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
49
from bzrlib.util import bencode
52
from bzrlib.decorators import needs_read_lock, needs_write_lock
53
from bzrlib.inter import InterObject
54
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
55
from bzrlib.symbol_versioning import (
58
from bzrlib.trace import mutter, mutter_callsite, note, warning
61
# Old formats display a warning, but only once
62
_deprecation_warning_done = False
65
######################################################################
68
class Repository(object):
69
"""Repository holding history for one or more branches.
71
The repository holds and retrieves historical information including
72
revisions and file history. It's normally accessed only by the Branch,
73
which views a particular line of development through that history.
75
The Repository builds on top of Stores and a Transport, which respectively
76
describe the disk data format and the way of accessing the (possibly
80
_file_ids_altered_regex = lazy_regex.lazy_compile(
81
r'file_id="(?P<file_id>[^"]+)"'
82
r'.*revision="(?P<revision_id>[^"]+)"'
85
def abort_write_group(self):
86
"""Commit the contents accrued within the current write group.
88
:seealso: start_write_group.
90
if self._write_group is not self.get_transaction():
91
# has an unlock or relock occured ?
92
raise errors.BzrError('mismatched lock context and write group.')
93
self._abort_write_group()
94
self._write_group = None
96
def _abort_write_group(self):
97
"""Template method for per-repository write group cleanup.
99
This is called during abort before the write group is considered to be
100
finished and should cleanup any internal state accrued during the write
101
group. There is no requirement that data handed to the repository be
102
*not* made available - this is not a rollback - but neither should any
103
attempt be made to ensure that data added is fully commited. Abort is
104
invoked when an error has occured so futher disk or network operations
105
may not be possible or may error and if possible should not be
110
def add_inventory(self, revision_id, inv, parents):
111
"""Add the inventory inv to the repository as revision_id.
113
:param parents: The revision ids of the parents that revision_id
114
is known to have and are in the repository already.
116
returns the sha1 of the serialized inventory.
118
revision_id = osutils.safe_revision_id(revision_id)
119
_mod_revision.check_not_reserved_id(revision_id)
120
assert inv.revision_id is None or inv.revision_id == revision_id, \
121
"Mismatch between inventory revision" \
122
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
123
assert inv.root is not None
124
inv_text = self.serialise_inventory(inv)
125
inv_sha1 = osutils.sha_string(inv_text)
126
inv_vf = self.control_weaves.get_weave('inventory',
127
self.get_transaction())
128
self._inventory_add_lines(inv_vf, revision_id, parents,
129
osutils.split_lines(inv_text))
132
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines):
134
for parent in parents:
136
final_parents.append(parent)
138
inv_vf.add_lines(revision_id, final_parents, lines)
141
def add_revision(self, revision_id, rev, inv=None, config=None):
142
"""Add rev to the revision store as revision_id.
144
:param revision_id: the revision id to use.
145
:param rev: The revision object.
146
:param inv: The inventory for the revision. if None, it will be looked
147
up in the inventory storer
148
:param config: If None no digital signature will be created.
149
If supplied its signature_needed method will be used
150
to determine if a signature should be made.
152
revision_id = osutils.safe_revision_id(revision_id)
153
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
155
_mod_revision.check_not_reserved_id(revision_id)
156
if config is not None and config.signature_needed():
158
inv = self.get_inventory(revision_id)
159
plaintext = Testament(rev, inv).as_short_text()
160
self.store_revision_signature(
161
gpg.GPGStrategy(config), plaintext, revision_id)
162
if not revision_id in self.get_inventory_weave():
164
raise errors.WeaveRevisionNotPresent(revision_id,
165
self.get_inventory_weave())
167
# yes, this is not suitable for adding with ghosts.
168
self.add_inventory(revision_id, inv, rev.parent_ids)
169
self._revision_store.add_revision(rev, self.get_transaction())
171
def _add_revision_text(self, revision_id, text):
172
revision = self._revision_store._serializer.read_revision_from_string(
174
self._revision_store._add_revision(revision, StringIO(text),
175
self.get_transaction())
178
def _all_possible_ids(self):
179
"""Return all the possible revisions that we could find."""
180
return self.get_inventory_weave().versions()
182
def all_revision_ids(self):
183
"""Returns a list of all the revision ids in the repository.
185
This is deprecated because code should generally work on the graph
186
reachable from a particular revision, and ignore any other revisions
187
that might be present. There is no direct replacement method.
189
return self._all_revision_ids()
192
def _all_revision_ids(self):
193
"""Returns a list of all the revision ids in the repository.
195
These are in as much topological order as the underlying store can
196
present: for weaves ghosts may lead to a lack of correctness until
197
the reweave updates the parents list.
199
if self._revision_store.text_store.listable():
200
return self._revision_store.all_revision_ids(self.get_transaction())
201
result = self._all_possible_ids()
202
# TODO: jam 20070210 Ensure that _all_possible_ids returns non-unicode
203
# ids. (It should, since _revision_store's API should change to
204
# return utf8 revision_ids)
205
return self._eliminate_revisions_not_present(result)
207
def break_lock(self):
208
"""Break a lock if one is present from another instance.
210
Uses the ui factory to ask for confirmation if the lock may be from
213
self.control_files.break_lock()
216
def _eliminate_revisions_not_present(self, revision_ids):
217
"""Check every revision id in revision_ids to see if we have it.
219
Returns a set of the present revisions.
222
for id in revision_ids:
223
if self.has_revision(id):
228
def create(a_bzrdir):
229
"""Construct the current default format repository in a_bzrdir."""
230
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
232
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
233
"""instantiate a Repository.
235
:param _format: The format of the repository on disk.
236
:param a_bzrdir: The BzrDir of the repository.
238
In the future we will have a single api for all stores for
239
getting file texts, inventories and revisions, then
240
this construct will accept instances of those things.
242
super(Repository, self).__init__()
243
self._format = _format
244
# the following are part of the public API for Repository:
245
self.bzrdir = a_bzrdir
246
self.control_files = control_files
247
self._revision_store = _revision_store
248
# backwards compatibility
249
self.weave_store = text_store
251
self._reconcile_does_inventory_gc = True
252
# not right yet - should be more semantically clear ?
254
self.control_store = control_store
255
self.control_weaves = control_store
256
# TODO: make sure to construct the right store classes, etc, depending
257
# on whether escaping is required.
258
self._warn_if_deprecated()
259
self._write_group = None
262
return '%s(%r)' % (self.__class__.__name__,
263
self.bzrdir.transport.base)
265
def has_same_location(self, other):
266
"""Returns a boolean indicating if this repository is at the same
267
location as another repository.
269
This might return False even when two repository objects are accessing
270
the same physical repository via different URLs.
272
if self.__class__ is not other.__class__:
274
return (self.control_files._transport.base ==
275
other.control_files._transport.base)
277
def is_in_write_group(self):
278
"""Return True if there is an open write group.
280
:seealso: start_write_group.
282
return self._write_group is not None
285
return self.control_files.is_locked()
287
def lock_write(self, token=None):
288
"""Lock this repository for writing.
290
This causes caching within the repository obejct to start accumlating
291
data during reads, and allows a 'write_group' to be obtained. Write
292
groups must be used for actual data insertion.
294
:param token: if this is already locked, then lock_write will fail
295
unless the token matches the existing lock.
296
:returns: a token if this instance supports tokens, otherwise None.
297
:raises TokenLockingNotSupported: when a token is given but this
298
instance doesn't support using token locks.
299
:raises MismatchedToken: if the specified token doesn't match the token
300
of the existing lock.
301
:seealso: start_write_group.
303
A token should be passed in if you know that you have locked the object
304
some other way, and need to synchronise this object's state with that
307
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
309
result = self.control_files.lock_write(token=token)
314
self.control_files.lock_read()
317
def get_physical_lock_status(self):
318
return self.control_files.get_physical_lock_status()
320
def leave_lock_in_place(self):
321
"""Tell this repository not to release the physical lock when this
324
If lock_write doesn't return a token, then this method is not supported.
326
self.control_files.leave_in_place()
328
def dont_leave_lock_in_place(self):
329
"""Tell this repository to release the physical lock when this
330
object is unlocked, even if it didn't originally acquire it.
332
If lock_write doesn't return a token, then this method is not supported.
334
self.control_files.dont_leave_in_place()
337
def gather_stats(self, revid=None, committers=None):
338
"""Gather statistics from a revision id.
340
:param revid: The revision id to gather statistics from, if None, then
341
no revision specific statistics are gathered.
342
:param committers: Optional parameter controlling whether to grab
343
a count of committers from the revision specific statistics.
344
:return: A dictionary of statistics. Currently this contains:
345
committers: The number of committers if requested.
346
firstrev: A tuple with timestamp, timezone for the penultimate left
347
most ancestor of revid, if revid is not the NULL_REVISION.
348
latestrev: A tuple with timestamp, timezone for revid, if revid is
349
not the NULL_REVISION.
350
revisions: The total revision count in the repository.
351
size: An estimate disk size of the repository in bytes.
