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
for line in w.iter_lines_added_or_present_in_versions(selected_revision_ids):
430
start = line.find('file_id="')+9
431
if start < 9: continue
432
end = line.find('"', start)
434
file_id = _unescape_xml(line[start:end])
436
start = line.find('revision="')+10
437
if start < 10: continue
438
end = line.find('"', start)
440
revision_id = _unescape_xml(line[start:end])
441
if revision_id in selected_revision_ids:
442
result.setdefault(file_id, set()).add(revision_id)
446
def get_inventory_weave(self):
447
return self.control_weaves.get_weave('inventory',
448
self.get_transaction())
451
def get_inventory(self, revision_id):
452
"""Get Inventory object by hash."""
453
return self.deserialise_inventory(
454
revision_id, self.get_inventory_xml(revision_id))
456
def deserialise_inventory(self, revision_id, xml):
457
"""Transform the xml into an inventory object.
459
:param revision_id: The expected revision id of the inventory.
460
:param xml: A serialised inventory.
462
result = self._serializer.read_inventory_from_string(xml)
463
result.root.revision = revision_id
466
def serialise_inventory(self, inv):
467
return self._serializer.write_inventory_to_string(inv)
470
def get_inventory_xml(self, revision_id):
471
"""Get inventory XML as a file object."""
473
assert isinstance(revision_id, basestring), type(revision_id)
474
iw = self.get_inventory_weave()
475
return iw.get_text(revision_id)
477
raise errors.HistoryMissing(self, 'inventory', revision_id)
480
def get_inventory_sha1(self, revision_id):
481
"""Return the sha1 hash of the inventory entry
483
return self.get_revision(revision_id).inventory_sha1
486
def get_revision_graph(self, revision_id=None):
487
"""Return a dictionary containing the revision graph.
489
:param revision_id: The revision_id to get a graph from. If None, then
490
the entire revision graph is returned. This is a deprecated mode of
491
operation and will be removed in the future.
492
:return: a dictionary of revision_id->revision_parents_list.
494
# special case NULL_REVISION
495
if revision_id == NULL_REVISION:
497
weave = self.get_inventory_weave()
498
all_revisions = self._eliminate_revisions_not_present(weave.versions())
499
entire_graph = dict([(node, weave.get_parents(node)) for
500
node in all_revisions])
501
if revision_id is None:
503
elif revision_id not in entire_graph:
504
raise errors.NoSuchRevision(self, revision_id)
506
# add what can be reached from revision_id
508
pending = set([revision_id])
509
while len(pending) > 0:
511
result[node] = entire_graph[node]
512
for revision_id in result[node]:
513
if revision_id not in result:
514
pending.add(revision_id)
518
def get_revision_graph_with_ghosts(self, revision_ids=None):
519
"""Return a graph of the revisions with ghosts marked as applicable.
521
:param revision_ids: an iterable of revisions to graph or None for all.
522
:return: a Graph object with the graph reachable from revision_ids.
526
pending = set(self.all_revision_ids())
529
pending = set(revision_ids)
530
# special case NULL_REVISION
531
if NULL_REVISION in pending:
532
pending.remove(NULL_REVISION)
533
required = set(pending)
536
revision_id = pending.pop()
538
rev = self.get_revision(revision_id)
539
except errors.NoSuchRevision:
540
if revision_id in required:
543
result.add_ghost(revision_id)
545
for parent_id in rev.parent_ids:
546
# is this queued or done ?
547
if (parent_id not in pending and
548
parent_id not in done):
550
pending.add(parent_id)
551
result.add_node(revision_id, rev.parent_ids)
552
done.add(revision_id)
556
def get_revision_inventory(self, revision_id):
557
"""Return inventory of a past revision."""
558
# TODO: Unify this with get_inventory()
559
# bzr 0.0.6 and later imposes the constraint that the inventory_id
560
# must be the same as its revision, so this is trivial.
561
if revision_id is None:
562
# This does not make sense: if there is no revision,
563
# then it is the current tree inventory surely ?!
564
# and thus get_root_id() is something that looks at the last
565
# commit on the branch, and the get_root_id is an inventory check.
566
raise NotImplementedError
567
# return Inventory(self.get_root_id())
569
return self.get_inventory(revision_id)
573
"""Return True if this repository is flagged as a shared repository."""
574
raise NotImplementedError(self.is_shared)
577
def reconcile(self, other=None, thorough=False):
578
"""Reconcile this repository."""
579
from bzrlib.reconcile import RepoReconciler
580
reconciler = RepoReconciler(self, thorough=thorough)
581
reconciler.reconcile()
585
def revision_tree(self, revision_id):
586
"""Return Tree for a revision on this branch.
588
`revision_id` may be None for the empty tree revision.
590
# TODO: refactor this to use an existing revision object
591
# so we don't need to read it in twice.
592
if revision_id is None or revision_id == NULL_REVISION:
593
return RevisionTree(self, Inventory(), NULL_REVISION)
595
inv = self.get_revision_inventory(revision_id)
596
return RevisionTree(self, inv, revision_id)
599
def revision_trees(self, revision_ids):
600
"""Return Tree for a revision on this branch.
602
`revision_id` may not be None or 'null:'"""
603
assert None not in revision_ids
604
assert NULL_REVISION not in revision_ids
605
texts = self.get_inventory_weave().get_texts(revision_ids)
606
for text, revision_id in zip(texts, revision_ids):
607
inv = self.deserialise_inventory(revision_id, text)
608
yield RevisionTree(self, inv, revision_id)
611
def get_ancestry(self, revision_id):
612
"""Return a list of revision-ids integrated by a revision.
614
The first element of the list is always None, indicating the origin
615
revision. This might change when we have history horizons, or
616
perhaps we should have a new API.
618
This is topologically sorted.
620
if revision_id is None:
622
if not self.has_revision(revision_id):
623
raise errors.NoSuchRevision(self, revision_id)
624
w = self.get_inventory_weave()
625
candidates = w.get_ancestry(revision_id)
626
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
629
def print_file(self, file, revision_id):
630
"""Print `file` to stdout.
632
FIXME RBC 20060125 as John Meinel points out this is a bad api
633
- it writes to stdout, it assumes that that is valid etc. Fix
634
by creating a new more flexible convenience function.
636
tree = self.revision_tree(revision_id)
637
# use inventory as it was in that revision
638
file_id = tree.inventory.path2id(file)
640
# TODO: jam 20060427 Write a test for this code path
641
# it had a bug in it, and was raising the wrong
643
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
644
tree.print_file(file_id)
646
def get_transaction(self):
647
return self.control_files.get_transaction()
649
def revision_parents(self, revid):
650
return self.get_inventory_weave().parent_names(revid)
653
def set_make_working_trees(self, new_value):
654
"""Set the policy flag for making working trees when creating branches.
656
This only applies to branches that use this repository.
658
The default is 'True'.
659
:param new_value: True to restore the default, False to disable making
662
raise NotImplementedError(self.set_make_working_trees)
664
def make_working_trees(self):
665
"""Returns the policy for making working trees on new branches."""
666
raise NotImplementedError(self.make_working_trees)
669
def sign_revision(self, revision_id, gpg_strategy):
670
plaintext = Testament.from_revision(self, revision_id).as_short_text()
671
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
674
def has_signature_for_revision_id(self, revision_id):
675
"""Query for a revision signature for revision_id in the repository."""
676
return self._revision_store.has_signature(revision_id,
677
self.get_transaction())
680
def get_signature_text(self, revision_id):
681
"""Return the text for a signature."""
682
return self._revision_store.get_signature_text(revision_id,
683
self.get_transaction())
686
def check(self, revision_ids):
687
"""Check consistency of all history of given revision_ids.
689
Different repository implementations should override _check().
691
:param revision_ids: A non-empty list of revision_ids whose ancestry
692
will be checked. Typically the last revision_id of a branch.
