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 lock_write(self, token=None):
662
"""Lock this repository for writing.
664
This causes caching within the repository obejct to start accumlating
665
data during reads, and allows a 'write_group' to be obtained. Write
666
groups must be used for actual data insertion.
668
:param token: if this is already locked, then lock_write will fail
669
unless the token matches the existing lock.
670
:returns: a token if this instance supports tokens, otherwise None.
671
:raises TokenLockingNotSupported: when a token is given but this
672
instance doesn't support using token locks.
673
:raises MismatchedToken: if the specified token doesn't match the token
674
of the existing lock.
675
:seealso: start_write_group.
677
A token should be passed in if you know that you have locked the object
678
some other way, and need to synchronise this object's state with that
681
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
683
result = self.control_files.lock_write(token=token)
688
self.control_files.lock_read()
691
def get_physical_lock_status(self):
692
return self.control_files.get_physical_lock_status()
694
def leave_lock_in_place(self):
695
"""Tell this repository not to release the physical lock when this
698
If lock_write doesn't return a token, then this method is not supported.
700
self.control_files.leave_in_place()
702
def dont_leave_lock_in_place(self):
703
"""Tell this repository to release the physical lock when this
704
object is unlocked, even if it didn't originally acquire it.
706
If lock_write doesn't return a token, then this method is not supported.
708
self.control_files.dont_leave_in_place()
711
def gather_stats(self, revid=None, committers=None):
712
"""Gather statistics from a revision id.
714
:param revid: The revision id to gather statistics from, if None, then
715
no revision specific statistics are gathered.
716
:param committers: Optional parameter controlling whether to grab
717
a count of committers from the revision specific statistics.
718
:return: A dictionary of statistics. Currently this contains:
719
committers: The number of committers if requested.
720
firstrev: A tuple with timestamp, timezone for the penultimate left
721
most ancestor of revid, if revid is not the NULL_REVISION.
722
latestrev: A tuple with timestamp, timezone for revid, if revid is
723
not the NULL_REVISION.
724
revisions: The total revision count in the repository.
725
size: An estimate disk size of the repository in bytes.
728
if revid and committers:
729
result['committers'] = 0
730
if revid and revid != _mod_revision.NULL_REVISION:
732
all_committers = set()
733
revisions = self.get_ancestry(revid)
734
# pop the leading None
736
first_revision = None
738
# ignore the revisions in the middle - just grab first and last
739
revisions = revisions[0], revisions[-1]
740
for revision in self.get_revisions(revisions):
741
if not first_revision:
742
first_revision = revision
744
all_committers.add(revision.committer)
745
last_revision = revision
747
result['committers'] = len(all_committers)
748
result['firstrev'] = (first_revision.timestamp,
749
first_revision.timezone)
750
result['latestrev'] = (last_revision.timestamp,
751
last_revision.timezone)
753
# now gather global repository information
754
if self.bzrdir.root_transport.listable():
755
c, t = self._revision_store.total_size(self.get_transaction())
756
result['revisions'] = c
760
def get_data_stream(self, revision_ids):
761
raise NotImplementedError(self.get_data_stream)
763
def insert_data_stream(self, stream):
764
for item_key, bytes in stream:
765
if item_key[0] == 'file':
766
(file_id,) = item_key[1:]
767
knit = self.weave_store.get_weave_or_empty(
768
file_id, self.get_transaction())
769
elif item_key == ('inventory',):
770
knit = self.get_inventory_weave()
771
elif item_key == ('revisions',):
772
knit = self._revision_store.get_revision_file(
773
self.get_transaction())
774
elif item_key == ('signatures',):
775
knit = self._revision_store.get_signature_file(
776
self.get_transaction())
778
raise RepositoryDataStreamError(
779
"Unrecognised data stream key '%s'" % (item_key,))
780
decoded_list = bencode.bdecode(bytes)
781
format = decoded_list.pop(0)
784
for version, options, parents, some_bytes in decoded_list:
785
data_list.append((version, options, len(some_bytes), parents))
786
knit_bytes += some_bytes
787
knit.insert_data_stream(
788
(format, data_list, StringIO(knit_bytes).read))
791
def missing_revision_ids(self, other, revision_id=None):
792
"""Return the revision ids that other has that this does not.
794
These are returned in topological order.
796
revision_id: only return revision ids included by revision_id.
798
return InterRepository.get(other, self).missing_revision_ids(revision_id)
802
"""Open the repository rooted at base.
804
For instance, if the repository is at URL/.bzr/repository,
805
Repository.open(URL) -> a Repository instance.
807
control = bzrdir.BzrDir.open(base)
808
return control.open_repository()
810
def copy_content_into(self, destination, revision_id=None):
811
"""Make a complete copy of the content in self into destination.
813
This is a destructive operation! Do not use it on existing
816
return InterRepository.get(self, destination).copy_content(revision_id)
818
def commit_write_group(self):
819
"""Commit the contents accrued within the current write group.
821
:seealso: start_write_group.
823
if self._write_group is not self.get_transaction():
824
# has an unlock or relock occured ?
825
raise errors.BzrError('mismatched lock context %r and '
827
(self.get_transaction(), self._write_group))
828
self._commit_write_group()
829
self._write_group = None
831
def _commit_write_group(self):
832
"""Template method for per-repository write group cleanup.
834
This is called before the write group is considered to be
835
finished and should ensure that all data handed to the repository
836
for writing during the write group is safely committed (to the
837
extent possible considering file system caching etc).
840
def fetch(self, source, revision_id=None, pb=None):
841
"""Fetch the content required to construct revision_id from source.
843
If revision_id is None all content is copied.
845
# fast path same-url fetch operations
846
if self.has_same_location(source):
847
# check that last_revision is in 'from' and then return a
849
if (revision_id is not None and
850
not _mod_revision.is_null(revision_id)):
851
self.get_revision(revision_id)
853
inter = InterRepository.get(source, self)
855
return inter.fetch(revision_id=revision_id, pb=pb)
856
except NotImplementedError:
857
raise errors.IncompatibleRepositories(source, self)
859
def create_bundle(self, target, base, fileobj, format=None):
860
return serializer.write_bundle(self, target, base, fileobj, format)
862
def get_commit_builder(self, branch, parents, config, timestamp=None,
863
timezone=None, committer=None, revprops=None,
865
"""Obtain a CommitBuilder for this repository.
867
:param branch: Branch to commit to.
868
:param parents: Revision ids of the parents of the new revision.
869
:param config: Configuration to use.
870
:param timestamp: Optional timestamp recorded for commit.
871
:param timezone: Optional timezone for timestamp.
872
:param committer: Optional committer to set for commit.
873
:param revprops: Optional dictionary of revision properties.
874
:param revision_id: Optional revision id.
876
result = self._commit_builder_class(self, parents, config,
877
timestamp, timezone, committer, revprops, revision_id)
878
self.start_write_group()
882
if (self.control_files._lock_count == 1 and
883
self.control_files._lock_mode == 'w'):
884
if self._write_group is not None:
885
raise errors.BzrError(
886
'Must end write groups before releasing write locks.')
887
self.control_files.unlock()
890
def clone(self, a_bzrdir, revision_id=None):
891
"""Clone this repository into a_bzrdir using the current format.
893
Currently no check is made that the format of this repository and
894
the bzrdir format are compatible. FIXME RBC 20060201.
896
:return: The newly created destination repository.
