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
class CommitBuilder(object):
66
"""Provides an interface to build up a commit.
68
This allows describing a tree to be committed without needing to
69
know the internals of the format of the repository.
72
# all clients should supply tree roots.
73
record_root_entry = True
74
# the default CommitBuilder does not manage trees whose root is versioned.
75
_versioned_root = False
77
def __init__(self, repository, parents, config, timestamp=None,
78
timezone=None, committer=None, revprops=None,
80
"""Initiate a CommitBuilder.
82
:param repository: Repository to commit to.
83
:param parents: Revision ids of the parents of the new revision.
84
:param config: Configuration to use.
85
:param timestamp: Optional timestamp recorded for commit.
86
:param timezone: Optional timezone for timestamp.
87
:param committer: Optional committer to set for commit.
88
:param revprops: Optional dictionary of revision properties.
89
:param revision_id: Optional revision id.
94
self._committer = self._config.username()
96
assert isinstance(committer, basestring), type(committer)
97
self._committer = committer
99
self.new_inventory = Inventory(None)
100
self._new_revision_id = revision_id
101
self.parents = parents
102
self.repository = repository
105
if revprops is not None:
106
self._revprops.update(revprops)
108
if timestamp is None:
109
timestamp = time.time()
110
# Restrict resolution to 1ms
111
self._timestamp = round(timestamp, 3)
114
self._timezone = osutils.local_time_offset()
116
self._timezone = int(timezone)
118
self._generate_revision_if_needed()
119
self._repo_graph = repository.get_graph()
121
def commit(self, message):
122
"""Make the actual commit.
124
:return: The revision id of the recorded revision.
126
rev = _mod_revision.Revision(
127
timestamp=self._timestamp,
128
timezone=self._timezone,
129
committer=self._committer,
131
inventory_sha1=self.inv_sha1,
132
revision_id=self._new_revision_id,
133
properties=self._revprops)
134
rev.parent_ids = self.parents
135
self.repository.add_revision(self._new_revision_id, rev,
136
self.new_inventory, self._config)
137
self.repository.commit_write_group()
138
return self._new_revision_id
141
"""Abort the commit that is being built.
143
self.repository.abort_write_group()
145
def revision_tree(self):
146
"""Return the tree that was just committed.
148
After calling commit() this can be called to get a RevisionTree
149
representing the newly committed tree. This is preferred to
150
calling Repository.revision_tree() because that may require
151
deserializing the inventory, while we already have a copy in
154
return RevisionTree(self.repository, self.new_inventory,
155
self._new_revision_id)
157
def finish_inventory(self):
158
"""Tell the builder that the inventory is finished."""
159
if self.new_inventory.root is None:
160
symbol_versioning.warn('Root entry should be supplied to'
161
' record_entry_contents, as of bzr 0.10.',
162
DeprecationWarning, stacklevel=2)
163
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
164
self.new_inventory.revision_id = self._new_revision_id
165
self.inv_sha1 = self.repository.add_inventory(
166
self._new_revision_id,
171
def _gen_revision_id(self):
172
"""Return new revision-id."""
173
return generate_ids.gen_revision_id(self._config.username(),
176
def _generate_revision_if_needed(self):
177
"""Create a revision id if None was supplied.
179
If the repository can not support user-specified revision ids
180
they should override this function and raise CannotSetRevisionId
181
if _new_revision_id is not None.
183
:raises: CannotSetRevisionId
185
if self._new_revision_id is None:
186
self._new_revision_id = self._gen_revision_id()
187
self.random_revid = True
189
self.random_revid = False
191
def _check_root(self, ie, parent_invs, tree):
192
"""Helper for record_entry_contents.
194
:param ie: An entry being added.
195
:param parent_invs: The inventories of the parent revisions of the
197
:param tree: The tree that is being committed.
199
# In this revision format, root entries have no knit or weave When
200
# serializing out to disk and back in root.revision is always
202
ie.revision = self._new_revision_id
204
def _get_delta(self, ie, basis_inv, path):
205
"""Get a delta against the basis inventory for ie."""
206
if ie.file_id not in basis_inv:
208
return (None, path, ie.file_id, ie)
209
elif ie != basis_inv[ie.file_id]:
211
# TODO: avoid tis id2path call.
212
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
217
def record_entry_contents(self, ie, parent_invs, path, tree,
219
"""Record the content of ie from tree into the commit if needed.
221
Side effect: sets ie.revision when unchanged
223
:param ie: An inventory entry present in the commit.
224
:param parent_invs: The inventories of the parent revisions of the
226
:param path: The path the entry is at in the tree.
227
:param tree: The tree which contains this entry and should be used to
229
:param content_summary: Summary data from the tree about the paths
230
content - stat, length, exec, sha/link target. This is only
231
accessed when the entry has a revision of None - that is when it is
232
a candidate to commit.
233
:return: A tuple (change_delta, version_recorded). change_delta is
234
an inventory_delta change for this entry against the basis tree of
235
the commit, or None if no change occured against the basis tree.
236
version_recorded is True if a new version of the entry has been
237
recorded. For instance, committing a merge where a file was only
238
changed on the other side will return (delta, False).
240
if self.new_inventory.root is None:
241
if ie.parent_id is not None:
242
raise errors.RootMissing()
243
self._check_root(ie, parent_invs, tree)
244
if ie.revision is None:
245
kind = content_summary[0]
247
# ie is carried over from a prior commit
249
# XXX: repository specific check for nested tree support goes here - if
250
# the repo doesn't want nested trees we skip it ?
251
if (kind == 'tree-reference' and
252
not self.repository._format.supports_tree_reference):
253
# mismatch between commit builder logic and repository:
254
# this needs the entry creation pushed down into the builder.
255
raise NotImplementedError('Missing repository subtree support.')
256
# transitional assert only, will remove before release.
257
assert ie.kind == kind
258
self.new_inventory.add(ie)
260
# TODO: slow, take it out of the inner loop.
262
basis_inv = parent_invs[0]
264
basis_inv = Inventory(root_id=None)
266
# ie.revision is always None if the InventoryEntry is considered
267
# for committing. We may record the previous parents revision if the
268
# content is actually unchanged against a sole head.
269
if ie.revision is not None:
270
if self._versioned_root or path != '':
271
# not considered for commit
274
# repositories that do not version the root set the root's
275
# revision to the new commit even when no change occurs, and
276
# this masks when a change may have occurred against the basis,
277
# so calculate if one happened.
278
if ie.file_id not in basis_inv:
280
delta = (None, path, ie.file_id, ie)
282
basis_id = basis_inv[ie.file_id]
283
if basis_id.name != '':
285
delta = (basis_inv.id2path(ie.file_id), path,
290
# not considered for commit, OR, for non-rich-root
291
return delta, ie.revision == self._new_revision_id and (path != '' or
292
self._versioned_root)
294
# XXX: Friction: parent_candidates should return a list not a dict
295
# so that we don't have to walk the inventories again.
296
parent_candiate_entries = ie.parent_candidates(parent_invs)
297
head_set = self._repo_graph.heads(parent_candiate_entries.keys())
299
for inv in parent_invs:
300
if ie.file_id in inv:
301
old_rev = inv[ie.file_id].revision
302
if old_rev in head_set:
303
heads.append(inv[ie.file_id].revision)
304
head_set.remove(inv[ie.file_id].revision)
307
# now we check to see if we need to write a new record to the
309
# We write a new entry unless there is one head to the ancestors, and
310
# the kind-derived content is unchanged.
312
# Cheapest check first: no ancestors, or more the one head in the
313
# ancestors, we write a new node.
317
# There is a single head, look it up for comparison
318
parent_entry = parent_candiate_entries[heads[0]]
319
# if the non-content specific data has changed, we'll be writing a
321
if (parent_entry.parent_id != ie.parent_id or
322
parent_entry.name != ie.name):
324
# now we need to do content specific checks:
326
# if the kind changed the content obviously has
327
if kind != parent_entry.kind:
331
if (# if the file length changed we have to store:
332
parent_entry.text_size != content_summary[1] or
333
# if the exec bit has changed we have to store:
334
parent_entry.executable != content_summary[2]):
336
elif parent_entry.text_sha1 == content_summary[3]:
337
# all meta and content is unchanged (using a hash cache
338
# hit to check the sha)
339
ie.revision = parent_entry.revision
340
ie.text_size = parent_entry.text_size
341
ie.text_sha1 = parent_entry.text_sha1
342
ie.executable = parent_entry.executable
343
return self._get_delta(ie, basis_inv, path), False
345
# Either there is only a hash change(no hash cache entry,
346
# or same size content change), or there is no change on
348
# Provide the parent's hash to the store layer, so that the
349
# content is unchanged we will not store a new node.
350
nostore_sha = parent_entry.text_sha1
352
# We want to record a new node regardless of the presence or
353
# absence of a content change in the file.
355
ie.executable = content_summary[2]
356
lines = tree.get_file(ie.file_id, path).readlines()
358
ie.text_sha1, ie.text_size = self._add_text_to_weave(
359
ie.file_id, lines, heads, nostore_sha)
360
except errors.ExistingContent:
361
# Turns out that the file content was unchanged, and we were
362
# only going to store a new node if it was changed. Carry over
364
ie.revision = parent_entry.revision
365
ie.text_size = parent_entry.text_size
366
ie.text_sha1 = parent_entry.text_sha1
367
ie.executable = parent_entry.executable
368
return self._get_delta(ie, basis_inv, path), False
369
elif kind == 'directory':
371
# all data is meta here, nothing specific to directory, so
373
ie.revision = parent_entry.revision
374
return self._get_delta(ie, basis_inv, path), False
376
self._add_text_to_weave(ie.file_id, lines, heads, None)
377
elif kind == 'symlink':
378
current_link_target = content_summary[3]
380
# symlink target is not generic metadata, check if it has
382
if current_link_target != parent_entry.symlink_target:
385
# unchanged, carry over.
386
ie.revision = parent_entry.revision
387
ie.symlink_target = parent_entry.symlink_target
388
return self._get_delta(ie, basis_inv, path), False
389
ie.symlink_target = current_link_target
391
self._add_text_to_weave(ie.file_id, lines, heads, None)
392
elif kind == 'tree-reference':
394
if content_summary[3] != parent_entry.reference_revision:
397
# unchanged, carry over.
398
ie.reference_revision = parent_entry.reference_revision
399
ie.revision = parent_entry.revision
400
return self._get_delta(ie, basis_inv, path), False
401
ie.reference_revision = content_summary[3]
403
self._add_text_to_weave(ie.file_id, lines, heads, None)
405
raise NotImplementedError('unknown kind')
406
ie.revision = self._new_revision_id
407
return self._get_delta(ie, basis_inv, path), True
409
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
410
versionedfile = self.repository.weave_store.get_weave_or_empty(
411
file_id, self.repository.get_transaction())
412
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
413
# than add_lines, and allows committing when a parent is ghosted for
415
# Note: as we read the content directly from the tree, we know its not
416
# been turned into unicode or badly split - but a broken tree
417
# implementation could give us bad output from readlines() so this is
418
# not a guarantee of safety. What would be better is always checking
419
# the content during test suite execution. RBC 20070912
421
return versionedfile.add_lines_with_ghosts(
422
self._new_revision_id, parents, new_lines,
423
nostore_sha=nostore_sha, random_id=self.random_revid,
424
check_content=False)[0:2]
426
versionedfile.clear_cache()
429
class RootCommitBuilder(CommitBuilder):
430
"""This commitbuilder actually records the root id"""
432
# the root entry gets versioned properly by this builder.
433
_versioned_root = True
435
def _check_root(self, ie, parent_invs, tree):
436
"""Helper for record_entry_contents.
438
:param ie: An entry being added.
439
:param parent_invs: The inventories of the parent revisions of the
441
:param tree: The tree that is being committed.
445
######################################################################
448
class Repository(object):
449
"""Repository holding history for one or more branches.
451
The repository holds and retrieves historical information including
452
revisions and file history. It's normally accessed only by the Branch,
453
which views a particular line of development through that history.
455
The Repository builds on top of Stores and a Transport, which respectively
456
describe the disk data format and the way of accessing the (possibly
460
# What class to use for a CommitBuilder. Often its simpler to change this
461
# in a Repository class subclass rather than to override
462
# get_commit_builder.
463
_commit_builder_class = CommitBuilder
464
# The search regex used by xml based repositories to determine what things
465
# where changed in a single commit.
466
_file_ids_altered_regex = lazy_regex.lazy_compile(
467
r'file_id="(?P<file_id>[^"]+)"'
468
r'.* revision="(?P<revision_id>[^"]+)"'
471
def abort_write_group(self):
472
"""Commit the contents accrued within the current write group.
474
:seealso: start_write_group.
476
if self._write_group is not self.get_transaction():
477
# has an unlock or relock occured ?
478
raise errors.BzrError('mismatched lock context and write group.')
479
self._abort_write_group()
480
self._write_group = None
482
def _abort_write_group(self):
483
"""Template method for per-repository write group cleanup.
485
This is called during abort before the write group is considered to be
486
finished and should cleanup any internal state accrued during the write
487
group. There is no requirement that data handed to the repository be
488
*not* made available - this is not a rollback - but neither should any
489
attempt be made to ensure that data added is fully commited. Abort is
490
invoked when an error has occured so futher disk or network operations
491
may not be possible or may error and if possible should not be
496
def add_inventory(self, revision_id, inv, parents):
497
"""Add the inventory inv to the repository as revision_id.
499
:param parents: The revision ids of the parents that revision_id
500
is known to have and are in the repository already.
502
returns the sha1 of the serialized inventory.
504
_mod_revision.check_not_reserved_id(revision_id)
505
assert inv.revision_id is None or inv.revision_id == revision_id, \
506
"Mismatch between inventory revision" \
507
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
508
assert inv.root is not None
509
inv_lines = self._serialise_inventory_to_lines(inv)
510
inv_vf = self.get_inventory_weave()
511
return self._inventory_add_lines(inv_vf, revision_id, parents,
512
inv_lines, check_content=False)
514
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
516
"""Store lines in inv_vf and return the sha1 of the inventory."""
518
for parent in parents:
520
final_parents.append(parent)
521
return inv_vf.add_lines(revision_id, final_parents, lines,
522
check_content=check_content)[0]
525
def add_revision(self, revision_id, rev, inv=None, config=None):
526
"""Add rev to the revision store as revision_id.
528
:param revision_id: the revision id to use.
529
:param rev: The revision object.
530
:param inv: The inventory for the revision. if None, it will be looked
531
up in the inventory storer
532
:param config: If None no digital signature will be created.
533
If supplied its signature_needed method will be used
534
to determine if a signature should be made.
536
# TODO: jam 20070210 Shouldn't we check rev.revision_id and
538
_mod_revision.check_not_reserved_id(revision_id)
539
if config is not None and config.signature_needed():
541
inv = self.get_inventory(revision_id)
542
plaintext = Testament(rev, inv).as_short_text()
543
self.store_revision_signature(
544
gpg.GPGStrategy(config), plaintext, revision_id)
545
if not revision_id in self.get_inventory_weave():
547
raise errors.WeaveRevisionNotPresent(revision_id,
548
self.get_inventory_weave())
550
# yes, this is not suitable for adding with ghosts.
551
self.add_inventory(revision_id, inv, rev.parent_ids)
552
self._revision_store.add_revision(rev, self.get_transaction())
554
def _add_revision_text(self, revision_id, text):
555
revision = self._revision_store._serializer.read_revision_from_string(
557
self._revision_store._add_revision(revision, StringIO(text),
558
self.get_transaction())
560
def all_revision_ids(self):
561
"""Returns a list of all the revision ids in the repository.
563
This is deprecated because code should generally work on the graph
564
reachable from a particular revision, and ignore any other revisions
565
that might be present. There is no direct replacement method.
567
if 'evil' in debug.debug_flags:
568
mutter_callsite(2, "all_revision_ids is linear with history.")
569
return self._all_revision_ids()
571
def _all_revision_ids(self):
572
"""Returns a list of all the revision ids in the repository.
574
These are in as much topological order as the underlying store can
577
raise NotImplementedError(self._all_revision_ids)
579
def break_lock(self):
580
"""Break a lock if one is present from another instance.
582
Uses the ui factory to ask for confirmation if the lock may be from
585
self.control_files.break_lock()
588
def _eliminate_revisions_not_present(self, revision_ids):
589
"""Check every revision id in revision_ids to see if we have it.
591
Returns a set of the present revisions.
594
for id in revision_ids:
595
if self.has_revision(id):
600
def create(a_bzrdir):
601
"""Construct the current default format repository in a_bzrdir."""
602
return RepositoryFormat.get_default_format().initialize(a_bzrdir)
604
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
605
"""instantiate a Repository.
607
:param _format: The format of the repository on disk.
608
:param a_bzrdir: The BzrDir of the repository.
610
In the future we will have a single api for all stores for
611
getting file texts, inventories and revisions, then
612
this construct will accept instances of those things.
614
super(Repository, self).__init__()
615
self._format = _format
616
# the following are part of the public API for Repository:
617
self.bzrdir = a_bzrdir
618
self.control_files = control_files
619
self._revision_store = _revision_store
620
# backwards compatibility
621
self.weave_store = text_store
623
self._reconcile_does_inventory_gc = True
624
self._reconcile_fixes_text_parents = False
625
# not right yet - should be more semantically clear ?
627
self.control_store = control_store
628
self.control_weaves = control_store
629
# TODO: make sure to construct the right store classes, etc, depending
630
# on whether escaping is required.
631
self._warn_if_deprecated()
632
self._write_group = None
633
self.base = control_files._transport.base
636
return '%s(%r)' % (self.__class__.__name__,
639
def has_same_location(self, other):
640
"""Returns a boolean indicating if this repository is at the same
641
location as another repository.
643
This might return False even when two repository objects are accessing
644
the same physical repository via different URLs.
646
if self.__class__ is not other.__class__:
648
return (self.control_files._transport.base ==
649
other.control_files._transport.base)
651
def is_in_write_group(self):
652
"""Return True if there is an open write group.
654
:seealso: start_write_group.
656
return self._write_group is not None
659
return self.control_files.is_locked()
661
def is_write_locked(self):
662
"""Return True if this object is write locked."""
663
return self.is_locked() and self.control_files._lock_mode == 'w'
665
def lock_write(self, token=None):
666
"""Lock this repository for writing.
668
This causes caching within the repository obejct to start accumlating
669
data during reads, and allows a 'write_group' to be obtained. Write
670
groups must be used for actual data insertion.
672
:param token: if this is already locked, then lock_write will fail
673
unless the token matches the existing lock.
674
:returns: a token if this instance supports tokens, otherwise None.
675
:raises TokenLockingNotSupported: when a token is given but this
676
instance doesn't support using token locks.
677
:raises MismatchedToken: if the specified token doesn't match the token
678
of the existing lock.
679
:seealso: start_write_group.
