1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
from cStringIO import StringIO
19
from bzrlib.lazy_import import lazy_import
20
lazy_import(globals(), """
39
revision as _mod_revision,
44
from bzrlib.bundle import serializer
45
from bzrlib.revisiontree import RevisionTree
46
from bzrlib.store.versioned import VersionedFileStore
47
from bzrlib.store.text import TextStore
48
from bzrlib.testament import Testament
51
from bzrlib.decorators import needs_read_lock, needs_write_lock
52
from bzrlib.inter import InterObject
53
from bzrlib.inventory import Inventory, InventoryDirectory, ROOT_ID
54
from bzrlib.symbol_versioning import (
57
from bzrlib.trace import mutter, mutter_callsite, note, warning
60
# Old formats display a warning, but only once
61
_deprecation_warning_done = False
64
class CommitBuilder(object):
65
"""Provides an interface to build up a commit.
67
This allows describing a tree to be committed without needing to
68
know the internals of the format of the repository.
71
# all clients should supply tree roots.
72
record_root_entry = True
73
# the default CommitBuilder does not manage trees whose root is versioned.
74
_versioned_root = False
76
def __init__(self, repository, parents, config, timestamp=None,
77
timezone=None, committer=None, revprops=None,
79
"""Initiate a CommitBuilder.
81
:param repository: Repository to commit to.
82
:param parents: Revision ids of the parents of the new revision.
83
:param config: Configuration to use.
84
:param timestamp: Optional timestamp recorded for commit.
85
:param timezone: Optional timezone for timestamp.
86
:param committer: Optional committer to set for commit.
87
:param revprops: Optional dictionary of revision properties.
88
:param revision_id: Optional revision id.
93
self._committer = self._config.username()
95
assert isinstance(committer, basestring), type(committer)
96
self._committer = committer
98
self.new_inventory = Inventory(None)
99
self._new_revision_id = osutils.safe_revision_id(revision_id)
100
self.parents = parents
101
self.repository = repository
104
if revprops is not None:
105
self._revprops.update(revprops)
107
if timestamp is None:
108
timestamp = time.time()
109
# Restrict resolution to 1ms
110
self._timestamp = round(timestamp, 3)
113
self._timezone = osutils.local_time_offset()
115
self._timezone = int(timezone)
117
self._generate_revision_if_needed()
119
def commit(self, message):
120
"""Make the actual commit.
122
:return: The revision id of the recorded revision.
124
rev = _mod_revision.Revision(
125
timestamp=self._timestamp,
126
timezone=self._timezone,
127
committer=self._committer,
129
inventory_sha1=self.inv_sha1,
130
revision_id=self._new_revision_id,
131
properties=self._revprops)
132
rev.parent_ids = self.parents
133
self.repository.add_revision(self._new_revision_id, rev,
134
self.new_inventory, self._config)
135
self.repository.commit_write_group()
136
return self._new_revision_id
139
"""Abort the commit that is being built.
141
self.repository.abort_write_group()
143
def revision_tree(self):
144
"""Return the tree that was just committed.
146
After calling commit() this can be called to get a RevisionTree
147
representing the newly committed tree. This is preferred to
148
calling Repository.revision_tree() because that may require
149
deserializing the inventory, while we already have a copy in
152
return RevisionTree(self.repository, self.new_inventory,
153
self._new_revision_id)
155
def finish_inventory(self):
156
"""Tell the builder that the inventory is finished."""
157
if self.new_inventory.root is None:
158
symbol_versioning.warn('Root entry should be supplied to'
159
' record_entry_contents, as of bzr 0.10.',
160
DeprecationWarning, stacklevel=2)
161
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
162
self.new_inventory.revision_id = self._new_revision_id
163
self.inv_sha1 = self.repository.add_inventory(
164
self._new_revision_id,
169
def _gen_revision_id(self):
170
"""Return new revision-id."""
171
return generate_ids.gen_revision_id(self._config.username(),
174
def _generate_revision_if_needed(self):
175
"""Create a revision id if None was supplied.
177
If the repository can not support user-specified revision ids
178
they should override this function and raise CannotSetRevisionId
179
if _new_revision_id is not None.
181
:raises: CannotSetRevisionId
183
if self._new_revision_id is None:
184
self._new_revision_id = self._gen_revision_id()
185
self.random_revid = True
187
self.random_revid = False
189
def _check_root(self, ie, parent_invs, tree):
190
"""Helper for record_entry_contents.
192
:param ie: An entry being added.
193
:param parent_invs: The inventories of the parent revisions of the
195
:param tree: The tree that is being committed.
197
# In this revision format, root entries have no knit or weave When
198
# serializing out to disk and back in root.revision is always
200
ie.revision = self._new_revision_id
202
def _get_delta(self, ie, basis_inv, path):
203
"""Get a delta against the basis inventory for ie."""
204
if ie.file_id not in basis_inv:
206
return (None, path, ie.file_id, ie)
207
elif ie != basis_inv[ie.file_id]:
209
# TODO: avoid tis id2path call.
210
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
215
def record_entry_contents(self, ie, parent_invs, path, tree,
217
"""Record the content of ie from tree into the commit if needed.
219
Side effect: sets ie.revision when unchanged
221
:param ie: An inventory entry present in the commit.
222
:param parent_invs: The inventories of the parent revisions of the
224
:param path: The path the entry is at in the tree.
225
:param tree: The tree which contains this entry and should be used to
227
:param content_summary: Summary data from the tree about the paths
228
content - stat, length, exec, sha/link target. This is only
229
accessed when the entry has a revision of None - that is when it is
230
a candidate to commit.
231
:return: A tuple (change_delta, version_recorded). change_delta is
232
an inventory_delta change for this entry against the basis tree of
233
the commit, or None if no change occured against the basis tree.
234
version_recorded is True if a new version of the entry has been
235
recorded. For instance, committing a merge where a file was only
236
changed on the other side will return (delta, False).
238
if self.new_inventory.root is None:
239
if ie.parent_id is not None:
240
raise errors.RootMissing()
241
self._check_root(ie, parent_invs, tree)
242
if ie.revision is None:
243
kind = content_summary[0]
245
# ie is carried over from a prior commit
247
# XXX: repository specific check for nested tree support goes here - if
248
# the repo doesn't want nested trees we skip it ?
249
if (kind == 'tree-reference' and
250
not self.repository._format.supports_tree_reference):
251
# mismatch between commit builder logic and repository:
252
# this needs the entry creation pushed down into the builder.
253
raise NotImplementedError('Missing repository subtree support.')
254
# transitional assert only, will remove before release.
255
assert ie.kind == kind
256
self.new_inventory.add(ie)
258
# TODO: slow, take it out of the inner loop.
260
basis_inv = parent_invs[0]
262
basis_inv = Inventory(root_id=None)
264
# ie.revision is always None if the InventoryEntry is considered
265
# for committing. We may record the previous parents revision if the
266
# content is actually unchanged against a sole head.
267
if ie.revision is not None:
268
if self._versioned_root or path != '':
269
# not considered for commit
272
# repositories that do not version the root set the root's
273
# revision to the new commit even when no change occurs, and
274
# this masks when a change may have occurred against the basis,
275
# so calculate if one happened.
276
if ie.file_id not in basis_inv:
278
delta = (None, path, ie.file_id, ie)
280
basis_id = basis_inv[ie.file_id]
281
if basis_id.name != '':
283
delta = (basis_inv.id2path(ie.file_id), path,
288
# not considered for commit, OR, for non-rich-root
289
return delta, ie.revision == self._new_revision_id and (path != '' or
290
self._versioned_root)
292
# XXX: Friction: parent_candidates should return a list not a dict
293
# so that we don't have to walk the inventories again.
294
parent_candiate_entries = ie.parent_candidates(parent_invs)
295
head_set = self.repository.get_graph().heads(parent_candiate_entries.keys())
297
for inv in parent_invs:
298
if ie.file_id in inv:
299
old_rev = inv[ie.file_id].revision
300
if old_rev in head_set:
301
heads.append(inv[ie.file_id].revision)
302
head_set.remove(inv[ie.file_id].revision)
305
# now we check to see if we need to write a new record to the
307
# We write a new entry unless there is one head to the ancestors, and
308
# the kind-derived content is unchanged.
310
# Cheapest check first: no ancestors, or more the one head in the
311
# ancestors, we write a new node.
315
# There is a single head, look it up for comparison
316
parent_entry = parent_candiate_entries[heads[0]]
317
# if the non-content specific data has changed, we'll be writing a
319
if (parent_entry.parent_id != ie.parent_id or
320
parent_entry.name != ie.name):
322
# now we need to do content specific checks:
324
# if the kind changed the content obviously has
325
if kind != parent_entry.kind:
329
if (# if the file length changed we have to store:
330
parent_entry.text_size != content_summary[1] or
331
# if the exec bit has changed we have to store:
332
parent_entry.executable != content_summary[2]):
334
elif parent_entry.text_sha1 == content_summary[3]:
