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 = 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()
118
self._repo_graph = repository.get_graph()
120
def commit(self, message):
121
"""Make the actual commit.
123
:return: The revision id of the recorded revision.
125
rev = _mod_revision.Revision(
126
timestamp=self._timestamp,
127
timezone=self._timezone,
128
committer=self._committer,
130
inventory_sha1=self.inv_sha1,
131
revision_id=self._new_revision_id,
132
properties=self._revprops)
133
rev.parent_ids = self.parents
134
self.repository.add_revision(self._new_revision_id, rev,
135
self.new_inventory, self._config)
136
self.repository.commit_write_group()
137
return self._new_revision_id
140
"""Abort the commit that is being built.
142
self.repository.abort_write_group()
144
def revision_tree(self):
145
"""Return the tree that was just committed.
147
After calling commit() this can be called to get a RevisionTree
148
representing the newly committed tree. This is preferred to
149
calling Repository.revision_tree() because that may require
150
deserializing the inventory, while we already have a copy in
153
return RevisionTree(self.repository, self.new_inventory,
154
self._new_revision_id)
156
def finish_inventory(self):
157
"""Tell the builder that the inventory is finished."""
158
if self.new_inventory.root is None:
159
symbol_versioning.warn('Root entry should be supplied to'
160
' record_entry_contents, as of bzr 0.10.',
161
DeprecationWarning, stacklevel=2)
162
self.new_inventory.add(InventoryDirectory(ROOT_ID, '', None))
163
self.new_inventory.revision_id = self._new_revision_id
164
self.inv_sha1 = self.repository.add_inventory(
165
self._new_revision_id,
170
def _gen_revision_id(self):
171
"""Return new revision-id."""
172
return generate_ids.gen_revision_id(self._config.username(),
175
def _generate_revision_if_needed(self):
176
"""Create a revision id if None was supplied.
178
If the repository can not support user-specified revision ids
179
they should override this function and raise CannotSetRevisionId
180
if _new_revision_id is not None.
182
:raises: CannotSetRevisionId
184
if self._new_revision_id is None:
185
self._new_revision_id = self._gen_revision_id()
186
self.random_revid = True
188
self.random_revid = False
190
def _check_root(self, ie, parent_invs, tree):
191
"""Helper for record_entry_contents.
193
:param ie: An entry being added.
194
:param parent_invs: The inventories of the parent revisions of the
196
:param tree: The tree that is being committed.
198
# In this revision format, root entries have no knit or weave When
199
# serializing out to disk and back in root.revision is always
201
ie.revision = self._new_revision_id
203
def _get_delta(self, ie, basis_inv, path):
204
"""Get a delta against the basis inventory for ie."""
205
if ie.file_id not in basis_inv:
207
return (None, path, ie.file_id, ie)
208
elif ie != basis_inv[ie.file_id]:
210
# TODO: avoid tis id2path call.
211
return (basis_inv.id2path(ie.file_id), path, ie.file_id, ie)
216
def record_entry_contents(self, ie, parent_invs, path, tree,
218
"""Record the content of ie from tree into the commit if needed.
220
Side effect: sets ie.revision when unchanged
222
:param ie: An inventory entry present in the commit.
223
:param parent_invs: The inventories of the parent revisions of the
225
:param path: The path the entry is at in the tree.
226
:param tree: The tree which contains this entry and should be used to
228
:param content_summary: Summary data from the tree about the paths
229
content - stat, length, exec, sha/link target. This is only
230
accessed when the entry has a revision of None - that is when it is
231
a candidate to commit.
232
:return: A tuple (change_delta, version_recorded). change_delta is
233
an inventory_delta change for this entry against the basis tree of
234
the commit, or None if no change occured against the basis tree.
235
version_recorded is True if a new version of the entry has been
236
recorded. For instance, committing a merge where a file was only
237
changed on the other side will return (delta, False).
239
if self.new_inventory.root is None:
240
if ie.parent_id is not None:
241
raise errors.RootMissing()
242
self._check_root(ie, parent_invs, tree)
243
if ie.revision is None:
244
kind = content_summary[0]
246
# ie is carried over from a prior commit
248
# XXX: repository specific check for nested tree support goes here - if
249
# the repo doesn't want nested trees we skip it ?
250
if (kind == 'tree-reference' and
251
not self.repository._format.supports_tree_reference):
252
# mismatch between commit builder logic and repository:
253
# this needs the entry creation pushed down into the builder.
254
raise NotImplementedError('Missing repository subtree support.')
255
# transitional assert only, will remove before release.
256
assert ie.kind == kind
257
self.new_inventory.add(ie)
259
# TODO: slow, take it out of the inner loop.
261
basis_inv = parent_invs[0]
263
basis_inv = Inventory(root_id=None)
265
# ie.revision is always None if the InventoryEntry is considered
266
# for committing. We may record the previous parents revision if the
267
# content is actually unchanged against a sole head.
268
if ie.revision is not None:
269
if self._versioned_root or path != '':
270
# not considered for commit
273
# repositories that do not version the root set the root's
274
# revision to the new commit even when no change occurs, and
275
# this masks when a change may have occurred against the basis,
276
# so calculate if one happened.
277
if ie.file_id not in basis_inv:
279
delta = (None, path, ie.file_id, ie)
281
basis_id = basis_inv[ie.file_id]
282
if basis_id.name != '':
284
delta = (basis_inv.id2path(ie.file_id), path,
289
# not considered for commit, OR, for non-rich-root
290
return delta, ie.revision == self._new_revision_id and (path != '' or
291
self._versioned_root)
293
# XXX: Friction: parent_candidates should return a list not a dict
294
# so that we don't have to walk the inventories again.
295
parent_candiate_entries = ie.parent_candidates(parent_invs)
296
head_set = self._repo_graph.heads(parent_candiate_entries.keys())
298
for inv in parent_invs:
299
if ie.file_id in inv:
300
old_rev = inv[ie.file_id].revision
301
if old_rev in head_set:
302
heads.append(inv[ie.file_id].revision)
303
head_set.remove(inv[ie.file_id].revision)
306
# now we check to see if we need to write a new record to the
308
# We write a new entry unless there is one head to the ancestors, and
309
# the kind-derived content is unchanged.
311
# Cheapest check first: no ancestors, or more the one head in the
312
# ancestors, we write a new node.
316
# There is a single head, look it up for comparison
317
parent_entry = parent_candiate_entries[heads[0]]
318
# if the non-content specific data has changed, we'll be writing a
320
if (parent_entry.parent_id != ie.parent_id or
321
parent_entry.name != ie.name):
323
# now we need to do content specific checks:
325
# if the kind changed the content obviously has
326
if kind != parent_entry.kind:
330
if (# if the file length changed we have to store:
331
parent_entry.text_size != content_summary[1] or
332
# if the exec bit has changed we have to store:
333
parent_entry.executable != content_summary[2]):
335
elif parent_entry.text_sha1 == content_summary[3]:
336
# all meta and content is unchanged (using a hash cache
337
# hit to check the sha)
338
ie.revision = parent_entry.revision
339
ie.text_size = parent_entry.text_size
340
ie.text_sha1 = parent_entry.text_sha1
341
ie.executable = parent_entry.executable
342
return self._get_delta(ie, basis_inv, path), False
344
# Either there is only a hash change(no hash cache entry,
345
# or same size content change), or there is no change on
347
# Provide the parent's hash to the store layer, so that the
348
# content is unchanged we will not store a new node.
349
nostore_sha = parent_entry.text_sha1
351
# We want to record a new node regardless of the presence or
352
# absence of a content change in the file.
354
ie.executable = content_summary[2]
355
lines = tree.get_file(ie.file_id, path).readlines()
357
ie.text_sha1, ie.text_size = self._add_text_to_weave(
358
ie.file_id, lines, heads, nostore_sha)
359
except errors.ExistingContent:
360
# Turns out that the file content was unchanged, and we were
361
# only going to store a new node if it was changed. Carry over
363
ie.revision = parent_entry.revision
364
ie.text_size = parent_entry.text_size
365
ie.text_sha1 = parent_entry.text_sha1
366
ie.executable = parent_entry.executable
367
return self._get_delta(ie, basis_inv, path), False
368
elif kind == 'directory':
370
# all data is meta here, nothing specific to directory, so
372
ie.revision = parent_entry.revision
373
return self._get_delta(ie, basis_inv, path), False
375
self._add_text_to_weave(ie.file_id, lines, heads, None)
376
elif kind == 'symlink':
377
current_link_target = content_summary[3]
379
# symlink target is not generic metadata, check if it has
381
if current_link_target != parent_entry.symlink_target:
384
# unchanged, carry over.
385
ie.revision = parent_entry.revision
386
ie.symlink_target = parent_entry.symlink_target
387
return self._get_delta(ie, basis_inv, path), False
388
ie.symlink_target = current_link_target
390
self._add_text_to_weave(ie.file_id, lines, heads, None)
391
elif kind == 'tree-reference':
393
if content_summary[3] != parent_entry.reference_revision:
396
# unchanged, carry over.
397
ie.reference_revision = parent_entry.reference_revision
398
ie.revision = parent_entry.revision
399
return self._get_delta(ie, basis_inv, path), False
400
ie.reference_revision = content_summary[3]
402
self._add_text_to_weave(ie.file_id, lines, heads, None)
404
raise NotImplementedError('unknown kind')
405
ie.revision = self._new_revision_id
406
return self._get_delta(ie, basis_inv, path), True
408
def _add_text_to_weave(self, file_id, new_lines, parents, nostore_sha):
409
versionedfile = self.repository.weave_store.get_weave_or_empty(
410
file_id, self.repository.get_transaction())
411
# Don't change this to add_lines - add_lines_with_ghosts is cheaper
412
# than add_lines, and allows committing when a parent is ghosted for
414
# Note: as we read the content directly from the tree, we know its not
415
# been turned into unicode or badly split - but a broken tree
416
# implementation could give us bad output from readlines() so this is
417
# not a guarantee of safety. What would be better is always checking
418
# the content during test suite execution. RBC 20070912
420
return versionedfile.add_lines_with_ghosts(
421
self._new_revision_id, parents, new_lines,
422
nostore_sha=nostore_sha, random_id=self.random_revid,
423
check_content=False)[0:2]
425
versionedfile.clear_cache()
428
class RootCommitBuilder(CommitBuilder):
429
"""This commitbuilder actually records the root id"""
431
# the root entry gets versioned properly by this builder.
432
_versioned_root = True
434
def _check_root(self, ie, parent_invs, tree):
435
"""Helper for record_entry_contents.
437
:param ie: An entry being added.
438
:param parent_invs: The inventories of the parent revisions of the
440
:param tree: The tree that is being committed.
444
######################################################################
447
class Repository(object):
448
"""Repository holding history for one or more branches.
450
The repository holds and retrieves historical information including
451
revisions and file history. It's normally accessed only by the Branch,
452
which views a particular line of development through that history.
454
The Repository builds on top of Stores and a Transport, which respectively
455
describe the disk data format and the way of accessing the (possibly
459
# What class to use for a CommitBuilder. Often its simpler to change this
460
# in a Repository class subclass rather than to override
461
# get_commit_builder.
462
_commit_builder_class = CommitBuilder
463
# The search regex used by xml based repositories to determine what things
464
# where changed in a single commit.
465
_file_ids_altered_regex = lazy_regex.lazy_compile(
466
r'file_id="(?P<file_id>[^"]+)"'
467
r'.* revision="(?P<revision_id>[^"]+)"'
470
def abort_write_group(self):
471
"""Commit the contents accrued within the current write group.
473
:seealso: start_write_group.
475
if self._write_group is not self.get_transaction():
476
# has an unlock or relock occured ?
477
raise errors.BzrError('mismatched lock context and write group.')
478
self._abort_write_group()
479
self._write_group = None
481
def _abort_write_group(self):
482
"""Template method for per-repository write group cleanup.
484
This is called during abort before the write group is considered to be
485
finished and should cleanup any internal state accrued during the write
486
group. There is no requirement that data handed to the repository be
487
*not* made available - this is not a rollback - but neither should any
488
attempt be made to ensure that data added is fully commited. Abort is
489
invoked when an error has occured so futher disk or network operations
490
may not be possible or may error and if possible should not be
495
def add_inventory(self, revision_id, inv, parents):
496
"""Add the inventory inv to the repository as revision_id.
498
:param parents: The revision ids of the parents that revision_id
499
is known to have and are in the repository already.
501
returns the sha1 of the serialized inventory.
503
assert self.is_in_write_group()
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
# not right yet - should be more semantically clear ?
626
self.control_store = control_store
627
self.control_weaves = control_store
628
# TODO: make sure to construct the right store classes, etc, depending
629
# on whether escaping is required.
630
self._warn_if_deprecated()
631
self._write_group = None
632
self.base = control_files._transport.base
635
return '%s(%r)' % (self.__class__.__name__,
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
return InterRepository.get(other, self).missing_revision_ids(revision_id)
771
"""Open the repository rooted at base.
773
For instance, if the repository is at URL/.bzr/repository,
774
Repository.open(URL) -> a Repository instance.
776
control = bzrdir.BzrDir.open(base)
777
return control.open_repository()
779
def copy_content_into(self, destination, revision_id=None):
780
"""Make a complete copy of the content in self into destination.
782
This is a destructive operation! Do not use it on existing
785
return InterRepository.get(self, destination).copy_content(revision_id)
787
def commit_write_group(self):
788
"""Commit the contents accrued within the current write group.
790
:seealso: start_write_group.
792
if self._write_group is not self.get_transaction():
793
# has an unlock or relock occured ?
794
raise errors.BzrError('mismatched lock context %r and '
796
(self.get_transaction(), self._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
# fast path same-url fetch operations
815
if self.has_same_location(source):
816
# check that last_revision is in 'from' and then return a
818
if (revision_id is not None and
819
not _mod_revision.is_null(revision_id)):
820
self.get_revision(revision_id)
822
inter = InterRepository.get(source, self)
824
return inter.fetch(revision_id=revision_id, pb=pb)
825
except NotImplementedError:
826
raise errors.IncompatibleRepositories(source, self)
828
def create_bundle(self, target, base, fileobj, format=None):
829
return serializer.write_bundle(self, target, base, fileobj, format)
831
def get_commit_builder(self, branch, parents, config, timestamp=None,
832
timezone=None, committer=None, revprops=None,
834
"""Obtain a CommitBuilder for this repository.
836
:param branch: Branch to commit to.
837
:param parents: Revision ids of the parents of the new revision.