695
raise ValueError("revision_ids must be non-empty in %s.check"
697
return self._check(revision_ids)
699
def _check(self, revision_ids):
700
result = check.Check(self)
704
def _warn_if_deprecated(self):
705
global _deprecation_warning_done
706
if _deprecation_warning_done:
708
_deprecation_warning_done = True
709
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
710
% (self._format, self.bzrdir.transport.base))
713
class AllInOneRepository(Repository):
714
"""Legacy support - the repository behaviour for all-in-one branches."""
716
def __init__(self, _format, a_bzrdir, _revision_store, control_store, text_store):
717
# we reuse one control files instance.
718
dir_mode = a_bzrdir._control_files._dir_mode
719
file_mode = a_bzrdir._control_files._file_mode
721
def get_store(name, compressed=True, prefixed=False):
722
# FIXME: This approach of assuming stores are all entirely compressed
723
# or entirely uncompressed is tidy, but breaks upgrade from
724
# some existing branches where there's a mixture; we probably
725
# still want the option to look for both.
726
relpath = a_bzrdir._control_files._escape(name)
727
store = TextStore(a_bzrdir._control_files._transport.clone(relpath),
728
prefixed=prefixed, compressed=compressed,
731
#if self._transport.should_cache():
732
# cache_path = os.path.join(self.cache_root, name)
733
# os.mkdir(cache_path)
734
# store = bzrlib.store.CachedStore(store, cache_path)
737
# not broken out yet because the controlweaves|inventory_store
738
# and text_store | weave_store bits are still different.
739
if isinstance(_format, RepositoryFormat4):
740
# cannot remove these - there is still no consistent api
741
# which allows access to this old info.
742
self.inventory_store = get_store('inventory-store')
743
text_store = get_store('text-store')
744
super(AllInOneRepository, self).__init__(_format, a_bzrdir, a_bzrdir._control_files, _revision_store, control_store, text_store)
748
"""AllInOne repositories cannot be shared."""
752
def set_make_working_trees(self, new_value):
753
"""Set the policy flag for making working trees when creating branches.
755
This only applies to branches that use this repository.
757
The default is 'True'.
758
:param new_value: True to restore the default, False to disable making
761
raise NotImplementedError(self.set_make_working_trees)
763
def make_working_trees(self):
764
"""Returns the policy for making working trees on new branches."""
768
def install_revision(repository, rev, revision_tree):
769
"""Install all revision data into a repository."""
772
for p_id in rev.parent_ids:
773
if repository.has_revision(p_id):
774
present_parents.append(p_id)
775
parent_trees[p_id] = repository.revision_tree(p_id)
777
parent_trees[p_id] = repository.revision_tree(None)
779
inv = revision_tree.inventory
781
# backwards compatability hack: skip the root id.
782
entries = inv.iter_entries()
784
# Add the texts that are not already present
785
for path, ie in entries:
786
w = repository.weave_store.get_weave_or_empty(ie.file_id,
787
repository.get_transaction())
788
if ie.revision not in w:
790
# FIXME: TODO: The following loop *may* be overlapping/duplicate
791
# with InventoryEntry.find_previous_heads(). if it is, then there
792
# is a latent bug here where the parents may have ancestors of each
794
for revision, tree in parent_trees.iteritems():
795
if ie.file_id not in tree:
797
parent_id = tree.inventory[ie.file_id].revision
798
if parent_id in text_parents:
800
text_parents.append(parent_id)
802
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
803
repository.get_transaction())
804
lines = revision_tree.get_file(ie.file_id).readlines()
805
vfile.add_lines(rev.revision_id, text_parents, lines)
807
# install the inventory
808
repository.add_inventory(rev.revision_id, inv, present_parents)
809
except errors.RevisionAlreadyPresent:
811
repository.add_revision(rev.revision_id, rev, inv)
814
class MetaDirRepository(Repository):
815
"""Repositories in the new meta-dir layout."""
817
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
818
super(MetaDirRepository, self).__init__(_format,
824
dir_mode = self.control_files._dir_mode
825
file_mode = self.control_files._file_mode
829
"""Return True if this repository is flagged as a shared repository."""
830
return self.control_files._transport.has('shared-storage')
833
def set_make_working_trees(self, new_value):
834
"""Set the policy flag for making working trees when creating branches.
836
This only applies to branches that use this repository.
838
The default is 'True'.
839
:param new_value: True to restore the default, False to disable making
844
self.control_files._transport.delete('no-working-trees')
845
except errors.NoSuchFile:
848
self.control_files.put_utf8('no-working-trees', '')
850
def make_working_trees(self):
851
"""Returns the policy for making working trees on new branches."""
852
return not self.control_files._transport.has('no-working-trees')
855
class KnitRepository(MetaDirRepository):
856
"""Knit format repository."""
858
def _warn_if_deprecated(self):
859
# This class isn't deprecated
862
def _inventory_add_lines(self, inv_vf, revid, parents, lines):
863
inv_vf.add_lines_with_ghosts(revid, parents, lines)
866
def _all_revision_ids(self):
867
"""See Repository.all_revision_ids()."""
868
# Knits get the revision graph from the index of the revision knit, so
869
# it's always possible even if they're on an unlistable transport.
870
return self._revision_store.all_revision_ids(self.get_transaction())
872
def fileid_involved_between_revs(self, from_revid, to_revid):
873
"""Find file_id(s) which are involved in the changes between revisions.
875
This determines the set of revisions which are involved, and then
876
finds all file ids affected by those revisions.
878
vf = self._get_revision_vf()
879
from_set = set(vf.get_ancestry(from_revid))
880
to_set = set(vf.get_ancestry(to_revid))
881
changed = to_set.difference(from_set)
882
return self._fileid_involved_by_set(changed)
884
def fileid_involved(self, last_revid=None):
885
"""Find all file_ids modified in the ancestry of last_revid.
887
:param last_revid: If None, last_revision() will be used.
890
changed = set(self.all_revision_ids())
892
changed = set(self.get_ancestry(last_revid))
895
return self._fileid_involved_by_set(changed)
898
def get_ancestry(self, revision_id):
899
"""Return a list of revision-ids integrated by a revision.
901
This is topologically sorted.
903
if revision_id is None:
905
vf = self._get_revision_vf()
907
return [None] + vf.get_ancestry(revision_id)
908
except errors.RevisionNotPresent:
909
raise errors.NoSuchRevision(self, revision_id)
912
def get_revision(self, revision_id):
913
"""Return the Revision object for a named revision"""
914
return self.get_revision_reconcile(revision_id)
917
def get_revision_graph(self, revision_id=None):
918
"""Return a dictionary containing the revision graph.
920
:param revision_id: The revision_id to get a graph from. If None, then
921
the entire revision graph is returned. This is a deprecated mode of
922
operation and will be removed in the future.
923
:return: a dictionary of revision_id->revision_parents_list.
925
# special case NULL_REVISION
926
if revision_id == NULL_REVISION:
928
weave = self._get_revision_vf()
929
entire_graph = weave.get_graph()
930
if revision_id is None:
931
return weave.get_graph()
932
elif revision_id not in weave:
933
raise errors.NoSuchRevision(self, revision_id)
935
# add what can be reached from revision_id
937
pending = set([revision_id])
938
while len(pending) > 0:
940
result[node] = weave.get_parents(node)
941
for revision_id in result[node]:
942
if revision_id not in result:
943
pending.add(revision_id)
947
def get_revision_graph_with_ghosts(self, revision_ids=None):
948
"""Return a graph of the revisions with ghosts marked as applicable.
950
:param revision_ids: an iterable of revisions to graph or None for all.
951
:return: a Graph object with the graph reachable from revision_ids.
954
vf = self._get_revision_vf()
955
versions = set(vf.versions())
957
pending = set(self.all_revision_ids())
960
pending = set(revision_ids)
961
# special case NULL_REVISION
962
if NULL_REVISION in pending:
963
pending.remove(NULL_REVISION)
964
required = set(pending)
967
revision_id = pending.pop()
968
if not revision_id in versions:
969
if revision_id in required:
970
raise errors.NoSuchRevision(self, revision_id)
972
result.add_ghost(revision_id)
973
# mark it as done so we don't try for it again.
974
done.add(revision_id)
976
parent_ids = vf.get_parents_with_ghosts(revision_id)
977
for parent_id in parent_ids:
978
# is this queued or done ?
979
if (parent_id not in pending and
980
parent_id not in done):
982
pending.add(parent_id)
983
result.add_node(revision_id, parent_ids)
984
done.add(revision_id)
987
def _get_revision_vf(self):
988
""":return: a versioned file containing the revisions."""