335
# all meta and content is unchanged (using a hash cache
336
# hit to check the sha)
337
ie.revision = parent_entry.revision
338
ie.text_size = parent_entry.text_size
339
ie.text_sha1 = parent_entry.text_sha1
340
ie.executable = parent_entry.executable
341
return self._get_delta(ie, basis_inv, path), False
343
# Either there is only a hash change(no hash cache entry,
344
# or same size content change), or there is no change on
346
# Provide the parent's hash to the store layer, so that the
347
# content is unchanged we will not store a new node.
348
nostore_sha = parent_entry.text_sha1
350
# We want to record a new node regardless of the presence or
351
# absence of a content change in the file.
353
ie.executable = content_summary[2]
354
lines = tree.get_file(ie.file_id, path).readlines()
356
ie.text_sha1, ie.text_size = self._add_text_to_weave(
357
ie.file_id, lines, heads, nostore_sha)
358
except errors.ExistingContent:
359
# Turns out that the file content was unchanged, and we were
360
# only going to store a new node if it was changed. Carry over
362
ie.revision = parent_entry.revision
363
ie.text_size = parent_entry.text_size
364
ie.text_sha1 = parent_entry.text_sha1
365
ie.executable = parent_entry.executable
366
return self._get_delta(ie, basis_inv, path), False
367
elif kind == 'directory':
369
# all data is meta here, nothing specific to directory, so
371
ie.revision = parent_entry.revision
372
return self._get_delta(ie, basis_inv, path), False
374
self._add_text_to_weave(ie.file_id, lines, heads, None)
375
elif kind == 'symlink':
376
current_link_target = content_summary[3]
378
# symlink target is not generic metadata, check if it has
380
if current_link_target != parent_entry.symlink_target:
383
# unchanged, carry over.
384
ie.revision = parent_entry.revision
385
ie.symlink_target = parent_entry.symlink_target
386
return self._get_delta(ie, basis_inv, path), False
387
ie.symlink_target = current_link_target
389
self._add_text_to_weave(ie.file_id, lines, heads, None)
390
elif kind == 'tree-reference':
392
if content_summary[3] != parent_entry.reference_revision:
395
# unchanged, carry over.
396
ie.reference_revision = parent_entry.reference_revision
397
ie.revision = parent_entry.revision
398
return self._get_delta(ie, basis_inv, path), False
399
ie.reference_revision = content_summary[3]
401
self._add_text_to_weave(ie.file_id, lines, heads, None)
403
raise NotImplementedError('unknown kind')
404
ie.revision = self._new_revision_id
405
return self._get_delta(ie, basis_inv, path), True
407
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
408
versionedfile = self.repository.weave_store.get_weave_or_empty(
409
file_id, self.repository.get_transaction())
410
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
411
# than add_lines, and allows committing when a parent is ghosted for
413
# Note: as we read the content directly from the tree, we know its not
414
# been turned into unicode or badly split - but a broken tree
415
# implementation could give us bad output from readlines() so this is
416
# not a guarantee of safety. What would be better is always checking
417
# the content during test suite execution. RBC 20070912
419
return versionedfile.add_lines_with_ghosts(
420
self._new_revision_id, parents, new_lines,
421
nostore_sha=nostore_sha, random_id=self.random_revid,
422
check_content=False)[0:2]
424
versionedfile.clear_cache()
427
class RootCommitBuilder(CommitBuilder):
428
"""This commitbuilder actually records the root id"""
430
# the root entry gets versioned properly by this builder.
431
_versioned_root = True
433
def _check_root(self, ie, parent_invs, tree):
434
"""Helper for record_entry_contents.
436
:param ie: An entry being added.
437
:param parent_invs: The inventories of the parent revisions of the
439
:param tree: The tree that is being committed.
443
######################################################################
446
class Repository(object):
447
"""Repository holding history for one or more branches.
449
The repository holds and retrieves historical information including
450
revisions and file history. It's normally accessed only by the Branch,
451
which views a particular line of development through that history.
453
The Repository builds on top of Stores and a Transport, which respectively
454
describe the disk data format and the way of accessing the (possibly
458
# What class to use for a CommitBuilder. Often its simpler to change this
459
# in a Repository class subclass rather than to override
460
# get_commit_builder.
461
_commit_builder_class = CommitBuilder
462
# The search regex used by xml based repositories to determine what things
463
# where changed in a single commit.
464
_file_ids_altered_regex = lazy_regex.lazy_compile(
465
r'file_id="(?P<file_id>[^"]+)"'
466
r'.* revision="(?P<revision_id>[^"]+)"'
469
def abort_write_group(self):
470
"""Commit the contents accrued within the current write group.
472
:seealso: start_write_group.
474
if self._write_group is not self.get_transaction():
475
# has an unlock or relock occured ?
476
raise errors.BzrError('mismatched lock context and write group.')
477
self._abort_write_group()
478
self._write_group = None
480
def _abort_write_group(self):
481
"""Template method for per-repository write group cleanup.
483
This is called during abort before the write group is considered to be
484
finished and should cleanup any internal state accrued during the write
485
group. There is no requirement that data handed to the repository be
486
*not* made available - this is not a rollback - but neither should any
487
attempt be made to ensure that data added is fully commited. Abort is
488
invoked when an error has occured so futher disk or network operations
489
may not be possible or may error and if possible should not be
494
def add_inventory(self, revision_id, inv, parents):
495
"""Add the inventory inv to the repository as revision_id.
497
:param parents: The revision ids of the parents that revision_id
498
is known to have and are in the repository already.
500
returns the sha1 of the serialized inventory.
502
revision_id = osutils.safe_revision_id(revision_id)
503
_mod_revision.check_not_reserved_id(revision_id)
504
assert inv.revision_id is None or inv.revision_id == revision_id, \
505
"Mismatch between inventory revision" \
506
" id and insertion revid (%r, %r)" % (inv.revision_id, revision_id)
507
assert inv.root is not None
508
inv_lines = self._serialise_inventory_to_lines(inv)
509
inv_vf = self.get_inventory_weave()
510
return self._inventory_add_lines(inv_vf, revision_id, parents,
511
inv_lines, check_content=False)
513
def _inventory_add_lines(self, inv_vf, revision_id, parents, lines,
515
"""Store lines in inv_vf and return the sha1 of the inventory."""
517
for parent in parents:
519
final_parents.append(parent)
520
return inv_vf.add_lines(revision_id, final_parents, lines,
521
check_content=check_content)[0]
524
def add_revision(self, revision_id, rev, inv=None, config=None):
525
"""Add rev to the revision store as revision_id.
527
:param revision_id: the revision id to use.
528
:param rev: The revision object.
529
:param inv: The inventory for the revision. if None, it will be looked
530
up in the inventory storer
531
:param config: If None no digital signature will be created.
532
If supplied its signature_needed method will be used
533
to determine if a signature should be made.
535
revision_id = osutils.safe_revision_id(revision_id)
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
635
return '%s(%r)' % (self.__class__.__name__,
636
self.bzrdir.transport.base)
638
def has_same_location(self, other):
639
"""Returns a boolean indicating if this repository is at the same
640
location as another repository.
642
This might return False even when two repository objects are accessing
643
the same physical repository via different URLs.
645
if self.__class__ is not other.__class__:
647
return (self.control_files._transport.base ==
648
other.control_files._transport.base)
650
def is_in_write_group(self):
651
"""Return True if there is an open write group.
653
:seealso: start_write_group.
655
return self._write_group is not None
658
return self.control_files.is_locked()
660
def lock_write(self, token=None):
661
"""Lock this repository for writing.
663
This causes caching within the repository obejct to start accumlating
664
data during reads, and allows a 'write_group' to be obtained. Write
665
groups must be used for actual data insertion.
667
:param token: if this is already locked, then lock_write will fail
668
unless the token matches the existing lock.
669
:returns: a token if this instance supports tokens, otherwise None.
670
:raises TokenLockingNotSupported: when a token is given but this
671
instance doesn't support using token locks.
672
:raises MismatchedToken: if the specified token doesn't match the token
673
of the existing lock.
674
:seealso: start_write_group.
676
A token should be passed in if you know that you have locked the object
677
some other way, and need to synchronise this object's state with that
680
XXX: this docstring is duplicated in many places, e.g. lockable_files.py
682
result = self.control_files.lock_write(token=token)
687
self.control_files.lock_read()
690
def get_physical_lock_status(self):
691
return self.control_files.get_physical_lock_status()
693
def leave_lock_in_place(self):
694
"""Tell this repository not to release the physical lock when this
697
If lock_write doesn't return a token, then this method is not supported.
699
self.control_files.leave_in_place()
701
def dont_leave_lock_in_place(self):
702
"""Tell this repository to release the physical lock when this
703
object is unlocked, even if it didn't originally acquire it.
705
If lock_write doesn't return a token, then this method is not supported.
707
self.control_files.dont_leave_in_place()
710
def gather_stats(self, revid=None, committers=None):
711
"""Gather statistics from a revision id.
713
:param revid: The revision id to gather statistics from, if None, then
714
no revision specific statistics are gathered.
715
:param committers: Optional parameter controlling whether to grab
716
a count of committers from the revision specific statistics.
717
:return: A dictionary of statistics. Currently this contains:
718
committers: The number of committers if requested.
719
firstrev: A tuple with timestamp, timezone for the penultimate left
720
most ancestor of revid, if revid is not the NULL_REVISION.
721
latestrev: A tuple with timestamp, timezone for revid, if revid is
722
not the NULL_REVISION.
723
revisions: The total revision count in the repository.
724
size: An estimate disk size of the repository in bytes.