898
# TODO: deprecate after 0.16; cloning this with all its settings is
899
# probably not very useful -- mbp 20070423
900
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
901
self.copy_content_into(dest_repo, revision_id)
904
def start_write_group(self):
905
"""Start a write group in the repository.
907
Write groups are used by repositories which do not have a 1:1 mapping
908
between file ids and backend store to manage the insertion of data from
909
both fetch and commit operations.
911
A write lock is required around the start_write_group/commit_write_group
912
for the support of lock-requiring repository formats.
914
One can only insert data into a repository inside a write group.
918
if not self.is_locked() or self.control_files._lock_mode != 'w':
919
raise errors.NotWriteLocked(self)
920
if self._write_group:
921
raise errors.BzrError('already in a write group')
922
self._start_write_group()
923
# so we can detect unlock/relock - the write group is now entered.
924
self._write_group = self.get_transaction()
926
def _start_write_group(self):
927
"""Template method for per-repository write group startup.
929
This is called before the write group is considered to be
934
def sprout(self, to_bzrdir, revision_id=None):
935
"""Create a descendent repository for new development.
937
Unlike clone, this does not copy the settings of the repository.
939
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
940
dest_repo.fetch(self, revision_id=revision_id)
943
def _create_sprouting_repo(self, a_bzrdir, shared):
944
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
945
# use target default format.
946
dest_repo = a_bzrdir.create_repository()
948
# Most control formats need the repository to be specifically
949
# created, but on some old all-in-one formats it's not needed
951
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
952
except errors.UninitializableFormat:
953
dest_repo = a_bzrdir.open_repository()
957
def has_revision(self, revision_id):
958
"""True if this repository has a copy of the revision."""
959
if 'evil' in debug.debug_flags:
960
mutter_callsite(3, "has_revision is a LBYL symptom.")
961
return self._revision_store.has_revision_id(revision_id,
962
self.get_transaction())
965
def get_revision(self, revision_id):
966
"""Return the Revision object for a named revision."""
967
return self.get_revisions([revision_id])[0]
970
def get_revision_reconcile(self, revision_id):
971
"""'reconcile' helper routine that allows access to a revision always.
973
This variant of get_revision does not cross check the weave graph
974
against the revision one as get_revision does: but it should only
975
be used by reconcile, or reconcile-alike commands that are correcting
976
or testing the revision graph.
978
return self._get_revisions([revision_id])[0]
981
def get_revisions(self, revision_ids):
982
"""Get many revisions at once."""
983
return self._get_revisions(revision_ids)
986
def _get_revisions(self, revision_ids):
987
"""Core work logic to get many revisions without sanity checks."""
988
for rev_id in revision_ids:
989
if not rev_id or not isinstance(rev_id, basestring):
990
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
991
revs = self._revision_store.get_revisions(revision_ids,
992
self.get_transaction())
994
assert not isinstance(rev.revision_id, unicode)
995
for parent_id in rev.parent_ids:
996
assert not isinstance(parent_id, unicode)
1000
def get_revision_xml(self, revision_id):
1001
# TODO: jam 20070210 This shouldn't be necessary since get_revision
1002
# would have already do it.
1003
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
1004
rev = self.get_revision(revision_id)
1005
rev_tmp = StringIO()
1006
# the current serializer..
1007
self._revision_store._serializer.write_revision(rev, rev_tmp)
1009
return rev_tmp.getvalue()
1012
def get_deltas_for_revisions(self, revisions):
1013
"""Produce a generator of revision deltas.
1015
Note that the input is a sequence of REVISIONS, not revision_ids.
1016
Trees will be held in memory until the generator exits.
1017
Each delta is relative to the revision's lefthand predecessor.
1019
required_trees = set()
1020
for revision in revisions:
1021
required_trees.add(revision.revision_id)
1022
required_trees.update(revision.parent_ids[:1])
1023
trees = dict((t.get_revision_id(), t) for
1024
t in self.revision_trees(required_trees))
1025
for revision in revisions:
1026
if not revision.parent_ids:
1027
old_tree = self.revision_tree(None)
1029
old_tree = trees[revision.parent_ids[0]]
1030
yield trees[revision.revision_id].changes_from(old_tree)
1033
def get_revision_delta(self, revision_id):
1034
"""Return the delta for one revision.
1036
The delta is relative to the left-hand predecessor of the
1039
r = self.get_revision(revision_id)
1040
return list(self.get_deltas_for_revisions([r]))[0]
1043
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1044
signature = gpg_strategy.sign(plaintext)
1045
self._revision_store.add_revision_signature_text(revision_id,
1047
self.get_transaction())
1049
def fileids_altered_by_revision_ids(self, revision_ids):
1050
"""Find the file ids and versions affected by revisions.
1052
:param revisions: an iterable containing revision ids.
1053
:return: a dictionary mapping altered file-ids to an iterable of
1054
revision_ids. Each altered file-ids has the exact revision_ids that
1055
altered it listed explicitly.
1057
assert self._serializer.support_altered_by_hack, \
1058
("fileids_altered_by_revision_ids only supported for branches "
1059
"which store inventory as unnested xml, not on %r" % self)
1060
selected_revision_ids = set(revision_ids)
1061
w = self.get_inventory_weave()
1064
# this code needs to read every new line in every inventory for the
1065
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1066
# not present in one of those inventories is unnecessary but not
1067
# harmful because we are filtering by the revision id marker in the
1068
# inventory lines : we only select file ids altered in one of those
1069
# revisions. We don't need to see all lines in the inventory because
1070
# only those added in an inventory in rev X can contain a revision=X
1072
unescape_revid_cache = {}
1073
unescape_fileid_cache = {}
1075
# jam 20061218 In a big fetch, this handles hundreds of thousands
1076
# of lines, so it has had a lot of inlining and optimizing done.
1077
# Sorry that it is a little bit messy.
1078
# Move several functions to be local variables, since this is a long
1080
search = self._file_ids_altered_regex.search
1081
unescape = _unescape_xml
1082
setdefault = result.setdefault
1083
pb = ui.ui_factory.nested_progress_bar()
1085
for line in w.iter_lines_added_or_present_in_versions(
1086
selected_revision_ids, pb=pb):
1087
match = search(line)
1090
# One call to match.group() returning multiple items is quite a
1091
# bit faster than 2 calls to match.group() each returning 1
1092
file_id, revision_id = match.group('file_id', 'revision_id')
1094
# Inlining the cache lookups helps a lot when you make 170,000
1095
# lines and 350k ids, versus 8.4 unique ids.
1096
# Using a cache helps in 2 ways:
1097
# 1) Avoids unnecessary decoding calls
1098
# 2) Re-uses cached strings, which helps in future set and
1100
# (2) is enough that removing encoding entirely along with
1101
# the cache (so we are using plain strings) results in no
1102
# performance improvement.
1104
revision_id = unescape_revid_cache[revision_id]
1106
unescaped = unescape(revision_id)
1107
unescape_revid_cache[revision_id] = unescaped
1108
revision_id = unescaped
1110
if revision_id in selected_revision_ids:
1112
file_id = unescape_fileid_cache[file_id]
1114
unescaped = unescape(file_id)
1115
unescape_fileid_cache[file_id] = unescaped
1117
setdefault(file_id, set()).add(revision_id)
1122
def iter_files_bytes(self, desired_files):
1123
"""Iterate through file versions.