354
if revid and committers:
355
result['committers'] = 0
356
if revid and revid != _mod_revision.NULL_REVISION:
358
all_committers = set()
359
revisions = self.get_ancestry(revid)
360
# pop the leading None
362
first_revision = None
364
# ignore the revisions in the middle - just grab first and last
365
revisions = revisions[0], revisions[-1]
366
for revision in self.get_revisions(revisions):
367
if not first_revision:
368
first_revision = revision
370
all_committers.add(revision.committer)
371
last_revision = revision
373
result['committers'] = len(all_committers)
374
result['firstrev'] = (first_revision.timestamp,
375
first_revision.timezone)
376
result['latestrev'] = (last_revision.timestamp,
377
last_revision.timezone)
379
# now gather global repository information
380
if self.bzrdir.root_transport.listable():
381
c, t = self._revision_store.total_size(self.get_transaction())
382
result['revisions'] = c
386
def get_data_stream(self, revision_ids):
387
raise NotImplementedError(self.get_data_stream)
389
def insert_data_stream(self, stream):
390
for item_key, bytes in stream:
391
if item_key[0] == 'file':
392
(file_id,) = item_key[1:]
393
knit = self.weave_store.get_weave_or_empty(
394
file_id, self.get_transaction())
395
elif item_key == ('inventory',):
396
knit = self.get_inventory_weave()
397
elif item_key == ('revisions',):
398
knit = self._revision_store.get_revision_file(
399
self.get_transaction())
400
elif item_key == ('signatures',):
401
knit = self._revision_store.get_signature_file(
402
self.get_transaction())
404
raise RepositoryDataStreamError(
405
"Unrecognised data stream key '%s'" % (item_key,))
406
decoded_list = bencode.bdecode(bytes)
407
format = decoded_list.pop(0)
410
for version, options, parents, some_bytes in decoded_list:
411
data_list.append((version, options, len(some_bytes), parents))
412
knit_bytes += some_bytes
413
knit.insert_data_stream(
414
(format, data_list, StringIO(knit_bytes).read))
417
def missing_revision_ids(self, other, revision_id=None):
418
"""Return the revision ids that other has that this does not.
420
These are returned in topological order.
422
revision_id: only return revision ids included by revision_id.
424
revision_id = osutils.safe_revision_id(revision_id)
425
return InterRepository.get(other, self).missing_revision_ids(revision_id)
429
"""Open the repository rooted at base.
431
For instance, if the repository is at URL/.bzr/repository,
432
Repository.open(URL) -> a Repository instance.
434
control = bzrdir.BzrDir.open(base)
435
return control.open_repository()
437
def copy_content_into(self, destination, revision_id=None):
438
"""Make a complete copy of the content in self into destination.
440
This is a destructive operation! Do not use it on existing
443
revision_id = osutils.safe_revision_id(revision_id)
444
return InterRepository.get(self, destination).copy_content(revision_id)
446
def commit_write_group(self):
447
"""Commit the contents accrued within the current write group.
449
:seealso: start_write_group.
451
if self._write_group is not self.get_transaction():
452
# has an unlock or relock occured ?
453
raise errors.BzrError('mismatched lock context and write group.')
454
self._commit_write_group()
455
self._write_group = None
457
def _commit_write_group(self):
458
"""Template method for per-repository write group cleanup.
460
This is called before the write group is considered to be
461
finished and should ensure that all data handed to the repository
462
for writing during the write group is safely committed (to the
463
extent possible considering file system caching etc).
466
def fetch(self, source, revision_id=None, pb=None):
467
"""Fetch the content required to construct revision_id from source.
469
If revision_id is None all content is copied.
471
revision_id = osutils.safe_revision_id(revision_id)
472
inter = InterRepository.get(source, self)
474
return inter.fetch(revision_id=revision_id, pb=pb)
475
except NotImplementedError:
476
raise errors.IncompatibleRepositories(source, self)
478
def create_bundle(self, target, base, fileobj, format=None):
479
return serializer.write_bundle(self, target, base, fileobj, format)
481
def get_commit_builder(self, branch, parents, config, timestamp=None,
482
timezone=None, committer=None, revprops=None,
484
"""Obtain a CommitBuilder for this repository.
486
:param branch: Branch to commit to.
487
:param parents: Revision ids of the parents of the new revision.
488
:param config: Configuration to use.
489
:param timestamp: Optional timestamp recorded for commit.
490
:param timezone: Optional timezone for timestamp.
491
:param committer: Optional committer to set for commit.
492
:param revprops: Optional dictionary of revision properties.
493
:param revision_id: Optional revision id.
495
revision_id = osutils.safe_revision_id(revision_id)
496
result =_CommitBuilder(self, parents, config, timestamp, timezone,
497
committer, revprops, revision_id)
498
self.start_write_group()
502
if (self.control_files._lock_count == 1 and
503
self.control_files._lock_mode == 'w'):
504
if self._write_group is not None:
505
raise errors.BzrError(
506
'Must end write groups before releasing write locks.')
507
self.control_files.unlock()
510
def clone(self, a_bzrdir, revision_id=None):
511
"""Clone this repository into a_bzrdir using the current format.
513
Currently no check is made that the format of this repository and
514
the bzrdir format are compatible. FIXME RBC 20060201.
516
:return: The newly created destination repository.
518
# TODO: deprecate after 0.16; cloning this with all its settings is
519
# probably not very useful -- mbp 20070423
520
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
521
self.copy_content_into(dest_repo, revision_id)
524
def start_write_group(self):
525
"""Start a write group in the repository.
527
Write groups are used by repositories which do not have a 1:1 mapping
528
between file ids and backend store to manage the insertion of data from
529
both fetch and commit operations.
531
A write lock is required around the start_write_group/commit_write_group
532
for the support of lock-requiring repository formats.
534
One can only insert data into a repository inside a write group.
538
if not self.is_locked() or self.control_files._lock_mode != 'w':
539
raise errors.NotWriteLocked(self)
540
if self._write_group:
541
raise errors.BzrError('already in a write group')
542
self._start_write_group()
543
# so we can detect unlock/relock - the write group is now entered.
544
self._write_group = self.get_transaction()
546
def _start_write_group(self):
547
"""Template method for per-repository write group startup.
549
This is called before the write group is considered to be
554
def sprout(self, to_bzrdir, revision_id=None):
555
"""Create a descendent repository for new development.
557
Unlike clone, this does not copy the settings of the repository.
559
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
560
dest_repo.fetch(self, revision_id=revision_id)
563
def _create_sprouting_repo(self, a_bzrdir, shared):
564
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
565
# use target default format.
566
dest_repo = a_bzrdir.create_repository()
568
# Most control formats need the repository to be specifically
569
# created, but on some old all-in-one formats it's not needed
571
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
572
except errors.UninitializableFormat:
573
dest_repo = a_bzrdir.open_repository()
577
def has_revision(self, revision_id):
578
"""True if this repository has a copy of the revision."""
579
if 'evil' in debug.debug_flags:
580
mutter_callsite(2, "has_revision is a LBYL symptom.")
581
revision_id = osutils.safe_revision_id(revision_id)
582
return self._revision_store.has_revision_id(revision_id,
583
self.get_transaction())
586
def get_revision_reconcile(self, revision_id):
587
"""'reconcile' helper routine that allows access to a revision always.
589
This variant of get_revision does not cross check the weave graph
590
against the revision one as get_revision does: but it should only
591
be used by reconcile, or reconcile-alike commands that are correcting
592
or testing the revision graph.
594
if not revision_id or not isinstance(revision_id, basestring):
595
raise errors.InvalidRevisionId(revision_id=revision_id,
597
return self.get_revisions([revision_id])[0]
600
def get_revisions(self, revision_ids):
601
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
602
revs = self._revision_store.get_revisions(revision_ids,
603
self.get_transaction())
605
assert not isinstance(rev.revision_id, unicode)
606
for parent_id in rev.parent_ids:
607
assert not isinstance(parent_id, unicode)
611
def get_revision_xml(self, revision_id):
612
# TODO: jam 20070210 This shouldn't be necessary since get_revision
613
# would have already do it.
614
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
615
revision_id = osutils.safe_revision_id(revision_id)
616
rev = self.get_revision(revision_id)
618
# the current serializer..
619
self._revision_store._serializer.write_revision(rev, rev_tmp)
621
return rev_tmp.getvalue()
624
def get_revision(self, revision_id):
625
"""Return the Revision object for a named revision"""
626
# TODO: jam 20070210 get_revision_reconcile should do this for us
627
revision_id = osutils.safe_revision_id(revision_id)
628
r = self.get_revision_reconcile(revision_id)
629
# weave corruption can lead to absent revision markers that should be
631
# the following test is reasonably cheap (it needs a single weave read)
632
# and the weave is cached in read transactions. In write transactions
633
# it is not cached but typically we only read a small number of
634
# revisions. For knits when they are introduced we will probably want
635
# to ensure that caching write transactions are in use.
636
inv = self.get_inventory_weave()
637
self._check_revision_parents(r, inv)
641
def get_deltas_for_revisions(self, revisions):
642
"""Produce a generator of revision deltas.
644
Note that the input is a sequence of REVISIONS, not revision_ids.
645
Trees will be held in memory until the generator exits.