727
if revid and committers:
728
result['committers'] = 0
729
if revid and revid != _mod_revision.NULL_REVISION:
731
all_committers = set()
732
revisions = self.get_ancestry(revid)
733
# pop the leading None
735
first_revision = None
737
# ignore the revisions in the middle - just grab first and last
738
revisions = revisions[0], revisions[-1]
739
for revision in self.get_revisions(revisions):
740
if not first_revision:
741
first_revision = revision
743
all_committers.add(revision.committer)
744
last_revision = revision
746
result['committers'] = len(all_committers)
747
result['firstrev'] = (first_revision.timestamp,
748
first_revision.timezone)
749
result['latestrev'] = (last_revision.timestamp,
750
last_revision.timezone)
752
# now gather global repository information
753
if self.bzrdir.root_transport.listable():
754
c, t = self._revision_store.total_size(self.get_transaction())
755
result['revisions'] = c
760
def missing_revision_ids(self, other, revision_id=None):
761
"""Return the revision ids that other has that this does not.
763
These are returned in topological order.
765
revision_id: only return revision ids included by revision_id.
767
revision_id = osutils.safe_revision_id(revision_id)
768
return InterRepository.get(other, self).missing_revision_ids(revision_id)
772
"""Open the repository rooted at base.
774
For instance, if the repository is at URL/.bzr/repository,
775
Repository.open(URL) -> a Repository instance.
777
control = bzrdir.BzrDir.open(base)
778
return control.open_repository()
780
def copy_content_into(self, destination, revision_id=None):
781
"""Make a complete copy of the content in self into destination.
783
This is a destructive operation! Do not use it on existing
786
revision_id = osutils.safe_revision_id(revision_id)
787
return InterRepository.get(self, destination).copy_content(revision_id)
789
def commit_write_group(self):
790
"""Commit the contents accrued within the current write group.
792
:seealso: start_write_group.
794
if self._write_group is not self.get_transaction():
795
# has an unlock or relock occured ?
796
raise errors.BzrError('mismatched lock context and write group.')
797
self._commit_write_group()
798
self._write_group = None
800
def _commit_write_group(self):
801
"""Template method for per-repository write group cleanup.
803
This is called before the write group is considered to be
804
finished and should ensure that all data handed to the repository
805
for writing during the write group is safely committed (to the
806
extent possible considering file system caching etc).
809
def fetch(self, source, revision_id=None, pb=None):
810
"""Fetch the content required to construct revision_id from source.
812
If revision_id is None all content is copied.
814
revision_id = osutils.safe_revision_id(revision_id)
815
inter = InterRepository.get(source, self)
817
return inter.fetch(revision_id=revision_id, pb=pb)
818
except NotImplementedError:
819
raise errors.IncompatibleRepositories(source, self)
821
def create_bundle(self, target, base, fileobj, format=None):
822
return serializer.write_bundle(self, target, base, fileobj, format)
824
def get_commit_builder(self, branch, parents, config, timestamp=None,
825
timezone=None, committer=None, revprops=None,
827
"""Obtain a CommitBuilder for this repository.
829
:param branch: Branch to commit to.
830
:param parents: Revision ids of the parents of the new revision.
831
:param config: Configuration to use.
832
:param timestamp: Optional timestamp recorded for commit.
833
:param timezone: Optional timezone for timestamp.
834
:param committer: Optional committer to set for commit.
835
:param revprops: Optional dictionary of revision properties.
836
:param revision_id: Optional revision id.
838
revision_id = osutils.safe_revision_id(revision_id)
839
result = self._commit_builder_class(self, parents, config,
840
timestamp, timezone, committer, revprops, revision_id)
841
self.start_write_group()
845
if (self.control_files._lock_count == 1 and
846
self.control_files._lock_mode == 'w'):
847
if self._write_group is not None:
848
raise errors.BzrError(
849
'Must end write groups before releasing write locks.')
850
self.control_files.unlock()
853
def clone(self, a_bzrdir, revision_id=None):
854
"""Clone this repository into a_bzrdir using the current format.
856
Currently no check is made that the format of this repository and
857
the bzrdir format are compatible. FIXME RBC 20060201.
859
:return: The newly created destination repository.
861
# TODO: deprecate after 0.16; cloning this with all its settings is
862
# probably not very useful -- mbp 20070423
863
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
864
self.copy_content_into(dest_repo, revision_id)
867
def start_write_group(self):
868
"""Start a write group in the repository.
870
Write groups are used by repositories which do not have a 1:1 mapping
871
between file ids and backend store to manage the insertion of data from
872
both fetch and commit operations.
874
A write lock is required around the start_write_group/commit_write_group
875
for the support of lock-requiring repository formats.
877
One can only insert data into a repository inside a write group.
881
if not self.is_locked() or self.control_files._lock_mode != 'w':
882
raise errors.NotWriteLocked(self)
883
if self._write_group:
884
raise errors.BzrError('already in a write group')
885
self._start_write_group()
886
# so we can detect unlock/relock - the write group is now entered.
887
self._write_group = self.get_transaction()
889
def _start_write_group(self):
890
"""Template method for per-repository write group startup.
892
This is called before the write group is considered to be
897
def sprout(self, to_bzrdir, revision_id=None):
898
"""Create a descendent repository for new development.
900
Unlike clone, this does not copy the settings of the repository.
902
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
903
dest_repo.fetch(self, revision_id=revision_id)
906
def _create_sprouting_repo(self, a_bzrdir, shared):
907
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
908
# use target default format.
909
dest_repo = a_bzrdir.create_repository()
911
# Most control formats need the repository to be specifically
912
# created, but on some old all-in-one formats it's not needed
914
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
915
except errors.UninitializableFormat:
916
dest_repo = a_bzrdir.open_repository()
920
def has_revision(self, revision_id):
921
"""True if this repository has a copy of the revision."""
922
if 'evil' in debug.debug_flags:
923
mutter_callsite(2, "has_revision is a LBYL symptom.")
924
revision_id = osutils.safe_revision_id(revision_id)
925
return self._revision_store.has_revision_id(revision_id,
926
self.get_transaction())
929
def get_revision(self, revision_id):
930
"""Return the Revision object for a named revision."""
931
return self.get_revisions([revision_id])[0]
934
def get_revision_reconcile(self, revision_id):
935
"""'reconcile' helper routine that allows access to a revision always.
937
This variant of get_revision does not cross check the weave graph
938
against the revision one as get_revision does: but it should only
939
be used by reconcile, or reconcile-alike commands that are correcting
940
or testing the revision graph.
942
return self._get_revisions([revision_id])[0]
945
def get_revisions(self, revision_ids):
946
"""Get many revisions at once."""
947
return self._get_revisions(revision_ids)
950
def _get_revisions(self, revision_ids):
951
"""Core work logic to get many revisions without sanity checks."""
952
revision_ids = [osutils.safe_revision_id(r) for r in revision_ids]
953
for rev_id in revision_ids:
954
if not rev_id or not isinstance(rev_id, basestring):
955
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
956
revs = self._revision_store.get_revisions(revision_ids,
957
self.get_transaction())
959
assert not isinstance(rev.revision_id, unicode)
960
for parent_id in rev.parent_ids:
961
assert not isinstance(parent_id, unicode)
965
def get_revision_xml(self, revision_id):
966
# TODO: jam 20070210 This shouldn't be necessary since get_revision
967
# would have already do it.
968
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
969
revision_id = osutils.safe_revision_id(revision_id)
970
rev = self.get_revision(revision_id)
972
# the current serializer..
973
self._revision_store._serializer.write_revision(rev, rev_tmp)
975
return rev_tmp.getvalue()
978
def get_deltas_for_revisions(self, revisions):
979
"""Produce a generator of revision deltas.
981
Note that the input is a sequence of REVISIONS, not revision_ids.
982
Trees will be held in memory until the generator exits.
983
Each delta is relative to the revision's lefthand predecessor.
985
required_trees = set()
986
for revision in revisions:
987
required_trees.add(revision.revision_id)
988
required_trees.update(revision.parent_ids[:1])
989
trees = dict((t.get_revision_id(), t) for
990
t in self.revision_trees(required_trees))
991
for revision in revisions:
992
if not revision.parent_ids:
993
old_tree = self.revision_tree(None)
995
old_tree = trees[revision.parent_ids[0]]
996
yield trees[revision.revision_id].changes_from(old_tree)
999
def get_revision_delta(self, revision_id):
1000
"""Return the delta for one revision.
1002
The delta is relative to the left-hand predecessor of the
1005
r = self.get_revision(revision_id)
1006
return list(self.get_deltas_for_revisions([r]))[0]
1009
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1010
revision_id = osutils.safe_revision_id(revision_id)
1011
signature = gpg_strategy.sign(plaintext)
1012
self._revision_store.add_revision_signature_text(revision_id,
1014
self.get_transaction())
1016
def fileids_altered_by_revision_ids(self, revision_ids):
1017
"""Find the file ids and versions affected by revisions.
1019
:param revisions: an iterable containing revision ids.
1020
:return: a dictionary mapping altered file-ids to an iterable of
1021
revision_ids. Each altered file-ids has the exact revision_ids that
1022
altered it listed explicitly.