1125
Files will not necessarily be returned in the order they occur in
1126
desired_files. No specific order is guaranteed.
1128
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1129
value supplied by the caller as part of desired_files. It should
1130
uniquely identify the file version in the caller's context. (Examples:
1131
an index number or a TreeTransform trans_id.)
1133
bytes_iterator is an iterable of bytestrings for the file. The
1134
kind of iterable and length of the bytestrings are unspecified, but for
1135
this implementation, it is a list of lines produced by
1136
VersionedFile.get_lines().
1138
:param desired_files: a list of (file_id, revision_id, identifier)
1141
transaction = self.get_transaction()
1142
for file_id, revision_id, callable_data in desired_files:
1144
weave = self.weave_store.get_weave(file_id, transaction)
1145
except errors.NoSuchFile:
1146
raise errors.NoSuchIdInRepository(self, file_id)
1147
yield callable_data, weave.get_lines(revision_id)
1149
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1150
"""Get an iterable listing the keys of all the data introduced by a set
1153
The keys will be ordered so that the corresponding items can be safely
1154
fetched and inserted in that order.
1156
:returns: An iterable producing tuples of (knit-kind, file-id,
1157
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1158
'revisions'. file-id is None unless knit-kind is 'file'.
1160
# XXX: it's a bit weird to control the inventory weave caching in this
1161
# generator. Ideally the caching would be done in fetch.py I think. Or
1162
# maybe this generator should explicitly have the contract that it
1163
# should not be iterated until the previously yielded item has been
1166
inv_w = self.get_inventory_weave()
1167
inv_w.enable_cache()
1169
# file ids that changed
1170
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1172
num_file_ids = len(file_ids)
1173
for file_id, altered_versions in file_ids.iteritems():
1174
if _files_pb is not None:
1175
_files_pb.update("fetch texts", count, num_file_ids)
1177
yield ("file", file_id, altered_versions)
1178
# We're done with the files_pb. Note that it finished by the caller,
1179
# just as it was created by the caller.
1183
yield ("inventory", None, revision_ids)
1187
revisions_with_signatures = set()
1188
for rev_id in revision_ids:
1190
self.get_signature_text(rev_id)
1191
except errors.NoSuchRevision:
1195
revisions_with_signatures.add(rev_id)
1197
yield ("signatures", None, revisions_with_signatures)
1200
yield ("revisions", None, revision_ids)
1203
def get_inventory_weave(self):
1204
return self.control_weaves.get_weave('inventory',
1205
self.get_transaction())
1208
def get_inventory(self, revision_id):
1209
"""Get Inventory object by hash."""
1210
return self.deserialise_inventory(
1211
revision_id, self.get_inventory_xml(revision_id))
1213
def deserialise_inventory(self, revision_id, xml):
1214
"""Transform the xml into an inventory object.
1216
:param revision_id: The expected revision id of the inventory.
1217
:param xml: A serialised inventory.
1219
return self._serializer.read_inventory_from_string(xml, revision_id)
1221
def serialise_inventory(self, inv):
1222
return self._serializer.write_inventory_to_string(inv)
1224
def _serialise_inventory_to_lines(self, inv):
1225
return self._serializer.write_inventory_to_lines(inv)
1227
def get_serializer_format(self):
1228
return self._serializer.format_num
1231
def get_inventory_xml(self, revision_id):
1232
"""Get inventory XML as a file object."""
1234
assert isinstance(revision_id, str), type(revision_id)
1235
iw = self.get_inventory_weave()
1236
return iw.get_text(revision_id)
1238
raise errors.HistoryMissing(self, 'inventory', revision_id)
1241
def get_inventory_sha1(self, revision_id):
1242
"""Return the sha1 hash of the inventory entry
1244
return self.get_revision(revision_id).inventory_sha1
1247
def get_revision_graph(self, revision_id=None):
1248
"""Return a dictionary containing the revision graph.
1250
NB: This method should not be used as it accesses the entire graph all
1251
at once, which is much more data than most operations should require.
1253
:param revision_id: The revision_id to get a graph from. If None, then
1254
the entire revision graph is returned. This is a deprecated mode of
1255
operation and will be removed in the future.
1256
:return: a dictionary of revision_id->revision_parents_list.
1258
raise NotImplementedError(self.get_revision_graph)
1261
def get_revision_graph_with_ghosts(self, revision_ids=None):
1262
"""Return a graph of the revisions with ghosts marked as applicable.
1264
:param revision_ids: an iterable of revisions to graph or None for all.
1265
:return: a Graph object with the graph reachable from revision_ids.
1267
if 'evil' in debug.debug_flags:
1269
"get_revision_graph_with_ghosts scales with size of history.")
1270
result = deprecated_graph.Graph()
1271
if not revision_ids:
1272
pending = set(self.all_revision_ids())
1275
pending = set(revision_ids)
1276
# special case NULL_REVISION
1277
if _mod_revision.NULL_REVISION in pending:
1278
pending.remove(_mod_revision.NULL_REVISION)
1279
required = set(pending)
1282
revision_id = pending.pop()
1284
rev = self.get_revision(revision_id)
1285
except errors.NoSuchRevision:
1286
if revision_id in required:
1289
result.add_ghost(revision_id)
1291
for parent_id in rev.parent_ids:
1292
# is this queued or done ?
1293
if (parent_id not in pending and
1294
parent_id not in done):
1296
pending.add(parent_id)
1297
result.add_node(revision_id, rev.parent_ids)
1298
done.add(revision_id)
1301
def _get_history_vf(self):
1302
"""Get a versionedfile whose history graph reflects all revisions.
1304
For weave repositories, this is the inventory weave.
1306
return self.get_inventory_weave()
1308
def iter_reverse_revision_history(self, revision_id):
1309
"""Iterate backwards through revision ids in the lefthand history
1311
:param revision_id: The revision id to start with. All its lefthand
1312
ancestors will be traversed.
1314
if revision_id in (None, _mod_revision.NULL_REVISION):
1316
next_id = revision_id
1317
versionedfile = self._get_history_vf()
1320
parents = versionedfile.get_parents(next_id)
1321
if len(parents) == 0:
1324
next_id = parents[0]
1327
def get_revision_inventory(self, revision_id):
1328
"""Return inventory of a past revision."""
1329
# TODO: Unify this with get_inventory()
1330
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1331
# must be the same as its revision, so this is trivial.
1332
if revision_id is None:
1333
# This does not make sense: if there is no revision,
1334
# then it is the current tree inventory surely ?!
1335
# and thus get_root_id() is something that looks at the last
1336
# commit on the branch, and the get_root_id is an inventory check.
1337
raise NotImplementedError
1338
# return Inventory(self.get_root_id())
1340
return self.get_inventory(revision_id)
1343
def is_shared(self):
1344
"""Return True if this repository is flagged as a shared repository."""
1345
raise NotImplementedError(self.is_shared)
1348
def reconcile(self, other=None, thorough=False):
1349
"""Reconcile this repository."""
1350
from bzrlib.reconcile import RepoReconciler
1351
reconciler = RepoReconciler(self, thorough=thorough)
1352
reconciler.reconcile()
1355
def _refresh_data(self):
1356
"""Helper called from lock_* to ensure coherency with disk.
1358
The default implementation does nothing; it is however possible
1359
for repositories to maintain loaded indices across multiple locks
1360
by checking inside their implementation of this method to see
1361
whether their indices are still valid. This depends of course on
1362
the disk format being validatable in this manner.