989
vf = self._revision_store.get_revision_file(self.get_transaction())
993
def reconcile(self, other=None, thorough=False):
994
"""Reconcile this repository."""
995
from bzrlib.reconcile import KnitReconciler
996
reconciler = KnitReconciler(self, thorough=thorough)
997
reconciler.reconcile()
1000
def revision_parents(self, revision_id):
1001
return self._get_revision_vf().get_parents(revision_id)
1004
class KnitRepository2(KnitRepository):
1006
def __init__(self, _format, a_bzrdir, control_files, _revision_store,
1007
control_store, text_store):
1008
KnitRepository.__init__(self, _format, a_bzrdir, control_files,
1009
_revision_store, control_store, text_store)
1010
self._serializer = xml6.serializer_v6
1012
def deserialise_inventory(self, revision_id, xml):
1013
"""Transform the xml into an inventory object.
1015
:param revision_id: The expected revision id of the inventory.
1016
:param xml: A serialised inventory.
1018
result = self._serializer.read_inventory_from_string(xml)
1019
assert result.root.revision is not None
1022
def serialise_inventory(self, inv):
1023
"""Transform the inventory object into XML text.
1025
:param revision_id: The expected revision id of the inventory.
1026
:param xml: A serialised inventory.
1028
assert inv.revision_id is not None
1029
assert inv.root.revision is not None
1030
return KnitRepository.serialise_inventory(self, inv)
1032
def get_commit_builder(self, branch, parents, config, timestamp=None,
1033
timezone=None, committer=None, revprops=None,
1035
"""Obtain a CommitBuilder for this repository.
1037
:param branch: Branch to commit to.
1038
:param parents: Revision ids of the parents of the new revision.
1039
:param config: Configuration to use.
1040
:param timestamp: Optional timestamp recorded for commit.
1041
:param timezone: Optional timezone for timestamp.
1042
:param committer: Optional committer to set for commit.
1043
:param revprops: Optional dictionary of revision properties.
1044
:param revision_id: Optional revision id.
1046
return RootCommitBuilder(self, parents, config, timestamp, timezone,
1047
committer, revprops, revision_id)
1050
class RepositoryFormat(object):
1051
"""A repository format.
1053
Formats provide three things:
1054
* An initialization routine to construct repository data on disk.
1055
* a format string which is used when the BzrDir supports versioned
1057
* an open routine which returns a Repository instance.
1059
Formats are placed in an dict by their format string for reference
1060
during opening. These should be subclasses of RepositoryFormat
1063
Once a format is deprecated, just deprecate the initialize and open
1064
methods on the format class. Do not deprecate the object, as the
1065
object will be created every system load.
1067
Common instance attributes:
1068
_matchingbzrdir - the bzrdir format that the repository format was
1069
originally written to work with. This can be used if manually
1070
constructing a bzrdir and repository, or more commonly for test suite
1074
_default_format = None
1075
"""The default format used for new repositories."""
1078
"""The known formats."""
1081
return "<%s>" % self.__class__.__name__
1084
def find_format(klass, a_bzrdir):
1085
"""Return the format for the repository object in a_bzrdir."""
1087
transport = a_bzrdir.get_repository_transport(None)
1088
format_string = transport.get("format").read()
1089
return klass._formats[format_string]
1090
except errors.NoSuchFile:
1091
raise errors.NoRepositoryPresent(a_bzrdir)
1093
raise errors.UnknownFormatError(format=format_string)
1095
def _get_control_store(self, repo_transport, control_files):
1096
"""Return the control store for this repository."""
1097
raise NotImplementedError(self._get_control_store)
1100
def get_default_format(klass):
1101
"""Return the current default format."""
1102
return klass._default_format
1104
def get_format_string(self):
1105
"""Return the ASCII format string that identifies this format.
1107
Note that in pre format ?? repositories the format string is
1108
not permitted nor written to disk.
1110
raise NotImplementedError(self.get_format_string)
1112
def get_format_description(self):
1113
"""Return the short description for this format."""
1114
raise NotImplementedError(self.get_format_description)
1116
def _get_revision_store(self, repo_transport, control_files):
1117
"""Return the revision store object for this a_bzrdir."""
1118
raise NotImplementedError(self._get_revision_store)
1120
def _get_text_rev_store(self,
1127
"""Common logic for getting a revision store for a repository.
1129
see self._get_revision_store for the subclass-overridable method to
1130
get the store for a repository.
1132
from bzrlib.store.revision.text import TextRevisionStore
1133
dir_mode = control_files._dir_mode
1134
file_mode = control_files._file_mode
1135
text_store =TextStore(transport.clone(name),
1137
compressed=compressed,
1139
file_mode=file_mode)
1140
_revision_store = TextRevisionStore(text_store, serializer)
1141
return _revision_store
1143
def _get_versioned_file_store(self,
1148
versionedfile_class=WeaveFile,
1149
versionedfile_kwargs={},
1151
weave_transport = control_files._transport.clone(name)
1152
dir_mode = control_files._dir_mode
1153
file_mode = control_files._file_mode
1154
return VersionedFileStore(weave_transport, prefixed=prefixed,
1156
file_mode=file_mode,
1157
versionedfile_class=versionedfile_class,
1158
versionedfile_kwargs=versionedfile_kwargs,
1161
def initialize(self, a_bzrdir, shared=False):
1162
"""Initialize a repository of this format in a_bzrdir.
1164
:param a_bzrdir: The bzrdir to put the new repository in it.
1165
:param shared: The repository should be initialized as a sharable one.
1167
This may raise UninitializableFormat if shared repository are not
1168
compatible the a_bzrdir.
1171
def is_supported(self):
1172
"""Is this format supported?
1174
Supported formats must be initializable and openable.
1175
Unsupported formats may not support initialization or committing or
1176
some other features depending on the reason for not being supported.
1180
def check_conversion_target(self, target_format):
1181
raise NotImplementedError(self.check_conversion_target)
1183
def open(self, a_bzrdir, _found=False):
1184
"""Return an instance of this format for the bzrdir a_bzrdir.
1186
_found is a private parameter, do not use it.
1188
raise NotImplementedError(self.open)
1191
def register_format(klass, format):
1192
klass._formats[format.get_format_string()] = format
1195
def set_default_format(klass, format):
1196
klass._default_format = format
1199
def unregister_format(klass, format):
1200
assert klass._formats[format.get_format_string()] is format
1201
del klass._formats[format.get_format_string()]
1204
class PreSplitOutRepositoryFormat(RepositoryFormat):
1205
"""Base class for the pre split out repository formats."""
1207
rich_root_data = False
1209
def initialize(self, a_bzrdir, shared=False, _internal=False):
1210
"""Create a weave repository.
1212
TODO: when creating split out bzr branch formats, move this to a common
1213
base for Format5, Format6. or something like that.
1215
from bzrlib.weavefile import write_weave_v5
1216
from bzrlib.weave import Weave
1219
raise errors.IncompatibleFormat(self, a_bzrdir._format)
1222
# always initialized when the bzrdir is.
1223
return self.open(a_bzrdir, _found=True)
1225
# Create an empty weave
1227
write_weave_v5(Weave(), sio)
1228
empty_weave = sio.getvalue()
1230
mutter('creating repository in %s.', a_bzrdir.transport.base)
1231
dirs = ['revision-store', 'weaves']
1232
files = [('inventory.weave', StringIO(empty_weave)),
1235
# FIXME: RBC 20060125 don't peek under the covers
1236
# NB: no need to escape relative paths that are url safe.
1237
control_files = LockableFiles(a_bzrdir.transport, 'branch-lock',
1239
control_files.create_lock()
1240
control_files.lock_write()
1241
control_files._transport.mkdir_multi(dirs,
1242
mode=control_files._dir_mode)
1244
for file, content in files:
1245
control_files.put(file, content)
1247
control_files.unlock()
1248
return self.open(a_bzrdir, _found=True)
1250
def _get_control_store(self, repo_transport, control_files):
1251
"""Return the control store for this repository."""