1024
assert self._serializer.support_altered_by_hack, \
1025
("fileids_altered_by_revision_ids only supported for branches "
1026
"which store inventory as unnested xml, not on %r" % self)
1027
selected_revision_ids = set(osutils.safe_revision_id(r)
1028
for r in revision_ids)
1029
w = self.get_inventory_weave()
1032
# this code needs to read every new line in every inventory for the
1033
# inventories [revision_ids]. Seeing a line twice is ok. Seeing a line
1034
# not present in one of those inventories is unnecessary but not
1035
# harmful because we are filtering by the revision id marker in the
1036
# inventory lines : we only select file ids altered in one of those
1037
# revisions. We don't need to see all lines in the inventory because
1038
# only those added in an inventory in rev X can contain a revision=X
1040
unescape_revid_cache = {}
1041
unescape_fileid_cache = {}
1043
# jam 20061218 In a big fetch, this handles hundreds of thousands
1044
# of lines, so it has had a lot of inlining and optimizing done.
1045
# Sorry that it is a little bit messy.
1046
# Move several functions to be local variables, since this is a long
1048
search = self._file_ids_altered_regex.search
1049
unescape = _unescape_xml
1050
setdefault = result.setdefault
1051
pb = ui.ui_factory.nested_progress_bar()
1053
for line in w.iter_lines_added_or_present_in_versions(
1054
selected_revision_ids, pb=pb):
1055
match = search(line)
1058
# One call to match.group() returning multiple items is quite a
1059
# bit faster than 2 calls to match.group() each returning 1
1060
file_id, revision_id = match.group('file_id', 'revision_id')
1062
# Inlining the cache lookups helps a lot when you make 170,000
1063
# lines and 350k ids, versus 8.4 unique ids.
1064
# Using a cache helps in 2 ways:
1065
# 1) Avoids unnecessary decoding calls
1066
# 2) Re-uses cached strings, which helps in future set and
1068
# (2) is enough that removing encoding entirely along with
1069
# the cache (so we are using plain strings) results in no
1070
# performance improvement.
1072
revision_id = unescape_revid_cache[revision_id]
1074
unescaped = unescape(revision_id)
1075
unescape_revid_cache[revision_id] = unescaped
1076
revision_id = unescaped
1078
if revision_id in selected_revision_ids:
1080
file_id = unescape_fileid_cache[file_id]
1082
unescaped = unescape(file_id)
1083
unescape_fileid_cache[file_id] = unescaped
1085
setdefault(file_id, set()).add(revision_id)
1090
def iter_files_bytes(self, desired_files):
1091
"""Iterate through file versions.
1093
Files will not necessarily be returned in the order they occur in
1094
desired_files. No specific order is guaranteed.
1096
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1097
value supplied by the caller as part of desired_files. It should
1098
uniquely identify the file version in the caller's context. (Examples:
1099
an index number or a TreeTransform trans_id.)
1101
bytes_iterator is an iterable of bytestrings for the file. The
1102
kind of iterable and length of the bytestrings are unspecified, but for
1103
this implementation, it is a list of lines produced by
1104
VersionedFile.get_lines().
1106
:param desired_files: a list of (file_id, revision_id, identifier)
1109
transaction = self.get_transaction()
1110
for file_id, revision_id, callable_data in desired_files:
1112
weave = self.weave_store.get_weave(file_id, transaction)
1113
except errors.NoSuchFile:
1114
raise errors.NoSuchIdInRepository(self, file_id)
1115
yield callable_data, weave.get_lines(revision_id)
1117
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1118
"""Get an iterable listing the keys of all the data introduced by a set
1121
The keys will be ordered so that the corresponding items can be safely
1122
fetched and inserted in that order.
1124
:returns: An iterable producing tuples of (knit-kind, file-id,
1125
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1126
'revisions'. file-id is None unless knit-kind is 'file'.
1128
# XXX: it's a bit weird to control the inventory weave caching in this
1129
# generator. Ideally the caching would be done in fetch.py I think. Or
1130
# maybe this generator should explicitly have the contract that it
1131
# should not be iterated until the previously yielded item has been
1133
inv_w = self.get_inventory_weave()
1134
inv_w.enable_cache()
1136
# file ids that changed
1137
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1139
num_file_ids = len(file_ids)
1140
for file_id, altered_versions in file_ids.iteritems():
1141
if _files_pb is not None:
1142
_files_pb.update("fetch texts", count, num_file_ids)
1144
yield ("file", file_id, altered_versions)
1145
# We're done with the files_pb. Note that it finished by the caller,
1146
# just as it was created by the caller.
1150
yield ("inventory", None, revision_ids)
1154
revisions_with_signatures = set()
1155
for rev_id in revision_ids:
1157
self.get_signature_text(rev_id)
1158
except errors.NoSuchRevision:
1162
revisions_with_signatures.add(rev_id)
1163
yield ("signatures", None, revisions_with_signatures)
1166
yield ("revisions", None, revision_ids)
1169
def get_inventory_weave(self):
1170
return self.control_weaves.get_weave('inventory',
1171
self.get_transaction())
1174
def get_inventory(self, revision_id):
1175
"""Get Inventory object by hash."""
1176
# TODO: jam 20070210 Technically we don't need to sanitize, since all
1177
# called functions must sanitize.
1178
revision_id = osutils.safe_revision_id(revision_id)
1179
return self.deserialise_inventory(
1180
revision_id, self.get_inventory_xml(revision_id))
1182
def deserialise_inventory(self, revision_id, xml):
1183
"""Transform the xml into an inventory object.
1185
:param revision_id: The expected revision id of the inventory.
1186
:param xml: A serialised inventory.
1188
revision_id = osutils.safe_revision_id(revision_id)
1189
result = self._serializer.read_inventory_from_string(xml)
1190
result.root.revision = revision_id
1193
def serialise_inventory(self, inv):
1194
return self._serializer.write_inventory_to_string(inv)
1196
def _serialise_inventory_to_lines(self, inv):
1197
return self._serializer.write_inventory_to_lines(inv)
1199
def get_serializer_format(self):
1200
return self._serializer.format_num
1203
def get_inventory_xml(self, revision_id):
1204
"""Get inventory XML as a file object."""
1205
revision_id = osutils.safe_revision_id(revision_id)
1207
assert isinstance(revision_id, str), type(revision_id)
1208
iw = self.get_inventory_weave()
1209
return iw.get_text(revision_id)
1211
raise errors.HistoryMissing(self, 'inventory', revision_id)
1214
def get_inventory_sha1(self, revision_id):
1215
"""Return the sha1 hash of the inventory entry
1217
# TODO: jam 20070210 Shouldn't this be deprecated / removed?
1218
revision_id = osutils.safe_revision_id(revision_id)
1219
return self.get_revision(revision_id).inventory_sha1
1222
def get_revision_graph(self, revision_id=None):
1223
"""Return a dictionary containing the revision graph.
1225
NB: This method should not be used as it accesses the entire graph all
1226
at once, which is much more data than most operations should require.
1228
:param revision_id: The revision_id to get a graph from. If None, then
1229
the entire revision graph is returned. This is a deprecated mode of
1230
operation and will be removed in the future.
1231
:return: a dictionary of revision_id->revision_parents_list.
1233
raise NotImplementedError(self.get_revision_graph)
1236
def get_revision_graph_with_ghosts(self, revision_ids=None):
1237
"""Return a graph of the revisions with ghosts marked as applicable.
1239
:param revision_ids: an iterable of revisions to graph or None for all.
1240
:return: a Graph object with the graph reachable from revision_ids.
1242
if 'evil' in debug.debug_flags:
1244
"get_revision_graph_with_ghosts scales with size of history.")
1245
result = deprecated_graph.Graph()
1246
if not revision_ids:
1247
pending = set(self.all_revision_ids())
1250
pending = set(osutils.safe_revision_id(r) for r in revision_ids)
1251
# special case NULL_REVISION
1252
if _mod_revision.NULL_REVISION in pending:
1253
pending.remove(_mod_revision.NULL_REVISION)
1254
required = set(pending)
1257
revision_id = pending.pop()
1259
rev = self.get_revision(revision_id)
1260
except errors.NoSuchRevision:
1261
if revision_id in required:
1264
result.add_ghost(revision_id)
1266
for parent_id in rev.parent_ids:
1267
# is this queued or done ?
1268
if (parent_id not in pending and
1269
parent_id not in done):
1271
pending.add(parent_id)
1272
result.add_node(revision_id, rev.parent_ids)
1273
done.add(revision_id)
1276
def _get_history_vf(self):
1277
"""Get a versionedfile whose history graph reflects all revisions.
1279
For weave repositories, this is the inventory weave.
1281
return self.get_inventory_weave()
1283
def iter_reverse_revision_history(self, revision_id):
1284
"""Iterate backwards through revision ids in the lefthand history
1286
:param revision_id: The revision id to start with. All its lefthand
1287
ancestors will be traversed.
1289
revision_id = osutils.safe_revision_id(revision_id)
1290
if revision_id in (None, _mod_revision.NULL_REVISION):
1292
next_id = revision_id
1293
versionedfile = self._get_history_vf()
1296
parents = versionedfile.get_parents(next_id)
1297
if len(parents) == 0:
1300
next_id = parents[0]
1303
def get_revision_inventory(self, revision_id):
1304
"""Return inventory of a past revision."""