838
:param config: Configuration to use.
839
:param timestamp: Optional timestamp recorded for commit.
840
:param timezone: Optional timezone for timestamp.
841
:param committer: Optional committer to set for commit.
842
:param revprops: Optional dictionary of revision properties.
843
:param revision_id: Optional revision id.
845
result = self._commit_builder_class(self, parents, config,
846
timestamp, timezone, committer, revprops, revision_id)
847
self.start_write_group()
851
if (self.control_files._lock_count == 1 and
852
self.control_files._lock_mode == 'w'):
853
if self._write_group is not None:
854
raise errors.BzrError(
855
'Must end write groups before releasing write locks.')
856
self.control_files.unlock()
859
def clone(self, a_bzrdir, revision_id=None):
860
"""Clone this repository into a_bzrdir using the current format.
862
Currently no check is made that the format of this repository and
863
the bzrdir format are compatible. FIXME RBC 20060201.
865
:return: The newly created destination repository.
867
# TODO: deprecate after 0.16; cloning this with all its settings is
868
# probably not very useful -- mbp 20070423
869
dest_repo = self._create_sprouting_repo(a_bzrdir, shared=self.is_shared())
870
self.copy_content_into(dest_repo, revision_id)
873
def start_write_group(self):
874
"""Start a write group in the repository.
876
Write groups are used by repositories which do not have a 1:1 mapping
877
between file ids and backend store to manage the insertion of data from
878
both fetch and commit operations.
880
A write lock is required around the start_write_group/commit_write_group
881
for the support of lock-requiring repository formats.
883
One can only insert data into a repository inside a write group.
887
if not self.is_locked() or self.control_files._lock_mode != 'w':
888
raise errors.NotWriteLocked(self)
889
if self._write_group:
890
raise errors.BzrError('already in a write group')
891
self._start_write_group()
892
# so we can detect unlock/relock - the write group is now entered.
893
self._write_group = self.get_transaction()
895
def _start_write_group(self):
896
"""Template method for per-repository write group startup.
898
This is called before the write group is considered to be
903
def sprout(self, to_bzrdir, revision_id=None):
904
"""Create a descendent repository for new development.
906
Unlike clone, this does not copy the settings of the repository.
908
dest_repo = self._create_sprouting_repo(to_bzrdir, shared=False)
909
dest_repo.fetch(self, revision_id=revision_id)
912
def _create_sprouting_repo(self, a_bzrdir, shared):
913
if not isinstance(a_bzrdir._format, self.bzrdir._format.__class__):
914
# use target default format.
915
dest_repo = a_bzrdir.create_repository()
917
# Most control formats need the repository to be specifically
918
# created, but on some old all-in-one formats it's not needed
920
dest_repo = self._format.initialize(a_bzrdir, shared=shared)
921
except errors.UninitializableFormat:
922
dest_repo = a_bzrdir.open_repository()
926
def has_revision(self, revision_id):
927
"""True if this repository has a copy of the revision."""
928
if 'evil' in debug.debug_flags:
929
mutter_callsite(3, "has_revision is a LBYL symptom.")
930
return self._revision_store.has_revision_id(revision_id,
931
self.get_transaction())
934
def get_revision(self, revision_id):
935
"""Return the Revision object for a named revision."""
936
return self.get_revisions([revision_id])[0]
939
def get_revision_reconcile(self, revision_id):
940
"""'reconcile' helper routine that allows access to a revision always.
942
This variant of get_revision does not cross check the weave graph
943
against the revision one as get_revision does: but it should only
944
be used by reconcile, or reconcile-alike commands that are correcting
945
or testing the revision graph.
947
return self._get_revisions([revision_id])[0]
950
def get_revisions(self, revision_ids):
951
"""Get many revisions at once."""
952
return self._get_revisions(revision_ids)
955
def _get_revisions(self, revision_ids):
956
"""Core work logic to get many revisions without sanity checks."""
957
for rev_id in revision_ids:
958
if not rev_id or not isinstance(rev_id, basestring):
959
raise errors.InvalidRevisionId(revision_id=rev_id, branch=self)
960
revs = self._revision_store.get_revisions(revision_ids,
961
self.get_transaction())
963
assert not isinstance(rev.revision_id, unicode)
964
for parent_id in rev.parent_ids:
965
assert not isinstance(parent_id, unicode)
969
def get_revision_xml(self, revision_id):
970
# TODO: jam 20070210 This shouldn't be necessary since get_revision
971
# would have already do it.
972
# TODO: jam 20070210 Just use _serializer.write_revision_to_string()
973
rev = self.get_revision(revision_id)
975
# the current serializer..
976
self._revision_store._serializer.write_revision(rev, rev_tmp)
978
return rev_tmp.getvalue()
981
def get_deltas_for_revisions(self, revisions):
982
"""Produce a generator of revision deltas.
984
Note that the input is a sequence of REVISIONS, not revision_ids.
985
Trees will be held in memory until the generator exits.
986
Each delta is relative to the revision's lefthand predecessor.
988
required_trees = set()
989
for revision in revisions:
990
required_trees.add(revision.revision_id)
991
required_trees.update(revision.parent_ids[:1])
992
trees = dict((t.get_revision_id(), t) for
993
t in self.revision_trees(required_trees))
994
for revision in revisions:
995
if not revision.parent_ids:
996
old_tree = self.revision_tree(None)
998
old_tree = trees[revision.parent_ids[0]]
999
yield trees[revision.revision_id].changes_from(old_tree)
1002
def get_revision_delta(self, revision_id):
1003
"""Return the delta for one revision.
1005
The delta is relative to the left-hand predecessor of the
1008
r = self.get_revision(revision_id)
1009
return list(self.get_deltas_for_revisions([r]))[0]
1012
def store_revision_signature(self, gpg_strategy, plaintext, revision_id):
1013
signature = gpg_strategy.sign(plaintext)
1014
self._revision_store.add_revision_signature_text(revision_id,
1016
self.get_transaction())
1018
def _find_file_ids_from_xml_inventory_lines(self, line_iterator,
1020
"""Helper routine for fileids_altered_by_revision_ids.
1022
This performs the translation of xml lines to revision ids.
1024
:param line_iterator: An iterator of lines
1025
:param revision_ids: The revision ids to filter for.
1026
:return: a dictionary mapping altered file-ids to an iterable of
1027
revision_ids. Each altered file-ids has the exact revision_ids that
1028
altered it listed explicitly.
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
for line in line_iterator:
1052
match = search(line)
1055
# One call to match.group() returning multiple items is quite a
1056
# bit faster than 2 calls to match.group() each returning 1
1057
file_id, revision_id = match.group('file_id', 'revision_id')
1059
# Inlining the cache lookups helps a lot when you make 170,000
1060
# lines and 350k ids, versus 8.4 unique ids.
1061
# Using a cache helps in 2 ways:
1062
# 1) Avoids unnecessary decoding calls
1063
# 2) Re-uses cached strings, which helps in future set and
1065
# (2) is enough that removing encoding entirely along with
1066
# the cache (so we are using plain strings) results in no
1067
# performance improvement.
1069
revision_id = unescape_revid_cache[revision_id]
1071
unescaped = unescape(revision_id)
1072
unescape_revid_cache[revision_id] = unescaped
1073
revision_id = unescaped
1075
if revision_id in revision_ids:
1077
file_id = unescape_fileid_cache[file_id]
1079
unescaped = unescape(file_id)
1080
unescape_fileid_cache[file_id] = unescaped
1082
setdefault(file_id, set()).add(revision_id)
1085
def fileids_altered_by_revision_ids(self, revision_ids):
1086
"""Find the file ids and versions affected by revisions.