1252
return self._get_versioned_file_store('',
1257
def _get_text_store(self, transport, control_files):
1258
"""Get a store for file texts for this format."""
1259
raise NotImplementedError(self._get_text_store)
1261
def open(self, a_bzrdir, _found=False):
1262
"""See RepositoryFormat.open()."""
1264
# we are being called directly and must probe.
1265
raise NotImplementedError
1267
repo_transport = a_bzrdir.get_repository_transport(None)
1268
control_files = a_bzrdir._control_files
1269
text_store = self._get_text_store(repo_transport, control_files)
1270
control_store = self._get_control_store(repo_transport, control_files)
1271
_revision_store = self._get_revision_store(repo_transport, control_files)
1272
return AllInOneRepository(_format=self,
1274
_revision_store=_revision_store,
1275
control_store=control_store,
1276
text_store=text_store)
1278
def check_conversion_target(self, target_format):
1282
class RepositoryFormat4(PreSplitOutRepositoryFormat):
1283
"""Bzr repository format 4.
1285
This repository format has:
1287
- TextStores for texts, inventories,revisions.
1289
This format is deprecated: it indexes texts using a text id which is
1290
removed in format 5; initialization and write support for this format
1295
super(RepositoryFormat4, self).__init__()
1296
self._matchingbzrdir = bzrdir.BzrDirFormat4()
1298
def get_format_description(self):
1299
"""See RepositoryFormat.get_format_description()."""
1300
return "Repository format 4"
1302
def initialize(self, url, shared=False, _internal=False):
1303
"""Format 4 branches cannot be created."""
1304
raise errors.UninitializableFormat(self)
1306
def is_supported(self):
1307
"""Format 4 is not supported.
1309
It is not supported because the model changed from 4 to 5 and the
1310
conversion logic is expensive - so doing it on the fly was not
1315
def _get_control_store(self, repo_transport, control_files):
1316
"""Format 4 repositories have no formal control store at this point.
1318
This will cause any control-file-needing apis to fail - this is desired.
1322
def _get_revision_store(self, repo_transport, control_files):
1323
"""See RepositoryFormat._get_revision_store()."""
1324
from bzrlib.xml4 import serializer_v4
1325
return self._get_text_rev_store(repo_transport,
1328
serializer=serializer_v4)
1330
def _get_text_store(self, transport, control_files):
1331
"""See RepositoryFormat._get_text_store()."""
1334
class RepositoryFormat5(PreSplitOutRepositoryFormat):
1335
"""Bzr control format 5.
1337
This repository format has:
1338
- weaves for file texts and inventory
1340
- TextStores for revisions and signatures.
1344
super(RepositoryFormat5, self).__init__()
1345
self._matchingbzrdir = bzrdir.BzrDirFormat5()
1347
def get_format_description(self):
1348
"""See RepositoryFormat.get_format_description()."""
1349
return "Weave repository format 5"
1351
def _get_revision_store(self, repo_transport, control_files):
1352
"""See RepositoryFormat._get_revision_store()."""
1353
"""Return the revision store object for this a_bzrdir."""
1354
return self._get_text_rev_store(repo_transport,
1359
def _get_text_store(self, transport, control_files):
1360
"""See RepositoryFormat._get_text_store()."""
1361
return self._get_versioned_file_store('weaves', transport, control_files, prefixed=False)
1364
class RepositoryFormat6(PreSplitOutRepositoryFormat):
1365
"""Bzr control format 6.
1367
This repository format has:
1368
- weaves for file texts and inventory
1369
- hash subdirectory based stores.
1370
- TextStores for revisions and signatures.
1374
super(RepositoryFormat6, self).__init__()
1375
self._matchingbzrdir = bzrdir.BzrDirFormat6()
1377
def get_format_description(self):
1378
"""See RepositoryFormat.get_format_description()."""
1379
return "Weave repository format 6"
1381
def _get_revision_store(self, repo_transport, control_files):
1382
"""See RepositoryFormat._get_revision_store()."""
1383
return self._get_text_rev_store(repo_transport,
1389
def _get_text_store(self, transport, control_files):
1390
"""See RepositoryFormat._get_text_store()."""
1391
return self._get_versioned_file_store('weaves', transport, control_files)
1394
class MetaDirRepositoryFormat(RepositoryFormat):
1395
"""Common base class for the new repositories using the metadir layout."""
1397
rich_root_data = False
1400
super(MetaDirRepositoryFormat, self).__init__()
1401
self._matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1403
def _create_control_files(self, a_bzrdir):
1404
"""Create the required files and the initial control_files object."""
1405
# FIXME: RBC 20060125 don't peek under the covers
1406
# NB: no need to escape relative paths that are url safe.
1407
repository_transport = a_bzrdir.get_repository_transport(self)
1408
control_files = LockableFiles(repository_transport, 'lock', LockDir)
1409
control_files.create_lock()
1410
return control_files
1412
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1413
"""Upload the initial blank content."""
1414
control_files = self._create_control_files(a_bzrdir)
1415
control_files.lock_write()
1417
control_files._transport.mkdir_multi(dirs,
1418
mode=control_files._dir_mode)
1419
for file, content in files:
1420
control_files.put(file, content)
1421
for file, content in utf8_files:
1422
control_files.put_utf8(file, content)
1424
control_files.put_utf8('shared-storage', '')
1426
control_files.unlock()
1429
class RepositoryFormat7(MetaDirRepositoryFormat):
1430
"""Bzr repository 7.
1432
This repository format has:
1433
- weaves for file texts and inventory
1434
- hash subdirectory based stores.
1435
- TextStores for revisions and signatures.
1436
- a format marker of its own
1437
- an optional 'shared-storage' flag
1438
- an optional 'no-working-trees' flag
1441
def _get_control_store(self, repo_transport, control_files):
1442
"""Return the control store for this repository."""
1443
return self._get_versioned_file_store('',
1448
def get_format_string(self):
1449
"""See RepositoryFormat.get_format_string()."""
1450
return "Bazaar-NG Repository format 7"
1452
def get_format_description(self):
1453
"""See RepositoryFormat.get_format_description()."""
1454
return "Weave repository format 7"
1456
def check_conversion_target(self, target_format):
1459
def _get_revision_store(self, repo_transport, control_files):
1460
"""See RepositoryFormat._get_revision_store()."""
1461
return self._get_text_rev_store(repo_transport,
1468
def _get_text_store(self, transport, control_files):
1469
"""See RepositoryFormat._get_text_store()."""
1470
return self._get_versioned_file_store('weaves',
1474
def initialize(self, a_bzrdir, shared=False):
1475
"""Create a weave repository.
1477
:param shared: If true the repository will be initialized as a shared
1480
from bzrlib.weavefile import write_weave_v5
1481
from bzrlib.weave import Weave
1483
# Create an empty weave
1485
write_weave_v5(Weave(), sio)
1486
empty_weave = sio.getvalue()
1488
mutter('creating repository in %s.', a_bzrdir.transport.base)
1489
dirs = ['revision-store', 'weaves']
1490
files = [('inventory.weave', StringIO(empty_weave)),
1492
utf8_files = [('format', self.get_format_string())]
1494
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1495
return self.open(a_bzrdir=a_bzrdir, _found=True)
1497
def open(self, a_bzrdir, _found=False, _override_transport=None):
1498
"""See RepositoryFormat.open().
1500
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1501
repository at a slightly different url
1502
than normal. I.e. during 'upgrade'.