1366
def revision_tree(self, revision_id):
1367
"""Return Tree for a revision on this branch.
1369
`revision_id` may be None for the empty tree revision.
1371
# TODO: refactor this to use an existing revision object
1372
# so we don't need to read it in twice.
1373
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1374
return RevisionTree(self, Inventory(root_id=None),
1375
_mod_revision.NULL_REVISION)
1377
inv = self.get_revision_inventory(revision_id)
1378
return RevisionTree(self, inv, revision_id)
1381
def revision_trees(self, revision_ids):
1382
"""Return Tree for a revision on this branch.
1384
`revision_id` may not be None or 'null:'"""
1385
assert None not in revision_ids
1386
assert _mod_revision.NULL_REVISION not in revision_ids
1387
texts = self.get_inventory_weave().get_texts(revision_ids)
1388
for text, revision_id in zip(texts, revision_ids):
1389
inv = self.deserialise_inventory(revision_id, text)
1390
yield RevisionTree(self, inv, revision_id)
1393
def get_ancestry(self, revision_id, topo_sorted=True):
1394
"""Return a list of revision-ids integrated by a revision.
1396
The first element of the list is always None, indicating the origin
1397
revision. This might change when we have history horizons, or
1398
perhaps we should have a new API.
1400
This is topologically sorted.
1402
if _mod_revision.is_null(revision_id):
1404
if not self.has_revision(revision_id):
1405
raise errors.NoSuchRevision(self, revision_id)
1406
w = self.get_inventory_weave()
1407
candidates = w.get_ancestry(revision_id, topo_sorted)
1408
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1411
"""Compress the data within the repository.
1413
This operation only makes sense for some repository types. For other
1414
types it should be a no-op that just returns.
1416
This stub method does not require a lock, but subclasses should use
1417
@needs_write_lock as this is a long running call its reasonable to
1418
implicitly lock for the user.
1422
def print_file(self, file, revision_id):
1423
"""Print `file` to stdout.
1425
FIXME RBC 20060125 as John Meinel points out this is a bad api
1426
- it writes to stdout, it assumes that that is valid etc. Fix
1427
by creating a new more flexible convenience function.
1429
tree = self.revision_tree(revision_id)
1430
# use inventory as it was in that revision
1431
file_id = tree.inventory.path2id(file)
1433
# TODO: jam 20060427 Write a test for this code path
1434
# it had a bug in it, and was raising the wrong
1436
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1437
tree.print_file(file_id)
1439
def get_transaction(self):
1440
return self.control_files.get_transaction()
1442
def revision_parents(self, revision_id):
1443
return self.get_inventory_weave().parent_names(revision_id)
1445
def get_parents(self, revision_ids):
1446
"""See StackedParentsProvider.get_parents"""
1448
for revision_id in revision_ids:
1449
if revision_id == _mod_revision.NULL_REVISION:
1453
parents = self.get_revision(revision_id).parent_ids
1454
except errors.NoSuchRevision:
1457
if len(parents) == 0:
1458
parents = [_mod_revision.NULL_REVISION]
1459
parents_list.append(parents)
1462
def _make_parents_provider(self):
1465
def get_graph(self, other_repository=None):
1466
"""Return the graph walker for this repository format"""
1467
parents_provider = self._make_parents_provider()
1468
if (other_repository is not None and
1469
other_repository.bzrdir.transport.base !=
1470
self.bzrdir.transport.base):
1471
parents_provider = graph._StackedParentsProvider(
1472
[parents_provider, other_repository._make_parents_provider()])
1473
return graph.Graph(parents_provider)
1475
def get_versioned_file_checker(self, revisions, revision_versions_cache):
1476
return VersionedFileChecker(revisions, revision_versions_cache, self)
1479
def set_make_working_trees(self, new_value):
1480
"""Set the policy flag for making working trees when creating branches.
1482
This only applies to branches that use this repository.
1484
The default is 'True'.
1485
:param new_value: True to restore the default, False to disable making
1488
raise NotImplementedError(self.set_make_working_trees)
1490
def make_working_trees(self):
1491
"""Returns the policy for making working trees on new branches."""
1492
raise NotImplementedError(self.make_working_trees)
1495
def sign_revision(self, revision_id, gpg_strategy):
1496
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1497
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1500
def has_signature_for_revision_id(self, revision_id):
1501
"""Query for a revision signature for revision_id in the repository."""
1502
return self._revision_store.has_signature(revision_id,
1503
self.get_transaction())
1506
def get_signature_text(self, revision_id):
1507
"""Return the text for a signature."""
1508
return self._revision_store.get_signature_text(revision_id,
1509
self.get_transaction())
1512
def check(self, revision_ids=None):
1513
"""Check consistency of all history of given revision_ids.
1515
Different repository implementations should override _check().
1517
:param revision_ids: A non-empty list of revision_ids whose ancestry
1518
will be checked. Typically the last revision_id of a branch.
1520
return self._check(revision_ids)
1522
def _check(self, revision_ids):
1523
result = check.Check(self)
1527
def _warn_if_deprecated(self):
1528
global _deprecation_warning_done
1529
if _deprecation_warning_done:
1531
_deprecation_warning_done = True
1532
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1533
% (self._format, self.bzrdir.transport.base))
1535
def supports_rich_root(self):
1536
return self._format.rich_root_data
1538
def _check_ascii_revisionid(self, revision_id, method):
1539
"""Private helper for ascii-only repositories."""
1540
# weave repositories refuse to store revisionids that are non-ascii.
1541
if revision_id is not None:
1542
# weaves require ascii revision ids.
1543
if isinstance(revision_id, unicode):
1545
revision_id.encode('ascii')
1546
except UnicodeEncodeError:
1547
raise errors.NonAsciiRevisionId(method, self)
1550
revision_id.decode('ascii')
1551
except UnicodeDecodeError:
1552
raise errors.NonAsciiRevisionId(method, self)
1554
def revision_graph_can_have_wrong_parents(self):
1555
"""Is it possible for this repository to have a revision graph with
1558
If True, then this repository must also implement
1559
_find_inconsistent_revision_parents so that check and reconcile can
1560
check for inconsistencies before proceeding with other checks that may
1561
depend on the revision index being consistent.
1563
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1565
# remove these delegates a while after bzr 0.15
1566
def __make_delegated(name, from_module):
1567
def _deprecated_repository_forwarder():
1568
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1569
% (name, from_module),
1572
m = __import__(from_module, globals(), locals(), [name])
1574
return getattr(m, name)
1575
except AttributeError:
1576
raise AttributeError('module %s has no name %s'
1578
globals()[name] = _deprecated_repository_forwarder
1581
'AllInOneRepository',
1582
'WeaveMetaDirRepository',
1583
'PreSplitOutRepositoryFormat',
1584
'RepositoryFormat4',
1585
'RepositoryFormat5',
1586
'RepositoryFormat6',
1587
'RepositoryFormat7',
1589
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1593
'RepositoryFormatKnit',
1594
'RepositoryFormatKnit1',
1596
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1599
def install_revision(repository, rev, revision_tree):
1600
"""Install all revision data into a repository."""
