1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
39
revision as _mod_revision,
44
from bzrlib.bundle import serializer
45
from bzrlib.revisiontree import RevisionTree
46
from bzrlib.store.versioned import VersionedFileStore
47
from bzrlib.store.text import TextStore
48
from bzrlib.testament import Testament
51
from bzrlib.decorators import needs_read_lock, needs_write_lock
52
from bzrlib.inter import InterObject
53
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
54
from bzrlib.symbol_versioning import (
57
from bzrlib.trace import mutter, mutter_callsite, note, warning
60
# Old formats display a warning, but only once
61
_deprecation_warning_done = False
64
######################################################################
67
class Repository(object):
68
"""Repository holding history for one or more branches.
70
The repository holds and retrieves historical information including
71
revisions and file history. It's normally accessed only by the Branch,
72
which views a particular line of development through that history.
74
The Repository builds on top of Stores and a Transport, which respectively
75
describe the disk data format and the way of accessing the (possibly
79
_file_ids_altered_regex = lazy_regex.lazy_compile(
80
r'file_id="(?P<file_id>[^"]+)"'
81
r'.* revision="(?P<revision_id>[^"]+)"'
84
def abort_write_group(self):
85
"""Commit the contents accrued within the current write group.
87
:seealso: start_write_group.
89
if self._write_group is not self.get_transaction():
90
# has an unlock or relock occured ?
91
raise errors.BzrError('mismatched lock context and write group.')
92
self._abort_write_group()
93
self._write_group = None
95
def _abort_write_group(self):
96
"""Template method for per-repository write group cleanup.
98
This is called during abort before the write group is considered to be
99
finished and should cleanup any internal state accrued during the write
100
group. There is no requirement that data handed to the repository be
101
*not* made available - this is not a rollback - but neither should any
102
attempt be made to ensure that data added is fully commited. Abort is
103
invoked when an error has occured so futher disk or network operations
104
may not be possible or may error and if possible should not be
109
def add_inventory(self, revision_id, inv, parents):
110
"""Add the inventory inv to the repository as revision_id.
112
:param parents: The revision ids of the parents that revision_id
113
is known to have and are in the repository already.
115
returns the sha1 of the serialized inventory.
117
revision_id = osutils.safe_revision_id(revision_id)
118
_mod_revision.check_not_reserved_id(revision_id)
119
assert inv.revision_id is None or inv.revision_id == revision_id, \
120
"Mismatch between inventory revision" \
121
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
122
assert inv.root is not None
123
inv_lines = self._serialise_inventory_to_lines(inv)
124
inv_vf = self.get_inventory_weave()
125
return self._inventory_add_lines(inv_vf, revision_id, parents,
126
inv_lines, check_content=False)
128
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
130
"""Store lines in inv_vf and return the sha1 of the inventory."""
132
for parent in parents:
134
final_parents.append(parent)
135
return inv_vf.add_lines(revision_id, final_parents, lines,
136
check_content=check_content)[0]
139
def add_revision(self, revision_id, rev, inv=None, config=None):
140
"""Add rev to the revision store as revision_id.
142
:param revision_id: the revision id to use.
143
:param rev: The revision object.
144
:param inv: The inventory for the revision. if None, it will be looked
145
up in the inventory storer
146
:param config: If None no digital signature will be created.
147
If supplied its signature_needed method will be used
148
to determine if a signature should be made.
150
revision_id = osutils.safe_revision_id(revision_id)
151
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
153
_mod_revision.check_not_reserved_id(revision_id)
154
if config is not None and config.signature_needed():
156
inv = self.get_inventory(revision_id)
157
plaintext = Testament(rev, inv).as_short_text()
158
self.store_revision_signature(
159
gpg.GPGStrategy(config), plaintext, revision_id)
160
if not revision_id in self.get_inventory_weave():
162
raise errors.WeaveRevisionNotPresent(revision_id,
163
self.get_inventory_weave())
165
# yes, this is not suitable for adding with ghosts.
166
self.add_inventory(revision_id, inv, rev.parent_ids)
167
self._revision_store.add_revision(rev, self.get_transaction())
169
def _add_revision_text(self, revision_id, text):
170
revision = self._revision_store._serializer.read_revision_from_string(
172
self._revision_store._add_revision(revision, StringIO(text),
173
self.get_transaction())
176
def _all_possible_ids(self):
177
"""Return all the possible revisions that we could find."""
178
return self.get_inventory_weave().versions()
180
def all_revision_ids(self):
181
"""Returns a list of all the revision ids in the repository.
183
This is deprecated because code should generally work on the graph
184
reachable from a particular revision, and ignore any other revisions
185
that might be present. There is no direct replacement method.
187
return self._all_revision_ids()
190
def _all_revision_ids(self):
191
"""Returns a list of all the revision ids in the repository.
193
These are in as much topological order as the underlying store can
194
present: for weaves ghosts may lead to a lack of correctness until
195
the reweave updates the parents list.
197
if self._revision_store.text_store.listable():
198
return self._revision_store.all_revision_ids(self.get_transaction())
199
result = self._all_possible_ids()
200
# TODO: jam 20070210 Ensure that _all_possible_ids returns non-unicode
201
# ids. (It should, since _revision_store's API should change to
202
# return utf8 revision_ids)
203
return self._eliminate_revisions_not_present(result)
205
def break_lock(self):
206
"""Break a lock if one is present from another instance.
208
Uses the ui factory to ask for confirmation if the lock may be from
211
self.control_files.break_lock()
214
def _eliminate_revisions_not_present(self, revision_ids):
215
"""Check every revision id in revision_ids to see if we have it.
217
Returns a set of the present revisions.
220
for id in revision_ids:
221
if self.has_revision(id):
226
def create(a_bzrdir):
227
"""Construct the current default format repository in a_bzrdir."""
228
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
230
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
231
"""instantiate a Repository.
233
:param _format: The format of the repository on disk.
234
:param a_bzrdir: The BzrDir of the repository.
236
In the future we will have a single api for all stores for
237
getting file texts, inventories and revisions, then
238
this construct will accept instances of those things.
240
super(Repository, self).__init__()
241
self._format = _format
242
# the following are part of the public API for Repository:
243
self.bzrdir = a_bzrdir
244
self.control_files = control_files
245
self._revision_store = _revision_store
246
# backwards compatibility
247
self.weave_store = text_store
249
self._reconcile_does_inventory_gc = True
250
# not right yet - should be more semantically clear ?
252
self.control_store = control_store
253
self.control_weaves = control_store
254
# TODO: make sure to construct the right store classes, etc, depending
255
# on whether escaping is required.
256
self._warn_if_deprecated()
257
self._write_group = None
260
return '%s(%r)' % (self.__class__.__name__,
261
self.bzrdir.transport.base)
263
def has_same_location(self, other):
264
"""Returns a boolean indicating if this repository is at the same
265
location as another repository.
267
This might return False even when two repository objects are accessing
268
the same physical repository via different URLs.
270
if self.__class__ is not other.__class__:
272
return (self.control_files._transport.base ==
273
other.control_files._transport.base)
275
def is_in_write_group(self):
276
"""Return True if there is an open write group.
278
:seealso: start_write_group.
280
return self._write_group is not None
283
return self.control_files.is_locked()
285
def lock_write(self, token=None):
286
"""Lock this repository for writing.
288
This causes caching within the repository obejct to start accumlating
289
data during reads, and allows a 'write_group' to be obtained. Write
290
groups must be used for actual data insertion.
292
:param token: if this is already locked, then lock_write will fail
293
unless the token matches the existing lock.
294
:returns: a token if this instance supports tokens, otherwise None.
295
:raises TokenLockingNotSupported: when a token is given but this
296
instance doesn't support using token locks.
297
:raises MismatchedToken: if the specified token doesn't match the token
298
of the existing lock.
299
:seealso: start_write_group.
301
A token should be passed in if you know that you have locked the object
302
some other way, and need to synchronise this object's state with that
305
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
307
result = self.control_files.lock_write(token=token)
312
self.control_files.lock_read()
315
def get_physical_lock_status(self):
316
return self.control_files.get_physical_lock_status()
318
def leave_lock_in_place(self):
319
"""Tell this repository not to release the physical lock when this
322
If lock_write doesn't return a token, then this method is not supported.
324
self.control_files.leave_in_place()
326
def dont_leave_lock_in_place(self):
327
"""Tell this repository to release the physical lock when this
328
object is unlocked, even if it didn't originally acquire it.
330
If lock_write doesn't return a token, then this method is not supported.
332
self.control_files.dont_leave_in_place()
335
def gather_stats(self, revid=None, committers=None):
336
"""Gather statistics from a revision id.
338
:param revid: The revision id to gather statistics from, if None, then
339
no revision specific statistics are gathered.
340
:param committers: Optional parameter controlling whether to grab
341
a count of committers from the revision specific statistics.
342
:return: A dictionary of statistics. Currently this contains:
343
committers: The number of committers if requested.