681
A token should be passed in if you know that you have locked the object
682
some other way, and need to synchronise this object's state with that
685
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
687
result = self.control_files.lock_write(token=token)
692
self.control_files.lock_read()
695
def get_physical_lock_status(self):
696
return self.control_files.get_physical_lock_status()
698
def leave_lock_in_place(self):
699
"""Tell this repository not to release the physical lock when this
702
If lock_write doesn't return a token, then this method is not supported.
704
self.control_files.leave_in_place()
706
def dont_leave_lock_in_place(self):
707
"""Tell this repository to release the physical lock when this
708
object is unlocked, even if it didn't originally acquire it.
710
If lock_write doesn't return a token, then this method is not supported.
712
self.control_files.dont_leave_in_place()
715
def gather_stats(self, revid=None, committers=None):
716
"""Gather statistics from a revision id.
718
:param revid: The revision id to gather statistics from, if None, then
719
no revision specific statistics are gathered.
720
:param committers: Optional parameter controlling whether to grab
721
a count of committers from the revision specific statistics.
722
:return: A dictionary of statistics. Currently this contains:
723
committers: The number of committers if requested.
724
firstrev: A tuple with timestamp, timezone for the penultimate left
725
most ancestor of revid, if revid is not the NULL_REVISION.
726
latestrev: A tuple with timestamp, timezone for revid, if revid is
727
not the NULL_REVISION.
728
revisions: The total revision count in the repository.
729
size: An estimate disk size of the repository in bytes.
732
if revid and committers:
733
result['committers'] = 0
734
if revid and revid != _mod_revision.NULL_REVISION:
736
all_committers = set()
737
revisions = self.get_ancestry(revid)
738
# pop the leading None
740
first_revision = None
742
# ignore the revisions in the middle - just grab first and last
743
revisions = revisions[0], revisions[-1]
744
for revision in self.get_revisions(revisions):
745
if not first_revision:
746
first_revision = revision
748
all_committers.add(revision.committer)
749
last_revision = revision
751
result['committers'] = len(all_committers)
752
result['firstrev'] = (first_revision.timestamp,
753
first_revision.timezone)
754
result['latestrev'] = (last_revision.timestamp,
755
last_revision.timezone)
757
# now gather global repository information
758
if self.bzrdir.root_transport.listable():
759
c, t = self._revision_store.total_size(self.get_transaction())
760
result['revisions'] = c
764
def get_data_stream(self, revision_ids):
765
raise NotImplementedError(self.get_data_stream)
767
def insert_data_stream(self, stream):
768
for item_key, bytes in stream:
769
if item_key[0] == 'file':
770
(file_id,) = item_key[1:]
771
knit = self.weave_store.get_weave_or_empty(
772
file_id, self.get_transaction())
773
elif item_key == ('inventory',):
774
knit = self.get_inventory_weave()
775
elif item_key == ('revisions',):
776
knit = self._revision_store.get_revision_file(
777
self.get_transaction())
778
elif item_key == ('signatures',):
779
knit = self._revision_store.get_signature_file(
780
self.get_transaction())
782
raise RepositoryDataStreamError(
783
"Unrecognised data stream key '%s'" % (item_key,))
784
decoded_list = bencode.bdecode(bytes)
785
format = decoded_list.pop(0)
788
for version, options, parents, some_bytes in decoded_list:
789
data_list.append((version, options, len(some_bytes), parents))
790
knit_bytes += some_bytes
791
knit.insert_data_stream(
792
(format, data_list, StringIO(knit_bytes).read))
795
def missing_revision_ids(self, other, revision_id=None):
796
"""Return the revision ids that other has that this does not.
798
These are returned in topological order.
800
revision_id: only return revision ids included by revision_id.
802
return InterRepository.get(other, self).missing_revision_ids(revision_id)
806
"""Open the repository rooted at base.
808
For instance, if the repository is at URL/.bzr/repository,
809
Repository.open(URL) -> a Repository instance.
811
control = bzrdir.BzrDir.open(base)
812
return control.open_repository()
814
def copy_content_into(self, destination, revision_id=None):
815
"""Make a complete copy of the content in self into destination.
817
This is a destructive operation! Do not use it on existing
820
return InterRepository.get(self, destination).copy_content(revision_id)
822
def commit_write_group(self):
823
"""Commit the contents accrued within the current write group.
825
:seealso: start_write_group.
827
if self._write_group is not self.get_transaction():
828
# has an unlock or relock occured ?
829
raise errors.BzrError('mismatched lock context %r and '
831
(self.get_transaction(), self._write_group))
832
self._commit_write_group()
833
self._write_group = None
835
def _commit_write_group(self):
836
"""Template method for per-repository write group cleanup.
838
This is called before the write group is considered to be
839
finished and should ensure that all data handed to the repository
840
for writing during the write group is safely committed (to the
841
extent possible considering file system caching etc).
844
def fetch(self, source, revision_id=None, pb=None):
845
"""Fetch the content required to construct revision_id from source.
847
If revision_id is None all content is copied.
849
# fast path same-url fetch operations
850
if self.has_same_location(source):
851
# check that last_revision is in 'from' and then return a
853
if (revision_id is not None and
854
not _mod_revision.is_null(revision_id)):
855
self.get_revision(revision_id)
857
inter = InterRepository.get(source, self)
859
return inter.fetch(revision_id=revision_id, pb=pb)
860
except NotImplementedError:
861
raise errors.IncompatibleRepositories(source, self)
863
def create_bundle(self, target, base, fileobj, format=None):
864
return serializer.write_bundle(self, target, base, fileobj, format)
866
def get_commit_builder(self, branch, parents, config, timestamp=None,
867
timezone=None, committer=None, revprops=None,
869
"""Obtain a CommitBuilder for this repository.
871
:param branch: Branch to commit to.
872
:param parents: Revision ids of the parents of the new revision.
873
:param config: Configuration to use.
874
:param timestamp: Optional timestamp recorded for commit.
875
:param timezone: Optional timezone for timestamp.
876
:param committer: Optional committer to set for commit.
877
:param revprops: Optional dictionary of revision properties.
878
:param revision_id: Optional revision id.
880
result = self._commit_builder_class(self, parents, config,
881
timestamp, timezone, committer, revprops, revision_id)
882
self.start_write_group()
886
if (self.control_files._lock_count == 1 and
887
self.control_files._lock_mode == 'w'):
888
if self._write_group is not None:
889
raise errors.BzrError(
890
'Must end write groups before releasing write locks.')
891
self.control_files.unlock()
894
def clone(self, a_bzrdir, revision_id=None):
895
"""Clone this repository into a_bzrdir using the current format.
897
Currently no check is made that the format of this repository and
898
the bzrdir format are compatible. FIXME RBC 20060201.
900
:return: The newly created destination repository.
902
# TODO: deprecate after 0.16; cloning this with all its settings is
903
# probably not very useful -- mbp 20070423
904
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
905
self.copy_content_into(dest_repo, revision_id)
908
def start_write_group(self):
909
"""Start a write group in the repository.
911
Write groups are used by repositories which do not have a 1:1 mapping
912
between file ids and backend store to manage the insertion of data from
913
both fetch and commit operations.
915
A write lock is required around the start_write_group/commit_write_group
916
for the support of lock-requiring repository formats.
918
One can only insert data into a repository inside a write group.
922
if not self.is_write_locked():
923
raise errors.NotWriteLocked(self)
924
if self._write_group:
925
raise errors.BzrError('already in a write group')
926
self._start_write_group()
927
# so we can detect unlock/relock - the write group is now entered.
928
self._write_group = self.get_transaction()
930
def _start_write_group(self):
931
"""Template method for per-repository write group startup.
933
This is called before the write group is considered to be
938
def sprout(self, to_bzrdir, revision_id=None):
939
"""Create a descendent repository for new development.
941
Unlike clone, this does not copy the settings of the repository.
943
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
944
dest_repo.fetch(self, revision_id=revision_id)
947
def _create_sprouting_repo(self, a_bzrdir, shared):
948
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
949
# use target default format.
950
dest_repo = a_bzrdir.create_repository()
952
# Most control formats need the repository to be specifically
953
# created, but on some old all-in-one formats it's not needed
955
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
956
except errors.UninitializableFormat:
957
dest_repo = a_bzrdir.open_repository()
961
def has_revision(self, revision_id):
962
"""True if this repository has a copy of the revision."""
963
if 'evil' in debug.debug_flags:
964
mutter_callsite(3, "has_revision is a LBYL symptom.")
965
return self._revision_store.has_revision_id(revision_id,
966
self.get_transaction())
969
def get_revision(self, revision_id):
970
"""Return the Revision object for a named revision."""
971
return self.get_revisions([revision_id])[0]
974
def get_revision_reconcile(self, revision_id):
975
"""'reconcile' helper routine that allows access to a revision always.
977
This variant of get_revision does not cross check the weave graph
978
against the revision one as get_revision does: but it should only
979
be used by reconcile, or reconcile-alike commands that are correcting
980
or testing the revision graph.
982
return self._get_revisions([revision_id])[0]
985
def get_revisions(self, revision_ids):
986
"""Get many revisions at once."""
987
return self._get_revisions(revision_ids)
990
def _get_revisions(self, revision_ids):
991
"""Core work logic to get many revisions without sanity checks."""
992
for rev_id in revision_ids:
993
if not rev_id or not isinstance(rev_id, basestring):
994
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
995
revs = self._revision_store.get_revisions(revision_ids,
996
self.get_transaction())
998
assert not isinstance(rev.revision_id, unicode)
999
for parent_id in rev.parent_ids:
1000
assert not isinstance(parent_id, unicode)
1004
def get_revision_xml(self, revision_id):
1005
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1006
# would have already do it.
1007
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1008
rev = self.get_revision(revision_id)
1009
rev_tmp = StringIO()
1010
# the current serializer..
1011
self._revision_store._serializer.write_revision(rev, rev_tmp)
1013
return rev_tmp.getvalue()
1016
def get_deltas_for_revisions(self, revisions):
1017
"""Produce a generator of revision deltas.