1305
# TODO: Unify this with get_inventory()
1306
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1307
# must be the same as its revision, so this is trivial.
1308
if revision_id is None:
1309
# This does not make sense: if there is no revision,
1310
# then it is the current tree inventory surely ?!
1311
# and thus get_root_id() is something that looks at the last
1312
# commit on the branch, and the get_root_id is an inventory check.
1313
raise NotImplementedError
1314
# return Inventory(self.get_root_id())
1316
return self.get_inventory(revision_id)
1319
def is_shared(self):
1320
"""Return True if this repository is flagged as a shared repository."""
1321
raise NotImplementedError(self.is_shared)
1324
def reconcile(self, other=None, thorough=False):
1325
"""Reconcile this repository."""
1326
from bzrlib.reconcile import RepoReconciler
1327
reconciler = RepoReconciler(self, thorough=thorough)
1328
reconciler.reconcile()
1331
def _refresh_data(self):
1332
"""Helper called from lock_* to ensure coherency with disk.
1334
The default implementation does nothing; it is however possible
1335
for repositories to maintain loaded indices across multiple locks
1336
by checking inside their implementation of this method to see
1337
whether their indices are still valid. This depends of course on
1338
the disk format being validatable in this manner.
1342
def revision_tree(self, revision_id):
1343
"""Return Tree for a revision on this branch.
1345
`revision_id` may be None for the empty tree revision.
1347
# TODO: refactor this to use an existing revision object
1348
# so we don't need to read it in twice.
1349
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1350
return RevisionTree(self, Inventory(root_id=None),
1351
_mod_revision.NULL_REVISION)
1353
revision_id = osutils.safe_revision_id(revision_id)
1354
inv = self.get_revision_inventory(revision_id)
1355
return RevisionTree(self, inv, revision_id)
1358
def revision_trees(self, revision_ids):
1359
"""Return Tree for a revision on this branch.
1361
`revision_id` may not be None or 'null:'"""
1362
assert None not in revision_ids
1363
assert _mod_revision.NULL_REVISION not in revision_ids
1364
texts = self.get_inventory_weave().get_texts(revision_ids)
1365
for text, revision_id in zip(texts, revision_ids):
1366
inv = self.deserialise_inventory(revision_id, text)
1367
yield RevisionTree(self, inv, revision_id)
1370
def get_ancestry(self, revision_id, topo_sorted=True):
1371
"""Return a list of revision-ids integrated by a revision.
1373
The first element of the list is always None, indicating the origin
1374
revision. This might change when we have history horizons, or
1375
perhaps we should have a new API.
1377
This is topologically sorted.
1379
if _mod_revision.is_null(revision_id):
1381
revision_id = osutils.safe_revision_id(revision_id)
1382
if not self.has_revision(revision_id):
1383
raise errors.NoSuchRevision(self, revision_id)
1384
w = self.get_inventory_weave()
1385
candidates = w.get_ancestry(revision_id, topo_sorted)
1386
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1389
"""Compress the data within the repository.
1391
This operation only makes sense for some repository types. For other
1392
types it should be a no-op that just returns.
1394
This stub method does not require a lock, but subclasses should use
1395
@needs_write_lock as this is a long running call its reasonable to
1396
implicitly lock for the user.
1400
def print_file(self, file, revision_id):
1401
"""Print `file` to stdout.
1403
FIXME RBC 20060125 as John Meinel points out this is a bad api
1404
- it writes to stdout, it assumes that that is valid etc. Fix
1405
by creating a new more flexible convenience function.
1407
revision_id = osutils.safe_revision_id(revision_id)
1408
tree = self.revision_tree(revision_id)
1409
# use inventory as it was in that revision
1410
file_id = tree.inventory.path2id(file)
1412
# TODO: jam 20060427 Write a test for this code path
1413
# it had a bug in it, and was raising the wrong
1415
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1416
tree.print_file(file_id)
1418
def get_transaction(self):
1419
return self.control_files.get_transaction()
1421
def revision_parents(self, revision_id):
1422
revision_id = osutils.safe_revision_id(revision_id)
1423
return self.get_inventory_weave().parent_names(revision_id)
1425
def get_parents(self, revision_ids):
1426
"""See StackedParentsProvider.get_parents"""
1428
for revision_id in revision_ids:
1429
if revision_id == _mod_revision.NULL_REVISION:
1433
parents = self.get_revision(revision_id).parent_ids
1434
except errors.NoSuchRevision:
1437
if len(parents) == 0:
1438
parents = [_mod_revision.NULL_REVISION]
1439
parents_list.append(parents)
1442
def _make_parents_provider(self):
1445
def get_graph(self, other_repository=None):
1446
"""Return the graph walker for this repository format"""
1447
parents_provider = self._make_parents_provider()
1448
if (other_repository is not None and
1449
other_repository.bzrdir.transport.base !=
1450
self.bzrdir.transport.base):
1451
parents_provider = graph._StackedParentsProvider(
1452
[parents_provider, other_repository._make_parents_provider()])
1453
return graph.Graph(parents_provider)
1455
def get_versioned_file_checker(self, revisions, revision_versions_cache):
1456
return WeaveChecker(revisions, revision_versions_cache, self)
1459
def set_make_working_trees(self, new_value):
1460
"""Set the policy flag for making working trees when creating branches.
1462
This only applies to branches that use this repository.
1464
The default is 'True'.
1465
:param new_value: True to restore the default, False to disable making
1468
raise NotImplementedError(self.set_make_working_trees)
1470
def make_working_trees(self):
1471
"""Returns the policy for making working trees on new branches."""
1472
raise NotImplementedError(self.make_working_trees)
1475
def sign_revision(self, revision_id, gpg_strategy):
1476
revision_id = osutils.safe_revision_id(revision_id)
1477
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1478
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1481
def has_signature_for_revision_id(self, revision_id):
1482
"""Query for a revision signature for revision_id in the repository."""
1483
revision_id = osutils.safe_revision_id(revision_id)
1484
return self._revision_store.has_signature(revision_id,
1485
self.get_transaction())
1488
def get_signature_text(self, revision_id):
1489
"""Return the text for a signature."""
1490
revision_id = osutils.safe_revision_id(revision_id)
1491
return self._revision_store.get_signature_text(revision_id,
1492
self.get_transaction())
1495
def check(self, revision_ids=None):
1496
"""Check consistency of all history of given revision_ids.
1498
Different repository implementations should override _check().
1500
:param revision_ids: A non-empty list of revision_ids whose ancestry
1501
will be checked. Typically the last revision_id of a branch.
1503
if revision_ids is not None:
1504
symbol_versioning.warn('revision_ids should not be supplied to'
1505
' Repostiory.check, as of bzr 0.92.',
1506
DeprecationWarning, stacklevel=3)
1507
return self._check(revision_ids)
1509
def _check(self, revision_ids):
1510
result = check.Check(self)
1514
def _warn_if_deprecated(self):
1515
global _deprecation_warning_done
1516
if _deprecation_warning_done:
1518
_deprecation_warning_done = True
1519
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1520
% (self._format, self.bzrdir.transport.base))
1522
def supports_rich_root(self):
1523
return self._format.rich_root_data
1525
def _check_ascii_revisionid(self, revision_id, method):
1526
"""Private helper for ascii-only repositories."""
1527
# weave repositories refuse to store revisionids that are non-ascii.
1528
if revision_id is not None:
1529
# weaves require ascii revision ids.
1530
if isinstance(revision_id, unicode):
1532
revision_id.encode('ascii')
1533
except UnicodeEncodeError:
1534
raise errors.NonAsciiRevisionId(method, self)
1537
revision_id.decode('ascii')
1538
except UnicodeDecodeError:
1539
raise errors.NonAsciiRevisionId(method, self)
1541
def revision_graph_can_have_wrong_parents(self):
1542
"""Is it possible for this repository to have a revision graph with
1545
If True, then this repository must also implement
1546
_find_inconsistent_revision_parents so that check and reconcile can
1547
check for inconsistencies before proceeding with other checks that may
1548
depend on the revision index being consistent.
1550
raise NotImplementedError(self.revision_graph_can_have_wrong_parents)
1552
# remove these delegates a while after bzr 0.15
1553
def __make_delegated(name, from_module):
1554
def _deprecated_repository_forwarder():
1555
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1556
% (name, from_module),
1559
m = __import__(from_module, globals(), locals(), [name])
1561
return getattr(m, name)
1562
except AttributeError:
1563
raise AttributeError('module %s has no name %s'
1565
globals()[name] = _deprecated_repository_forwarder
1568
'AllInOneRepository',
1569
'WeaveMetaDirRepository',
1570
'PreSplitOutRepositoryFormat',
1571
'RepositoryFormat4',
1572
'RepositoryFormat5',
1573
'RepositoryFormat6',
1574
'RepositoryFormat7',
1576
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1580
'RepositoryFormatKnit',
1581
'RepositoryFormatKnit1',
1583
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1586
def install_revision(repository, rev, revision_tree):
1587
"""Install all revision data into a repository."""