646
Each delta is relative to the revision's lefthand predecessor.
648
required_trees = set()
649
for revision in revisions:
650
required_trees.add(revision.revision_id)
651
required_trees.update(revision.parent_ids[:1])
652
trees = dict((t.get_revision_id(), t) for
653
t in self.revision_trees(required_trees))
654
for revision in revisions:
655
if not revision.parent_ids:
656
old_tree = self.revision_tree(None)
658
old_tree = trees[revision.parent_ids[0]]
659
yield trees[revision.revision_id].changes_from(old_tree)
662
def get_revision_delta(self, revision_id):
663
"""Return the delta for one revision.
665
The delta is relative to the left-hand predecessor of the
668
r = self.get_revision(revision_id)
669
return list(self.get_deltas_for_revisions([r]))[0]
671
def _check_revision_parents(self, revision, inventory):
672
"""Private to Repository and Fetch.
674
This checks the parentage of revision in an inventory weave for
675
consistency and is only applicable to inventory-weave-for-ancestry
676
using repository formats & fetchers.
678
weave_parents = inventory.get_parents(revision.revision_id)
679
weave_names = inventory.versions()
680
for parent_id in revision.parent_ids:
681
if parent_id in weave_names:
682
# this parent must not be a ghost.
683
if not parent_id in weave_parents:
685
raise errors.CorruptRepository(self)
688
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
689
revision_id = osutils.safe_revision_id(revision_id)
690
signature = gpg_strategy.sign(plaintext)
691
self._revision_store.add_revision_signature_text(revision_id,
693
self.get_transaction())
695
def fileids_altered_by_revision_ids(self, revision_ids):
696
"""Find the file ids and versions affected by revisions.
698
:param revisions: an iterable containing revision ids.
699
:return: a dictionary mapping altered file-ids to an iterable of
700
revision_ids. Each altered file-ids has the exact revision_ids that
701
altered it listed explicitly.
703
assert self._serializer.support_altered_by_hack, \
704
("fileids_altered_by_revision_ids only supported for branches "
705
"which store inventory as unnested xml, not on %r" % self)
706
selected_revision_ids = set(osutils.safe_revision_id(r)
707
for r in revision_ids)
708
w = self.get_inventory_weave()
711
# this code needs to read every new line in every inventory for the
712
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
713
# not present in one of those inventories is unnecessary but not
714
# harmful because we are filtering by the revision id marker in the
715
# inventory lines : we only select file ids altered in one of those
716
# revisions. We don't need to see all lines in the inventory because
717
# only those added in an inventory in rev X can contain a revision=X
719
unescape_revid_cache = {}
720
unescape_fileid_cache = {}
722
# jam 20061218 In a big fetch, this handles hundreds of thousands
723
# of lines, so it has had a lot of inlining and optimizing done.
724
# Sorry that it is a little bit messy.
725
# Move several functions to be local variables, since this is a long
727
search = self._file_ids_altered_regex.search
728
unescape = _unescape_xml
729
setdefault = result.setdefault
730
pb = ui.ui_factory.nested_progress_bar()
732
for line in w.iter_lines_added_or_present_in_versions(
733
selected_revision_ids, pb=pb):
737
# One call to match.group() returning multiple items is quite a
738
# bit faster than 2 calls to match.group() each returning 1
739
file_id, revision_id = match.group('file_id', 'revision_id')
741
# Inlining the cache lookups helps a lot when you make 170,000
742
# lines and 350k ids, versus 8.4 unique ids.
743
# Using a cache helps in 2 ways:
744
# 1) Avoids unnecessary decoding calls
745
# 2) Re-uses cached strings, which helps in future set and
747
# (2) is enough that removing encoding entirely along with
748
# the cache (so we are using plain strings) results in no
749
# performance improvement.
751
revision_id = unescape_revid_cache[revision_id]
753
unescaped = unescape(revision_id)
754
unescape_revid_cache[revision_id] = unescaped
755
revision_id = unescaped
757
if revision_id in selected_revision_ids:
759
file_id = unescape_fileid_cache[file_id]
761
unescaped = unescape(file_id)
762
unescape_fileid_cache[file_id] = unescaped
764
setdefault(file_id, set()).add(revision_id)
769
def iter_files_bytes(self, desired_files):
770
"""Iterate through file versions.
772
Files will not necessarily be returned in the order they occur in
773
desired_files. No specific order is guaranteed.
775
Yields pairs of identifier, bytes_iterator. identifier is an opaque
776
value supplied by the caller as part of desired_files. It should
777
uniquely identify the file version in the caller's context. (Examples:
778
an index number or a TreeTransform trans_id.)
780
bytes_iterator is an iterable of bytestrings for the file. The
781
kind of iterable and length of the bytestrings are unspecified, but for
782
this implementation, it is a list of lines produced by
783
VersionedFile.get_lines().
785
:param desired_files: a list of (file_id, revision_id, identifier)
788
transaction = self.get_transaction()
789
for file_id, revision_id, callable_data in desired_files:
791
weave = self.weave_store.get_weave(file_id, transaction)
792
except errors.NoSuchFile:
793
raise errors.NoSuchIdInRepository(self, file_id)
794
yield callable_data, weave.get_lines(revision_id)
796
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
797
"""Get an iterable listing the keys of all the data introduced by a set
800
The keys will be ordered so that the corresponding items can be safely
801
fetched and inserted in that order.
803
:returns: An iterable producing tuples of (knit-kind, file-id,
804
versions). knit-kind is one of 'file', 'inventory', 'signatures',
805
'revisions'. file-id is None unless knit-kind is 'file'.
807
# XXX: it's a bit weird to control the inventory weave caching in this
808
# generator. Ideally the caching would be done in fetch.py I think. Or
809
# maybe this generator should explicitly have the contract that it
810
# should not be iterated until the previously yielded item has been
813
inv_w = self.get_inventory_weave()
816
# file ids that changed
817
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
819
num_file_ids = len(file_ids)
820
for file_id, altered_versions in file_ids.iteritems():
821
if _files_pb is not None:
822
_files_pb.update("fetch texts", count, num_file_ids)
824
yield ("file", file_id, altered_versions)
825
# We're done with the files_pb. Note that it finished by the caller,
826
# just as it was created by the caller.
830
yield ("inventory", None, revision_ids)
834
revisions_with_signatures = set()
835
for rev_id in revision_ids:
837
self.get_signature_text(rev_id)
838
except errors.NoSuchRevision:
842
revisions_with_signatures.add(rev_id)
844
yield ("signatures", None, revisions_with_signatures)
847
yield ("revisions", None, revision_ids)
850
def get_inventory_weave(self):
851
return self.control_weaves.get_weave('inventory',
852
self.get_transaction())
855
def get_inventory(self, revision_id):
856
"""Get Inventory object by hash."""
857
# TODO: jam 20070210 Technically we don't need to sanitize, since all
858
# called functions must sanitize.
859
revision_id = osutils.safe_revision_id(revision_id)
860
return self.deserialise_inventory(
861
revision_id, self.get_inventory_xml(revision_id))
863
def deserialise_inventory(self, revision_id, xml):
864
"""Transform the xml into an inventory object.
866
:param revision_id: The expected revision id of the inventory.
867
:param xml: A serialised inventory.
869
revision_id = osutils.safe_revision_id(revision_id)
870
result = self._serializer.read_inventory_from_string(xml)
871
result.root.revision = revision_id
874
def serialise_inventory(self, inv):
875
return self._serializer.write_inventory_to_string(inv)
877
def get_serializer_format(self):
878
return self._serializer.format_num
881
def get_inventory_xml(self, revision_id):
882
"""Get inventory XML as a file object."""
883
revision_id = osutils.safe_revision_id(revision_id)
885
assert isinstance(revision_id, str), type(revision_id)
886
iw = self.get_inventory_weave()
887
return iw.get_text(revision_id)
889
raise errors.HistoryMissing(self, 'inventory', revision_id)
892
def get_inventory_sha1(self, revision_id):
893
"""Return the sha1 hash of the inventory entry
895
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
896
revision_id = osutils.safe_revision_id(revision_id)
897
return self.get_revision(revision_id).inventory_sha1
900
def get_revision_graph(self, revision_id=None):
901
"""Return a dictionary containing the revision graph.
903
:param revision_id: The revision_id to get a graph from. If None, then
904
the entire revision graph is returned. This is a deprecated mode of
905
operation and will be removed in the future.
906
:return: a dictionary of revision_id->revision_parents_list.
908
if 'evil' in debug.debug_flags:
910
"get_revision_graph scales with size of history.")
911
# special case NULL_REVISION
912
if revision_id == _mod_revision.NULL_REVISION:
914
revision_id = osutils.safe_revision_id(revision_id)
915
a_weave = self.get_inventory_weave()
916
all_revisions = self._eliminate_revisions_not_present(
918
entire_graph = dict([(node, tuple(a_weave.get_parents(node))) for
919
node in all_revisions])
920
if revision_id is None:
922
elif revision_id not in entire_graph:
923
raise errors.NoSuchRevision(self, revision_id)
925
# add what can be reached from revision_id
927
pending = set([revision_id])
928
while len(pending) > 0:
930
result[node] = entire_graph[node]
931
for revision_id in result[node]:
932
if revision_id not in result:
933
pending.add(revision_id)
937
def get_revision_graph_with_ghosts(self, revision_ids=None):
938
"""Return a graph of the revisions with ghosts marked as applicable.