1019
Note that the input is a sequence of REVISIONS, not revision_ids.
1020
Trees will be held in memory until the generator exits.
1021
Each delta is relative to the revision's lefthand predecessor.
1023
required_trees = set()
1024
for revision in revisions:
1025
required_trees.add(revision.revision_id)
1026
required_trees.update(revision.parent_ids[:1])
1027
trees = dict((t.get_revision_id(), t) for
1028
t in self.revision_trees(required_trees))
1029
for revision in revisions:
1030
if not revision.parent_ids:
1031
old_tree = self.revision_tree(None)
1033
old_tree = trees[revision.parent_ids[0]]
1034
yield trees[revision.revision_id].changes_from(old_tree)
1037
def get_revision_delta(self, revision_id):
1038
"""Return the delta for one revision.
1040
The delta is relative to the left-hand predecessor of the
1043
r = self.get_revision(revision_id)
1044
return list(self.get_deltas_for_revisions([r]))[0]
1047
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1048
signature = gpg_strategy.sign(plaintext)
1049
self._revision_store.add_revision_signature_text(revision_id,
1051
self.get_transaction())
1053
def fileids_altered_by_revision_ids(self, revision_ids):
1054
"""Find the file ids and versions affected by revisions.
1056
:param revisions: an iterable containing revision ids.
1057
:return: a dictionary mapping altered file-ids to an iterable of
1058
revision_ids. Each altered file-ids has the exact revision_ids that
1059
altered it listed explicitly.
1061
assert self._serializer.support_altered_by_hack, \
1062
("fileids_altered_by_revision_ids only supported for branches "
1063
"which store inventory as unnested xml, not on %r" % self)
1064
selected_revision_ids = set(revision_ids)
1065
w = self.get_inventory_weave()
1068
# this code needs to read every new line in every inventory for the
1069
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1070
# not present in one of those inventories is unnecessary but not
1071
# harmful because we are filtering by the revision id marker in the
1072
# inventory lines : we only select file ids altered in one of those
1073
# revisions. We don't need to see all lines in the inventory because
1074
# only those added in an inventory in rev X can contain a revision=X
1076
unescape_revid_cache = {}
1077
unescape_fileid_cache = {}
1079
# jam 20061218 In a big fetch, this handles hundreds of thousands
1080
# of lines, so it has had a lot of inlining and optimizing done.
1081
# Sorry that it is a little bit messy.
1082
# Move several functions to be local variables, since this is a long
1084
search = self._file_ids_altered_regex.search
1085
unescape = _unescape_xml
1086
setdefault = result.setdefault
1087
pb = ui.ui_factory.nested_progress_bar()
1089
for line in w.iter_lines_added_or_present_in_versions(
1090
selected_revision_ids, pb=pb):
1091
match = search(line)
1094
# One call to match.group() returning multiple items is quite a
1095
# bit faster than 2 calls to match.group() each returning 1
1096
file_id, revision_id = match.group('file_id', 'revision_id')
1098
# Inlining the cache lookups helps a lot when you make 170,000
1099
# lines and 350k ids, versus 8.4 unique ids.
1100
# Using a cache helps in 2 ways:
1101
# 1) Avoids unnecessary decoding calls
1102
# 2) Re-uses cached strings, which helps in future set and
1104
# (2) is enough that removing encoding entirely along with
1105
# the cache (so we are using plain strings) results in no
1106
# performance improvement.
1108
revision_id = unescape_revid_cache[revision_id]
1110
unescaped = unescape(revision_id)
1111
unescape_revid_cache[revision_id] = unescaped
1112
revision_id = unescaped
1114
if revision_id in selected_revision_ids:
1116
file_id = unescape_fileid_cache[file_id]
1118
unescaped = unescape(file_id)
1119
unescape_fileid_cache[file_id] = unescaped
1121
setdefault(file_id, set()).add(revision_id)
1126
def iter_files_bytes(self, desired_files):
1127
"""Iterate through file versions.
1129
Files will not necessarily be returned in the order they occur in
1130
desired_files. No specific order is guaranteed.
1132
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1133
value supplied by the caller as part of desired_files. It should
1134
uniquely identify the file version in the caller's context. (Examples:
1135
an index number or a TreeTransform trans_id.)
1137
bytes_iterator is an iterable of bytestrings for the file. The
1138
kind of iterable and length of the bytestrings are unspecified, but for
1139
this implementation, it is a list of lines produced by
1140
VersionedFile.get_lines().
1142
:param desired_files: a list of (file_id, revision_id, identifier)
1145
transaction = self.get_transaction()
1146
for file_id, revision_id, callable_data in desired_files:
1148
weave = self.weave_store.get_weave(file_id, transaction)
1149
except errors.NoSuchFile:
1150
raise errors.NoSuchIdInRepository(self, file_id)
1151
yield callable_data, weave.get_lines(revision_id)
1153
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1154
"""Get an iterable listing the keys of all the data introduced by a set
1157
The keys will be ordered so that the corresponding items can be safely
1158
fetched and inserted in that order.
1160
:returns: An iterable producing tuples of (knit-kind, file-id,
1161
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1162
'revisions'. file-id is None unless knit-kind is 'file'.
1164
# XXX: it's a bit weird to control the inventory weave caching in this
1165
# generator. Ideally the caching would be done in fetch.py I think. Or
1166
# maybe this generator should explicitly have the contract that it
1167
# should not be iterated until the previously yielded item has been
1170
inv_w = self.get_inventory_weave()
1171
inv_w.enable_cache()
1173
# file ids that changed
1174
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1176
num_file_ids = len(file_ids)
1177
for file_id, altered_versions in file_ids.iteritems():
1178
if _files_pb is not None:
1179
_files_pb.update("fetch texts", count, num_file_ids)
1181
yield ("file", file_id, altered_versions)
1182
# We're done with the files_pb. Note that it finished by the caller,
1183
# just as it was created by the caller.
1187
yield ("inventory", None, revision_ids)
1191
revisions_with_signatures = set()
1192
for rev_id in revision_ids:
1194
self.get_signature_text(rev_id)
1195
except errors.NoSuchRevision:
1199
revisions_with_signatures.add(rev_id)
1201
yield ("signatures", None, revisions_with_signatures)
1204
yield ("revisions", None, revision_ids)
1207
def get_inventory_weave(self):
1208
return self.control_weaves.get_weave('inventory',
1209
self.get_transaction())
1212
def get_inventory(self, revision_id):
1213
"""Get Inventory object by hash."""
1214
return self.deserialise_inventory(
1215
revision_id, self.get_inventory_xml(revision_id))
1217
def deserialise_inventory(self, revision_id, xml):
1218
"""Transform the xml into an inventory object.
1220
:param revision_id: The expected revision id of the inventory.
1221
:param xml: A serialised inventory.
1223
return self._serializer.read_inventory_from_string(xml, revision_id)
1225
def serialise_inventory(self, inv):
1226
return self._serializer.write_inventory_to_string(inv)
1228
def _serialise_inventory_to_lines(self, inv):
1229
return self._serializer.write_inventory_to_lines(inv)
1231
def get_serializer_format(self):
1232
return self._serializer.format_num
1235
def get_inventory_xml(self, revision_id):
1236
"""Get inventory XML as a file object."""
1238
assert isinstance(revision_id, str), type(revision_id)
1239
iw = self.get_inventory_weave()
1240
return iw.get_text(revision_id)
1242
raise errors.HistoryMissing(self, 'inventory', revision_id)
1245
def get_inventory_sha1(self, revision_id):
1246
"""Return the sha1 hash of the inventory entry
1248
return self.get_revision(revision_id).inventory_sha1
1251
def get_revision_graph(self, revision_id=None):
1252
"""Return a dictionary containing the revision graph.
1254
NB: This method should not be used as it accesses the entire graph all
1255
at once, which is much more data than most operations should require.
1257
:param revision_id: The revision_id to get a graph from. If None, then
1258
the entire revision graph is returned. This is a deprecated mode of
1259
operation and will be removed in the future.
1260
:return: a dictionary of revision_id->revision_parents_list.
1262
raise NotImplementedError(self.get_revision_graph)
1265
def get_revision_graph_with_ghosts(self, revision_ids=None):
1266
"""Return a graph of the revisions with ghosts marked as applicable.
1268
:param revision_ids: an iterable of revisions to graph or None for all.
1269
:return: a Graph object with the graph reachable from revision_ids.
1271
if 'evil' in debug.debug_flags:
1273
"get_revision_graph_with_ghosts scales with size of history.")
1274
result = deprecated_graph.Graph()
1275
if not revision_ids:
1276
pending = set(self.all_revision_ids())
1279
pending = set(revision_ids)
1280
# special case NULL_REVISION
1281
if _mod_revision.NULL_REVISION in pending:
1282
pending.remove(_mod_revision.NULL_REVISION)
1283
required = set(pending)
1286
revision_id = pending.pop()
1288
rev = self.get_revision(revision_id)
1289
except errors.NoSuchRevision:
1290
if revision_id in required:
1293
result.add_ghost(revision_id)
1295
for parent_id in rev.parent_ids:
1296
# is this queued or done ?
1297
if (parent_id not in pending and
1298
parent_id not in done):
1300
pending.add(parent_id)
1301
result.add_node(revision_id, rev.parent_ids)
1302
done.add(revision_id)
1305
def _get_history_vf(self):
1306
"""Get a versionedfile whose history graph reflects all revisions.