1088
:param revisions: an iterable containing revision ids.
1089
:return: a dictionary mapping altered file-ids to an iterable of
1090
revision_ids. Each altered file-ids has the exact revision_ids that
1091
altered it listed explicitly.
1093
assert self._serializer.support_altered_by_hack, \
1094
("fileids_altered_by_revision_ids only supported for branches "
1095
"which store inventory as unnested xml, not on %r" % self)
1096
selected_revision_ids = set(revision_ids)
1097
w = self.get_inventory_weave()
1098
pb = ui.ui_factory.nested_progress_bar()
1100
return self._find_file_ids_from_xml_inventory_lines(
1101
w.iter_lines_added_or_present_in_versions(
1102
selected_revision_ids, pb=pb),
1103
selected_revision_ids)
1107
def iter_files_bytes(self, desired_files):
1108
"""Iterate through file versions.
1110
Files will not necessarily be returned in the order they occur in
1111
desired_files. No specific order is guaranteed.
1113
Yields pairs of identifier, bytes_iterator. identifier is an opaque
1114
value supplied by the caller as part of desired_files. It should
1115
uniquely identify the file version in the caller's context. (Examples:
1116
an index number or a TreeTransform trans_id.)
1118
bytes_iterator is an iterable of bytestrings for the file. The
1119
kind of iterable and length of the bytestrings are unspecified, but for
1120
this implementation, it is a list of lines produced by
1121
VersionedFile.get_lines().
1123
:param desired_files: a list of (file_id, revision_id, identifier)
1126
transaction = self.get_transaction()
1127
for file_id, revision_id, callable_data in desired_files:
1129
weave = self.weave_store.get_weave(file_id, transaction)
1130
except errors.NoSuchFile:
1131
raise errors.NoSuchIdInRepository(self, file_id)
1132
yield callable_data, weave.get_lines(revision_id)
1134
def item_keys_introduced_by(self, revision_ids, _files_pb=None):
1135
"""Get an iterable listing the keys of all the data introduced by a set
1138
The keys will be ordered so that the corresponding items can be safely
1139
fetched and inserted in that order.
1141
:returns: An iterable producing tuples of (knit-kind, file-id,
1142
versions). knit-kind is one of 'file', 'inventory', 'signatures',
1143
'revisions'. file-id is None unless knit-kind is 'file'.
1145
# XXX: it's a bit weird to control the inventory weave caching in this
1146
# generator. Ideally the caching would be done in fetch.py I think. Or
1147
# maybe this generator should explicitly have the contract that it
1148
# should not be iterated until the previously yielded item has been
1150
inv_w = self.get_inventory_weave()
1151
inv_w.enable_cache()
1153
# file ids that changed
1154
file_ids = self.fileids_altered_by_revision_ids(revision_ids)
1156
num_file_ids = len(file_ids)
1157
for file_id, altered_versions in file_ids.iteritems():
1158
if _files_pb is not None:
1159
_files_pb.update("fetch texts", count, num_file_ids)
1161
yield ("file", file_id, altered_versions)
1162
# We're done with the files_pb. Note that it finished by the caller,
1163
# just as it was created by the caller.
1167
yield ("inventory", None, revision_ids)
1171
revisions_with_signatures = set()
1172
for rev_id in revision_ids:
1174
self.get_signature_text(rev_id)
1175
except errors.NoSuchRevision:
1179
revisions_with_signatures.add(rev_id)
1180
yield ("signatures", None, revisions_with_signatures)
1183
yield ("revisions", None, revision_ids)
1186
def get_inventory_weave(self):
1187
return self.control_weaves.get_weave('inventory',
1188
self.get_transaction())
1191
def get_inventory(self, revision_id):
1192
"""Get Inventory object by hash."""
1193
return self.deserialise_inventory(
1194
revision_id, self.get_inventory_xml(revision_id))
1196
def deserialise_inventory(self, revision_id, xml):
1197
"""Transform the xml into an inventory object.
1199
:param revision_id: The expected revision id of the inventory.
1200
:param xml: A serialised inventory.
1202
return self._serializer.read_inventory_from_string(xml, revision_id)
1204
def serialise_inventory(self, inv):
1205
return self._serializer.write_inventory_to_string(inv)
1207
def _serialise_inventory_to_lines(self, inv):
1208
return self._serializer.write_inventory_to_lines(inv)
1210
def get_serializer_format(self):
1211
return self._serializer.format_num
1214
def get_inventory_xml(self, revision_id):
1215
"""Get inventory XML as a file object."""
1217
assert isinstance(revision_id, str), type(revision_id)
1218
iw = self.get_inventory_weave()
1219
return iw.get_text(revision_id)
1221
raise errors.HistoryMissing(self, 'inventory', revision_id)
1224
def get_inventory_sha1(self, revision_id):
1225
"""Return the sha1 hash of the inventory entry
1227
return self.get_revision(revision_id).inventory_sha1
1230
def get_revision_graph(self, revision_id=None):
1231
"""Return a dictionary containing the revision graph.
1233
NB: This method should not be used as it accesses the entire graph all
1234
at once, which is much more data than most operations should require.
1236
:param revision_id: The revision_id to get a graph from. If None, then
1237
the entire revision graph is returned. This is a deprecated mode of
1238
operation and will be removed in the future.
1239
:return: a dictionary of revision_id->revision_parents_list.
1241
raise NotImplementedError(self.get_revision_graph)
1244
def get_revision_graph_with_ghosts(self, revision_ids=None):
1245
"""Return a graph of the revisions with ghosts marked as applicable.
1247
:param revision_ids: an iterable of revisions to graph or None for all.
1248
:return: a Graph object with the graph reachable from revision_ids.
1250
if 'evil' in debug.debug_flags:
1252
"get_revision_graph_with_ghosts scales with size of history.")
1253
result = deprecated_graph.Graph()
1254
if not revision_ids:
1255
pending = set(self.all_revision_ids())
1258
pending = set(revision_ids)
1259
# special case NULL_REVISION
1260
if _mod_revision.NULL_REVISION in pending:
1261
pending.remove(_mod_revision.NULL_REVISION)
1262
required = set(pending)
1265
revision_id = pending.pop()
1267
rev = self.get_revision(revision_id)
1268
except errors.NoSuchRevision:
1269
if revision_id in required:
1272
result.add_ghost(revision_id)
1274
for parent_id in rev.parent_ids:
1275
# is this queued or done ?
1276
if (parent_id not in pending and
1277
parent_id not in done):
1279
pending.add(parent_id)
1280
result.add_node(revision_id, rev.parent_ids)
1281
done.add(revision_id)
1284
def _get_history_vf(self):
1285
"""Get a versionedfile whose history graph reflects all revisions.
1287
For weave repositories, this is the inventory weave.
1289
return self.get_inventory_weave()
1291
def iter_reverse_revision_history(self, revision_id):
1292
"""Iterate backwards through revision ids in the lefthand history
1294
:param revision_id: The revision id to start with. All its lefthand
1295
ancestors will be traversed.