940
:param revision_ids: an iterable of revisions to graph or None for all.
941
:return: a Graph object with the graph reachable from revision_ids.
943
if 'evil' in debug.debug_flags:
945
"get_revision_graph_with_ghosts scales with size of history.")
946
result = deprecated_graph.Graph()
948
pending = set(self.all_revision_ids())
951
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
952
# special case NULL_REVISION
953
if _mod_revision.NULL_REVISION in pending:
954
pending.remove(_mod_revision.NULL_REVISION)
955
required = set(pending)
958
revision_id = pending.pop()
960
rev = self.get_revision(revision_id)
961
except errors.NoSuchRevision:
962
if revision_id in required:
965
result.add_ghost(revision_id)
967
for parent_id in rev.parent_ids:
968
# is this queued or done ?
969
if (parent_id not in pending and
970
parent_id not in done):
972
pending.add(parent_id)
973
result.add_node(revision_id, rev.parent_ids)
974
done.add(revision_id)
977
def _get_history_vf(self):
978
"""Get a versionedfile whose history graph reflects all revisions.
980
For weave repositories, this is the inventory weave.
982
return self.get_inventory_weave()
984
def iter_reverse_revision_history(self, revision_id):
985
"""Iterate backwards through revision ids in the lefthand history
987
:param revision_id: The revision id to start with. All its lefthand
988
ancestors will be traversed.
990
revision_id = osutils.safe_revision_id(revision_id)
991
if revision_id in (None, _mod_revision.NULL_REVISION):
993
next_id = revision_id
994
versionedfile = self._get_history_vf()
997
parents = versionedfile.get_parents(next_id)
998
if len(parents) == 0:
1001
next_id = parents[0]
1004
def get_revision_inventory(self, revision_id):
1005
"""Return inventory of a past revision."""
1006
# TODO: Unify this with get_inventory()
1007
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1008
# must be the same as its revision, so this is trivial.
1009
if revision_id is None:
1010
# This does not make sense: if there is no revision,
1011
# then it is the current tree inventory surely ?!
1012
# and thus get_root_id() is something that looks at the last
1013
# commit on the branch, and the get_root_id is an inventory check.
1014
raise NotImplementedError
1015
# return Inventory(self.get_root_id())
1017
return self.get_inventory(revision_id)
1020
def is_shared(self):
1021
"""Return True if this repository is flagged as a shared repository."""
1022
raise NotImplementedError(self.is_shared)
1025
def reconcile(self, other=None, thorough=False):
1026
"""Reconcile this repository."""
1027
from bzrlib.reconcile import RepoReconciler
1028
reconciler = RepoReconciler(self, thorough=thorough)
1029
reconciler.reconcile()
1032
def _refresh_data(self):
1033
"""Helper called from lock_* to ensure coherency with disk.
1035
The default implementation does nothing; it is however possible
1036
for repositories to maintain loaded indices across multiple locks
1037
by checking inside their implementation of this method to see
1038
whether their indices are still valid. This depends of course on
1039
the disk format being validatable in this manner.
1043
def revision_tree(self, revision_id):
1044
"""Return Tree for a revision on this branch.
1046
`revision_id` may be None for the empty tree revision.
1048
# TODO: refactor this to use an existing revision object
1049
# so we don't need to read it in twice.
1050
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1051
return RevisionTree(self, Inventory(root_id=None),
1052
_mod_revision.NULL_REVISION)
1054
revision_id = osutils.safe_revision_id(revision_id)
1055
inv = self.get_revision_inventory(revision_id)
1056
return RevisionTree(self, inv, revision_id)
1059
def revision_trees(self, revision_ids):
1060
"""Return Tree for a revision on this branch.
1062
`revision_id` may not be None or 'null:'"""
1063
assert None not in revision_ids
1064
assert _mod_revision.NULL_REVISION not in revision_ids
1065
texts = self.get_inventory_weave().get_texts(revision_ids)
1066
for text, revision_id in zip(texts, revision_ids):
1067
inv = self.deserialise_inventory(revision_id, text)
1068
yield RevisionTree(self, inv, revision_id)
1071
def get_ancestry(self, revision_id, topo_sorted=True):
1072
"""Return a list of revision-ids integrated by a revision.
1074
The first element of the list is always None, indicating the origin
1075
revision. This might change when we have history horizons, or
1076
perhaps we should have a new API.
1078
This is topologically sorted.
1080
if _mod_revision.is_null(revision_id):
1082
revision_id = osutils.safe_revision_id(revision_id)
1083
if not self.has_revision(revision_id):
1084
raise errors.NoSuchRevision(self, revision_id)
1085
w = self.get_inventory_weave()
1086
candidates = w.get_ancestry(revision_id, topo_sorted)
1087
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1090
"""Compress the data within the repository.
1092
This operation only makes sense for some repository types. For other
1093
types it should be a no-op that just returns.
1095
This stub method does not require a lock, but subclasses should use
1096
@needs_write_lock as this is a long running call its reasonable to
1097
implicitly lock for the user.
1101
def print_file(self, file, revision_id):
1102
"""Print `file` to stdout.
1104
FIXME RBC 20060125 as John Meinel points out this is a bad api
1105
- it writes to stdout, it assumes that that is valid etc. Fix
1106
by creating a new more flexible convenience function.
1108
revision_id = osutils.safe_revision_id(revision_id)
1109
tree = self.revision_tree(revision_id)
1110
# use inventory as it was in that revision
1111
file_id = tree.inventory.path2id(file)
1113
# TODO: jam 20060427 Write a test for this code path
1114
# it had a bug in it, and was raising the wrong
1116
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1117
tree.print_file(file_id)
1119
def get_transaction(self):
1120
return self.control_files.get_transaction()
1122
def revision_parents(self, revision_id):
1123
revision_id = osutils.safe_revision_id(revision_id)
1124
return self.get_inventory_weave().parent_names(revision_id)
1126
def get_parents(self, revision_ids):
1127
"""See StackedParentsProvider.get_parents"""
1129
for revision_id in revision_ids:
1130
if revision_id == _mod_revision.NULL_REVISION:
1134
parents = self.get_revision(revision_id).parent_ids
1135
except errors.NoSuchRevision:
1138
if len(parents) == 0:
1139
parents = [_mod_revision.NULL_REVISION]
1140
parents_list.append(parents)
1143
def _make_parents_provider(self):
1146
def get_graph(self, other_repository=None):
1147
"""Return the graph walker for this repository format"""
1148
parents_provider = self._make_parents_provider()
1149
if (other_repository is not None and
1150
other_repository.bzrdir.transport.base !=
1151
self.bzrdir.transport.base):
1152
parents_provider = graph._StackedParentsProvider(
1153
[parents_provider, other_repository._make_parents_provider()])
1154
return graph.Graph(parents_provider)
1157
def set_make_working_trees(self, new_value):
1158
"""Set the policy flag for making working trees when creating branches.
1160
This only applies to branches that use this repository.
1162
The default is 'True'.
1163
:param new_value: True to restore the default, False to disable making
1166
raise NotImplementedError(self.set_make_working_trees)
1168
def make_working_trees(self):
1169
"""Returns the policy for making working trees on new branches."""
1170
raise NotImplementedError(self.make_working_trees)
1173
def sign_revision(self, revision_id, gpg_strategy):
1174
revision_id = osutils.safe_revision_id(revision_id)
1175
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1176
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1179
def has_signature_for_revision_id(self, revision_id):
1180
"""Query for a revision signature for revision_id in the repository."""
1181
revision_id = osutils.safe_revision_id(revision_id)
1182
return self._revision_store.has_signature(revision_id,
1183
self.get_transaction())
1186
def get_signature_text(self, revision_id):
1187
"""Return the text for a signature."""
1188
revision_id = osutils.safe_revision_id(revision_id)
1189
return self._revision_store.get_signature_text(revision_id,
1190
self.get_transaction())
1193
def check(self, revision_ids):
1194
"""Check consistency of all history of given revision_ids.
1196
Different repository implementations should override _check().
1198
:param revision_ids: A non-empty list of revision_ids whose ancestry
1199
will be checked. Typically the last revision_id of a branch.
1201
if not revision_ids:
1202
raise ValueError("revision_ids must be non-empty in %s.check"
1204
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
1205
return self._check(revision_ids)
1207
def _check(self, revision_ids):
1208
result = check.Check(self)
1212
def _warn_if_deprecated(self):
1213
global _deprecation_warning_done
1214
if _deprecation_warning_done:
1216
_deprecation_warning_done = True
1217
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1218
% (self._format, self.bzrdir.transport.base))
1220
def supports_rich_root(self):
1221
return self._format.rich_root_data
1223
def _check_ascii_revisionid(self, revision_id, method):
1224
"""Private helper for ascii-only repositories."""
