config, timestamp=None,
2255
timezone=None, committer=None, revprops=None,
2257
"""Initiate a CommitBuilder.
2259
:param repository: Repository to commit to.
2260
:param parents: Revision ids of the parents of the new revision.
2261
:param config: Configuration to use.
2262
:param timestamp: Optional timestamp recorded for commit.
2263
:param timezone: Optional timezone for timestamp.
2264
:param committer: Optional committer to set for commit.
2265
:param revprops: Optional dictionary of revision properties.
2266
:param revision_id: Optional revision id.
2268
self._config = config
2270
if committer is None:
2271
self._committer = self._config.username()
2273
assert isinstance(committer, basestring), type(committer)
2274
self._committer = committer
2276
self.new_inventory = Inventory(None)
2277
self._new_revision_id = revision_id
2278
self.parents = parents
2279
self.repository = repository
2282
if revprops is not None:
2283
self._revprops.update(revprops)
2285
if timestamp is None:
2286
timestamp = time.time()
2287
# Restrict resolution to 1ms
2288
self._timestamp = round(timestamp, 3)
2290
if timezone is None:
2291
self._timezone = local_time_offset()
2293
self._timezone = int(timezone)
2295
self._generate_revision_if_needed()
2297
def commit(self, message):
2298
"""Make the actual commit.
2300
:return: The revision id of the recorded revision.
2302
rev = Revision(timestamp=self._timestamp,
2303
timezone=self._timezone,
2304
committer=self._committer,
2306
inventory_sha1=self.inv_sha1,
2307
revision_id=self._new_revision_id,
2308
properties=self._revprops)
2309
rev.parent_ids = self.parents
2310
self.repository.add_revision(self._new_revision_id, rev,
2311
self.new_inventory, self._config)
2312
return self._new_revision_id
2314
def revision_tree(self):
2315
"""Return the tree that was just committed.
2317
After calling commit() this can be called to get a RevisionTree
2318
representing the newly committed tree. This is preferred to
2319
calling Repository.revision_tree() because that may require
2320
deserializing the inventory, while we already have a copy in
2323
return RevisionTree(self.repository, self.new_inventory,
2324
self._new_revision_id)
2326
def finish_inventory(self):
2327
"""Tell the builder that the inventory is finished."""
2328
if self.new_inventory.root is None:
2329
symbol_versioning.warn('Root entry should be supplied to'
2330
' record_entry_contents, as of bzr 0.10.',
2331
DeprecationWarning, stacklevel=2)
2332
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
2333
self.new_inventory.revision_id = self._new_revision_id
2334
self.inv_sha1 = self.repository.add_inventory(
2335
self._new_revision_id,
2340
def _gen_revision_id(self):
2341
"""Return new revision-id."""
2342
s = '%s-%s-' % (self._config.user_email(),
2343
compact_date(self._timestamp))
2344
s += hexlify(rand_bytes(8))
2347
def _generate_revision_if_needed(self):
2348
"""Create a revision id if None was supplied.
2350
If the repository can not support user-specified revision ids
2351
they should override this function and raise UnsupportedOperation
2352
if _new_revision_id is not None.
2354
:raises: UnsupportedOperation
2356
if self._new_revision_id is None:
2357
self._new_revision_id = self._gen_revision_id()
2359
def record_entry_contents(self, ie, parent_invs, path, tree):
2360
"""Record the content of ie from tree into the commit if needed.
2362
Side effect: sets ie.revision when unchanged
2364
:param ie: An inventory entry present in the commit.
2365
:param parent_invs: The inventories of the parent revisions of the
2367
:param path: The path the entry is at in the tree.
2368
:param tree: The tree which contains this entry and should be used to
2371
if self.new_inventory.root is None and ie.parent_id is not None:
2372
symbol_versioning.warn('Root entry should be supplied to'
2373
' record_entry_contents, as of bzr 0.10.',
2374
DeprecationWarning, stacklevel=2)
2375
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
2377
self.new_inventory.add(ie)
2379
# ie.revision is always None if the InventoryEntry is considered
2380
# for committing. ie.snapshot will record the correct revision
2381
# which may be the sole parent if it is untouched.
2382
if ie.revision is not None:
2385
# In this revision format, root entries have no knit or weave
2386
if ie is self.new_inventory.root:
2387
# When serializing out to disk and back in
2388
# root.revision is always _new_revision_id
2389
ie.revision = self._new_revision_id
2391
previous_entries = ie.find_previous_heads(
2393
self.repository.weave_store,
2394
self.repository.get_transaction())
2395
# we are creating a new revision for ie in the history store
2397
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2399
def modified_directory(self, file_id, file_parents):
2400
"""Record the presence of a symbolic link.