1588
present_parents = []
1590
for p_id in rev.parent_ids:
1591
if repository.has_revision(p_id):
1592
present_parents.append(p_id)
1593
parent_trees[p_id] = repository.revision_tree(p_id)
1595
parent_trees[p_id] = repository.revision_tree(None)
1597
inv = revision_tree.inventory
1598
entries = inv.iter_entries()
1599
# backwards compatibility hack: skip the root id.
1600
if not repository.supports_rich_root():
1601
path, root = entries.next()
1602
if root.revision != rev.revision_id:
1603
raise errors.IncompatibleRevision(repr(repository))
1604
# Add the texts that are not already present
1605
for path, ie in entries:
1606
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1607
repository.get_transaction())
1608
if ie.revision not in w:
1610
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1611
# with InventoryEntry.find_previous_heads(). if it is, then there
1612
# is a latent bug here where the parents may have ancestors of each
1614
for revision, tree in parent_trees.iteritems():
1615
if ie.file_id not in tree:
1617
parent_id = tree.inventory[ie.file_id].revision
1618
if parent_id in text_parents:
1620
text_parents.append(parent_id)
1622
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1623
repository.get_transaction())
1624
lines = revision_tree.get_file(ie.file_id).readlines()
1625
vfile.add_lines(rev.revision_id, text_parents, lines)
1627
# install the inventory
1628
repository.add_inventory(rev.revision_id, inv, present_parents)
1629
except errors.RevisionAlreadyPresent:
1631
repository.add_revision(rev.revision_id, rev, inv)
1634
class MetaDirRepository(Repository):
1635
"""Repositories in the new meta-dir layout."""
1637
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1638
super(MetaDirRepository, self).__init__(_format,
1644
dir_mode = self.control_files._dir_mode
1645
file_mode = self.control_files._file_mode
1648
def is_shared(self):
1649
"""Return True if this repository is flagged as a shared repository."""
1650
return self.control_files._transport.has('shared-storage')
1653
def set_make_working_trees(self, new_value):
1654
"""Set the policy flag for making working trees when creating branches.
1656
This only applies to branches that use this repository.
1658
The default is 'True'.
1659
:param new_value: True to restore the default, False to disable making
1664
self.control_files._transport.delete('no-working-trees')
1665
except errors.NoSuchFile:
1668
self.control_files.put_utf8('no-working-trees', '')
1670
def make_working_trees(self):
1671
"""Returns the policy for making working trees on new branches."""
1672
return not self.control_files._transport.has('no-working-trees')
1675
class RepositoryFormatRegistry(registry.Registry):
1676
"""Registry of RepositoryFormats.
1679
def get(self, format_string):
1680
r = registry.Registry.get(self, format_string)
1686
format_registry = RepositoryFormatRegistry()
1687
"""Registry of formats, indexed by their identifying format string.
1689
This can contain either format instances themselves, or classes/factories that
1690
can be called to obtain one.
1694
#####################################################################
1695
# Repository Formats
1697
class RepositoryFormat(object):
1698
"""A repository format.
1700
Formats provide three things:
1701
* An initialization routine to construct repository data on disk.
1702
* a format string which is used when the BzrDir supports versioned
1704
* an open routine which returns a Repository instance.
1706
Formats are placed in an dict by their format string for reference
1707
during opening. These should be subclasses of RepositoryFormat
1710
Once a format is deprecated, just deprecate the initialize and open
1711
methods on the format class. Do not deprecate the object, as the
1712
object will be created every system load.
1714
Common instance attributes:
1715
_matchingbzrdir - the bzrdir format that the repository format was
1716
originally written to work with. This can be used if manually
1717
constructing a bzrdir and repository, or more commonly for test suite
1722
return "<%s>" % self.__class__.__name__
1724
def __eq__(self, other):
1725
# format objects are generally stateless
1726
return isinstance(other, self.__class__)
1728
def __ne__(self, other):
1729
return not self == other
1732
def find_format(klass, a_bzrdir):
1733
"""Return the format for the repository object in a_bzrdir.
1735
This is used by bzr native formats that have a "format" file in
1736
the repository. Other methods may be used by different types of
1740
transport = a_bzrdir.get_repository_transport(None)
1741
format_string = transport.get("format").read()
1742
return format_registry.get(format_string)
1743
except errors.NoSuchFile:
1744
raise errors.NoRepositoryPresent(a_bzrdir)
1746
raise errors.UnknownFormatError(format=format_string)
1749
def register_format(klass, format):
1750
format_registry.register(format.get_format_string(), format)
1753
def unregister_format(klass, format):
1754
format_registry.remove(format.get_format_string())
1757
def get_default_format(klass):
1758
"""Return the current default format."""
1759
from bzrlib import bzrdir
1760
return bzrdir.format_registry.make_bzrdir('default').repository_format
1762
def _get_control_store(self, repo_transport, control_files):
1763
"""Return the control store for this repository."""
1764
raise NotImplementedError(self._get_control_store)
1766
def get_format_string(self):
1767
"""Return the ASCII format string that identifies this format.
1769
Note that in pre format ?? repositories the format string is
1770
not permitted nor written to disk.
1772
raise NotImplementedError(self.get_format_string)
1774
def get_format_description(self):
1775
"""Return the short description for this format."""
1776
raise NotImplementedError(self.get_format_description)
1778
def _get_revision_store(self, repo_transport, control_files):
1779
"""Return the revision store object for this a_bzrdir."""
1780
raise NotImplementedError(self._get_revision_store)
1782
def _get_text_rev_store(self,
1789
"""Common logic for getting a revision store for a repository.
1791
see self._get_revision_store for the subclass-overridable method to
1792
get the store for a repository.
1794
from bzrlib.store.revision.text import TextRevisionStore
1795
dir_mode = control_files._dir_mode
1796
file_mode = control_files._file_mode
1797
text_store = TextStore(transport.clone(name),
1799
compressed=compressed,
1801
file_mode=file_mode)
1802
_revision_store = TextRevisionStore(text_store, serializer)
1803
return _revision_store
1805
# TODO: this shouldn't be in the base class, it's specific to things that
1806
# use weaves or knits -- mbp 20070207
1807
def _get_versioned_file_store(self,
1812
versionedfile_class=None,
1813
versionedfile_kwargs={},
1815
if versionedfile_class is None:
1816
versionedfile_class = self._versionedfile_class
1817
weave_transport = control_files._transport.clone(name)
1818
dir_mode = control_files._dir_mode
1819
file_mode = control_files._file_mode
1820
return VersionedFileStore(weave_transport, prefixed=prefixed,
1822
file_mode=file_mode,
1823
versionedfile_class=versionedfile_class,
1824
versionedfile_kwargs=versionedfile_kwargs,
1827
def initialize(self, a_bzrdir, shared=False):
1828
"""Initialize a repository of this format in a_bzrdir.
1830
:param a_bzrdir: The bzrdir to put the new repository in it.
1831
:param shared: The repository should be initialized as a sharable one.
1832
:returns: The new repository object.
1834
This may raise UninitializableFormat if shared repository are not
1835
compatible the a_bzrdir.
1837
raise NotImplementedError(self.initialize)
1839
def is_supported(self):
1840
"""Is this format supported?
1842
Supported formats must be initializable and openable.
1843
Unsupported formats may not support initialization or committing or
1844
some other features depending on the reason for not being supported.
1848
def check_conversion_target(self, target_format):
1849
raise NotImplementedError(self.check_conversion_target)
1851
def open(self, a_bzrdir, _found=False):
1852
"""Return an instance of this format for the bzrdir a_bzrdir.
1854
_found is a private parameter, do not use it.
1856
raise NotImplementedError(self.open)
1859
class MetaDirRepositoryFormat(RepositoryFormat):
1860
"""Common base class for the new repositories using the metadir layout."""
1862
rich_root_data = False
1863
supports_tree_reference = False
1864
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1867
super(MetaDirRepositoryFormat, self).__init__()
1869
def _create_control_files(self, a_bzrdir):
1870
"""Create the required files and the initial control_files object."""
1871
# FIXME: RBC 20060125 don't peek under the covers
1872
# NB: no need to escape relative paths that are url safe.
1873
repository_transport = a_bzrdir.get_repository_transport(self)
1874
control_files = lockable_files.LockableFiles(repository_transport,
1875
'lock', lockdir.LockDir)
1876
control_files.create_lock()
1877
return control_files
1879
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1880
"""Upload the initial blank content."""
1881
control_files = self._create_control_files(a_bzrdir)
1882
control_files.lock_write()
1884
control_files._transport.mkdir_multi(dirs,
1885
mode=control_files._dir_mode)
1886
for file, content in files:
1887
control_files.put(file, content)
1888
for file, content in utf8_files:
1889
control_files.put_utf8(file, content)
1891
control_files.put_utf8('shared-storage', '')
1893
control_files.unlock()
1896
# formats which have no format string are not discoverable
1897
# and not independently creatable, so are not registered. They're
1898
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1899
# needed, it's constructed directly by the BzrDir. Non-native formats where
1900
# the repository is not separately opened are similar.
1902
format_registry.register_lazy(
1903
'Bazaar-NG Repository format 7',
1904
'bzrlib.repofmt.weaverepo',
1907
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1908
# default control directory format
1910
format_registry.register_lazy(
1911
'Bazaar-NG Knit Repository Format 1',
1912
'bzrlib.repofmt.knitrepo',
1913
'RepositoryFormatKnit1',
1915
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1917
format_registry.register_lazy(
1918
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1919
'bzrlib.repofmt.knitrepo',
1920
'RepositoryFormatKnit3',
1924
class InterRepository(InterObject):
1925
"""This class represents operations taking place between two repositories.