1225
# weave repositories refuse to store revisionids that are non-ascii.
1226
if revision_id is not None:
1227
# weaves require ascii revision ids.
1228
if isinstance(revision_id, unicode):
1230
revision_id.encode('ascii')
1231
except UnicodeEncodeError:
1232
raise errors.NonAsciiRevisionId(method, self)
1235
revision_id.decode('ascii')
1236
except UnicodeDecodeError:
1237
raise errors.NonAsciiRevisionId(method, self)
1241
# remove these delegates a while after bzr 0.15
1242
def __make_delegated(name, from_module):
1243
def _deprecated_repository_forwarder():
1244
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1245
% (name, from_module),
1248
m = __import__(from_module, globals(), locals(), [name])
1250
return getattr(m, name)
1251
except AttributeError:
1252
raise AttributeError('module %s has no name %s'
1254
globals()[name] = _deprecated_repository_forwarder
1257
'AllInOneRepository',
1258
'WeaveMetaDirRepository',
1259
'PreSplitOutRepositoryFormat',
1260
'RepositoryFormat4',
1261
'RepositoryFormat5',
1262
'RepositoryFormat6',
1263
'RepositoryFormat7',
1265
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1269
'RepositoryFormatKnit',
1270
'RepositoryFormatKnit1',
1272
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1275
def install_revision(repository, rev, revision_tree):
1276
"""Install all revision data into a repository."""
1277
present_parents = []
1279
for p_id in rev.parent_ids:
1280
if repository.has_revision(p_id):
1281
present_parents.append(p_id)
1282
parent_trees[p_id] = repository.revision_tree(p_id)
1284
parent_trees[p_id] = repository.revision_tree(None)
1286
inv = revision_tree.inventory
1287
entries = inv.iter_entries()
1288
# backwards compatibility hack: skip the root id.
1289
if not repository.supports_rich_root():
1290
path, root = entries.next()
1291
if root.revision != rev.revision_id:
1292
raise errors.IncompatibleRevision(repr(repository))
1293
# Add the texts that are not already present
1294
for path, ie in entries:
1295
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1296
repository.get_transaction())
1297
if ie.revision not in w:
1299
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1300
# with InventoryEntry.find_previous_heads(). if it is, then there
1301
# is a latent bug here where the parents may have ancestors of each
1303
for revision, tree in parent_trees.iteritems():
1304
if ie.file_id not in tree:
1306
parent_id = tree.inventory[ie.file_id].revision
1307
if parent_id in text_parents:
1309
text_parents.append(parent_id)
1311
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1312
repository.get_transaction())
1313
lines = revision_tree.get_file(ie.file_id).readlines()
1314
vfile.add_lines(rev.revision_id, text_parents, lines)
1316
# install the inventory
1317
repository.add_inventory(rev.revision_id, inv, present_parents)
1318
except errors.RevisionAlreadyPresent:
1320
repository.add_revision(rev.revision_id, rev, inv)
1323
class MetaDirRepository(Repository):
1324
"""Repositories in the new meta-dir layout."""
1326
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1327
super(MetaDirRepository, self).__init__(_format,
1333
dir_mode = self.control_files._dir_mode
1334
file_mode = self.control_files._file_mode
1337
def is_shared(self):
1338
"""Return True if this repository is flagged as a shared repository."""
1339
return self.control_files._transport.has('shared-storage')
1342
def set_make_working_trees(self, new_value):
1343
"""Set the policy flag for making working trees when creating branches.
1345
This only applies to branches that use this repository.
1347
The default is 'True'.
1348
:param new_value: True to restore the default, False to disable making
1353
self.control_files._transport.delete('no-working-trees')
1354
except errors.NoSuchFile:
1357
self.control_files.put_utf8('no-working-trees', '')
1359
def make_working_trees(self):
1360
"""Returns the policy for making working trees on new branches."""
1361
return not self.control_files._transport.has('no-working-trees')
1364
class RepositoryFormatRegistry(registry.Registry):
1365
"""Registry of RepositoryFormats.
1368
def get(self, format_string):
1369
r = registry.Registry.get(self, format_string)
1375
format_registry = RepositoryFormatRegistry()
1376
"""Registry of formats, indexed by their identifying format string.
1378
This can contain either format instances themselves, or classes/factories that
1379
can be called to obtain one.
1383
#####################################################################
1384
# Repository Formats
1386
class RepositoryFormat(object):
1387
"""A repository format.
1389
Formats provide three things:
1390
* An initialization routine to construct repository data on disk.
1391
* a format string which is used when the BzrDir supports versioned
1393
* an open routine which returns a Repository instance.
1395
Formats are placed in an dict by their format string for reference
1396
during opening. These should be subclasses of RepositoryFormat
1399
Once a format is deprecated, just deprecate the initialize and open
1400
methods on the format class. Do not deprecate the object, as the
1401
object will be created every system load.
1403
Common instance attributes:
1404
_matchingbzrdir - the bzrdir format that the repository format was
1405
originally written to work with. This can be used if manually
1406
constructing a bzrdir and repository, or more commonly for test suite
1411
return "<%s>" % self.__class__.__name__
1413
def __eq__(self, other):
1414
# format objects are generally stateless
1415
return isinstance(other, self.__class__)
1417
def __ne__(self, other):
1418
return not self == other
1421
def find_format(klass, a_bzrdir):
1422
"""Return the format for the repository object in a_bzrdir.
1424
This is used by bzr native formats that have a "format" file in
1425
the repository. Other methods may be used by different types of
1429
transport = a_bzrdir.get_repository_transport(None)
1430
format_string = transport.get("format").read()
1431
return format_registry.get(format_string)
1432
except errors.NoSuchFile:
1433
raise errors.NoRepositoryPresent(a_bzrdir)
1435
raise errors.UnknownFormatError(format=format_string)
1438
def register_format(klass, format):
1439
format_registry.register(format.get_format_string(), format)
1442
def unregister_format(klass, format):
1443
format_registry.remove(format.get_format_string())
1446
def get_default_format(klass):
1447
"""Return the current default format."""
1448
from bzrlib import bzrdir
1449
return bzrdir.format_registry.make_bzrdir('default').repository_format
1451
def _get_control_store(self, repo_transport, control_files):
1452
"""Return the control store for this repository."""
1453
raise NotImplementedError(self._get_control_store)
1455
def get_format_string(self):
1456
"""Return the ASCII format string that identifies this format.
1458
Note that in pre format ?? repositories the format string is
1459
not permitted nor written to disk.
1461
raise NotImplementedError(self.get_format_string)
1463
def get_format_description(self):
1464
"""Return the short description for this format."""
1465
raise NotImplementedError(self.get_format_description)
1467
def _get_revision_store(self, repo_transport, control_files):
1468
"""Return the revision store object for this a_bzrdir."""
1469
raise NotImplementedError(self._get_revision_store)
1471
def _get_text_rev_store(self,
1478
"""Common logic for getting a revision store for a repository.
1480
see self._get_revision_store for the subclass-overridable method to
1481
get the store for a repository.
1483
from bzrlib.store.revision.text import TextRevisionStore
1484
dir_mode = control_files._dir_mode
1485
file_mode = control_files._file_mode
1486
text_store = TextStore(transport.clone(name),
1488
compressed=compressed,
1490
file_mode=file_mode)
1491
_revision_store = TextRevisionStore(text_store, serializer)
1492
return _revision_store
1494
# TODO: this shouldn't be in the base class, it's specific to things that
1495
# use weaves or knits -- mbp 20070207
1496
def _get_versioned_file_store(self,
1501
versionedfile_class=None,
1502
versionedfile_kwargs={},
1504
if versionedfile_class is None:
1505
versionedfile_class = self._versionedfile_class
1506
weave_transport = control_files._transport.clone(name)
1507
dir_mode = control_files._dir_mode
1508
file_mode = control_files._file_mode
1509
return VersionedFileStore(weave_transport, prefixed=prefixed,
1511
file_mode=file_mode,
1512
versionedfile_class=versionedfile_class,
1513
versionedfile_kwargs=versionedfile_kwargs,
1516
def initialize(self, a_bzrdir, shared=False):
1517
"""Initialize a repository of this format in a_bzrdir.
1519
:param a_bzrdir: The bzrdir to put the new repository in it.
1520
:param shared: The repository should be initialized as a sharable one.
1521
:returns: The new repository object.
1523
This may raise UninitializableFormat if shared repository are not
1524
compatible the a_bzrdir.
1526
raise NotImplementedError(self.initialize)
1528
def is_supported(self):
1529
"""Is this format supported?
1531
Supported formats must be initializable and openable.
1532
Unsupported formats may not support initialization or committing or
1533
some other features depending on the reason for not being supported.
1537
def check_conversion_target(self, target_format):
1538
raise NotImplementedError(self.check_conversion_target)
1540
def open(self, a_bzrdir, _found=False):
1541
"""Return an instance of this format for the bzrdir a_bzrdir.
1543
_found is a private parameter, do not use it.
1545
raise NotImplementedError(self.open)
1548
class MetaDirRepositoryFormat(RepositoryFormat):
1549
"""Common base class for the new repositories using the metadir layout."""