1297
if revision_id in (None, _mod_revision.NULL_REVISION):
1299
next_id = revision_id
1300
versionedfile = self._get_history_vf()
1303
parents = versionedfile.get_parents(next_id)
1304
if len(parents) == 0:
1307
next_id = parents[0]
1310
def get_revision_inventory(self, revision_id):
1311
"""Return inventory of a past revision."""
1312
# TODO: Unify this with get_inventory()
1313
# bzr 0.0.6 and later imposes the constraint that the inventory_id
1314
# must be the same as its revision, so this is trivial.
1315
if revision_id is None:
1316
# This does not make sense: if there is no revision,
1317
# then it is the current tree inventory surely ?!
1318
# and thus get_root_id() is something that looks at the last
1319
# commit on the branch, and the get_root_id is an inventory check.
1320
raise NotImplementedError
1321
# return Inventory(self.get_root_id())
1323
return self.get_inventory(revision_id)
1326
def is_shared(self):
1327
"""Return True if this repository is flagged as a shared repository."""
1328
raise NotImplementedError(self.is_shared)
1331
def reconcile(self, other=None, thorough=False):
1332
"""Reconcile this repository."""
1333
from bzrlib.reconcile import RepoReconciler
1334
reconciler = RepoReconciler(self, thorough=thorough)
1335
reconciler.reconcile()
1338
def _refresh_data(self):
1339
"""Helper called from lock_* to ensure coherency with disk.
1341
The default implementation does nothing; it is however possible
1342
for repositories to maintain loaded indices across multiple locks
1343
by checking inside their implementation of this method to see
1344
whether their indices are still valid. This depends of course on
1345
the disk format being validatable in this manner.
1349
def revision_tree(self, revision_id):
1350
"""Return Tree for a revision on this branch.
1352
`revision_id` may be None for the empty tree revision.
1354
# TODO: refactor this to use an existing revision object
1355
# so we don't need to read it in twice.
1356
if revision_id is None or revision_id == _mod_revision.NULL_REVISION:
1357
return RevisionTree(self, Inventory(root_id=None),
1358
_mod_revision.NULL_REVISION)
1360
inv = self.get_revision_inventory(revision_id)
1361
return RevisionTree(self, inv, revision_id)
1364
def revision_trees(self, revision_ids):
1365
"""Return Tree for a revision on this branch.
1367
`revision_id` may not be None or 'null:'"""
1368
assert None not in revision_ids
1369
assert _mod_revision.NULL_REVISION not in revision_ids
1370
texts = self.get_inventory_weave().get_texts(revision_ids)
1371
for text, revision_id in zip(texts, revision_ids):
1372
inv = self.deserialise_inventory(revision_id, text)
1373
yield RevisionTree(self, inv, revision_id)
1376
def get_ancestry(self, revision_id, topo_sorted=True):
1377
"""Return a list of revision-ids integrated by a revision.
1379
The first element of the list is always None, indicating the origin
1380
revision. This might change when we have history horizons, or
1381
perhaps we should have a new API.
1383
This is topologically sorted.
1385
if _mod_revision.is_null(revision_id):
1387
if not self.has_revision(revision_id):
1388
raise errors.NoSuchRevision(self, revision_id)
1389
w = self.get_inventory_weave()
1390
candidates = w.get_ancestry(revision_id, topo_sorted)
1391
return [None] + candidates # self._eliminate_revisions_not_present(candidates)
1394
"""Compress the data within the repository.
1396
This operation only makes sense for some repository types. For other
1397
types it should be a no-op that just returns.
1399
This stub method does not require a lock, but subclasses should use
1400
@needs_write_lock as this is a long running call its reasonable to
1401
implicitly lock for the user.
1405
def print_file(self, file, revision_id):
1406
"""Print `file` to stdout.
1408
FIXME RBC 20060125 as John Meinel points out this is a bad api
1409
- it writes to stdout, it assumes that that is valid etc. Fix
1410
by creating a new more flexible convenience function.
1412
tree = self.revision_tree(revision_id)
1413
# use inventory as it was in that revision
1414
file_id = tree.inventory.path2id(file)
1416
# TODO: jam 20060427 Write a test for this code path
1417
# it had a bug in it, and was raising the wrong
1419
raise errors.BzrError("%r is not present in revision %s" % (file, revision_id))
1420
tree.print_file(file_id)
1422
def get_transaction(self):
1423
return self.control_files.get_transaction()
1425
def revision_parents(self, revision_id):
1426
return self.get_inventory_weave().parent_names(revision_id)
1428
def get_parents(self, revision_ids):
1429
"""See StackedParentsProvider.get_parents"""
1431
for revision_id in revision_ids:
1432
if revision_id == _mod_revision.NULL_REVISION:
1436
parents = self.get_revision(revision_id).parent_ids
1437
except errors.NoSuchRevision:
1440
if len(parents) == 0:
1441
parents = [_mod_revision.NULL_REVISION]
1442
parents_list.append(parents)
1445
def _make_parents_provider(self):
1448
def get_graph(self, other_repository=None):
1449
"""Return the graph walker for this repository format"""
1450
parents_provider = self._make_parents_provider()
1451
if (other_repository is not None and
1452
other_repository.bzrdir.transport.base !=
1453
self.bzrdir.transport.base):
1454
parents_provider = graph._StackedParentsProvider(
1455
[parents_provider, other_repository._make_parents_provider()])
1456
return graph.Graph(parents_provider)
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
plaintext = Testament.from_revision(self, revision_id).as_short_text()
1477
self.store_revision_signature(gpg_strategy, plaintext, revision_id)
1480
def has_signature_for_revision_id(self, revision_id):
1481
"""Query for a revision signature for revision_id in the repository."""
1482
return self._revision_store.has_signature(revision_id,
1483
self.get_transaction())
1486
def get_signature_text(self, revision_id):
1487
"""Return the text for a signature."""
1488
return self._revision_store.get_signature_text(revision_id,
1489
self.get_transaction())
1492
def check(self, revision_ids):
1493
"""Check consistency of all history of given revision_ids.
1495
Different repository implementations should override _check().
1497
:param revision_ids: A non-empty list of revision_ids whose ancestry
1498
will be checked. Typically the last revision_id of a branch.
1500
if not revision_ids:
1501
raise ValueError("revision_ids must be non-empty in %s.check"
1503
return self._check(revision_ids)
1505
def _check(self, revision_ids):
1506
result = check.Check(self)
1510
def _warn_if_deprecated(self):
1511
global _deprecation_warning_done
1512
if _deprecation_warning_done:
1514
_deprecation_warning_done = True
1515
warning("Format %s for %s is deprecated - please use 'bzr upgrade' to get better performance"
1516
% (self._format, self.bzrdir.transport.base))
1518
def supports_rich_root(self):
1519
return self._format.rich_root_data
1521
def _check_ascii_revisionid(self, revision_id, method):
1522
"""Private helper for ascii-only repositories."""