344
firstrev: A tuple with timestamp, timezone for the penultimate left
345
most ancestor of revid, if revid is not the NULL_REVISION.
346
latestrev: A tuple with timestamp, timezone for revid, if revid is
347
not the NULL_REVISION.
348
revisions: The total revision count in the repository.
349
size: An estimate disk size of the repository in bytes.
352
if revid and committers:
353
result['committers'] = 0
354
if revid and revid != _mod_revision.NULL_REVISION:
356
all_committers = set()
357
revisions = self.get_ancestry(revid)
358
# pop the leading None
360
first_revision = None
362
# ignore the revisions in the middle - just grab first and last
363
revisions = revisions[0], revisions[-1]
364
for revision in self.get_revisions(revisions):
365
if not first_revision:
366
first_revision = revision
368
all_committers.add(revision.committer)
369
last_revision = revision
371
result['committers'] = len(all_committers)
372
result['firstrev'] = (first_revision.timestamp,
373
first_revision.timezone)
374
result['latestrev'] = (last_revision.timestamp,
375
last_revision.timezone)
377
# now gather global repository information
378
if self.bzrdir.root_transport.listable():
379
c, t = self._revision_store.total_size(self.get_transaction())
380
result['revisions'] = c
385
def missing_revision_ids(self, other, revision_id=None):
386
"""Return the revision ids that other has that this does not.
388
These are returned in topological order.
390
revision_id: only return revision ids included by revision_id.
392
revision_id = osutils.safe_revision_id(revision_id)
393
return InterRepository.get(other, self).missing_revision_ids(revision_id)
397
"""Open the repository rooted at base.
399
For instance, if the repository is at URL/.bzr/repository,
400
Repository.open(URL) -> a Repository instance.
402
control = bzrdir.BzrDir.open(base)
403
return control.open_repository()
405
def copy_content_into(self, destination, revision_id=None):
406
"""Make a complete copy of the content in self into destination.
408
This is a destructive operation! Do not use it on existing
411
revision_id = osutils.safe_revision_id(revision_id)
412
return InterRepository.get(self, destination).copy_content(revision_id)
414
def commit_write_group(self):
415
"""Commit the contents accrued within the current write group.
417
:seealso: start_write_group.
419
if self._write_group is not self.get_transaction():
420
# has an unlock or relock occured ?
421
raise errors.BzrError('mismatched lock context and write group.')
422
self._commit_write_group()
423
self._write_group = None
425
def _commit_write_group(self):
426
"""Template method for per-repository write group cleanup.
428
This is called before the write group is considered to be
429
finished and should ensure that all data handed to the repository
430
for writing during the write group is safely committed (to the
431
extent possible considering file system caching etc).
434
def fetch(self, source, revision_id=None, pb=None):
435
"""Fetch the content required to construct revision_id from source.
437
If revision_id is None all content is copied.
439
revision_id = osutils.safe_revision_id(revision_id)
440
inter = InterRepository.get(source, self)
442
return inter.fetch(revision_id=revision_id, pb=pb)
443
except NotImplementedError:
444
raise errors.IncompatibleRepositories(source, self)
446
def create_bundle(self, target, base, fileobj, format=None):
447
return serializer.write_bundle(self, target, base, fileobj, format)
449
def get_commit_builder(self, branch, parents, config, timestamp=None,
450
timezone=None, committer=None, revprops=None,
452
"""Obtain a CommitBuilder for this repository.
454
:param branch: Branch to commit to.
455
:param parents: Revision ids of the parents of the new revision.
456
:param config: Configuration to use.
457
:param timestamp: Optional timestamp recorded for commit.
458
:param timezone: Optional timezone for timestamp.
459
:param committer: Optional committer to set for commit.
460
:param revprops: Optional dictionary of revision properties.
461
:param revision_id: Optional revision id.
463
revision_id = osutils.safe_revision_id(revision_id)
464
result = CommitBuilder(self, parents, config, timestamp, timezone,
465
committer, revprops, revision_id)
466
self.start_write_group()
470
if (self.control_files._lock_count == 1 and
471
self.control_files._lock_mode == 'w'):
472
if self._write_group is not None:
473
raise errors.BzrError(
474
'Must end write groups before releasing write locks.')
475
self.control_files.unlock()
478
def clone(self, a_bzrdir, revision_id=None):
479
"""Clone this repository into a_bzrdir using the current format.
481
Currently no check is made that the format of this repository and
482
the bzrdir format are compatible. FIXME RBC 20060201.
484
:return: The newly created destination repository.
486
# TODO: deprecate after 0.16; cloning this with all its settings is
487
# probably not very useful -- mbp 20070423
488
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
489
self.copy_content_into(dest_repo, revision_id)
492
def start_write_group(self):
493
"""Start a write group in the repository.
495
Write groups are used by repositories which do not have a 1:1 mapping
496
between file ids and backend store to manage the insertion of data from
497
both fetch and commit operations.
499
A write lock is required around the start_write_group/commit_write_group
500
for the support of lock-requiring repository formats.
502
One can only insert data into a repository inside a write group.
506
if not self.is_locked() or self.control_files._lock_mode != 'w':
507
raise errors.NotWriteLocked(self)
508
if self._write_group:
509
raise errors.BzrError('already in a write group')
510
self._start_write_group()
511
# so we can detect unlock/relock - the write group is now entered.
512
self._write_group = self.get_transaction()
514
def _start_write_group(self):
515
"""Template method for per-repository write group startup.
517
This is called before the write group is considered to be
522
def sprout(self, to_bzrdir, revision_id=None):
523
"""Create a descendent repository for new development.
525
Unlike clone, this does not copy the settings of the repository.
527
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
528
dest_repo.fetch(self, revision_id=revision_id)
531
def _create_sprouting_repo(self, a_bzrdir, shared):
532
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
533
# use target default format.
534
dest_repo = a_bzrdir.create_repository()
536
# Most control formats need the repository to be specifically
537
# created, but on some old all-in-one formats it's not needed
539
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
540
except errors.UninitializableFormat:
541
dest_repo = a_bzrdir.open_repository()
545
def has_revision(self, revision_id):
546
"""True if this repository has a copy of the revision."""
547
if 'evil' in debug.debug_flags:
548
mutter_callsite(2, "has_revision is a LBYL symptom.")
549
revision_id = osutils.safe_revision_id(revision_id)
550
return self._revision_store.has_revision_id(revision_id,
551
self.get_transaction())
554
def get_revision_reconcile(self, revision_id):
555
"""'reconcile' helper routine that allows access to a revision always.
557
This variant of get_revision does not cross check the weave graph
558
against the revision one as get_revision does: but it should only
559
be used by reconcile, or reconcile-alike commands that are correcting
560
or testing the revision graph.
562
if not revision_id or not isinstance(revision_id, basestring):
563
raise errors.InvalidRevisionId(revision_id=revision_id,
565
return self.get_revisions([revision_id])[0]
568
def get_revisions(self, revision_ids):
569
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
570
revs = self._revision_store.get_revisions(revision_ids,
571
self.get_transaction())
573
assert not isinstance(rev.revision_id, unicode)
574
for parent_id in rev.parent_ids:
575
assert not isinstance(parent_id, unicode)
579
def get_revision_xml(self, revision_id):
580
# TODO: jam 20070210 This shouldn't be necessary since get_revision
581
# would have already do it.
582
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
583
revision_id = osutils.safe_revision_id(revision_id)
584
rev = self.get_revision(revision_id)
586
# the current serializer..
587
self._revision_store._serializer.write_revision(rev, rev_tmp)
589
return rev_tmp.getvalue()
592
def get_revision(self, revision_id):
593
"""Return the Revision object for a named revision"""
594
# TODO: jam 20070210 get_revision_reconcile should do this for us
595
revision_id = osutils.safe_revision_id(revision_id)
596
r = self.get_revision_reconcile(revision_id)
597
# weave corruption can lead to absent revision markers that should be
599
# the following test is reasonably cheap (it needs a single weave read)
600
# and the weave is cached in read transactions. In write transactions
601
# it is not cached but typically we only read a small number of
602
# revisions. For knits when they are introduced we will probably want
603
# to ensure that caching write transactions are in use.
604
inv = self.get_inventory_weave()
605
self._check_revision_parents(r, inv)
609
def get_deltas_for_revisions(self, revisions):
610
"""Produce a generator of revision deltas.
612
Note that the input is a sequence of REVISIONS, not revision_ids.
613
Trees will be held in memory until the generator exits.
614
Each delta is relative to the revision's lefthand predecessor.
616
required_trees = set()
617
for revision in revisions:
618
required_trees.add(revision.revision_id)
619
required_trees.update(revision.parent_ids[:1])
620
trees = dict((t.get_revision_id(), t) for
621
t in self.revision_trees(required_trees))
622
for revision in revisions:
623
if not revision.parent_ids:
624
old_tree = self.revision_tree(None)
626
old_tree = trees[revision.parent_ids[0]]
627
yield trees[revision.revision_id].changes_from(old_tree)
630
def get_revision_delta(self, revision_id):
631
"""Return the delta for one revision.