1551
rich_root_data = False
1552
supports_tree_reference = False
1553
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1556
super(MetaDirRepositoryFormat, self).__init__()
1558
def _create_control_files(self, a_bzrdir):
1559
"""Create the required files and the initial control_files object."""
1560
# FIXME: RBC 20060125 don't peek under the covers
1561
# NB: no need to escape relative paths that are url safe.
1562
repository_transport = a_bzrdir.get_repository_transport(self)
1563
control_files = lockable_files.LockableFiles(repository_transport,
1564
'lock', lockdir.LockDir)
1565
control_files.create_lock()
1566
return control_files
1568
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1569
"""Upload the initial blank content."""
1570
control_files = self._create_control_files(a_bzrdir)
1571
control_files.lock_write()
1573
control_files._transport.mkdir_multi(dirs,
1574
mode=control_files._dir_mode)
1575
for file, content in files:
1576
control_files.put(file, content)
1577
for file, content in utf8_files:
1578
control_files.put_utf8(file, content)
1580
control_files.put_utf8('shared-storage', '')
1582
control_files.unlock()
1585
# formats which have no format string are not discoverable
1586
# and not independently creatable, so are not registered. They're
1587
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1588
# needed, it's constructed directly by the BzrDir. Non-native formats where
1589
# the repository is not separately opened are similar.
1591
format_registry.register_lazy(
1592
'Bazaar-NG Repository format 7',
1593
'bzrlib.repofmt.weaverepo',
1596
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1597
# default control directory format
1599
format_registry.register_lazy(
1600
'Bazaar-NG Knit Repository Format 1',
1601
'bzrlib.repofmt.knitrepo',
1602
'RepositoryFormatKnit1',
1604
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1606
format_registry.register_lazy(
1607
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1608
'bzrlib.repofmt.knitrepo',
1609
'RepositoryFormatKnit3',
1613
class InterRepository(InterObject):
1614
"""This class represents operations taking place between two repositories.
1616
Its instances have methods like copy_content and fetch, and contain
1617
references to the source and target repositories these operations can be
1620
Often we will provide convenience methods on 'repository' which carry out
1621
operations with another repository - they will always forward to
1622
InterRepository.get(other).method_name(parameters).
1626
"""The available optimised InterRepository types."""
1628
def copy_content(self, revision_id=None):
1629
raise NotImplementedError(self.copy_content)
1631
def fetch(self, revision_id=None, pb=None):
1632
"""Fetch the content required to construct revision_id.
1634
The content is copied from self.source to self.target.
1636
:param revision_id: if None all content is copied, if NULL_REVISION no
1638
:param pb: optional progress bar to use for progress reports. If not
1639
provided a default one will be created.
1641
Returns the copied revision count and the failed revisions in a tuple:
1644
raise NotImplementedError(self.fetch)
1647
def missing_revision_ids(self, revision_id=None):
1648
"""Return the revision ids that source has that target does not.
1650
These are returned in topological order.
1652
:param revision_id: only return revision ids included by this
1655
# generic, possibly worst case, slow code path.
1656
target_ids = set(self.target.all_revision_ids())
1657
if revision_id is not None:
1658
# TODO: jam 20070210 InterRepository is internal enough that it
1659
# should assume revision_ids are already utf-8
1660
revision_id = osutils.safe_revision_id(revision_id)
1661
source_ids = self.source.get_ancestry(revision_id)
1662
assert source_ids[0] is None
1665
source_ids = self.source.all_revision_ids()
1666
result_set = set(source_ids).difference(target_ids)
1667
# this may look like a no-op: its not. It preserves the ordering
1668
# other_ids had while only returning the members from other_ids
1669
# that we've decided we need.
1670
return [rev_id for rev_id in source_ids if rev_id in result_set]
1673
class InterSameDataRepository(InterRepository):
1674
"""Code for converting between repositories that represent the same data.
1676
Data format and model must match for this to work.
1680
def _get_repo_format_to_test(self):
1681
"""Repository format for testing with."""
1682
return RepositoryFormat.get_default_format()
1685
def is_compatible(source, target):
1686
if source.supports_rich_root() != target.supports_rich_root():
1688
if source._serializer != target._serializer:
1693
def copy_content(self, revision_id=None):
1694
"""Make a complete copy of the content in self into destination.
1696
This copies both the repository's revision data, and configuration information
1697
such as the make_working_trees setting.
1699
This is a destructive operation! Do not use it on existing
1702
:param revision_id: Only copy the content needed to construct
1703
revision_id and its parents.
1706
self.target.set_make_working_trees(self.source.make_working_trees())
1707
except NotImplementedError:
1709
# TODO: jam 20070210 This is fairly internal, so we should probably
1710
# just assert that revision_id is not unicode.
1711
revision_id = osutils.safe_revision_id(revision_id)
1712
# but don't bother fetching if we have the needed data now.
1713
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1714
self.target.has_revision(revision_id)):
1716
self.target.fetch(self.source, revision_id=revision_id)
1719
def fetch(self, revision_id=None, pb=None):
1720
"""See InterRepository.fetch()."""
1721
from bzrlib.fetch import GenericRepoFetcher
1722
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1723
self.source, self.source._format, self.target,
1724
self.target._format)
1725
# TODO: jam 20070210 This should be an assert, not a translate
1726
revision_id = osutils.safe_revision_id(revision_id)
1727
f = GenericRepoFetcher(to_repository=self.target,
1728
from_repository=self.source,
1729
last_revision=revision_id,
1731
return f.count_copied, f.failed_revisions
1734
class InterWeaveRepo(InterSameDataRepository):
1735
"""Optimised code paths between Weave based repositories."""
1738
def _get_repo_format_to_test(self):
1739
from bzrlib.repofmt import weaverepo
1740
return weaverepo.RepositoryFormat7()
1743
def is_compatible(source, target):
1744
"""Be compatible with known Weave formats.
1746
We don't test for the stores being of specific types because that
1747
could lead to confusing results, and there is no need to be
1750
from bzrlib.repofmt.weaverepo import (
1756
return (isinstance(source._format, (RepositoryFormat5,
1758
RepositoryFormat7)) and
1759
isinstance(target._format, (RepositoryFormat5,
1761
RepositoryFormat7)))
1762
except AttributeError:
1766
def copy_content(self, revision_id=None):
1767
"""See InterRepository.copy_content()."""
1768
# weave specific optimised path:
1769
# TODO: jam 20070210 Internal, should be an assert, not translate
1770
revision_id = osutils.safe_revision_id(revision_id)
1772
self.target.set_make_working_trees(self.source.make_working_trees())
1773
except NotImplementedError:
1775
# FIXME do not peek!
1776
if self.source.control_files._transport.listable():
1777
pb = ui.ui_factory.nested_progress_bar()
1779
self.target.weave_store.copy_all_ids(
1780
self.source.weave_store,
1782
from_transaction=self.source.get_transaction(),
1783
to_transaction=self.target.get_transaction())
1784
pb.update('copying inventory', 0, 1)
1785
self.target.control_weaves.copy_multi(
1786
self.source.control_weaves, ['inventory'],
1787
from_transaction=self.source.get_transaction(),
1788
to_transaction=self.target.get_transaction())
1789
self.target._revision_store.text_store.copy_all_ids(
1790
self.source._revision_store.text_store,
1795
self.target.fetch(self.source, revision_id=revision_id)
1798
def fetch(self, revision_id=None, pb=None):
1799
"""See InterRepository.fetch()."""
1800
from bzrlib.fetch import GenericRepoFetcher
1801
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1802
self.source, self.source._format, self.target, self.target._format)
1803
# TODO: jam 20070210 This should be an assert, not a translate
1804
revision_id = osutils.safe_revision_id(revision_id)
1805
f = GenericRepoFetcher(to_repository=self.target,
1806
from_repository=self.source,
1807
last_revision=revision_id,
1809
return f.count_copied, f.failed_revisions
1812
def missing_revision_ids(self, revision_id=None):
1813
"""See InterRepository.missing_revision_ids()."""
1814
# we want all revisions to satisfy revision_id in source.
1815
# but we don't want to stat every file here and there.
1816
# we want then, all revisions other needs to satisfy revision_id
1817
# checked, but not those that we have locally.
1818
# so the first thing is to get a subset of the revisions to
1819
# satisfy revision_id in source, and then eliminate those that
1820
# we do already have.
1821
# this is slow on high latency connection to self, but as as this
1822
# disk format scales terribly for push anyway due to rewriting
1823
# inventory.weave, this is considered acceptable.
1825
if revision_id is not None:
1826
source_ids = self.source.get_ancestry(revision_id)
1827
assert source_ids[0] is None
1830
source_ids = self.source._all_possible_ids()
1831
source_ids_set = set(source_ids)
1832
# source_ids is the worst possible case we may need to pull.
1833
# now we want to filter source_ids against what we actually
1834
# have in target, but don't try to check for existence where we know
1835
# we do not have a revision as that would be pointless.
1836
target_ids = set(self.target._all_possible_ids())
1837
possibly_present_revisions = target_ids.intersection(source_ids_set)
1838
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1839
required_revisions = source_ids_set.difference(actually_present_revisions)
1840
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1841
if revision_id is not None:
1842
# we used get_ancestry to determine source_ids then we are assured all
1843
# revisions referenced are present as they are installed in topological order.
1844
# and the tip revision was validated by get_ancestry.
1845
return required_topo_revisions
1847
# if we just grabbed the possibly available ids, then
1848
# we only have an estimate of whats available and need to validate
1849
# that against the revision records.
1850
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1853
class InterKnitRepo(InterSameDataRepository):
1854
"""Optimised code paths between Knit based repositories."""