1308
For weave repositories, this is the inventory weave.
1310
return self.get_inventory_weave()
1312
def iter_reverse_revision_history(self, revision_id):
1313
"""Iterate backwards through revision ids in the lefthand history
1315
:param revision_id: The revision id to start with. All its lefthand
1316
ancestors will be traversed.
1318
if revision_id in (None, _mod_revision.NULL_REVISION):
1320
next_id = revision_id
1321
versionedfile = self._get_history_vf()
1324
parents = versionedfile.get_parents(next_id)
1325
if len(parents) == 0:
1328
next_id = parents[0]
1331
def get_revision_inventory(self, revision_id):
1332
"""Return inventory of a past revision."""
1333
# TODO: Unify this with get_inventory()
1334
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1335
# must be the same as its revision, so this is trivial.
1336
if revision_id is None:
1337
# This does not make sense: if there is no revision,
1338
# then it is the current tree inventory surely ?!
1339
# and thus get_root_id() is something that looks at the last
1340
# commit on the branch, and the get_root_id is an inventory check.
1341
raise NotImplementedError
1342
# return Inventory(self.get_root_id())
1344
return self.get_inventory(revision_id)
1347
def is_shared(self):
1348
"""Return True if this repository is flagged as a shared repository."""
1349
raise NotImplementedError(self.is_shared)
1352
def reconcile(self, other=None, thorough=False):
1353
"""Reconcile this repository."""
1354
from bzrlib.reconcile import RepoReconciler
1355
reconciler = RepoReconciler(self, thorough=thorough)
1356
reconciler.reconcile()
1359
def _refresh_data(self):
1360
"""Helper called from lock_* to ensure coherency with disk.
1362
The default implementation does nothing; it is however possible
1363
for repositories to maintain loaded indices across multiple locks
1364
by checking inside their implementation of this method to see
1365
whether their indices are still valid. This depends of course on
1366
the disk format being validatable in this manner.
1370
def revision_tree(self, revision_id):
1371
"""Return Tree for a revision on this branch.
1373
`revision_id` may be None for the empty tree revision.
1375
# TODO: refactor this to use an existing revision object
1376
# so we don't need to read it in twice.
1377
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1378
return RevisionTree(self, Inventory(root_id=None),
1379
_mod_revision.NULL_REVISION)
1381
inv = self.get_revision_inventory(revision_id)
1382
return RevisionTree(self, inv, revision_id)
1385
def revision_trees(self, revision_ids):
1386
"""Return Tree for a revision on this branch.
1388
`revision_id` may not be None or 'null:'"""
1389
assert None not in revision_ids
1390
assert _mod_revision.NULL_REVISION not in revision_ids
1391
texts = self.get_inventory_weave().get_texts(revision_ids)
1392
for text, revision_id in zip(texts, revision_ids):
1393
inv = self.deserialise_inventory(revision_id, text)
1394
yield RevisionTree(self, inv, revision_id)
1397
def get_ancestry(self, revision_id, topo_sorted=True):
1398
"""Return a list of revision-ids integrated by a revision.
1400
The first element of the list is always None, indicating the origin
1401
revision. This might change when we have history horizons, or
1402
perhaps we should have a new API.
1404
This is topologically sorted.
1406
if _mod_revision.is_null(revision_id):
1408
if not self.has_revision(revision_id):
1409
raise errors.NoSuchRevision(self, revision_id)
1410
w = self.get_inventory_weave()
1411
candidates = w.get_ancestry(revision_id, topo_sorted)
1412
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1415
"""Compress the data within the repository.
1417
This operation only makes sense for some repository types. For other
1418
types it should be a no-op that just returns.
1420
This stub method does not require a lock, but subclasses should use
1421
@needs_write_lock as this is a long running call its reasonable to
1422
implicitly lock for the user.
1426
def print_file(self, file, revision_id):
1427
"""Print `file` to stdout.
1429
FIXME RBC 20060125 as John Meinel points out this is a bad api
1430
- it writes to stdout, it assumes that that is valid etc. Fix
1431
by creating a new more flexible convenience function.
1433
tree = self.revision_tree(revision_id)
1434
# use inventory as it was in that revision
1435
file_id = tree.inventory.path2id(file)
1437
# TODO: jam 20060427 Write a test for this code path
1438
# it had a bug in it, and was raising the wrong
1440
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1441
tree.print_file(file_id)
1443
def get_transaction(self):
1444
return self.control_files.get_transaction()
1446
def revision_parents(self, revision_id):
1447
return self.get_inventory_weave().parent_names(revision_id)
1449
def get_parents(self, revision_ids):
1450
"""See StackedParentsProvider.get_parents"""
1452
for revision_id in revision_ids:
1453
if revision_id == _mod_revision.NULL_REVISION:
1457
parents = self.get_revision(revision_id).parent_ids
1458
except errors.NoSuchRevision:
1461
if len(parents) == 0:
1462
parents = [_mod_revision.NULL_REVISION]
1463
parents_list.append(parents)
1466
def _make_parents_provider(self):
1469
def get_graph(self, other_repository=None):
1470
"""Return the graph walker for this repository format"""
1471
parents_provider = self._make_parents_provider()
1472
if (other_repository is not None and
1473
other_repository.bzrdir.transport.base !=
1474
self.bzrdir.transport.base):
1475
parents_provider = graph._StackedParentsProvider(
1476
[parents_provider, other_repository._make_parents_provider()])
1477
return graph.Graph(parents_provider)
1479
def get_versioned_file_checker(self, revisions, revision_versions_cache):
1480
return VersionedFileChecker(revisions, revision_versions_cache, self)
1483
def set_make_working_trees(self, new_value):
1484
"""Set the policy flag for making working trees when creating branches.
1486
This only applies to branches that use this repository.
1488
The default is 'True'.
1489
:param new_value: True to restore the default, False to disable making
1492
raise NotImplementedError(self.set_make_working_trees)
1494
def make_working_trees(self):
1495
"""Returns the policy for making working trees on new branches."""
1496
raise NotImplementedError(self.make_working_trees)
1499
def sign_revision(self, revision_id, gpg_strategy):
1500
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1501
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1504
def has_signature_for_revision_id(self, revision_id):
1505
"""Query for a revision signature for revision_id in the repository."""
1506
return self._revision_store.has_signature(revision_id,
1507
self.get_transaction())
1510
def get_signature_text(self, revision_id):
1511
"""Return the text for a signature."""
1512
return self._revision_store.get_signature_text(revision_id,
1513
self.get_transaction())
1516
def check(self, revision_ids=None):
1517
"""Check consistency of all history of given revision_ids.
1519
Different repository implementations should override _check().
1521
:param revision_ids: A non-empty list of revision_ids whose ancestry
1522
will be checked. Typically the last revision_id of a branch.
1524
return self._check(revision_ids)
1526
def _check(self, revision_ids):
1527
result = check.Check(self)
1531
def _warn_if_deprecated(self):
1532
global _deprecation_warning_done
1533
if _deprecation_warning_done:
1535
_deprecation_warning_done = True
1536
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1537
% (self._format, self.bzrdir.transport.base))
1539
def supports_rich_root(self):
1540
return self._format.rich_root_data
1542
def _check_ascii_revisionid(self, revision_id, method):
1543
"""Private helper for ascii-only repositories."""
1544
# weave repositories refuse to store revisionids that are non-ascii.
1545
if revision_id is not None:
1546
# weaves require ascii revision ids.
1547
if isinstance(revision_id, unicode):
1549
revision_id.encode('ascii')
1550
except UnicodeEncodeError:
1551
raise errors.NonAsciiRevisionId(method, self)
1554
revision_id.decode('ascii')
1555
except UnicodeDecodeError:
1556
raise errors.NonAsciiRevisionId(method, self)
1558
def revision_graph_can_have_wrong_parents(self):
1559
"""Is it possible for this repository to have a revision graph with
1562
If True, then this repository must also implement
1563
_find_inconsistent_revision_parents so that check and reconcile can
1564
check for inconsistencies before proceeding with other checks that may
1565
depend on the revision index being consistent.
1567
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1569
# remove these delegates a while after bzr 0.15
1570
def __make_delegated(name, from_module):
1571
def _deprecated_repository_forwarder():
1572
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1573
% (name, from_module),
1576
m = __import__(from_module, globals(), locals(), [name])
1578
return getattr(m, name)
1579
except AttributeError:
1580
raise AttributeError('module %s has no name %s'
1582
globals()[name] = _deprecated_repository_forwarder
1585
'AllInOneRepository',
1586
'WeaveMetaDirRepository',
1587
'PreSplitOutRepositoryFormat',
1588
'RepositoryFormat4',
1589
'RepositoryFormat5',
1590
'RepositoryFormat6',
1591
'RepositoryFormat7',
1593
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1597
'RepositoryFormatKnit',
1598
'RepositoryFormatKnit1',
1600
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1603
def install_revision(repository, rev, revision_tree):
1604
"""Install all revision data into a repository."""