633
The delta is relative to the left-hand predecessor of the
636
r = self.get_revision(revision_id)
637
return list(self.get_deltas_for_revisions([r]))[0]
639
def _check_revision_parents(self, revision, inventory):
640
"""Private to Repository and Fetch.
642
This checks the parentage of revision in an inventory weave for
643
consistency and is only applicable to inventory-weave-for-ancestry
644
using repository formats & fetchers.
646
weave_parents = inventory.get_parents(revision.revision_id)
647
weave_names = inventory.versions()
648
for parent_id in revision.parent_ids:
649
if parent_id in weave_names:
650
# this parent must not be a ghost.
651
if not parent_id in weave_parents:
653
raise errors.CorruptRepository(self)
656
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
657
revision_id = osutils.safe_revision_id(revision_id)
658
signature = gpg_strategy.sign(plaintext)
659
self._revision_store.add_revision_signature_text(revision_id,
661
self.get_transaction())
663
def fileids_altered_by_revision_ids(self, revision_ids):
664
"""Find the file ids and versions affected by revisions.
666
:param revisions: an iterable containing revision ids.
667
:return: a dictionary mapping altered file-ids to an iterable of
668
revision_ids. Each altered file-ids has the exact revision_ids that
669
altered it listed explicitly.
671
assert self._serializer.support_altered_by_hack, \
672
("fileids_altered_by_revision_ids only supported for branches "
673
"which store inventory as unnested xml, not on %r" % self)
674
selected_revision_ids = set(osutils.safe_revision_id(r)
675
for r in revision_ids)
676
w = self.get_inventory_weave()
679
# this code needs to read every new line in every inventory for the
680
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
681
# not present in one of those inventories is unnecessary but not
682
# harmful because we are filtering by the revision id marker in the
683
# inventory lines : we only select file ids altered in one of those
684
# revisions. We don't need to see all lines in the inventory because
685
# only those added in an inventory in rev X can contain a revision=X
687
unescape_revid_cache = {}
688
unescape_fileid_cache = {}
690
# jam 20061218 In a big fetch, this handles hundreds of thousands
691
# of lines, so it has had a lot of inlining and optimizing done.
692
# Sorry that it is a little bit messy.
693
# Move several functions to be local variables, since this is a long
695
search = self._file_ids_altered_regex.search
696
unescape = _unescape_xml
697
setdefault = result.setdefault
698
pb = ui.ui_factory.nested_progress_bar()
700
for line in w.iter_lines_added_or_present_in_versions(
701
selected_revision_ids, pb=pb):
705
# One call to match.group() returning multiple items is quite a
706
# bit faster than 2 calls to match.group() each returning 1
707
file_id, revision_id = match.group('file_id', 'revision_id')
709
# Inlining the cache lookups helps a lot when you make 170,000
710
# lines and 350k ids, versus 8.4 unique ids.
711
# Using a cache helps in 2 ways:
712
# 1) Avoids unnecessary decoding calls
713
# 2) Re-uses cached strings, which helps in future set and
715
# (2) is enough that removing encoding entirely along with
716
# the cache (so we are using plain strings) results in no
717
# performance improvement.
719
revision_id = unescape_revid_cache[revision_id]
721
unescaped = unescape(revision_id)
722
unescape_revid_cache[revision_id] = unescaped
723
revision_id = unescaped
725
if revision_id in selected_revision_ids:
727
file_id = unescape_fileid_cache[file_id]
729
unescaped = unescape(file_id)
730
unescape_fileid_cache[file_id] = unescaped
732
setdefault(file_id, set()).add(revision_id)
737
def iter_files_bytes(self, desired_files):
738
"""Iterate through file versions.
740
Files will not necessarily be returned in the order they occur in
741
desired_files. No specific order is guaranteed.
743
Yields pairs of identifier, bytes_iterator. identifier is an opaque
744
value supplied by the caller as part of desired_files. It should
745
uniquely identify the file version in the caller's context. (Examples:
746
an index number or a TreeTransform trans_id.)
748
bytes_iterator is an iterable of bytestrings for the file. The
749
kind of iterable and length of the bytestrings are unspecified, but for
750
this implementation, it is a list of lines produced by
751
VersionedFile.get_lines().
753
:param desired_files: a list of (file_id, revision_id, identifier)
756
transaction = self.get_transaction()
757
for file_id, revision_id, callable_data in desired_files:
759
weave = self.weave_store.get_weave(file_id, transaction)
760
except errors.NoSuchFile:
761
raise errors.NoSuchIdInRepository(self, file_id)
762
yield callable_data, weave.get_lines(revision_id)
764
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
765
"""Get an iterable listing the keys of all the data introduced by a set
768
The keys will be ordered so that the corresponding items can be safely
769
fetched and inserted in that order.
771
:returns: An iterable producing tuples of (knit-kind, file-id,
772
versions). knit-kind is one of 'file', 'inventory', 'signatures',
773
'revisions'. file-id is None unless knit-kind is 'file'.
775
# XXX: it's a bit weird to control the inventory weave caching in this
776
# generator. Ideally the caching would be done in fetch.py I think. Or
777
# maybe this generator should explicitly have the contract that it
778
# should not be iterated until the previously yielded item has been
780
inv_w = self.get_inventory_weave()
783
# file ids that changed
784
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
786
num_file_ids = len(file_ids)
787
for file_id, altered_versions in file_ids.iteritems():
788
if _files_pb is not None:
789
_files_pb.update("fetch texts", count, num_file_ids)
791
yield ("file", file_id, altered_versions)
792
# We're done with the files_pb. Note that it finished by the caller,
793
# just as it was created by the caller.
797
yield ("inventory", None, revision_ids)
801
revisions_with_signatures = set()
802
for rev_id in revision_ids:
804
self.get_signature_text(rev_id)
805
except errors.NoSuchRevision:
809
revisions_with_signatures.add(rev_id)
810
yield ("signatures", None, revisions_with_signatures)
813
yield ("revisions", None, revision_ids)
816
def get_inventory_weave(self):
817
return self.control_weaves.get_weave('inventory',
818
self.get_transaction())
821
def get_inventory(self, revision_id):
822
"""Get Inventory object by hash."""
823
# TODO: jam 20070210 Technically we don't need to sanitize, since all
824
# called functions must sanitize.
825
revision_id = osutils.safe_revision_id(revision_id)
826
return self.deserialise_inventory(
827
revision_id, self.get_inventory_xml(revision_id))
829
def deserialise_inventory(self, revision_id, xml):
830
"""Transform the xml into an inventory object.
832
:param revision_id: The expected revision id of the inventory.
833
:param xml: A serialised inventory.
835
revision_id = osutils.safe_revision_id(revision_id)
836
result = self._serializer.read_inventory_from_string(xml)
837
result.root.revision = revision_id
840
def serialise_inventory(self, inv):
841
return self._serializer.write_inventory_to_string(inv)
843
def _serialise_inventory_to_lines(self, inv):
844
return self._serializer.write_inventory_to_lines(inv)
846
def get_serializer_format(self):
847
return self._serializer.format_num
850
def get_inventory_xml(self, revision_id):
851
"""Get inventory XML as a file object."""
852
revision_id = osutils.safe_revision_id(revision_id)
854
assert isinstance(revision_id, str), type(revision_id)
855
iw = self.get_inventory_weave()
856
return iw.get_text(revision_id)
858
raise errors.HistoryMissing(self, 'inventory', revision_id)
861
def get_inventory_sha1(self, revision_id):
862
"""Return the sha1 hash of the inventory entry
864
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
865
revision_id = osutils.safe_revision_id(revision_id)
866
return self.get_revision(revision_id).inventory_sha1
869
def get_revision_graph(self, revision_id=None):
870
"""Return a dictionary containing the revision graph.
872
:param revision_id: The revision_id to get a graph from. If None, then
873
the entire revision graph is returned. This is a deprecated mode of
874
operation and will be removed in the future.
875
:return: a dictionary of revision_id->revision_parents_list.
877
if 'evil' in debug.debug_flags:
879
"get_revision_graph scales with size of history.")
880
# special case NULL_REVISION
881
if revision_id == _mod_revision.NULL_REVISION:
883
revision_id = osutils.safe_revision_id(revision_id)
884
a_weave = self.get_inventory_weave()
885
all_revisions = self._eliminate_revisions_not_present(
887
entire_graph = dict([(node, tuple(a_weave.get_parents(node))) for
888
node in all_revisions])
889
if revision_id is None:
891
elif revision_id not in entire_graph:
892
raise errors.NoSuchRevision(self, revision_id)
894
# add what can be reached from revision_id
896
pending = set([revision_id])
897
while len(pending) > 0:
899
result[node] = entire_graph[node]
900
for revision_id in result[node]:
901
if revision_id not in result:
902
pending.add(revision_id)
906
def get_revision_graph_with_ghosts(self, revision_ids=None):
907
"""Return a graph of the revisions with ghosts marked as applicable.