1857
def _get_repo_format_to_test(self):
1858
from bzrlib.repofmt import knitrepo
1859
return knitrepo.RepositoryFormatKnit1()
1862
def is_compatible(source, target):
1863
"""Be compatible with known Knit formats.
1865
We don't test for the stores being of specific types because that
1866
could lead to confusing results, and there is no need to be
1869
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
1871
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1872
isinstance(target._format, (RepositoryFormatKnit1)))
1873
except AttributeError:
1877
def fetch(self, revision_id=None, pb=None):
1878
"""See InterRepository.fetch()."""
1879
from bzrlib.fetch import KnitRepoFetcher
1880
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1881
self.source, self.source._format, self.target, self.target._format)
1882
# TODO: jam 20070210 This should be an assert, not a translate
1883
revision_id = osutils.safe_revision_id(revision_id)
1884
f = KnitRepoFetcher(to_repository=self.target,
1885
from_repository=self.source,
1886
last_revision=revision_id,
1888
return f.count_copied, f.failed_revisions
1891
def missing_revision_ids(self, revision_id=None):
1892
"""See InterRepository.missing_revision_ids()."""
1893
if revision_id is not None:
1894
source_ids = self.source.get_ancestry(revision_id)
1895
assert source_ids[0] is None
1898
source_ids = self.source._all_possible_ids()
1899
source_ids_set = set(source_ids)
1900
# source_ids is the worst possible case we may need to pull.
1901
# now we want to filter source_ids against what we actually
1902
# have in target, but don't try to check for existence where we know
1903
# we do not have a revision as that would be pointless.
1904
target_ids = set(self.target._all_possible_ids())
1905
possibly_present_revisions = target_ids.intersection(source_ids_set)
1906
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
1907
required_revisions = source_ids_set.difference(actually_present_revisions)
1908
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
1909
if revision_id is not None:
1910
# we used get_ancestry to determine source_ids then we are assured all
1911
# revisions referenced are present as they are installed in topological order.
1912
# and the tip revision was validated by get_ancestry.
1913
return required_topo_revisions
1915
# if we just grabbed the possibly available ids, then
1916
# we only have an estimate of whats available and need to validate
1917
# that against the revision records.
1918
return self.source._eliminate_revisions_not_present(required_topo_revisions)
1921
class InterModel1and2(InterRepository):
1924
def _get_repo_format_to_test(self):
1928
def is_compatible(source, target):
1929
if not source.supports_rich_root() and target.supports_rich_root():
1935
def fetch(self, revision_id=None, pb=None):
1936
"""See InterRepository.fetch()."""
1937
from bzrlib.fetch import Model1toKnit2Fetcher
1938
# TODO: jam 20070210 This should be an assert, not a translate
1939
revision_id = osutils.safe_revision_id(revision_id)
1940
f = Model1toKnit2Fetcher(to_repository=self.target,
1941
from_repository=self.source,
1942
last_revision=revision_id,
1944
return f.count_copied, f.failed_revisions
1947
def copy_content(self, revision_id=None):
1948
"""Make a complete copy of the content in self into destination.
1950
This is a destructive operation! Do not use it on existing
1953
:param revision_id: Only copy the content needed to construct
1954
revision_id and its parents.
1957
self.target.set_make_working_trees(self.source.make_working_trees())
1958
except NotImplementedError:
1960
# TODO: jam 20070210 Internal, assert, don't translate
1961
revision_id = osutils.safe_revision_id(revision_id)
1962
# but don't bother fetching if we have the needed data now.
1963
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
1964
self.target.has_revision(revision_id)):
1966
self.target.fetch(self.source, revision_id=revision_id)
1969
class InterKnit1and2(InterKnitRepo):
1972
def _get_repo_format_to_test(self):
1976
def is_compatible(source, target):
1977
"""Be compatible with Knit1 source and Knit3 target"""
1978
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
1980
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
1981
RepositoryFormatKnit3
1982
return (isinstance(source._format, (RepositoryFormatKnit1)) and
1983
isinstance(target._format, (RepositoryFormatKnit3)))
1984
except AttributeError:
1988
def fetch(self, revision_id=None, pb=None):
1989
"""See InterRepository.fetch()."""
1990
from bzrlib.fetch import Knit1to2Fetcher
1991
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
1992
self.source, self.source._format, self.target,
1993
self.target._format)
1994
# TODO: jam 20070210 This should be an assert, not a translate
1995
revision_id = osutils.safe_revision_id(revision_id)
1996
f = Knit1to2Fetcher(to_repository=self.target,
1997
from_repository=self.source,
1998
last_revision=revision_id,
2000
return f.count_copied, f.failed_revisions
2003
class InterRemoteToOther(InterRepository):
2005
def __init__(self, source, target):
2006
InterRepository.__init__(self, source, target)
2007
self._real_inter = None
2010
def is_compatible(source, target):
2011
if not isinstance(source, remote.RemoteRepository):
2013
source._ensure_real()
2014
real_source = source._real_repository
2015
# Is source's model compatible with target's model, and are they the
2016
# same format? Currently we can only optimise fetching from an
2017
# identical model & format repo.
2018
assert not isinstance(real_source, remote.RemoteRepository), (
2019
"We don't support remote repos backed by remote repos yet.")
2020
return real_source._format == target._format
2023
def fetch(self, revision_id=None, pb=None):
2024
"""See InterRepository.fetch()."""
2025
from bzrlib.fetch import RemoteToOtherFetcher
2026
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2027
self.source, self.target, self.target._format)
2028
# TODO: jam 20070210 This should be an assert, not a translate
2029
revision_id = osutils.safe_revision_id(revision_id)
2030
f = RemoteToOtherFetcher(to_repository=self.target,
2031
from_repository=self.source,
2032
last_revision=revision_id,
2034
return f.count_copied, f.failed_revisions
2037
def _get_repo_format_to_test(self):
2041
class InterOtherToRemote(InterRepository):
2043
def __init__(self, source, target):
2044
InterRepository.__init__(self, source, target)
2045
self._real_inter = None
2048
def is_compatible(source, target):
2049
if isinstance(target, remote.RemoteRepository):
2053
def _ensure_real_inter(self):
2054
if self._real_inter is None:
2055
self.target._ensure_real()
2056
real_target = self.target._real_repository
2057
self._real_inter = InterRepository.get(self.source, real_target)
2059
def copy_content(self, revision_id=None):
2060
self._ensure_real_inter()
2061
self._real_inter.copy_content(revision_id=revision_id)
2063
def fetch(self, revision_id=None, pb=None):
2064
self._ensure_real_inter()
2065
self._real_inter.fetch(revision_id=revision_id, pb=pb)
2068
def _get_repo_format_to_test(self):
2072
InterRepository.register_optimiser(InterSameDataRepository)
2073
InterRepository.register_optimiser(InterWeaveRepo)
2074
InterRepository.register_optimiser(InterKnitRepo)
2075
InterRepository.register_optimiser(InterModel1and2)
2076
InterRepository.register_optimiser(InterKnit1and2)
2077
InterRepository.register_optimiser(InterRemoteToOther)
2078
InterRepository.register_optimiser(InterOtherToRemote)
2081
class CopyConverter(object):
2082
"""A repository conversion tool which just performs a copy of the content.
2084
This is slow but quite reliable.
2087
def __init__(self, target_format):
2088
"""Create a CopyConverter.
2090
:param target_format: The format the resulting repository should be.
2092
self.target_format = target_format
2094
def convert(self, repo, pb):
2095
"""Perform the conversion of to_convert, giving feedback via pb.
2097
:param to_convert: The disk object to convert.
2098
:param pb: a progress bar to use for progress information.
2103
# this is only useful with metadir layouts - separated repo content.
2104
# trigger an assertion if not such
2105
repo._format.get_format_string()
2106
self.repo_dir = repo.bzrdir
2107
self.step('Moving repository to repository.backup')
2108
self.repo_dir.transport.move('repository', 'repository.backup')
2109
backup_transport = self.repo_dir.transport.clone('repository.backup')
2110
repo._format.check_conversion_target(self.target_format)
2111
self.source_repo = repo._format.open(self.repo_dir,
2113
_override_transport=backup_transport)
2114
self.step('Creating new repository')
2115
converted = self.target_format.initialize(self.repo_dir,
2116
self.source_repo.is_shared())
2117
converted.lock_write()
2119
self.step('Copying content into repository.')
2120
self.source_repo.copy_content_into(converted)
2123
self.step('Deleting old repository content.')
2124
self.repo_dir.transport.delete_tree('repository.backup')
2125
self.pb.note('repository converted')
2127
def step(self, message):
2128
"""Update the pb by a step."""
2130
self.pb.update(message, self.count, self.total)
2133
class CommitBuilder(object):
2134
"""Provides an interface to build up a commit.