1601
present_parents = []
1603
for p_id in rev.parent_ids:
1604
if repository.has_revision(p_id):
1605
present_parents.append(p_id)
1606
parent_trees[p_id] = repository.revision_tree(p_id)
1608
parent_trees[p_id] = repository.revision_tree(None)
1610
inv = revision_tree.inventory
1611
entries = inv.iter_entries()
1612
# backwards compatibility hack: skip the root id.
1613
if not repository.supports_rich_root():
1614
path, root = entries.next()
1615
if root.revision != rev.revision_id:
1616
raise errors.IncompatibleRevision(repr(repository))
1617
# Add the texts that are not already present
1618
for path, ie in entries:
1619
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1620
repository.get_transaction())
1621
if ie.revision not in w:
1623
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1624
# with InventoryEntry.find_previous_heads(). if it is, then there
1625
# is a latent bug here where the parents may have ancestors of each
1627
for revision, tree in parent_trees.iteritems():
1628
if ie.file_id not in tree:
1630
parent_id = tree.inventory[ie.file_id].revision
1631
if parent_id in text_parents:
1633
text_parents.append(parent_id)
1635
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1636
repository.get_transaction())
1637
lines = revision_tree.get_file(ie.file_id).readlines()
1638
vfile.add_lines(rev.revision_id, text_parents, lines)
1640
# install the inventory
1641
repository.add_inventory(rev.revision_id, inv, present_parents)
1642
except errors.RevisionAlreadyPresent:
1644
repository.add_revision(rev.revision_id, rev, inv)
1647
class MetaDirRepository(Repository):
1648
"""Repositories in the new meta-dir layout."""
1650
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1651
super(MetaDirRepository, self).__init__(_format,
1657
dir_mode = self.control_files._dir_mode
1658
file_mode = self.control_files._file_mode
1661
def is_shared(self):
1662
"""Return True if this repository is flagged as a shared repository."""
1663
return self.control_files._transport.has('shared-storage')
1666
def set_make_working_trees(self, new_value):
1667
"""Set the policy flag for making working trees when creating branches.
1669
This only applies to branches that use this repository.
1671
The default is 'True'.
1672
:param new_value: True to restore the default, False to disable making
1677
self.control_files._transport.delete('no-working-trees')
1678
except errors.NoSuchFile:
1681
self.control_files.put_utf8('no-working-trees', '')
1683
def make_working_trees(self):
1684
"""Returns the policy for making working trees on new branches."""
1685
return not self.control_files._transport.has('no-working-trees')
1688
class RepositoryFormatRegistry(registry.Registry):
1689
"""Registry of RepositoryFormats."""
1691
def get(self, format_string):
1692
r = registry.Registry.get(self, format_string)
1698
format_registry = RepositoryFormatRegistry()
1699
"""Registry of formats, indexed by their identifying format string.
1701
This can contain either format instances themselves, or classes/factories that
1702
can be called to obtain one.
1706
#####################################################################
1707
# Repository Formats
1709
class RepositoryFormat(object):
1710
"""A repository format.
1712
Formats provide three things:
1713
* An initialization routine to construct repository data on disk.
1714
* a format string which is used when the BzrDir supports versioned
1716
* an open routine which returns a Repository instance.
1718
There is one and only one Format subclass for each on-disk format. But
1719
there can be one Repository subclass that is used for several different
1720
formats. The _format attribute on a Repository instance can be used to
1721
determine the disk format.
1723
Formats are placed in an dict by their format string for reference
1724
during opening. These should be subclasses of RepositoryFormat
1727
Once a format is deprecated, just deprecate the initialize and open
1728
methods on the format class. Do not deprecate the object, as the
1729
object will be created every system load.
1731
Common instance attributes:
1732
_matchingbzrdir - the bzrdir format that the repository format was
1733
originally written to work with. This can be used if manually
1734
constructing a bzrdir and repository, or more commonly for test suite
1739
return "<%s>" % self.__class__.__name__
1741
def __eq__(self, other):
1742
# format objects are generally stateless
1743
return isinstance(other, self.__class__)
1745
def __ne__(self, other):
1746
return not self == other
1749
def find_format(klass, a_bzrdir):
1750
"""Return the format for the repository object in a_bzrdir.
1752
This is used by bzr native formats that have a "format" file in
1753
the repository. Other methods may be used by different types of
1757
transport = a_bzrdir.get_repository_transport(None)
1758
format_string = transport.get("format").read()
1759
return format_registry.get(format_string)
1760
except errors.NoSuchFile:
1761
raise errors.NoRepositoryPresent(a_bzrdir)
1763
raise errors.UnknownFormatError(format=format_string)
1766
def register_format(klass, format):
1767
format_registry.register(format.get_format_string(), format)
1770
def unregister_format(klass, format):
1771
format_registry.remove(format.get_format_string())
1774
def get_default_format(klass):
1775
"""Return the current default format."""
1776
from bzrlib import bzrdir
1777
return bzrdir.format_registry.make_bzrdir('default').repository_format
1779
def _get_control_store(self, repo_transport, control_files):
1780
"""Return the control store for this repository."""
1781
raise NotImplementedError(self._get_control_store)
1783
def get_format_string(self):
1784
"""Return the ASCII format string that identifies this format.
1786
Note that in pre format ?? repositories the format string is
1787
not permitted nor written to disk.
1789
raise NotImplementedError(self.get_format_string)
1791
def get_format_description(self):
1792
"""Return the short description for this format."""
1793
raise NotImplementedError(self.get_format_description)
1795
def _get_revision_store(self, repo_transport, control_files):
1796
"""Return the revision store object for this a_bzrdir."""
1797
raise NotImplementedError(self._get_revision_store)
1799
def _get_text_rev_store(self,
1806
"""Common logic for getting a revision store for a repository.
1808
see self._get_revision_store for the subclass-overridable method to
1809
get the store for a repository.
1811
from bzrlib.store.revision.text import TextRevisionStore
1812
dir_mode = control_files._dir_mode
1813
file_mode = control_files._file_mode
1814
text_store = TextStore(transport.clone(name),
1816
compressed=compressed,
1818
file_mode=file_mode)
1819
_revision_store = TextRevisionStore(text_store, serializer)
1820
return _revision_store
1822
# TODO: this shouldn't be in the base class, it's specific to things that
1823
# use weaves or knits -- mbp 20070207
1824
def _get_versioned_file_store(self,
1829
versionedfile_class=None,
1830
versionedfile_kwargs={},
1832
if versionedfile_class is None:
1833
versionedfile_class = self._versionedfile_class
1834
weave_transport = control_files._transport.clone(name)
1835
dir_mode = control_files._dir_mode
1836
file_mode = control_files._file_mode
1837
return VersionedFileStore(weave_transport, prefixed=prefixed,
1839
file_mode=file_mode,
1840
versionedfile_class=versionedfile_class,
1841
versionedfile_kwargs=versionedfile_kwargs,
1844
def initialize(self, a_bzrdir, shared=False):
1845
"""Initialize a repository of this format in a_bzrdir.
1847
:param a_bzrdir: The bzrdir to put the new repository in it.
1848
:param shared: The repository should be initialized as a sharable one.
1849
:returns: The new repository object.
1851
This may raise UninitializableFormat if shared repository are not
1852
compatible the a_bzrdir.
1854
raise NotImplementedError(self.initialize)
1856
def is_supported(self):
1857
"""Is this format supported?
1859
Supported formats must be initializable and openable.
1860
Unsupported formats may not support initialization or committing or
1861
some other features depending on the reason for not being supported.