909
:param revision_ids: an iterable of revisions to graph or None for all.
910
:return: a Graph object with the graph reachable from revision_ids.
912
if 'evil' in debug.debug_flags:
914
"get_revision_graph_with_ghosts scales with size of history.")
915
result = deprecated_graph.Graph()
917
pending = set(self.all_revision_ids())
920
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
921
# special case NULL_REVISION
922
if _mod_revision.NULL_REVISION in pending:
923
pending.remove(_mod_revision.NULL_REVISION)
924
required = set(pending)
927
revision_id = pending.pop()
929
rev = self.get_revision(revision_id)
930
except errors.NoSuchRevision:
931
if revision_id in required:
934
result.add_ghost(revision_id)
936
for parent_id in rev.parent_ids:
937
# is this queued or done ?
938
if (parent_id not in pending and
939
parent_id not in done):
941
pending.add(parent_id)
942
result.add_node(revision_id, rev.parent_ids)
943
done.add(revision_id)
946
def _get_history_vf(self):
947
"""Get a versionedfile whose history graph reflects all revisions.
949
For weave repositories, this is the inventory weave.
951
return self.get_inventory_weave()
953
def iter_reverse_revision_history(self, revision_id):
954
"""Iterate backwards through revision ids in the lefthand history
956
:param revision_id: The revision id to start with. All its lefthand
957
ancestors will be traversed.
959
revision_id = osutils.safe_revision_id(revision_id)
960
if revision_id in (None, _mod_revision.NULL_REVISION):
962
next_id = revision_id
963
versionedfile = self._get_history_vf()
966
parents = versionedfile.get_parents(next_id)
967
if len(parents) == 0:
973
def get_revision_inventory(self, revision_id):
974
"""Return inventory of a past revision."""
975
# TODO: Unify this with get_inventory()
976
# bzr 0.0.6 and later imposes the constraint that the inventory_id
977
# must be the same as its revision, so this is trivial.
978
if revision_id is None:
979
# This does not make sense: if there is no revision,
980
# then it is the current tree inventory surely ?!
981
# and thus get_root_id() is something that looks at the last
982
# commit on the branch, and the get_root_id is an inventory check.
983
raise NotImplementedError
984
# return Inventory(self.get_root_id())
986
return self.get_inventory(revision_id)
990
"""Return True if this repository is flagged as a shared repository."""
991
raise NotImplementedError(self.is_shared)
994
def reconcile(self, other=None, thorough=False):
995
"""Reconcile this repository."""
996
from bzrlib.reconcile import RepoReconciler
997
reconciler = RepoReconciler(self, thorough=thorough)
998
reconciler.reconcile()
1001
def _refresh_data(self):
1002
"""Helper called from lock_* to ensure coherency with disk.
1004
The default implementation does nothing; it is however possible
1005
for repositories to maintain loaded indices across multiple locks
1006
by checking inside their implementation of this method to see
1007
whether their indices are still valid. This depends of course on
1008
the disk format being validatable in this manner.
1012
def revision_tree(self, revision_id):
1013
"""Return Tree for a revision on this branch.
1015
`revision_id` may be None for the empty tree revision.
1017
# TODO: refactor this to use an existing revision object
1018
# so we don't need to read it in twice.
1019
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1020
return RevisionTree(self, Inventory(root_id=None),
1021
_mod_revision.NULL_REVISION)
1023
revision_id = osutils.safe_revision_id(revision_id)
1024
inv = self.get_revision_inventory(revision_id)
1025
return RevisionTree(self, inv, revision_id)
1028
def revision_trees(self, revision_ids):
1029
"""Return Tree for a revision on this branch.
1031
`revision_id` may not be None or 'null:'"""
1032
assert None not in revision_ids
1033
assert _mod_revision.NULL_REVISION not in revision_ids
1034
texts = self.get_inventory_weave().get_texts(revision_ids)
1035
for text, revision_id in zip(texts, revision_ids):
1036
inv = self.deserialise_inventory(revision_id, text)
1037
yield RevisionTree(self, inv, revision_id)
1040
def get_ancestry(self, revision_id, topo_sorted=True):
1041
"""Return a list of revision-ids integrated by a revision.
1043
The first element of the list is always None, indicating the origin
1044
revision. This might change when we have history horizons, or
1045
perhaps we should have a new API.
1047
This is topologically sorted.
1049
if _mod_revision.is_null(revision_id):
1051
revision_id = osutils.safe_revision_id(revision_id)
1052
if not self.has_revision(revision_id):
1053
raise errors.NoSuchRevision(self, revision_id)
1054
w = self.get_inventory_weave()
1055
candidates = w.get_ancestry(revision_id, topo_sorted)
1056
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1059
"""Compress the data within the repository.
1061
This operation only makes sense for some repository types. For other
1062
types it should be a no-op that just returns.
1064
This stub method does not require a lock, but subclasses should use
1065
@needs_write_lock as this is a long running call its reasonable to
1066
implicitly lock for the user.
1070
def print_file(self, file, revision_id):
1071
"""Print `file` to stdout.
1073
FIXME RBC 20060125 as John Meinel points out this is a bad api
1074
- it writes to stdout, it assumes that that is valid etc. Fix
1075
by creating a new more flexible convenience function.
1077
revision_id = osutils.safe_revision_id(revision_id)
1078
tree = self.revision_tree(revision_id)
1079
# use inventory as it was in that revision
1080
file_id = tree.inventory.path2id(file)
1082
# TODO: jam 20060427 Write a test for this code path
1083
# it had a bug in it, and was raising the wrong
1085
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1086
tree.print_file(file_id)
1088
def get_transaction(self):
1089
return self.control_files.get_transaction()
1091
def revision_parents(self, revision_id):
1092
revision_id = osutils.safe_revision_id(revision_id)
1093
return self.get_inventory_weave().parent_names(revision_id)
1095
def get_parents(self, revision_ids):
1096
"""See StackedParentsProvider.get_parents"""
1098
for revision_id in revision_ids:
1099
if revision_id == _mod_revision.NULL_REVISION:
1103
parents = self.get_revision(revision_id).parent_ids
1104
except errors.NoSuchRevision:
1107
if len(parents) == 0:
1108
parents = [_mod_revision.NULL_REVISION]
1109
parents_list.append(parents)
1112
def _make_parents_provider(self):
1115
def get_graph(self, other_repository=None):
1116
"""Return the graph walker for this repository format"""
1117
parents_provider = self._make_parents_provider()
1118
if (other_repository is not None and
1119
other_repository.bzrdir.transport.base !=
1120
self.bzrdir.transport.base):
1121
parents_provider = graph._StackedParentsProvider(
1122
[parents_provider, other_repository._make_parents_provider()])
1123
return graph.Graph(parents_provider)
1126
def set_make_working_trees(self, new_value):
1127
"""Set the policy flag for making working trees when creating branches.
1129
This only applies to branches that use this repository.
1131
The default is 'True'.
1132
:param new_value: True to restore the default, False to disable making
1135
raise NotImplementedError(self.set_make_working_trees)
1137
def make_working_trees(self):
1138
"""Returns the policy for making working trees on new branches."""
1139
raise NotImplementedError(self.make_working_trees)
1142
def sign_revision(self, revision_id, gpg_strategy):
1143
revision_id = osutils.safe_revision_id(revision_id)
1144
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1145
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1148
def has_signature_for_revision_id(self, revision_id):
1149
"""Query for a revision signature for revision_id in the repository."""
1150
revision_id = osutils.safe_revision_id(revision_id)
1151
return self._revision_store.has_signature(revision_id,
1152
self.get_transaction())
1155
def get_signature_text(self, revision_id):
1156
"""Return the text for a signature."""
1157
revision_id = osutils.safe_revision_id(revision_id)
1158
return self._revision_store.get_signature_text(revision_id,
1159
self.get_transaction())
1162
def check(self, revision_ids):
1163
"""Check consistency of all history of given revision_ids.
1165
Different repository implementations should override _check().
1167
:param revision_ids: A non-empty list of revision_ids whose ancestry
1168
will be checked. Typically the last revision_id of a branch.
1170
if not revision_ids:
1171
raise ValueError("revision_ids must be non-empty in %s.check"
1173
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
1174
return self._check(revision_ids)
1176
def _check(self, revision_ids):
1177
result = check.Check(self)
1181
def _warn_if_deprecated(self):
1182
global _deprecation_warning_done
1183
if _deprecation_warning_done:
1185
_deprecation_warning_done = True
1186
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1187
% (self._format, self.bzrdir.transport.base))
1189
def supports_rich_root(self):
1190
return self._format.rich_root_data
1192
def _check_ascii_revisionid(self, revision_id, method):
1193
"""Private helper for ascii-only repositories."""