1505
format = RepositoryFormat.find_format(a_bzrdir)
1506
assert format.__class__ == self.__class__
1507
if _override_transport is not None:
1508
repo_transport = _override_transport
1510
repo_transport = a_bzrdir.get_repository_transport(None)
1511
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1512
text_store = self._get_text_store(repo_transport, control_files)
1513
control_store = self._get_control_store(repo_transport, control_files)
1514
_revision_store = self._get_revision_store(repo_transport, control_files)
1515
return MetaDirRepository(_format=self,
1517
control_files=control_files,
1518
_revision_store=_revision_store,
1519
control_store=control_store,
1520
text_store=text_store)
1523
class RepositoryFormatKnit(MetaDirRepositoryFormat):
1524
"""Bzr repository knit format (generalized).
1526
This repository format has:
1527
- knits for file texts and inventory
1528
- hash subdirectory based stores.
1529
- knits for revisions and signatures
1530
- TextStores for revisions and signatures.
1531
- a format marker of its own
1532
- an optional 'shared-storage' flag
1533
- an optional 'no-working-trees' flag
1537
def _get_control_store(self, repo_transport, control_files):
1538
"""Return the control store for this repository."""
1539
return VersionedFileStore(
1542
file_mode=control_files._file_mode,
1543
versionedfile_class=KnitVersionedFile,
1544
versionedfile_kwargs={'factory':KnitPlainFactory()},
1547
def _get_revision_store(self, repo_transport, control_files):
1548
"""See RepositoryFormat._get_revision_store()."""
1549
from bzrlib.store.revision.knit import KnitRevisionStore
1550
versioned_file_store = VersionedFileStore(
1552
file_mode=control_files._file_mode,
1555
versionedfile_class=KnitVersionedFile,
1556
versionedfile_kwargs={'delta':False, 'factory':KnitPlainFactory(),},
1559
return KnitRevisionStore(versioned_file_store)
1561
def _get_text_store(self, transport, control_files):
1562
"""See RepositoryFormat._get_text_store()."""
1563
return self._get_versioned_file_store('knits',
1566
versionedfile_class=KnitVersionedFile,
1567
versionedfile_kwargs={
1568
'create_parent_dir':True,
1569
'delay_create':True,
1570
'dir_mode':control_files._dir_mode,
1574
def initialize(self, a_bzrdir, shared=False):
1575
"""Create a knit format 1 repository.
1577
:param a_bzrdir: bzrdir to contain the new repository; must already
1579
:param shared: If true the repository will be initialized as a shared
1582
mutter('creating repository in %s.', a_bzrdir.transport.base)
1583
dirs = ['revision-store', 'knits']
1585
utf8_files = [('format', self.get_format_string())]
1587
self._upload_blank_content(a_bzrdir, dirs, files, utf8_files, shared)
1588
repo_transport = a_bzrdir.get_repository_transport(None)
1589
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1590
control_store = self._get_control_store(repo_transport, control_files)
1591
transaction = transactions.WriteTransaction()
1592
# trigger a write of the inventory store.
1593
control_store.get_weave_or_empty('inventory', transaction)
1594
_revision_store = self._get_revision_store(repo_transport, control_files)
1595
_revision_store.has_revision_id('A', transaction)
1596
_revision_store.get_signature_file(transaction)
1597
return self.open(a_bzrdir=a_bzrdir, _found=True)
1599
def open(self, a_bzrdir, _found=False, _override_transport=None):
1600
"""See RepositoryFormat.open().
1602
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1603
repository at a slightly different url
1604
than normal. I.e. during 'upgrade'.
1607
format = RepositoryFormat.find_format(a_bzrdir)
1608
assert format.__class__ == self.__class__
1609
if _override_transport is not None:
1610
repo_transport = _override_transport
1612
repo_transport = a_bzrdir.get_repository_transport(None)
1613
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1614
text_store = self._get_text_store(repo_transport, control_files)
1615
control_store = self._get_control_store(repo_transport, control_files)
1616
_revision_store = self._get_revision_store(repo_transport, control_files)
1617
return KnitRepository(_format=self,
1619
control_files=control_files,
1620
_revision_store=_revision_store,
1621
control_store=control_store,
1622
text_store=text_store)
1625
class RepositoryFormatKnit1(RepositoryFormatKnit):
1626
"""Bzr repository knit format 1.
1628
This repository format has:
1629
- knits for file texts and inventory
1630
- hash subdirectory based stores.
1631
- knits for revisions and signatures
1632
- TextStores for revisions and signatures.
1633
- a format marker of its own
1634
- an optional 'shared-storage' flag
1635
- an optional 'no-working-trees' flag
1638
This format was introduced in bzr 0.8.
1640
def get_format_string(self):
1641
"""See RepositoryFormat.get_format_string()."""
1642
return "Bazaar-NG Knit Repository Format 1"
1644
def get_format_description(self):
1645
"""See RepositoryFormat.get_format_description()."""
1646
return "Knit repository format 1"
1648
def check_conversion_target(self, target_format):
1652
class RepositoryFormatKnit2(RepositoryFormatKnit):
1653
"""Bzr repository knit format 2.
1655
THIS FORMAT IS EXPERIMENTAL
1656
This repository format has:
1657
- knits for file texts and inventory
1658
- hash subdirectory based stores.
1659
- knits for revisions and signatures
1660
- TextStores for revisions and signatures.
1661
- a format marker of its own
1662
- an optional 'shared-storage' flag
1663
- an optional 'no-working-trees' flag
1665
- Support for recording full info about the tree root
1669
rich_root_data = True
1671
def get_format_string(self):
1672
"""See RepositoryFormat.get_format_string()."""
1673
return "Bazaar Knit Repository Format 2\n"
1675
def get_format_description(self):
1676
"""See RepositoryFormat.get_format_description()."""
1677
return "Knit repository format 2"
1679
def check_conversion_target(self, target_format):
1680
if not target_format.rich_root_data:
1681
raise errors.BadConversionTarget(
1682
'Does not support rich root data.', target_format)
1684
def open(self, a_bzrdir, _found=False, _override_transport=None):
1685
"""See RepositoryFormat.open().
1687
:param _override_transport: INTERNAL USE ONLY. Allows opening the
1688
repository at a slightly different url
1689
than normal. I.e. during 'upgrade'.
1692
format = RepositoryFormat.find_format(a_bzrdir)
1693
assert format.__class__ == self.__class__
1694
if _override_transport is not None:
1695
repo_transport = _override_transport
1697
repo_transport = a_bzrdir.get_repository_transport(None)
1698
control_files = LockableFiles(repo_transport, 'lock', LockDir)
1699
text_store = self._get_text_store(repo_transport, control_files)
1700
control_store = self._get_control_store(repo_transport, control_files)
1701
_revision_store = self._get_revision_store(repo_transport, control_files)
1702
return KnitRepository2(_format=self,
1704
control_files=control_files,
1705
_revision_store=_revision_store,
1706
control_store=control_store,
1707
text_store=text_store)
1711
# formats which have no format string are not discoverable
1712
# and not independently creatable, so are not registered.
1713
RepositoryFormat.register_format(RepositoryFormat7())
1714
_default_format = RepositoryFormatKnit1()
1715
RepositoryFormat.register_format(_default_format)
1716
RepositoryFormat.register_format(RepositoryFormatKnit2())
1717
RepositoryFormat.set_default_format(_default_format)
1718
_legacy_formats = [RepositoryFormat4(),
1719
RepositoryFormat5(),
1720
RepositoryFormat6()]
1723
class InterRepository(InterObject):
1724
"""This class represents operations taking place between two repositories.
1726
Its instances have methods like copy_content and fetch, and contain
1727
references to the source and target repositories these operations can be
1730
Often we will provide convenience methods on 'repository' which carry out
1731
operations with another repository - they will always forward to
1732
InterRepository.get(other).method_name(parameters).
1736
"""The available optimised InterRepository types."""
1738
def copy_content(self, revision_id=None, basis=None):
1739
raise NotImplementedError(self.copy_content)
1741
def fetch(self, revision_id=None, pb=None):
1742
raise NotImplementedError(self.fetch)
1745
def missing_revision_ids(self, revision_id=None):
1746
"""Return the revision ids that source has that target does not.