1523
# weave repositories refuse to store revisionids that are non-ascii.
1524
if revision_id is not None:
1525
# weaves require ascii revision ids.
1526
if isinstance(revision_id, unicode):
1528
revision_id.encode('ascii')
1529
except UnicodeEncodeError:
1530
raise errors.NonAsciiRevisionId(method, self)
1533
revision_id.decode('ascii')
1534
except UnicodeDecodeError:
1535
raise errors.NonAsciiRevisionId(method, self)
1539
# remove these delegates a while after bzr 0.15
1540
def __make_delegated(name, from_module):
1541
def _deprecated_repository_forwarder():
1542
symbol_versioning.warn('%s moved to %s in bzr 0.15'
1543
% (name, from_module),
1546
m = __import__(from_module, globals(), locals(), [name])
1548
return getattr(m, name)
1549
except AttributeError:
1550
raise AttributeError('module %s has no name %s'
1552
globals()[name] = _deprecated_repository_forwarder
1555
'AllInOneRepository',
1556
'WeaveMetaDirRepository',
1557
'PreSplitOutRepositoryFormat',
1558
'RepositoryFormat4',
1559
'RepositoryFormat5',
1560
'RepositoryFormat6',
1561
'RepositoryFormat7',
1563
__make_delegated(_name, 'bzrlib.repofmt.weaverepo')
1567
'RepositoryFormatKnit',
1568
'RepositoryFormatKnit1',
1570
__make_delegated(_name, 'bzrlib.repofmt.knitrepo')
1573
def install_revision(repository, rev, revision_tree):
1574
"""Install all revision data into a repository."""
1575
repository.start_write_group()
1577
_install_revision(repository, rev, revision_tree)
1579
repository.abort_write_group()
1582
repository.commit_write_group()
1585
def _install_revision(repository, rev, revision_tree):
1586
"""Install all revision data into a repository."""
1587
present_parents = []
1589
for p_id in rev.parent_ids:
1590
if repository.has_revision(p_id):
1591
present_parents.append(p_id)
1592
parent_trees[p_id] = repository.revision_tree(p_id)
1594
parent_trees[p_id] = repository.revision_tree(None)
1596
inv = revision_tree.inventory
1597
entries = inv.iter_entries()
1598
# backwards compatibility hack: skip the root id.
1599
if not repository.supports_rich_root():
1600
path, root = entries.next()
1601
if root.revision != rev.revision_id:
1602
raise errors.IncompatibleRevision(repr(repository))
1603
# Add the texts that are not already present
1604
for path, ie in entries:
1605
w = repository.weave_store.get_weave_or_empty(ie.file_id,
1606
repository.get_transaction())
1607
if ie.revision not in w:
1609
# FIXME: TODO: The following loop *may* be overlapping/duplicate
1610
# with InventoryEntry.find_previous_heads(). if it is, then there
1611
# is a latent bug here where the parents may have ancestors of each
1613
for revision, tree in parent_trees.iteritems():
1614
if ie.file_id not in tree:
1616
parent_id = tree.inventory[ie.file_id].revision
1617
if parent_id in text_parents:
1619
text_parents.append(parent_id)
1621
vfile = repository.weave_store.get_weave_or_empty(ie.file_id,
1622
repository.get_transaction())
1623
lines = revision_tree.get_file(ie.file_id).readlines()
1624
vfile.add_lines(rev.revision_id, text_parents, lines)
1626
# install the inventory
1627
repository.add_inventory(rev.revision_id, inv, present_parents)
1628
except errors.RevisionAlreadyPresent:
1630
repository.add_revision(rev.revision_id, rev, inv)
1633
class MetaDirRepository(Repository):
1634
"""Repositories in the new meta-dir layout."""
1636
def __init__(self, _format, a_bzrdir, control_files, _revision_store, control_store, text_store):
1637
super(MetaDirRepository, self).__init__(_format,
1643
dir_mode = self.control_files._dir_mode
1644
file_mode = self.control_files._file_mode
1647
def is_shared(self):
1648
"""Return True if this repository is flagged as a shared repository."""
1649
return self.control_files._transport.has('shared-storage')
1652
def set_make_working_trees(self, new_value):
1653
"""Set the policy flag for making working trees when creating branches.
1655
This only applies to branches that use this repository.
1657
The default is 'True'.
1658
:param new_value: True to restore the default, False to disable making
1663
self.control_files._transport.delete('no-working-trees')
1664
except errors.NoSuchFile:
1667
self.control_files.put_utf8('no-working-trees', '')
1669
def make_working_trees(self):
1670
"""Returns the policy for making working trees on new branches."""
1671
return not self.control_files._transport.has('no-working-trees')
1674
class RepositoryFormatRegistry(registry.Registry):
1675
"""Registry of RepositoryFormats."""
1677
def get(self, format_string):
1678
r = registry.Registry.get(self, format_string)
1684
format_registry = RepositoryFormatRegistry()
1685
"""Registry of formats, indexed by their identifying format string.
1687
This can contain either format instances themselves, or classes/factories that
1688
can be called to obtain one.
1692
#####################################################################
1693
# Repository Formats
1695
class RepositoryFormat(object):
1696
"""A repository format.
1698
Formats provide three things:
1699
* An initialization routine to construct repository data on disk.
1700
* a format string which is used when the BzrDir supports versioned
1702
* an open routine which returns a Repository instance.
1704
There is one and only one Format subclass for each on-disk format. But
1705
there can be one Repository subclass that is used for several different
1706
formats. The _format attribute on a Repository instance can be used to
1707
determine the disk format.
1709
Formats are placed in an dict by their format string for reference
1710
during opening. These should be subclasses of RepositoryFormat
1713
Once a format is deprecated, just deprecate the initialize and open
1714
methods on the format class. Do not deprecate the object, as the
1715
object will be created every system load.
1717
Common instance attributes:
1718
_matchingbzrdir - the bzrdir format that the repository format was
1719
originally written to work with. This can be used if manually
1720
constructing a bzrdir and repository, or more commonly for test suite
1725
return "<%s>" % self.__class__.__name__
1727
def __eq__(self, other):
1728
# format objects are generally stateless
1729
return isinstance(other, self.__class__)
1731
def __ne__(self, other):
1732
return not self == other
1735
def find_format(klass, a_bzrdir):
1736
"""Return the format for the repository object in a_bzrdir.
1738
This is used by bzr native formats that have a "format" file in
1739
the repository. Other methods may be used by different types of
1743
transport = a_bzrdir.get_repository_transport(None)
1744
format_string = transport.get("format").read()
1745
return format_registry.get(format_string)
1746
except errors.NoSuchFile:
1747
raise errors.NoRepositoryPresent(a_bzrdir)
1749
raise errors.UnknownFormatError(format=format_string)
1752
def register_format(klass, format):
1753
format_registry.register(format.get_format_string(), format)
1756
def unregister_format(klass, format):
1757
format_registry.remove(format.get_format_string())
1760
def get_default_format(klass):
1761
"""Return the current default format."""
1762
from bzrlib import bzrdir
1763
return bzrdir.format_registry.make_bzrdir('default').repository_format
1765
def _get_control_store(self, repo_transport, control_files):
1766
"""Return the control store for this repository."""
1767
raise NotImplementedError(self._get_control_store)
1769
def get_format_string(self):
1770
"""Return the ASCII format string that identifies this format.
1772
Note that in pre format ?? repositories the format string is
1773
not permitted nor written to disk.
1775
raise NotImplementedError(self.get_format_string)
1777
def get_format_description(self):
1778
"""Return the short description for this format."""
1779
raise NotImplementedError(self.get_format_description)
1781
def _get_revision_store(self, repo_transport, control_files):
1782
"""Return the revision store object for this a_bzrdir."""