1194
# weave repositories refuse to store revisionids that are non-ascii.
1195
if revision_id is not None:
1196
# weaves require ascii revision ids.
1197
if isinstance(revision_id, unicode):
1199
revision_id.encode('ascii')
1200
except UnicodeEncodeError:
1201
raise errors.NonAsciiRevisionId(method, self)
1204
revision_id.decode('ascii')
1205
except UnicodeDecodeError:
1206
raise errors.NonAsciiRevisionId(method, self)
1210
# remove these delegates a while after bzr 0.15
1211
def __make_delegated(name, from_module):
1212
def _deprecated_repository_forwarder():
1213
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1214
% (name, from_module),
1217
m = __import__(from_module, globals(), locals(), [name])
1219
return getattr(m, name)
1220
except AttributeError:
1221
raise AttributeError('module %s has no name %s'
1223
globals()[name] = _deprecated_repository_forwarder
1226
'AllInOneRepository',
1227
'WeaveMetaDirRepository',
1228
'PreSplitOutRepositoryFormat',
1229
'RepositoryFormat4',
1230
'RepositoryFormat5',
1231
'RepositoryFormat6',
1232
'RepositoryFormat7',
1234
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1238
'RepositoryFormatKnit',
1239
'RepositoryFormatKnit1',
1241
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1244
def install_revision(repository, rev, revision_tree):
1245
"""Install all revision data into a repository."""
1246
present_parents = []
1248
for p_id in rev.parent_ids:
1249
if repository.has_revision(p_id):
1250
present_parents.append(p_id)
1251
parent_trees[p_id] = repository.revision_tree(p_id)
1253
parent_trees[p_id] = repository.revision_tree(None)
1255
inv = revision_tree.inventory
1256
entries = inv.iter_entries()
1257
# backwards compatibility hack: skip the root id.
1258
if not repository.supports_rich_root():
1259
path, root = entries.next()
1260
if root.revision != rev.revision_id:
1261
raise errors.IncompatibleRevision(repr(repository))
1262
# Add the texts that are not already present
1263
for path, ie in entries:
1264
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1265
repository.get_transaction())
1266
if ie.revision not in w:
1268
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1269
# with InventoryEntry.find_previous_heads(). if it is, then there
1270
# is a latent bug here where the parents may have ancestors of each
1272
for revision, tree in parent_trees.iteritems():
1273
if ie.file_id not in tree:
1275
parent_id = tree.inventory[ie.file_id].revision
1276
if parent_id in text_parents:
1278
text_parents.append(parent_id)
1280
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1281
repository.get_transaction())
1282
lines = revision_tree.get_file(ie.file_id).readlines()
1283
vfile.add_lines(rev.revision_id, text_parents, lines)
1285
# install the inventory
1286
repository.add_inventory(rev.revision_id, inv, present_parents)
1287
except errors.RevisionAlreadyPresent:
1289
repository.add_revision(rev.revision_id, rev, inv)
1292
class MetaDirRepository(Repository):
1293
"""Repositories in the new meta-dir layout."""
1295
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1296
super(MetaDirRepository, self).__init__(_format,
1302
dir_mode = self.control_files._dir_mode
1303
file_mode = self.control_files._file_mode
1306
def is_shared(self):
1307
"""Return True if this repository is flagged as a shared repository."""
1308
return self.control_files._transport.has('shared-storage')
1311
def set_make_working_trees(self, new_value):
1312
"""Set the policy flag for making working trees when creating branches.
1314
This only applies to branches that use this repository.
1316
The default is 'True'.
1317
:param new_value: True to restore the default, False to disable making
1322
self.control_files._transport.delete('no-working-trees')
1323
except errors.NoSuchFile:
1326
self.control_files.put_utf8('no-working-trees', '')
1328
def make_working_trees(self):
1329
"""Returns the policy for making working trees on new branches."""
1330
return not self.control_files._transport.has('no-working-trees')
1333
class RepositoryFormatRegistry(registry.Registry):
1334
"""Registry of RepositoryFormats.
1337
def get(self, format_string):
1338
r = registry.Registry.get(self, format_string)
1344
format_registry = RepositoryFormatRegistry()
1345
"""Registry of formats, indexed by their identifying format string.
1347
This can contain either format instances themselves, or classes/factories that
1348
can be called to obtain one.
1352
#####################################################################
1353
# Repository Formats
1355
class RepositoryFormat(object):
1356
"""A repository format.
1358
Formats provide three things:
1359
* An initialization routine to construct repository data on disk.
1360
* a format string which is used when the BzrDir supports versioned
1362
* an open routine which returns a Repository instance.
1364
Formats are placed in an dict by their format string for reference
1365
during opening. These should be subclasses of RepositoryFormat
1368
Once a format is deprecated, just deprecate the initialize and open
1369
methods on the format class. Do not deprecate the object, as the
1370
object will be created every system load.
1372
Common instance attributes:
1373
_matchingbzrdir - the bzrdir format that the repository format was
1374
originally written to work with. This can be used if manually
1375
constructing a bzrdir and repository, or more commonly for test suite
1380
return "<%s>" % self.__class__.__name__
1382
def __eq__(self, other):
1383
# format objects are generally stateless
1384
return isinstance(other, self.__class__)
1386
def __ne__(self, other):
1387
return not self == other
1390
def find_format(klass, a_bzrdir):
1391
"""Return the format for the repository object in a_bzrdir.
1393
This is used by bzr native formats that have a "format" file in
1394
the repository. Other methods may be used by different types of
1398
transport = a_bzrdir.get_repository_transport(None)
1399
format_string = transport.get("format").read()
1400
return format_registry.get(format_string)
1401
except errors.NoSuchFile:
1402
raise errors.NoRepositoryPresent(a_bzrdir)
1404
raise errors.UnknownFormatError(format=format_string)
1407
def register_format(klass, format):
1408
format_registry.register(format.get_format_string(), format)
1411
def unregister_format(klass, format):
1412
format_registry.remove(format.get_format_string())
1415
def get_default_format(klass):
1416
"""Return the current default format."""
1417
from bzrlib import bzrdir
1418
return bzrdir.format_registry.make_bzrdir('default').repository_format
1420
def _get_control_store(self, repo_transport, control_files):
1421
"""Return the control store for this repository."""
1422
raise NotImplementedError(self._get_control_store)
1424
def get_format_string(self):
1425
"""Return the ASCII format string that identifies this format.
1427
Note that in pre format ?? repositories the format string is
1428
not permitted nor written to disk.
1430
raise NotImplementedError(self.get_format_string)
1432
def get_format_description(self):
1433
"""Return the short description for this format."""
1434
raise NotImplementedError(self.get_format_description)
1436
def _get_revision_store(self, repo_transport, control_files):
1437
"""Return the revision store object for this a_bzrdir."""
1438
raise NotImplementedError(self._get_revision_store)
1440
def _get_text_rev_store(self,
1447
"""Common logic for getting a revision store for a repository.
1449
see self._get_revision_store for the subclass-overridable method to
1450
get the store for a repository.
1452
from bzrlib.store.revision.text import TextRevisionStore
1453
dir_mode = control_files._dir_mode
1454
file_mode = control_files._file_mode
1455
text_store = TextStore(transport.clone(name),
1457
compressed=compressed,
1459
file_mode=file_mode)
1460
_revision_store = TextRevisionStore(text_store, serializer)
1461
return _revision_store
1463
# TODO: this shouldn't be in the base class, it's specific to things that
1464
# use weaves or knits -- mbp 20070207
1465
def _get_versioned_file_store(self,
1470
versionedfile_class=None,
1471
versionedfile_kwargs={},
1473
if versionedfile_class is None:
1474
versionedfile_class = self._versionedfile_class
1475
weave_transport = control_files._transport.clone(name)
1476
dir_mode = control_files._dir_mode
1477
file_mode = control_files._file_mode
1478
return VersionedFileStore(weave_transport, prefixed=prefixed,
1480
file_mode=file_mode,
1481
versionedfile_class=versionedfile_class,
1482
versionedfile_kwargs=versionedfile_kwargs,
1485
def initialize(self, a_bzrdir, shared=False):
1486
"""Initialize a repository of this format in a_bzrdir.
1488
:param a_bzrdir: The bzrdir to put the new repository in it.
1489
:param shared: The repository should be initialized as a sharable one.
1490
:returns: The new repository object.
1492
This may raise UninitializableFormat if shared repository are not
1493
compatible the a_bzrdir.
1495
raise NotImplementedError(self.initialize)
1497
def is_supported(self):
1498
"""Is this format supported?
1500
Supported formats must be initializable and openable.
1501
Unsupported formats may not support initialization or committing or
1502
some other features depending on the reason for not being supported.
1506
def check_conversion_target(self, target_format):
1507
raise NotImplementedError(self.check_conversion_target)
1509
def open(self, a_bzrdir, _found=False):
1510
"""Return an instance of this format for the bzrdir a_bzrdir.