1748
These are returned in topological order.
1750
:param revision_id: only return revision ids included by this
1753
# generic, possibly worst case, slow code path.
1754
target_ids = set(self.target.all_revision_ids())
1755
if revision_id is not None:
1756
source_ids = self.source.get_ancestry(revision_id)
1757
assert source_ids[0] is None
1760
source_ids = self.source.all_revision_ids()
1761
result_set = set(source_ids).difference(target_ids)
1762
# this may look like a no-op: its not. It preserves the ordering
1763
# other_ids had while only returning the members from other_ids
1764
# that we've decided we need.
1765
return [rev_id for rev_id in source_ids if rev_id in result_set]
1768
class InterSameDataRepository(InterRepository):
1769
"""Code for converting between repositories that represent the same data.
1771
Data format and model must match for this to work.
1774
_matching_repo_format = RepositoryFormat4()
1775
"""Repository format for testing with."""
1778
def is_compatible(source, target):
1779
if not isinstance(source, Repository):
1781
if not isinstance(target, Repository):
1783
if source._format.rich_root_data == target._format.rich_root_data:
1789
def copy_content(self, revision_id=None, basis=None):
1790
"""Make a complete copy of the content in self into destination.
1792
This is a destructive operation! Do not use it on existing
1795
:param revision_id: Only copy the content needed to construct
1796
revision_id and its parents.
1797
:param basis: Copy the needed data preferentially from basis.
1800
self.target.set_make_working_trees(self.source.make_working_trees())
1801
except NotImplementedError:
1803
# grab the basis available data
1804
if basis is not None:
1805
self.target.fetch(basis, revision_id=revision_id)
1806
# but don't bother fetching if we have the needed data now.
1807
if (revision_id not in (None, NULL_REVISION) and
1808
self.target.has_revision(revision_id)):
1810
self.target.fetch(self.source, revision_id=revision_id)
1813
def fetch(self, revision_id=None, pb=None):
1814
"""Fetch the content required to construct revision_id.
1816
The content is copied from source to target.
1818
:param revision_id: if None all content is copied, if NULL_REVISION no
1820
:param pb: optional progress bar to use for progress reports. If not
1821
provided a default one will be created.
1823
Returns the copied revision count and the failed revisions in a tuple:
1826
from bzrlib.fetch import GenericRepoFetcher
1827
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1828
self.source, self.source._format, self.target,
1829
self.target._format)
1830
f = GenericRepoFetcher(to_repository=self.target,
1831
from_repository=self.source,
1832
last_revision=revision_id,
1834
return f.count_copied, f.failed_revisions
1837
class InterWeaveRepo(InterSameDataRepository):
1838
"""Optimised code paths between Weave based repositories."""
1840
_matching_repo_format = RepositoryFormat7()
1841
"""Repository format for testing with."""
1844
def is_compatible(source, target):
1845
"""Be compatible with known Weave formats.
1847
We don't test for the stores being of specific types because that
1848
could lead to confusing results, and there is no need to be
1852
return (isinstance(source._format, (RepositoryFormat5,
1854
RepositoryFormat7)) and
1855
isinstance(target._format, (RepositoryFormat5,
1857
RepositoryFormat7)))
1858
except AttributeError:
1862
def copy_content(self, revision_id=None, basis=None):
1863
"""See InterRepository.copy_content()."""
1864
# weave specific optimised path:
1865
if basis is not None:
1866
# copy the basis in, then fetch remaining data.
1867
basis.copy_content_into(self.target, revision_id)
1868
# the basis copy_content_into could miss-set this.
1870
self.target.set_make_working_trees(self.source.make_working_trees())
1871
except NotImplementedError:
1873
self.target.fetch(self.source, revision_id=revision_id)
1876
self.target.set_make_working_trees(self.source.make_working_trees())
1877
except NotImplementedError:
1879
# FIXME do not peek!
1880
if self.source.control_files._transport.listable():
1881
pb = ui.ui_factory.nested_progress_bar()
1883
self.target.weave_store.copy_all_ids(
1884
self.source.weave_store,
1886
from_transaction=self.source.get_transaction(),
1887
to_transaction=self.target.get_transaction())
1888
pb.update('copying inventory', 0, 1)
1889
self.target.control_weaves.copy_multi(
1890
self.source.control_weaves, ['inventory'],
1891
from_transaction=self.source.get_transaction(),
1892
to_transaction=self.target.get_transaction())
1893
self.target._revision_store.text_store.copy_all_ids(
1894
self.source._revision_store.text_store,
1899
self.target.fetch(self.source, revision_id=revision_id)
1902
def fetch(self, revision_id=None, pb=None):
1903
"""See InterRepository.fetch()."""
1904
from bzrlib.fetch import GenericRepoFetcher
1905
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1906
self.source, self.source._format, self.target, self.target._format)
1907
f = GenericRepoFetcher(to_repository=self.target,
1908
from_repository=self.source,
1909
last_revision=revision_id,
1911
return f.count_copied, f.failed_revisions
1914
def missing_revision_ids(self, revision_id=None):
1915
"""See InterRepository.missing_revision_ids()."""
1916
# we want all revisions to satisfy revision_id in source.
1917
# but we don't want to stat every file here and there.
1918
# we want then, all revisions other needs to satisfy revision_id
1919
# checked, but not those that we have locally.
1920
# so the first thing is to get a subset of the revisions to
1921
# satisfy revision_id in source, and then eliminate those that
1922
# we do already have.
1923
# this is slow on high latency connection to self, but as as this
1924
# disk format scales terribly for push anyway due to rewriting
1925
# inventory.weave, this is considered acceptable.
1927
if revision_id is not None:
1928
source_ids = self.source.get_ancestry(revision_id)
1929
assert source_ids[0] is None
1932
source_ids = self.source._all_possible_ids()
1933
source_ids_set = set(source_ids)
1934
# source_ids is the worst possible case we may need to pull.
1935
# now we want to filter source_ids against what we actually
1936
# have in target, but don't try to check for existence where we know
1937
# we do not have a revision as that would be pointless.
1938
target_ids = set(self.target._all_possible_ids())
1939
possibly_present_revisions = target_ids.intersection(source_ids_set)
1940
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1941
required_revisions = source_ids_set.difference(actually_present_revisions)
1942
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1943
if revision_id is not None:
1944
# we used get_ancestry to determine source_ids then we are assured all
1945
# revisions referenced are present as they are installed in topological order.
1946
# and the tip revision was validated by get_ancestry.
1947
return required_topo_revisions
1949
# if we just grabbed the possibly available ids, then
1950
# we only have an estimate of whats available and need to validate
1951
# that against the revision records.
1952
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1955
class InterKnitRepo(InterSameDataRepository):
1956
"""Optimised code paths between Knit based repositories."""
1958
_matching_repo_format = RepositoryFormatKnit1()
1959
"""Repository format for testing with."""
1962
def is_compatible(source, target):
1963
"""Be compatible with known Knit formats.
1965
We don't test for the stores being of specific types because that
1966
could lead to confusing results, and there is no need to be
1970
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1971
isinstance(target._format, (RepositoryFormatKnit1)))
1972
except AttributeError:
1976
def fetch(self, revision_id=None, pb=None):
1977
"""See InterRepository.fetch()."""
1978
from bzrlib.fetch import KnitRepoFetcher
1979
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1980
self.source, self.source._format, self.target, self.target._format)
1981
f = KnitRepoFetcher(to_repository=self.target,
1982
from_repository=self.source,
1983
last_revision=revision_id,
1985
return f.count_copied, f.failed_revisions
1988
def missing_revision_ids(self, revision_id=None):
1989
"""See InterRepository.missing_revision_ids()."""