1927
Its instances have methods like copy_content and fetch, and contain
1928
references to the source and target repositories these operations can be
1931
Often we will provide convenience methods on 'repository' which carry out
1932
operations with another repository - they will always forward to
1933
InterRepository.get(other).method_name(parameters).
1937
"""The available optimised InterRepository types."""
1939
def copy_content(self, revision_id=None):
1940
raise NotImplementedError(self.copy_content)
1942
def fetch(self, revision_id=None, pb=None):
1943
"""Fetch the content required to construct revision_id.
1945
The content is copied from self.source to self.target.
1947
:param revision_id: if None all content is copied, if NULL_REVISION no
1949
:param pb: optional progress bar to use for progress reports. If not
1950
provided a default one will be created.
1952
Returns the copied revision count and the failed revisions in a tuple:
1955
raise NotImplementedError(self.fetch)
1958
def missing_revision_ids(self, revision_id=None):
1959
"""Return the revision ids that source has that target does not.
1961
These are returned in topological order.
1963
:param revision_id: only return revision ids included by this
1966
# generic, possibly worst case, slow code path.
1967
target_ids = set(self.target.all_revision_ids())
1968
if revision_id is not None:
1969
# TODO: jam 20070210 InterRepository is internal enough that it
1970
# should assume revision_ids are already utf-8
1971
revision_id = osutils.safe_revision_id(revision_id)
1972
source_ids = self.source.get_ancestry(revision_id)
1973
assert source_ids[0] is None
1976
source_ids = self.source.all_revision_ids()
1977
result_set = set(source_ids).difference(target_ids)
1978
# this may look like a no-op: its not. It preserves the ordering
1979
# other_ids had while only returning the members from other_ids
1980
# that we've decided we need.
1981
return [rev_id for rev_id in source_ids if rev_id in result_set]
1984
class InterSameDataRepository(InterRepository):
1985
"""Code for converting between repositories that represent the same data.
1987
Data format and model must match for this to work.
1991
def _get_repo_format_to_test(self):
1992
"""Repository format for testing with.
1994
InterSameData can pull from subtree to subtree and from non-subtree to
1995
non-subtree, so we test this with the richest repository format.
1997
from bzrlib.repofmt import knitrepo
1998
return knitrepo.RepositoryFormatKnit3()
2001
def is_compatible(source, target):
2002
if source.supports_rich_root() != target.supports_rich_root():
2004
if source._serializer != target._serializer:
2009
def copy_content(self, revision_id=None):
2010
"""Make a complete copy of the content in self into destination.
2012
This copies both the repository's revision data, and configuration information
2013
such as the make_working_trees setting.
2015
This is a destructive operation! Do not use it on existing
2018
:param revision_id: Only copy the content needed to construct
2019
revision_id and its parents.
2022
self.target.set_make_working_trees(self.source.make_working_trees())
2023
except NotImplementedError:
2025
# TODO: jam 20070210 This is fairly internal, so we should probably
2026
# just assert that revision_id is not unicode.
2027
revision_id = osutils.safe_revision_id(revision_id)
2028
# but don't bother fetching if we have the needed data now.
2029
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2030
self.target.has_revision(revision_id)):
2032
self.target.fetch(self.source, revision_id=revision_id)
2035
def fetch(self, revision_id=None, pb=None):
2036
"""See InterRepository.fetch()."""
2037
from bzrlib.fetch import GenericRepoFetcher
2038
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2039
self.source, self.source._format, self.target,
2040
self.target._format)
2041
# TODO: jam 20070210 This should be an assert, not a translate
2042
revision_id = osutils.safe_revision_id(revision_id)
2043
f = GenericRepoFetcher(to_repository=self.target,
2044
from_repository=self.source,
2045
last_revision=revision_id,
2047
return f.count_copied, f.failed_revisions
2050
class InterWeaveRepo(InterSameDataRepository):
2051
"""Optimised code paths between Weave based repositories.
2053
This should be in bzrlib/repofmt/weaverepo.py but we have not yet
2054
implemented lazy inter-object optimisation.
2058
def _get_repo_format_to_test(self):
2059
from bzrlib.repofmt import weaverepo
2060
return weaverepo.RepositoryFormat7()
2063
def is_compatible(source, target):
2064
"""Be compatible with known Weave formats.
2066
We don't test for the stores being of specific types because that
2067
could lead to confusing results, and there is no need to be
2070
from bzrlib.repofmt.weaverepo import (
2076
return (isinstance(source._format, (RepositoryFormat5,
2078
RepositoryFormat7)) and
2079
isinstance(target._format, (RepositoryFormat5,
2081
RepositoryFormat7)))
2082
except AttributeError:
2086
def copy_content(self, revision_id=None):
2087
"""See InterRepository.copy_content()."""
2088
# weave specific optimised path:
2089
# TODO: jam 20070210 Internal, should be an assert, not translate
2090
revision_id = osutils.safe_revision_id(revision_id)
2092
self.target.set_make_working_trees(self.source.make_working_trees())
2093
except NotImplementedError:
2095
# FIXME do not peek!
2096
if self.source.control_files._transport.listable():
2097
pb = ui.ui_factory.nested_progress_bar()
2099
self.target.weave_store.copy_all_ids(
2100
self.source.weave_store,
2102
from_transaction=self.source.get_transaction(),
2103
to_transaction=self.target.get_transaction())
2104
pb.update('copying inventory', 0, 1)
2105
self.target.control_weaves.copy_multi(
2106
self.source.control_weaves, ['inventory'],
2107
from_transaction=self.source.get_transaction(),
2108
to_transaction=self.target.get_transaction())
2109
self.target._revision_store.text_store.copy_all_ids(
2110
self.source._revision_store.text_store,
2115
self.target.fetch(self.source, revision_id=revision_id)
2118
def fetch(self, revision_id=None, pb=None):
2119
"""See InterRepository.fetch()."""
2120
from bzrlib.fetch import GenericRepoFetcher
2121
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2122
self.source, self.source._format, self.target, self.target._format)
2123
# TODO: jam 20070210 This should be an assert, not a translate
2124
revision_id = osutils.safe_revision_id(revision_id)
2125
f = GenericRepoFetcher(to_repository=self.target,
2126
from_repository=self.source,
2127
last_revision=revision_id,
2129
return f.count_copied, f.failed_revisions
2132
def missing_revision_ids(self, revision_id=None):
2133
"""See InterRepository.missing_revision_ids()."""
2134
# we want all revisions to satisfy revision_id in source.
2135
# but we don't want to stat every file here and there.
2136
# we want then, all revisions other needs to satisfy revision_id
2137
# checked, but not those that we have locally.
2138
# so the first thing is to get a subset of the revisions to
2139
# satisfy revision_id in source, and then eliminate those that
2140
# we do already have.
2141
# this is slow on high latency connection to self, but as as this
2142
# disk format scales terribly for push anyway due to rewriting
2143
# inventory.weave, this is considered acceptable.
2145
if revision_id is not None:
2146
source_ids = self.source.get_ancestry(revision_id)
2147
assert source_ids[0] is None
2150
source_ids = self.source._all_possible_ids()
2151
source_ids_set = set(source_ids)
2152
# source_ids is the worst possible case we may need to pull.
2153
# now we want to filter source_ids against what we actually
2154
# have in target, but don't try to check for existence where we know
2155
# we do not have a revision as that would be pointless.
2156
target_ids = set(self.target._all_possible_ids())
2157
possibly_present_revisions = target_ids.intersection(source_ids_set)
2158
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2159
required_revisions = source_ids_set.difference(actually_present_revisions)
2160
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2161
if revision_id is not None:
2162
# we used get_ancestry to determine source_ids then we are assured all
2163
# revisions referenced are present as they are installed in topological order.
2164
# and the tip revision was validated by get_ancestry.
2165
return required_topo_revisions
2167
# if we just grabbed the possibly available ids, then
2168
# we only have an estimate of whats available and need to validate
2169
# that against the revision records.
2170
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2173
class InterKnitRepo(InterSameDataRepository):
2174
"""Optimised code paths between Knit based repositories."""
2177
def _get_repo_format_to_test(self):
2178
from bzrlib.repofmt import knitrepo
2179
return knitrepo.RepositoryFormatKnit1()
2182
def is_compatible(source, target):
2183
"""Be compatible with known Knit formats.
2185
We don't test for the stores being of specific types because that
2186
could lead to confusing results, and there is no need to be
2189
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1
2191
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2192
isinstance(target._format, (RepositoryFormatKnit1)))
2193
except AttributeError:
2197
def fetch(self, revision_id=None, pb=None):
2198
"""See InterRepository.fetch()."""
2199
from bzrlib.fetch import KnitRepoFetcher
2200
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2201
self.source, self.source._format, self.target, self.target._format)
2202
# TODO: jam 20070210 This should be an assert, not a translate
2203
revision_id = osutils.safe_revision_id(revision_id)
2204
f = KnitRepoFetcher(to_repository=self.target,
2205
from_repository=self.source,
2206
last_revision=revision_id,
2208
return f.count_copied, f.failed_revisions
2211
def missing_revision_ids(self, revision_id=None):
2212
"""See InterRepository.missing_revision_ids()."""