1512
_found is a private parameter, do not use it.
1514
raise NotImplementedError(self.open)
1517
class MetaDirRepositoryFormat(RepositoryFormat):
1518
"""Common base class for the new repositories using the metadir layout."""
1520
rich_root_data = False
1521
supports_tree_reference = False
1522
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1525
super(MetaDirRepositoryFormat, self).__init__()
1527
def _create_control_files(self, a_bzrdir):
1528
"""Create the required files and the initial control_files object."""
1529
# FIXME: RBC 20060125 don't peek under the covers
1530
# NB: no need to escape relative paths that are url safe.
1531
repository_transport = a_bzrdir.get_repository_transport(self)
1532
control_files = lockable_files.LockableFiles(repository_transport,
1533
'lock', lockdir.LockDir)
1534
control_files.create_lock()
1535
return control_files
1537
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1538
"""Upload the initial blank content."""
1539
control_files = self._create_control_files(a_bzrdir)
1540
control_files.lock_write()
1542
control_files._transport.mkdir_multi(dirs,
1543
mode=control_files._dir_mode)
1544
for file, content in files:
1545
control_files.put(file, content)
1546
for file, content in utf8_files:
1547
control_files.put_utf8(file, content)
1549
control_files.put_utf8('shared-storage', '')
1551
control_files.unlock()
1554
# formats which have no format string are not discoverable
1555
# and not independently creatable, so are not registered. They're
1556
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1557
# needed, it's constructed directly by the BzrDir. Non-native formats where
1558
# the repository is not separately opened are similar.
1560
format_registry.register_lazy(
1561
'Bazaar-NG Repository format 7',
1562
'bzrlib.repofmt.weaverepo',
1565
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1566
# default control directory format
1568
format_registry.register_lazy(
1569
'Bazaar-NG Knit Repository Format 1',
1570
'bzrlib.repofmt.knitrepo',
1571
'RepositoryFormatKnit1',
1573
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1575
format_registry.register_lazy(
1576
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1577
'bzrlib.repofmt.knitrepo',
1578
'RepositoryFormatKnit3',
1582
class InterRepository(InterObject):
1583
"""This class represents operations taking place between two repositories.
1585
Its instances have methods like copy_content and fetch, and contain
1586
references to the source and target repositories these operations can be
1589
Often we will provide convenience methods on 'repository' which carry out
1590
operations with another repository - they will always forward to
1591
InterRepository.get(other).method_name(parameters).
1595
"""The available optimised InterRepository types."""
1597
def copy_content(self, revision_id=None):
1598
raise NotImplementedError(self.copy_content)
1600
def fetch(self, revision_id=None, pb=None):
1601
"""Fetch the content required to construct revision_id.
1603
The content is copied from self.source to self.target.
1605
:param revision_id: if None all content is copied, if NULL_REVISION no
1607
:param pb: optional progress bar to use for progress reports. If not
1608
provided a default one will be created.
1610
Returns the copied revision count and the failed revisions in a tuple:
1613
raise NotImplementedError(self.fetch)
1616
def missing_revision_ids(self, revision_id=None):
1617
"""Return the revision ids that source has that target does not.
1619
These are returned in topological order.
1621
:param revision_id: only return revision ids included by this
1624
# generic, possibly worst case, slow code path.
1625
target_ids = set(self.target.all_revision_ids())
1626
if revision_id is not None:
1627
# TODO: jam 20070210 InterRepository is internal enough that it
1628
# should assume revision_ids are already utf-8
1629
revision_id = osutils.safe_revision_id(revision_id)
1630
source_ids = self.source.get_ancestry(revision_id)
1631
assert source_ids[0] is None
1634
source_ids = self.source.all_revision_ids()
1635
result_set = set(source_ids).difference(target_ids)
1636
# this may look like a no-op: its not. It preserves the ordering
1637
# other_ids had while only returning the members from other_ids
1638
# that we've decided we need.
1639
return [rev_id for rev_id in source_ids if rev_id in result_set]
1642
class InterSameDataRepository(InterRepository):
1643
"""Code for converting between repositories that represent the same data.
1645
Data format and model must match for this to work.
1649
def _get_repo_format_to_test(self):
1650
"""Repository format for testing with.
1652
InterSameData can pull from subtree to subtree and from non-subtree to
1653
non-subtree, so we test this with the richest repository format.
1655
from bzrlib.repofmt import knitrepo
1656
return knitrepo.RepositoryFormatKnit3()
1659
def is_compatible(source, target):
1660
if source.supports_rich_root() != target.supports_rich_root():
1662
if source._serializer != target._serializer:
1667
def copy_content(self, revision_id=None):
1668
"""Make a complete copy of the content in self into destination.
1670
This copies both the repository's revision data, and configuration information
1671
such as the make_working_trees setting.
1673
This is a destructive operation! Do not use it on existing
1676
:param revision_id: Only copy the content needed to construct
1677
revision_id and its parents.
1680
self.target.set_make_working_trees(self.source.make_working_trees())
1681
except NotImplementedError:
1683
# TODO: jam 20070210 This is fairly internal, so we should probably
1684
# just assert that revision_id is not unicode.
1685
revision_id = osutils.safe_revision_id(revision_id)
1686
# but don't bother fetching if we have the needed data now.
1687
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1688
self.target.has_revision(revision_id)):
1690
self.target.fetch(self.source, revision_id=revision_id)
1693
def fetch(self, revision_id=None, pb=None):
1694
"""See InterRepository.fetch()."""
1695
from bzrlib.fetch import GenericRepoFetcher
1696
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1697
self.source, self.source._format, self.target,
1698
self.target._format)
1699
# TODO: jam 20070210 This should be an assert, not a translate
1700
revision_id = osutils.safe_revision_id(revision_id)
1701
f = GenericRepoFetcher(to_repository=self.target,
1702
from_repository=self.source,
1703
last_revision=revision_id,
1705
return f.count_copied, f.failed_revisions
1708
class InterWeaveRepo(InterSameDataRepository):
1709
"""Optimised code paths between Weave based repositories."""
1712
def _get_repo_format_to_test(self):
1713
from bzrlib.repofmt import weaverepo
1714
return weaverepo.RepositoryFormat7()
1717
def is_compatible(source, target):
1718
"""Be compatible with known Weave formats.
1720
We don't test for the stores being of specific types because that
1721
could lead to confusing results, and there is no need to be
1724
from bzrlib.repofmt.weaverepo import (
1730
return (isinstance(source._format, (RepositoryFormat5,
1732
RepositoryFormat7)) and
1733
isinstance(target._format, (RepositoryFormat5,
1735
RepositoryFormat7)))
1736
except AttributeError:
1740
def copy_content(self, revision_id=None):
1741
"""See InterRepository.copy_content()."""
1742
# weave specific optimised path:
1743
# TODO: jam 20070210 Internal, should be an assert, not translate
1744
revision_id = osutils.safe_revision_id(revision_id)
1746
self.target.set_make_working_trees(self.source.make_working_trees())
1747
except NotImplementedError:
1749
# FIXME do not peek!
1750
if self.source.control_files._transport.listable():
1751
pb = ui.ui_factory.nested_progress_bar()
1753
self.target.weave_store.copy_all_ids(
1754
self.source.weave_store,
1756
from_transaction=self.source.get_transaction(),
1757
to_transaction=self.target.get_transaction())
1758
pb.update('copying inventory', 0, 1)
1759
self.target.control_weaves.copy_multi(
1760
self.source.control_weaves, ['inventory'],
1761
from_transaction=self.source.get_transaction(),
1762
to_transaction=self.target.get_transaction())
1763
self.target._revision_store.text_store.copy_all_ids(
1764
self.source._revision_store.text_store,
1769
self.target.fetch(self.source, revision_id=revision_id)
1772
def fetch(self, revision_id=None, pb=None):
1773
"""See InterRepository.fetch()."""
1774
from bzrlib.fetch import GenericRepoFetcher
1775
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1776
self.source, self.source._format, self.target, self.target._format)
1777
# TODO: jam 20070210 This should be an assert, not a translate
1778
revision_id = osutils.safe_revision_id(revision_id)
1779
f = GenericRepoFetcher(to_repository=self.target,
1780
from_repository=self.source,
1781
last_revision=revision_id,
1783
return f.count_copied, f.failed_revisions
1786
def missing_revision_ids(self, revision_id=None):
1787
"""See InterRepository.missing_revision_ids()."""