1990
if revision_id is not None:
1991
source_ids = self.source.get_ancestry(revision_id)
1992
assert source_ids[0] is None
1995
source_ids = self.source._all_possible_ids()
1996
source_ids_set = set(source_ids)
1997
# source_ids is the worst possible case we may need to pull.
1998
# now we want to filter source_ids against what we actually
1999
# have in target, but don't try to check for existence where we know
2000
# we do not have a revision as that would be pointless.
2001
target_ids = set(self.target._all_possible_ids())
2002
possibly_present_revisions = target_ids.intersection(source_ids_set)
2003
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2004
required_revisions = source_ids_set.difference(actually_present_revisions)
2005
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2006
if revision_id is not None:
2007
# we used get_ancestry to determine source_ids then we are assured all
2008
# revisions referenced are present as they are installed in topological order.
2009
# and the tip revision was validated by get_ancestry.
2010
return required_topo_revisions
2012
# if we just grabbed the possibly available ids, then
2013
# we only have an estimate of whats available and need to validate
2014
# that against the revision records.
2015
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2018
class InterModel1and2(InterRepository):
2020
_matching_repo_format = None
2023
def is_compatible(source, target):
2024
if not isinstance(source, Repository):
2026
if not isinstance(target, Repository):
2028
if not source._format.rich_root_data and target._format.rich_root_data:
2034
def fetch(self, revision_id=None, pb=None):
2035
"""See InterRepository.fetch()."""
2036
from bzrlib.fetch import Model1toKnit2Fetcher
2037
f = Model1toKnit2Fetcher(to_repository=self.target,
2038
from_repository=self.source,
2039
last_revision=revision_id,
2041
return f.count_copied, f.failed_revisions
2044
def copy_content(self, revision_id=None, basis=None):
2045
"""Make a complete copy of the content in self into destination.
2047
This is a destructive operation! Do not use it on existing
2050
:param revision_id: Only copy the content needed to construct
2051
revision_id and its parents.
2052
:param basis: Copy the needed data preferentially from basis.
2055
self.target.set_make_working_trees(self.source.make_working_trees())
2056
except NotImplementedError:
2058
# grab the basis available data
2059
if basis is not None:
2060
self.target.fetch(basis, revision_id=revision_id)
2061
# but don't bother fetching if we have the needed data now.
2062
if (revision_id not in (None, NULL_REVISION) and
2063
self.target.has_revision(revision_id)):
2065
self.target.fetch(self.source, revision_id=revision_id)
2068
class InterKnit1and2(InterKnitRepo):
2070
_matching_repo_format = None
2073
def is_compatible(source, target):
2074
"""Be compatible with Knit1 source and Knit2 target"""
2076
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2077
isinstance(target._format, (RepositoryFormatKnit2)))
2078
except AttributeError:
2082
def fetch(self, revision_id=None, pb=None):
2083
"""See InterRepository.fetch()."""
2084
from bzrlib.fetch import Knit1to2Fetcher
2085
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2086
self.source, self.source._format, self.target,
2087
self.target._format)
2088
f = Knit1to2Fetcher(to_repository=self.target,
2089
from_repository=self.source,
2090
last_revision=revision_id,
2092
return f.count_copied, f.failed_revisions
2095
InterRepository.register_optimiser(InterSameDataRepository)
2096
InterRepository.register_optimiser(InterWeaveRepo)
2097
InterRepository.register_optimiser(InterKnitRepo)
2098
InterRepository.register_optimiser(InterModel1and2)
2099
InterRepository.register_optimiser(InterKnit1and2)
2102
class RepositoryTestProviderAdapter(object):
2103
"""A tool to generate a suite testing multiple repository formats at once.
2105
This is done by copying the test once for each transport and injecting
2106
the transport_server, transport_readonly_server, and bzrdir_format and
2107
repository_format classes into each copy. Each copy is also given a new id()
2108
to make it easy to identify.
2111
def __init__(self, transport_server, transport_readonly_server, formats):
2112
self._transport_server = transport_server
2113
self._transport_readonly_server = transport_readonly_server
2114
self._formats = formats
2116
def adapt(self, test):
2117
result = TestSuite()
2118
for repository_format, bzrdir_format in self._formats:
2119
new_test = deepcopy(test)
2120
new_test.transport_server = self._transport_server
2121
new_test.transport_readonly_server = self._transport_readonly_server
2122
new_test.bzrdir_format = bzrdir_format
2123
new_test.repository_format = repository_format
2124
def make_new_test_id():
2125
new_id = "%s(%s)" % (new_test.id(), repository_format.__class__.__name__)
2126
return lambda: new_id
2127
new_test.id = make_new_test_id()
2128
result.addTest(new_test)
2132
class InterRepositoryTestProviderAdapter(object):
2133
"""A tool to generate a suite testing multiple inter repository formats.
2135
This is done by copying the test once for each interrepo provider and injecting
2136
the transport_server, transport_readonly_server, repository_format and
2137
repository_to_format classes into each copy.
2138
Each copy is also given a new id() to make it easy to identify.
2141
def __init__(self, transport_server, transport_readonly_server, formats):
2142
self._transport_server = transport_server
2143
self._transport_readonly_server = transport_readonly_server
2144
self._formats = formats
2146
def adapt(self, test):
2147
result = TestSuite()
2148
for interrepo_class, repository_format, repository_format_to in self._formats:
2149
new_test = deepcopy(test)
2150
new_test.transport_server = self._transport_server
2151
new_test.transport_readonly_server = self._transport_readonly_server
2152
new_test.interrepo_class = interrepo_class
2153
new_test.repository_format = repository_format
2154
new_test.repository_format_to = repository_format_to
2155
def make_new_test_id():
2156
new_id = "%s(%s)" % (new_test.id(), interrepo_class.__name__)
2157
return lambda: new_id
2158
new_test.id = make_new_test_id()
2159
result.addTest(new_test)
2163
def default_test_list():
2164
"""Generate the default list of interrepo permutations to test."""
2166
# test the default InterRepository between format 6 and the current
2168
# XXX: robertc 20060220 reinstate this when there are two supported
2169
# formats which do not have an optimal code path between them.
2170
#result.append((InterRepository,
2171
# RepositoryFormat6(),
2172
# RepositoryFormatKnit1()))
2173
for optimiser in InterRepository._optimisers:
2174
if optimiser._matching_repo_format is not None:
2175
result.append((optimiser,
2176
optimiser._matching_repo_format,
2177
optimiser._matching_repo_format
2179
# if there are specific combinations we want to use, we can add them
2181
result.append((InterModel1and2, RepositoryFormat5(),
2182
RepositoryFormatKnit2()))
2183
result.append((InterKnit1and2, RepositoryFormatKnit1(),
2184
RepositoryFormatKnit2()))
2188
class CopyConverter(object):
2189
"""A repository conversion tool which just performs a copy of the content.
2191
This is slow but quite reliable.
2194
def __init__(self, target_format):
2195
"""Create a CopyConverter.
2197
:param target_format: The format the resulting repository should be.
2199
self.target_format = target_format
2201
def convert(self, repo, pb):
2202
"""Perform the conversion of to_convert, giving feedback via pb.
2204
:param to_convert: The disk object to convert.
2205
:param pb: a progress bar to use for progress information.
2210
# this is only useful with metadir layouts - separated repo content.
2211
# trigger an assertion if not such
2212
repo._format.get_format_string()
2213
self.repo_dir = repo.bzrdir
2214
self.step('Moving repository to repository.backup')
2215
self.repo_dir.transport.move('repository', 'repository.backup')
2216
backup_transport = self.repo_dir.transport.clone('repository.backup')
2217
repo._format.check_conversion_target(self.target_format)
2218
self.source_repo = repo._format.open(self.repo_dir,
2220
_override_transport=backup_transport)
2221
self.step('Creating new repository')
2222
converted = self.target_format.initialize(self.repo_dir,
2223
self.source_repo.is_shared())
2224
converted.lock_write()
2226
self.step('Copying content into repository.')
2227
self.source_repo.copy_content_into(converted)
2230
self.step('Deleting old repository content.')
2231
self.repo_dir.transport.delete_tree('repository.backup')
2232
self.pb.note('repository converted')
2234
def step(self, message):
2235
"""Update the pb by a step."""
2237
self.pb.update(message, self.count, self.total)
2240
class CommitBuilder(object):
2241
"""Provides an interface to build up a commit.