2136
This allows describing a tree to be committed without needing to
2137
know the internals of the format of the repository.
2140
record_root_entry = False
2141
def __init__(self, repository, parents, config, timestamp=None,
2142
timezone=None, committer=None, revprops=None,
2144
"""Initiate a CommitBuilder.
2146
:param repository: Repository to commit to.
2147
:param parents: Revision ids of the parents of the new revision.
2148
:param config: Configuration to use.
2149
:param timestamp: Optional timestamp recorded for commit.
2150
:param timezone: Optional timezone for timestamp.
2151
:param committer: Optional committer to set for commit.
2152
:param revprops: Optional dictionary of revision properties.
2153
:param revision_id: Optional revision id.
2155
self._config = config
2157
if committer is None:
2158
self._committer = self._config.username()
2160
assert isinstance(committer, basestring), type(committer)
2161
self._committer = committer
2163
self.new_inventory = Inventory(None)
2164
self._new_revision_id = osutils.safe_revision_id(revision_id)
2165
self.parents = parents
2166
self.repository = repository
2169
if revprops is not None:
2170
self._revprops.update(revprops)
2172
if timestamp is None:
2173
timestamp = time.time()
2174
# Restrict resolution to 1ms
2175
self._timestamp = round(timestamp, 3)
2177
if timezone is None:
2178
self._timezone = osutils.local_time_offset()
2180
self._timezone = int(timezone)
2182
self._generate_revision_if_needed()
2184
def commit(self, message):
2185
"""Make the actual commit.
2187
:return: The revision id of the recorded revision.
2189
rev = _mod_revision.Revision(
2190
timestamp=self._timestamp,
2191
timezone=self._timezone,
2192
committer=self._committer,
2194
inventory_sha1=self.inv_sha1,
2195
revision_id=self._new_revision_id,
2196
properties=self._revprops)
2197
rev.parent_ids = self.parents
2198
self.repository.add_revision(self._new_revision_id, rev,
2199
self.new_inventory, self._config)
2200
self.repository.commit_write_group()
2201
return self._new_revision_id
2203
def revision_tree(self):
2204
"""Return the tree that was just committed.
2206
After calling commit() this can be called to get a RevisionTree
2207
representing the newly committed tree. This is preferred to
2208
calling Repository.revision_tree() because that may require
2209
deserializing the inventory, while we already have a copy in
2212
return RevisionTree(self.repository, self.new_inventory,
2213
self._new_revision_id)
2215
def finish_inventory(self):
2216
"""Tell the builder that the inventory is finished."""
2217
if self.new_inventory.root is None:
2218
symbol_versioning.warn('Root entry should be supplied to'
2219
' record_entry_contents, as of bzr 0.10.',
2220
DeprecationWarning, stacklevel=2)
2221
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
2222
self.new_inventory.revision_id = self._new_revision_id
2223
self.inv_sha1 = self.repository.add_inventory(
2224
self._new_revision_id,
2229
def _gen_revision_id(self):
2230
"""Return new revision-id."""
2231
return generate_ids.gen_revision_id(self._config.username(),
2234
def _generate_revision_if_needed(self):
2235
"""Create a revision id if None was supplied.
2237
If the repository can not support user-specified revision ids
2238
they should override this function and raise CannotSetRevisionId
2239
if _new_revision_id is not None.
2241
:raises: CannotSetRevisionId
2243
if self._new_revision_id is None:
2244
self._new_revision_id = self._gen_revision_id()
2246
def record_entry_contents(self, ie, parent_invs, path, tree):
2247
"""Record the content of ie from tree into the commit if needed.
2249
Side effect: sets ie.revision when unchanged
2251
:param ie: An inventory entry present in the commit.
2252
:param parent_invs: The inventories of the parent revisions of the
2254
:param path: The path the entry is at in the tree.
2255
:param tree: The tree which contains this entry and should be used to
2258
if self.new_inventory.root is None and ie.parent_id is not None:
2259
symbol_versioning.warn('Root entry should be supplied to'
2260
' record_entry_contents, as of bzr 0.10.',
2261
DeprecationWarning, stacklevel=2)
2262
self.record_entry_contents(tree.inventory.root.copy(), parent_invs,
2264
self.new_inventory.add(ie)
2266
# ie.revision is always None if the InventoryEntry is considered
2267
# for committing. ie.snapshot will record the correct revision
2268
# which may be the sole parent if it is untouched.
2269
if ie.revision is not None:
2272
# In this revision format, root entries have no knit or weave
2273
if ie is self.new_inventory.root:
2274
# When serializing out to disk and back in
2275
# root.revision is always _new_revision_id
2276
ie.revision = self._new_revision_id
2278
previous_entries = ie.find_previous_heads(
2280
self.repository.weave_store,
2281
self.repository.get_transaction())
2282
# we are creating a new revision for ie in the history store
2284
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2286
def modified_directory(self, file_id, file_parents):
2287
"""Record the presence of a symbolic link.
2289
:param file_id: The file_id of the link to record.
2290
:param file_parents: The per-file parent revision ids.
2292
self._add_text_to_weave(file_id, [], file_parents.keys())
2294
def modified_reference(self, file_id, file_parents):
2295
"""Record the modification of a reference.
2297
:param file_id: The file_id of the link to record.
2298
:param file_parents: The per-file parent revision ids.
2300
self._add_text_to_weave(file_id, [], file_parents.keys())
2302
def modified_file_text(self, file_id, file_parents,
2303
get_content_byte_lines, text_sha1=None,
2305
"""Record the text of file file_id
2307
:param file_id: The file_id of the file to record the text of.
2308
:param file_parents: The per-file parent revision ids.
2309
:param get_content_byte_lines: A callable which will return the byte
2311
:param text_sha1: Optional SHA1 of the file contents.
2312
:param text_size: Optional size of the file contents.
2314
# mutter('storing text of file {%s} in revision {%s} into %r',
2315
# file_id, self._new_revision_id, self.repository.weave_store)
2316
# special case to avoid diffing on renames or
2318
if (len(file_parents) == 1
2319
and text_sha1 == file_parents.values()[0].text_sha1
2320
and text_size == file_parents.values()[0].text_size):
2321
previous_ie = file_parents.values()[0]
2322
versionedfile = self.repository.weave_store.get_weave(file_id,
2323
self.repository.get_transaction())
2324
versionedfile.clone_text(self._new_revision_id,
2325
previous_ie.revision, file_parents.keys())
2326
return text_sha1, text_size
2328
new_lines = get_content_byte_lines()
2329
# TODO: Rather than invoking sha_strings here, _add_text_to_weave
2330
# should return the SHA1 and size
2331
self._add_text_to_weave(file_id, new_lines, file_parents.keys())
2332
return osutils.sha_strings(new_lines), \
2333
sum(map(len, new_lines))
2335
def modified_link(self, file_id, file_parents, link_target):
2336
"""Record the presence of a symbolic link.
2338
:param file_id: The file_id of the link to record.
2339
:param file_parents: The per-file parent revision ids.
2340
:param link_target: Target location of this link.
2342
self._add_text_to_weave(file_id, [], file_parents.keys())
2344
def _add_text_to_weave(self, file_id, new_lines, parents):
2345
versionedfile = self.repository.weave_store.get_weave_or_empty(
2346
file_id, self.repository.get_transaction())
2347
versionedfile.add_lines(self._new_revision_id, parents, new_lines)
2348
versionedfile.clear_cache()
2351
class _CommitBuilder(CommitBuilder):
2352
"""Temporary class so old CommitBuilders are detected properly
2354
Note: CommitBuilder works whether or not root entry is recorded.
2357
record_root_entry = True
2360
class RootCommitBuilder(CommitBuilder):
2361
"""This commitbuilder actually records the root id"""
2363
record_root_entry = True
2365
def record_entry_contents(self, ie, parent_invs, path, tree):
2366
"""Record the content of ie from tree into the commit if needed.
2368
Side effect: sets ie.revision when unchanged
2370
:param ie: An inventory entry present in the commit.
2371
:param parent_invs: The inventories of the parent revisions of the
2373
:param path: The path the entry is at in the tree.
2374
:param tree: The tree which contains this entry and should be used to
2377
assert self.new_inventory.root is not None or ie.parent_id is None
2378
self.new_inventory.add(ie)
2380
# ie.revision is always None if the InventoryEntry is considered
2381
# for committing. ie.snapshot will record the correct revision
2382
# which may be the sole parent if it is untouched.
2383
if ie.revision is not None:
2386
previous_entries = ie.find_previous_heads(
2388
self.repository.weave_store,
2389
self.repository.get_transaction())
2390
# we are creating a new revision for ie in the history store
2392
ie.snapshot(self._new_revision_id, path, previous_entries, tree, self)
2404
def _unescaper(match, _map=_unescape_map):
2405
code = match.group(1)
2409
if not code.startswith('#'):
2411
return unichr(int(code[1:])).encode('utf8')
2417
def _unescape_xml(data):
2418
"""Unescape predefined XML entities in a string of data."""
2420
if _unescape_re is None:
2421
_unescape_re = re.compile('\&([^;]*);')
2422
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