1605
present_parents = []
1607
for p_id in rev.parent_ids:
1608
if repository.has_revision(p_id):
1609
present_parents.append(p_id)
1610
parent_trees[p_id] = repository.revision_tree(p_id)
1612
parent_trees[p_id] = repository.revision_tree(None)
1614
inv = revision_tree.inventory
1615
entries = inv.iter_entries()
1616
# backwards compatibility hack: skip the root id.
1617
if not repository.supports_rich_root():
1618
path, root = entries.next()
1619
if root.revision != rev.revision_id:
1620
raise errors.IncompatibleRevision(repr(repository))
1621
# Add the texts that are not already present
1622
for path, ie in entries:
1623
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1624
repository.get_transaction())
1625
if ie.revision not in w:
1627
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1628
# with InventoryEntry.find_previous_heads(). if it is, then there
1629
# is a latent bug here where the parents may have ancestors of each
1631
for revision, tree in parent_trees.iteritems():
1632
if ie.file_id not in tree:
1634
parent_id = tree.inventory[ie.file_id].revision
1635
if parent_id in text_parents:
1637
text_parents.append(parent_id)
1639
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1640
repository.get_transaction())
1641
lines = revision_tree.get_file(ie.file_id).readlines()
1642
vfile.add_lines(rev.revision_id, text_parents, lines)
1644
# install the inventory
1645
repository.add_inventory(rev.revision_id, inv, present_parents)
1646
except errors.RevisionAlreadyPresent:
1648
repository.add_revision(rev.revision_id, rev, inv)
1651
class MetaDirRepository(Repository):
1652
"""Repositories in the new meta-dir layout."""
1654
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1655
super(MetaDirRepository, self).__init__(_format,
1661
dir_mode = self.control_files._dir_mode
1662
file_mode = self.control_files._file_mode
1665
def is_shared(self):
1666
"""Return True if this repository is flagged as a shared repository."""
1667
return self.control_files._transport.has('shared-storage')
1670
def set_make_working_trees(self, new_value):
1671
"""Set the policy flag for making working trees when creating branches.
1673
This only applies to branches that use this repository.
1675
The default is 'True'.
1676
:param new_value: True to restore the default, False to disable making
1681
self.control_files._transport.delete('no-working-trees')
1682
except errors.NoSuchFile:
1685
self.control_files.put_utf8('no-working-trees', '')
1687
def make_working_trees(self):
1688
"""Returns the policy for making working trees on new branches."""
1689
return not self.control_files._transport.has('no-working-trees')
1692
class RepositoryFormatRegistry(registry.Registry):
1693
"""Registry of RepositoryFormats."""
1695
def get(self, format_string):
1696
r = registry.Registry.get(self, format_string)
1702
format_registry = RepositoryFormatRegistry()
1703
"""Registry of formats, indexed by their identifying format string.
1705
This can contain either format instances themselves, or classes/factories that
1706
can be called to obtain one.
1710
#####################################################################
1711
# Repository Formats
1713
class RepositoryFormat(object):
1714
"""A repository format.
1716
Formats provide three things:
1717
* An initialization routine to construct repository data on disk.
1718
* a format string which is used when the BzrDir supports versioned
1720
* an open routine which returns a Repository instance.
1722
There is one and only one Format subclass for each on-disk format. But
1723
there can be one Repository subclass that is used for several different
1724
formats. The _format attribute on a Repository instance can be used to
1725
determine the disk format.
1727
Formats are placed in an dict by their format string for reference
1728
during opening. These should be subclasses of RepositoryFormat
1731
Once a format is deprecated, just deprecate the initialize and open
1732
methods on the format class. Do not deprecate the object, as the
1733
object will be created every system load.
1735
Common instance attributes:
1736
_matchingbzrdir - the bzrdir format that the repository format was
1737
originally written to work with. This can be used if manually
1738
constructing a bzrdir and repository, or more commonly for test suite
1743
return "<%s>" % self.__class__.__name__
1745
def __eq__(self, other):
1746
# format objects are generally stateless
1747
return isinstance(other, self.__class__)
1749
def __ne__(self, other):
1750
return not self == other
1753
def find_format(klass, a_bzrdir):
1754
"""Return the format for the repository object in a_bzrdir.
1756
This is used by bzr native formats that have a "format" file in
1757
the repository. Other methods may be used by different types of
1761
transport = a_bzrdir.get_repository_transport(None)
1762
format_string = transport.get("format").read()
1763
return format_registry.get(format_string)
1764
except errors.NoSuchFile:
1765
raise errors.NoRepositoryPresent(a_bzrdir)
1767
raise errors.UnknownFormatError(format=format_string)
1770
def register_format(klass, format):
1771
format_registry.register(format.get_format_string(), format)
1774
def unregister_format(klass, format):
1775
format_registry.remove(format.get_format_string())
1778
def get_default_format(klass):
1779
"""Return the current default format."""
1780
from bzrlib import bzrdir
1781
return bzrdir.format_registry.make_bzrdir('default').repository_format
1783
def _get_control_store(self, repo_transport, control_files):
1784
"""Return the control store for this repository."""
1785
raise NotImplementedError(self._get_control_store)
1787
def get_format_string(self):
1788
"""Return the ASCII format string that identifies this format.
1790
Note that in pre format ?? repositories the format string is
1791
not permitted nor written to disk.
1793
raise NotImplementedError(self.get_format_string)
1795
def get_format_description(self):
1796
"""Return the short description for this format."""
1797
raise NotImplementedError(self.get_format_description)
1799
def _get_revision_store(self, repo_transport, control_files):
1800
"""Return the revision store object for this a_bzrdir."""
1801
raise NotImplementedError(self._get_revision_store)
1803
def _get_text_rev_store(self,
1810
"""Common logic for getting a revision store for a repository.
1812
see self._get_revision_store for the subclass-overridable method to
1813
get the store for a repository.
1815
from bzrlib.store.revision.text import TextRevisionStore
1816
dir_mode = control_files._dir_mode
1817
file_mode = control_files._file_mode
1818
text_store = TextStore(transport.clone(name),
1820
compressed=compressed,
1822
file_mode=file_mode)
1823
_revision_store = TextRevisionStore(text_store, serializer)
1824
return _revision_store
1826
# TODO: this shouldn't be in the base class, it's specific to things that
1827
# use weaves or knits -- mbp 20070207
1828
def _get_versioned_file_store(self,
1833
versionedfile_class=None,
1834
versionedfile_kwargs={},
1836
if versionedfile_class is None:
1837
versionedfile_class = self._versionedfile_class
1838
weave_transport = control_files._transport.clone(name)
1839
dir_mode = control_files._dir_mode
1840
file_mode = control_files._file_mode
1841
return VersionedFileStore(weave_transport, prefixed=prefixed,
1843
file_mode=file_mode,
1844
versionedfile_class=versionedfile_class,
1845
versionedfile_kwargs=versionedfile_kwargs,
1848
def initialize(self, a_bzrdir, shared=False):
1849
"""Initialize a repository of this format in a_bzrdir.
1851
:param a_bzrdir: The bzrdir to put the new repository in it.
1852
:param shared: The repository should be initialized as a sharable one.
1853
:returns: The new repository object.
1855
This may raise UninitializableFormat if shared repository are not
1856
compatible the a_bzrdir.
1858
raise NotImplementedError(self.initialize)
1860
def is_supported(self):
1861
"""Is this format supported?
1863
Supported formats must be initializable and openable.
1864
Unsupported formats may not support initialization or committing or
1865
some other features depending on the reason for not being supported.
1869
def check_conversion_target(self, target_format):
1870
raise NotImplementedError(self.check_conversion_target)
1872
def open(self, a_bzrdir, _found=False):
1873
"""Return an instance of this format for the bzrdir a_bzrdir.
1875
_found is a private parameter, do not use it.
1877
raise NotImplementedError(self.open)
1880
class MetaDirRepositoryFormat(RepositoryFormat):
1881
"""Common base class for the new repositories using the metadir layout."""
1883
rich_root_data = False
1884
supports_tree_reference = False
1885
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1888
super(MetaDirRepositoryFormat, self).__init__()
1890
def _create_control_files(self, a_bzrdir):
1891
"""Create the required files and the initial control_files object."""
1892
# FIXME: RBC 20060125 don't peek under the covers
1893
# NB: no need to escape relative paths that are url safe.
1894
repository_transport = a_bzrdir.get_repository_transport(self)
1895
control_files = lockable_files.LockableFiles(repository_transport,
1896
'lock', lockdir.LockDir)
1897
control_files.create_lock()
1898
return control_files
1900
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1901
"""Upload the initial blank content."""
1902
control_files = self._create_control_files(a_bzrdir)
1903
control_files.lock_write()
1905
control_files._transport.mkdir_multi(dirs,
1906
mode=control_files._dir_mode)
1907
for file, content in files:
1908
control_files.put(file, content)
1909
for file, content in utf8_files:
1910
control_files.put_utf8(file, content)
1912
control_files.put_utf8('shared-storage', '')
1914
control_files.unlock()
1917
# formats which have no format string are not discoverable
1918
# and not independently creatable, so are not registered. They're
1919
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1920
# needed, it's constructed directly by the BzrDir. Non-native formats where
1921
# the repository is not separately opened are similar.
1923
format_registry.register_lazy(
1924
'Bazaar-NG Repository format 7',
1925
'bzrlib.repofmt.weaverepo',
1929
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1930
# default control directory format
1931
format_registry.register_lazy(
1932
'Bazaar-NG Knit Repository Format 1',
1933
'bzrlib.repofmt.knitrepo',
1934
'RepositoryFormatKnit1',
1936
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1938
format_registry.register_lazy(
1939
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1940
'bzrlib.repofmt.knitrepo',
1941
'RepositoryFormatKnit3',
1945
class InterRepository(InterObject):
1946
"""This class represents operations taking place between two repositories.
1948
Its instances have methods like copy_content and fetch, and contain
1949
references to the source and target repositories these operations can be
1952
Often we will provide convenience methods on 'repository' which carry out
1953
operations with another repository - they will always forward to
1954
InterRepository.get(other).method_name(parameters).