1865
def check_conversion_target(self, target_format):
1866
raise NotImplementedError(self.check_conversion_target)
1868
def open(self, a_bzrdir, _found=False):
1869
"""Return an instance of this format for the bzrdir a_bzrdir.
1871
_found is a private parameter, do not use it.
1873
raise NotImplementedError(self.open)
1876
class MetaDirRepositoryFormat(RepositoryFormat):
1877
"""Common base class for the new repositories using the metadir layout."""
1879
rich_root_data = False
1880
supports_tree_reference = False
1881
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1884
super(MetaDirRepositoryFormat, self).__init__()
1886
def _create_control_files(self, a_bzrdir):
1887
"""Create the required files and the initial control_files object."""
1888
# FIXME: RBC 20060125 don't peek under the covers
1889
# NB: no need to escape relative paths that are url safe.
1890
repository_transport = a_bzrdir.get_repository_transport(self)
1891
control_files = lockable_files.LockableFiles(repository_transport,
1892
'lock', lockdir.LockDir)
1893
control_files.create_lock()
1894
return control_files
1896
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1897
"""Upload the initial blank content."""
1898
control_files = self._create_control_files(a_bzrdir)
1899
control_files.lock_write()
1901
control_files._transport.mkdir_multi(dirs,
1902
mode=control_files._dir_mode)
1903
for file, content in files:
1904
control_files.put(file, content)
1905
for file, content in utf8_files:
1906
control_files.put_utf8(file, content)
1908
control_files.put_utf8('shared-storage', '')
1910
control_files.unlock()
1913
# formats which have no format string are not discoverable
1914
# and not independently creatable, so are not registered. They're
1915
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1916
# needed, it's constructed directly by the BzrDir. Non-native formats where
1917
# the repository is not separately opened are similar.
1919
format_registry.register_lazy(
1920
'Bazaar-NG Repository format 7',
1921
'bzrlib.repofmt.weaverepo',
1925
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1926
# default control directory format
1927
format_registry.register_lazy(
1928
'Bazaar-NG Knit Repository Format 1',
1929
'bzrlib.repofmt.knitrepo',
1930
'RepositoryFormatKnit1',
1932
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1934
format_registry.register_lazy(
1935
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1936
'bzrlib.repofmt.knitrepo',
1937
'RepositoryFormatKnit3',
1941
class InterRepository(InterObject):
1942
"""This class represents operations taking place between two repositories.
1944
Its instances have methods like copy_content and fetch, and contain
1945
references to the source and target repositories these operations can be
1948
Often we will provide convenience methods on 'repository' which carry out
1949
operations with another repository - they will always forward to
1950
InterRepository.get(other).method_name(parameters).
1954
"""The available optimised InterRepository types."""
1956
def copy_content(self, revision_id=None):
1957
raise NotImplementedError(self.copy_content)
1959
def fetch(self, revision_id=None, pb=None):
1960
"""Fetch the content required to construct revision_id.
1962
The content is copied from self.source to self.target.
1964
:param revision_id: if None all content is copied, if NULL_REVISION no
1966
:param pb: optional progress bar to use for progress reports. If not
1967
provided a default one will be created.
1969
Returns the copied revision count and the failed revisions in a tuple:
1972
raise NotImplementedError(self.fetch)
1975
def missing_revision_ids(self, revision_id=None):
1976
"""Return the revision ids that source has that target does not.
1978
These are returned in topological order.
1980
:param revision_id: only return revision ids included by this
1983
# generic, possibly worst case, slow code path.
1984
target_ids = set(self.target.all_revision_ids())
1985
if revision_id is not None:
1986
source_ids = self.source.get_ancestry(revision_id)
1987
assert source_ids[0] is None
1990
source_ids = self.source.all_revision_ids()
1991
result_set = set(source_ids).difference(target_ids)
1992
# this may look like a no-op: its not. It preserves the ordering
1993
# other_ids had while only returning the members from other_ids
1994
# that we've decided we need.
1995
return [rev_id for rev_id in source_ids if rev_id in result_set]
1998
def _same_model(source, target):
1999
"""True if source and target have the same data representation."""
2000
if source.supports_rich_root() != target.supports_rich_root():
2002
if source._serializer != target._serializer:
2007
class InterSameDataRepository(InterRepository):
2008
"""Code for converting between repositories that represent the same data.
2010
Data format and model must match for this to work.
2014
def _get_repo_format_to_test(self):
2015
"""Repository format for testing with.
2017
InterSameData can pull from subtree to subtree and from non-subtree to
2018
non-subtree, so we test this with the richest repository format.
2020
from bzrlib.repofmt import knitrepo
2021
return knitrepo.RepositoryFormatKnit3()
2024
def is_compatible(source, target):
2025
return InterRepository._same_model(source, target)
2028
def copy_content(self, revision_id=None):
2029
"""Make a complete copy of the content in self into destination.
2031
This copies both the repository's revision data, and configuration information
2032
such as the make_working_trees setting.
2034
This is a destructive operation! Do not use it on existing
2037
:param revision_id: Only copy the content needed to construct
2038
revision_id and its parents.
2041
self.target.set_make_working_trees(self.source.make_working_trees())
2042
except NotImplementedError:
2044
# but don't bother fetching if we have the needed data now.
2045
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2046
self.target.has_revision(revision_id)):
2048
self.target.fetch(self.source, revision_id=revision_id)
2051
def fetch(self, revision_id=None, pb=None):
2052
"""See InterRepository.fetch()."""
2053
from bzrlib.fetch import GenericRepoFetcher
2054
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2055
self.source, self.source._format, self.target,
2056
self.target._format)
2057
f = GenericRepoFetcher(to_repository=self.target,
2058
from_repository=self.source,
2059
last_revision=revision_id,
2061
return f.count_copied, f.failed_revisions
2064
class InterWeaveRepo(InterSameDataRepository):
2065
"""Optimised code paths between Weave based repositories.
2067
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2068
implemented lazy inter-object optimisation.
2072
def _get_repo_format_to_test(self):
2073
from bzrlib.repofmt import weaverepo
2074
return weaverepo.RepositoryFormat7()
2077
def is_compatible(source, target):
2078
"""Be compatible with known Weave formats.
2080
We don't test for the stores being of specific types because that
2081
could lead to confusing results, and there is no need to be
2084
from bzrlib.repofmt.weaverepo import (
2090
return (isinstance(source._format, (RepositoryFormat5,
2092
RepositoryFormat7)) and
2093
isinstance(target._format, (RepositoryFormat5,
2095
RepositoryFormat7)))
2096
except AttributeError:
2100
def copy_content(self, revision_id=None):
2101
"""See InterRepository.copy_content()."""
2102
# weave specific optimised path:
2104
self.target.set_make_working_trees(self.source.make_working_trees())
2105
except NotImplementedError:
2107
# FIXME do not peek!
2108
if self.source.control_files._transport.listable():
2109
pb = ui.ui_factory.nested_progress_bar()
2111
self.target.weave_store.copy_all_ids(
2112
self.source.weave_store,
2114
from_transaction=self.source.get_transaction(),
2115
to_transaction=self.target.get_transaction())
2116
pb.update('copying inventory', 0, 1)
2117
self.target.control_weaves.copy_multi(
2118
self.source.control_weaves, ['inventory'],
2119
from_transaction=self.source.get_transaction(),
2120
to_transaction=self.target.get_transaction())
2121
self.target._revision_store.text_store.copy_all_ids(
2122
self.source._revision_store.text_store,
2127
self.target.fetch(self.source, revision_id=revision_id)
2130
def fetch(self, revision_id=None, pb=None):
2131
"""See InterRepository.fetch()."""