2402
:param file_id: The file_id of the link to record.
2403
:param file_parents: The per-file parent revision ids.
2405
self._add_text_to_weave(file_id, [], file_parents.keys())
2407
def modified_file_text(self, file_id, file_parents,
2408
get_content_byte_lines, text_sha1=None,
2410
"""Record the text of file file_id
2412
:param file_id: The file_id of the file to record the text of.
2413
:param file_parents: The per-file parent revision ids.
2414
:param get_content_byte_lines: A callable which will return the byte
2416
:param text_sha1: Optional SHA1 of the file contents.
2417
:param text_size: Optional size of the file contents.
2419
# mutter('storing text of file {%s} in revision {%s} into %r',
2420
# file_id, self._new_revision_id, self.repository.weave_store)
2421
# special case to avoid diffing on renames or
2423
if (len(file_parents) == 1
2424
and text_sha1 == file_parents.values()[0].text_sha1
2425
and text_size == file_parents.values()[0].text_size):
2426
previous_ie = file_parents.values()[0]
2427
versionedfile = self.repository.weave_store.get_weave(file_id,
2428
self.repository.get_transaction())
2429
versionedfile.clone_text(self._new_revision_id,
2430
previous_ie.revision, file_parents.keys())
2431
return text_sha1, text_size
2433
new_lines = get_content_byte_lines()
2434
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2435
# should return the SHA1 and size
2436
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2437
return osutils.sha_strings(new_lines), \
2438
sum(map(len, new_lines))
2440
def modified_link(self, file_id, file_parents, link_target):
2441
"""Record the presence of a symbolic link.
2443
:param file_id: The file_id of the link to record.
2444
:param file_parents: The per-file parent revision ids.
2445
:param link_target: Target location of this link.
2447
self._add_text_to_weave(file_id, [], file_parents.keys())
2449
def _add_text_to_weave(self, file_id, new_lines, parents):
2450
versionedfile = self.repository.weave_store.get_weave_or_empty(
2451
file_id, self.repository.get_transaction())
2452
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2453
versionedfile.clear_cache()
2456
class _CommitBuilder(CommitBuilder):
2457
"""Temporary class so old CommitBuilders are detected properly
2459
Note: CommitBuilder works whether or not root entry is recorded.
2462
record_root_entry = True
2465
class RootCommitBuilder(CommitBuilder):
2466
"""This commitbuilder actually records the root id"""
2468
record_root_entry = True
2470
def record_entry_contents(self, ie, parent_invs, path, tree):
2471
"""Record the content of ie from tree into the commit if needed.
2473
Side effect: sets ie.revision when unchanged
2475
:param ie: An inventory entry present in the commit.
2476
:param parent_invs: The inventories of the parent revisions of the
2478
:param path: The path the entry is at in the tree.
2479
:param tree: The tree which contains this entry and should be used to
2482
assert self.new_inventory.root is not None or ie.parent_id is None
2483
self.new_inventory.add(ie)
2485
# ie.revision is always None if the InventoryEntry is considered
2486
# for committing. ie.snapshot will record the correct revision
2487
# which may be the sole parent if it is untouched.
2488
if ie.revision is not None:
2491
previous_entries = ie.find_previous_heads(
2493
self.repository.weave_store,
2494
self.repository.get_transaction())
2495
# we are creating a new revision for ie in the history store
2497
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2509
def _unescaper(match, _map=_unescape_map):
2510
return _map[match.group(1)]
2516
def _unescape_xml(data):
2517
"""Unescape predefined XML entities in a string of data."""
2519
if _unescape_re is None:
2520
_unescape_re = re.compile('\&([^;]*);')
2521
return _unescape_re.sub(_unescaper, data)