1958
"""The available optimised InterRepository types."""
1960
def copy_content(self, revision_id=None):
1961
raise NotImplementedError(self.copy_content)
1963
def fetch(self, revision_id=None, pb=None):
1964
"""Fetch the content required to construct revision_id.
1966
The content is copied from self.source to self.target.
1968
:param revision_id: if None all content is copied, if NULL_REVISION no
1970
:param pb: optional progress bar to use for progress reports. If not
1971
provided a default one will be created.
1973
Returns the copied revision count and the failed revisions in a tuple:
1976
raise NotImplementedError(self.fetch)
1979
def missing_revision_ids(self, revision_id=None):
1980
"""Return the revision ids that source has that target does not.
1982
These are returned in topological order.
1984
:param revision_id: only return revision ids included by this
1987
# generic, possibly worst case, slow code path.
1988
target_ids = set(self.target.all_revision_ids())
1989
if revision_id is not None:
1990
source_ids = self.source.get_ancestry(revision_id)
1991
assert source_ids[0] is None
1994
source_ids = self.source.all_revision_ids()
1995
result_set = set(source_ids).difference(target_ids)
1996
# this may look like a no-op: its not. It preserves the ordering
1997
# other_ids had while only returning the members from other_ids
1998
# that we've decided we need.
1999
return [rev_id for rev_id in source_ids if rev_id in result_set]
2002
def _same_model(source, target):
2003
"""True if source and target have the same data representation."""
2004
if source.supports_rich_root() != target.supports_rich_root():
2006
if source._serializer != target._serializer:
2011
class InterSameDataRepository(InterRepository):
2012
"""Code for converting between repositories that represent the same data.
2014
Data format and model must match for this to work.
2018
def _get_repo_format_to_test(self):
2019
"""Repository format for testing with.
2021
InterSameData can pull from subtree to subtree and from non-subtree to
2022
non-subtree, so we test this with the richest repository format.
2024
from bzrlib.repofmt import knitrepo
2025
return knitrepo.RepositoryFormatKnit3()
2028
def is_compatible(source, target):
2029
return InterRepository._same_model(source, target)
2032
def copy_content(self, revision_id=None):
2033
"""Make a complete copy of the content in self into destination.
2035
This copies both the repository's revision data, and configuration information
2036
such as the make_working_trees setting.
2038
This is a destructive operation! Do not use it on existing
2041
:param revision_id: Only copy the content needed to construct
2042
revision_id and its parents.
2045
self.target.set_make_working_trees(self.source.make_working_trees())
2046
except NotImplementedError:
2048
# but don't bother fetching if we have the needed data now.
2049
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2050
self.target.has_revision(revision_id)):
2052
self.target.fetch(self.source, revision_id=revision_id)
2055
def fetch(self, revision_id=None, pb=None):
2056
"""See InterRepository.fetch()."""
2057
from bzrlib.fetch import GenericRepoFetcher
2058
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2059
self.source, self.source._format, self.target,
2060
self.target._format)
2061
f = GenericRepoFetcher(to_repository=self.target,
2062
from_repository=self.source,
2063
last_revision=revision_id,
2065
return f.count_copied, f.failed_revisions
2068
class InterWeaveRepo(InterSameDataRepository):
2069
"""Optimised code paths between Weave based repositories.
2071
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2072
implemented lazy inter-object optimisation.
2076
def _get_repo_format_to_test(self):
2077
from bzrlib.repofmt import weaverepo
2078
return weaverepo.RepositoryFormat7()
2081
def is_compatible(source, target):
2082
"""Be compatible with known Weave formats.
2084
We don't test for the stores being of specific types because that
2085
could lead to confusing results, and there is no need to be
2088
from bzrlib.repofmt.weaverepo import (
2094
return (isinstance(source._format, (RepositoryFormat5,
2096
RepositoryFormat7)) and
2097
isinstance(target._format, (RepositoryFormat5,
2099
RepositoryFormat7)))
2100
except AttributeError:
2104
def copy_content(self, revision_id=None):
2105
"""See InterRepository.copy_content()."""
2106
# weave specific optimised path:
2108
self.target.set_make_working_trees(self.source.make_working_trees())
2109
except NotImplementedError:
2111
# FIXME do not peek!
2112
if self.source.control_files._transport.listable():
2113
pb = ui.ui_factory.nested_progress_bar()
2115
self.target.weave_store.copy_all_ids(
2116
self.source.weave_store,
2118
from_transaction=self.source.get_transaction(),
2119
to_transaction=self.target.get_transaction())
2120
pb.update('copying inventory', 0, 1)
2121
self.target.control_weaves.copy_multi(
2122
self.source.control_weaves, ['inventory'],
2123
from_transaction=self.source.get_transaction(),
2124
to_transaction=self.target.get_transaction())
2125
self.target._revision_store.text_store.copy_all_ids(
2126
self.source._revision_store.text_store,
2131
self.target.fetch(self.source, revision_id=revision_id)
2134
def fetch(self, revision_id=None, pb=None):
2135
"""See InterRepository.fetch()."""
2136
from bzrlib.fetch import GenericRepoFetcher
2137
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2138
self.source, self.source._format, self.target, self.target._format)
2139
f = GenericRepoFetcher(to_repository=self.target,
2140
from_repository=self.source,
2141
last_revision=revision_id,
2143
return f.count_copied, f.failed_revisions
2146
def missing_revision_ids(self, revision_id=None):
2147
"""See InterRepository.missing_revision_ids()."""
2148
# we want all revisions to satisfy revision_id in source.
2149
# but we don't want to stat every file here and there.
2150
# we want then, all revisions other needs to satisfy revision_id
2151
# checked, but not those that we have locally.
2152
# so the first thing is to get a subset of the revisions to
2153
# satisfy revision_id in source, and then eliminate those that
2154
# we do already have.
2155
# this is slow on high latency connection to self, but as as this
2156
# disk format scales terribly for push anyway due to rewriting
2157
# inventory.weave, this is considered acceptable.
2159
if revision_id is not None:
2160
source_ids = self.source.get_ancestry(revision_id)
2161
assert source_ids[0] is None
2164
source_ids = self.source._all_possible_ids()
2165
source_ids_set = set(source_ids)
2166
# source_ids is the worst possible case we may need to pull.
2167
# now we want to filter source_ids against what we actually
2168
# have in target, but don't try to check for existence where we know
2169
# we do not have a revision as that would be pointless.
2170
target_ids = set(self.target._all_possible_ids())
2171
possibly_present_revisions = target_ids.intersection(source_ids_set)
2172
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2173
required_revisions = source_ids_set.difference(actually_present_revisions)
2174
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2175
if revision_id is not None:
2176
# we used get_ancestry to determine source_ids then we are assured all
2177
# revisions referenced are present as they are installed in topological order.
2178
# and the tip revision was validated by get_ancestry.
2179
return required_topo_revisions
2181
# if we just grabbed the possibly available ids, then
2182
# we only have an estimate of whats available and need to validate
2183
# that against the revision records.
2184
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2187
class InterKnitRepo(InterSameDataRepository):
2188
"""Optimised code paths between Knit based repositories."""
2191
def _get_repo_format_to_test(self):
2192
from bzrlib.repofmt import knitrepo
2193
return knitrepo.RepositoryFormatKnit1()
2196
def is_compatible(source, target):
2197
"""Be compatible with known Knit formats.
2199
We don't test for the stores being of specific types because that
2200
could lead to confusing results, and there is no need to be
2203
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2205
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2206
isinstance(target._format, RepositoryFormatKnit))
2207
except AttributeError:
2209
return are_knits and InterRepository._same_model(source, target)
2212
def fetch(self, revision_id=None, pb=None):
2213
"""See InterRepository.fetch()."""
2214
from bzrlib.fetch import KnitRepoFetcher
2215
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2216
self.source, self.source._format, self.target, self.target._format)
2217
f = KnitRepoFetcher(to_repository=self.target,
2218
from_repository=self.source,
2219
last_revision=revision_id,
2221
return f.count_copied, f.failed_revisions
2224
def missing_revision_ids(self, revision_id=None):
2225
"""See InterRepository.missing_revision_ids()."""
2226
if revision_id is not None:
2227
source_ids = self.source.get_ancestry(revision_id)
2228
assert source_ids[0] is None
2231
source_ids = self.source.all_revision_ids()
2232
source_ids_set = set(source_ids)
2233
# source_ids is the worst possible case we may need to pull.
2234
# now we want to filter source_ids against what we actually
2235
# have in target, but don't try to check for existence where we know
2236
# we do not have a revision as that would be pointless.
2237
target_ids = set(self.target.all_revision_ids())
2238
possibly_present_revisions = target_ids.intersection(source_ids_set)
2239
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2240
required_revisions = source_ids_set.difference(actually_present_revisions)
2241
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2242
if revision_id is not None:
2243
# we used get_ancestry to determine source_ids then we are assured all
2244
# revisions referenced are present as they are installed in topological order.
2245
# and the tip revision was validated by get_ancestry.
2246
return required_topo_revisions
2248
# if we just grabbed the possibly available ids, then
2249
# we only have an estimate of whats available and need to validate
2250
# that against the revision records.
2251
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2254
class InterModel1and2(InterRepository):
2257
def _get_repo_format_to_test(self):
2261
def is_compatible(source, target):
2262
if not source.supports_rich_root() and target.supports_rich_root():
2268
def fetch(self, revision_id=None, pb=None):
2269
"""See InterRepository.fetch()."""