2132
from bzrlib.fetch import GenericRepoFetcher
2133
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2134
self.source, self.source._format, self.target, self.target._format)
2135
f = GenericRepoFetcher(to_repository=self.target,
2136
from_repository=self.source,
2137
last_revision=revision_id,
2139
return f.count_copied, f.failed_revisions
2142
def missing_revision_ids(self, revision_id=None):
2143
"""See InterRepository.missing_revision_ids()."""
2144
# we want all revisions to satisfy revision_id in source.
2145
# but we don't want to stat every file here and there.
2146
# we want then, all revisions other needs to satisfy revision_id
2147
# checked, but not those that we have locally.
2148
# so the first thing is to get a subset of the revisions to
2149
# satisfy revision_id in source, and then eliminate those that
2150
# we do already have.
2151
# this is slow on high latency connection to self, but as as this
2152
# disk format scales terribly for push anyway due to rewriting
2153
# inventory.weave, this is considered acceptable.
2155
if revision_id is not None:
2156
source_ids = self.source.get_ancestry(revision_id)
2157
assert source_ids[0] is None
2160
source_ids = self.source._all_possible_ids()
2161
source_ids_set = set(source_ids)
2162
# source_ids is the worst possible case we may need to pull.
2163
# now we want to filter source_ids against what we actually
2164
# have in target, but don't try to check for existence where we know
2165
# we do not have a revision as that would be pointless.
2166
target_ids = set(self.target._all_possible_ids())
2167
possibly_present_revisions = target_ids.intersection(source_ids_set)
2168
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2169
required_revisions = source_ids_set.difference(actually_present_revisions)
2170
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2171
if revision_id is not None:
2172
# we used get_ancestry to determine source_ids then we are assured all
2173
# revisions referenced are present as they are installed in topological order.
2174
# and the tip revision was validated by get_ancestry.
2175
return required_topo_revisions
2177
# if we just grabbed the possibly available ids, then
2178
# we only have an estimate of whats available and need to validate
2179
# that against the revision records.
2180
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2183
class InterKnitRepo(InterSameDataRepository):
2184
"""Optimised code paths between Knit based repositories."""
2187
def _get_repo_format_to_test(self):
2188
from bzrlib.repofmt import knitrepo
2189
return knitrepo.RepositoryFormatKnit1()
2192
def is_compatible(source, target):
2193
"""Be compatible with known Knit formats.
2195
We don't test for the stores being of specific types because that
2196
could lead to confusing results, and there is no need to be
2199
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit
2201
are_knits = (isinstance(source._format, RepositoryFormatKnit) and
2202
isinstance(target._format, RepositoryFormatKnit))
2203
except AttributeError:
2205
return are_knits and InterRepository._same_model(source, target)
2208
def fetch(self, revision_id=None, pb=None):
2209
"""See InterRepository.fetch()."""
2210
from bzrlib.fetch import KnitRepoFetcher
2211
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2212
self.source, self.source._format, self.target, self.target._format)
2213
f = KnitRepoFetcher(to_repository=self.target,
2214
from_repository=self.source,
2215
last_revision=revision_id,
2217
return f.count_copied, f.failed_revisions
2220
def missing_revision_ids(self, revision_id=None):
2221
"""See InterRepository.missing_revision_ids()."""
2222
if revision_id is not None:
2223
source_ids = self.source.get_ancestry(revision_id)
2224
assert source_ids[0] is None
2227
source_ids = self.source.all_revision_ids()
2228
source_ids_set = set(source_ids)
2229
# source_ids is the worst possible case we may need to pull.
2230
# now we want to filter source_ids against what we actually
2231
# have in target, but don't try to check for existence where we know
2232
# we do not have a revision as that would be pointless.
2233
target_ids = set(self.target.all_revision_ids())
2234
possibly_present_revisions = target_ids.intersection(source_ids_set)
2235
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2236
required_revisions = source_ids_set.difference(actually_present_revisions)
2237
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2238
if revision_id is not None:
2239
# we used get_ancestry to determine source_ids then we are assured all
2240
# revisions referenced are present as they are installed in topological order.
2241
# and the tip revision was validated by get_ancestry.
2242
return required_topo_revisions
2244
# if we just grabbed the possibly available ids, then
2245
# we only have an estimate of whats available and need to validate
2246
# that against the revision records.
2247
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2250
class InterModel1and2(InterRepository):
2253
def _get_repo_format_to_test(self):
2257
def is_compatible(source, target):
2258
if not source.supports_rich_root() and target.supports_rich_root():
2264
def fetch(self, revision_id=None, pb=None):
2265
"""See InterRepository.fetch()."""
2266
from bzrlib.fetch import Model1toKnit2Fetcher
2267
f = Model1toKnit2Fetcher(to_repository=self.target,
2268
from_repository=self.source,
2269
last_revision=revision_id,
2271
return f.count_copied, f.failed_revisions
2274
def copy_content(self, revision_id=None):
2275
"""Make a complete copy of the content in self into destination.
2277
This is a destructive operation! Do not use it on existing
2280
:param revision_id: Only copy the content needed to construct
2281
revision_id and its parents.
2284
self.target.set_make_working_trees(self.source.make_working_trees())
2285
except NotImplementedError:
2287
# but don't bother fetching if we have the needed data now.
2288
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2289
self.target.has_revision(revision_id)):
2291
self.target.fetch(self.source, revision_id=revision_id)
2294
class InterKnit1and2(InterKnitRepo):
2297
def _get_repo_format_to_test(self):
2301
def is_compatible(source, target):
2302
"""Be compatible with Knit1 source and Knit3 target"""
2303
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2305
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
2306
RepositoryFormatKnit3
2307
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2308
isinstance(target._format, (RepositoryFormatKnit3)))
2309
except AttributeError:
2313
def fetch(self, revision_id=None, pb=None):
2314
"""See InterRepository.fetch()."""
2315
from bzrlib.fetch import Knit1to2Fetcher
2316
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2317
self.source, self.source._format, self.target,
2318
self.target._format)
2319
f = Knit1to2Fetcher(to_repository=self.target,
2320
from_repository=self.source,
2321
last_revision=revision_id,
2323
return f.count_copied, f.failed_revisions
2326
class InterRemoteToOther(InterRepository):
2328
def __init__(self, source, target):
2329
InterRepository.__init__(self, source, target)
2330
self._real_inter = None
2333
def is_compatible(source, target):
2334
if not isinstance(source, remote.RemoteRepository):
2336
source._ensure_real()
2337
real_source = source._real_repository
2338
# Is source's model compatible with target's model, and are they the
2339
# same format? Currently we can only optimise fetching from an
2340
# identical model & format repo.
2341
assert not isinstance(real_source, remote.RemoteRepository), (
2342
"We don't support remote repos backed by remote repos yet.")
2343
return real_source._format == target._format
2346
def fetch(self, revision_id=None, pb=None):
2347
"""See InterRepository.fetch()."""