1783
raise NotImplementedError(self._get_revision_store)
1785
def _get_text_rev_store(self,
1792
"""Common logic for getting a revision store for a repository.
1794
see self._get_revision_store for the subclass-overridable method to
1795
get the store for a repository.
1797
from bzrlib.store.revision.text import TextRevisionStore
1798
dir_mode = control_files._dir_mode
1799
file_mode = control_files._file_mode
1800
text_store = TextStore(transport.clone(name),
1802
compressed=compressed,
1804
file_mode=file_mode)
1805
_revision_store = TextRevisionStore(text_store, serializer)
1806
return _revision_store
1808
# TODO: this shouldn't be in the base class, it's specific to things that
1809
# use weaves or knits -- mbp 20070207
1810
def _get_versioned_file_store(self,
1815
versionedfile_class=None,
1816
versionedfile_kwargs={},
1818
if versionedfile_class is None:
1819
versionedfile_class = self._versionedfile_class
1820
weave_transport = control_files._transport.clone(name)
1821
dir_mode = control_files._dir_mode
1822
file_mode = control_files._file_mode
1823
return VersionedFileStore(weave_transport, prefixed=prefixed,
1825
file_mode=file_mode,
1826
versionedfile_class=versionedfile_class,
1827
versionedfile_kwargs=versionedfile_kwargs,
1830
def initialize(self, a_bzrdir, shared=False):
1831
"""Initialize a repository of this format in a_bzrdir.
1833
:param a_bzrdir: The bzrdir to put the new repository in it.
1834
:param shared: The repository should be initialized as a sharable one.
1835
:returns: The new repository object.
1837
This may raise UninitializableFormat if shared repository are not
1838
compatible the a_bzrdir.
1840
raise NotImplementedError(self.initialize)
1842
def is_supported(self):
1843
"""Is this format supported?
1845
Supported formats must be initializable and openable.
1846
Unsupported formats may not support initialization or committing or
1847
some other features depending on the reason for not being supported.
1851
def check_conversion_target(self, target_format):
1852
raise NotImplementedError(self.check_conversion_target)
1854
def open(self, a_bzrdir, _found=False):
1855
"""Return an instance of this format for the bzrdir a_bzrdir.
1857
_found is a private parameter, do not use it.
1859
raise NotImplementedError(self.open)
1862
class MetaDirRepositoryFormat(RepositoryFormat):
1863
"""Common base class for the new repositories using the metadir layout."""
1865
rich_root_data = False
1866
supports_tree_reference = False
1867
_matchingbzrdir = bzrdir.BzrDirMetaFormat1()
1870
super(MetaDirRepositoryFormat, self).__init__()
1872
def _create_control_files(self, a_bzrdir):
1873
"""Create the required files and the initial control_files object."""
1874
# FIXME: RBC 20060125 don't peek under the covers
1875
# NB: no need to escape relative paths that are url safe.
1876
repository_transport = a_bzrdir.get_repository_transport(self)
1877
control_files = lockable_files.LockableFiles(repository_transport,
1878
'lock', lockdir.LockDir)
1879
control_files.create_lock()
1880
return control_files
1882
def _upload_blank_content(self, a_bzrdir, dirs, files, utf8_files, shared):
1883
"""Upload the initial blank content."""
1884
control_files = self._create_control_files(a_bzrdir)
1885
control_files.lock_write()
1887
control_files._transport.mkdir_multi(dirs,
1888
mode=control_files._dir_mode)
1889
for file, content in files:
1890
control_files.put(file, content)
1891
for file, content in utf8_files:
1892
control_files.put_utf8(file, content)
1894
control_files.put_utf8('shared-storage', '')
1896
control_files.unlock()
1899
# formats which have no format string are not discoverable
1900
# and not independently creatable, so are not registered. They're
1901
# all in bzrlib.repofmt.weaverepo now. When an instance of one of these is
1902
# needed, it's constructed directly by the BzrDir. Non-native formats where
1903
# the repository is not separately opened are similar.
1905
format_registry.register_lazy(
1906
'Bazaar-NG Repository format 7',
1907
'bzrlib.repofmt.weaverepo',
1911
# KEEP in sync with bzrdir.format_registry default, which controls the overall
1912
# default control directory format
1913
format_registry.register_lazy(
1914
'Bazaar-NG Knit Repository Format 1',
1915
'bzrlib.repofmt.knitrepo',
1916
'RepositoryFormatKnit1',
1918
format_registry.default_key = 'Bazaar-NG Knit Repository Format 1'
1920
format_registry.register_lazy(
1921
'Bazaar Knit Repository Format 3 (bzr 0.15)\n',
1922
'bzrlib.repofmt.knitrepo',
1923
'RepositoryFormatKnit3',
1926
# Experimental formats. These make no guarantee about data stability.
1927
# There is one format for pre-subtrees, and one for post-subtrees to
1928
# allow ease of testing.
1929
format_registry.register_lazy(
1930
'Bazaar Experimental no-subtrees\n',
1931
'bzrlib.repofmt.pack_repo',
1932
'RepositoryFormatGraphKnit1',
1934
format_registry.register_lazy(
1935
'Bazaar Experimental subtrees\n',
1936
'bzrlib.repofmt.pack_repo',
1937
'RepositoryFormatGraphKnit3',
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 InterPackRepo(InterSameDataRepository):
2251
"""Optimised code paths between Pack based repositories."""
2254
def _get_repo_format_to_test(self):
2255
from bzrlib.repofmt import pack_repo
2256
return pack_repo.RepositoryFormatGraphKnit1()
2259
def is_compatible(source, target):
2260
"""Be compatible with known Pack formats.
2262
We don't test for the stores being of specific types because that
2263
could lead to confusing results, and there is no need to be
2266
from bzrlib.repofmt.pack_repo import RepositoryFormatPack
2268
are_packs = (isinstance(source._format, RepositoryFormatPack) and
2269
isinstance(target._format, RepositoryFormatPack))
2270
except AttributeError:
2272
return are_packs and InterRepository._same_model(source, target)
2275
def fetch(self, revision_id=None, pb=None):
2276
"""See InterRepository.fetch()."""
2277
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2278
self.source, self.source._format, self.target, self.target._format)
2279
# TODO: jam 20070210 This should be an assert, not a translate
2280
revision_id = osutils.safe_revision_id(revision_id)
2281
self.count_copied = 0
2282
if revision_id is None:
2284
# everything to do - use pack logic
2285
# to fetch from all packs to one without
2286
# inventory parsing etc, IFF nothing to be copied is in the target.
2288
revision_ids = self.source.all_revision_ids()
2289
# implementing the TODO will involve:
2290
# - detecting when all of a pack is selected
2291
# - avoiding as much as possible pre-selection, so the
2292
# more-core routines such as create_pack_from_packs can filter in
2293
# a just-in-time fashion. (though having a HEADS list on a
2294
# repository might make this a lot easier, because we could
2295
# sensibly detect 'new revisions' without doing a full index scan.
2296
elif _mod_revision.is_null(revision_id):
2301
revision_ids = self.missing_revision_ids(revision_id)
2302
except errors.NoSuchRevision:
2303
raise errors.InstallFailed([revision_id])
2304
packs = self.source._packs.all_packs()
2305
pack = self.target._packs.create_pack_from_packs(
2306
packs, '.fetch', revision_ids,
2308
if pack is not None:
2309
self.target._packs._save_pack_names()
2310
self.target._packs.add_pack_to_memory(pack)
2311
# Trigger an autopack. This may duplicate effort as we've just done
2312
# a pack creation, but for now it is simpler to think about as
2313
# 'upload data, then repack if needed'.
2314
self.target._packs.autopack()
2315
return pack.get_revision_count()
2320
def missing_revision_ids(self, revision_id=None):
2321
"""See InterRepository.missing_revision_ids()."""