2270
from bzrlib.fetch import Model1toKnit2Fetcher
2271
f = Model1toKnit2Fetcher(to_repository=self.target,
2272
from_repository=self.source,
2273
last_revision=revision_id,
2275
return f.count_copied, f.failed_revisions
2278
def copy_content(self, revision_id=None):
2279
"""Make a complete copy of the content in self into destination.
2281
This is a destructive operation! Do not use it on existing
2284
:param revision_id: Only copy the content needed to construct
2285
revision_id and its parents.
2288
self.target.set_make_working_trees(self.source.make_working_trees())
2289
except NotImplementedError:
2291
# but don't bother fetching if we have the needed data now.
2292
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2293
self.target.has_revision(revision_id)):
2295
self.target.fetch(self.source, revision_id=revision_id)
2298
class InterKnit1and2(InterKnitRepo):
2301
def _get_repo_format_to_test(self):
2305
def is_compatible(source, target):
2306
"""Be compatible with Knit1 source and Knit3 target"""
2307
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2309
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
2310
RepositoryFormatKnit3
2311
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2312
isinstance(target._format, (RepositoryFormatKnit3)))
2313
except AttributeError:
2317
def fetch(self, revision_id=None, pb=None):
2318
"""See InterRepository.fetch()."""
2319
from bzrlib.fetch import Knit1to2Fetcher
2320
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2321
self.source, self.source._format, self.target,
2322
self.target._format)
2323
f = Knit1to2Fetcher(to_repository=self.target,
2324
from_repository=self.source,
2325
last_revision=revision_id,
2327
return f.count_copied, f.failed_revisions
2330
class InterRemoteToOther(InterRepository):
2332
def __init__(self, source, target):
2333
InterRepository.__init__(self, source, target)
2334
self._real_inter = None
2337
def is_compatible(source, target):
2338
if not isinstance(source, remote.RemoteRepository):
2340
source._ensure_real()
2341
real_source = source._real_repository
2342
# Is source's model compatible with target's model, and are they the
2343
# same format? Currently we can only optimise fetching from an
2344
# identical model & format repo.
2345
assert not isinstance(real_source, remote.RemoteRepository), (
2346
"We don't support remote repos backed by remote repos yet.")
2347
return real_source._format == target._format
2350
def fetch(self, revision_id=None, pb=None):
2351
"""See InterRepository.fetch()."""
2352
from bzrlib.fetch import RemoteToOtherFetcher
2353
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2354
self.source, self.target, self.target._format)
2355
# TODO: jam 20070210 This should be an assert, not a translate
2356
revision_id = osutils.safe_revision_id(revision_id)
2357
f = RemoteToOtherFetcher(to_repository=self.target,
2358
from_repository=self.source,
2359
last_revision=revision_id,
2361
return f.count_copied, f.failed_revisions
2364
def _get_repo_format_to_test(self):
2368
class InterOtherToRemote(InterRepository):
2370
def __init__(self, source, target):
2371
InterRepository.__init__(self, source, target)
2372
self._real_inter = None
2375
def is_compatible(source, target):
2376
if isinstance(target, remote.RemoteRepository):
2380
def _ensure_real_inter(self):
2381
if self._real_inter is None:
2382
self.target._ensure_real()
2383
real_target = self.target._real_repository
2384
self._real_inter = InterRepository.get(self.source, real_target)
2386
def copy_content(self, revision_id=None):
2387
self._ensure_real_inter()
2388
self._real_inter.copy_content(revision_id=revision_id)
2390
def fetch(self, revision_id=None, pb=None):
2391
self._ensure_real_inter()
2392
self._real_inter.fetch(revision_id=revision_id, pb=pb)
2395
def _get_repo_format_to_test(self):
2399
InterRepository.register_optimiser(InterSameDataRepository)
2400
InterRepository.register_optimiser(InterWeaveRepo)
2401
InterRepository.register_optimiser(InterKnitRepo)
2402
InterRepository.register_optimiser(InterModel1and2)
2403
InterRepository.register_optimiser(InterKnit1and2)
2404
InterRepository.register_optimiser(InterRemoteToOther)
2405
InterRepository.register_optimiser(InterOtherToRemote)
2408
class CopyConverter(object):
2409
"""A repository conversion tool which just performs a copy of the content.
2411
This is slow but quite reliable.
2414
def __init__(self, target_format):
2415
"""Create a CopyConverter.
2417
:param target_format: The format the resulting repository should be.
2419
self.target_format = target_format
2421
def convert(self, repo, pb):
2422
"""Perform the conversion of to_convert, giving feedback via pb.
2424
:param to_convert: The disk object to convert.
2425
:param pb: a progress bar to use for progress information.
2430
# this is only useful with metadir layouts - separated repo content.
2431
# trigger an assertion if not such
2432
repo._format.get_format_string()
2433
self.repo_dir = repo.bzrdir
2434
self.step('Moving repository to repository.backup')
2435
self.repo_dir.transport.move('repository', 'repository.backup')
2436
backup_transport = self.repo_dir.transport.clone('repository.backup')
2437
repo._format.check_conversion_target(self.target_format)
2438
self.source_repo = repo._format.open(self.repo_dir,
2440
_override_transport=backup_transport)
2441
self.step('Creating new repository')
2442
converted = self.target_format.initialize(self.repo_dir,
2443
self.source_repo.is_shared())
2444
converted.lock_write()
2446
self.step('Copying content into repository.')
2447
self.source_repo.copy_content_into(converted)
2450
self.step('Deleting old repository content.')
2451
self.repo_dir.transport.delete_tree('repository.backup')
2452
self.pb.note('repository converted')
2454
def step(self, message):
2455
"""Update the pb by a step."""
2457
self.pb.update(message, self.count, self.total)
2469
def _unescaper(match, _map=_unescape_map):
2470
code = match.group(1)
2474
if not code.startswith('#'):
2476
return unichr(int(code[1:])).encode('utf8')
2482
def _unescape_xml(data):
2483
"""Unescape predefined XML entities in a string of data."""
2485
if _unescape_re is None:
2486
_unescape_re = re.compile('\&([^;]*);')
2487
return _unescape_re.sub(_unescaper, data)
2490
class _RevisionTextVersionCache(object):
2491
"""A cache of the versionedfile versions for revision and file-id."""
2493
def __init__(self, repository):
2494
self.repository = repository
2495
self.revision_versions = {}
2496
self.revision_parents = {}
2497
self.repo_graph = self.repository.get_graph()
2500
def add_revision_text_versions(self, tree):
2501
"""Cache text version data from the supplied revision tree"""
2503
for path, entry in tree.iter_entries_by_dir():
2504
inv_revisions[entry.file_id] = entry.revision
2505
self.revision_versions[tree.get_revision_id()] = inv_revisions
2506
return inv_revisions
2508
def get_text_version(self, file_id, revision_id):
2509
"""Determine the text version for a given file-id and revision-id"""
2511
inv_revisions = self.revision_versions[revision_id]
2514
tree = self.repository.revision_tree(revision_id)
2515
except errors.RevisionNotPresent:
2516
self.revision_versions[revision_id] = inv_revisions = {}
2518
inv_revisions = self.add_revision_text_versions(tree)
2519
return inv_revisions.get(file_id)
2521
def prepopulate_revs(self, revision_ids):
2522
# Filter out versions that we don't have an inventory for, so that the
2523
# revision_trees() call won't fail.
2524
inv_weave = self.repository.get_inventory_weave()
2525
revs = [r for r in revision_ids if inv_weave.has_version(r)]
2526
# XXX: this loop is very similar to
2527
# bzrlib.fetch.Inter1and2Helper.iter_rev_trees.
2529
for tree in self.repository.revision_trees(revs[:100]):
2530
if tree.inventory.revision_id is None:
2531
tree.inventory.revision_id = tree.get_revision_id()
2532
self.add_revision_text_versions(tree)
2535
def get_parents(self, revision_id):
2537
return self.revision_parents[revision_id]
2539
parents = self.repository.get_parents([revision_id])[0]
2540
self.revision_parents[revision_id] = parents
2543
def heads(self, revision_ids):
2544
revision_ids = tuple(revision_ids)
2546
return self.rev_heads[revision_ids]
2548
heads = self.repo_graph.heads(revision_ids)
2549
self.rev_heads[revision_ids] = heads
2552
class VersionedFileChecker(object):
2554
def __init__(self, planned_revisions, revision_versions, repository):
2555
self.planned_revisions = planned_revisions
2556
self.revision_versions = revision_versions
2557
self.repository = repository
2559
def calculate_file_version_parents(self, revision_id, file_id):
2560
text_revision = self.revision_versions.get_text_version(
2561
file_id, revision_id)
2562
if text_revision is None:
2564
parents_of_text_revision = self.revision_versions.get_parents(
2566
parents_from_inventories = []
2567
for parent in parents_of_text_revision:
2568
if parent == _mod_revision.NULL_REVISION:
2570
introduced_in = self.revision_versions.get_text_version(file_id,
2572
if introduced_in is not None:
2573
parents_from_inventories.append(introduced_in)
2574
heads = set(self.revision_versions.heads(parents_from_inventories))
2576
for parent in parents_from_inventories:
2577
if parent in heads and parent not in new_parents:
2578
new_parents.append(parent)
2581
def check_file_version_parents(self, weave, file_id):
2583
for num, revision_id in enumerate(self.planned_revisions):
2584
correct_parents = self.calculate_file_version_parents(
2585
revision_id, file_id)
2586
if correct_parents is None:
2588
text_revision = self.revision_versions.get_text_version(
2589
file_id, revision_id)
2590
knit_parents = weave.get_parents(text_revision)
2591
if correct_parents != knit_parents:
2592
result[revision_id] = (knit_parents, correct_parents)
added
removed
bzrlib/repository.py