2348
from bzrlib.fetch import RemoteToOtherFetcher
2349
mutter("Using fetch logic to copy between %s(remote) and %s(%s)",
2350
self.source, self.target, self.target._format)
2351
# TODO: jam 20070210 This should be an assert, not a translate
2352
revision_id = osutils.safe_revision_id(revision_id)
2353
f = RemoteToOtherFetcher(to_repository=self.target,
2354
from_repository=self.source,
2355
last_revision=revision_id,
2357
return f.count_copied, f.failed_revisions
2360
def _get_repo_format_to_test(self):
2364
class InterOtherToRemote(InterRepository):
2366
def __init__(self, source, target):
2367
InterRepository.__init__(self, source, target)
2368
self._real_inter = None
2371
def is_compatible(source, target):
2372
if isinstance(target, remote.RemoteRepository):
2376
def _ensure_real_inter(self):
2377
if self._real_inter is None:
2378
self.target._ensure_real()
2379
real_target = self.target._real_repository
2380
self._real_inter = InterRepository.get(self.source, real_target)
2382
def copy_content(self, revision_id=None):
2383
self._ensure_real_inter()
2384
self._real_inter.copy_content(revision_id=revision_id)
2386
def fetch(self, revision_id=None, pb=None):
2387
self._ensure_real_inter()
2388
self._real_inter.fetch(revision_id=revision_id, pb=pb)
2391
def _get_repo_format_to_test(self):
2395
InterRepository.register_optimiser(InterSameDataRepository)
2396
InterRepository.register_optimiser(InterWeaveRepo)
2397
InterRepository.register_optimiser(InterKnitRepo)
2398
InterRepository.register_optimiser(InterModel1and2)
2399
InterRepository.register_optimiser(InterKnit1and2)
2400
InterRepository.register_optimiser(InterRemoteToOther)
2401
InterRepository.register_optimiser(InterOtherToRemote)
2404
class CopyConverter(object):
2405
"""A repository conversion tool which just performs a copy of the content.
2407
This is slow but quite reliable.
2410
def __init__(self, target_format):
2411
"""Create a CopyConverter.
2413
:param target_format: The format the resulting repository should be.
2415
self.target_format = target_format
2417
def convert(self, repo, pb):
2418
"""Perform the conversion of to_convert, giving feedback via pb.
2420
:param to_convert: The disk object to convert.
2421
:param pb: a progress bar to use for progress information.
2426
# this is only useful with metadir layouts - separated repo content.
2427
# trigger an assertion if not such
2428
repo._format.get_format_string()
2429
self.repo_dir = repo.bzrdir
2430
self.step('Moving repository to repository.backup')
2431
self.repo_dir.transport.move('repository', 'repository.backup')
2432
backup_transport = self.repo_dir.transport.clone('repository.backup')
2433
repo._format.check_conversion_target(self.target_format)
2434
self.source_repo = repo._format.open(self.repo_dir,
2436
_override_transport=backup_transport)
2437
self.step('Creating new repository')
2438
converted = self.target_format.initialize(self.repo_dir,
2439
self.source_repo.is_shared())
2440
converted.lock_write()
2442
self.step('Copying content into repository.')
2443
self.source_repo.copy_content_into(converted)
2446
self.step('Deleting old repository content.')
2447
self.repo_dir.transport.delete_tree('repository.backup')
2448
self.pb.note('repository converted')
2450
def step(self, message):
2451
"""Update the pb by a step."""
2453
self.pb.update(message, self.count, self.total)
2465
def _unescaper(match, _map=_unescape_map):
2466
code = match.group(1)
2470
if not code.startswith('#'):
2472
return unichr(int(code[1:])).encode('utf8')
2478
def _unescape_xml(data):
2479
"""Unescape predefined XML entities in a string of data."""
2481
if _unescape_re is None:
2482
_unescape_re = re.compile('\&([^;]*);')
2483
return _unescape_re.sub(_unescaper, data)
2486
class _RevisionTextVersionCache(object):
2487
"""A cache of the versionedfile versions for revision and file-id."""
2489
def __init__(self, repository):
2490
self.repository = repository
2491
self.revision_versions = {}
2492
self.revision_parents = {}
2493
self.repo_graph = self.repository.get_graph()
2496
def add_revision_text_versions(self, tree):
2497
"""Cache text version data from the supplied revision tree"""
2499
for path, entry in tree.iter_entries_by_dir():
2500
inv_revisions[entry.file_id] = entry.revision
2501
self.revision_versions[tree.get_revision_id()] = inv_revisions
2502
return inv_revisions
2504
def get_text_version(self, file_id, revision_id):
2505
"""Determine the text version for a given file-id and revision-id"""
2507
inv_revisions = self.revision_versions[revision_id]
2510
tree = self.repository.revision_tree(revision_id)
2511
except errors.RevisionNotPresent:
2512
self.revision_versions[revision_id] = inv_revisions = {}
2514
inv_revisions = self.add_revision_text_versions(tree)
2515
return inv_revisions.get(file_id)
2517
def prepopulate_revs(self, revision_ids):
2518
# Filter out versions that we don't have an inventory for, so that the
2519
# revision_trees() call won't fail.
2520
inv_weave = self.repository.get_inventory_weave()
2521
revs = [r for r in revision_ids if inv_weave.has_version(r)]
2522
# XXX: this loop is very similar to
2523
# bzrlib.fetch.Inter1and2Helper.iter_rev_trees.
2525
for tree in self.repository.revision_trees(revs[:100]):
2526
if tree.inventory.revision_id is None:
2527
tree.inventory.revision_id = tree.get_revision_id()
2528
self.add_revision_text_versions(tree)
2531
def get_parents(self, revision_id):
2533
return self.revision_parents[revision_id]
2535
parents = self.repository.get_parents([revision_id])[0]
2536
self.revision_parents[revision_id] = parents
2539
def heads(self, revision_ids):
2540
revision_ids = tuple(revision_ids)
2542
return self.rev_heads[revision_ids]
2544
heads = self.repo_graph.heads(revision_ids)
2545
self.rev_heads[revision_ids] = heads
2548
class VersionedFileChecker(object):
2550
def __init__(self, planned_revisions, revision_versions, repository):
2551
self.planned_revisions = planned_revisions
2552
self.revision_versions = revision_versions
2553
self.repository = repository
2555
def calculate_file_version_parents(self, revision_id, file_id):
2556
text_revision = self.revision_versions.get_text_version(
2557
file_id, revision_id)
2558
if text_revision is None:
2560
parents_of_text_revision = self.revision_versions.get_parents(
2562
parents_from_inventories = []
2563
for parent in parents_of_text_revision:
2564
if parent == _mod_revision.NULL_REVISION:
2566
introduced_in = self.revision_versions.get_text_version(file_id,
2568
if introduced_in is not None:
2569
parents_from_inventories.append(introduced_in)
2570
heads = set(self.revision_versions.heads(parents_from_inventories))
2572
for parent in parents_from_inventories:
2573
if parent in heads and parent not in new_parents:
2574
new_parents.append(parent)
2577
def check_file_version_parents(self, weave, file_id):
2579
for num, revision_id in enumerate(self.planned_revisions):
2580
correct_parents = self.calculate_file_version_parents(
2581
revision_id, file_id)
2582
if correct_parents is None:
2584
text_revision = self.revision_versions.get_text_version(
2585
file_id, revision_id)
2586
knit_parents = weave.get_parents(text_revision)
2587
if correct_parents != knit_parents:
2588
result[revision_id] = (knit_parents, correct_parents)
added
removed
bzrlib/repository.py