2243
This allows describing a tree to be committed without needing to
2244
know the internals of the format of the repository.
2247
record_root_entry = False
2248
def __init__(self, repository, parents, config, timestamp=None,
2249
timezone=None, committer=None, revprops=None,
2251
"""Initiate a CommitBuilder.
2253
:param repository: Repository to commit to.
2254
:param parents: Revision ids of the parents of the new revision.
2255
:param config: Configuration to use.
2256
:param timestamp: Optional timestamp recorded for commit.
2257
:param timezone: Optional timezone for timestamp.
2258
:param committer: Optional committer to set for commit.
2259
:param revprops: Optional dictionary of revision properties.
2260
:param revision_id: Optional revision id.
2262
self._config = config
2264
if committer is None:
2265
self._committer = self._config.username()
2267
assert isinstance(committer, basestring), type(committer)
2268
self._committer = committer
2270
self.new_inventory = Inventory(None)
2271
self._new_revision_id = revision_id
2272
self.parents = parents
2273
self.repository = repository
2276
if revprops is not None:
2277
self._revprops.update(revprops)
2279
if timestamp is None:
2280
timestamp = time.time()
2281
# Restrict resolution to 1ms
2282
self._timestamp = round(timestamp, 3)
2284
if timezone is None:
2285
self._timezone = local_time_offset()
2287
self._timezone = int(timezone)
2289
self._generate_revision_if_needed()
2291
def commit(self, message):
2292
"""Make the actual commit.
2294
:return: The revision id of the recorded revision.
2296
rev = Revision(timestamp=self._timestamp,
2297
timezone=self._timezone,
2298
committer=self._committer,
2300
inventory_sha1=self.inv_sha1,
2301
revision_id=self._new_revision_id,
2302
properties=self._revprops)
2303
rev.parent_ids = self.parents
2304
self.repository.add_revision(self._new_revision_id, rev,
2305
self.new_inventory, self._config)
2306
return self._new_revision_id
2308
def finish_inventory(self):
2309
"""Tell the builder that the inventory is finished."""
2310
if self.new_inventory.root is None:
2311
symbol_versioning.warn('Root entry should be supplied to'
2312
' record_entry_contents, as of bzr 0.10.',
2313
DeprecationWarning, stacklevel=2)
2314
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
2315
self.new_inventory.revision_id = self._new_revision_id
2316
self.inv_sha1 = self.repository.add_inventory(
2317
self._new_revision_id,
2322
def _gen_revision_id(self):
2323
"""Return new revision-id."""
2324
s = '%s-%s-' % (self._config.user_email(),
2325
compact_date(self._timestamp))
2326
s += hexlify(rand_bytes(8))
2329
def _generate_revision_if_needed(self):
2330
"""Create a revision id if None was supplied.
2332
If the repository can not support user-specified revision ids
2333
they should override this function and raise UnsupportedOperation
2334
if _new_revision_id is not None.
2336
:raises: UnsupportedOperation
2338
if self._new_revision_id is None:
2339
self._new_revision_id = self._gen_revision_id()
2341
def record_entry_contents(self, ie, parent_invs, path, tree):
2342
"""Record the content of ie from tree into the commit if needed.
2344
Side effect: sets ie.revision when unchanged
2346
:param ie: An inventory entry present in the commit.
2347
:param parent_invs: The inventories of the parent revisions of the
2349
:param path: The path the entry is at in the tree.
2350
:param tree: The tree which contains this entry and should be used to
2353
if self.new_inventory.root is None and ie.parent_id is not None:
2354
symbol_versioning.warn('Root entry should be supplied to'
2355
' record_entry_contents, as of bzr 0.10.',
2356
DeprecationWarning, stacklevel=2)
2357
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
2359
self.new_inventory.add(ie)
2361
# ie.revision is always None if the InventoryEntry is considered
2362
# for committing. ie.snapshot will record the correct revision
2363
# which may be the sole parent if it is untouched.
2364
if ie.revision is not None:
2367
# In this revision format, root entries have no knit or weave
2368
if ie is self.new_inventory.root:
2369
if len(parent_invs):
2370
ie.revision = parent_invs[0].root.revision
2374
previous_entries = ie.find_previous_heads(
2376
self.repository.weave_store,
2377
self.repository.get_transaction())
2378
# we are creating a new revision for ie in the history store
2380
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2382
def modified_directory(self, file_id, file_parents):
2383
"""Record the presence of a symbolic link.
2385
:param file_id: The file_id of the link to record.
2386
:param file_parents: The per-file parent revision ids.
2388
self._add_text_to_weave(file_id, [], file_parents.keys())
2390
def modified_file_text(self, file_id, file_parents,
2391
get_content_byte_lines, text_sha1=None,
2393
"""Record the text of file file_id
2395
:param file_id: The file_id of the file to record the text of.
2396
:param file_parents: The per-file parent revision ids.
2397
:param get_content_byte_lines: A callable which will return the byte
2399
:param text_sha1: Optional SHA1 of the file contents.
2400
:param text_size: Optional size of the file contents.
2402
# mutter('storing text of file {%s} in revision {%s} into %r',
2403
# file_id, self._new_revision_id, self.repository.weave_store)
2404
# special case to avoid diffing on renames or
2406
if (len(file_parents) == 1
2407
and text_sha1 == file_parents.values()[0].text_sha1
2408
and text_size == file_parents.values()[0].text_size):
2409
previous_ie = file_parents.values()[0]
2410
versionedfile = self.repository.weave_store.get_weave(file_id,
2411
self.repository.get_transaction())
2412
versionedfile.clone_text(self._new_revision_id,
2413
previous_ie.revision, file_parents.keys())
2414
return text_sha1, text_size
2416
new_lines = get_content_byte_lines()
2417
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2418
# should return the SHA1 and size
2419
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2420
return osutils.sha_strings(new_lines), \
2421
sum(map(len, new_lines))
2423
def modified_link(self, file_id, file_parents, link_target):
2424
"""Record the presence of a symbolic link.
2426
:param file_id: The file_id of the link to record.
2427
:param file_parents: The per-file parent revision ids.
2428
:param link_target: Target location of this link.
2430
self._add_text_to_weave(file_id, [], file_parents.keys())
2432
def _add_text_to_weave(self, file_id, new_lines, parents):
2433
versionedfile = self.repository.weave_store.get_weave_or_empty(
2434
file_id, self.repository.get_transaction())
2435
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2436
versionedfile.clear_cache()
2439
class _CommitBuilder(CommitBuilder):
2440
"""Temporary class so old CommitBuilders are detected properly
2442
Note: CommitBuilder works whether or not root entry is recorded.
2445
record_root_entry = True
2448
class RootCommitBuilder(CommitBuilder):
2449
"""This commitbuilder actually records the root id"""
2451
record_root_entry = True
2453
def record_entry_contents(self, ie, parent_invs, path, tree):
2454
"""Record the content of ie from tree into the commit if needed.
2456
Side effect: sets ie.revision when unchanged
2458
:param ie: An inventory entry present in the commit.
2459
:param parent_invs: The inventories of the parent revisions of the
2461
:param path: The path the entry is at in the tree.
2462
:param tree: The tree which contains this entry and should be used to
2465
assert self.new_inventory.root is not None or ie.parent_id is None
2466
self.new_inventory.add(ie)
2468
# ie.revision is always None if the InventoryEntry is considered
2469
# for committing. ie.snapshot will record the correct revision
2470
# which may be the sole parent if it is untouched.
2471
if ie.revision is not None:
2474
previous_entries = ie.find_previous_heads(
2476
self.repository.weave_store,
2477
self.repository.get_transaction())
2478
# we are creating a new revision for ie in the history store
2480
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2492
def _unescaper(match, _map=_unescape_map):
2493
return _map[match.group(1)]
2499
def _unescape_xml(data):
2500
"""Unescape predefined XML entities in a string of data."""
2502
if _unescape_re is None:
2503
_unescape_re = re.compile('\&([^;]*);')
2504
return _unescape_re.sub(_unescaper, data)