1788
# we want all revisions to satisfy revision_id in source.
1789
# but we don't want to stat every file here and there.
1790
# we want then, all revisions other needs to satisfy revision_id
1791
# checked, but not those that we have locally.
1792
# so the first thing is to get a subset of the revisions to
1793
# satisfy revision_id in source, and then eliminate those that
1794
# we do already have.
1795
# this is slow on high latency connection to self, but as as this
1796
# disk format scales terribly for push anyway due to rewriting
1797
# inventory.weave, this is considered acceptable.
1799
if revision_id is not None:
1800
source_ids = self.source.get_ancestry(revision_id)
1801
assert source_ids[0] is None
1804
source_ids = self.source._all_possible_ids()
1805
source_ids_set = set(source_ids)
1806
# source_ids is the worst possible case we may need to pull.
1807
# now we want to filter source_ids against what we actually
1808
# have in target, but don't try to check for existence where we know
1809
# we do not have a revision as that would be pointless.
1810
target_ids = set(self.target._all_possible_ids())
1811
possibly_present_revisions = target_ids.intersection(source_ids_set)
1812
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1813
required_revisions = source_ids_set.difference(actually_present_revisions)
1814
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1815
if revision_id is not None:
1816
# we used get_ancestry to determine source_ids then we are assured all
1817
# revisions referenced are present as they are installed in topological order.
1818
# and the tip revision was validated by get_ancestry.
1819
return required_topo_revisions
1821
# if we just grabbed the possibly available ids, then
1822
# we only have an estimate of whats available and need to validate
1823
# that against the revision records.
1824
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1827
class InterKnitRepo(InterSameDataRepository):
1828
"""Optimised code paths between Knit based repositories."""
1831
def _get_repo_format_to_test(self):
1832
from bzrlib.repofmt import knitrepo
1833
return knitrepo.RepositoryFormatKnit1()
1836
def is_compatible(source, target):
1837
"""Be compatible with known Knit formats.
1839
We don't test for the stores being of specific types because that
1840
could lead to confusing results, and there is no need to be
1843
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1845
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1846
isinstance(target._format, (RepositoryFormatKnit1)))
1847
except AttributeError:
1851
def fetch(self, revision_id=None, pb=None):
1852
"""See InterRepository.fetch()."""
1853
from bzrlib.fetch import KnitRepoFetcher
1854
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1855
self.source, self.source._format, self.target, self.target._format)
1856
# TODO: jam 20070210 This should be an assert, not a translate
1857
revision_id = osutils.safe_revision_id(revision_id)
1858
f = KnitRepoFetcher(to_repository=self.target,
1859
from_repository=self.source,
1860
last_revision=revision_id,
1862
return f.count_copied, f.failed_revisions
1865
def missing_revision_ids(self, revision_id=None):
1866
"""See InterRepository.missing_revision_ids()."""
1867
if revision_id is not None:
1868
source_ids = self.source.get_ancestry(revision_id)
1869
assert source_ids[0] is None
1872
source_ids = self.source._all_possible_ids()
1873
source_ids_set = set(source_ids)
1874
# source_ids is the worst possible case we may need to pull.
1875
# now we want to filter source_ids against what we actually
1876
# have in target, but don't try to check for existence where we know
1877
# we do not have a revision as that would be pointless.
1878
target_ids = set(self.target._all_possible_ids())
1879
possibly_present_revisions = target_ids.intersection(source_ids_set)
1880
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1881
required_revisions = source_ids_set.difference(actually_present_revisions)
1882
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1883
if revision_id is not None:
1884
# we used get_ancestry to determine source_ids then we are assured all
1885
# revisions referenced are present as they are installed in topological order.
1886
# and the tip revision was validated by get_ancestry.
1887
return required_topo_revisions
1889
# if we just grabbed the possibly available ids, then
1890
# we only have an estimate of whats available and need to validate
1891
# that against the revision records.
1892
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1895
class InterModel1and2(InterRepository):
1898
def _get_repo_format_to_test(self):
1902
def is_compatible(source, target):
1903
if not source.supports_rich_root() and target.supports_rich_root():
1909
def fetch(self, revision_id=None, pb=None):
1910
"""See InterRepository.fetch()."""
1911
from bzrlib.fetch import Model1toKnit2Fetcher
1912
# TODO: jam 20070210 This should be an assert, not a translate
1913
revision_id = osutils.safe_revision_id(revision_id)
1914
f = Model1toKnit2Fetcher(to_repository=self.target,
1915
from_repository=self.source,
1916
last_revision=revision_id,
1918
return f.count_copied, f.failed_revisions
1921
def copy_content(self, revision_id=None):
1922
"""Make a complete copy of the content in self into destination.
1924
This is a destructive operation! Do not use it on existing
1927
:param revision_id: Only copy the content needed to construct
1928
revision_id and its parents.
1931
self.target.set_make_working_trees(self.source.make_working_trees())
1932
except NotImplementedError:
1934
# TODO: jam 20070210 Internal, assert, don't translate
1935
revision_id = osutils.safe_revision_id(revision_id)
1936
# but don't bother fetching if we have the needed data now.
1937
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1938
self.target.has_revision(revision_id)):
1940
self.target.fetch(self.source, revision_id=revision_id)
1943
class InterKnit1and2(InterKnitRepo):
1946
def _get_repo_format_to_test(self):
1950
def is_compatible(source, target):
1951
"""Be compatible with Knit1 source and Knit3 target"""
1952
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
1954
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1955
RepositoryFormatKnit3
1956
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1957
isinstance(target._format, (RepositoryFormatKnit3)))
1958
except AttributeError:
1962
def fetch(self, revision_id=None, pb=None):
1963
"""See InterRepository.fetch()."""
1964
from bzrlib.fetch import Knit1to2Fetcher
1965
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1966
self.source, self.source._format, self.target,
1967
self.target._format)
1968
# TODO: jam 20070210 This should be an assert, not a translate
1969
revision_id = osutils.safe_revision_id(revision_id)
1970
f = Knit1to2Fetcher(to_repository=self.target,
1971
from_repository=self.source,
1972
last_revision=revision_id,
1974
return f.count_copied, f.failed_revisions
1977
class InterRemoteRepository(InterRepository):
1978
"""Code for converting between RemoteRepository objects.
1980
This just gets an non-remote repository from the RemoteRepository, and calls
1981
InterRepository.get again.
1984
def __init__(self, source, target):
1985
if isinstance(source, remote.RemoteRepository):
1986
source._ensure_real()
1987
real_source = source._real_repository
1989
real_source = source
1990
if isinstance(target, remote.RemoteRepository):
1991
target._ensure_real()
1992
real_target = target._real_repository
1994
real_target = target
1995
self.real_inter = InterRepository.get(real_source, real_target)
1998
def is_compatible(source, target):
1999
if isinstance(source, remote.RemoteRepository):
2001
if isinstance(target, remote.RemoteRepository):
2005
def copy_content(self, revision_id=None):
2006
self.real_inter.copy_content(revision_id=revision_id)
2008
def fetch(self, revision_id=None, pb=None):
2009
self.real_inter.fetch(revision_id=revision_id, pb=pb)
2012
def _get_repo_format_to_test(self):
2016
InterRepository.register_optimiser(InterSameDataRepository)
2017
InterRepository.register_optimiser(InterWeaveRepo)
2018
InterRepository.register_optimiser(InterKnitRepo)
2019
InterRepository.register_optimiser(InterModel1and2)
2020
InterRepository.register_optimiser(InterKnit1and2)
2021
InterRepository.register_optimiser(InterRemoteRepository)
2024
class CopyConverter(object):
2025
"""A repository conversion tool which just performs a copy of the content.
2027
This is slow but quite reliable.
2030
def __init__(self, target_format):
2031
"""Create a CopyConverter.
2033
:param target_format: The format the resulting repository should be.
2035
self.target_format = target_format
2037
def convert(self, repo, pb):
2038
"""Perform the conversion of to_convert, giving feedback via pb.
2040
:param to_convert: The disk object to convert.
2041
:param pb: a progress bar to use for progress information.
2046
# this is only useful with metadir layouts - separated repo content.
2047
# trigger an assertion if not such
2048
repo._format.get_format_string()
2049
self.repo_dir = repo.bzrdir
2050
self.step('Moving repository to repository.backup')
2051
self.repo_dir.transport.move('repository', 'repository.backup')
2052
backup_transport = self.repo_dir.transport.clone('repository.backup')
2053
repo._format.check_conversion_target(self.target_format)
2054
self.source_repo = repo._format.open(self.repo_dir,
2056
_override_transport=backup_transport)
2057
self.step('Creating new repository')
2058
converted = self.target_format.initialize(self.repo_dir,
2059
self.source_repo.is_shared())
2060
converted.lock_write()
2062
self.step('Copying content into repository.')
2063
self.source_repo.copy_content_into(converted)
2066
self.step('Deleting old repository content.')
2067
self.repo_dir.transport.delete_tree('repository.backup')
2068
self.pb.note('repository converted')
2070
def step(self, message):
2071
"""Update the pb by a step."""
2073
self.pb.update(message, self.count, self.total)
2076
class CommitBuilder(object):
2077
"""Provides an interface to build up a commit.
2079
This allows describing a tree to be committed without needing to
2080
know the internals of the format of the repository.