2213
if revision_id is not None:
2214
source_ids = self.source.get_ancestry(revision_id)
2215
assert source_ids[0] is None
2218
source_ids = self.source.all_revision_ids()
2219
source_ids_set = set(source_ids)
2220
# source_ids is the worst possible case we may need to pull.
2221
# now we want to filter source_ids against what we actually
2222
# have in target, but don't try to check for existence where we know
2223
# we do not have a revision as that would be pointless.
2224
target_ids = set(self.target.all_revision_ids())
2225
possibly_present_revisions = target_ids.intersection(source_ids_set)
2226
actually_present_revisions = set(self.target._eliminate_revisions_not_present(possibly_present_revisions))
2227
required_revisions = source_ids_set.difference(actually_present_revisions)
2228
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2229
if revision_id is not None:
2230
# we used get_ancestry to determine source_ids then we are assured all
2231
# revisions referenced are present as they are installed in topological order.
2232
# and the tip revision was validated by get_ancestry.
2233
return required_topo_revisions
2235
# if we just grabbed the possibly available ids, then
2236
# we only have an estimate of whats available and need to validate
2237
# that against the revision records.
2238
return self.source._eliminate_revisions_not_present(required_topo_revisions)
2241
class InterModel1and2(InterRepository):
2244
def _get_repo_format_to_test(self):
2248
def is_compatible(source, target):
2249
if not source.supports_rich_root() and target.supports_rich_root():
2255
def fetch(self, revision_id=None, pb=None):
2256
"""See InterRepository.fetch()."""
2257
from bzrlib.fetch import Model1toKnit2Fetcher
2258
# TODO: jam 20070210 This should be an assert, not a translate
2259
revision_id = osutils.safe_revision_id(revision_id)
2260
f = Model1toKnit2Fetcher(to_repository=self.target,
2261
from_repository=self.source,
2262
last_revision=revision_id,
2264
return f.count_copied, f.failed_revisions
2267
def copy_content(self, revision_id=None):
2268
"""Make a complete copy of the content in self into destination.
2270
This is a destructive operation! Do not use it on existing
2273
:param revision_id: Only copy the content needed to construct
2274
revision_id and its parents.
2277
self.target.set_make_working_trees(self.source.make_working_trees())
2278
except NotImplementedError:
2280
# TODO: jam 20070210 Internal, assert, don't translate
2281
revision_id = osutils.safe_revision_id(revision_id)
2282
# but don't bother fetching if we have the needed data now.
2283
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2284
self.target.has_revision(revision_id)):
2286
self.target.fetch(self.source, revision_id=revision_id)
2289
class InterKnit1and2(InterKnitRepo):
2292
def _get_repo_format_to_test(self):
2296
def is_compatible(source, target):
2297
"""Be compatible with Knit1 source and Knit3 target"""
2298
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2300
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit1, \
2301
RepositoryFormatKnit3
2302
return (isinstance(source._format, (RepositoryFormatKnit1)) and
2303
isinstance(target._format, (RepositoryFormatKnit3)))
2304
except AttributeError:
2308
def fetch(self, revision_id=None, pb=None):
2309
"""See InterRepository.fetch()."""
2310
from bzrlib.fetch import Knit1to2Fetcher
2311
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2312
self.source, self.source._format, self.target,
2313
self.target._format)
2314
# TODO: jam 20070210 This should be an assert, not a translate
2315
revision_id = osutils.safe_revision_id(revision_id)
2316
f = Knit1to2Fetcher(to_repository=self.target,
2317
from_repository=self.source,
2318
last_revision=revision_id,
2320
return f.count_copied, f.failed_revisions
2323
class InterRemoteRepository(InterRepository):
2324
"""Code for converting between RemoteRepository objects.
2326
This just gets an non-remote repository from the RemoteRepository, and calls
2327
InterRepository.get again.
2330
def __init__(self, source, target):
2331
if isinstance(source, remote.RemoteRepository):
2332
source._ensure_real()
2333
real_source = source._real_repository
2335
real_source = source
2336
if isinstance(target, remote.RemoteRepository):
2337
target._ensure_real()
2338
real_target = target._real_repository
2340
real_target = target
2341
self.real_inter = InterRepository.get(real_source, real_target)
2344
def is_compatible(source, target):
2345
if isinstance(source, remote.RemoteRepository):
2347
if isinstance(target, remote.RemoteRepository):
2351
def copy_content(self, revision_id=None):
2352
self.real_inter.copy_content(revision_id=revision_id)
2354
def fetch(self, revision_id=None, pb=None):
2355
self.real_inter.fetch(revision_id=revision_id, pb=pb)
2358
def _get_repo_format_to_test(self):
2362
InterRepository.register_optimiser(InterSameDataRepository)
2363
InterRepository.register_optimiser(InterWeaveRepo)
2364
InterRepository.register_optimiser(InterKnitRepo)
2365
InterRepository.register_optimiser(InterModel1and2)
2366
InterRepository.register_optimiser(InterKnit1and2)
2367
InterRepository.register_optimiser(InterRemoteRepository)
2370
class CopyConverter(object):
2371
"""A repository conversion tool which just performs a copy of the content.
2373
This is slow but quite reliable.
2376
def __init__(self, target_format):
2377
"""Create a CopyConverter.
2379
:param target_format: The format the resulting repository should be.
2381
self.target_format = target_format
2383
def convert(self, repo, pb):
2384
"""Perform the conversion of to_convert, giving feedback via pb.
2386
:param to_convert: The disk object to convert.
2387
:param pb: a progress bar to use for progress information.
2392
# this is only useful with metadir layouts - separated repo content.
2393
# trigger an assertion if not such
2394
repo._format.get_format_string()
2395
self.repo_dir = repo.bzrdir
2396
self.step('Moving repository to repository.backup')
2397
self.repo_dir.transport.move('repository', 'repository.backup')
2398
backup_transport = self.repo_dir.transport.clone('repository.backup')
2399
repo._format.check_conversion_target(self.target_format)
2400
self.source_repo = repo._format.open(self.repo_dir,
2402
_override_transport=backup_transport)
2403
self.step('Creating new repository')
2404
converted = self.target_format.initialize(self.repo_dir,
2405
self.source_repo.is_shared())
2406
converted.lock_write()
2408
self.step('Copying content into repository.')
2409
self.source_repo.copy_content_into(converted)
2412
self.step('Deleting old repository content.')
2413
self.repo_dir.transport.delete_tree('repository.backup')
2414
self.pb.note('repository converted')
2416
def step(self, message):
2417
"""Update the pb by a step."""
2419
self.pb.update(message, self.count, self.total)
2431
def _unescaper(match, _map=_unescape_map):
2432
code = match.group(1)
2436
if not code.startswith('#'):
2438
return unichr(int(code[1:])).encode('utf8')
2444
def _unescape_xml(data):
2445
"""Unescape predefined XML entities in a string of data."""
2447
if _unescape_re is None:
2448
_unescape_re = re.compile('\&([^;]*);')
2449
return _unescape_re.sub(_unescaper, data)
2452
class _RevisionTextVersionCache(object):
2453
"""A cache of the versionedfile versions for revision and file-id"""
2455
def __init__(self, repository):
2456
self.repository = repository
2457
self.revision_versions = {}
2459
def add_revision_text_version(self, tree):
2460
"""Cache text version data from the supplied revision tree"""
2462
for path, entry in tree.iter_entries_by_dir():
2463
inv_revisions[entry.file_id] = entry.revision
2464
self.revision_versions[tree.get_revision_id()] = inv_revisions
2465
return inv_revisions
2467
def get_text_version(self, file_id, revision_id):
2468
"""Determine the text version for a given file-id and revision-id"""
2470
inv_revisions = self.revision_versions[revision_id]
2472
tree = self.repository.revision_tree(revision_id)
2473
inv_revisions = self.add_revision_text_version(tree)
2474
return inv_revisions.get(file_id)
2477
class WeaveChecker(object):
2479
def __init__(self, planned_revisions, revision_versions, repository):
2480
self.planned_revisions = planned_revisions
2481
self.revision_versions = revision_versions
2482
self.repository = repository
2484
def calculate_file_version_parents(self, revision_id, file_id):
2485
text_revision = self.revision_versions.get_text_version(
2486
file_id, revision_id)
2487
if text_revision is None:
2489
parents_of_text_revision = self.repository.get_parents(
2491
parents_from_inventories = []
2492
for parent in parents_of_text_revision:
2493
if parent == _mod_revision.NULL_REVISION:
2496
inventory = self.repository.get_inventory(parent)
2497
except errors.RevisionNotPresent:
2500
introduced_in = inventory[file_id].revision
2501
parents_from_inventories.append(introduced_in)
2502
mutter('%r:%r introduced in: %r',
2503
file_id, revision_id, parents_from_inventories)
2504
graph = self.repository.get_graph()
2505
heads = set(graph.heads(parents_from_inventories))
2506
mutter(' heads: %r', heads)
2508
for parent in parents_from_inventories:
2509
if parent in heads and parent not in new_parents:
2510
new_parents.append(parent)
2513
def check_file_version_parents(self, weave, file_id):
2515
for num, revision_id in enumerate(self.planned_revisions):
2516
correct_parents = self.calculate_file_version_parents(
2517
revision_id, file_id)
2518
if correct_parents is None:
2520
text_revision = self.revision_versions.get_text_version(
2521
file_id, revision_id)
2522
knit_parents = weave.get_parents(text_revision)
2523
if correct_parents != knit_parents:
2524
result[revision_id] = (knit_parents, correct_parents)
added
removed
bzrlib/repository.py