2322
if revision_id is not None:
2323
source_ids = self.source.get_ancestry(revision_id)
2324
assert source_ids[0] is None
2327
source_ids = self.source.all_revision_ids()
2328
source_ids_set = set(source_ids)
2329
# source_ids is the worst possible case we may need to pull.
2330
# now we want to filter source_ids against what we actually
2331
# have in target, but don't try to check for existence where we know
2332
# we do not have a revision as that would be pointless.
2333
target_ids = set(self.target.all_revision_ids())
2334
actually_present_revisions = target_ids.intersection(source_ids_set)
2335
required_revisions = source_ids_set.difference(actually_present_revisions)
2336
required_topo_revisions = [rev_id for rev_id in source_ids if rev_id in required_revisions]
2337
return required_topo_revisions
2340
class InterModel1and2(InterRepository):
2343
def _get_repo_format_to_test(self):
2347
def is_compatible(source, target):
2348
if not source.supports_rich_root() and target.supports_rich_root():
2354
def fetch(self, revision_id=None, pb=None):
2355
"""See InterRepository.fetch()."""
2356
from bzrlib.fetch import Model1toKnit2Fetcher
2357
f = Model1toKnit2Fetcher(to_repository=self.target,
2358
from_repository=self.source,
2359
last_revision=revision_id,
2361
return f.count_copied, f.failed_revisions
2364
def copy_content(self, revision_id=None):
2365
"""Make a complete copy of the content in self into destination.
2367
This is a destructive operation! Do not use it on existing
2370
:param revision_id: Only copy the content needed to construct
2371
revision_id and its parents.
2374
self.target.set_make_working_trees(self.source.make_working_trees())
2375
except NotImplementedError:
2377
# but don't bother fetching if we have the needed data now.
2378
if (revision_id not in (None, _mod_revision.NULL_REVISION) and
2379
self.target.has_revision(revision_id)):
2381
self.target.fetch(self.source, revision_id=revision_id)
2384
class InterKnit1and2(InterKnitRepo):
2387
def _get_repo_format_to_test(self):
2391
def is_compatible(source, target):
2392
"""Be compatible with Knit1 source and Knit3 target"""
2393
from bzrlib.repofmt.knitrepo import RepositoryFormatKnit3
2395
from bzrlib.repofmt.knitrepo import (RepositoryFormatKnit1,
2396
RepositoryFormatKnit3)
2397
from bzrlib.repofmt.pack_repo import (RepositoryFormatGraphKnit1,
2398
RepositoryFormatGraphKnit3)
2399
return (isinstance(source._format, RepositoryFormatKnit1) and
2400
isinstance(target._format, RepositoryFormatKnit3) or
2401
isinstance(source._format, (RepositoryFormatGraphKnit1)) and
2402
isinstance(target._format, (RepositoryFormatGraphKnit3))
2404
except AttributeError:
2408
def fetch(self, revision_id=None, pb=None):
2409
"""See InterRepository.fetch()."""
2410
from bzrlib.fetch import Knit1to2Fetcher
2411
mutter("Using fetch logic to copy between %s(%s) and %s(%s)",
2412
self.source, self.source._format, self.target,
2413
self.target._format)
2414
f = Knit1to2Fetcher(to_repository=self.target,
2415
from_repository=self.source,
2416
last_revision=revision_id,
2418
return f.count_copied, f.failed_revisions
2421
class InterRemoteRepository(InterRepository):
2422
"""Code for converting between RemoteRepository objects.
2424
This just gets an non-remote repository from the RemoteRepository, and calls
2425
InterRepository.get again.
2428
def __init__(self, source, target):
2429
if isinstance(source, remote.RemoteRepository):
2430
source._ensure_real()
2431
real_source = source._real_repository
2433
real_source = source
2434
if isinstance(target, remote.RemoteRepository):
2435
target._ensure_real()
2436
real_target = target._real_repository
2438
real_target = target
2439
self.real_inter = InterRepository.get(real_source, real_target)
2442
def is_compatible(source, target):
2443
if isinstance(source, remote.RemoteRepository):
2445
if isinstance(target, remote.RemoteRepository):
2449
def copy_content(self, revision_id=None):
2450
self.real_inter.copy_content(revision_id=revision_id)
2452
def fetch(self, revision_id=None, pb=None):
2453
self.real_inter.fetch(revision_id=revision_id, pb=pb)
2456
def _get_repo_format_to_test(self):
2460
InterRepository.register_optimiser(InterSameDataRepository)
2461
InterRepository.register_optimiser(InterWeaveRepo)
2462
InterRepository.register_optimiser(InterKnitRepo)
2463
InterRepository.register_optimiser(InterModel1and2)
2464
InterRepository.register_optimiser(InterKnit1and2)
2465
InterRepository.register_optimiser(InterPackRepo)
2466
InterRepository.register_optimiser(InterRemoteRepository)
2469
class CopyConverter(object):
2470
"""A repository conversion tool which just performs a copy of the content.
2472
This is slow but quite reliable.
2475
def __init__(self, target_format):
2476
"""Create a CopyConverter.
2478
:param target_format: The format the resulting repository should be.
2480
self.target_format = target_format
2482
def convert(self, repo, pb):
2483
"""Perform the conversion of to_convert, giving feedback via pb.
2485
:param to_convert: The disk object to convert.
2486
:param pb: a progress bar to use for progress information.
2491
# this is only useful with metadir layouts - separated repo content.
2492
# trigger an assertion if not such
2493
repo._format.get_format_string()
2494
self.repo_dir = repo.bzrdir
2495
self.step('Moving repository to repository.backup')
2496
self.repo_dir.transport.move('repository', 'repository.backup')
2497
backup_transport = self.repo_dir.transport.clone('repository.backup')
2498
repo._format.check_conversion_target(self.target_format)
2499
self.source_repo = repo._format.open(self.repo_dir,
2501
_override_transport=backup_transport)
2502
self.step('Creating new repository')
2503
converted = self.target_format.initialize(self.repo_dir,
2504
self.source_repo.is_shared())
2505
converted.lock_write()
2507
self.step('Copying content into repository.')
2508
self.source_repo.copy_content_into(converted)
2511
self.step('Deleting old repository content.')
2512
self.repo_dir.transport.delete_tree('repository.backup')
2513
self.pb.note('repository converted')
2515
def step(self, message):
2516
"""Update the pb by a step."""
2518
self.pb.update(message, self.count, self.total)
2530
def _unescaper(match, _map=_unescape_map):
2531
code = match.group(1)
2535
if not code.startswith('#'):
2537
return unichr(int(code[1:])).encode('utf8')
2543
def _unescape_xml(data):
2544
"""Unescape predefined XML entities in a string of data."""
2546
if _unescape_re is None:
2547
_unescape_re = re.compile('\&([^;]*);')
2548
return _unescape_re.sub(_unescaper, data)
added
removed
bzrlib/repository.py