2083
# all clients should supply tree roots.
2084
record_root_entry = True
2086
def __init__(self, repository, parents, config, timestamp=None,
2087
timezone=None, committer=None, revprops=None,
2089
"""Initiate a CommitBuilder.
2091
:param repository: Repository to commit to.
2092
:param parents: Revision ids of the parents of the new revision.
2093
:param config: Configuration to use.
2094
:param timestamp: Optional timestamp recorded for commit.
2095
:param timezone: Optional timezone for timestamp.
2096
:param committer: Optional committer to set for commit.
2097
:param revprops: Optional dictionary of revision properties.
2098
:param revision_id: Optional revision id.
2100
self._config = config
2102
if committer is None:
2103
self._committer = self._config.username()
2105
assert isinstance(committer, basestring), type(committer)
2106
self._committer = committer
2108
self.new_inventory = Inventory(None)
2109
self._new_revision_id = osutils.safe_revision_id(revision_id)
2110
self.parents = parents
2111
self.repository = repository
2114
if revprops is not None:
2115
self._revprops.update(revprops)
2117
if timestamp is None:
2118
timestamp = time.time()
2119
# Restrict resolution to 1ms
2120
self._timestamp = round(timestamp, 3)
2122
if timezone is None:
2123
self._timezone = osutils.local_time_offset()
2125
self._timezone = int(timezone)
2127
self._generate_revision_if_needed()
2129
def commit(self, message):
2130
"""Make the actual commit.
2132
:return: The revision id of the recorded revision.
2134
rev = _mod_revision.Revision(
2135
timestamp=self._timestamp,
2136
timezone=self._timezone,
2137
committer=self._committer,
2139
inventory_sha1=self.inv_sha1,
2140
revision_id=self._new_revision_id,
2141
properties=self._revprops)
2142
rev.parent_ids = self.parents
2143
self.repository.add_revision(self._new_revision_id, rev,
2144
self.new_inventory, self._config)
2145
self.repository.commit_write_group()
2146
return self._new_revision_id
2149
"""Abort the commit that is being built.
2151
self.repository.abort_write_group()
2153
def revision_tree(self):
2154
"""Return the tree that was just committed.
2156
After calling commit() this can be called to get a RevisionTree
2157
representing the newly committed tree. This is preferred to
2158
calling Repository.revision_tree() because that may require
2159
deserializing the inventory, while we already have a copy in
2162
return RevisionTree(self.repository, self.new_inventory,
2163
self._new_revision_id)
2165
def finish_inventory(self):
2166
"""Tell the builder that the inventory is finished."""
2167
if self.new_inventory.root is None:
2168
symbol_versioning.warn('Root entry should be supplied to'
2169
' record_entry_contents, as of bzr 0.10.',
2170
DeprecationWarning, stacklevel=2)
2171
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
2172
self.new_inventory.revision_id = self._new_revision_id
2173
self.inv_sha1 = self.repository.add_inventory(
2174
self._new_revision_id,
2179
def _gen_revision_id(self):
2180
"""Return new revision-id."""
2181
return generate_ids.gen_revision_id(self._config.username(),
2184
def _generate_revision_if_needed(self):
2185
"""Create a revision id if None was supplied.
2187
If the repository can not support user-specified revision ids
2188
they should override this function and raise CannotSetRevisionId
2189
if _new_revision_id is not None.
2191
:raises: CannotSetRevisionId
2193
if self._new_revision_id is None:
2194
self._new_revision_id = self._gen_revision_id()
2195
self.random_revid = True
2197
self.random_revid = False
2199
def _check_root(self, ie, parent_invs, tree):
2200
"""Helper for record_entry_contents.
2202
:param ie: An entry being added.
2203
:param parent_invs: The inventories of the parent revisions of the
2205
:param tree: The tree that is being committed.
2207
if ie.parent_id is not None:
2208
# if ie is not root, add a root automatically.
2209
symbol_versioning.warn('Root entry should be supplied to'
2210
' record_entry_contents, as of bzr 0.10.',
2211
DeprecationWarning, stacklevel=2)
2212
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
2215
# In this revision format, root entries have no knit or weave When
2216
# serializing out to disk and back in root.revision is always
2218
ie.revision = self._new_revision_id
2220
def record_entry_contents(self, ie, parent_invs, path, tree):
2221
"""Record the content of ie from tree into the commit if needed.
2223
Side effect: sets ie.revision when unchanged
2225
:param ie: An inventory entry present in the commit.
2226
:param parent_invs: The inventories of the parent revisions of the
2228
:param path: The path the entry is at in the tree.
2229
:param tree: The tree which contains this entry and should be used to
2232
if self.new_inventory.root is None:
2233
self._check_root(ie, parent_invs, tree)
2234
self.new_inventory.add(ie)
2236
# ie.revision is always None if the InventoryEntry is considered
2237
# for committing. ie.snapshot will record the correct revision
2238
# which may be the sole parent if it is untouched.
2239
if ie.revision is not None:
2242
parent_candiate_entries = ie.parent_candidates(parent_invs)
2243
heads = self.repository.get_graph().heads(parent_candiate_entries.keys())
2244
# XXX: Note that this is unordered - and this is tolerable because
2245
# the previous code was also unordered.
2246
previous_entries = dict((head, parent_candiate_entries[head]) for head
2248
# we are creating a new revision for ie in the history store and
2250
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2252
def modified_directory(self, file_id, file_parents):
2253
"""Record the presence of a symbolic link.
2255
:param file_id: The file_id of the link to record.
2256
:param file_parents: The per-file parent revision ids.
2258
self._add_text_to_weave(file_id, [], file_parents.keys())
2260
def modified_reference(self, file_id, file_parents):
2261
"""Record the modification of a reference.
2263
:param file_id: The file_id of the link to record.
2264
:param file_parents: The per-file parent revision ids.
2266
self._add_text_to_weave(file_id, [], file_parents.keys())
2268
def modified_file_text(self, file_id, file_parents,
2269
get_content_byte_lines, text_sha1=None,
2271
"""Record the text of file file_id
2273
:param file_id: The file_id of the file to record the text of.
2274
:param file_parents: The per-file parent revision ids.
2275
:param get_content_byte_lines: A callable which will return the byte
2277
:param text_sha1: Optional SHA1 of the file contents.
2278
:param text_size: Optional size of the file contents.
2280
# mutter('storing text of file {%s} in revision {%s} into %r',
2281
# file_id, self._new_revision_id, self.repository.weave_store)
2282
# special case to avoid diffing on renames or
2284
if (len(file_parents) == 1
2285
and text_sha1 == file_parents.values()[0].text_sha1
2286
and text_size == file_parents.values()[0].text_size):
2287
previous_ie = file_parents.values()[0]
2288
versionedfile = self.repository.weave_store.get_weave(file_id,
2289
self.repository.get_transaction())
2290
versionedfile.clone_text(self._new_revision_id,
2291
previous_ie.revision, file_parents.keys())
2292
return text_sha1, text_size
2294
new_lines = get_content_byte_lines()
2295
return self._add_text_to_weave(file_id, new_lines,
2296
file_parents.keys())
2298
def modified_link(self, file_id, file_parents, link_target):
2299
"""Record the presence of a symbolic link.
2301
:param file_id: The file_id of the link to record.
2302
:param file_parents: The per-file parent revision ids.
2303
:param link_target: Target location of this link.
2305
self._add_text_to_weave(file_id, [], file_parents.keys())
2307
def _add_text_to_weave(self, file_id, new_lines, parents):
2308
versionedfile = self.repository.weave_store.get_weave_or_empty(
2309
file_id, self.repository.get_transaction())
2310
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
2311
# than add_lines, and allows committing when a parent is ghosted for
2313
# Note: as we read the content directly from the tree, we know its not
2314
# been turned into unicode or badly split - but a broken tree
2315
# implementation could give us bad output from readlines() so this is
2316
# not a guarantee of safety. What would be better is always checking
2317
# the content during test suite execution. RBC 20070912
2318
result = versionedfile.add_lines_with_ghosts(
2319
self._new_revision_id, parents, new_lines,
2320
random_id=self.random_revid, check_content=False)[0:2]
2321
versionedfile.clear_cache()
2325
class RootCommitBuilder(CommitBuilder):
2326
"""This commitbuilder actually records the root id"""
2328
def _check_root(self, ie, parent_invs, tree):
2329
"""Helper for record_entry_contents.
2331
:param ie: An entry being added.
2332
:param parent_invs: The inventories of the parent revisions of the
2334
:param tree: The tree that is being committed.
2336
# ie must be root for this builder
2337
assert ie.parent_id is None
2349
def _unescaper(match, _map=_unescape_map):
2350
code = match.group(1)
2354
if not code.startswith('#'):
2356
return unichr(int(code[1:])).encode('utf8')
2362
def _unescape_xml(data):
2363
"""Unescape predefined XML entities in a string of data."""
2365
if _unescape_re is None:
2366
_unescape_re = re.compile('\&([^;]*);')
2367
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
